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

Brief written August 30, 2026 from GPS Intelligence System data.(DS)

The Scoreboard Problem

Georgia's Department of Public Health inspects prison kitchens under the same 100-point system it uses for restaurants, hospitals, and schools. O.C.G.A. § 26-2-370 defines "food service establishment" to include "institutions, both public and private," and the DPH Interpretation Manual explicitly interprets that language to encompass correctional facilities. DPH Rule 511-6-1-.01 exempts food service on federal property but pointedly does not exempt state facilities — the deliberate inclusion of federal property while omitting state institutions makes clear that Georgia prisons fall within DPH jurisdiction. GDC's own Standard Operating Procedure 409.04.26, titled "Food Service Permits-Health Department Inspections," confirms that prison kitchens obtain permits from and undergo inspections by county health departments.

GPS reviewed the statewide inspection database and found a striking distribution: the overwhelming majority of GDC facilities score in the A range, often repeatedly. Central State Prison scored a perfect 100 in both June and November 2025. Baldwin State Prison scored 100 in June 2025. Hancock State Prison scored 96 in May 2025 while its employee dining hall scored 100. Calhoun State Prison posted perfect scores across multiple inspections in 2023 and 2024. The pattern holds across the system — Walker, Lee, Dooly, Coffee, Rutledge, Riverbend, Long Unit, and dozens of others routinely land at 95 or above.

That distribution is the first analytical problem. A system where nearly every kitchen scores an A while three facilities have failed outright since 2022 — Johnson State Prison at 64 in December 2023, Smith State Prison at 68 in May 2022, and Pulaski State Prison at 67 in January 2026 — suggests either that most kitchens are genuinely excellent, or that the inspection regime is not measuring what it claims to measure. GPS's reporting points to the latter.

What a Failing Inspection Actually Finds

The three failing inspections are worth reading closely, because they reveal what the scoring system is capable of capturing when it does capture something.

Johnson State Prison received the lowest documented DPH food-safety score for any GDC facility: 64 on December 11, 2023. The inspector found multiple rats and roaches throughout the kitchen, a problem described in the report as ongoing "with little to no change." Bulk food items — oil, flour, and rice bran — had holes gnawed through bags with visible rat droppings and urine. Multiple cold-holding foods exceeded 41°F and were discarded during the inspection. Five cooking ovens, one tilting skillet, one cooking kettle, one griddle, one freezer unit, and one bulk ice machine were all broken and in need of repair or replacement. Holes were found in floors, walls, and ceilings. A follow-up inspection on December 20, 2023 scored 67 — still failing — with six violations including food separation, cold holding, and surface cleanability.

Pulaski State Prison failed on January 29, 2026, with a score of 67. The facility's only designated handwashing sink was nonfunctional — plumbing had been ripped from the wall with the pipe smashed inward. Sewage was backing up through floor drains, marked as a repeat violation. Employees switched between tasks without washing hands. Nacho meat registered 65°F and sauce at 123°F — both hot-holding violations marked as repeats. The trajectory into that failure was visible in the data: 83 in February 2025, 73 in August 2025, 78 in a September 2025 follow-up, then 67 in January 2026. A follow-up inspection on February 6, 2026 brought the score back up to 96.

Smith State Prison scored 68 on May 27, 2022 — a failing grade. By February 16, 2026, the score had risen only to 72, a "C" grade. GPS's review of the inspection record found rodent activity noted in every inspection from 2022 through 2025. The February 2026 inspection found roach activity in the bakery and tray-making station (marked as a 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 (also marked as a repeat).

The wide variance in scores — from 64 at Johnson to 100 at Central — demonstrates that adequate food safety is achievable within the GDC system. The failures at Johnson, Pulaski, and Smith are therefore a matter of management and resource allocation rather than inherent impossibility. But the more damning pattern is what happens after a failure is documented.

The Broken Response Loop

The "documented failure → return → find again" pattern is the strongest evidence that the regime is broken at the response stage, not the discovery stage. Inspectors are finding the problems. The problems are not being fixed.

