This retrospective cohort analysis included patients selected from the HealthCore Integrated Research Environment (HIRE), a database of medical, laboratory, and pharmacy claims of commercially insured patients. The HIRE contains administrative claims data from 14 major commercial health plans across the United States, representing 45 million lives.
As a non-interventional, observational analysis, no patients were directly involved in the study. The study was conducted in compliance with state and federal laws, including the Health Insurance Portability and Accountability Act (HIPAA) of 1996.
A chart review was included in this study and approval for a complete waiver of HIPAA authorization was granted by a central Institutional Review Board (Quorum Review IRB, waiver #26699) prior to chart collection. All claims were from a limited dataset with anonymized patient information.
Patients between the ages of 13 and 64 years who had two or more claims for schizophrenia (International Classification of Diseases, 9th edition code 295.xx) from January 1, 2007 through to April 30, 2010 were included in the study. The younger patients were included in order to be able to describe the treatment patterns not only for chronic schizophrenia patients, but also for those who are at the earliest stages of their diagnosis and disease. Patients were followed for one year subsequent to the first claim for schizophrenia (the index date).
Antipsychotic medications were identified from pharmacy claims closest to the index date (±90 days). All first-generation (FGA) and second-generation (SGA) antipsychotics were included. Patients with a prescription for an FGA and/or SGA within 90 days before or after the index date were stratified by type of therapy into the monotherapy or polypharmacy cohorts. The first of these prescriptions was defined as the index medication. Monotherapy was defined as a single antipsychotic medication, whereas polypharmacy was defined as having two or more antipsychotics within a 45-day period. Short-acting antipsychotics given by injection for acute treatment were excluded to eliminate the possibility of mis-categorizing patients taking medications in an inpatient setting. Long-acting injections were not included because we found their use to be low in the claims (<2% overall, data not shown).
From the list of all schizophrenia patients included in the study, a random sample of 200 was targeted for medical record review. Patients were required to have documented schizophrenia in their medical record.
Because clinical outcome measures are not available in administrative claims data, the primary outcome measure used in this analysis was duration of therapy, measured over one year following the date antipsychotic treatment was initiated. Treatment duration was defined as the number of days without a 90-day gap in treatment. Discontinuation was defined as a 90-day gap in antipsychotic therapy prior to the end of the one year follow-up period. In the polypharmacy group, discontinuation was defined as a 90-day gap in at least one of the antipsychotic medications during follow-up.
Secondary outcomes included the type and number of psychiatric and somatic comorbidities during the one year follow-up period. Psychiatric comorbidities included depression, anxiety, bipolar, schizoaffective, attention-deficit and personality disorders. Analysis of somatic comorbidities included the Deyo-Charlson Comorbidity Index (DCI) score and the proportion of patients with claims for cardiovascular diseases, cerebrovascular disease, diabetes, hyperglycemia, or hyperlipidemia. Concomitant psychiatric medications other than antipsychotics, including antidepressants, anxiolytics, hypnotics and sedatives, mood stabilizers, and stimulants, were recorded. The type and number of somatic medications, including lipid-modifying agents, anti-diabetic agents, anti-Parkinson disease and movement disorder medications, cardiovascular agents, and anti-obesity drugs were also collected.
A random sample of 200 patients within the larger claims cohort was selected for medical chart review to estimate schizophrenia severity because this is not available in claims data. Severity within the sub-sample was determined from physician documentation in the charts and described as mild, moderate or severe. Due to the small number of patients with documented schizophrenia severity, comparisons between polypharmacy and monotherapy groups were descriptive only.
Measures were described as means (± standard deviation [SD]) for continuous variables, and frequencies for categorical variables. Differences in mean length of therapy were compared using t tests, and χ tests were used to compare the proportion of patients discontinuing treatment and the prevalence of comorbidities. Ordinary Least Squares (OLS) regression was used to predict the length of therapy (by type of antipsychotic therapy regimen), when controlling for gender, region, number of somatic and psychiatric comorbidities, DCI score, and the number of psychiatric and somatic medications.