Polygenic Prediction of Suicide: Clinical, Ethical and Psychosocial Impact

New York State Psychiatric Institute Dba Research Foundation For Mental Hygiene, Inc, New York NY

Investigators

Abstract

Suicide is a leading cause of death in the U.S. and a major public health problem, with hundreds of thousands of suicide attempts and tens of thousands of deaths annually. Unlike several other leading causes of death, progress in reducing suicide rates has been limited, in part because of an extremely limited ability to predict which individuals will attempt or die by suicide. Indeed, a recent meta-analysis of studies that attempted to predict suicidality based on putative risk factors found that predictive accuracy was barely better than chance and had shown no significant improvement over five decades. In this context, recent years have seen increasing enthusiasm regarding the prospect of using genetic information to predict suicide risk. In particular, researchers have recently developed polygenic risk scores (PRS), which use data from genome-wide association studies (GWAS) to estimate a person's relative genetic susceptibility to a particular health outcome (in this case, suicide) by aggregating all risk variants present in the person's genome, weighted by the strength of each variant's association with the outcome in question. Early research suggests that PRS for suicide have significant predictive power, and the prospect of using PRS to predict individuals' risk of suicide has generated significant excitement as a possible precision medicine tool for identifying people in need of intervention and ultimately preventing suicide deaths. However, there are also significant reasons for concern. For instance, if high PRS for suicide are misinterpreted in a fatalistic manner, this could lead to hopelessness among patients or premature reliance on intrusive or coercive approaches to risk-reduction among family members and clinicians. Low PRS could potentially lead to false reassurance. PRS could also affect treatment-related beliefs, patient-provider relationships, and stigmatization of at-risk individuals. The proposed research will investigate attitudes and beliefs about the prospect and implications of using PRS to predict suicide risk in clinical care, using surveys and qualitative interviews of primary care patients with symptoms of depression (Aim 1) and of primary care physicians and psychiatrists (Aim 2). Among the parameters examined for patients will be willingness to receive genetic predictions of suicide risk and emotional reactions to the receipt of such predictions; anticipated impact of suicide PRS on self-perceptions, reproductive decision-making, life planning, familial and interpersonal relationships, treatment-seeking, patient-provider relationships, and actual suicide risk; and other perceived risks and benefits. Physicians will be asked about perceived clinical utility and likely impact on clinical decisions, along with other perceived benefits and risks. We will also conduct an experimental study of physicians, examining the effects of PRS-based suicide risk information (i.e., mock electronic health record summaries) on putative decisions about clinical care (Aim 3). By investigating how the prospect of PRS-based suicide prediction is received by patients and physicians, the proposed research will help to anticipate the ethical, clinical, and societal implications of polygenic prediction of suicide.