Relationship between evaluation factors and star ratings for Japanese community healthcare institutions in electronic word-of-mouth reviews: an observational study.

Abstract

Background: Internet reviews have become increasingly crucial for both healthcare providers and patients. Electronic word-of-mouth (eWOM) reviews on internet sites often comprise textual content with numerical ratings. In this study, we aimed to identify the evaluation factors of community healthcare institutions regarding eWOM reviews and the impact of each evaluation factor on the institution's ratings.

Methods: An observational study was conducted to qualitatively and quantitatively analyse eWOM data posted on Google for randomly selected healthcare institutions in Hirosaki, Japan from September to October 2022. For qualitative data, the authors repeatedly read the eWOM text, coded it, and categorised related sections. For quantitative analysis, a multivariate analysis using a linear regression model was conducted with the categorised factors from the qualitative analysis as explanatory variables and eWOM ratings as response variables.

Results: Twenty medical institutions (two hospitals and 18 clinics) were randomly extracted from the registry. A total of 147 eWOM texts from each institution were analysed, and coding was performed for 474 segments in the texts. In the qualitative analysis, the evaluated factors in eWOM texts for medical institutions were categorised as communication (evaluation factor for communication with healthcare providers), clinical practice (evaluation factor for the clinical practice of healthcare providers), and medical institution (evaluation factor for the characteristics of medical institutions). According to the multiple regression analysis, the partial regression coefficients for the explanatory variables of positive communication, clinical practice, and medical institution evaluations were 1.04 (95% confidence interval 0.70 to 1.39), 0.65 (95% confidence interval 0.27 to 1.04), and 0.75 (95% confidence interval 0.35 to 1.15), respectively, with the number of ratings as the response variable. Partial regression coefficients for the explanatory variables of negative communication, clinical practice, and medical institution evaluations were - 1.52 (95% confidence interval - 1.88 to -1.16), -0.90 (95% confidence interval - 1.28 to -0.52), and - 0.24 (95% confidence interval - 0.60 to 0.12), respectively.

Conclusion: We quantitatively and qualitatively analysed the eWOM reviews and ratings of healthcare institutions posted on Google. Three evaluation factors were identified: communication, clinical practice, and medical institution. Our study revealed that communication significantly impacts ratings.