Why observational studies should be among the tools used in comparative effectiveness research.

Doctors, patients, and other decision makers need access to the best available clinical evidence, which can come from systematic reviews, experimental trials, and observational research. Despite methodological challenges, high-quality observational studies have an important role in comparative effectiveness research because they can address issues that are otherwise difficult or impossible to study. In addition, many clinical and policy decisions do not require the very high levels of certainty provided by large, rigorous randomized trials. This paper provides insights and a framework to guide good decision making that involves the full range of high-quality comparative effectiveness research techniques, including observational research.

The prevalence and geographic distribution of Crohn's disease and ulcerative colitis in the United States.

BACKGROUND & AIMS: Previous US studies of inflammatory bowel disease (IBD) prevalence have sampled small, geographically restricted populations and may not be generalizable to the entire nation. This study sought to determine the prevalence of Crohn's disease (CD) and ulcerative colitis (UC) in a large national sample and to compare the prevalence across geographic regions and other sociodemographic characteristics. METHODS: We analyzed the health insurance claims for 9 million Americans, pooled from 87 health plans in 33 states, and identified cases of CD and UC using diagnosis codes. Prevalence was determined by dividing the number of cases by the number of persons enrolled for 2 years. Logistic regression was used to compare prevalence estimates by geographic region, age, sex, and insurance type (Medicaid vs commercial). RESULTS: The prevalence of CD and UC in children younger than 20 years was 43 (95% confidence interval [CI], 40-45) and 28 (95% CI, 26-30) per 100,000, respectively. In adults, the prevalence of CD and UC was 201 (95% CI, 197-204) and 238 (95% CI, 234-241), respectively. The prevalence of both conditions was lower in the South, compared with the Northeast, Midwest, and West. IBD appears to be more common in commercially insured individuals, compared with those insured by Medicaid. CONCLUSIONS: This estimation of the prevalence of IBD in the US should help quantify the overall burden of disease and inform the planning of appropriate clinical services.

Risperidone versus Conventional Antipsychotics for Schizophrenia and Schizoaffective Disorder : Symptoms, Quality of Life and Resource Use under Customary Clinical Care.

OBJECTIVE: To prospectively compare risperidone with conventional antipsychotic agents among schizophrenia patients treated under usual practice conditions. DESIGN: One-year, multicentre, open-label, randomised trial carried out in 21 centres in 17 states of the US. PATIENTS: 684 patients were followed from 1995 to 1997, and must have experienced a symptom relapse at study start. INTERVENTIONS: Patients were randomly assigned to risperidone therapy or their physician's 'best choice' of any one of the 13 conventional antipsychotic medications approved in the US. MAIN OUTCOME MEASURES AND RESULTS: Outcomes measured were changes in psychiatric symptoms, side effects, satisfaction with drug therapy, quality of life (including health-related quality of life [HRQOL]) and resource utilisation. A subgroup analysis of the non-switchers was also conducted. Irrespective of treatment group, treatment switching and days with no drug therapy were observed. Compared with patients on conventional antipsychotics, those in the risperidone group achieved statistically superior scores on the Positive and Negative Syndrome Scale for Schizophrenia (PANSS) [PANSS total score improved from 83.32 to 61.80 vs 81.42 to 66.99 in the risperidone and conventional groups, respectively), Barnes Akathisia Scale (scores improved from 0.89 to 0.55 vs 0.87 to 0.81 in the risperidone and conventional groups, respectively), and 36-Item Short Form Health Survey (SF-36) scale (scores improved from 32.83 to 39.92 vs 32.55 to 37.22 in the risperidone and conventional groups, respectively) during the 1-year treatment period. A significantly higher percentage of risperidone- treated patients had a 60% improvement in PANSS scores at 12 months (20.9% of patients compared with 10.7% in the risperidone and conventional groups, respectively). There was no statistically significant difference in resource utilisation between the two groups. Among non-switchers, patients in the risperidone group had lower total costs and more clinical benefits. CONCLUSIONS: Conditions of usual practice resulted in a high degree of non-treatment, treatment changing and multi-antipsychotic drug therapy. Patients in the risperidone group had better clinical outcomes (e.g. reduced psychiatric symptoms and side effects) and improved HRQOL. There were no significant differences in healthcare utilisation between the two study groups.