Scores Without Sanitation: Why Georgia's Prison Food-Safety Numbers Don't Reflect What Inmates Eat From

Brief written August 2, 2026 from GPS Intelligence System data.

Georgia’s prison kitchens receive regular health inspections from county sanitarian offices, scored on a 100-point scale and published in a statewide public database. On paper, most facilities pass comfortably. Scores in the A range predominate, and many kitchens routinely receive perfect 100s. But the scores measure a scheduled, hour-long walkthrough — not what arrives on the tray three times a day, not what happens between inspections, and not what the men and women who work inside those kitchens see every shift. This gap between the posted number and the lived reality is the central finding of Georgia Prisoners’ Speak (GPS) reporting, corroborated by independent news investigations and the written accounts of incarcerated witnesses.

A Scorecard That Misses the Meal

The Georgia Department of Public Health (DPH) classifies prison kitchens as food service establishments under O.C.G.A. § 26‑2‑370, and DPH Rule 511‑6‑1 explicitly includes “correction facilities” within its scope. GDC’s own policy, SOP 409.04.26, confirms that kitchens obtain permits from and undergo inspections by local health departments. The resulting scores are publicly searchable alongside restaurants, hospitals, and schools; they are not, however, collected or analyzed in any centralized GDC‑specific portal. The data are there, but no agency systematically tracks prison kitchen trends across the state.

GPS reviewed the full public inspection record for every major GDC facility and found that most facilities consistently earn A grades. Central State Prison received a perfect 100 twice in 2025. Hancock State Prison scored 96 for its inmate kitchen and 100 for the employee dining hall in the same May 2025 cycle. Autry State Prison, Baldwin, Calhoun, Dooly, Lee, Montgomery, and Riverbend all produced multiple 100s over the past two years. The dataset suggests a well‑functioning system.

But a narrow band of facilities tells a different story. Three prisons have fallen below the 70‑point passing threshold since 2022: Johnson State Prison (64, December 2023), Pulaski State Prison (67, January 2026), and Smith State Prison (68, May 2022). The variance — from 64 to 100 — demonstrates that adequate food safety is achievable within the GDC system. The failures at the bottom end are, as GPS’s investigation found, a matter of management and resource allocation, not of impossible constraints.

Failing Scores and the Same Findings Again

On December 11, 2023, a routine inspection of Johnson State Prison in Wrightsville recorded a score of 64 — the lowest documented for any GDC kitchen. The inspector noted multiple rats and roaches throughout the kitchen, a problem described as ongoing “with little to no change.” Bulk food items — oil, flour, rice bran — had been gnawed through by rodents, with visible droppings and urine contamination. Five cooking ovens, a tilting skillet, a cooking kettle, a griddle, a freezer unit, and a bulk ice machine were all broken and in need of repair or replacement. Holes were found in floors, walls, and ceilings. Multiple cold‑holding foods exceeded 41°F and were discarded during the walkthrough. Follow‑up inspections brought the score back to 91, but the underlying equipment failures that enabled the infestation were not addressed within the inspection cycle.

At Pulaski State Prison in Hawkinsville, the January 29, 2026 routine inspection returned a failing 67, the culmination of a steady decline: 83 in February 2025, 73 in August 2025, a follow‑up 78 in September, and then the failure. The single designated handwashing sink was non‑functional — plumbing had been ripped from the wall, the pipe smashed inward. Sewage was backing up through floor drains, a repeat violation from prior inspections. Employees moved between tasks without washing their hands. Nacho meat registered 65 °F and sauce at 123 °F, both hot‑holding violations marked as repeats. A follow‑up on February 6 scored 96 — a swing of 29 points in seven days — which raises the question of what, exactly, was measured the first time and why it was not remedied before the inspector arrived.

Smith State Prison in Glennville failed with a 68 in May 2022 and had crawled only to 72 by February 2026. Every inspection from 2022 through 2025 noted rodent activity. The February 2026 report added roach activity in the bakery and tray‑making station (repeat), broken handwashing sinks, broken plumbing in four sinks, a broken dishwasher faucet, clogged floor drains, and damaged walls with mildew on walls, floors, and ceiling (repeat).

The pattern across these three facilities is the same: documentation of severe, repeating violations; a follow‑up score that superficially corrects the written deficiencies; and no sustained change. The inspection system is structured to detect failures; it is not structured to enforce long‑term correction.

