Lehigh County District Attorney's Office

The Lehigh County District Attorney's Office entered the investigation of Charles Cullen after a nurse’s suspicions about unusual deaths at Saint Luke’s Medical Center moved from hospital channels into law enforcement. The office pursued the case with the Pennsylvania State Police, called staff to account for what they had seen, and retained a forensic pathologist to examine patient records. Yet the inquiry ended without a prosecutable conclusion. Its significance lies in that tension: witness accounts and patterns of unexpected deaths raised grave questions, while the documentary and medical evidence assembled for review could not establish that Cullen had killed a particular patient.

Referral and investigation

On August 29, 2002, Saint Luke’s nurse Pat Medellin took her concerns to an acquaintance in the Easton, Pennsylvania, Police Department. She had observed unusual deaths on her unit and had also seen Cullen removed from the hospital for diverting dangerous medications. The hospital had told her that Cullen had not harmed any patients and had closed its own investigation; Medellin was not satisfied with that answer. Her disclosure passed through local police command to the coroner and then to the office of Lehigh County district attorney James B. Martin.

The investigation expanded beyond Medellin’s account. Pennsylvania State Police troopers interviewed Saint Luke’s staff, while coroners Zachary Lysek and Scott Grim worked through medical charts. The nurses’ accounts were described as damningly consistent: Lynn Tester noticed that patients who had seemed to be recovering died suddenly after Cullen changed their IVs; Robin Saulsberry recalled seeing him leave a patient’s room shortly before an unexpected code and death. Tester also compared Cullen’s share of working hours with his presence at deaths: he worked 26 percent of the hours but was present for 58 percent of the deaths. These observations made the suspected pattern more concrete, but they did not, by themselves, establish what had happened to any one patient.

Medical review and the limits of the record

To assess the cases, the district attorney’s office retained Dr. Isidore Mihalakis, a forensic pathologist who had also been involved years earlier in the investigation of Helen Dean’s suspicious death at Warren Hospital. Mihalakis spent months reviewing charts for seventeen patients selected by Saint Luke’s. His review did not produce a finding that Cullen had harmed anyone, and the paperwork offered no prosecutable proof.

The selection and context of those records mattered. Mihalakis was not given a written list of the medications that had been diverted, and apparently did not speak with the nurses or review their statements about their suspicions. Without those sources, the chart review could not connect the clinical record to the staff’s observations or the suspected drug access. The investigation thus contained two kinds of information that did not readily join: nurses described recurring, alarming circumstances, while the selected medical files did not supply a clear evidentiary trail showing that a specific patient had received a lethal dose from Cullen.

The distinction between suspicion and proof shaped the office’s predicament. Investigators could regard the pattern as compelling and still lack a medical conclusion that would support a charge. Even the possibility that a medication had reached a patient was not equivalent to demonstrating, in a form usable in court, who administered it and with what result. The record’s gaps were not a resolution of the nurses’ concerns; they were a barrier to turning those concerns into a provable murder case.

Closure and the case’s aftermath

The district attorney’s investigation lasted eight months and was closed on May 18, 2003. By the time of closure, the office had spent months pursuing the Saint Luke’s deaths, but the pathologist’s review had not established that Cullen harmed any of the seventeen patients whose charts were examined. The inquiry was therefore unsuccessful from a prosecutorial standpoint, despite the accumulating suspicions reported by hospital staff.

Cullen had already moved on from Saint Luke’s to Somerset Medical Center, with neutral references. Later, he confessed to killing five Saint Luke’s patients involved in the cases at issue in subsequent litigation; seven patients in total were specifically identified as having been killed by him at Saint Luke’s. Those later disclosures establish the scale of the harm, but they should not be mistaken for evidence available to the district attorney’s office when it closed its 2003 case. At that point, the office had no confession and no conclusive chart review tying Cullen to a particular death.

Investigative significance

The office’s case illustrates how a criminal investigation can gather troubling witness accounts without acquiring the clinical evidence needed to prove individual acts. The nurses’ reports supplied a pattern and pointed toward a suspect; the chart review, conducted without the diverted-medication list and without the nurses’ statements, did not resolve the central question of causation. That mismatch helps explain why a case that appeared alarming to staff could still be closed without a prosecutable finding.

The closure also forms a sharp contrast with the later investigation in New Jersey, where the question of how Cullen could be made to confess became central. The Lehigh County inquiry had already assembled testimony, medical records, and expert review, yet the case remained unproved. Its outcome was not that the suspicions had been disproved, but that the available evidence had not crossed the threshold required to establish murder. The eventual admissions and identification of victims make the earlier failure more consequential, while preserving the crucial distinction between what became known later and what the office could establish at the time.