Yarmuth Wilsdon - Investigative Report Re WA DOC Early Release of Offenders Vol 1, 2016
Download original document:
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
INVESTIGATIVE REPORT RE: DEPARTMENT OF CORRECTIONS – EARLY RELEASE OF OFFENDERS VOLUME 1 – REPORT Department of Corrections – Early Release of Offenders TABLE OF CONTENTS I. INTRODUCTION .................................................................................................................1 II. EXECUTIVE SUMMARY ...................................................................................................2 III. IV. A. Synopsis of Factual Findings .....................................................................................2 B. Summary of Conclusions ...........................................................................................4 C. Summary of Recommendations .................................................................................6 BACKGROUND ...................................................................................................................8 A. DOC Organizational Structure ...................................................................................8 B. Offender Release Dates and the King Decision .........................................................10 C. OMNI Case Management System .............................................................................11 D. OMNI Maintenance and Upgrades ............................................................................12 CHRONOLOGY OF OFFENDER EARLY RELEASE ISSUE ...........................................13 A. July 25, 2002 - March 2003: DOC reprogrammed OBTS to comply with the King decision. .....................................................................................................................13 B. March 2003 - December 2012: The programming error was undetected for over ten years. ....................................................................................................................14 C. December 2012: DOC discovered the computational error after a victim’s family complained about the assailant’s early release .........................................................15 1. December 7, 2012: Assistant Attorney General Ronda Larson advised DOC on the early release issue. .....................................................................15 2. December 7, 2012 – January 2013: Ms. Stigall notified her supervisor and others of the early release issue and submitted an IT request to fix the error. ...............................................................................................................17 D. January 3, 2013 - March 25, 2013: Sue Schuler conducted an IT Consultation regarding the King programming fix. ........................................................................24 E. April 3, 2013 - November 3, 2015: The King fix was pending in IT and -1- Delayed 16 M-releases without any significant progress. .........................................25 V. 1. April 3, 2013: Sue Schuler logged the King fix in Clearquest and scheduled it for M34, which had a release of September 13, 2013 ...............25 2. July 2013 – September 2013: While the King fix was pending, Wendy Stigall oversaw a statewide audit of jail good time calculations ...................26 3. November 27, 2013: Mark Ardiel first worked on the King programming fix. … .............................................................................................................29 4. February 2015 - September 2015: Mark Ardiel took paternity leave and the project was delayed several M-releases ...................................................31 F. November 2015: Work finally resumed on the King fix ...........................................32 G. December 2015: DOC Executive Staff and then Governor Inslee learned of the problem. .....................................................................................................................33 H. January 12, 2016: The King programming fix was completed ..................................35 CONCLUSIONS....................................................................................................................36 A. DOC’s failure to correctly calculate offender release dates was not intentional or malicious. ...................................................................................................................36 B. The advice tendered to DOC employees by the attorney general’s office was flawed .........................................................................................................................36 C. Ms. Larson’s Memorandum was not subject to supervisory review .........................38 D. DOC failed to implement a protocol of hand calculating offender release dates when errors were discovered .....................................................................................39 E. DOC management systematically failed to address the miscalculation of offender early release dates ......................................................................................................40 1. Wendy Stigall, Senior Records Manager .......................................................40 2. Denise Doty, Assistant Secretary, Administrative Services Division ...........42 3. Kathy Gastreich, Risk Management/Safety Director.....................................42 -2- Dough Hoffer, Chief Information Technology Officer .................................42 5. Clela Steelhammer, Manager, Legislative and Policy Coordination .............43 6. David Dunnington, IT Business Manager .....................................................43 7. Sue Schuler, IT Business Analyst ..................................................................44 F. The information technology group lacked a meaningful system for prioritizing work ...........................................................................................................................45 G. Lack of contractor resources did not cause the interminable delay in correcting OMNI’s calculation of early release dates .................................................................46 H. Other IT priorities, inordinately high turnover in DOC management and the IT group, and DOC budget concerns may have compounded the delays in addressing the King decision change request ............................................................46 I. Neither DOC secretary Dan Palcholke, nor his predecessor, Bernie Warner, was aware of the offender release date issue prior to mid-December 2015......................47 J. VI. 4. 1. Bernie Warner ................................................................................................47 2. Dan Palcholke ................................................................................................48 Neither Governor Jay Inslee nor members of his staff was aware of the offender release date issue prior to mid-December 2015 .........................................................48 RECOMMENDATIONS .......................................................................................................49 A. All AG opinions to DOC should be subject to supervisory review and approval .....49 B. The IT governance process should be re-structured ..................................................49 C. DOC should appoint an outside monitor ...................................................................50 D. DOC should adopt a policy requiring the hand calculation of release dates when problems are detected ................................................................................................50 E. DOC should adopt a policy requiring the immediate notification of the appropriate Assistant Secretaries of any system-wide error that affects the Sentencing, release, or supervision of offenders .......................................................50 -3- F. A Second Programmer who is proficient on each component of OMNI should be available ................................................................................................................51 G. DOC management should emphasize to all employees that its core mission is public safety ...............................................................................................................51 H. The DOC should create an ombudsman position ......................................................51 891.01 qb240102 -4- M EMORANDUM T 206.516.3800 F 206.516.3888 TO: The Honorable Jay Inslee Governor, State of Washington FROM: Carl Blackstone Robert Westinghouse DATE: February 19, 2016 RE: Department of Corrections – Early Release of Offenders I. INTRODUCTION Late on the afternoon of December 17, 2015, Governor Jay Inslee first learned that for over 13 years the Washington State Department of Corrections (“DOC”) had been releasing certain prison inmates earlier than their sentences allowed. After a brief investigation, on December 22, 2015, Governor Inslee notified the public that DOC had erroneously released more than 3,000 inmates earlier than they should have been released. On the same day he made this announcement, Governor Inslee hired Carl Blackstone and Robert Westinghouse of the Yarmuth Wilsdon law firm to conduct an investigation of how this problem occurred, who was responsible, and how similar problems could be prevented in the future. Over the next seven weeks, we interviewed – in some instances, on multiple occasions – 58 witnesses, including every current and former DOC employee who our investigation showed had any involvement in the early release problem; attorneys in the Washington State Attorney General’s office who advised DOC or supervised attorneys who advised DOC in conjunction with the early release of offenders’ problem; and Governor Jay Inslee and members of his staff who learned about the problem when it was discovered in December 2015. 1 (Attached hereto as Exhibit 1 is a list of all the witnesses we interviewed.) We also reviewed and/or electronically searched over 134,000 pages of documents which were primarily obtained from DOC. (Attached hereto as Exhibit 2 is an index of the documents we reviewed/searched). 2 1 We also met on two occasions with Washington State Senators Mike Padden, who chairs the Senate Law and Justice Committee and Steve O’Ban, a member of that committee. During these meetings we discussed the nature, scope, and timing of our investigation. We also solicited from them suggestions for questions that they might wish for us to address. Thereafter, the Senate Facilities and Operations Committee approved the hiring of an outside investigator. 2 Diana Breaux, a partner at Yarmuth Wilsdon, assisted with our investigation. 818 STEW ART STREET, SUITE 1400 SEATTLE W ASHINGTON 98101 II. EXECUTIVE SUMMARY The early release of thousands of prisoners over 13 years was caused by a series of errors coupled with bureaucratic incompetence, systemic failures of process and management, and an inexplicable failure both on an institutional and individual level to appreciate the fact that releasing even one inmate early, let alone thousands, undermined the core mission of the Department of Corrections, which is to protect the public. A. Synopsis of Factual Findings Our investigation developed facts to support each of these findings: 1. When the Supreme Court ruled on July 3, 2002, in In re King, 146 Wn. 2d 658 (2002) (decision attached hereto as Exhibit 3), that the DOC was erroneously denying offenders credit for early release time earned during presentence detention, DOC incorrectly reprogrammed its computer tracking system, thereby resulting in the early release of offenders with sentencing enhancements. The programming error went undetected for over ten years. To date, DOC has determined that as many as 2,176 offenders were released early between July 2002 and December 11, 2011. 3 2. The programming error was discovered in December 2012 when the family of an assault victim notified DOC that the release date for their son’s assailant, Curtis Robinson, appeared to be earlier than warranted by statute. 4 Steve Eckstrom, the DOC Victim Services Program Manager contacted Wendy Stigall, the DOC Records Manager, and Ronda Larson, an Assistant Attorney General who regularly advised DOC on legal matters, to advise them that DOC had been miscalculating offender early release dates. 3 On December 7, 2012, Ms. Stigall and Ms. Larson discussed the problem. Later that day, Ms. Larson sent an e-mail to Ms. Stigall in which she advised that DOC should hand-calculate Robinson’s release date, but that it could wait on the reprogramming of the computer tracking system for other offenders. 4. Wendy Stigall and others in DOC followed this advice. Ms. Stigall directed Elaine Downey, the records manger at the institution where Mr. Robinson was incarcerated to hand calculate his release date. Ms. Stigall also decided, in accordance with Ms. Larson’s advice, that it was not necessary to hand calculate 3 DOC has not yet reviewed each of these offenders’ files to determine if they were actually released earlier than they should have been. DOC estimates that as many as 25% of these offenders may not have been released early for a variety of reasons including the fact that an offender might have lost good time credit for misbehaving. Assuming this to be the case, then the total number of offenders released early during this period of time would be approximately 1,620. (2/16/16 e-mail from Peter Graham to Carl Blackstone attached as Exhibit 4). 4 Mr. Robinson’s release date was originally calculated as February 5, 2013. After the error was discovered, his release date was re-calculated by hand. The correct release date was March 22, 2013, or 45 days later than he had originally been scheduled for release. -2- the release dates for all other offenders serving enhanced sentences pending the reprogramming of the computer tracking system. 5 5. Ms. Stigall took various steps to notify her superiors at DOC about the early release problem, including advising her supervisor, Assistant Secretary Denise Doty, of the problem during a meeting on either December 10 or 11, 2012; sending an e-mail to Kathy Gastreich, DOC Risk Manager, notifying her of the problem and requesting her input; raising the issue at two management meetings attended by Ms. Doty, Ms. Gastreich, 5 Doug Hoffer, DOC’s Chief Information Officer, and others. Additionally, Ms. Stigall forwarded Ms. Larson’s December 7, 2012, opinion to Denise Doty, Kathy Gastreich, and Clela Steelhammer, DOC Legislative and Policy Coordinator. None of these individuals had any specific recollection of receiving this information from Ms. Stigall, although Ms. Doty and Mr. Hoffer each had a vague recollection that Ms. Stigall had notified them of the problem. None of these individuals took any further steps to address the problem or to follow up to be certain that the computer error had been corrected. 6 6. On December 27, 2012, Ms. Stigall submitted a “Change Request” to the DOC Information Technology (“IT”) group to correct the programming error. 7. David Dunnington, the IT Business Manager, assigned this Change Request to IT Business Analyst Sue Schuler to conduct an IT Consultation to determine the nature of the problem and the time it would take to fix it. Although IT Consultations are supposed to be completed promptly, Ms. Schuler took over three months to complete her consultation. Ms. Schuler received estimates from the programmer and tester that it would take approximately 36 hours to correct the programming error. 8. The IT Change Request was finally approved on April 3, 2013. The programming error was originally projected to be fixed by September 13, 2013. This did not happen. Instead, the fix was delayed 16 times over the next 30 months. David Dunnington was primarily responsible for repeatedly delaying this project. He was unable to explain the delays, at least in part because he failed to make any record of the reasons for the delays. 9. On November 2, 2015, Wendy Stigall met with Ira Feuer, who had recently been hired as the DOC Chief Information Officer. Ms. Stigall told Mr. Feuer that IT had failed to address the early release problem. Mr. Feuer told Ms. Stigall that he would look into it. Prompted by Mr. Feuer’s inquiries, IT finally began a fullscale effort to correct the programming error on November 3, 2015. Ms. Gastreich attended only one of these meetings. 6 Ms. Steelhammer apparently provided a brief synopsis of the problem at either DOC Executive Staff or a Senior Leadership meeting, but suggested that the problem was limited to only one or a small number of offenders and had been corrected. -3- B. 10. On December 15, 2015, the DOC Secretary, Dan Pacholke, and his leadership team, were made aware of the programming error and the nearly three-year delay in addressing this problem for the first time. 11. During the period between December 11, 2011, and December 15, 2015, 1,137 offenders were released earlier than they should have been. (Exhibit 4). 12. On December 16, 2015, Governor Inslee’s staff was made aware of the early release issue. Governor Inslee first learned of the issue himself late in the afternoon on December 17, 2015. 13. On December 22, 2015, Governor Inslee held a press conference notifying the public of the early release problem. 14. A final “fix” of the programming error was not implemented until January 12, 2016 - more than three years after the problem was identified. Summary of Conclusions 1. The failure to correctly calculate offender early release dates was not intentional or malicious. In response to the King decision, DOC reprogrammed its computer system to use an incorrect formula to calculate offenders’ release dates. This programming flaw went undetected for over 10 years. Once it was detected, it took DOC over three additional years to correct the problem. 2. The advice tendered to DOC employees by the Attorney General’s office was seriously flawed. Assistant Attorney General Ronda Larson’s advice that DOC need not hand calculate offender release dates pending a fix to the computer system and her advice that DOC could continue releasing inmates early until the programming error was corrected created a risk to public safety and undermined the core mission of DOC. Her advice appears to have played a part in DOC’s lethargic response to this problem. 3. Ronda Larson’s advice, which was set forth in a memorandum which she prepared on December 7, 2012, did not receive supervisory review. Ms. Larson was supervised by Assistant Attorney General Paul Weisser. Although she did not discuss her advice on the early release issue with Mr. Weisser, she did copy him on her December 7, 2012, e-mail. Mr. Weisser did not recall this e-mail even though it is reasonable to conclude that it should have caught his attention. 4. DOC failed to follow its normal protocol to hand-calculate offender release dates when errors were discovered. In the past when errors in release dates occurred, DOC typically hand calculated an offender’s release date until the error was corrected in the computer system. In this case, however, DOC failed to follow this procedure. -4- 5. There was a systematic failure of management at DOC to address the miscalculation of offender early release dates. The following DOC employees were aware of the problem. All but Wendy Stigall failed to address the problem: a. Wendy Stigall, the DOC Records Manager, learned of the problem on December 7, 2012. Thereafter she did an admirable job of notifying DOC managers and others of the problem and in initiating the computer fix. Ms. Stigall waited for almost three years, however, to intervene or seek management involvement in light of the repeated delays by IT in correcting the early release date programming error. b. Denise Doty, Assistant Secretary of the Administrative Services Division, learned of the problem from Wendy Stigall and did not notify the DOC Secretary or other Assistant Secretaries. She did not discuss the matter with the Chief Information Technology Officer, Doug Hoffer, who reported to her; nor did she direct Ms. Stigall to hand calculate offender release dates pending the fix of the programming error. c. Kathy Gastreich, DOC Risk Management Director, was notified by Wendy Stigall of the issue via e-mail. She also was sent a copy of Ronda Larson’s memorandum. She took no action. Ms. Gastreich claimed to have no recollection of this issue being raised. d. Clela Steelhammer, DOC Legislative Liaison, was also notified by Ms. Stigall of the issue via e-mail and was provided with a copy of Ms. Larson’s memorandum. She took no action beyond providing a cursory report at a subsequent Senior Leadership meeting. Ms. Steelhammer claimed to have no recollection of this issue being raised. e. Doug Hoffer, DOC Chief Information Officer, managed the Information Technology Unit. He did not have an effective process for being kept apprised of IT work on correcting defects and implementing enhancements to DOC’s computer tracking system. He was not informed about the offender early release problem or its many delays. He took no action to involve himself in this particular problem. f. Dave Dunnington, the IT Business Manager, failed to recognize the significance of the early release problem or to properly prioritize the programming fix required to correct it. Instead, he repeatedly pushed it back to later and later completion dates. This resulted in a three-year delay in correcting the problem. g. Sue Schuler, the IT business analyst responsible for shepherding the early release programming fix to completion failed to identify the importance of the project or to effectively manage its progress. She did not provide adequate interface between the business user, Ms. Stigall, and the IT group. -5- 6. C. The IT group lacked a meaningful system for prioritizing work. The most glaring flaw in the system was the failure to set priorities based on an assessment of the impact a particular defect or enhancement would have on public safety. Additional flaws in the prioritization system included the following: a. For the most part priorities were set by IT representatives rather than by DOC personnel who actually worked in the operating divisions, e.g., prisons and community corrections, and had much greater appreciation as to how IT problems affected the mission of DOC; b. Senior DOC management was minimally involved in setting priorities. Instead, it appears that priorities were set largely based on whoever “squeaked” the loudest. This “squeaky wheel” phenomenon was a poor substitute for a logical ordering of work based on its importance to DOC and its impact on the community; c. There was little if any oversight of this process by Doug Hoffer, the Chief Information Officer, and little communication between him and Dave Dunnington, the IT Business Manager for Prisons. 