At Johnson State Prison, the December 2023 inspection found rats, roaches, broken equipment, and temperature violations. The follow-up nine days later still scored 67. The underlying conditions — the deferred maintenance, the pest infestation, the broken equipment — were not remediated in any durable way. GPS's systemic finding documents that most GDC facilities are 30-40+ years old, with deferred maintenance producing systemwide infrastructure failures including broken kitchen sanitization equipment and pest infestations. The DOJ October 2024 findings, the Guidehouse 2024 assessment, and Commissioner Oliver's public "end of life" statements all corroborate the pattern.

At Pulaski, the decline from 83 to 67 over eleven months was visible in the public record. Each inspection documented specific violations — handwashing facilities, hot holding temperatures, plumbing, backflow devices — and each subsequent inspection found the same categories of violations still present. The January 2026 failure was not a surprise; it was the endpoint of a documented trajectory.

At Smith, rodent activity was noted in every inspection from 2022 through 2025. The February 2026 inspection found roach activity in the bakery and tray-making station marked as a repeat violation. Broken plumbing in four sinks. A broken dishwasher faucet. Clogged floor drains. Mildew on walls, floors, and ceiling — repeat. The facility scored 68 in May 2022 and 72 in February 2026. Four years, a four-point improvement, and the same violations recurring.

The Marshall Project reported in March 2025 that health departments generally must arrange prison inspections in advance due to security protocols, and that even when violations are found, inspectors are reluctant to shut down prison kitchens since incarcerated people have no alternative food source. That structural reality — announced inspections, no meaningful enforcement lever — means the inspection regime functions as documentation rather than regulation.

Small-County Regulatory Capture

The structural dynamics of small-county inspection are not an accusation against any individual inspector. They are a description of how the system is built.

Prison inspections are scheduled in advance. GDC's own SOP 409.04.26 confirms that prison kitchens obtain permits from and undergo inspections by county health departments — the same county health departments that inspect local restaurants, schools, and nursing homes. In the small counties where many GDC facilities are located, the prison is often the largest employer. The sanitarian who inspects the prison kitchen is a county employee. The facility staff they inspect are county residents. The professional and social overlap is structural, not personal.

GPS has documented professional overlap between inspectors and facility staff in small-county settings — the regulatory-capture dynamic that sits at the center of this investigation. The mechanism is not corruption; it is proximity. An inspector who sees the prison's kitchen staff at the grocery store, whose children attend school with the warden's children, who knows that a failing score could mean budget pressure on the county's largest employer — that inspector faces a different set of incentives than one inspecting an anonymous restaurant in Atlanta.

The scheduled nature of the inspections compounds this. A kitchen that knows the inspector is coming can clean, can fix the most visible problems, can ensure the handwashing sink is functional for the one day it matters. GPS's systemic finding is blunt: inspections are scheduled walkthroughs that do not assess equipment under load. A dishwasher that works for a demonstration and fails under the load of 1,500 trays is invisible to the inspection regime.

The Contradiction: High Scores, Contaminated Trays

The analytical center of this investigation is the contradiction between what DPH scores say and what inmate maintenance workers observe.

GPS has documented a systemic pattern of food-service sanitation failure across GDC kitchens that DPH inspection scores systematically fail to capture. Tray-sanitizing dishwashers broken for sustained periods. Inmate-maintenance worker accounts collected at Dooly State Prison describe thousands of roaches inside kitchen equipment. Sustained roach and rodent infestation in kitchen and serving areas. Meals served on visibly contaminated trays.

GPS's investigative reporting, published as "Dunked, Stacked, and Served," documents photographs from Johnson State Prison showing contaminated food trays — the result of broken dishwashers, chemical-barrel washing, and 30 years of deferred maintenance. The trays show dark residue along the seams of the compartments, brown and black buildup adhering to the edges, stains embedded in the plastic that no amount of rinsing removes.

