Official evidence review · November 7, 2026
Independent Kilby audit: findings and reform implications
The state examiners identified deficiencies in ADOC governance and at Kilby, including oversight cooperation, disciplinary records, incident documentation, medical privacy, facility conditions and reentry preparation. The report is an independent oversight review, not a judicial ruling. Its recommendations are advisory to the Legislature. ALPRP policy actions listed below are proposals, not enacted ADOC requirements.
Documented findings and ALPRP responses
| Area | Reported audit finding | Report pages | Proposed ALPRP safeguard |
|---|---|---|---|
| Independent oversight | The examiners reported significant difficulty obtaining ADOC records, issued a subpoena for certain documents, and described restricted access to personnel. | pp. 10–12 | Require independent access, documented agency responses, corrective-action registers, and confidential reporting pathways. |
| Employee discipline | In the sampled Kilby personnel records, 26% of disciplinary actions were absent from the centralized eCAS data. One missing prior use-of-force infraction materially affected the subsequent discipline recommendation. | pp. 53–54 | Reconcile disciplinary systems and personnel files; require documented review of prior conduct and supervisory signoff. |
| Use-of-force process | Among sampled use-of-force files, 57% lacked evidence of required notifications and 80% lacked evidence of timely forwarding of preliminary review. | pp. 82–83 | Track notifications, reviewer assignments, deadlines, escalation and independently verified closures. |
| Death-reporting integrity | Three datasets initially listed 66, 72 and 68 Kilby-associated deaths for October 2023–September 2025. Auditors reconciled the differences, which involved transfer and facility attribution. | pp. 89–90 | Publish reconciled records with standardized definitions and documented transfer attribution; never sum the three counts. |
| Medical privacy and conditions | Auditors observed exposed clinical encounters and unsanitary conditions in the infirmary, including ventilation and maintenance concerns. | pp. 67–69 | Add privacy controls, inspection logs, vendor responsibility, repair escalation and clinical oversight. |
| Rehabilitation and reentry | Kilby had no dedicated reentry employee. Auditors could not verify the full volume and attendance of certain programs from records supplied. ADOC lacked trackable reentry plans for most incarcerated people. | pp. 32–34, 74–76 | Create individual intake-to-release plans and track access, waitlists, attendance, completion, identification, housing and transportation. |
| Reentry transportation and ID | Auditors noted the limited statewide rollout of photo ID assistance and recommended more flexible transportation to stable post-release destinations. | pp. 33–34 | Measure whether released residents obtain identification and can reach safe housing and support. |
Audit percentages are drawn from the auditors' samples, not statewide prevalence estimates. The three initial death counts were reconciled by the auditors. Page numbers refer to printed report pages.
From audit findings to measurable reform
The following crosswalk groups the 13 selected Kilby findings into seven implementation topics. It identifies the documented evidence, ALPRP's proposed response and how a future authorized pilot would verify delivery. These are planned indicators—not measured Pilot 100 results.
