Association between Visiting Time and Data Quality of Community Health Survey

Abstract

Introduction: Community health survey is one of main health statistics used by community health centers for making local health improvement program. However, the environment around this survey is getting hard to perform because of many social reasons, so strategy on when to visit household was argued as one of methods to enhance survey performance. However, verification on whether visiting time can affect to data quality should be advanced before making visiting time strategy. This studys objective is to examine that visiting time can affect to data quality and provide appropriate evidence and materials to make new strategies.

Methods: 2015 community health survey data (n= 228,588) telephone inspection data (n= 24,545, 10% of the original data was sampled for re-checking main survey by telephone) were used for this study. As representatives of data quality, short time survey (10 mins or under) and question concordance were used through main data and telephone survey data each. Logistic regression considering clustered sampling was used to find the association between visiting time and data quality.

Results: For short time survey, the data was analyzed separately according to employment status. In employed group, there was higher likelihood for short time survey to be happened during 11-13 and 14-16 hour than 8 – 10 hour. In employed group, there was higher likelihood that short time survey happens than 8 – 10 hour. Concordance was not clearly statistically associated with visiting time.

Conclusions: Community health survey needs quite longer time for survey than other surveys. So, Short time survey could mean respondents insincerity response or data that was collected inappropriate way. So, more targeted monitoring on certain time frame will be needed and better guideline and education for interviewers will be needed to block short time survey event to be happened.