Comparison of OLS Regression and Quantile Regression on Factors Affecting Cultural Competency of General Hospital Nurses

Wednesday, 24 July 2013

Yun-Hee Park, MPH, RN
College of Nursing, Yonsei University, Seoul, South Korea
Chung Yul Lee, PhD, RN
College of Nursing, Yonsei University, Seodaemunku, South Korea

Learning Objective 1: The learner will be able to understand the effective factors influencing nurses’ cultural competency improvement.

Learning Objective 2: The learner will be able to understand the difference between OLS regression analysis and Quantile regression analysis approach.

Purpose:

 The Purpose of this descriptive study was to examine the factors that contribute to the level of cultural competency and how various factors impact different quantiles of the cultural competency distribution.

Methods:

 A convenience sample of 327 nurses from general hospitals in the Seoul metropolitan area was recruited by using a self-reporting questionnaire.  The cultural competency was measured using the Korean version of the Caffrey Cultural Competence in Healthcare Scale (CCCHS) and Cultural Competence Assessment (CCA). Using the STATA software package, OLS (Original Least Square) Regression and quantile Regression analysis were used to know how various factors impact on different quantiles of the cultural competency distribution.

Results:

 OLS regression analysis showed that cultural competency was affected by cultural experience; participating in foreign language classes (p=.004), working with foreign professionals (p=.01) and language fluency ( p=.019) and insufficient job control (p<.001) among job stress variables. By using quantile regression, cultural experience variable was constantly associated with the cultural competency level at different 3 quantile (25, 50, 75). Job stress with cultural experience variable was associated with the cultural competency level at 50 quantile. The coefficients of experience of participating in foreign language classes among cultural experience variables increased gradually according to level of quantiles; 7 in the 25 quantile, 9 in the 50 quantile, and 10 in the 75 quantile.

Conclusion:  

While OLS regression can identify the factors influencing cultural competency, Quantile regression enables researchers estimating heterogeneous effect in different levels of cultural competency. This study suggests that interventions to strengthen cultural competency for the 50 quantile group should include consideration of job stress. In addition, it is advisable that policy considerations which expands of taking foreign language classes to improve the cultural competency level.