7. The cause of the interminable delay in correcting the early release programming error was not due to a lack of contractor resources. 8. Other IT priorities, inordinately high turnover in DOC management and the IT group, and DOC budget concerns may have compounded the delays in addressing the early release programming error. 9. Neither DOC Secretary Dan Pacholke nor his predecessor, Bernie Warner, was aware of the offender release date issue prior to mid-December 2015. 10. Neither Governor Jay Inslee nor members of his staff was aware of the early issue prior to mid-December 2015. Summary of Recommendations 1. All AG Opinions to DOC should be subject to supervisory review and approval. Even though Assistant Attorney General Ronda Larson copied her supervisor on her December 7, 2012, advice to DOC, he failed to read it. We are convinced that had Mr. Weisser been aware of Ms. Larson’s advice, he likely would have intervened and modified her advice. The Attorney General’s Office should adopt a policy requiring that advice given by an Assistant Attorney General to DOC relating to release dates and other significant matters must be subject to supervisory review and approval. -6- 2. DOC should restructure its IT governance process. DOC lacks an effective governance process for prioritizing IT work. The process needs to be restructured to ensure that public safety is the primary criterion used to prioritize projects. To accomplish that we recommend the following changes: 3. a. Business Users Should Set Priorities - Priorities should not be set by the IT group. Rather, the business users who understand the impact of IT defects and enhancements should have primary responsibility for setting priorities. The business users must be represented by senior managers from the operational divisions, e.g., prisons and community corrections, who have sufficient authority to set priorities and be held accountable for their decisions. IT, on the other hand, must promptly provide technical input such as estimates of time required to code, test, and implement a fix, and regularly update business users on progress. b. Documentation - An effective prioritization process must include sufficient documentation to track the progress of IT work and include an explanation for all delays in that work. This tracking system must be easily accessible by the business users so that they can be kept apprised of the status of the work. c. Accountability - Managers must effectively monitor work to ensure that it is being performed promptly. At a minimum, the CIO and Assistant Secretaries should be provided with a spreadsheet setting forth the status of all pending projects. d. Failsafe Mechanism - The system must include a failsafe mechanism that requires supervisory review and approval of further delay in a project if it has not been completed in a fixed time period. DOC should appoint an Outside Monitor. It would be appropriate to appoint an outside monitor to oversee the restructuring of the IT prioritization process and then oversee its operation, at least for a limited period of time. 4. Offender release dates should be hand calculated pending a sentencingrelated fix in the computer system. DOC should adopt a policy requiring that once a sentence calculation error is identified, offenders’ release dates should be hand calculated until the error is corrected in the computer system. -7- 5. DOC Assistant Secretaries must be notified of any system-wide errors affecting sentencing, release or supervision of offenders. DOC should adopt a policy that whenever a defect in calculating an offender’s sentence, release date, or supervision term comes to the attention of any DOC employee, that employee must forward the information regarding the defect to the Assistant Secretaries in charge of the Administrative Services, Prison, and Community Corrections divisions, and the Statewide Records Manager. 6. A second programmer capable of fixing sentencing errors should be available. DOC has relied on an outside contractor to fix the more complex sentencing errors. At present, DOC relies primarily on one programmer working for this contractor. In the future, DOC should ensure that at least a second programmer is available to address sentencing issues. 7. DOC management should emphasize to all employees that its core mission is public safety. The early release of offenders, particularly after the issue was discovered in December 2012, was based in part on an inability by DOC employees to appreciate the impact that the early release of offenders would have on public safety. The new DOC Secretary should make it a top priority to encourage all DOC employees to be more attentive to public safety and the other objectives of the Department. 8. DOC should create an ombudsman position. We have detected a reluctance by some DOC employees to come forward and voice concerns or complaints about the department. The appointment of an Ombudsman would encourage more DOC employees to express their concerns. III. A. BACKGROUND DOC Organizational Structure The Washington State DOC was created in 1981 by the Washington State Legislature. DOC is responsible for managing adult prison facilities and supervising adult offenders residing in the community. The law creating DOC provides that “the system should ensure the public safety.” RCW 72.09.010. DOC currently operates 12 prisons in the State of Washington, ten of which house male inmates and the other two house female inmates. Presently there are approximately 16,000 inmates incarcerated in these prisons. DOC employs approximately 8,100 individuals and has an annual operating budget of approximately $1.7 billion. DOC is the third largest cabinet agency in Washington State. The Secretary of DOC is a cabinet-level position appointed by the Governor. The current Secretary is Dan Pacholke, who -8- has held this position since October 2015. 7 Prior to Secretary Pacholke, Bernie Warner served as the DOC Secretary from July 2011 to October 2015. DOC is organized into the following five divisions, each of which is headed by an Assistant Secretary, as identified below: Administrative Services Division – Julie Martin; Community Corrections Division – Anmarie Aylward; Health Services Division – Kevin Bovenkamp; Offender Change Division – Keri Waterland; and Prisons Division – Steve Sinclair. The Administrative Services Division (“ASD”) has been the primary focus of this investigation, because the three departments principally responsible for calculating, tracking and addressing issues with prisoner release dates – Records, Information Technology (“IT”) and Risk Management – all are within this division. The Statewide Records Manager, the Chief Information Officer (“CIO”) in charge of the Information Technology (“IT”) group, and the Director of Risk Management/Safety all report to the ASD Assistant Secretary. At the time the early offender release problem came to light in December 2012, Denise Doty was the Assistant Secretary for ASD. She left DOC in January 2014 to take a position with the Office of Financial Management (“OFM”). 8 Following Ms. Doty’s departure from DOC, Brian Tinney served as the interim ASD Assistant Secretary until November 2015, when he left DOC. Julie Martin was appointed to be ASD Administrative Secretary in November 2015. She remains in that position today. Wendy Stigall was the Statewide Records Manager at the time the problem was discovered in December 2012. She remains in that position today. The CIO in December 2012 was Doug Hoffer. In March 2014, he left DOC and assumed his current position as Assistant Director, Telecommunications Division, at the Washington Technology Office. Thereafter, the following individuals served as CIO: Peter Jekel (March 2014 - April 2014); Jibu Jacob (April 2014 - July 2014); David Switzer (July 2014 - May 2015); and Lee Baublitz (May 2015 - August 2015). The current CIO, Ira Feuer, was appointed by Secretary Bernie Warner in August 2015. The Risk Management Director at the time the problem surfaced was Kathy Gastreich. She remains in the same position, but no longer reports to the ASD Assistant Secretary. 7 On February 6, 2016, Secretary Pacholke announced that he would be resigning from the Department of Corrections. 8 Ms. Doty announced her resignation from OFM on February 12, 2016. -9- B. Offender Release Dates and the King Decision Once an individual is arrested for a crime, he or she is either released on bail or detained in a local county jail pending trial. If the defendant subsequently pleads guilty or is convicted following trial, a judge imposes the defendant’s sentence. If the court’s sentence of imprisonment exceeds one year, the offender is transferred to DOC custody to serve his or her sentence in prison. The offender is entitled to receive credit for any time served in jail. He or she may also earn credit for “good time” during local incarceration. DOC is then responsible for determining the offender’s date of release from prison. This can be a complicated process. In its simplest form, however, the process begins with the sentence imposed by the court. DOC will then reduce the sentence by the number of days an offender spent in jail (“actual time”) and by the number of days of accumulated good time. While in DOC custody, an offender can further reduce his or her sentence through “earned release time.” Earned release time consists of both good time credit, which is awarded to inmates who stay out of trouble, and earned release time, which is awarded to offenders who participate in approved programs, including work and school. For most criminal offenses, earned release time is capped at one-third of an offender’s sentence. For certain more serious offenses, earned release time is limited to ten or fifteen percent of the base sentence. Since 1995, Washington state has adopted a series of laws which impose mandatory prison sentences for certain offenses (“mandatory minimums”) and also impose sentencing enhancements if an offender uses a firearm, a deadly weapon or engages in a crime with a sexual motivation. For instance, if an offender uses a firearm in connection with a robbery he or she will receive a five-year enhancement on top of his or her prison sentence for the robbery. An offender is not eligible for either good conduct time or earned release time on a mandatory minimum or enhanced sentence. The sentence must be served in its entirety (“day for day”). Prior to the King Decision, DOC took the position that an offender’s sentence began to run on the date that he or she was incarcerated in a local jail. If the defendant was subsequently sentenced to an enhanced sentence, DOC would run that sentence first and give the offender no credit for earned release time that was earned in jail or in prison. Once the enhanced sentence was served, the defendant would then begin serving his or her base sentence, for which he or she could receive earned release credit of up to one-third of the sentence. On July 3, 2002, the Washington Supreme Court decided In Re King, holding that DOC had been improperly calculating release dates for those offenders who were serving enhanced sentences. (Exhibit 3). In King, the offender was serving a 190-month base sentence and an additional 60-month enhancement. 9 When he was transferred to DOC custody following conviction, the Snohomish County jail certified that he had served 287 days in pre-sentence custody, for which he had earned 57 days of early release time. Because DOC calculated the offender’s release date by treating the time served in local custody as part of his enhanced sentence, for which he could not earn early release time, the Supreme Court ruled that he was unjustly deprived of his jail early release time. The King decision directed DOC to begin - 10 - running the enhancement on the day the offender was transferred to State custody, applying the time in local custody against the offender’s base sentence. 10 The proper application of King is to deduct the time in local custody and early release time earned in local custody from the base sentence before calculating early release time available on the balance of the base sentence. For example, if an offender received a base sentence of 1,490 days, served 138 days in jail, and earned 69 days earned release time in jail, he or she should have 1,283 days remaining on his or her base sentence when he or she enters prison. The correct computation of the remaining potential earned release time would be one third of 1,283, or 427.66 days. This maximum of 427.66 days when added to the 69 days of earned release time from local custody equals 496.66 total days of earned release time. This is exactly one-third of the original base sentence of 1,490 days. Under this calculation, the offender’s estimated early release date would be 855 days (1282 – 427.66 = 855.34) after the offender begins serving his or her remaining base sentence in state custody. At the time of the King decision in July 2002, DOC utilized the Offender Based Tracking System (“OBTS”) to electronically track all offenders in DOC custody or under its supervision. In reprogramming OBTS after the King decision, an offender’s earned release time was wrongly computed against the base sentence after reducing the number of days in the base sentence by the number of good time days credited while in local custody. 11 This computation resulted in offenders receiving more than the statutory maximum good time. More specifically, the additional maximum earned release time was determined by first subtracting the good time credit received by the offender in local custody from the total base sentence (1,490 – 69 = 1,421 days) and then computing one third of the remaining base sentence (1,421 days x 1/3 = 473.66 days). When added to the good time release days that the offender had earned in local custody, it was determined that the offender’s maximum earned release time was 542.66 days (473.66 + 69 = 542.66 days). Thus, the offender in this example would have to serve 817.34 additional days after entering DOC custody. (1490 – (138 + 69 + 473.66) = 809.34 days). His or her estimated release date would come 46 days sooner than allowed by statute (542.66 - 496.66 = 46 days) and the offender would actually serve 46 fewer days in custody than permitted by statute. (The incorrect and correct calculation sheets using this example are attached as Exhibit 5). C. OMNI Case Management System In approximately 2000, DOC began work on a new web-based electronic tracking system, intended as an upgrade to OBTS, which came to be known as Offender Management Network Information (“OMNI”). From the outset, OMNI proved to be a challenge to design and implement. Initially, International Business Machines (“IBM”) was awarded the contract to 10 The incorrect calculation of an offender’s earned release date did not occur for offenders without enhancements because the offender’s service of the base sentence was not interrupted when he was transferred to State custody. Without an enhancement, the base sentence begun in local custody presentence was simply continued after transfer to DOC custody with the maximum earned release time calculated against the remaining base sentence after transfer. 11 The number of days the offender actually served in local custody, in this example 130 days, was applied against the enhancement portion of the sentence. This was contrary to the mandate of the King decision. - 11 - design and implement this updated electronic tracking system, but it was unable to complete the task. In 2006 the contract was moved to Sierra Systems (now Sierra Cedar), which successfully designed and implemented the system. OMNI was brought on line in August 2008. After implementation, there were a continuing series of programming glitches, many of which affected sentencing structure and time accounting (“SSTA”). Over time, as OMNI has been refined, the number of SSTA problems have been substantially reduced. Nevertheless continuing legislative changes to this State’s sentencing laws, coupled with a slowed, but not eliminated, set of mechanical calculation defects, has caused IT to have to complete a never ending string of SSTA fixes to OMNI. D. OMNI Maintenance and Upgrades The DOC IT Department oversees maintenance of and upgrades to OMNI, with programming support from Sierra Cedar. OMNI programming changes typically are rolled out every eight weeks as part of regularly scheduled maintenance or “M” releases. The M-release dates are selected in advance. Each M-release is given a number. (Attached hereto as Exhibit 6 is list of dates for M-release 32 through M-release 50, hereafter referred to as M34, etc.) Urgent projects or “hot fixes” are performed on an as-needed basis between M releases. Individual OMNI programming changes typically are initiated by an IT change request submitted by a DOC employee. The same IT request form is used for all requests, whether the issue is minor or critical. We heard from several people that high-ranking DOC employees (e.g., Assistant Secretaries) often bypass the IT request process and make their requests directly to IT business analysts (“drive-by requests”). Change requests are submitted to an IT gatekeeper who forwards the request to the IT Triage Team for review. If the Triage team approves the request, it is assigned to an IT business analyst, who enters the request into Clearquest (“CQ”), a DOC database that tracks OMNI programming changes. 12 Clearquest requires that the business analyst characterize the programming change as either a defect or an enhancement. DOC has no written definition of either a defect or an enhancement. As explained to us by various IT personnel, a defect occurs when a particular IT function fails to operate in accordance with its requirement. An enhancement occurs when an IT function operates in accordance with the underlying business requirements, but it is later determined that the requirements were incorrect and need to be changed. The business analyst assigns a severity rating to each IT project added to Clearquest. The severity levels range from level 1, the most severe (“Critical impact”), to level 4, the least severe (“Minimal Business Impact”). (Exhibit 7). Finally, the business analyst also enters into Clearquest a priority date as to when the defect or enhancement should be fixed, and places the request into a queue for a particular M-release. Approved change requests are assigned to a programmer or coder, either a DOC employee or a Sierra Cedar programmer who undertakes the fix. Once the programmer completes his or her work, the matter is referred for testing to ensure that the problem has been corrected. Once the testing is successfully completed, the change is implemented and the matter is closed. 