| Question to answer | Official evidence and limits | Proposed standard | What a pilot would verify |
|---|---|---|---|
| Can independent oversight obtain and verify records without avoidable barriers? | The auditors reported restricted record and staff access, including a subpoena for some records. An audit-specific finding, not a measured statewide denial rate. | Independent oversight access, confidentiality, corrective-action ownership and anti-retaliation procedures. | Audit trail completeness and timely oversight responses Dated request/response logs, exception review, independent sampled verification. |
| Do corrective decisions account for the full disciplinary history? | 26% of actions in sampled personnel files did not appear in central eCAS data. Sampled file discrepancy, not misconduct prevalence across ADOC. | Reconcile discipline across case management and personnel records, review prior conduct, audit exceptions. | Disciplinary record reconciliation completion Matched source records and documented discrepancies/corrections. |
| Can reviewers verify required incident notifications and timelines? | Sampled case files lacked evidence of some notifications and timely preliminary review forwarding. Missing records do not independently establish that notice or reviews never occurred. | Time-stamped notification receipts, clear review assignments and escalation rules. | On-time documented incident notification and review Source-level audit trails and independent reconciliation of incident records. |
| Are death reports consistent across transfers and systems? | Initial lists showed 66, 72 and 68 Kilby-associated deaths over the same period; auditors reconciled attribution differences. These initial counts were reconciled, are not additive, and do not establish a final three-way disagreement. | Standardized unique death identifiers, transfer attribution and reconciled reporting. | Death-case reconciliation and notifications Signed cross-source reconciliation report and audit trail. |
| Do healthcare conditions and privacy protections hold in practice? | Auditors observed privacy limitations and deficiencies in infirmary cleanliness and maintenance. Direct observations at Kilby; not a measurement of every ADOC healthcare setting. | Protected clinical services, hygienic facilities, contract compliance and rapid corrective actions. | Healthcare referral/grievance timeliness and independently verified critical repairs Privacy-aware audit logs, inspections, independent closure checks. |
| Are rehabilitation programs available and measurable, rather than only listed? | Kilby lacked dedicated reentry staff; the audit documented constrained workshop capacity and incomplete participation records. The reported roughly 120 interested people and 20 seats describe one workshop, not statewide demand. | Track eligibility, referrals, waitlist, attendance, completion and transfer-related interruption. | Reentry plan coverage and access-to-program completion Waitlist, sign-in, credential and case-plan reconciliation. |
| Can people leaving at sentence completion access ID, transport and support? | The audit cited 15–20% end-of-sentence releases in FY2023–FY2025 and recommended improving ID and transportation assistance. A historical ADOC interval, not a current 2026 release rate. | Individualized, verified release readiness independent of parole eligibility. | Reentry plan coverage and verified post-release linkage Dated release checklist, referrals, verification and follow-up where lawful. |
Inspect the complete evidence-to-action registry · Policy standard · Pilot 100 evaluation · Research context
The registry identifies source and period, implementation ownership needs, methods for independent verification and privacy-safe public indicators. Report references and underlying audit samples remain distinct from statewide performance measures.
Five-stage application
Revive: begin with clinical privacy, appropriate intake assessment and reliable medical referrals. Rehabilitate: track access to effective programs and interrupted treatment. Rebuild: measure education and career-readiness opportunities. Restore: prepare identification, housing, family connections and transportation. Release: verify individualized reentry arrangements instead of assuming readiness from time served alone.
Three levels of accountability
- Policy: Is a clear standard adopted and available to the responsible personnel?
- Execution: Did the staff, contractor or agency perform the required action?
- Verification: Does a reliable record demonstrate timely completion, review and correction when necessary?
ALPRP recommends reporting these measures separately. A missing document is a documentation deficiency; it should not automatically be described as proof that the underlying action never occurred.
Agency response and review boundaries
ADOC's included response cites modernization initiatives, infrastructure constraints, pending offender-management improvements, funding limitations and Kilby's planned closure under the state's bed-replacement plan. These considerations should be presented alongside the examiners' findings. Immediate life-safety and care obligations should be distinguished from long-term capital investments.
The report's Kilby observations should not automatically be generalized to every Alabama prison. Proposed technology should maintain privacy, human review, data minimization and independent oversight, and should never autonomously determine discipline, diagnosis, housing or parole.
ALPRP implementation priorities
1. Independent oversight and transparent corrective-action tracking. 2. Unified disciplinary, death and use-of-force records with reconciled definitions. 3. Medical privacy, sanitary conditions and accountable vendor performance. 4. Individual intake-to-release plans with verified program access and post-release support. 5. Workforce training, retention and consistent discipline.
Review ALPRP governance proposals · Evidence standards · Systemwide safeguards
Download and inspect the Kilby-only audit findings dataset
ALPRP separately publishes a structured compilation of 13 selected findings from Report 27X-001 and six proposed Pilot 100 performance indicators. The file identifies observation types, source page numbers, dates and limitations. It is not part of the statewide verified-facts series and is not a collection of current Pilot 100 results.
Open the labeled Kilby audit JSON dataset · Policy foundation addendum · Pilot 100 independent verification measures