Photo: Olga Kononenko / Unsplash — hospital corridor, undisclosed location

A Good Night

Raymond Carroll served two tours before the back gave out. He came home to Harrison County in 1987 and spent the next thirty years doing what you do when you come home from somewhere like that — you don't talk about it, you find a job that uses your hands, you go to the VFW on Thursdays. By the time he was admitted to the Louis A. Johnson VA Medical Center in Clarksburg, he was seventy-three years old, his kidneys were giving him trouble, and the care was free. That's what they told him when he signed up. That part, at least, was true.

His daughter, Tammy, worked second shift at the Dollar General on Pike Street. She couldn't always make the drive up to the hospital on weekdays, so she called. The nurses told her what nurses tell you — that he'd had a good night, that he was resting, that the doctor had been by. She believed them. You believe them. They're nurses. They're at the VA. He served. There is a system built specifically for this situation, and you trust it because the alternative is too heavy to carry while you're also ringing up cleaning supplies at the Dollar General.

In the autumn of 2017, Raymond was on the night ward in Building One. The staff rotation was tight — the VA had been running short for years, and the overnight shift ran leaner than anyone was officially comfortable with. There was a nursing assistant assigned to that ward whose name was Reta Mays. She had been with the facility for a few years. She knew the layout, the rhythms, the hours when nobody was watching the supply closet.

The nights in Clarksburg in October get cold fast. The hills go dark early, and the lights in the parking lot at the VA cast that particular amber that makes everything look like a photograph taken in the wrong decade. Tammy would think about that later — the light, the parking lot, the drive she didn't make enough. She would think about whether there was a moment she could have known, some change in her father's voice, some hesitation in the nurse's answer. There wasn't. The system was designed to be legible to the people inside it. Everyone outside was operating on faith.

Raymond Carroll died in November 2017. The death certificate said cardiac complications. His kidneys had been failing, so that was consistent. The attending physician signed it. The paperwork was filed. His name was added to the list of patients who had died on that ward that quarter, and the list was not, at that moment, remarkable enough to anyone to look at twice.

He was the third.

Stay with that.

Tammy held a small service at the VFW. People who knew Raymond from the plant came. His pastor. A few people from the Thursday group. Nobody mentioned the VA, because there was nothing, at that point, to mention. He'd been sick. The hospital had tried. That's the story Tammy had, and it was the only story available. She drove home past the Dollar General, past the amber light in the parking lot, past the hill where the hospital sat on the ridge above town, and she went back to work the next morning. The system had produced a death and a silence in equal measure, and the silence was the part that would last longest.

What the Monitoring System Was Built Not to See

Between July 2017 and June 2018, Reta Mays — a ward nursing assistant, not a licensed nurse — entered patient rooms on the overnight shift at the Louis A. Johnson VA Medical Center and administered unlicensed injections of insulin to patients who were not diabetic. Seven veterans died. The Department of Justice would later determine that Mays acted with premeditation, that the deaths were murders, and that the victims were specifically chosen because they were elderly, ill, and unlikely to be checked on again before morning.

The mechanism was not Mays alone. The mechanism was what the VA had built — or failed to build — around her.

Notice what the VA's staffing structure made possible.

At the time of the murders, the Louis A. Johnson VA Medical Center was operating with documented staffing shortfalls on overnight rotations. Nursing assistants like Mays held significant unsupervised access to patient rooms during hours when registered nursing staff were spread thin across multiple wards. The facility's medication management system — specifically the protocols governing insulin access and logging — had gaps that allowed medication to be removed from supply stores without triggering immediate audit alerts. This was not unique to Clarksburg: a 2019 VA Office of Inspector General report examining multiple facilities identified medication tracking inconsistencies as a systemic risk factor across the agency's healthcare network.

The first suspicious death occurred in July 2017. The seventh occurred in June 2018. That is eleven months during which six more veterans were killed after the first death should, in a properly functioning monitoring system, have triggered a review. What happened in those nine months — what did not happen — is the structural story.

The VA's internal death review process is designed to catch statistically anomalous patterns: a ward where deaths are clustering at unusual rates, at unusual hours, among patients who were not in critical condition. The Louis A. Johnson facility had such a pattern. The signal was present in the data. What the system lacked was a mechanism to elevate that pattern to anyone with the authority and incentive to investigate it. Death reviews in VA facilities are conducted by committees that include the clinical staff of the unit under review — a structural conflict of interest that, when deaths are caused by someone within that unit, creates an environment where the committee is reviewing its own conduct.

This is the mechanism: a staffing model that underpaid and under-supervised overnight care, a medication access system with insufficient logging controls, a death review process structurally incapable of investigating deaths caused by its own members, and a reporting culture at a facility that — like many VA facilities serving post-industrial rural communities — was chronically under-resourced and under-inspected.

Cool. Now explain how nine months passed.

