Use and validation of administrative data for suicide research
Loading...
Date
2017-09-01
Authors
Randall, Jason Robert
Journal Title
Journal ISSN
Volume Title
Publisher
International journal of methods in psychiatric research
Abstract
Health system administrative databases are valuable sources of data about health system use. Most of these databases now extend multiple decades and therefore can be used to perform birth cohort studies on children, youth and young adults. The goal of this thesis was to use the administrative data available in Manitoba, Canada to assess the incidence of mental health disorders, suicide attempts, and deaths, as well as to assess the validity of detecting suicide attempts with these data.
Chapter I examined the incidence of diagnosis with anxiety, mood and adjustment, personality, schizophrenia, and substance use disorders in a cohort of individuals born in Manitoba who were living in the city of Winnipeg on their tenth birthday. The estimates were higher than those provided by studies relying on recall of diagnosis. The results supported the idea that recall for diagnosis with mental disorders might be a poor method of determining lifetime history of these illnesses.
Chapter II examined the occurrence of suicide attempts/deaths in the cohort from Chapter I. This chapter showed that the suicide attempts were fairly common in the sample. It also noted that individuals with personality and schizophrenia disorders had the highest occurrence of attempts/deaths and that the occurrence of behaviours was particularly high after the first instance of diagnosis with one of the five disorders examined in the previous chapter.
Chapter III assessed the validity of using International Classification of Diseases (ICD) codes for detecting suicide attempts from hospital discharge abstracts. We found that these codes have good specificity and positive predictive validity, but miss most of the suicide attempts that are admitted.
Chapter IV used latent class analysis to examine whether the current data in the hospital discharge abstracts, medical claims and emergency department information system (EDIS) can accurately identify individuals presenting with self-harm to the emergency department. This study found that these data sources are currently insufficient at identifying these individuals.
Overall, this thesis used administrative data to perform an epidemiological cohort study on mental illness and suicide attempts, but also highlighted some of the limitations of this method of epidemiological study.
Description
Keywords
epidemiology, mental health, suicide, self-harm, validation, administrative data
Citation
Emergency department and inpatient coding for self‐harm and suicide attempts: Validation using clinician assessment data JR Randall, LL Roos, LM Lix, LY Katz, JM Bolton - International journal of methods in psychiatric research, 2017