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

Brief written June 7, 2026 from GPS Intelligence System data.

The Score Distribution: Mostly A’s, With a Few Sharp Exceptions

Georgia Department of Public Health inspection records for state prison kitchens, compiled by Georgia Prisoners’ Speak (GPS), show a striking distribution: the vast majority of facilities receive scores in the 90s or a perfect 100, earning “A” grades. Among the facilities that have posted a 100 within the last year are Autry, Calhoun, Central State, Coffee Correctional, Dooly, Hancock, Lee, Montgomery, Rutledge, Walker, and several women’s facilities. These numbers suggest that achieving a passing score is not only possible but routine under the current inspection regime.

Yet the same dataset also contains three failing grades — scores below 70 — at Johnson State Prison (64 in December 2023), Pulaski State Prison (67 in January 2026), and Smith State Prison (68 in May 2022). The range from a perfect 100 to a failing 64, documented across a system of roughly three dozen facilities, raises a central question: if so many kitchens can score in the A range, why are a few failing repeatedly — and what does the sudden appearance of a failing score actually tell us about the daily conditions incarcerated people experience?

A GPS review of the DPH inspection records found that at Smith State Prison, rodent activity was cited in every inspection from 2022 through 2025. The February 2026 inspection noted roach activity in the bakery and tray-making station, listed as a repeat violation, alongside 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 repeat. These are not isolated discoveries; they are documentation of an environment that has been failing inspections for years and is still in the same condition at the next walkthrough.

Three Failing Kitchens and the Repeat-Violation Trap

The failing scores are not outliers that immediately self-correct. Instead, they follow a pattern: the inspector finds severe conditions, issues a failing grade, and the facility stages a cleanup for the follow-up — but the underlying problems return by the next routine inspection.

At Johnson State Prison in Wrightsville, a December 11, 2023 inspection found a score of 64. The inspector documented multiple rats and roaches throughout the kitchen, described as an ongoing problem “with little to no change.” Bulk food items — oil, flour, rice bran — had holes gnawed through bags with visible droppings and urine. 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 perforated the floors, walls, and ceilings. Cold-holding violations forced the discard of multiple food items on the spot. A follow-up inspection just nine days later, on December 20, 2023, produced a score of 67 — still an F, despite the intervening time. The facility’s scores continued to oscillate: 75 (C) in June 2023, 67 (F) in December 2023, 91 (A) after a follow-up in July 2023, then back to 86 (B) in March 2024, 96 (A) in December 2024, only to slide again to 80 (B) in March 2025 and 88 (B) in October 2025. The F was not a one-time event; it was the bottom of a long oscillation that the inspection system has been unable to break.

Pulaski State Prison in Hawkinsville followed a similar trajectory. Its scores declined from 92 (A) in June 2023 to 82 (B) in June 2024, then 83 (B) in February 2025, 73 (C) in August 2025, and 78 (C) in a September 2025 follow-up. In January 2026, it failed outright with a 67. The inspector found that the facility’s only designated handwashing sink was nonfunctional — the 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 were observed switching between tasks without washing hands. Nacho meat held at 65°F and sauce at 123°F were documented as repeat hot-holding violations. Then, a follow-up inspection on February 6, 2026, brought the score to 96 — an A, but achieved by rapid remediation that the facility’s own track record suggests will not hold.

Smith State Prison in Glennville scored 68 in May 2022, a failing grade. By January 2024, it had risen only to 82 (B), then 85 (B) in June 2025, and 72 (C) in February 2026 — still below 80 after nearly four years of inspection findings. The pattern at all three facilities is not that the inspector finds new problems each time; it is that the same classes of violation — cold-holding, pest infestation, plumbing, equipment disrepair — appear again and again, documented, ostensibly corrected for the follow-up, and then reemerge.

The Scheduled-Inspection Trap

The reason a facility can fail, pass a follow-up with a 96, and then fail again lies in how the inspections themselves are conducted. As The Marshall Project reported in March 2025, health departments generally must arrange prison inspections in advance because of security protocols. The kitchen staff knows the inspector is coming. Moreover, even when violations are found, inspectors are reluctant to shut down prison kitchens, because incarcerated people have no alternative food source — a kitchen closure is not a temporary inconvenience for a restaurant that closes for the afternoon; it is a logistical crisis for a prison.

This structural dynamic is reinforced by the economic geography of the counties where prisons operate. Many GDC facilities are the largest employer in small rural counties. The health department sanitarians who conduct the inspections often work, socialize, and worship in the same communities as the prison staff whose kitchens they are scoring. The pressure not to disrupt the largest local institution is not a matter of individual corruption but of structural embeddedness: a sanitarian who issues a failing grade to the county’s primary economic engine faces a fundamentally different set of professional and social consequences than an inspector in a large metro health district. No individual inspector is accused of wrongdoing; the design of the system itself — scheduled, overlapping social networks, no independent backup — effectively discourages the kind of enforcement that would force sustained remediation.

Compounding this, the Georgia Department of Corrections conducts its own internal food-service inspections under Board of Corrections Rule 125-1-2-.10, which mandates “regularly scheduled and unannounced evaluative inspections.” These GDC inspections are not made public. Unlike the DPH reports, which are integrated into the statewide restaurant-inspection database, GDC’s internal findings are invisible. The only public window into kitchen conditions is the DPH score — a scheduled snapshot from an agency that lacks the authority to unilaterally close a kitchen and whose inspectors would hesitate to do so even if they had it.

