iWe present two case studies of medical criminals, a doctor and a nurses aid, to highlight the lessons learned and opportunities for effective peer review, credentialing, and early intervention to protect patient safety.
“Doctors are Good!” This statement reflects a foundational assumption of most people when accessing medical care for themselves or their family. Although rare, some doctors do bad things; in fact, some doctors and other health care providers are criminals. By studying the doctors and other health care providers who are caught and prosecuted for criminal wrongdoing relating to patient care, we have a distinct opportunity to learn where systems may be improved to identify and prevent the “bad doctors and other healthcare providers” from treating our patients, friends, family, and the unwary.
Robert Morris Levy, M.D.
“How did [the former Dr.] Levy[1] misdiagnose, or non-diagnose, 3,000 cases without having been discovered by audits or second opinions?” This public comment published with The Washington Post’s article on former pathologist Robert Morris Levy strikes to the heart of what every medical staff professional seeks to protect against: a bad provider who is not identified through initial credentialing, recredentialing, or ongoing peer review. After being arrested and charged with 15 counts of manslaughter, in January 2021, Levy pleaded guilty to three counts of manslaughter and was sentenced to 20 years in prison.
How did this happen? The Department of Veterans Affairs Office of Inspector General (“OIG”) undertook a comprehensive analysis of Levy’s decade-long tenure as the Chief of Pathology at Veterans Health Care System of the Ozarks in Fayetteville, Arkansas, resulting in a lengthy report published on June 2, 2021.[2] The findings were stunning and pointed out multiple missed opportunities to identify and intervene during Levy’s initial application, periodic recredentialing, and multiple staff reports of impairment when on duty.
OIG Investigation
After Levy was fired in 2018, the VA sent in a team of pathologists to conduct a “look-back[3]” and who reviewed 33,902 cases from 2005 to 2017. The team confirmed 3,619 misdiagnoses, 589 of which were determined to have the potential for negative impact on the veterans’ patient care/treatment. Unfortunately, there was not merely the “potential” for negative impact but actual harm, with 15 fatalities directly attributable to Levy’s erroneous diagnoses. Levy’s misdiagnoses and falsification of records, through which he falsely claimed his peer concurred with his diagnoses, led to at least 15 deaths where veterans were either treated incorrectly or not treated at all. His mistakes delayed medical care for an unknown number of veterans and led to unnecessary treatment for others.
The retrospective review identified intervention opportunities that may have averted this tragedy, starting with Levy’s initial application in 2005. He disclosed a previous conviction in 1996 related to driving while intoxicated. His background revealed he was only employed at his prior hospital for a period of eight months. Three of his peer references were from his most recent employment, with only one reference going back more than two years. Levy was initially hired as a locum tenens provider and almost immediately elevated to a service chief position. During his first two years of practice, his probationary period, he had a high error rate that was only identified during the look-back. As only one of two pathologists[4] there was no process in place to review Levy’s pathology practice and establish his clinical competency during the probationary period. During reappointments, Levy’s “peer” references were all from physicians who were not pathologists and who could not be expected to reasonably attest to his clinical competency.
Additionally, Levy was the chair of the pathology quality management committee and responsible for reporting the number of major diagnostic discrepancies (those that would affect the patient’s clinical care) to the medical staff Tissue Committee and Quality Management Committee. The Tissue Committee meeting minutes from November 2009 through September 2017 reflected zero major discrepancies, without any discussion challenging the data. This same information was reported to the hospital leaders. The hospital’s quality management staff had the ability to generate reports of major diagnostic discrepancy cases from the computer system but failed to do so during Levy’s tenure. This occurred despite hospital leadership’s knowledge that a patient received incorrect treatment in 2014, resulting from an amended pathology report that reflected a major diagnostic discrepancy. Surprisingly, there was no increased scrutiny of Levy’s pathology cases following discovery.
Throughout his tenure Levy was subject to peer review as part of the medical staff’s ongoing provider practice evaluation (“OPPE”). But that peer review was limited to a random selection of 10 percent of his cases which were overread by another external pathologist. However, this random selection OPPE process would not and did not detect the error rate identified during the retrospective review. The College of American Pathologists (“CAP”) provides the following guidance for peer review: “there is evidence that targeted review (review of a specific type of case) is more efficient at finding important diagnostic discrepancies or errors than randomly selecting cases for review.”[5] Unfortunately, the VA did not follow the CAP guidance which would have provided an opportunity for earlier detection.
