Validation of key behaviourally based mental health diagnoses in administrative data: suicide attempt, alcohol abuse, illicit drug abuse and tobacco use
1 Center for Statistical Consultation and Research, 3555 Rackham, University of Michigan, Ann Arbor, MI 48109-1070, USA
2 Department of Veterans Affairs, Ann Arbor Center of Excellence (COE), Serious Mental Illness Treatment Resource and Evaluation Center (SMITREC), Ann Arbor, MI, 48105, USA
3 Center for Health Quality, Outcomes and Economic Research, Edith Nourse Rogers Memorial VA Hospital, 200 Springs Road, Bedford, MA 01730, USA
4 Department of Psychiatry, 55 Lake Avenue, University of Massachusetts Medical School, Worcester, MA 01655, USA
5 Department of Psychiatry, 4250 Plymouth Rd., University of Michigan Medical School, Ann Arbor, MI, 48109-5765, USA
BMC Health Services Research 2012, 12:18 doi:10.1186/1472-6963-12-18Published: 23 January 2012
Observational research frequently uses administrative codes for mental health or substance use diagnoses and for important behaviours such as suicide attempts. We sought to validate codes (International Classification of Diseases, 9th edition, clinical modification diagnostic and E-codes) entered in Veterans Health Administration administrative data for patients with depression versus a gold standard of electronic medical record text ("chart notation").
Three random samples of patients were selected, each stratified by geographic region, gender, and year of cohort entry, from a VHA depression treatment cohort from April 1, 1999 to September 30, 2004. The first sample was selected from patients who died by suicide, the second from patients who remained alive on the date of death of suicide cases, and the third from patients with a new start of a commonly used antidepressant medication. Four variables were assessed using administrative codes in the year prior to the index date: suicide attempt, alcohol abuse/dependence, drug abuse/dependence and tobacco use.
Specificity was high (≥ 90%) for all four administrative codes, regardless of the sample. Sensitivity was ≤75% and was particularly low for suicide attempt (≤ 17%). Positive predictive values for alcohol dependence/abuse and tobacco use were high, but barely better than flipping a coin for illicit drug abuse/dependence. Sensitivity differed across the three samples, but was highest in the suicide death sample.
Administrative data-based diagnoses among VHA records have high specificity, but low sensitivity. The accuracy level varies by different diagnosis and by different patient subgroup.