12 Clearquest is a tracking system that is internal to the IT group. Other DOC employees cannot access this tracking system. - 12 - DOC does a not have a formal written process setting forth the procedure to be followed in prioritizing OMNI defects or enhancements. As explained to us, business analysts assigned responsibility for individual change requests are to work with business users who submitted the change requests to develop specifications for the changes; to represent the interests of the business users in helping to set priorities; and to keep the business users apprised as to the progress of the IT group in completing the programming, testing, and implementation of the changes. A group called the OMNI team meets twice a week – on Mondays from 11:00 a.m. to 12:00 p.m. and on Wednesdays from 1:00 to 2:00 p.m. – to review the status of projects pending for an upcoming M-release. During these meetings, programming changes scheduled for the current M-release are the predominant focus of the review and discussion. The meetings are led by Dave Dunnington, the IT Business Manager for Prisons. Other attendees during the period December 2012 through December 2015 typically included Luann Kawata, IT Business Manager for Community Corrections; all IT Business analysts; Deepak Sandanan, the lead DOC tester; Trang Nguyen, the lead DOC programmer; and Mark Ardiel of Sierra Cedar, who typically participated by phone. There are no minutes or recorded notes of OMNI team meetings. During these meetings the OMNI team determines if they will be able to fix a particular defect or enhancement in time for the upcoming M-release. If they conclude that they cannot address the problem or complete the work, the matter will typically be moved to a later M-release date. Only Dave Dunnington and Deepak Sandanandan had the authority to move a CQ item to a later M-release date. Messrs. Ardiel and Sandanandan provided information as to the time required to complete the programming and testing for defects and enhancements raised in change requests assigned to the current M-release. Mr. Dunnington then made the final decision as to what work could be completed in the current M release cycle and which change requests needed to be pushed back to a later M release. To allow adequate time for testing, the programming work for a particular change request generally needed to be completed by the end of the fourth week of the eight-week cycle, after which a “code-freeze” was imposed. Change requests for which programming and testing could not be completed were transferred to a later M release. IV. CHRONOLGY OF OFFENDER EARLY RELEASE ISSUE A. July 25, 2002 - March 2003: DOC reprogrammed OBTS to comply with the King decision. In response to the King decision, DOC recognized that it had to re-program OBTS to account for an offender’s jail time and jail good time. Assistant Attorney General Paul Weisser advised DOC Secretary Joe Lehman on July 3, 2002, that The [King] decision probably won’t result in the offenders serving more or less time than they otherwise would have served, but implementing the rule the court adopted might involve changes to OBTS and record keeping functions. The correctional records managers may have their hands full with this one. I suspect that many offenders’ (hundreds or thousands) time structure will have to be individually recalculated, because I don’t think OBTS can accommodate the rule the court announced in King on a system wide basis. - 13 - (July 3, 2002 e-mail attached as Exhibit 8). Secretary Lehman responded to Mr. Weisser’s e-mail that “unfortunately he [Mr. Weisser] points out a real potential with work that will have to be done by records staff.” (Exhibit 8). The matter was then referred to Janice McMann, the DOC State Records Manager. On July 25, 2002, Ms. McMann submitted a Work Request to the DOC Information Services requesting a programming change to OBTS to comply with the King decision. (Work Request attached hereto as Exhibit 9). It is not exactly clear as to when DOC made the change to OBTS, but we believe it was most likely completed by March 2003. It must be assumed that when this request was made no one recognized that the computation would be incorrect if the maximum earned release time was computed after instead of before the reduction from the base sentence of the good time earned in local custody. Thus, the resulting change was based upon an incorrect formula for calculating an offender’s early release date. Because the formula was incorrect, it did not matter whether the offender’s release date was calculated electronically through OBTS (or OMNI), or by hand. Until someone identified the error in the formula, the miscalculated early release dates for all offenders with enhancements remained undetected. B. March 2003 - December 2012: The programming error was undetected for over ten years. There was apparently no complaint about the miscalculation of early release dates during this ten-year period. It appears that as many as 2,176 offenders were released before their correct early release dates during the period between July 2002 and December 11, 2011. We have not uncovered any evidence indicating that any DOC official was aware of the programming error during this ten-year period. However, DOC is still producing e-mails and other documents to us; to the extent those records reflect that any DOC officials were aware of the programming error, we will supplement this report. An e-mail dated October 12, 2007 from Rhonda Larson to Leora McDonald, then the Records Supervisor in the Warrants Unit, addresses the application of good time credit for time served in local custody. (Exhibit 10). As Ms. Larson noted in her e-mail, the DOC practice after the King decision was to apply an offender’s actual time served in local custody first against any enhancement. She questioned why in this situation, DOC gave the offender credit for good time against his base sentence when the offender was statutorily precluded from earning good time while serving an enhancement or a mandatory minimum sentence. Although Ms. Larson’s e-mail correctly identified a fundamental problem underlying the early release issue, she did not grasp that the formula produced a calculation error that caused offenders to receive more than the statutory maximum good time, resulting in their early release. Nor have we uncovered any evidence that Ms. Larson or any other DOC employee appreciated or understood that this issue affected offender release dates. - 14 - C. December 2012: DOC discovered the computational error after a victim’s family complained about the assailant’s early release. In early December 2012, Steve Eckstrom, the Victim Services Program Manager, 13 received a call 14 from the father of a stabbing victim who was questioning why his child’s assailant, Curtis Robinson, was to be released significantly earlier than the victim’s family anticipated. Mr. Ekstrom completed his own calculations and concluded that the parent was correct. Because of this call, Mr. Ekstrom was concerned that DOC had not been calculating offender early release dates correctly. He called Wendy Stigall to advise her of this problem. Ms. Stigall told Mr. Ekstrom that she was going to request that the problem be corrected by the IT group. 1. December 7, 2012: Assistant Attorney General Ronda Larson advised DOC on the early release issue. On December 7, 2012, after speaking with Ms. Stigall, Mr. Ekstrom called Ronda Larson, the Assistant Attorney General in the Corrections Division of the Attorney General’s office who most regularly advised DOC employees on legal matters. He left her a voice message and followed up with an e-mail at 10:44 a.m. explaining the situation. He noted that DOC either needed to determine that its current calculation was correct or change it. (Mr. Eckstrom’s e-mail is attached as Exhibit 11.) Mr. Ekstrom did not follow-up on these initial contacts. After speaking with Ms. Stigall, he assumed that the problem had been fixed. He talked to Ms. Stigall a couple of times each month thereafter, but he never raised the subject again. He never received the Ronda Larson memorandum described below. He assumed that he had alerted his supervisor, Merlyn Miller, the Community Corrections Program Administrator, of the problem, but he cannot recall doing so. In the hours immediately following Steve Ekstrom’s calls, Wendy Stigall and Ronda Larson discussed the problem. Ms. Larson prepared a memorandum setting forth her understanding of the earned release date problem. She also offered her opinion as to how this problem should be addressed. She sent this memorandum to Wendy Stigall via e-mail at 2:29 p.m. Ms. Larson copied her supervisor in the Attorney General’s office, Paul Weisser. Ms. Larson’s memorandum, attached hereto as Exhibit 12, is significant in several respects. First, she observes that when an offender’s base sentence is short, the error in calculating the earned release date will result in the offender receiving more “good time” than warranted. This observation is incorrect. The calculation error impacted all offenders with enhanced sentences, not just those with short base sentences. Thus, the pool of offenders impacted by the miscalculation was significantly larger than Ms. Larson anticipated. 13 The Victim Services Program is responsible for notifying victims, who have registered, of an offender’s release date and for serving as a liaison with victims to establish safety plans following an offender’s release. 14 We do not know the exact date of this call because Mr. Eckstrom did not make notes of the conversation. He believed, however, that he received the call in early December 2012. - 15 - Second, although she recommended that DOC hand calculate the earned release date for Mr. Robinson, she continued “I don’t believe it is necessary, from a risk management perspective, to do hand calculations now of everyone in prison with an enhancement. Waiting for OMNI to be reprogrammed should be sufficient, except for in (sic) Robinson’s case.” At the end of her memorandum Ms. Larson repeats this opinion: As to the long process of reprogramming OMNI, it would be reasonable to not manually fix the hundreds of sentences that have enhancements and instead wait for the reprogramming to occur so that OMNI can do the recalculation automatically. Although this will result in offenders being released earlier than the law allows for the time being, until OMNI gets fixed, the DOC has been releasing them earlier for a decade (since the In re King decision), and a few more months is not going [sic] to make that much difference in light of this (with the exception of Robinson’s case.) Furthermore, this is something that the DOC has identified internally, rather than something that is being forced upon it by an outside entity such as the court. It is therefore not so urgent as to require the large input of personnel resources to do handcalculations of hundreds of sentences. (Exhibit 12 at WS_000018). Ms. Larson also recognized the significant liability that the State might face if Mr. Robinson was to be released early. As she wrote, If the DOC does not fix Robinson’s sentence, the likelihood that DOC will be sued and lose in a tort lawsuit is unreasonably high, if Robinson were to [be] releas[ed] and immediately go and kill the victim, for example. In such a scenario, because the DOC knew that Robinson was getting 58% good time illegally, and didn’t fix it, the DOC would lose such a lawsuit and sustain a lot of monetary damages. (Exhibit12 at WS_000017). It is difficult to reconcile this conclusion with Ms. Larson’s advice that DOC need not hand calculate release dates for offenders other than Mr. Robinson. Ms. Larson clearly understood that there were potentially hundreds of offenders currently in prison whose earned release dates had been miscalculated, yet she failed to consider the community’s safety risk if any of these offenders were to be released early and reoffend. Her acknowledgement that the problem had existed for a decade provides scant support for her opinion that DOC could wait for the OMNI fix. Nor does her recognition that an OMNI fix may take months to implement explain why she did not believe it to be imperative to stop the practice of early releases immediately. Ms. Larson understood that there was an alternative – hand calculating the earned release dates for offenders who were about to be released – but she only found this remedy to be necessary for Mr. Robinson, whose victim’s family had complained. 15 15 Ms. Larson resigned from the Attorney General’s Office on February 12, 2016. - 16 - Paul Weisser has no recollection of receiving Ms. Larson’s e-mail and attached memorandum. Ms. Larson does not recall having spoken with her supervisor about the issue. Although it seems, in retrospect, like the problem was of a sufficient magnitude to command some attention, it did not. On December 7, 2012, at 2:38 p.m., Ms. Stigall forwarded the Ronda Larson e-mail and attached memorandum to Sue Schuler, a business analyst in the IT group with whom she regularly worked on SSTA programming issues. (Exhibit 13). Ms. Schuler served as the liaison between business users such as Ms. Stigall, and the IT group to facilitate the programming of OMNI enhancements and defects. Ms. Schuler advised us that she did not think much about the advice in the Larson memorandum. Her responsibility was to understand the specifications of the change in the OMNI system that the business user was seeking. She added that she did not know the number of offenders whose earned release dates had been incorrectly calculated, but based on the Larson memorandum, she at the very least understood that the problem affected only those offenders with short base sentences, which Ms. Larson identified as “hundreds of sentences.” Also on December 7, 2012, at 3:08 p.m., Ms. Stigall sent Ms. Larson’s e-mail and memorandum to Elaine Downey, the records manager at Cedar Creek prison where Mr. Robinson was incarcerated. (Exhibit 14). She advised Ms. Downey that she needed to manually adjust Mr. Robinson’s release date. Ms. Downey responded that she would address the problem when she returned to work the next Tuesday. Ms. Downey told us that she and Ms. Stigall recalculated Mr. Robinson’s release date to ensure that he was not released early. Ms. Downey said that Ms. Stigall told her that she was going to have the problem fixed in OMNI. Ms. Downey assumed that the problem had been fixed and did not follow-up with Ms. Stigall or hand calculate other offenders’ release dates. Ms. Downey told us that the amount of time it took to hand calculate an offender’s release date varied from five to ten minutes in a simple case to several hours for a more complex case. 2. December 2012 - January 3013: Ms. Stigall notified her supervisor and others of the early release issue and submitted an IT request to fix the error. Ms. Stigall told us that on either December 10 or 11, 2012, she met with her supervisor, Assistant Secretary Denise Doty. Ms. Stigall said that during that meeting she told Ms. Doty about the programming error and the fact that it had resulted in the early release of offenders. According to Ms. Stigall, Ms. Doty directed her to notify the DOC risk manager, Kathy Gastreich, of this issue. Ms. Doty told us that she has no recollection of this meeting, although she did have a vague recollection that at some point Ms. Stigall told her about the problem. Despite Ms. Doty’s lack of recollection, we find that Ms. Stigall did in fact notify her of the issue on or before December 11, 2012. We base this conclusion on two reasons. First, given the significant nature of the problem, it is certainly something that Ms. Stigall would have felt the need to bring to the attention of her supervisor. Second, and more importantly, Ms. Stigall’s recollection that Ms. Doty directed her to notify Kathy Gastreich is corroborated by the fact that Ms. Stigall did just that. On December 11 at 11:24 a.m., Ms. Stigall sent an e-mail to Kathy - 17 - Gastreich apprising her of the early release problem and providing her with a copy of the Ronda Larson memo. (E-mail attached as Exhibit 15). Ms. Stigall also copied Ms. Doty on this e-mail which is consistent with the fact that she would have already discussed the matter with Ms. Doty. Both Ms. Gastreich and Ms. Doty have told us that neither one of them has any recollection of this e-mail. This is troubling for several reasons. As an initial matter, Ms. Stigall’s e-mail was directed to Ms. Gastreich and copied to Ms. Doty, who was Ms. Gastreich’s supervisor. One would have expected that this at the very least would have gotten Ms. Gastreich’s attention and caused her to read the e-mail. Moreover, the e-mail attached a memorandum from an Assistant Attorney General, which contained the warning “AttorneyClient Privileged Communication.” This too should have made an impression on both Ms. Gastreich and Ms. Doty. Additionally, Ms. Stigall’s e-mail raised a significant issue and specifically requested a response from Ms. Gastreich. As Ms. Stigall writes: Before I do an IT request to have the programming changed I wanted to run this past you for your input. OMNI has been calculating these sentences the same for approximately 10 years now (since the King decision). We are going to manually adjust Robinson’s case but this has the potential to add time to several hundred offenders. We are not talking huge amounts of time but in this case as an example it will add about a month. Implementing these changes is always a lot easier if it is going the offender favor. (Exhibit 15, emphasis added). This was a brief but extremely important e-mail that requested Ms. Gastrich’s input. Her failure to remember the e-mail or respond to it is mystifying. During the course of our investigation we have reviewed several e-mail exchanges between Ms. Stigall and Ms. Gastreich related to other matters. Those exchanges typically began with a request from Ms. Stigall for input or other information from Ms. Gastreich. In each instance, Ms. Gastreich responded promptly to Ms. Stigall. (E-mails attached as Exhibit 16). We are at a loss to understand how Ms. Gastreich apparently failed to read Ms. Stigall’s December 11 e-mail. It is particularly troubling that Ms. Stigall’s brief e-mail raised issues which Ms. Gastreich was specifically responsible for addressing. She is the DOC risk manager and, as such, it is her job to assess the public safety and financial risks associated with proposed DOC action. In this case, Ms. Gastreich was being advised that DOC had been releasing prisoners early for over ten years and that an Assistant Attorney General believed that DOC “from a risk management perspective” could continue releasing possibly “hundreds” of inmates for “a few more months” until the problem was fixed in OMNI. This should have gotten Ms. Gastreich’s full attention but it did not. During our interview of Ms. Gastreich she forthrightly acknowledged that she should have read Ms. Stigall’s e-mail. Had she done so, she claims that she would have disagreed with Ms. Larson’s advice and would have directed DOC to hand calculate release dates pending the OMNI fix. - 18 - Assistant Attorney General Dan Judge had a recollection of talking to Ms. Gastreich about this matter. Mr. Judge works in the Attorney General’s Torts Division and he is responsible for representing DOC in tort lawsuits filed by crime victims against DOC for negligent supervision of offenders. In this capacity, Mr. Judge had frequent contact with Ms. Gastreich. Mr. Judge told us that he recalled a conversation he had with Ms. Gastreich on Friday December 7, 2012, which he claimed took place during the early afternoon between 1 p.m. and 1:30 p.m. Mr. Judge took no notes of this conversation and recalled all of the details simply by memory. Mr. Judge said that during the conversation Ms. Gastreich advised him that DOC was working on a computer glitch which had resulted in the early release of inmates. Ms. Gastreich said that she was working with Ronda Larson, who had expressed the view that DOC could wait to address the problem until after the computer fix was completed. Ms. Gastreich wanted to know what Mr. Judge thought of this advice from a risk management point of view. Mr. Judge claims he told Ms. Gastreich that DOC needed to fix the problem or they would be facing tort liability. Ms. Gastreich told us that she had no recollection of this conversation with Mr. Judge. We do have some concerns as to Mr. Judge’s recollection. He admitted to us that in late December 2015 he first learned through the extensive media coverage that DOC had been releasing thousands of inmates early. At that time he did not recall his conversation with Ms. Gastreich. Mr. Judge admitted that he was able to read various documents posted on line by DOC, including Ronda Larson’s December 7, 2012, 2:29 p.m. e-mail to Wendy Stigall with a copy to Paul Weisser. Mr. Judge claimed that over the course of several days he started to remember his conversation