An incarcerated writer who entered the Georgia system in 2015 and has spent nearly a decade eating from those trays described the conditions in GPS's Tell My Story series. At the diagnostic prison in Jackson, roaches were everywhere — in the bottom of the trays, and because the trays are stacked, that meant they were also on top of the trays. They scattered when the trays were picked up.

The Marshall Project's May 16, 2026 investigation independently corroborated the pattern, reporting rats in kitchens, insects in food, moldy trays, and visible malnutrition across Georgia facilities. The Marshall Project quoted GPS's Bernard Christian connecting chronic hunger and degraded food quality to the violence pattern the DOJ documented in October 2024.

The contradiction is the story: high DPH scores at GDC facilities coexist with sustained witness reports of equipment failure and food contamination. The scores measure a single announced walkthrough on a single day. They do not measure what arrives on the tray, what happens between inspections, or what the maintenance workers see inside the equipment.

The SMU Caveat

One clarification is necessary regarding Gumm v. Ford, the class action challenging food in the Special Management Unit at Georgia Diagnostic and Classification Prison. Timothy Gumm alleged that SMU inmates received food that was "inedible or have less nutritional content," undercooked, cold, rotten, or beyond its expiration date, and that he lost 40 pounds. A class settlement was approved on May 7, 2019, requiring SMU prisoners to receive the same food access as the general population, with attorney fees totaling $425,000.

The relevant point for this analysis is that SMU food chains typically do not run through the inspected main kitchen. Food for restrictive housing units is often prepared separately, held longer, and transported differently than general-population meals. The DPH inspection of the main kitchen at Georgia Diagnostic — which scored 100 in March 2025 — would not capture conditions in the SMU food chain. The Gumm litigation is therefore relevant to the broader question of food adequacy in Georgia prisons, but it does not directly bear on the reliability of DPH kitchen inspection scores, because the SMU food chain generally bypasses the inspected kitchen.

What the Scores Don't Measure

The inspection regime is not broken because inspectors are incompetent. It is broken because the regime is structurally incapable of measuring what it claims to measure.

A DPH food-safety inspection is a point-in-time walkthrough of a kitchen. It checks whether food is stored at proper temperatures at the moment of inspection. It checks whether handwashing sinks are functional at the moment of inspection. It checks whether the dishwasher is working at the moment of inspection. It does not check whether the dishwasher works under the load of 1,500 trays three times a day. It does not check whether the roaches return an hour after the inspector leaves. It does not check whether the trays are actually clean when they come out of the machine, or whether the chemical concentration in the sanitizing rinse is adequate, or whether the maintenance worker who repairs the equipment has been told to prioritize something else.

The scores also do not measure the food itself. A kitchen can score 100 while serving meals that are nutritionally inadequate, contaminated at the point of service, or prepared under conditions that the inspection did not observe. The score measures the kitchen's compliance with a checklist on a single day. It does not measure what incarcerated people actually eat.

GPS's systemic finding documents that GDC spends approximately $1.69 per person per day on food — under 60 cents per meal — versus the FDA Thrifty Food Plan estimate of roughly $10 per day for an adult man's nutritionally adequate diet. The state spends approximately 14 times more on medical care for incarcerated people than on their food. That budgetary reality is the context in which the inspection scores must be read. A system that spends 60 cents per meal is not a system that is investing in food safety. The inspection scores, however high, are measuring compliance with a checklist, not the quality of what is served.

Sources

This analysis draws on the Georgia Department of Public Health's statewide food-service inspection database, including full inspection reports for Johnson State Prison, Pulaski State Prison, Smith State Prison, and dozens of other GDC facilities; GDC Standard Operating Procedures including SOP 409.04.26 and SOP 409.04.13; Georgia statutes including O.C.G.A. § 26-2-370 and § 42-4-32; DPH Rule 511-6-1; federal court records in Gumm v. Ford; the U.S. Department of Justice's October 2024 findings report; reporting from The Marshall Project, the Atlanta Journal-Constitution, Truthout, and Filter Magazine; GPS's own investigative reporting including "Dunked, Stacked, and Served"; and inmate and family accounts collected by GPS staff.