What a Walkthrough Doesn’t See

The scores themselves rarely reflect the internal environment that incarcerated kitchen workers describe. GPS’s investigation “Dunked, Stacked, and Served” — a fact‑checked, published article in the GPS intelligence repository — drew on photographs from Johnson State Prison that show contaminated food trays with dark residue embedded along the compartment joints, brown and black buildup that no rinse cycle can remove. The article documented that tray‑sanitizing dishwashers had been broken for sustained periods, with workers forced to wash trays in chemical barrels. Photographs of roach‑lined serving tables and mouse droppings on food contact surfaces were described by inmate‑maintenance workers at facilities that carry A grades.

Inmate accounts collected by GPS describe a recurring scene: thousands of roaches inside kitchen equipment, including inside the seams of tray carts; mice nesting in dry‑goods storage; and meals served on trays that are stacked wet, without a functioning heat‑sanitization cycle. At Coastal State Prison, GPS received a resident’s account of visibly contaminated trays. These conditions persist at facilities that receive high inspection marks because no inspector sees the kitchen under load. The walkthrough occurs after advance notice is given, after the deepest cleaning the facility can produce, and before the 4 a.m. shift cranks up the steam tables.

The contradiction between a 96 posted on the wall and the roach‑lined tray it nominally certifies is the analytical center of GPS’s documentation. It reveals that the inspection system, however well‑intentioned, operates as a ceremonial process that cannot reach the operational reality of a prison kitchen.

Small Counties, Scheduled Visits, and Structural Capture

A significant structural factor exacerbates the gap. Georgia prison inspections are, by necessity, scheduled in advance due to security protocols. The Marshall Project, in a March 2025 article, noted that health departments must coordinate with facilities before arrival, and even when violations are found, inspectors are reluctant to shut down a prison kitchen because incarcerated people have no alternative food source. That advance notice, combined with the absence of unannounced evaluative visits by an independent oversight body, gives kitchens the inherent advantage: they know when the inspector is coming and can present a staged version of the operation.

In many of Georgia’s rural counties, the prison is the largest employer and the single most dominant institution in the local economy. The county sanitarian office that conducts the inspection is embedded in the same small‑town professional and social networks as the facility’s administrative staff. This structural overlap — not an accusation against any individual inspector — creates a dynamic in which the arms‑length adversarial relationship that a health inspection requires is difficult to sustain. GPS’s investigation found professional overlap between inspectors and facility personnel in small‑county settings, a regulatory‑capture pattern that makes it harder for a routine visit to produce meaningful findings. The situation is compounded by the fact that GDC’s own internal food inspections, mandated by Board of Corrections Rule 125‑1‑2‑.10, are not publicly accessible, so there is no cross‑referencing mechanism to compare what a DPH sanitarian sees with what GDC’s own evaluators observe.

A Brief Legal Dimension: Gumm v. Ford

Timothy Gumm’s case in the Middle District of Georgia (Case No. 5:15‑CV‑41) provides a narrow but relevant caution about the limits of inspection coverage. Gumm alleged that men in the Special Management Unit at Georgia Diagnostic and Classification Prison received food that was “inedible or have less nutritional content,” undercooked, cold, rotten, or beyond its expiration date. He lost 40 pounds. The class settlement approved in 2019 required SMU prisoners to receive the same food access as the general population, with attorney fees of $425,000. It should be noted that the food served in SMU units is often prepared in a separate area and may not be covered by the main kitchen inspection that DPH conducts. Gumm v. Ford therefore demonstrates the severity of food‑related harm in Georgia prisons while also illustrating that the inspection scores for the main kitchen do not necessarily capture conditions in segregated housing.

What the Scores Obscure

Georgia’s DPH inspection regime for prison kitchens, standing alone, offers a picture of broad compliance. The existence of that public data is, in fact, better than what exists in many states. But the scores are a measure of a single kitchen on a single, announced day — and they are generated by a small‑county inspection apparatus that operates without independent oversight, without unannounced evaluative visits, and without any mechanism to track or publish recurring patterns across facilities. GPS has documented 1,859 deaths in GDC custody since 2020, a catastrophic violence and mortality crisis that has drawn a Department of Justice investigation and a still‑unresolved pattern‑or‑practice finding issued in October 2024. The food served inside those walls is not the sole cause of that crisis, but the inspection regime’s failure to capture the sanitation reality of prison kitchens — broken dishwashers, rodent‑gnawed dry goods, trays stacked with roach residue — is one piece of a larger system that has lost the capacity to assure basic safety.

Sources

This analysis draws on the Georgia Department of Public Health’s publicly accessible food‑service inspection database; GPS’s own investigation “Dunked, Stacked, and Served”; GPS’s internal documentation of inmate‑witness accounts; The Marshall Project’s May 2026 investigation of prison food conditions in Georgia; and federal district court records in Gumm v. Ford. Inspection scores and violation details are reproduced as recorded in the DPH system; facility‑specific observations are drawn from GPS‑collected firsthand reports and published accounts.