with Ms. Gastreich. As his memory got better he was able to remember the exact date the conversation took place, “December 7”; the actual day of the week the conversation occurred, “a Friday”; and the time of the conversation, “early afternoon between 1 and 1:30 p.m.” We asked Mr. Judge how he knew the conversation took place on a Friday and he said he “just knew” it happened on a Friday. We also asked him why he thought the conversation had occurred on December 7, and he said that it was because that day was a Friday and the following Friday, December 14, was the date of the Sandy Hook shooting and he knew that the conversation did not take place then. Our collective experience as former federal prosecutors is that most people’s memory gets worse over time not better. Given that Mr. Judge had no notes of his conversation with Ms. Gastreich and that over three years had passed since the conversation allegedly occurred, we are at a loss to understand Mr. Judge’s certainty that his conversation with Ms. Gastreich took place on Friday December 7 in the early afternoon. Moreover, there is simply no evidence that Ms. Gastreich was even aware of the issue on December 7. Ms. Stigall has told us that the first and only time she communicated with Ms. Gastreich was via e-mail, at Denise Doty’s direction, on December 11, 2012. Additionally, we are troubled by the fact that Mr. Judge failed to follow up with Ms. Gastreich after the December 7 conversation. Although he met with Ms. Gastreich on a fairly regular basis, he did not ask her whether DOC had fixed the problem for almost three years. Mr. Judge did say that he recalled a fairly recent conversation with Ms. Gastreich in which he asked her if the problem had been fixed and she said that it had not. Unlike the December 7 - 19 - conversation, Mr. Judge could not recall the precise date and time of this conversation. He claimed it could have taken place in either August or September 2015. Ms. Gastreich has no recollection of this conversation. Mr. Judge also told us that he recalled an e-mail exchange and conversation on December 7, 2012, with Tim Lang, an Assistant Attorney General, who supervised the Attorney General’s Corrections Division. Mr. Lang supervised Ronda Larson and also attended weekly DOC executive staff meetings. Mr. Judge claims that after he talked to Kathy Gastreich on December 7, 2012, he either called Mr. Lang or sent him an e-mail indicating that he had had a “distressing call” from Ms. Gastreich. Mr. Judge said that Mr. Lang called him and he told Mr. Lang about his conversation with Ms. Gastreich, relaying to Mr. Lang that there had been a computer error over a lengthy period of time which had resulted in the early release of offenders. Mr. Judge also told Mr. Lang that Ronda Larson was working on providing advice to DOC but that it had not yet been provided. Mr. Judge said he wanted Mr. Lang to be aware of the issue and he advised Mr. Lang to check in with Ms. Larson. According to Mr. Judge, Mr. Lang asked Mr. Judge to send him a follow-up e-mail reminding Mr. Lang of the conversation. Mr. Judge was able to locate some e-mail communications he had with Mr. Lang on December 7, 2012. Mr. Judge recalled that on that day he had asked Mr. Lang to provide him with some legal briefs relating to a matter that Mr. Judge and others in his office were addressing. In response to this request, Mr. Lang sent an e-mail to Mr. Judge at 4:34 p.m. on December 7, 2012, attaching copies of two legal memoranda. (E-mail attached hereto as Exhibit 17). Mr. Judge replied to this e-mail at 4:49 p.m. on December 7, 2012. Mr. Judge claims that in this reply he mentioned the “distressing call” he had received from Ms. Gastreich. Mr. Judge deleted this e-mail at some point and the Attorney General’s office has been unable to retrieve Mr. Judge’s reply. 16 Mr. Lang has denied that Mr. Judge ever told him either over the phone or by e-mail about his conversation with Ms. Gastreich. Mr. Lang was quite certain that if Mr. Judge had told him about the early release of offenders he would have raised it with Ms. Larson and with DOC executive staff. We are inclined to agree with Mr. Lang on this point. As explained above, we are concerned about Mr. Judge’s recollection. Mr. Judge told us that in late December 2015 he read media reports relating to the early release of offenders. At that time he said he did not recall any communication with Mr. Lang relating to this issue. On either December 29, 30, or 31 he met with his supervisor Pam Anderson to tell her that he had recalled talking to Kathy Gastreich about the early release issue. However, when he talked to Ms. Anderson, Mr. Judge said he had no recollection of discussing the matter with Tim Lang. Mr. Judge then told us that on January 5. 2016, he did in fact recall his December 7, 2012, conversation with Mr. Lang. He could not explain why he was able to recall this conversation later. Again, we are perplexed that Mr. Judge’s memory seems to have improved dramatically over time and he can offer no plausible explanation as to why this occurred. Moreover, Tim Lang is a very conscientious and well-respected public servant. We are confident that had Mr. Judge in fact told him that prisoners were being released early and that Ms. Larson was preparing 16 Mr. Lang sent another e-mail to Mr. Judge at 4:44 p.m. on the same day attaching another legal memorandum to Mr. Judge. (e-mail attached hereto as Exhibit 18) - 20 - some sort of advice memo, Mr. Lang would have followed up with Ms. Larson and raised the matter with DOC executive staff. December 11 - 27, 2012 Ms. Stigall was out of the office on medical leave from December 11 through December 25, 2012. She returned to the office on December 26, 2012, and turned her attention once again to the early release matter. At 12:25 p.m. she sent an e-mail to Bryan Irwin requesting that he provide her with a “list of all active prison offenders that have a mandatory or an enhancement (please). We are looking at having to change how the jail credits are applied to all of these offenders.” (E-mail attached as Exhibit 19). This e-mail is significant because it indicates that Ms. Stigall understood that the early release problem was not limited only to those inmates who were serving a short base sentence, as Ms. Larson opined in her memo, but rather affected every inmate serving either a mandatory sentence or enhanced sentence. The same day, December 26, 2012, at 3:58 p.m., she sent an e-mail to Ronda Larson, with a copy to Paul Weisser, asking for Ms. Larson’s guidance in coming up with “the information that I need to have OMNI programmed.” Ms. Stigall provided Ms. Larson with three examples she had constructed to properly calculate an offender’s release date. (E-mail attached as Exhibit 20). Ms. Stigall requested that Ms. Larson review these examples and “give me a call to discuss before I request programming changes. My supervisors are interested in this because it will be adding time to offender’s sentences.” 17 Ms. Stigall then concluded her e-mail with the statement “I also find that it is making a change regardless if this is a short base sentence or not.” This information was contrary to Ms. Larson’s belief that the programming error affected only those offenders who were serving an enhanced sentence followed by a short base sentence. At 5:37 p.m. Ms. Larson responded to Ms. Stigall’s e-mail and advised Ms. Stigall as to which of her examples was the most “desirable from a policy perspective.” (E-mail attached as Exhibit 21). Ms. Larson also copied Paul Weisser on her response. Mr. Weisser told us that he has no recollection of these e-mails and recalls no conversation with Ms. Larson about the matter raised in the e-mails. On December 27, 2012, Ms. Stigall prepared an Information Technology Service Request, which is commonly referred to as a “change request” or “CR.” (Exhibit 22). A CR is the first step that must be taken to request IT’s assistance in fixing a programming error. Ms. Stigall’s request stated that it was based on the advice of an Assistant Attorney General. This should have alerted the IT group that this was an important request. Ms. Stigall also explained that “the application of jail credits in OMNI when there is a mandatory/enhancement that are being served as flat time needs to be changed. The current programming is allowing more than the maximum amount of good time to be applied to the base sentences.” Ms. Stigall noted that this problem needed to be addressed “ASAP This needs to be a Records/SSTA priority [because] all current [release dates] when there is a mandatory/enhancement are in error.” 17 This is further corroboration that Ms. Stigall had raised the issue with Denise Doty and Kathy Gastreich. - 21 - Ms. Stigall sent her Change Request to two individuals. At 7:27 a.m. she sent the CR to the DOC IT Gatekeeper (Exhibit 23). Several seconds later. she sent the request to IT Business Analyst Sue Schuler. (Exhibit 24). At 10:14 a.m. Ms. Stigall sent four documents to Humberto Holguin who worked with the IT gatekeeper. (Exhibit 25). These documents included the December 7, 2012, e-mail from Ronda Larson to Wendy Stigall and the exchange of e-mails between Ms. Stigall and Ms. Larson on December 26, 2012. Mr. Holguin then entered Ms. Stigall’s CR into a DOC software program called Service Desk Express (“SDE”). (SDE printout attached as Exhibit 26). SDE tracked the progress of a CR and allowed a business user to determine the status of a CR. Mr. Holquin also added the four documents provided by Ms. Stigall to the SDE file. It is our understanding that some DOC employees outside the IT group could access SDE to check on the status of any pending CR. It is unclear whether DOC employees regularly used the SDE tracking tool to follow the progress of CR items. Once logged into SDE, Ms. Stigall’s CR was referred by the IT Gatekeeper to the IT Triage Team. (Exhibit 27). This team was supposed to assess each CR to determine whether it would be approved for work by the IT group. The Triage Team reviewed Ms. Stigall’s CR on December 31, 2012, and apparently they did not have enough information to approve the CR. As a result, the CR was referred to the IT Business Unit for a “consultation.” In this case, David Dunnington, the IT Business Unit Manager, directed Sue Schuler to conduct the consultation January 2, 2013 On January 2, 2013, Ms. Stigall attended a weekly ASD Management Team Meeting. Denise Doty, CIO Doug Hoffer, Budget Director Sarian Scott, and several other ASD managers attended this meeting. (Minutes attached as Exhibit 28). Ms. Stigall said that she raised the early release problem during this meeting. She said this was reflected in the minutes as “Wendy said they are working on the new mandatory sentencing enhancements.” Neither Ms. Doty nor Mr. Hoffer has any recollection that Ms. Stigall raised the issue during this meeting. We do believe that Ms. Stigall did in fact raise the issue during this meeting, because on January 2, 2013 at 2:06 p.m. she sent an e-mail to Clela Steelmaker, the DOC Legislative Policy Director stating “I was talking about this programming change at Denise’s direct report meeting and Sarian asked if I had shared it with you. Please review and give me your input.” (E-mail attached as Exhibit 29). Ms. Stigall attached Ronda Larson’s December 7, 2012, memorandum to this e-mail. Sarian Scott was the ASD budget director and she had attended the January ASD Management Team Meeting. This e-mail corroborates Ms. Stigall’s recollection that she raised the early release matter with the ASD management team. Clela Steelhammer had no recollection of this e-mail. Ms. Steelhammer would have been a logical person to have disclosed this problem to the DOC upper management. She was DOC’s liaison with the legislature and dealt with all new legislation affecting DOC, some of which brought about changes in SSTA. It is reasonable to conclude that this error in calculating offender release dates should have drawn her attention, because it, too, raised an SSTA issue. She offered as a possible explanation for having no memory of this e-mail that it was at the beginning of that year’s legislative session and she was very busy. - 22 - Ms. Steelhammer participated in weekly Executive Staff meetings and Senior Leadership Team meetings, which were attended by the DOC secretary, all of the assistant secretaries, and other high-level DOC officials. It does appear that Ms. Steelhammer did in fact raise the SSTA issue at least in a cursory manner to at least some senior managers. Sandy Mullins, who worked as the DOC Director of Executive Policy, also attended the weekly Executive Staff and Senior Leadership Team meetings. Ms. Mullins told us that she thought she recalled that Clela Steelhammer told her either before or after one of these meetings that there had been a problem with an inmate being released early, but the matter had been addressed by the Attorney General’s office, and implied that the problem had been fixed. Secretary Warner also told us that he had a vague recollection that someone, who he could not identify, had mentioned that a prisoner had been released early, but he believed it was an isolated problem which had been fixed. January 3, 2013 On January 3, 2013, DOC employee Geoffrey Nelson provided Ms. Stigall with a list of all inmates serving mandatory and enhanced sentences. That list included the names of 2,786 offenders and each inmate’s projected release date. (List attached as Exhibit 30). Ms. Stigall admitted that she did not share this list with anyone else at DOC. DOC subsequently provided us with a list of these inmates, sorted based on their projected month and year of release. (Exhibit 31). Based on our review of the list of inmates and their projected release dates, we were able to determine that 25 inmates were to be released in January 2013, 23 in February 2013, 31 in March 2013, and 24 in April 2013. To the extent that Ms. Larson and Ms. Stigall believed that it would have taken a “few months” to fix the programming error in OMNI, it would not have been overly burdensome for DOC to hand calculate the release dates for this relatively small group of offenders. Ms. Stigall told us that she did not consider hand calculating the release dates for only those offenders who were projected to be released during the period while the programming problem was being fixed in OMNI. January 9, 2013 On January 9, 2013, Ms. Stigall attended another ASD Management Team meeting. Among those attending were Denise Doty, Doug Hoffer, Brian Tinney, Sarian Scott, and Kathy Gastreich. Ms. Stigall told us that she again raised the early release issue during this meeting citing to the minutes which reflect that “Wendy said that they are still working on the mandatory sentencing enhancements.” (Minutes attached as Exhibit 32). Neither Ms. Doty nor Ms. Gastreich had any recollection of Ms. Stigall raising the issue during this meeting. 18 Doug Hoffer has told us that he has some recollection that Ms. Stigall did raise the issue about inmates being released early. He did not know when she raised the issued but he thought it could have been during one of the ASD weekly management meetings. 18 There is further evidence that Ms. Stigall raised the early release issue with Ms. Doty. On February 26, 2013, Ms. Stigall sent an e-mail to Ms. Doty listing her work goals for 2013. One of the goals listed was “Programming for King Decision/Jail Credits.” (Email attached hereto as Exhibit 33). Ms. Doty had no recollection of this e-mail. She did admit that she would have met with Ms. Stigall to review her goals for 2013 but Ms. Doty had no recollection of discussing the King programming matter with Ms. Stigall. - 23 - D. January 3, 2013 - March 25, 2013: Sue Schuler conducted an IT Consultation regarding the King programing fix. After Ms. Stigall submitted her IT Change Request on December 27, 2012, David Dunnington directed Sue Schuler to conduct an IT Consultation. This occurred on January 3, 2013. (Exhibit 26). A consultation is supposed to be conducted promptly, typically in a week or two. The purpose of the consultation is to determine the nature of the problem, the solution desired by the business user, and an estimate of the time it will take to fix the problem. As part of the consultation process, Ms. Schuler met with Wendy Stigall, who claims that she told Ms. Schuler that the programming problem was serious and that it needed to be given a high priority and fixed as soon as possible. Ms. Schuler does not have the same recollection. Ms. Schuler believed that the programming error affected a relatively small number of inmates who were serving enhanced sentences followed by a short base sentence. Ms. Schuler did not complete the consultation until March 25, 2013 – almost three months after it was first assigned to her. (Exhibit 35 at GOV_000075). Ms. Schuler maintains that during this three month hiatus she was attempting to gain clarification from Ms Stigall as to nature of the problem and the specifications for programming the required change. She asserts that there were numerous e-mails between Ms. Stigall and her relating to this consultation. Although we have requested all e-mails between Ms. Stigall and Ms. Schuler during the relevant period, we have found none other than the afore-mentioned December 7 e-mail and the December 27, 2012, e-mail, forwarding a copy of the change request to Ms. Schuler. Ms. Stigall denies that she was ever asked for specifications and said that she provided Ms. Schuler with all of the information she requested. 19 Ms. Schuler told us that she had still not received the necessary information from Ms. Stigall at the time she prepared her IT Consultation Form. Ms. Schuler said that she prepared this form knowing that she was lacking information she needed. Ms. Schuler’s explanation is dubious. Her IT Consultation form does not indicate that the business requirements were incomplete because she had not received critical information from Ms. Stigall. Moreover, Ms. Schuler’s supervisor, David Dunnington, told us that it would have been inappropriate for Ms. Schuler to have submitted an IT Consultation Form until she had received all necessary information from the business user. In the Consultation Form, Ms. Schuler recommended that Sierra Cedar, the outside contractor which held the maintenance contract for OMNI, be used to complete the programming work. We were told repeatedly that the “go-to” person within Sierra for all SSTA programming needs was Mark Ardiel. 19 Moreover, the consultation form completed by Ms. Schuler provides only a general description of the problem with no specifications for the programming. Ms. Schuler periodic reports to Mr. Dunnington during this period only note that she had conferred with Steve Collins and Ms. Stigall regarding the calculations of jail time credit and sought work estimate for this change request. Exhibits 34. - 24 - Ms. Schuler forwarded her consultation form to Mark Ardiel and to Deepak Sadanandan, an IT employee who supervised the IT testing team for all OMNI programming. She asked each of them to provide her with an estimate of the time required to complete the programming and testing stages for this change request. (Exhibit 35). On March 28, 2013, Mr. Ardiel responded with an estimate of twenty hours to complete the programming work. (Exhibit 36). On April 1, 2013, Mr. Sadanandan responded with an estimate of 16 hours to complete the testing. (Exhibit 37). Once the consultation had been completed, the change request was returned to the Triage Team for its approval. For this request, the approval to move forward was given on April 3, 2013. (Exhibit 38). E. April 3, 2013 - November 3, 2015: The King fix was pending in IT and delayed 16 M-releases without any significant progress. 1. April 3, 2013: Sue Schuler logged the King fix in Clearquest and scheduled it for M34, which had a release of September 13, 2013. On April 3, 2013, Sue Schuler opened a file for the approved change request in Clearquest, the DOC database that tracks OMNI programming changes. The project was assigned a Clearquest ID number (“OMNI00024910”) and became part of the OMNI programming queue. Schuler assigned the project the name “CR#6307 SSTA – Application of Jail Credits per King Decision.” (Clearquest report attached as Exhibit 39). Ms. Schuler originally gave the King programming fix a severity rating of 2 (“Serious Impact”), (Exhibit 39 at TN_000346), but in February 2014, Mr. Dunnington demoted the project to 3 (“Moderate Impact”), (Exhibit 39 at TN_000341), and the project remained classified as severity 3 until it was completed in January 2016. Ms. Schuler also scheduled the King programming fix to be included in M34, with a release date of September 12, 2013. (Exhibit 6). Ms. Schuler said she selected M34 because the two earlier M-releases (M32 and M33) were already filled up. 20 Ms. Schuler’s explanation is hard to square with the fact that on April 12, 2013, nine days after she opened the King fix in Clearquest, she opened another item in Clearquest to fix a programming error related to good time calculations for offenders who engaged in persistent misbehavior (“PM”). Ms. Schuler also gave the PM programming fix a severity 2 rating. (PM Clearquest report attached hereto as Exhibit 41 at IF_000953). Ms. Schuler scheduled the PM project for an earlier release, M33, which was scheduled for release on July 18, 2013. 