Health Means Different across Cultures: A Multilevel Model Analyzing Self-Report Health Status Using World Values Survey

       Objective: Health is a culture-bound construct. Given the different attitudes people hold toward mental illness across cultures, physical and mental health can be weighed differently. Using a multilevel approach in analyzing a cross-cultural survey under Hofstede’s framework, the current study aims to identify cultural influences on individual’s self-report health status (SRHS), and shed light to international public health research.

       Method: Based on the international World Values Survey taken in 2010-2012, the analysis included 54,913 participants from 37 countries. Overall health status was the outcome variable. Social capital, life satisfaction, and happiness, which measures individual’s mental health status, were included as the individual-level predictors, while Hofstede’s cultural dimensions as the country-level predictors. A series of models were built in terms of increasing complexity to test the effects of cultural dimensions.

       Results: Significant cross-level interactions were found after controlling for demographic and socioeconomic differences: (1) Social capital’s influence on SRHS was significantly predicted by the culture’s uncertainty avoidance (p< .05); (2) life-satisfaction’s influence on SRHS was significantly predicted by the culture’s individualism-collectivism (p< .001) and masculinity-femininity (p< .05); (3) perceived happiness’s influence was significantly predicted by the culture’s individualism-collectivism (p< .05); Using full maximum likelihood estimation, model comparison showed that the country-level cultural predictors significantly explained the model deviance ( ).

       Significance: Results of the current study demonstrate the cultural differences in understanding health status. By proposing and testing a multilevel model, the current study provide guidance to researchers using worldwide questionnaires to take into account cultural differences in analyzing health-related data.