Purpose: The aims of this study were to investigate regional differences in hepatitis C virus (HCV) infection treatment with peginterferon and ribavirin in Japan and to develop and validate statistical models for analysis of regional differences, using generalized linear mixed models.
Methods: Individuals with chronic HCV infection were identified from the Japanese Interferon Database (registered from December 2009 to April 2013). The total sustained virologic response rate and the rate in each prefecture were calculated. In the analysis using generalized linear mixed models, the following four models were constructed: 1) prefecture as a fixed effect, 2) prefecture and other confounding variables as fixed effects, 3) prefecture as a random effect, and 4) prefecture as a random effect and other confounding variables as fixed effects. The quality of the model fit was assessed using the Akaike information criterion and the Bayesian information criterion. All statistical analyses were performed using SAS Version 9.4 for Windows.
Results: From 36 prefectures, 16,349 cases were recorded in the study period. Of these, 4,677 were excluded according to certain criteria. The total sustained virologic response rate was 59.9% (range, 43.9%-71.6%). The statistical model including prefecture as a random effect and other confounding variables as fixed effects showed the best fit based on the Akaike information criterion (13,830.92) and Bayesian information criterion (13,845.17).
Conclusion: Regional differences may exist in HCV infection treatment in Japan. The model including prefecture as a random effect and other confounding variables as fixed effects was appropriate for analysis of such regional differences. Additional studies considering the medical situations of each patient would provide useful information that could contribute to improve and standardize HCV infection treatment.
Keywords: generalized linear mixed model; hepatitis C virus; interferon; nationwide database; regional differences.