The FBI was brought in following the physician's report in 2018. Mays was identified as the primary suspect. She was arrested in July 2019. The federal grand jury indictment charged her with seven counts of second-degree murder and one count of assault with intent to commit murder. On July 14, 2020, she pleaded guilty. On May 11, 2021, U.S. District Judge Thomas Kleeh sentenced her to seven consecutive life sentences plus twenty years — ineligible for parole. The DOJ press release described the case as one of the worst crimes ever prosecuted in the Northern District of West Virginia. The VA issued a statement expressing condolences. The families of the seven veterans received notification by letter.

Compare this to the Charles Cullen case in New Jersey, where a licensed nurse murdered between 29 and 40 patients across nine hospitals between 1988 and 2003. Cullen was flagged as problematic at multiple facilities. He was never fully investigated, and hospitals — eager to avoid liability exposure — provided neutral references that allowed him to move to the next position. The pattern holds: institutions treat internal death investigations as liability management problems, not as public health emergencies. The difference between Cullen and Mays is not that the systems were different. It's that Mays was in a VA facility where an outside physician had access to the mortality data and was willing to use it.

The structural conditions at the Louis A. Johnson Medical Center in 2017 were not exceptional by VA standards. They were representative. A Government Accountability Office report from 2015 identified rural VA facilities as systematically under-resourced relative to their patient populations, with higher nurse vacancy rates, longer fill times for open positions, and less access to psychiatric and specialty oversight than urban or suburban facilities. The Clarksburg facility was operating within those documented norms. The problem wasn't deviation from the system. The problem was compliance with it.

What Mays required to operate for nine months was not opportunity she created — it was opportunity the institution had already engineered. Access to medications without clinical licensing verification. Overnight rounds without supervisory check-ins. A death review committee that included the floor's clinical staff. And a facility so understaffed that adding one more task — actually auditing mortality patterns — was genuinely beyond what the night shift had capacity to do. The system was not failing. It was functioning precisely as designed for a facility at its resource level. That is the part that the DOJ conviction, the life sentence, and the VA press release did not address. That is what needs addressing.

The Infrastructure of Inattention

West Virginia has the fourth-highest percentage of veterans per capita of any state in the country, according to the 2022 American Community Survey. That concentration — built over decades of military recruitment from economically limited communities — means the VA system in West Virginia is providing care for a disproportionately large share of the state's most medically complex population, in a state with the lowest median household income in the nation and among the worst rates of chronic disease.

Read that again.

The highest veteran density. The lowest median income. The worst chronic disease rates. And a VA system that, as of 2023, was carrying approximately 49,000 unfilled positions nationwide — a staffing deficit documented in the VA's own annual performance reports. The facilities serving rural Appalachian communities like Clarksburg are not, within that system, the ones that attract the most resources, the most oversight attention, or the most competitive applicants for nursing and clinical roles.

Research on hospital-based serial killers — a small but well-documented category in criminology — consistently identifies the same enabling conditions: understaffed night shifts, inadequate medication controls, weak peer reporting cultures, and death review processes that incentivize quiet closure over investigation. A 2019 study published in the Journal of Patient Safety estimated that between 210,000 and 440,000 patients suffer preventable harm in U.S. hospitals annually, with a significant share of those harms occurring during overnight and weekend shifts when supervision ratios drop. The VA's rural facilities operate at the far end of that supervision gap.

What the Clarksburg case reveals is not that individual bad actors can exist in healthcare settings — that is known. What it reveals is the specific conditions under which a bad actor can operate for nine months without detection: a facility understaffed enough to make oversight difficult, a medication system permissive enough to make access possible, and a death review culture institutional enough to make pattern recognition someone else's problem.

The VA's response to the Mays case included updates to background check protocols and medication access procedures at the Clarksburg facility. Those reforms are documented. What is not documented — because it is not measured — is whether the underlying staffing and oversight ratios that enabled the murders have been corrected, or whether they represent a structural condition of rural VA care that individual facility reforms cannot reach.

In the Cullen case, New Jersey eventually passed legislation requiring hospitals to report nurses to the state licensing board when they are terminated or resign under investigation. That law was passed in 2005, after Cullen had already killed for seventeen years. The mechanism of reform was a single individual's willingness to confess. Not a system that caught him.

The signal in Clarksburg is not that Reta Mays was a murderer. The signal is that the system built to protect Raymond Carroll and six other veterans was designed, by budget, by staffing ratio, by geographic priority, and by internal review culture, to be unequal to what it was required to do — and that this inequality is not an anomaly in rural VA care. It is the condition.

The VA's mission — "to care for him who shall have borne the battle" — is an obligation that was formalized in 1865 and has been underfunded, inconsistently distributed, and unevenly inspected ever since. West Virginia's veterans did not choose the state's economy or its geography. They did not choose to be served by a facility carrying sixteen vacancies and a history of deficiency findings. They chose to serve the country. The country built a system in return, and in Clarksburg, between July 2017 and June 2018, that system killed seven of them. The accountability conversation ended with seven life sentences. The structural conversation has not meaningfully begun.