21 (Exhibit 6). 20 Schuler added as background to the Clearquest file for the King programming fix the IT Service Request, the IT Consultation Form, and a spreadsheet illustrating three examples of how good time should properly be calculated. (Attached as Exhibit 40.) 21 It is worth noting that Wendy Stigall submitted the IT change request to fix the PM problem on February 7, 2013. She identified the problem as “OMNI is currently not programmed to apply Persistent Misbehavior (PM) sanction losses correctly, causing errors in release date,” (Exhibit 42). This matter was referred to Sue Schuler to conduct an IT consultation, which she apparently completed by February 14, 2013. (Exhibit 43). Ms. Stigall has told us that the PM error “cut both ways” in that in some cases it resulted in earlier releases date and at other times offenders received release dates beyond what they should have served. We were advised by Julie Martin, ASD Assistant Secretary, that this programming error did not result in the early release of any inmates. - 25 - On April 3, 2013, approximately two hours after Schuler created a file for the King programming fix in Clearquest, David Dunnington logged the fix as “Required for” M34. (Exhibit 39 at TN_000345). In fact, the fix was not implemented until M49, which was implemented on January 12, 2016. As summarized below, between April 3, 2013 and October 7, 2015, IT staff (most often, David Dunnington) pushed the programming fix out to later M Release dates and ultimately delayed the fix by sixteen M-releases. These changes often coincided with regularly scheduled OMNI team meetings, and sometimes the changes were recorded in Clearquest during the scheduled meeting times. Although users are supposed to enter notes in Clearquest explaining status changes, with rare exceptions, no explanations or justifications were entered for the delays in the King project. June 12, 2013 On June 12, 2013, David Dunnington reassigned the King programming fix from M34 to M35 without recording any explanation for the change. (Exhibit 39 at TN_000345). The effect of this change was an eight-week delay: M34 had a release date of September 12, 2013, whereas M35 had a release date of November 7, 2013. (Exhibit 6). David Dunnington entered the change at 1:55 p.m., the tail end of the regularly scheduled OMNI team meeting. There are no entries in Clearquest indicating whether any work was done on the project between April 3 and June 12, nor have we located any e-mails indicating that any activity took place during this period. Mr. Dunnington told us that he could not recall why he made this change. 2. July 2013 - September 2013: While the King fix was pending, Wendy Stigall oversaw a statewide audit of jail good time calculations. In June, 2013, the Washington State Legislature passed Senate Bill 5892, which revised certain jail good time rules, and gave DOC the authority to calculate maximum jail good time. The bill was retroactive, and in response DOC conducted an audit of jail good time calculations for all sentences. Wendy Stigall oversaw the jail good time reviews. She distributed six pages of detailed instructions to records staff on how to audit jail good time. (Exhibit 44). The audit was to be completed by September 1, 2013. In a July 22, 2013 e-mail, Ms. Stigall notes with respect to the September 1 deadline, “This deadline was not set by the legislature but that deadline was set to ensure that any offenders who may be eligible to release would be released on the earliest possible date to meet the projected savings.” (Attached as Exhibit 45). Ms. Stigall’s instructions make no mention of the fact that, for prisoners with enhancements, jail good time was being improperly deducted off of the base sentence, which resulted in incorrect early release date calculations. July 10, 2013 On July 10, 2013, Wendy Stigall sent an e-mail to Sue Schuler identifying various IT fixes which she “would like to see prioritized.” Included on this list was the King fix identified as number 24910. (Exhibit 46). - 26 - August 7, 2013 On August 7, 2013, Mr. Dunnington moved the King programming fix to M36 which was scheduled for January 10, 2014. (Exhibit 39 at TN_000344). Mr. Dunnington told us that it was his practice to clear every item out of an upcoming M Release and then request that each IT business analyst provide him with a list of items which they believed should be included in the empty M Release. Consistent with this practice, on August 7, 2013, Mr. Dunnington sent an email to Sue Schuler and the other business analysts requesting that they provide him with a list of items each analyst would like to see included in the upcoming M35 release. (Exhibit 47 at GOV_000190). August 8, 2013 On August 8, 2013, Sue Schuler sent Mr. Dunnington an e-mail requesting that item 24910 “Application of Jail Credits per King Decision” be included in the M35 Release. (Exhibit 47 at GOV_000189). Based on this request, Mr. Dunnington moved the King programming fix back to the M35 release (Exhibit 39 at TN_000344), scheduled for November 7, 2013. (Exhibit 6). August 15, 2013 On August 15, 2013 Wendy Stigall held a meeting of all DOC records managers. The minutes of that meeting, which are attached hereto as Exhibit 48, reflect that Secretary Bernie Warner and Assistant Secretary Denise Doty were present at the beginning of this meeting. The minutes further reflect that Ms. Stigall advised the group about the King Decision fix, stating that King Decision: I sent an e-mail on August 19, 2013 explaining the calculation issues under the King Decision. Programming will be implemented that will correct this error. (Exhibit 48 at 000226). Ms. Stigall also made a PowerPoint presentation during this meeting. (Exhibit 49) One of the slides addresses the fix to the King decision stating that, “It was recently discovered that the programming of offenders with mandatories and enhancements is not correct. The offenders are getting too much good time.” (Exhibit 49 at 000462). According to Ms. Stigall, Secretary Warner and Assistant Secretary Doty only stayed for a brief portion of the meeting and neither was present when she addressed the King issue. Ms. Doty also told us that she and Secretary Warner only spent a short amount of time at that meeting and left well before Ms. Stigall began discussing the King issue. September 18, 2013 On September 18, 2013, Sue Davis added a document with the file name, “King Decision example from Wendy Stigall” to the Clearquest file. (Exhibit 50). In it, Stigall illustrates how, by deducting total jail good time from the base sentence, OMNI gave prisoners excess good time. In the example, the prisoner has a base sentence of 20 months, which – as Stigall notes – - 27 - should carry a maximum good time of 202 days (based on a maximum earned release percentage of 33 1/3). As Ms. Stigall explains, the problem occurred because OMNI subtracted the jail good time from the base sentence – in the example, OMNI credits the prisoner with 231 days of jail good time, which is 29 days more than the maximum good time the offender should be able to receive. After deducting the jail good time from the base sentence, OMNI then calculated additional earned release time at 33% on the remaining 377 days, leaving the offender with only 252 days to serve instead of 406, and a total earned time percentage of 58, well in excess of the statutory maximum of 33 1/3. September 27, 2013 On September 27, 2013, Mark Ardiel, Sierra Cedar’s lead programmer, first logged into the King Clearquest file (Exhibit 39 at TN_000343). Mr. Ardiel’s entry describes the programming changes in general terms (“We will need to change the sentence calculations so that it splits out the application of the Jail Credits, so that it can apply them in a different order”) and repeats instructions as to the order in which credits should be applied, which he had received by e-mail from Stigall the previous day. There is no indication in this entry that he needs any further instructions, or that the project is larger in scope or complexity than originally forecast. The day before he made this entry, Mr. Ardiel and Ms. Stigall had exchanged brief e-mails about the fix. (Exhibit 51). September 30, 2013 On September 30, 2013, Deepak Sadanandan, a DOC Supervisor in the Application Development Team, pushed the project to M37 (with a release date of March 16, 2014), as a “MUST FIX.” (Exhibit at 39 at TN_000343). The “MUST FIX” designation is significant. Projects identified as “MUST FIX” were supposed to be given priority over projects not so designated, and “MUST FIX” projects were separately tracked for each M-release. Mr. Sadanandan also added a note, “fix this CR [change request] first, before PM.” PM or “Persistent Misconduct” was a separate SSTA programming fix, which was logged in Clearquest on April 12, 2013, 9 days after King. Like King, PM was assigned to Mark Ardiel. (Exhibit 41 at IF_000945). In the end, King was not fixed before PM: PM was implemented on March 24, 2014, nearly 22 months before the King programming issues were resolved. (Exhibit 41 at IF_000943). We asked Mr. Dunnington why the PM item was fixed first and he could offer no explanation. October 1 - 8, 2013 On October 1, 2013, the day after Mr. Sadanandan had characterized the project as “MUST FIX,” Mr. Dunnington deleted the “MUST FIX” designation and reassigned the King programming to M38. (Exhibit 39 at TN_000342). On October 8, 2013, Sue Schuler sent Dunnington an e-mail with a list of SSTA items to include in M37. This list includes the King programming fix. (Exhibit 52). That same day, Mr. Dunnington moved the project back to M 37, but without the “MUST FIX” designation. (Exhibit 39 at TN_000342). We have asked Mr. Dunnington to provide us records relating to any other CQ projects in which he deleted the “must fix” designation. To date he has not provided this information. - 28 - November 13, 2013 On November 13, 2013, David Dunnington made an entry into the SDE tracking system stating that the King fix had been “implemented” and as a result SDE reflected that the King fix had been “completed” and then “closed” on that day. (Exhibit 26). This was not true. Mr. Dunnington has told us that he inadvertently made this entry. He speculated that he had intended to close another item but had erroneously entered the item number for the King fix. 3. November 27, 2013: Mark Ardiel first worked on the King programming fix. On November 27, 2013, more than eight months after Sue Schuler first opened the Clearquest file for the King project, Mark Ardiel logged his second entry. Mr. Ardiel changed the project state from “Submitted” to “Assigned,” and assigned himself as the project owner. (Exhibit 39 at TN_000342). Mark Ardiel would remain the assigned project owner until December 3, 2015, when the project owner was changed to David Gale, a tester in DOC’s IT group. (Exhibit 39 at TN_000336). Mr. Ardiel told us that he worked on this project over the course of the next month or so. He did not keep a log of the dates he worked on the project; he estimated, however, that he spent about 80 to 100 hours working on the project. There is no entry in Clearquest reflecting that any work was done on this project. He told us that he thought his work was almost complete but that he had run into a few problems. He said that he needed additional information from Wendy Stigall and that she was very busy and difficult to get a hold of. Mr. Ardiel eventually directed his efforts elsewhere and he was not told to complete this project. There are no notes reflecting why he discontinued work on the item or suggesting that he was waiting for additional information. January 29, 2014 The next Clearquest entry was not until January 29, 2014. (Exhibit 39 at TN_000341). At 1:00 p.m., the scheduled start of the Wednesday afternoon OMNI meeting, Mr. Ardiel changed the status from “Assigned” to developing code (“Dev Code”) – a status that is supposed to indicate that the project is being actively worked on. Five minutes later, Deepak Sadanandan moved the fix to M38 as a “MUST FIX,” without any explanation for the reassignment. M38 had a release date of May 1, 2014 (Exhibit 6) – more than a year after the King programming fix had first been logged into Clearquest. February 3 - 4, 2014 On February 3, 2014, Mr. Dunnington deleted the “MUST FIX” designation and moved the project to M39. (Exhibit 39 at TN_000341). The next day, after Sue Schuler sent Mr. Dunnington an e-mail identifying SSTA items for M 38, which included King programming, (Exhibit 53), Mr. Dunnington moved the project back to M38 without the “MUST FIX” designation. (Exhibit 39 at TN_000341). - 29 - The same day, February 4, 2014, Mr. Dunnington changed the severity level of the King programming fix from 2 (“Serious Impact”) to 3 (“Moderate Impact”). (Exhibit 39 at TN_000341). Mr. Dunnington did not record any justification for the diminished severity level. Mr. Dunnington claimed to us that the change to severity level 3 was made as part of a “policy change” that all enhancements were to be uniformly classified as severity 3. The only support for this explanation comes from a November 21, 2013, e-mail from Mr. Dunnington to the architecture team, which attached an updated draft of the “Defect/Enhancement Severity Level Definitions.” The document included the following note: “Enhancements: All System Enhancements entered into CQ [Clearquest] should be assigned Severity Level 3 for consistency.” (Exhibit 54). Mr. Dunnington’s e-mail notes that the document is “[f]or review on Friday, 11/22[/2013].” It is unclear whether or when this policy was implemented, or whether the change in severity Mr. Dunnington made to the King programming fix two and a half months later reflected any such policy change. 22 We asked Mr. Dunnington if he could recall any other instances in which he reduced the severity level on an enhancement from a level 2 to a level 3. He could not recall but he said he would check to see if he could find any other instances in which that happened. He has not provided us with any other instances in which he made a change in severity level. March 28 -31, 2014 On Friday, March 28, 2014, Mr. Dunnington sent Ms. Schuler an e-mail asking whether 24910 could be moved to M 39. (Exhibit 55). Ms. Schuler responded early the following Monday, March 31, 2014, “If we have to.” Mr. Dunnington replied, “Are there any big concerns? Will Wendy be ok with it?” Ms. Schuler responded, “If she has to be – I talked to her about it today.” (Exhibit 55 ). That same day, Mr. Dunnington moved the King programming fix to M39 as a “MUST FIX.” (Exhibit 39 at TN_000340). He included a note in the entry: “Updated to M39 due to code freeze.” “Code freeze” deadlines occur roughly one month before an M-release. Defects and enhancements that have not been successfully coded prior to the code freeze are relegated to later M-releases. There is no indication of the status of the King programming fix as of the time of the M38 code freeze. May 20, 2014 On May 20, 2014 – roughly one month before the scheduled release for M-39 – Mr. Sadanandan pushed the project to M40, “as per Sue Schuler’s request.” He also deleted the “MUST FIX” designation. (Exhibit 39 at TN_000340). July 16, 2014 On July 16, 2014, Mr. Dunnington moved the project to M41, (Exhibit 39 at TN_000340), which had a release date of October 16, 2014. (Exhibit 6). There is no 22 Ira Feuer, the current CIO, told us that it made no sense to treat every enhancement as a severity level 3. He pointed out that depending on the nature of the enhancement it could easily be categorized as either a severity level 1 or 2. - 30 - explanation in ClearQuest for the delay. Mr. Dunnington logged this change just after a scheduled OMNI meeting. September 3, 2014 On September 3, 2014, nine minutes into the regularly scheduled Wednesday afternoon OMNI meeting, Mr. Sadanandan reassigned the project to M42, and again classified the project as a “MUST FIX.” The only explanation recorded in Clearquest is, “Moved to M42 per OMNI meeting.” (Exhibit 39 at TN_000339). M42 had a release date of December 11, 2014. (Exhibit 6 ). September 11, 2014 On September 11, 2014, Mr. Dunnington moved the project to M 43, without the “MUST FIX” designation. (Exhibit 39 at TN_000339 ). No explanation is given for the delay and there is no indication that any work had been performed on the project. M43 had a release date of February 5, 2015. (Exhibit 6). September 26, 2014 On September 26, 2014, Wendy Stigall attended a records meeting at DOC headquarters with Records Management Supervisors, Correctional Records Supervisors, and Correctional Records Technicians from prison facilities and regional records offices. The minutes (attached hereto as Exhibit 56) reflect that Ms. Stigall presented a list of outstanding IT requests that the records department was waiting on. The “King Decision” is listed as one of nine items, and is described as, “Programming for jail credits and jail good time when there is a mandatory or enhancement.” November 2014 - December 2014 Mr. Dunnington twice more delayed the King programming fix – on November 4, 2014, (Exhibit 39 at TN_000339), and December 22, 2014, (Exhibit 39 at TN_000339). The first two times he pushed the fix back a single release (to M43 and then M44, respectively). M44 had a release date of April 2, 2015. 