Partial compliance and risk of rehospitalization among California Medicaid patients with schizophrenia

Weiden PJ, Kozma C, Grogg A and Locklear J; Psychiatric Services 55 (8); 886-891

Commented by Dr Stefan Leucht, 24 Aug 2004

Aim of the study

It is well known that noncompliance with medication is a major problem in patients with schizophrenia that can lead to relapse and rehospitalisation. However, noncompliance is often not a dichotomous yes-no phenomenon. In the daily routine it may be common that patients take medication irregularly rather than stopping it completely.

However, the impact of such partial noncompliance on relapse risk has to date not been examined. The aim of the authors therefore was to examine the association between various degrees of noncompliance with antipsychotic medication and risk of hospitalisation in Californian Medicaid patients with schizophrenia.

Methods

The authors had access to a random sample of 20% (n=4,325) of California Medicaid outpatients with schizophrenia (mean age 44.2 years, 58.5% men) who were treated with antipsychotic drugs in the years 1999 to 2001. Patients’ compliance was assessed retrospectively by analysing pharmacy refills and medical claims. Four different definitions were used to define compliance: gaps in medication therapy, medication consistency, medication persistence, and a medication possession ratio.

Since all measures provided consistent results, the primary focus was "gaps in medication therapy" which was calculated as the longest period during which no medication seemed to be available. On average the patients were followed for one year and had a 19.1 dispensing events. The odds of at least one hospitalisation was assessed for each compliance definition using logistic regression models.

Results

The main result of the analysis was a significant correlation between hospitalization risk and degree of compliance. Lower compliance was associated with a higher hospitalization risk according to all compliance definitions and it seemed that non-compliance was a more important factor than any other statistically significant risk factors such as Medicare eligibility and age. Even gaps in medication coverage as small as one to ten days were associated with increased risk of hospitalization (odds ratio [OR]=1.98). When the gaps ranged between 11 to 30 days the OR was 2.81, and gaps of more than 30 days were associated with an OR of 3.96.

Discussion

The new finding of the study is that there is a direct correlation between the degree of noncompliance and risk of hospitalisation in patients with schizophrenia. Thus, while previous studies often considered compliance as a dichotomous yes/no phenomenon, the analysis showed that compliance is rather a continuum and that even partial noncompliance increases the hospitalisation risk. The main methodological advantage of using pharmacy refills is the possibility to analyse a large group of patients during a long period of time.

However, there are also a number of methodological limitations: Reaons of noncompliance can not be detected (was medication stopped intentionally or unintentionally?), the causality between partial compliance and hospitalisation is not completely clear (it could also be that when patients start to relapse, they just forget to take their medication more often, thus inefficacy of treatment being in part the reason for noncompliance), Medicaid patients may not be representative for all patients etc.

The main clinical implication of the study is that complete medication coverage must be the treatment goal and that even partial noncompliance must be avoided. Depot antipsychotic drugs certainly play a crucial role here, because they guarantee that patients are taking all of the prescribed medication.

Last updated: 24.08.2004