Abstract
BACKGROUND: Schizophrenia is associated with excess mortality and multimorbidity, which is possibly associated with difficulty in coordinating care for multiple mental and physical comorbidities. We analyzed the receipt by patients with schizophrenia of 11 types of guideline-concordant care and the association of such care with survival. METHODS: Guideline-concordant care over an 8-year period (financial years 2002 to 2009) was examined in a nationwide sample of 49,173 male veterans with schizophrenia, who were aged 50 years or older. Administrative databases from the electronic medical record system of the Veterans Health Administration (VA) provided comprehensive measures of patient demographics and medical information. Relying on the 2004 American Psychiatric Association guidelines, patterns in 11 types of care were identified and cluster-analyzed. Care types included cardiovascular, metabolic, weight management, nicotine dependence, infectious diseases, vision, and mental health counseling (individual, family, drugs/alcohol, psychiatric medication, and compensated work therapy). Survival analysis estimated association of care patterns with survival, adjusting for clinical and demographic covariates. RESULTS: There was an average of four chronic diseases in addition to schizophrenia in the cohort, notably hypertension (43%) and dyslipidemia (29%). Three longitudinal trajectories (clusters) were identified: 'high-consistent' (averaging 5.4 types of care annually), 'moderate-consistent' (averaging 3.8), and 'poor-decreasing' (averaging 1.9). Most veterans were receiving cardiovascular care (67 to 76%), hepatic and renal function assays (79 to 84%), individual counseling (72 to 85%) and psychiatry consults (66 to 82%), with the proportion receiving care varying by cluster group. After adjustment for age, baseline comorbidity, and other covariates, there was a greater survival rate for those with poor-decreasing care compared with high-consistent care, and for high-consistent compared with moderate-consistent care. CONCLUSIONS: Relatively low levels of guideline-concordant care were seen for older VA patients with schizophrenia, and trajectories of care over time were associated with survival in a non-intuitive pattern. The group with the lowest and decreasing levels of care was also the oldest, but nonetheless had the best age-adjusted and other covariate-adjusted survival rates, possibly because they were requiring less care relative to younger, sicker veterans, and thus their comorbidity burden was markedly lower. Notably, in the group with the sickest individuals (that is those with the highest comorbidity scores, who were very disabled), receiving guideline-concordant care was associated with improved survival in adjusted models compared with those patients receiving only moderate levels of care.