Impairment
The failure to follow-up on the items in Levy’s initial application related to his prior DUI, lack of “peer” references of the same specialty, and failure to follow the CAP guidance on peer review, were signs that might have led to an earlier intervention and potential discovery of the problem. Perhaps the most concerning misstep was the failure of medical staff leadership to effectively intervene in light of multiple reports of Levy’s possible impairment. The chief of staff received periodic informal reports that Levy smelled like alcohol. On one occasion in 2014, the chief of staff investigated, and Levy said he had been drinking a lot of juice, which caused the odor. Levy was not required to take a blood or urine test and no further investigation occurred at that time. In 2015, there were two additional reports to the chief of staff that Levy had the odor of alcohol, red glassy eyes, and hand tremors. Levy voluntarily agreed to testing, but due to misinformation received from human resources, the testing was not performed. In 2016, Levy was again reported for signs of impairment and agreed to testing. The testing showed Levy to have a high blood alcohol level. Levy was immediately removed from clinical care and sent for rehabilitation. After completing a three-month inpatient alcohol treatment program, Levy was allowed to return to work and later received a bonus for stellar performance. However, the medical staff leadership did not conduct ongoing monitoring after Levy returned to clinical practice. By this time, Levy was ordering a substance on the internet known as 2-methyl-2-butanol (2M2B), which would mask the alcohol level in his blood in the event a blood test was required.
Part of the justification for allowing Levy to return to work was the lack of evidence of adverse clinical outcomes. However, the medical staff had not conducted a comprehensive review of his cases. Additionally, leadership did not investigate previously reported concerns that Levy had subverted the pathology and laboratory quality management program, repeatedly misrepresented second reviews of cases, and was deficient in communication with providers when there were significant changes in diagnoses. Although medical staff policy required a second read by a different pathologist when there was a new diagnosis of cancer, Levy was entering concurrence statements into some patients records when a second pathologist had not reviewed the pathology slides or agreed with the interpretation or diagnosis. This pattern was not discovered until the retrospective review.
Termination (finally…)
Levy was removed from clinical practice and terminated in 2018 when he was arrested for driving under the influence. Only then did the medical staff engage in the retrospective review, with its findings followed by Levy’s arrest and charges for manslaughter and mail fraud.[6]
Reta Mays: Medical Serial Killer
“Medical Serial Killers” are a rare but known phenomenon.[7] Reta Mays was a nursing assistant at the Louis A. Johnson VA Medical Center in Clarksburg, West Virginia. On July 14, 2020, she pleaded guilty to seven counts of second-degree murder and one count of assault with the intent to commit murder of patients at the VA Medical Center. Ms. Mays pleaded guilty to deliberately administering insulin to these patients in 2017 and 2018, resulting in hypoglycemia and death. Ms. Mays admitted that she intentionally administered the insulin because she believed the patients were suffering and she wanted them to pass “gently.”
After the Office of Inspector General (“OIG”) became involved, a lengthy investigation ensued which encompassed review of 200 patient deaths, thousands of documents, and 75 interviews. “The OIG found that the facility had serious, pervasive, and deep-rooted clinical and administrative failures that contributed to Ms. Mays’s criminal actions not being identified and stopped earlier.”[8] The OIG blamed weaknesses in many of the facility’s systems and culture of safety for the devasting consequences.
Background
Prior to joining the VA, Ms. Mays had been with the Department of Corrections where she had been the subject of multiple allegations of excessive force. However, before hiring Ms. Mays in June 2015, no one at the VA contacted her prior employer to determine her employment status, skills, and performance. Ms. Mays’s background information was not reviewed even though it had been flagged. Potentially, had this simple step been followed, Ms. Mays might not have been hired.
Additionally, while Ms. Mays was given excellent performance ratings, according to her supervisors, she was verbally counseled for improperly accessing a blood sample from a blood tube, leaving a patient in soiled bedding at the end of her shift, repeated incidents of disconnecting IV lines, and turning off or clearing IV pump data.
According to facility policies, Ms. Mays was not allowed access to the medication room nor medication carts. However, the investigators learned that, in practice, Ms. Mays and other nursing assistants had full access to the medication room. The pharmacy failed to identify the high use of both insulin and the medications used to reverse hypoglycemia at the facility, both of which were evident on retrospective review, but were never identified and brought to leadership’s attention.