The absence of a centralized GDC-specific inspection database further erodes accountability. Prison inspection records are mixed into the general food-service database alongside restaurants, schools, and hospitals. There is no mechanism for tracking prison-specific trends, comparing facilities, or identifying systemic failures across the system from the DPH data alone — GPS assembled the dataset used in this analysis by extracting and collating the records manually.

What the Scores Don’t Capture: Inmate Maintenance Workers Speak

If the failing scores reveal kitchens that cannot even pass a scheduled walkthrough, the real test of the inspection regime is what it misses in the facilities that score in the high nineties or a perfect 100. Here, the accounts of the incarcerated workers who actually operate the kitchens carry the weight of contradiction.

GPS has received multiple consistent reports from incarcerated kitchen workers at facilities that routinely receive A grades under the DPH system. These accounts, provided to GPS staff over time, describe conditions that would never appear on an announced one-day inspection: serving tables lined with roaches, mice active in dry storage, food trays washed by hand in repurposed chemical barrels because the dishwashers have been inoperable for years. The workers who report these conditions are the same people tasked with making the kitchen inspection-ready when the sanitarian arrives — a temporary cosmetic fix that masks what becomes routine again the following day.

At Baldwin State Prison, the DPH inspection records show perfect 100s in September 2023, April 2024, December 2024, June 2025, and December 2025. Central State Prison received a 100 in July 2023, January 2024, July 2024, December 2024, June 2025, and November 2025. On paper, these are impeccably maintained food-service operations. Yet the worker accounts collected by GPS describe the very vermin activity and equipment failures that the DPH inspector, arriving on a pre-arranged date and walking a prepared path, did not see.

This contradiction is not new. Carla Simmons, incarcerated at three Georgia women’s facilities since 2004, wrote publicly about finding “roach legs in cornbread and rats climbing over the dry goods in the warehouse.” Her account, published in Truthout, spans multiple facilities — including those that, like Arrendale State Prison, have posted 100 (April 2025) and 92 (April 2025) on their DPH reports. Similarly, GPS’s own investigative reporting documented photographs from Johnson State Prison showing contaminated food trays being washed in chemical barrels — the result, the reporting found, of broken industrial dishwashers and thirty years of deferred maintenance. Even a failing inspection, which catches the rat-gnawed bags and inoperable equipment, does not capture the daily reality of workers being forced to serve food on trays dunked in chemical residue because no working alternative exists.

The Broken Response Loop

The legal architecture of Georgia’s prison food inspection exists on paper in sufficient detail. 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 to encompass correctional facilities. DPH Rule 511-6-1-.01 deliberately exempts federal property — military bases — but does not exempt state government facilities, placing Georgia prisons squarely within DPH jurisdiction. GDC’s own Standard Operating Procedure SOP 409.04.26, titled “Food Service Permits-Health Department Inspections,” acknowledges that prison kitchens obtain permits from county health departments and undergo their inspections.

But none of this legal machinery contains a mechanism to force lasting change. The inspections find violations. The violations are cited. A follow-up is scheduled. The facility cleans, patches, discards, and passes. Then the sanitarian leaves and the cycle begins again — because there is no independent enforcement body that can compel a state prison to permanently repair its plumbing, replace its ovens, or maintain pest exclusion between inspections. The DPH cannot realistically close a prison kitchen. GDC’s own internal inspections are sealed from public view. No outside oversight body — no correctional ombudsman, no legislative inspector general for food safety, no consent decree — carries a mandate to audit the follow-up.

The civil litigation record provides one cautionary note. In Gumm v. Ford, a federal class-action case at the Georgia Diagnostic and Classification Prison, incarcerated individuals in the Special Management Unit alleged that they received food that was undercooked, cold, rotten, or beyond its expiration date, causing one plaintiff to lose 40 pounds. A settlement approved in May 2019 required that SMU prisoners receive the same food access as the general population. But the SMU food chain typically runs through a separate meal-delivery pathway that does not originate in the main kitchen inspected by DPH, meaning the DPH score for the facility’s primary kitchen offers no window into what SMU prisoners actually eat.

Toward Real Oversight

GPS’s review of the inspection landscape suggests that Georgia is in a minority of states where an independent health authority inspects prison kitchens at all — most state systems self-inspect without external oversight. Other states are grappling with the same gap: Ohio’s Correctional Institution Inspection Committee, a bipartisan legislative body, conducted 65 prison inspections in 2023-2024 and publishes its findings in public biennial reports. Massachusetts has proposed legislation to create an independent Inspector of Correctional Food Services. Alabama is considering giving its Department of Public Health enforcement authority over prison kitchens, an authority it currently lacks.

Georgia has the foundation: an existing DPH inspection apparatus that, despite its flaws, provides scores that the public can access. What it lacks — and what the repeated oscillation between failing and passing scores at Johnson, Pulaski, and Smith demonstrates — is a mechanism to turn a one-day snapshot into sustained accountability. Until inspections are genuinely unannounced, until the scores are paired with an enforcement authority that can demand permanent remediation, and until the voices of the incarcerated workers who keep the kitchens running are treated as evidence, the A on the wall will continue to have little to do with what reaches the tray.

Sources: This analysis draws on DPH food-safety inspection records for Georgia Department of Corrections facilities, compiled and analyzed by GPS; reporting from The Marshall Project, Truthout, and Filter Magazine; GDC standard operating procedures and board rules; the civil docket in Gumm v. Ford; and inmate-worker accounts collected by GPS staff. The full dataset of DPH scores was extracted from the Georgia health inspections public database and from documents obtained through open records requests.