4. February 2015 - September 2015: Mark Ardiel took paternity leave and the project was delayed several M-releases. Mark Ardiel began paternity leave on February 19, 2015. One week later, on February 26, 2015, Mr. Dunnington pushed the project out two M-releases, to M47, (Exhibit 39 at TN_000338) which had a release date of September 17, 2015. (Exhibit 6). Mr. Dunnington did not characterize the project as a “MUST FIX.” Notably, Mr. Ardiel originally was scheduled to return from paternity leave on July 1, 2015. Mr. Ardiel ultimately extended his leave through September 1, 2015. On April 24, 2015, Mr. Dunnington reassigned the project to M50, likely as a result of Mr. Ardiel’s extension of his paternity leave. (Exhibit 39 at TN_000338). M50 had a release date of March 3, 2016. (Exhibit 61). - 31 - It is important to note that Sierra Cedar had several other programmers available who could have worked on the King fix while Mark Ardiel was on paternity leave. However, neither Mr. Dunnington nor any other DOC employee requested that Sierra assign a programmer to work on the King fix during Mr. Ardiel’s absence. May 26, 2015 As explained above, DOC tracked the progress of IT Change Requests through a system called Service Desk Express (“SDE”). DOC replaced this tracking system with a new system called Easy Vista. On May 26, 2015, Sue Schuler made an entry into Easy Vista reflecting that she was “Closing this incident [item 24910] as the fix has been implemented in OMNI.” (Easy Vista entry attached hereto as Exhibit 57). This was not correct, as the King fix had not even been addressed by this date. Ms. Schuler has acknowledged that she made an erroneous entry. She claims that this entry was made inadvertently and that she had intended to close another item but had mistakenly entered item number 24910. We must note, however, the somewhat improbable coincidence that both Mr. Dunnington, (who closed the item in SDE on November 13, 2013, Exhibit 26), and Ms. Schuler made inadvertent entries into the DOC tracking system incorrectly reflecting that the King fix had been implemented. September 1, 2015 Mark Ardiel returned to work on September 1, 2015. Mr. Ardiel was not directed to work on the King fix but instead began working on other matters. October 7, 2015 On October 7, 2015, Tonya Christen, another IT Business Analyst, moved the project back to M49. (Exhibit 39 at TN_000337). M49 had a release date of January 12, 2016. There is no explanation for this change. On December 7, 2015, Christen reclassified the project as “M49_must fix.” (Exhibit 39 at TN_000336). F. November 2015: Work finally resumed on the King Fix. November 2, 2015 After the project had languished for almost thirty months, IT finally began to address the issue in November 2015. A meeting on November 2, 2015, between the new CIO, Ira Feuer, and Wendy Stigall, appears to have instigated this work. As the new CIO, Mr. Feuer was meeting with a number of ASD employees to determine their IT needs. During his meeting with Ms. Stigall, she told Mr. Feurer that she had an old IT Change Request which IT had not fixed in over three years. Ms. Stigall told Mr. Feuer that the Change Request related to “offender release dates.” Mr. Feuer also noticed that Ms. Stigall had written these words in red on a whiteboard in her office. Ms. Stigall did not explain the nature of the problem to Mr. Feurer, nor did she tell him that offenders were being released early. Mr. Feuer told Ms. Stigall that he would look into the matter for her. - 32 - Mr. Feuer then met with David Dunnington, also on November 2. Mr. Dunnington assured Mr. Feuer that the problem was being worked on. Mr. Feuer also met with Jay Ahn, the lead IT programmer, who also assured him that the matter was been worked on. As far as we can determine, this matter was not being addressed as of November 2. It appears that work began on November 3, 2013. November 3, 2015 Sue Schuler sent David Gale and Wendy Stigall a meeting request for November 3, 2015, from 8:00 a.m. to 9:00 a.m. The subject line of the request was “King Decision,” and the body of the request contained a single word in all-caps: “HELP”. (Exhibit 58). At 8:53 a.m. on November 3, 2015, Wendy Stigall sent Sue Schuler and David Gale an e-mail with the subject line “King Decision.” The e-mail reads, “I found what I was looking for as soon as you left. Hopefully some of this will help the confusion.” Ms. Stigall attached to the e-mail a PDF she had e-mailed herself at 7:44 that morning. The PDF includes her e-mail correspondence with Ronda Larson of December 7 and December 26, 2012, and the original work request from July 25, 2002. (Exhibit 59). Thirteen minutes later, at 9:06 a.m. on November 3, 2015, David Gale e-mailed Mark Ardiel, cc’ing Wendy Stigall, Sue Schuler, and Brian Alonso. In his e-mail, Mr. Gale asks Mr. Ardiel if he can provide any information on the changes for 24910 and whether Mr. Ardiel has started coding the changes. Mr. Gale also offers a high-level explanation of the coding change based on his interpretation of the defect “and this morning’s meeting.” (Exhibit 60). Mr. Ardiel responded four hours later, at 1:11 p.m., writing: “I haven’t gotten back into coding the changes for this yet, but will be doing that soon.” (Exhibit 60). November 3 – November 30, 2015 Mark Ardiel was responsible for making the programming changes. During the months of November and December, he estimated that he spent 307 hours to make the fix. Clearquest reflects that he completed his coding on November 30, 2013. (Exhibit 39 at TN_000337). Thereafter the fix was subjected to testing, which initially failed and required considerable additional development work by Mr. Ardiel and others between December 2015 and January 2016. (Exhibit 39 at TN_000323-000336). G. December 2015: DOC Executive Staff and then Governor Inslee learned of the problem. December 11, 2015 Julie Martin became the ASD Assistant Secretary in November 2015. On December 11, 2015, she sent an e-mail to all of her direct reports and told them that she was planning to meet with the head of the Prisons Division, Steve Sinclair, and his team. Ms. Martin wanted to know - 33 - if any of her direct reports had any matters they wanted to address at this meeting. Wendy Stigall responded to this e-mail stating that “I do have a calculation issue that will be affecting prison sentences for offenders, with mandatory enhancements. It will be increasing their sentences.” (E-mails attached hereto as Exhibit 61). Ms. Martin invited Ms Stigall to attend the meeting December 14, 2015 Ms. Martin and Ms. Stigall met with Steve Sinclair and his team on December 14, 2015. During that meeting, Ms. Stigall discussed that she had been waiting for some time for IT to fix a sentencing problem that had resulted in offenders being released earlier than they should have been. Both Ms. Martin and Mr. Sinclair grasped the seriousness of this issue and they decided to bring the issue to the attention of the DOC Senior Leadership Team December 15, 2015 On December 15, 2015, Wendy Stigall attended a DOC Senior Leadership Team Meeting. Secretary Pacholke, the Assistant Secretaries, and other DOC senior administrators were in attendance. (Agenda attached hereto as Exhibit 62). 23 Ms. Stigall briefed the Leadership Team on the programming problem, which had resulted in the early release of offenders over many years. Secretary Pacholke has told us that this was the first time that he learned of the problem, and we have uncovered no evidence indicating otherwise. December 16, 2015: Governor Inslee and his staff learned of the problem. Members of the Governor’s staff learned of the early release issue on December 16, 2015. At approximately 5:30 in the afternoon, Sandy Mullins notified Matt Steuerwalt, the Governor’s Policy Director, of the issue. Ms. Mullins provided Mr. Steuerwalt with a brief description of the problem. Later that evening, she provided Mr. Steuerwalt with additional information. Mr. Steuerwalt said he had no prior knowledge of this issue and our investigation has uncovered no evidence indicating otherwise. Later that evening, Mr. Steuerwalt talked with David Postman, the Governor’s Chief of Staff, and Nick Brown, the Governor’s counsel, and briefed both of them on the issue. Both Mr. Postman and Mr. Brown told us that this was the first time they had learned of the early release issue and our investigation has uncovered no evidence indicating otherwise. December 17, 2015 At 9 a.m. on December 17, 2015, the Governor’s executive staff met. David Postman, Nick Brown, Matt Steuerwalt, and others were present during that meeting. Also present were Jody Becker-Green, DOC Deputy Secretary; Ira Feuer, DOC Chief Information Officer; Wendy Stigall; Alex McBain, DOC Policy Director; Jeremy Barclay, DOC Communications Director; and Tim Lang, Assistant Attorney General. During that meeting the DOC employees provided additional information regarding the history of DOC’s failure to address the early release issue. 23 Apparently no minutes were prepared of this meeting. - 34 - At approximately 5:30 p.m., David Postman, Matt Steuerwalt, and Nick Brown met with Governor Inslee and briefed him on the matter. Governor Inslee has told us that it took him a “nano second” to realize the seriousness of the problem. Governor Inslee stated that he was extremely upset to learn of this problem and directed DOC to fix it immediately to protect the public. Governor Inslee told us that this was the first time he had learned of the problem. Our investigation has uncovered no evidence indicating otherwise. December 18, 2015 On December 18, 2015, Governor Inslee met with DOC Secretary Pacholke. Governor Inslee was quite upset and asked Secretary Pacholke a number of questions, including: how the problem occurred; why it went undetected for years; and how many offenders had been erroneously released. Governor Inslee also asked Secretary Pacholke about the culture at DOC. Secretary Pacholke was not able to provide all of the information requested by the Governor, leading to even more frustration on the Governor’s part. DOC also ordered a manual calculation of sentences of all affected offenders before they were released. December 20, 2015 On Sunday December 20, 2015, at approximately 7 p.m., Governor Inslee and members of his staff, including David Postman, Nick Brown, Joby Shimomura, Kelly Wicker, and David Schumacher went to DOC headquarters 24 to participate in a meeting with Secretary Pacholke and a number of DOC employees, including Dan Pacholke, Jody Becker-Green, Peter Graham, Sue Schuler, Wendy Stigall, Kathy Gastreich, and Alex McBain. Assistant Attorney General Paul Weisser also participated in the meeting. During this meeting, which lasted approximately two hours, DOC provided Governor Inslee and his team with more information relating to the early release issue. At the conclusion of the meeting Governor Inslee directed Secretary Pacholke to bring back into custody those offenders who had been improperly released. December 22, 2015: Governee Inslee publicly disclosed the problem. On December 22, 2015, Governor Inslee held a press conference during which he publicly disclosed that for a number of years DOC had been releasing inmates earlier than they should have been released. (Attached hereto as Exhibit 63 is a copy of the Governor’s press release and related documents which were provided to the media.) H. January 12, 2016: The King programming fix was completed. M49 was released on January 12, 2016. The release report reflects that the King fix had finally been completed. (M49 attached hereto as Exhibit 64). From November 30, 2015 through January 20, 2016, there are 21 separate entries in Clearquest for the King programming fix. These entries reflect that the code initially was 24 Matt Steuerwalt participated in the meeting by phone. - 35 - deemed complete by Mark Ardiel on November 30, 2015. Finally, on January 20, 2016, more than three years after the problem had first been identified, the programming was deemed fixed and the Clearquest file closed. The final entry, input by Schuler on January 20, 2016, reads: “Action: Closed.” (Exhibit 39 at TN_000323). V. A. CONCLUSIONS DOC’s failure to correctly calculate offender release dates was not intentional or malicious. In response to the King Decision, DOC adopted an incorrect formula to compute the release dates for offenders serving enhanced sentences. This computational error was first programmed into OBTS and then carried over to OMNI. For over ten years this problem went undetected, resulting in the early release of thousands of offenders. Once the problem was finally discovered in December 2012, it took DOC over three years to fix the problem. In hindsight, even after acknowledging that sentencing law in this State is a complex maze of frequently amended and modified statutes, it is difficult to understand how this miscalculation went undetected for more than a decade. 25 Clearly, the erroneous formula was applied without further thought until a victim’s family brought the error to DOC’s attention. No evidence has come to light during this investigation that suggests that any individual acted intentionally or maliciously either at the time DOC employees attempted to reprogram OBTS to reflect the Supreme Court holding in the King decision or after the calculation error was discovered in December 2012. Indeed, it is our impression that all of the DOC employees with whom we spoke were dedicated public servants, who labored under large workloads in a stressful environment. As discussed elsewhere in this report, a litany of errors, oversights, and mismanagement was responsible for both the initial miscalculation and the failure to correct the formula when the problem was discovered. Nothing, however, suggests that the root cause of these problems was deliberate or prompted by any ulterior motive. B. The advice tendered to DOC employees by the attorney general’s office was flawed. DOC’s failure to address the early release problem for over three years can be traced to Assistant Attorney General Ronda Larson’s December 7, 2012, memorandum. In that memorandum, Ms. Larson recommended that DOC hand calculate only Mr. Robinson’s sentence and advised that it was not necessary “from a risk management perspective, to do a hand calculation of everyone in prison with an enhancement . . . [pending] the long process of reprogramming OMNI for everyone else.” Ms. Larson candidly admitted in her memorandum that her advice “will result in offenders being released earlier than the law allows for the time being, until OMNI gets fixed, the DOC has been releasing them earlier for a decade (since the In 25 A repeated theme of many DOC employees who submitted comments to the Senate-sponsored website http://fixdoc.org/ in the aftermath of the discovery of the offender early release issue is that the complexity of the sentencing structure in this State, built upon an ever expanding web of legislative changes and modifications to our sentencing laws, makes it difficult to confidently calculate offender release dates. This does not excuse the delay in correcting the calculation error once it surfaced, but it provides some context as to the length of time that passed between the implementation of the King decision in 2002 and the discovery of the miscalculation error in 2012. - 36 - re King decision), and a few more months is not going to make that much difference in light of this (with the exception of Robinson’s case).” (Exhibit 12 at WS_000018). Ms. Larson’s advice regarding other offenders was not a legal opinion, but a business or operational judgment, which DOC should have rejected. Nevertheless, Ms. Larson’s advice undoubtedly carried added weight because it was given by a trusted legal advisor who had been the “go-to” person in the Corrections Division of the Attorney General’s Office for advice on DOC legal matters for at least the prior two years. We consistently have been told by DOC employees involved in this matter that they held Ms. Larson in high regard and relied upon her opinions on this and other issues. Unfortunately, the advice that Ms. Larson tendered to DOC on this occasion was deeply flawed for a number of reasons. First, Ms. Larson based her advice in part on the fact that correcting the early release dates would have required “the large input of personnel resources to do hand calculations of hundreds of sentences.” Her concern is misguided. DOC’s core mission is to protect the public and the early release of inmates puts the public at risk. In this situation, all necessary DOC resources should have been deployed to immediately correct this problem. Moreover, if resources truly were at issue, Ms. Larson should have advised DOC to conduct a piecemeal recalculation of inmate release dates, based upon the date of expected release. Assuming that Ms. Larson’s belief that the coding could be fixed within a “few months” had been correct, she should have advised DOC to initially hand calculate the release date for prisoners projected to be released during the months of December 2012 and January 2013. 26 If the problem was not fixed in January, then DOC could have hand calculated the release date for prisoners projected to be released in February 2013. This process could have continued until the problem was fixed in OMNI. Second, Ms. Larson erred in advising DOC that it could continue to release inmates early until the problem was fixed in OMNI. Ms. Larson clearly understood the potentially catastrophic impact of releasing even one inmate early. She acknowledged as much when she wrote that if Mr. Robinson was to be released early and “immediately go and kill the victim, for example,” DOC would be sued and would lose such a lawsuit. This acknowledgement makes clear that she understood that if DOC released one defendant earlier than the law allows this could potentially cause harm to a victim and significant liability to the taxpayers. The early release of “hundreds” of inmates would greatly increase the odds that innocent people would be victimized and the taxpayers would be compelled to pay significant damages. The only appropriate advice would have been to direct DOC to immediately ensure that no inmates were released earlier than the law allows. Ms. Larson told us that her conclusion was based on her view that her responsibility as an Assistant Attorney General was to provide advice to DOC and represent that Department against claims and lawsuits filed by prisoners. It was not her responsibility to handle tort claims filed by victims against DOC. Nevertheless, Ms. Larson clearly knew that DOC’s primary mission is to protect public safety. Her advice is strangely inconsistent with this reality. Rather than focusing 26 According to DOC records 25 inmates were scheduled to be released in January 2013, 23 in February 2013 and 31 in March 2013. (Exhibit 31). Thus, the task of hand calculating the release dates for this small number of offenders would not have been terribly time consuming. - 37 - on risk, Ms. Larson appears to have relied on an artificial distinction noting, “[T]his is something that DOC has identified internally, rather than something that is being forced upon it by an outside entity such as the court.” Third, Ms. Larson failed to properly assess the magnitude of the problem. Ms. Larson claims that she believed the problem affected only those inmates who were serving an enhanced sentence followed by a short base sentence. As a result, Ms. Larson stated that she believed that only a few hundred inmates would be released early. We are at a loss to understand how or why Ms. Larson reached this conclusion. She clearly understood that the early release problem was the product of a flawed formula used to calculate release dates. If the formula were flawed, then it should have been obvious to Ms. Larson that every inmate serving an enhanced sentence could potentially be the beneficiary of an early release date. Moreover, it would have been reasonable to seek additional information before jumping to conclusions. Wendy Stigall did just this, by requesting a list of all inmates serving enhanced sentences. She received that list on January 3, 2013, and it contained the names of 2,786 prisoners. (Exhibit 30). Ms. Larson now acknowledges that her assumption was unwarranted and that in fact thousands of inmates were released early. Fourth, Ms. Larson also failed to make any effort to determine if DOC actually fixed the problem. Although she had regular dealings with Wendy Stigall, she never asked Ms. Stigall if the problem had been fixed. Ms. Larson claimed that it was not her responsibility to ensure that DOC fixed the problem. While this is technically true, Ms. Larson’s job was to provide sound legal advice to DOC. She had identified a serious problem and advised DOC to fix it. At the very least, it is reasonable to expect that she would have followed through to ensure that the problem was fixed promptly. We must note, however, that Ms. Larson’s memorandum, while deeply, flawed was merely her “advice” to DOC. Although her advice carried considerable weight with DOC, DOC was by no means required to follow that advice. The fact that neither Wendy Stigall, Denise Doty, Kathy Gastreich, or others questioned that advice is troubling. C. Ms. Larson’s Memorandum was not subject to supervisory review. Ms. Larson may also be faulted for her failure to effectively raise this issue with her supervisor in the Corrections Division, Paul Weisser. It is true that Ms. Larson forwarded copies of both her December 7 and December 26 e-mails to Wendy Stigall, together with her memoranda to Mr. Weisser, (Exhibits 12 and 21) but he has no recollection of reading either email at the time. Mr. Weisser told us that he conducted meetings once a month with his team and that there were countless “hallway” encounters in which legal issues of the day were discussed. He did not mention any discussion of this issue at the time, although he would have expected Ms. Larson to run her advice by him. This certainly seems to be in the category of matters that might command supervisory attention. It is also reasonable to question why an experienced supervisor such as Mr. Weisser did not pick up on the matter from the e-mails sent to him. The subject line of both e-mails was “Should DOC reprogram OMNI to run jail time off base rather than off enhancement?