Peer Review Deficiencies
The medical workup of the patients with hypoglycemia failed to adequately evaluate the cause of hypoglycemia in patients. Hypoglycemia is rare in patients who are not receiving medication for diabetes. External causes should have been considered when the initial evaluation failed to identify a clinical cause for the patient’s hypoglycemia. In one case, the hospitalist considered surreptitious insulin administration and discussed his concerns with the nurse manager. However, the hospital and medical staff conducted no further follow-up. The hospitalists also failed to order an endocrinology consultation, which may have led to further scrutiny and detection of external insulin administration.
Medical record documentation was found to be scarce and sometimes lacking altogether, which further led to the failure to detect Ms. Mays’s activities. Lack of adequate and comprehensive documentation, combined with a lack of communication regarding patients and their unexplained episodes of hypoglycemia during hand-off to the next hospitalist, was a significant contributing factor to the failure to detect Ms. Mays’s acts. The fact that the hospitalists were working varying shifts and the absence of meaningful communication contributed to the hospitalists failing to identify the pattern of hypoglycemic episodes and deaths at the hospital.
Other contributing factors to Ms. Mays’s ability to continue to harm patients was the failure to follow policies on reporting of adverse events. The unexplained hypoglycemic episodes and deaths would have qualified for adverse event reporting but were not reported. The quality program did not contain critical quality information. The mortality review conducted by the risk manager was superficial even though the hospital had a cluster of deaths attributed to hypoglycemia.
Discovering a Medical Serial Killer
In 2018, the VA Facility Director identified the pattern of unexplained deaths from hypoglycemia and called for an investigation that ultimately identified Ms. Mays and her involvement. Ms. Mays was removed from her employment once the investigation was completed. Ms. Mays was later criminally charged and received seven life sentences for her acts.
Lessons Learned (again)
Not unlike the Texas criminal prosecution of Christopher Duntsch, aka Dr. Death, for two patient deaths and multiple adverse outcomes, Levy and Ms. Mays presented several opportunities for the hospital, its medical staff, and leadership to detect and prevent Levy and Ms. Mays from engaging in further clinical practice. And, like Duntsch, it was the criminal justice system that brought final resolution for the patients and their families when the credentialing, peer review and quality systems failed.
Conclusion
The cases of the former Dr. Levy and Ms. Mays emphasize the need for careful credentialing, peer review, and identification of red flags. Background checks must be appropriately scrutinized. Peer references must be from peers who can attest to an individual’s clinical competence and those references should cover at least the past five years. Hospitals and healthcare organizations must have an effective peer review system that includes a meaningful review of cases using peer review indicators or processes that are published, supported by objective evidence, and proven to be effective. If the hospital has a small number of providers in the specialty under review, make arrangements with other institutions or external review organizations to conduct periodic reviews targeted to identify problems. When long periods of internal review of a service never identify a problem or opportunity for improvement, scrutinize, and revamp the review process. Assume that perfect clinical outcomes for prolonged periods are not the norm and find new ways to look at the data. Conduct prompt and thorough mortality reviews. If the death is unexplained, look further and always look for patterns. Most importantly, take complaints about providers related to competence, quality, or impairment seriously and follow up. Know that a blood or urine test may not reveal drug or alcohol use – the clinical signs of impairment must be taken just as seriously as objective diagnostic tests.
No peer review system is perfect. Continued vigilance and avoiding complacency are the best defenses to rare criminal medical providers.
FOOTNOTES
[1] When Levy lost his license, he was no longer allowed to use the title of “Doctor”
[2] www.va.gov/oig/publications/report-summary.asp?id=4945
[3] “Look-back” was the term used in the OIG report to describe the retrospective review of cases.
[4] The other pathologist was Levy’s subordinate and not in a position to perform peer evaluation.
[5] College of American Pathologists, “About Us,” The College of American Pathologists is an organization of board-certified pathologists. Association of Directors of Anatomic and Surgical Pathology, “Interpretive Diagnostic Error Reduction in Surgical Pathology and Cytology,”
[6] Mail fraud was the result of purchasing 2M2B on the internet.
[7] Medical Serial Killers: The So-Called Angels of Mercy, Crimetraveller.org, May 31, 2021
[8] Care and Oversight Deficiencies Related to Multiple Homicides at the Louis A. Johnson VA Medical Center in Clarksburg, West Virginia, Veterans Health Administration, Office of Healthcare Inspections, Report #20-03593-140, May 11, 2021.
