Somerset Medical Center patient-incident cluster

The Somerset Medical Center patient-incident cluster was a series of unexplained, life-threatening crises in the hospital’s Critical Care Unit, marked by abnormal laboratory findings and, in five cases, death. Its significance lies not only in the possibility that patients were deliberately poisoned, but also in the widening gap between clinical evidence, the hospital’s internal inquiry, and the urgent warnings from New Jersey Poison Control Center. The cluster raised a question that the institution could not resolve by treating each patient as an isolated medical puzzle: whether someone was harming patients, and who had a duty to act before that possibility was proved.

Background and Emerging Pattern

The hospital’s account to investigators eventually described six incidents over five months. The first five named patients were Joseph Lehman on May 28, Frances Kane on June 4, Jin Kyung Han on June 16, Reverend Florian Gall on June 28, and Francis Agoada on August 27. McKinley Crews was the sixth; his death, only four days before the hospital called the prosecutor, prompted the delayed referral. The institution reported that five of the six patients had died. It did not say that the incidents were connected, and its own five-month investigation had not identified their source.

The signs were medically different but disturbing in combination. Somerset later reported that Gall and Han had digoxin findings, while the other four incidents involved insulin. On July 8, 2003, director Dr. Steven Marcus of New Jersey Poison Control Center confronted the hospital’s administrators by conference call. Risk manager Mary Lund had already spoken with Marcus once; now senior administrators heard his warning that suspected deliberate harm carried a legal reporting obligation. Bruce Ruck, also of Poison Control, asked whether the hospital had measured C-peptide and insulin levels in two other patients with unusual glucose results. Lund confirmed that the results correlated. Dr. William Cors then described an endocrinologist’s difficulty explaining at least one case without an external influence. The clinicians’ evidence had moved the concern beyond an unexplained fluctuation: an outside source of drug exposure was a serious possibility.

The book sets this pattern against earlier hospital warning signs. At Saint Barnabas Hospital, tests of patients who had suffered unexplained insulin crashes found unusually high levels of insulin, much of it foreign to the body. In one investigated IV bag, testing detected insulin, and microscopic examination found tiny needle punctures, including on the bag’s edges. At Saint Luke’s, nurses had described patients who seemed to be recovering before suddenly coding, sometimes soon after Charles Cullen changed their IVs. These accounts do not by themselves establish what happened in each Somerset case; they show why the cluster’s repeated abnormalities, drug findings, and timing demanded more than routine explanations.

Investigation and Reporting Dispute

Somerset’s administrators continued to weigh the possibility of human error against a more sinister explanation. They had checked for manufacturer recalls and adverse drug interactions, verified IV equipment and bedside monitors, and repeated laboratory tests to rule out lab error. They also tightened pharmacy controls on insulin and digoxin, making nurses accountable for these commonly used medicines. The hospital said it was reviewing its documentation systems, including Pyxis MedStation 2000 for drug records and Cerner for patient charts, and that it was interviewing involved staff. Yet it had not found an innocent error in the records, nor had it identified a source for the incidents.

Marcus regarded the delay as a danger in itself. He told Somerset that suspected patient poisonings had to be reported to the state within twenty-four hours, a deadline the hospital had already missed. The administration replied that it would not report the cases until it had completed a thorough internal investigation. Marcus’s concern was informed by earlier cases in which hospitals treated repeated patient collapses as a medical pattern to study while administrators and lawyers treated them as a potential lawsuit. In his view, waiting for a complete institutional inquiry before contacting authorities repeated a pattern in which patients could die while officials deliberated.

The conflict sharpened when Marcus said that the calls had been recorded. He had already contacted Eddy Bresnitz, the state epidemiologist and assistant commissioner of the New Jersey Department of Health and Senior Services, describing a cluster that might have resulted from a criminal act. He also emailed assistant commissioner Amie Thornton about four untoward clinical events and Somerset’s reluctance to report them before finishing its own inquiry. A few hours after Marcus disclosed the recordings, Lund reported the four patient incidents to the department by fax and email. That report described the hospital’s checks and new drug controls, while acknowledging that the available explanations were running out. The sequence exposed a fundamental disagreement: Somerset wanted to complete its own investigation first; Poison Control considered prompt reporting part of the effort to prevent further harm.

Police Inquiry and the Limits of Proof

When Somerset finally notified the Somerset County Prosecutor’s Office, its presentation was careful. Counsel Paul Nittoly said that the hospital was not reporting a homicide, not exactly. Cors described abnormal laboratory findings and life-threatening symptoms, but avoided asserting that the cases had a common cause. For investigators Tim Braun and Danny Baldwin, the briefing offered names, dates, and medical terminology but no conventional crime scene, weapon, or eyewitness. The case began in uncertainty: the hospital had identified a series of alarming incidents, yet had not established whether they were crimes or how they might be linked.

The investigators faced a concrete evidentiary obstacle. They were told that Pyxis retained drug-access data for only thirty days, while Reverend Gall had been dead nearly four times that long. Without a record tying access to a drug to a particular worker and time, the detectives could not connect a suspected drug exposure, a patient, and a suspect as directly as they might connect a bullet to a gun. The reported limit made crucial records appear lost and left the developing case short of proof, even as clinical findings and the accumulating timeline strengthened suspicion.

The book also records how staff observations helped turn isolated events into a pattern. At Saint Luke’s, Nurse Lynn Tester said that patients who seemed to be improving had died suddenly, often after Cullen changed their IVs. Nurse Robin Saulsberry recalled seeing him leave a patient’s room shortly before that patient coded, and said he watched the electrocardiogram monitor with unusual attention. Tester calculated that Cullen had worked 26 percent of the hours but was present for 58 percent of the deaths. These recollections concerned another hospital and were not proof of the Somerset incidents; in the broader investigation, they showed why staff accounts, timing, laboratory evidence, and medication records all mattered, while none alone could settle the case.

Staff, Trust, and Consequences

The cluster altered ordinary work on the Somerset unit before its cause was known. At Saint Barnabas, Amy Loughren had already objected when managers required nurses to sign an insulin adjustment form and electronically attest to an estimate of how much insulin remained in a vial. She considered the estimate unreliable and asked why insulin was being treated like a narcotic; her supervisor would not explain the change. Later, rumors linked Charlie, insulin, and a death, and colleagues avoided Amy because she was known as his friend and had challenged the new protocol. Her protest had registered a safety concern, but the institution had not given her the information that would have explained it.

When Danny Baldwin showed her a Pyxis printout from the night Reverend Gall died, Amy recognized Charlie’s activity and could no longer dismiss the possibility that he had done something terrible during those shifts. The document changed the meaning of their friendship without resolving the personal conflict: she could begin to believe the evidence, but could not easily reconcile it with the person she trusted. She weighed the risk of lying to Charlie and her employer against the possibility that the detectives were mistaken or manipulating her. Despite fear for her job and the isolation of breaking ranks, she began helping the investigators. She later identified Cerner records as important and supplied paperwork while detectives pursued the patient charts.

The cluster therefore developed along two connected tracks: clinical signs that pointed toward an external drug exposure, and institutional decisions that governed how quickly those signs reached investigators and health officials. Somerset’s tighter medication controls and eventual report were responses, but they did not erase the earlier delay or establish who caused the incidents. For Amy, the inquiry made loyalty itself a source of danger; for investigators, the difference between a strong pattern and a provable individual act remained decisive. The episode ends its initial phase not with a settled explanation, but with the case moving beyond the hospital’s internal review and into a wider investigation.