/ Robinson#357042.” Arguably for anyone familiar with the King decision – a Supreme court - 38 - decision that was well known in both the Corrections Division and DOC, this subject heading should have conjured up the calculation of release date issues addressed in that Supreme Court opinion. Undoubtedly, Mr. Weisser received countless e-mails on a daily basis, but a supervisor is expected to recognize and separate the “wheat from the chaff.” Here, he had two chances to catch the issue. At the very least, it seems reasonable to conclude that Mr. Weisser in his supervisory capacity might have skimmed one or both of the e-mails and memoranda and been alerted to the problem and the problematic advice. This apparently did not happen. Nor, as already noted, does it appear Mr. Weisser received any oral prompt from Ms. Larson that would have helped to avoid the oversight. Mr. Weisser acknowledged that in retrospect this was a significant matter. If he had talked to Ms. Larson at the time, he would have concluded that DOC was taking appropriate steps to address the problem, but he would have questioned the advice to only hand calculate the release date for one offender. He would have wanted to know how quickly the OMNI fix could have been implemented and how confident DOC was that the fix was correct. Mr. Weisser noted that although he had no specific experience, he thought it was a time-consuming process to correct a programming problem in OMNI. Mr. Weisser stated that he also would have advised Ms. Stigall that triage needed to be completed to determine which offenders would immediately be affected by the calculation error. Because Mr. Weisser apparently did not become aware of the problem, he did not raise it with his supervisor Tim Lang, who heads the Corrections Division in the Attorney General’s Office. With the single exception of Dan Judge’s vague recollection that he raised the issue with Mr. Lang after he was called by Kathy Gastreich, there is no evidence that Mr. Lang knew of the problem. Finally, other than Mr. Weisser, our investigation has uncovered no evidence that anyone else in the Attorney General’s Office was made aware of Ms. Larson’s opinion. D. DOC failed to implement a protocol of hand calculating offender release dates when errors were discovered. Prior to December 2012, DOC had discovered other errors in calculating an offender’s release date. Carrie Fleming, former DOC Records Manager, 27 told us that when these errors occurred, DOC typically hand calculated the release date for that offender and for others similarly situated until such time as either OBTS (more recently, OMNI) was reprogrammed. Although it is undoubtedly the case that most calculation errors affected only a single offender or a small set of offenders, it does not seem unreasonable to conclude that this procedure would be applied no matter how many offenders were affected. This procedure was clearly understood by former Records Manager Carrie Flemming, who stated that particularly during the transition from OBTS to OMNI, there were numerous SSTA errors that needed to be addressed. It is unclear whether Wendy Stigall also understood hand calculation was the default process when errors were discovered. She orally told us that her prior experience with this procedure involved only single offenders. We then received an e-mail from Ms. Stigall stating 27 Ms. Fleming retired from DOC on January 31, 2012, at which time Wendy Stigall became the Records Manager. - 39 - that “the only sentences that I can recall having hand calculated are statutory maximum sentences . . . where there are consecutive relationships and the prison maximum expiration date is longer than the statutory maximum date.” (Exhibit 65). In this case, however, Ms. Stigall relied upon the recommendation of Ronda Larson and only hand calculated the release date for Mr. Robinson when this problem came to light. Admittedly, the hand calculation of release dates for hundreds, if not thousands, of offenders, would have been onerous. 28 As emphasized earlier in this report, however, such calculations could have been approached on a month-by-month basis until OMNI was reprogrammed. This was apparently not done because Ms. Larson, Ms. Stigall, and everyone from senior management who learned of this problem believed that the computer fix would occur within a matter of months. The decision to let even one additional offender leave prison before having completed his court-ordered sentence was problematic. It was compounded by each additional offender who left DOC custody and by each delay of the OMNI programming fix. E. DOC management systematically failed to address the miscalculation of offender early release dates. A key focus of our investigation was to determine which individuals were aware of, or should have been aware of, the early release problem and what, if any, actions they took to address the problem. As a result of our investigation, the evidence supports the following conclusions: 1. Wendy Stigall, Senior Records Manager Wendy Stigall first learned of the problem on December 7, 2012. Thereafter, she did an admirable job of notifying DOC managers and others of the problem. She met with her supervisor, Assistant Secretary Denise Doty, on either December 10 or 11, and notified her of the issue. On December 11, 2012, she e-mailed Ronda Larson’s memorandum to both Ms. Doty and Kathy Gastreich, DOC risk management director. (Exhibit 15). She also requested input from Ms. Gastreich. On January 2, 2013, Ms. Stigall raised the issue at the weekly Administrative Services Divisions meeting, which was attended by Ms. Doty, CIO Doug Hoffer, Brian Tinney, and Budget Director Sarian Scott. (Exhibit 28). Following that meeting Ms. Stigall forwarded Ms. Larson’s e-mail to Clela Steelhammer. (Exhibit 29). Then, on January 9, 2013, Ms. Stigall again raised the issue at the following week’s Administrative Services Division meeting. Ms. Doty, Mr. Hoffer, Ms. Scott, and Ms. Gastreich were all present at this meeting. (Exhibit 32). 28 We have been advised that although the release dates for offenders with relatively straightforward sentences can be calculated by hand in a matter of five to ten minutes, more complex cases can take hours or even most of a day. When DOC set about to hand-calculate all offender sentences after this problem was reported, it assigned numerous DOC employees, who worked day after day for many days, to complete the task. - 40 - Nevertheless, none of these senior management personnel has any recollection of this issue coming to their attention. It is difficult to explain this collective absence of memory. For those who only learned of the problem in the direct report meetings, we do not know what explanation Ms. Stigall gave as to the nature or scope of the problem, or what tone or level of urgency accompanied her report. It is reasonable to assume, however, that at a minimum Ms. Stigall mentioned the early release of offenders and the need to fix OMNI. At the least, Ms. Scott heard enough to suggest to Ms. Stigall that she share the information with Ms. Steelhammer. (Exhibit 29). It is reasonable to conclude that Mr. Hoffer, who was in charge of IT, would have given special attention to this matter because it needed to be addressed by his team. However, it does not appear it received any attention from him. Ms. Stigall should also be credited with having submitted a timely change request to IT on December 27, 2012, to start the process for correcting OMNI. (Exhibit 22). In addition, she should be recognized as the person who, on November 2, 2015, finally identified the problem to Ira Feuer, the new Chief of Information Technology, as he canvassed the DOC personnel for IT needs, thereby bringing about the push to complete work on the programming and testing to correct the SSTA problem. Unfortunately, Ms. Stigall appears to have done little 29 to nothing in the intervening period of almost three years to get this problem corrected. She failed to act even though she had sought and been provided a list of all offenders whose release dates may have been impacted by the miscalculation. From these records, she knew that there were potentially thousands of such offenders. In the face of this information, Ms. Stigall chose to remain silent. On one level, it was not Ms. Stigall’s fault that the change request that she submitted fell into an interminable delay. Ms. Stigall told us that she did not press for action at an earlier time because she assumed that management had decided that other IT projects and change requests should be given priority. She specifically noted that she believed there was a Governance Committee in place that set priorities for IT work. This was a belief shared by others with whom we spoke. The reality, however, as we have explained in our review of the IT process, was very different. Although there had apparently been a Governance Committee, at least to set priorities for larger projects in the past, it had fallen by the wayside long before this problem surfaced. Instead, priorities, at least for most change requests, were set by Dave Dunnington through the OMNI Meetings. On the other hand, Ms. Stigall was the business user who had the greatest interest in seeing that the OMNI correction was made and she was the person who most clearly understood the consequences of not getting the problem taken care of immediately. She “owned” the problem. As an experienced DOC employee with immediate responsibility for ensuring that offenders’ records were correctly maintained, she should have acted sooner. 30 29 We did find an e-mail dated July 10, 2013, from Ms. Stigall to Sue Schuler stating that Ms. Stigall “would like to see [the King fix] prioritized.” (Exhibit 46). Other than that e-mail we found no other records indicating that Ms. Stigall had pushed IT to address the King matter. 30 Ms. Stigall received uniformly high marks for her expertise in handling all facets of offender records. She has a vast experience in SSTA matters. On the other hand, Mr. Pacholke offered the assessment that she had problems “seeing the forest for the trees.” As an exempt DOC employee, she was not subject to mandatory evaluations by her - 41 - 2. Denise Doty, Assistant Secretary, Administrative Services Division There is ample evidence indicating that Denise Doty was aware of the problem and failed to address it. Wendy Stigall met with her on either December 10 or 11, 2012, and told her of the problem, although Ms. Doty has no recollection of this meeting. She received a copy of the Ronda Larson memorandum on December 11, 2012, (Exhibit 15) and had no recollection of it. Ms. Doty attended two ASD management meetings during which Ms. Stigall raised the early release problem, (Exhibits 28 and 32) and again had no recollection of the subject being raised. Then, on February 26, 2013, Ms. Stigall sent Ms. Doty her goals for the year, one of which was “programming for King Decision/Jail credits.” (Exhibit 33). Ms. Doty’s claim that she only had a vague recollection of the problem is belied by Ms. Stigall’s repeated efforts to bring the matter to Ms. Doty’s attention. Ms. Doty was clearly in a position to have been able to promptly address the problem raised by Ms. Stigall. She attended weekly Executive Team and Leadership team meetings, and yet she never advised Secretary Warner and other members of the DOC management team of the problem. She supervised the IT Department and yet she never directed the CIO, who was her direct report, to immediately address the problem. Finally, she supervised the Records Department and yet failed to direct Ms. Stigall to hand calculate offenders’ release dates pending the fix in OMNI. We do not believe that Ms. Doty intentionally failed to take these actions; rather, her failure to act was based on her inability to either grasp or appreciate what Ms. Stigall was telling her. This is especially perplexing because those who have reached management positions should be adept at identifying the most significant issues and, once identified, at ensuring that those issues are addressed. Nonetheless, this problem was overlooked, ignored, or forgotten for more than three years. 3. Kathy Gastreich, Risk Management/Safety Director On December 11, 2012, Wendy Stigall sent Ms. Gastreich Ronda Larson’s memo and requested Ms. Gastreich’s “input” on the issue. (Exhibit 15). Ms. Gastreich has no recollection of this e-mail and claims she was not aware of the problem until very recently. Ms. Gastreich also attended the January 9, 2013, ASD Management meeting (Exhibit 32) during which Wendy Stigall raised the issue. Ms. Gastreich claimed to have no recollection of this meeting. Ms. Gastreich’s failure to recall these events is troubling. Ms. Gastreich is the person directly responsible for risk management. She had a practice of promptly responding to Ms. Stigall’s e-mails. Ms. Stigall’s December 11 e-mail which attached Ronda Larson’s e-mail expressly addresses the risk of early release not only of Mr. Robinson, but potentially hundreds of other inmates. superiors. She did, however receive an evaluation from Denise Doty for the 2012 calendar year which was excellent in all respects. - 42 - 4. Doug Hoffer, Chief Information Technology Officer Doug Hoffer had only a vague recollection of the problem. This is quite surprising on a number of levels. First, as already noted, it is reasonable to conclude Mr. Hoffer should have focused on the issue – and the critical role of IT in correcting it – when Ms. Stigall brought it up during the weekly Administrative Service Division meetings. Second, and perhaps more importantly, Mr. Hoffer should have had a management oversight process in place so that he knew, and those who he supervised in IT told him, that Ms. Stigall’s change request was in the queue and was constantly being delayed. The failure of Mr. Hoffer to be “in the loop” is a significant management failure. 5. Clela Steelhammer, Manager, Legislative and Policy Coordination Ms. Steelhammer was DOC’s primary liaison with the state legislature. It is reasonable to conclude that she would have given particular attention to a systematic failure to comply with a statutory mandate regarding earned release time. One possible explanation for why this issue did not make more of an impression on Ms. Steelhammer and others in senior management is the reaction of Sandy Mullins to Ms. Steelhammer’s mention of the problem in one of their regular Senior Management Meetings. According to Ms. Mullins, she concluded from Ms. Steelhammer’s report that the problem was narrow in scope and had been corrected. Although Ms. Steelhammer had been sent a copy of Ms. Larson’s memorandum, the seriousness of the matter apparently did not register with her and was not conveyed by her to senior management. 6. David Dunnington, IT Business Manager David Dunnington, as the IT Business Manager, 31 bore primary responsibility for repeatedly delaying the IT group from fixing the early release problem. Mr. Dunnington, who began his career with DOC as a correctional officer, clearly understood that OMNI was miscalculating release dates for offenders who were serving enhanced or mandatory sentences. 32 He also understood that this computer error was causing offenders to be released into the community earlier than they should have been. Mr. Dunnington told us, however, that he was not aware of the “magnitude” of the problem until December of 2015. Even if this is true, Mr. Dunnington should have appreciated the fact that the early release of even one offender posed a potential risk to the public. As such, he should have understood that the early release problem needed to be addressed as quickly as possible. Mr. Dunnington’s actions were largely the result of DOC’s deficient system for prioritizing IT fixes. As a result, Mr. Dunnington had the authority to unilaterally decide the timing as to when IT defects and enhancements would be addressed. We are confident that had Mr. Dunnington decided that the early release problem should be fixed as soon as possible, it would have gotten fixed much earlier than it was. Instead, we found that Mr. Dunnington repeatedly delayed the fix. A review of the Clearquest record reveals that on 13 occasions Mr. Dunnington moved the early release problem to a later M Release. On at least two occasions he 31 In early January of this year, DOC promoted Mr. Dunnington to Deputy CIO. 32 During our investigation we did not hear a negative word about Mr. Dunnington. Rather, he was uniformly praised as a competent, dedicated and hard-working public servant. - 43 - removed the “must fix” designation from the project. Although we have asked Mr. Dunnington to provide us with other instances in which he removed the “must fix”designation, he has failed to do so. In contrast, we note that Mr. Dunnington elevated the Persistent Misbehavior matter, to a “must fix” and never removed that designation. (Exhibit 41 at IF_000953). Mr. Dunnington also downgraded the early release item from a Severity level 2 to a Severity level 3. We have also asked Mr. Dunnington to provide us with any other instances in which he reduced the severity level for any other defect or enhancement and he has failed to do so. Again, in contrast, we note that the Persistent Misbehavior item, which was also considered to be an enhancement, was given a Severity Level 2 and never reduced by Mr. Dunnington to a Severity level 3. (Exhibit 41). Most striking is the fact that Mr. Dunnington could not provide us with a reason as to why he kept delaying this project. Although there was a note field in Clearquest in which Mr. Dunnington could have recorded the reason as to why this project kept getting moved to later M Release date, Mr. Dunnington failed to make any entries into the note field. Nor did we find any evidence that Mr. Dunnington ever raised this matter to the CIO level and never sought guidance or input from any senior manager as to its importance. Instead he told us that he relied exclusively on information provided in the OMNI meetings by Sue Schuler, the business analyst assigned to this project. Although this practice may have been defensible for most change requests, the sheer number of delays and bumps in M release dates made it indefensible to continue these postponements ad infinitum, without seeking advice from above. 7. Sue Schuler, IT Business Analyst Sue Schuler also bears some responsibility for delaying this project. Ms. Schuler, like Mr. Dunnington, understood that the programming error in OMNI was causing the early release of offenders. Ms. Stigall also provided her with a copy of Ms. Larson’s memo. (Exhibit 13). At the very least Ms. Schuler should have known that the programming error might cause the release of “hundreds” of offenders. Ms. Stigall also told Ms. Schuler that the programming error was significant and needed to be fixed sooner rather than later. Ms. Schuler apparently failed to grasp the severity of the problem. It took her over three months to complete the IT consultation. Her claim that the delay was caused by Ms. Stigall’s failure to provide her with information is not credible. We have found no e-mail communication from Ms. Schuler requesting additional information from Ms. Stigall. Ms. Stigall has also told us that Ms. Schuler had not asked for additional information. Also, the IT Consultation form completed by Ms. Schuler did not indicate that any information was missing, even though Ms. Schuler told us she had submitted the form even though she was lacking all of the necessary information. Once the item was placed into Clearquest, Ms. Schuler did little to ensure that the item would get fixed promptly. As an IT business analyst, Ms. Schuler’s role in part was to represent the business user’s interest. Ms. Schuler attended the twice weekly OMNI team meetings. As - 44 - Ms. Stigall’s representative she should have pushed the OMNI team to address the early release problem. She failed to do that. F. The IT group lacked a meaningful system for prioritizing work. It is now abundantly clear that the early release problem should have been a top priority for the DOC IT group. Instead, it took over three months for the IT consultation to be conducted and then the problem went unaddressed for over 30 months. This lengthy and inexplicable delay was the result of a deficient process for prioritizing IT work. The overarching flaw in the process was its failure to prioritize IT defects and enhancements based on their impact on public safety. The early release of offenders posed an obvious risk to public safety, and as such, it should have been given the highest priority by IT. This was not the case for several reasons. First, priorities were essentially set by IT representatives. Although the business analyst assigned to a particular change request was to speak for the business user, it is not clear that these analysts truly understood the practical significance of particular items. In this instance, for example, it is certainly not clear from the Consultation Form prepared by Ms. Schuler that this calculation defect potentially impacted hundreds of offenders. This is the situation even though Ms. Schuler had been provided with the Larson e-mail and opinion. Clearly, the interface between the IT business analysts and the business users was not always effective in providing the user’s assessment of need and impact. This failure can reasonably be attributed to several causes: a) the business analysts did not completely understand the importance of the requested change; b) they did not effectively communicate this importance to the OMNI team; and, c) it was overlooked or given too little weight by Dave Dunnington. Second, there does not appear to have been adequate involvement by senior management responsible for business operations in the setting of priorities. Indeed, it is unclear what system was used to set priorities. It appears that Mr. Dunnington essentially had the power to set priorities based on whatever factors he deemed appropriate. Mr. Dunnington volunteered in our initial interview of him, that it was the business analyst who shouted the loudest whose change request was pushed to the top of the heap. This “squeaky wheel” phenomenon is a poor substitute for a logical ordering of work on change requests