During a 13-year period, approximately 2.2% of older adults who died by suicide were associated with long-term care (LTC), according to the results of a cross-sectional epidemiologic study published in JAMA Network Open.
Investigators assessed suicide rates among adults aged 55 years and older, using data collected by the National Violent Death Reporting System (NVDRS) between 2003 and 2015. Along with quantitative information on the decedent, the NVDRS also supplies case narratives abstracted from coroner/medical examiner reports. A machine learning algorithm was used to identify keywords associated with LTC in these case narratives. Adults for whom the algorithm identified keywords were classified as having LTC exposure. Outcome measures included patient demographics, health history, and means of injury. Number of suicides associated with long-term care was also calculated.
Among 47,759 deaths across 27 states (median age, 64 years; 77.6% men; 90.0% non-Hispanic white), the algorithm identified 1037 (2.2%) suicides or undetermined deaths associated with LTC. Decedents whose death was associated with LTC had a median age of 79 years, 73.8% were men, and 94.3% were non-Hispanic white. Among these deaths, 428 occurred while the decedent was living in a long-term care facility and 449 occurred during the transition into or out of long-term care, with 160 noted as “otherwise associated with LTC.”
Investigators noted 2 separate NVDRS variables for location of injury and death. The closest value to LTC for injury is “supervised residential facility”; of the 263 suicidal injuries coded with this value, 106 (40.3%) were identified as occurring in LTC by the algorithm. The agreement among the algorithm, the injury location code (κ statistic, 0.30; 95% CI, 0.26-0.35), and the death location code LTC/nursing home (κ statistic, 0.17; 95% CI, 0.14-0.20) was poor. As such, investigators advocated for more specific injury coding within the NVDRS.
The NVDRS does not include data from all 50 US states, including Florida, which is home to a large number of older adults.
These data underscore the necessity of mental health care for older adults living in long-term facilities, as well as those transitioning into and out of LTCs. In addition, algorithms developed through natural language processing may provide a means to improve variable coding in the NVDRS.
Mezuk B, Ko TM, Kalesnikava VA, Jurgens D. Suicide among older adults living in or transitioning to residential long-term care, 2003 to 2015. JAMA Netw Open. 2019;2(6):e195627.