Development and Evaluation of Patient-Reported Outcome Measures for use in a Diverse Society

Patient-reported outcome measures are increasingly being used in health services research as an essential component of the total health evaluation of populations. Increased diversity in the U.S. population and the globalization of research pose a challenge to researchers who are concerned with inclusiveness and comparisons among different segments of the population. Health-Related Quality of Life (HRQOL) outcome measures are based on surveys designed to capture subject's own perceptions of functioning and well-being. So the challenge is psychometric equivalence in survey data collected in a multi-cultural and diverse population.;Readability refers to the ease of reading and understanding text. The concept of readability is important in survey research because when items are not appropriately understood, subjects may respond inaccurately or not respond at all. The first manuscript of this dissertation is a readability analysis of seven extensively used HRQOL instruments at the item level. Results showed that all seven surveys have a substantial number of items with high readability levels that may not be appropriate for the general population.;Item Response Theory (IRT) models describe the probabilistic relationship of responses to an item and the underlying trait. These models are helpful in evaluation psychometric equivalence when surveys are used in different cultures. Item bias or differential item functioning is important in survey research because it deals with identifying significant differences across groups, in the proportion of subjects endorsing an item, who have the same ability level. The second manuscript of this dissertation evaluates the English and Spanish language versions of the Center for Epidemiological Studies-Depression (CES-D) for differential item functioning among Latinos. The results of this analysis showed that English-speaking and Spanish-speaking Latinos with the same level of depression responded differently to three of the eleven items in the CES-D.;Item banks can provide an adequate pool to ensure a sufficient number of survey items are available that are applicable for a variety of health service applications in diverse subgroups of the population. The NIH has several initiatives for research involving development and measurement of patient-reported outcomes. The third manuscript of this dissertation summarizes the initial steps in the creation of a vision-targeted HRQOL item bank -- one of the tools in the NIH Toolbox initiative. The goal of this item bank is to create a list of items that measure the health-related impact of vision impairment, vision loss or any chronic eye disease. The results of this manuscript are a vision-targeted item bank comprised of 8 domains and 53 items ready to be field tested in a sample of 800 subjects.;In conclusion, given the increase in diversity in the U.S. population and the globalization of research, equivalence in measurement when surveys are used across different segments of the population and cultures is warranted.