based on the importance to the Department and impact on the community. Moreover, it appears that there was no reasonable means of blending work on the major projects with the myriad of less demanding change requests. Third, although we were told that items classified as a “must fix” were placed at the top of the current M release given the top priority for completion the pending cycle, this did not seem to hold true for all items. As previously noted, the change request related to correcting the calculation of offender release dates was given a “must fix” classification at least twice, and each time relegated to a later M release cycle with the “must fix” classification deleted. Fourth, there was a breakdown in communications between the IT Prison business unit managed by Dave Dunnington, and the Chief Information Officer and his deputy. Mr. Dunnington and Mr. Hoffer both told us that they met only infrequently. It appears that the CIO essentially abdicated his oversight role to Mr. Dunington who rarely communicated to him the status of work on various project and change requests. Although the CIO and his deputy could - 45 - have consulted Clearquest to learn the status of a specific project, this was no substitute for regular briefings and information sharing on work performed and work delayed. Finally, there was an almost complete failure to enter in Clearquest information explaining work performed on the offender release date defect, and more importantly in this instance, reasons and explanation for the myriad of delays. Without such entries, attempts to historically assess what happened, why it happened, and how it can be changed going forward are doomed to failure. G. Lack of contractor resources did not cause the interminable delay in correcting OMNI’s calculation of early release dates. By all accounts, SSTA defects in the OMNI system were referred to Sierra Cedar, the outside contractor holding the maintenance contract for OMNI. Sierra had been the principal designer of OMNI and commanded a thorough knowledge of its system. Within Sierra, the programmer identified by all as the “go-to” person for SSTA matters was Mark Ardiel. Several people told us SSTA changes required by far the most complex programming of all OMNI functional areas. DOC personnel repeatedly told us that they had confidence in him and went to him more or less by default, especially when there was a concern, as with this defect, that one programming change might affect other aspects of SSTA in OMNI. We were also told, however, both by representatives of Sierra and by DOC personnel that there were four or five other programmers assigned by Sierra to the OMNI maintenance contract and that one or more of these programmers could also have undertaken the offender release date correction. Neither Dave Dunnington, nor anyone else from DOC’s IT group asked Sierra to provide resources in addition to Mr. Ardiel to address this particular defect. Mr. Ardiel also stated that he took his directions as to which change requests to address exclusively from the OMNI meeting and if he had been instructed to concentrate on the early release date problem, he would have done so. He also focused his work on the change request items included in the current M release. No one from DOC contradicted Mr. Ardiel on this point. Even Mr. Ardiel’s unavailability for a seven month period from February to September 2015 while on paternity leave has not been tendered by anyone as a reason for the delay in addressing the offender release date problem. It is significant that during Mr. Ardiel’s absence, DOC never requested that Sierra provide another programmer to address the early release problem. Although Mr. Ardiel and others have acknowledged that the re-programming task to correct the early release problem proved more difficult and time consuming than initially anticipated, none have suggested that the work could not have been completed much earlier if it had been given a higher priority. H. Other IT priorities, inordinately high turnover in DOC management and the IT group, and DOC budget concerns may have compounded the delays in addressing the King decision change request. Although there were some concerns voiced during the investigation that there was an overly aggressive push from senior management to design and implement a more robust offender risk management system, called the “Strong R,” and, implicitly, that this caused delay in addressing the King defect, we did not find solid evidence to support this assertion. The Strong - 46 - R push was primarily centered in the early months of 2014, long after the King decision change request had been submitted. As described to us, this initiative commanded a concentrated focus of management for a period of a few weeks, but it did not require a significant allocation of resources from the IT programmers and testers. Indeed it appears that virtually all of the IT work associated with this project was done by an outside consulting firm, Assessments.com. Assessments.com had no involvement whatsoever in maintaining and upgrading OMNI. The Strong R project may be more relevant in assessing the atmosphere at DOC, and particularly, in the IT group. A consistent theme of many past and present DOC employees was that the Strong R project was a source of friction and dissent within DOC. We were also told that IT personnel felt underappreciated, perhaps in part because of the handling of the Strong R project, and this disgruntlement or other frustrations with the emphasis on this approach to offender risk assessment was a cause of employee turnover. There is no doubt that the IT group specifically, and DOC management generally, suffered significant personnel change, especially during 2014 and 2015. This loss of senior leadership undoubtedly contributed to failed oversight and, perhaps, the constant delays in correcting the King defect. Moreover, it has also been noted that the nation and State’s economic downturn in 2008 and the following years led to significant budget cuts and belt tightening among all governmental departments and agencies. The effect was that the continuing flow of work was done by fewer people who experienced greater stress in the workplace. We cannot conclusively say whether this should be added to the milieu of other factors leading to the three-year hiatus in fixing the King decision defect. It is safe to conclude, however, that the outflow of experienced leaders and IT personnel with years of technical experience and the general dwindling of resources could not have helped the situation. I. Neither DOC secretary Dan Pacholke nor his predecessor, Bernie Warner, was aware of the offender release date issue prior to mid-December 2015. 1. Bernie Warner Bernie Warner told us that he was not aware of the early release problem. He did say that he had a vague recollection that an issue was raised regarding the early release of one inmate but that the matter had been corrected. We found no evidence to contradict this recollection. As Ms. Mullins explained there was nothing about the report that would have alarmed then Secretary Warner or prompted further action on the part of those attending the meeting. We have identified the minutes of one Correction Records Management meeting on August 15, 2013, in which Wendy Stigall presented a series of slides, one of which focused on the defect in allocating the good time credits earned by an offender in presentence detention in local jails and its impact of offender release dates. These minutes reflect that Bernie Warner and Denise Doty were present for some part of the meeting. Ms Stigall has stated that the visit by Mr. Warner and Ms. Doty could be characterized as a brief “meet and greet” and they did not stay for her slide presentation. This was confirmed by Mr. Doty. No other evidence suggests that Mr. Warner knew of the problem - 47 - 2. Dan Pacholke Dan Pacholke was the Assistant Secretary for the Prisons Division when this problem surfaced in December 2012. He has no recollection of the problem coming to his attention at the time. When we met Dan Pacholke, who until his resignation on February 6, 2016, was the Secretary of DOC, he opined that there was a sense of apathy within the department brought about by his predecessor, Bernie Warner. 33 He claimed that Mr. Warner was primarily focused on national penology issues and was frequently away from DOC headquarters speaking at national and international events. Whether this is an accurate description of the mindset of DOC employees during this period, and whether, if accurate, it has anything to do with the failure to address the early release of offender issue in a timely fashion, is unclear. No one who had involvement with this particular problem with whom we have spoken, suggested that this was a cause for either the original error or its perpetuation over the years. We have been advised by at least one DOC employee who worked under Mr. Pacholke when he headed the Prisons Division, that there was a “toxic atmosphere” at DOC and this made employees reluctant to speak up. This employee identified Dan Pacholke as the principal problem. This employee gave one anecdotal example where Mr. Pacholke, as the Director of the Prisons Division, reacted negatively to anyone contradicting him, thereby chilling the willingness of employees to speak up. A second employee told us that Mr. Pacholke favored DOC employees who had Prisons’ experience and was dismissive of others. Again, no one with whom we have spoken who had involvement with this particular problem offered these factors as an explanation for why word of the problem or the repeated delays in fixing the problem was not passed up the ladder to top management. J. Neither Governor Jay Inslee nor members of his staff was aware of the offender release date issue prior to mid-December 2015. We have interviewed Governor Jay Inslee who has advised us that he learned of the offender release date issue for the first time on December 17, when members of his staff brought it to his attention. The next day he met with Secretary Pacholke and others from DOC senior management for a more complete briefing on the issue. Chief of Staff David Postman, General Counsel Nicholas Brown, and Policy Director Matt Steuerwalt, advised us that they were informed about the problem late in the afternoon on December 16, 2015. They worked until late in the evening gathering additional information about the problem. The next day they brought the problem to the Governor’s attention. We were retained to conduct this investigation less than one week later. 33 According to a Seattle Times article which appeared on February 7, 2016, Sen. Mark Miloscia, who chairs the Senate’s Accountability and Reform Committee, opined that Secretary Pacholke “inherited a culture of apathy at DOC.” It is unclear from the article whether this opinion is based upon information received from Secretary Pacholke or other sources. Some of the comments submitted to the fixdoc.org website offered support for this assessment, at least in the context of the working environment for corrections officers working in the various State penal institutions. - 48 - Our investigation has demonstrated that neither the Governor nor any of his staff referenced above knew of this matter in advance of the times disclosed to us. As noted earlier in this report, Sandy Mullins, who from December 2013 forward, served as a Senior Policy Advisor for Government Operations and Public Safety in the Governor’s office, did recall, after the problem came to light in December 2015, that Ms. Steelhammer had briefly referenced it before or after a DOC Senior Leadership meeting. She has not stated, and there is no evidence to suggest otherwise, that she ever brought this matter to the attention of anyone in the Governor’s office prior to mid-December 2015. VI. A. RECOMMENDATIONS All AG opinions to DOC should be subject to supervisory review and approval. Ronda Larson’s flawed advice to DOC escaped supervisory review. In order to prevent this problem from occurring in the future we recommend that the Attorney General’s Office institute a policy requiring that advice given by an Assistant Attorney General to DOC relating to release dates and other significant issues, must be subject to supervisory review and approval. B. The IT governance process should be re-structured. As previously addressed, the process for prioritizing IT change requests was seriously flawed during the relevant time frame. A system based on which business user squeaks the loudest as the method for assigning work to the IT programmers and testers is the equivalent of no system at all. It is our understanding that steps are currently underway to put a new prioritization process in place. This apparently involves using new software called Decision Lens and periodic meetings of senior management, known as the Tiger Team, to review all pending IT Change Requests. Although this may be a positive change, it is clear more must be done. First, the prioritization process must be driven by the overarching mission of DOC, which is to protect public safety. Every IT defect and enhancement must be assessed in light of its impact on public safety and prioritized accordingly. Second, the IT group should play a limited role in setting priorities. The business users who understand the impact of IT defects and enhancements on DOC operations are in the best position to assess the impact of such changes on the outside world. Business users must be represented in the prioritizing process by senior managers from operations who can knowingly represent the needs of the business users and who have sufficient authority to make prioritization decisions and be held accountable for their decisions. IT, on the other hand, must provide technical input on the programming challenges; estimates of time required to complete and test such change requests; and ongoing progress reports. Third, an effective prioritization system must include sufficient documentation to allow the progress of a change request to be tracked, as well as to provide a mechanism for retrieving a comprehensive accounting of work completed and delays and problems encountered. It must also include substantive information explaining any interruptions of work on the item. - 49 - Fourth, an effective prioritization system also requires that business users be kept fully informed as to the status of their change requests and be given a means of voicing concerns if unacceptable delays are encountered. This could be accomplished by creating a tracking system that allows involved parties to determine what has been done, by whom, and when for each project and change request. It must also facilitate the historical recreation of the path taken by any project or change request. Fifth, there must be an accountability component to this prioritization system that permits managers to effectively monitor work and productivity and to identify and address unwarranted problems and delays. The IT business manager should, at a minimum, provide the CIO with a spreadsheet reviewing the status of all pending projects and change requests. This spreadsheet should be organized by date of submission of the project or change request so that the CIO could immediately focus on items that have been in the queue for extended periods. Finally, the system must include a failsafe mechanism that requires supervisory review and sign off on any project or change request that has not been completed in a fixed period of time, for example, within two M releases or within four months of submission. It would not be difficult to program Clearquest or some other computer tracking system to “flag” change requests that remain in the queue after this period of time. Any such delayed change request could, by policy, require the signature of the CIO and the appropriate assistant secretary of the division whose personnel had submitted the change request, before further time passed or it was moved to an even later M release. Such a failsafe mechanism would ensure that unreasonable delays and postponements could not occur without management knowledge and approval. C. DOC should appoint an outside monitor. It might be appropriate for a limited period of time to appoint an outside monitor to oversee the restructuring of the IT prioritization process and then oversee its operation. Because this is viewed as a critical step in ensuring that the King decision defect experience is not repeated, someone from outside DOC with particular experience in IT management might help ensure that an adequate prioritization system is implemented with DOC’s IT group. D. DOC should adopt a policy requiring the hand calculation of release dates when problems are detected. The early release of thousand of offenders could have been easily prevented had DOC promptly started to hand calculate the release dates for all offenders serving enhanced sentences. DOC should immediately implement a new rule requiring that whenever a sentencing defect impacts release dates, offenders release dates must be hand calculated until the defect is fixed. E. DOC should adopt a policy requiring the immediate notification of the appropriate Assistant Secretaries of any system-wide error that affects the sentencing, release, or supervision of offenders. In addition, there should be a policy that whenever a defect in calculating an offender’s sentence, release date, or supervision term comes to the attention of any DOC employee, that employee must forward information regarding the defect to a senior manager in the institution, to - 50 - the institution’s record manager, to other appropriate managers, and to the Statewide Records Manager. These managers should then have the responsibility to analyze the defect to determine if it is an isolated occurrence or a more systemic problem that must be broadly addressed. Systems-wide defects should be brought to the attention of the appropriate Division chief to oversee the adjustment or correction F. A second programmer who is proficient on each component of OMNI should be available. Again, this recommendation is obvious, but it apparently has not been implemented either within DOC’s IT group or within Sierra. A commitment should be made to immediately begin to train back-up programmers not only for SSTA issues, but for every aspect of OMNI. Such back-ups should, whenever possible, come from DOC’s own IT staff, so that the reliance on outside contractors’ work can, over time, be reduced. Moreover, a more effective record keeping system should be implemented to document the use of outside contractor resources. Although no one voiced concern regarding the ability of Sierra Cedar to complete all assigned work, there was no obvious mechanism in place to track the work performed by Sierra Cedar personnel. Without data, it is impossible to assess whether the outside contractor’s personnel are being effectively utilized; what work each has been assigned or completed; and what if any delays have occurred. G. DOC management should emphasize to all employees that its core mission is public safety. It is likely that most, if not all, employees recognize that the core mission of the Department of Corrections is the safety of the citizens of this State. That said, the events related to the handling of the offender release date calculation defect suggest that many, including some in key management positions either lost sight of this mission objective or failed to correctly analyze the probable impact of their decisions on the community. The consequences were tragic. With the appointment of a new DOC Secretary, there is an obvious opportunity to give renewed emphasis to this core mission objective. It may be appropriate to implement a training module that would encourage all DOC employees to be more attentive to this, and other, objectives of the Department. It may also be appropriate to begin some form of Department wide campaign that promotes community, employee, and offender safety. The purpose of such training and promotion is to be certain that there is a greater likelihood that DOC employees make decisions using the proper criteria and standards. H. DOC should create an ombudsman position. A constant theme of many comments submitted by DOC past and present employees to the fixdoc.org website is that they are reluctant to come forward and voice concerns or complaints for fear of retaliation. These concerns appear to have come primarily from DOC employees in the individual institutions, but there may be relevance at the headquarters level, too. Clearly, there was some reluctance on the part of those attempting to wrestle with the offender release date defect to press for the problem to be fixed. An ombudsman might help - 51 - break down barriers to facilitate more open communications between lower level personnel and management. 891.01 qb240101 - 52 -