Determinants of Health Seeking Behavior in Pakistan: A Complex Health Survey Design

Background It is important to understand the health seeking behavior of the population and trend of health services utilization in Pakistan. To investigate the determinants of health seeking behavior in Pakistan we suggest... [ view full abstract ]

Background
It is important to understand the health seeking behavior of the population and trend of health services utilization in Pakistan. To investigate the determinants of health seeking behavior in Pakistan we suggest a multilevel pseudo maximum likelihood (MPML) approach to estimates model parameters for the complex survey design.

Method
The sampling strategy of the National Health Survey was stratified two stage cluster sampling. Overall 18,315 subjects were interviewed. This is three level data with PSUs at the third level, household at the second level and persons at the first level. Health care utilization was considered as a binary outcome.
Results
We found age, gender, marital status, household ownership of durable goods, urban/rural status, community development index, and province as significant predictors of health care utilization (p-value <0.05). We also found two significant interactions; between gender and marital status (p-value<0.005), and between the community development index and urban/rural status (p-value <0.045).
The variances of the random intercepts are estimated as 0.135 for PSU level and 0.224 for households. The results are significantly different from zero (p-value<0.05) and indicate considerable heterogeneity in health care utilization with respect to HHs and PSUs.
Conclusion
Our study results reveal the inequalities between socio-economic groups and between urban and rural residents of Pakistan especially in terms of health care utilization. There is also a need to consider gender sensitive programs. Policy makers in Pakistan frequently make decisions without data. This study builds the case for Pakistan that data can safely be used for decision makings. It also gives support to health advocates who use data to promote health reform. This study gives advocates a stronger position in relation to decision makers in the government, as they marshal data to promote their policies for reform.