Analyzing Demographic Variabilities Associated With Age of Diagnosis of Schizophrenia Among Patients in Controlled Clinical Trials

Abstract

Background

Prior literature and epidemiological data suggests that the age of diagnosis of schizophrenia (AOD) follows a bimodal and trimodal distribution for males and females, respectively. The studies used to generate these findings were often small and relied on self-reported patient data from a single geographic region in addition to other methodological limitations. We replicated these studies using a modern big data approach by combining raw data from large randomized controlled clinical trials.

Methods

Patient-level data from 15 similarly designed, randomized, double-blind, placebo-controlled, crossover studies with patients using paliperidone extended-release tablets, paliperidone palmitate 1-month, and paliperidone palmitate 3-month, were obtained through the Yale Open Data Access Initiative (YODA). Descriptive statistics and histograms were calculated for continuous variables. A multivariable linear regression was performed with AOD as the outcome variable. Race and sex were used as predictor variables.

Results

Our final analysis included 7881 patients consisting of male (n=4962) and female (n=2919) patients among different racial demographics. Race was consolidated into the following groups: Asian (n=949), Black (n=1692), Hispanic (n=3), Southeast Asian (n=17), White (n=4769), and other (n=343) based on patient self-identification on the YODA datasets. A chi-square test revealed that there is a statistically significant association between patient sex and AOD (x²=295.61, df=68, p < 0.0001). By proxy, this likely means that sex affects age of onset (AOS) as well. Our linear regression output with sex as a predictor of AOD revealed that only the male variable was found to have a statistically significant relationship (p<0.0001) with AOD and resulted in a lower AOD. Histograms generated with the frequency of occurrences against AOD for both male and female patients appeared to be unimodally distributed and skewed right. However, the AOD for female and male patients were found to be 28.79 and 25.44 years old, respectively. This demonstrates that while both male and female AOD are distributed unimodally, there are slight differences in their distributions.

Conclusion

Our analysis differs from previous studies and finds that AOD for male and female patients are seen in a unimodal distribution as compared to previous literature that shows a bimodal and trimodal distribution. Our findings not only call for a re-evaluation of previous epidemiological understandings of AOD but may support future efforts in understanding the origins and typical clinical presentations of patients with newly developed symptomatology of schizophrenia as well as support clinicians’ perspectives as part of clarifying differential diagnoses. Further studies can also continue to evaluate possible correlations among different races.

Funding

No Funding