Relative accuracy of social and medical determinants of suicide in electronic health records

Abstract

Objective: This paper compares the accuracy of predicting suicide from Social Determinants of Health (SDoH) or history of illness. Population Studied: 5 313 965 Veterans who at least had two primary care visits between 2008 and 2016. Study Design: The dependent variable was suicide or intentional self-injury. The independent variables were 10 495 International Classification of Disease (ICD) Version 9 codes, age, and gender. The ICD codes included 40 V-codes used for measuring SDoH, such as family disruption, family history of substance abuse, lack of education, legal impediments, social isolation, unemployment, and homelessness. The sample was randomly divided into training (90 percent) and validation (10 percent) sets. Area under the receiver operating characteristic (AROC) was used to measure accuracy of predictions in the validation set. Principal Findings: Separate analyses were done for inpatient and outpatient codes; the results were similar. In the hospitalized group, the mean age was 67.2 years, and 92.1 percent were male. The mean number of medical diagnostic codes during the study period was 37; and 12.9 percent had at least one SDoH V-code. At least one episode of suicide or intentional self-injury occurred in 1.89 percent of cases. SDoH V-codes, on average, elevated the risk of suicide or intentional self-injury by 24-fold (ranging from 4to 86-fold). An index of 40 SDoH codes predicted suicide or intentional self-injury with an AROC of 0.64. An index of 10 445 medical diagnoses, without SDoH V-codes, had AROC of 0.77. The combined SDoH and medical diagnoses codes also had AROC of 0.77. Conclusion: In predicting suicide or intentional self-harm, SDoH V-codes add negligible information beyond what is already available in medical diagnosis codes. Implications for Practice: Policies that affect SDoH (eg, housing policies, resilience training) may not have an impact on suicide rates, if they do not change the underlying medical causes of SDoH. K E Y W O R D S high dimensional data, hierarchical modeling, mental health, naive Bayes, observational data/ quasi-experiments, self-harm, social determinants of health, stressful life events, suicide, VA health care system This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Evaluation of a Model to Target High-risk Psychiatric Inpatients for an Intensive Postdischarge Suicide Prevention Intervention

“…Our finding that NLP and LexisNexis SDOH variables improve model prediction accuracy is broadly consistent with prior studies of NLP and SDOH variables predicting suicide, although to our knowledge, no prior studies investigated these associations among postdischarge psychiatric inpatients. Four results about these predictors are especially noteworthy in terms of potential clinical implications.…”

Hospitalizations for COVID-19 Among US People Experiencing Incarceration or Homelessness

“…The main limitation is our ability to identify PEI and PEH from ICD-10-CM and admission codes. Our findings might not be generalizable to all PEI and PEH hospitalized for COVID-19, and ICD-10-CM codes are likely insufficient to identify all hospitalized PEI . We included admission codes in our PEI definition to broaden the scope, which provided more conservative estimates than ICD-10-CM codes alone.…”

“…Our findings might not be generalizable to all PEI and PEH hospitalized for COVID-19, and ICD-10-CM codes are likely insufficient to identify all hospitalized PEI. 35 , 36 We included admission codes in our PEI definition to broaden the scope, which provided more conservative estimates than ICD-10-CM codes alone. For PEH, ICD-10-CM codes are a more established method of identification but are underused, and hospitals might not have standardized methods for recording housing status in electronic health records.…”

Association Between Adverse Social Determinants of Health and Suicide Death

“…The primary exposures of interest were adverse SDoH documented using ICD-9-CM codes captured during clinical visits in the year before the index date. Adverse SDoH were classified into 9 categories developed by Torres et al 5 : (1) alcoholism and drug addiction in the family; (2) encounter for mental health services for victim and perpetrator of abuse or history of abuse; (3) other problems related to primary support group, including family circumstances; (4) problems related to care provider dependency; (5) problems related to education and literacy; (6) problems related to employment and unemployment or occupational maladjustment; (7) problems relating to housing and economic circumstances; (8) problems related to legal circumstances; and (9) problems related to other psychosocial circumstances. See Torres et al 5 for the list of specific ICD-9-CM codes included in each category.…”

“…The main purposes of this study were to identify the prevalence of ICD-9-CM codes related to adverse SDoH among individuals who died by suicide between 2000 and 2015 and to examine the associations between documented adverse SDoH and suicide. Although we anticipated that the prevalence of adverse SDoH-related diagnosis codes would be low, given that these codes may have been underutilized during the ICD-9 era, [5][6][7] we hypothesized that, as supported by past literature, [7][8][9][10][11][12][13] adverse SDoH would be associated with increased suicide risk. Observing such associations would provide additional support for increased standardized documentation of SDoH in health care settings, including for use in suicide prevention efforts.…”