Advancing methods for patient-centred measurement in acute in-patient and emergency department sectors of care
Analysis 1. Side-by-side comparison of PROMs
Aim: To compare scores and validity evidence on the VR-12 and the EQ-5D.
1. How do the rates of missing data compare for the EQ-5D and VR-12?
2. How do the distributions of scores from the VR-12 and EQ-5D compare? For example, is there evidence that one instrument is less susceptible to producing skewed distributions or a restricted range of scores?
3. Are there significant differences in the scores from the VR-12 and EQ-5D for different groups and settings?
4. Is there evidence of measurement bias (differential item functioning) for one or both PROMs? If so, is there more bias for one of the PROMs?
5. What is the association between each of the PROMs and other measures of health status? Does one instrument appear to be more strongly correlated with health status indicators, both concurrent and predictive?
Analysis 2: Associations between patient-reported experiences and patient-reported outcomes
Aim: to describe the associations between generic PROMs (VR-12 and EQ-5D) and PREMs.
1. To what extent are PROM and PREM scores related?
2. Are some PROM and PREM domains more highly related than others?
3. Can the Wilson and Cleary (1995) Framework for PROMs be adapted to be inclusive of PREMs?
Analysis 3: Key driver analyses of PREMs
Aim: to investigate statistical methods for determining the relative importance of patient-reported experiences in relation to global patient experience and patient-reported outcomes.
1. What is the relative importance of PREM dimensions and items explaining variance in global patient experience?
2. What is the relative importance of PREM dimensions and items explaining variance in patient-reported physical and mental health outcomes?
3. What is the relative importance of PREM dimensions and items explaining variance in patient-reported health utility scores?
Analysis 4: Computerized Adaptive Testing and PREMs
Aim: to test the application of computer adaptive testing (CAT) methods to patient-reported experience measures.
1. Identify a statistical dimensional structure for the CPES-IC items that is compatible with CAT.
2. Examine heterogeneity in the population with respect to the identified dimensional structure.
3. Establish and validate CAT scoring algorithms.
Analysis 5: Missing data and representativeness
Aim: Examine sample under-representation in patient-centred measurement data, including type of missing data and sampling information characteristic of surveys in our province).
1. Describe and examine the extent of sample under-representation.
2. Investigate methodological approaches to address the impact and consequences of sample under-representation.
3. Develop expository resources to support analysts and applied researchers in understanding the problem and implementing methods for addressing sample under-representation.
Wilson, I. B., & Cleary, P. D. (1995). Linking clinical variables with health-related quality of life. A conceptual model of patient outcomes. JAMA, 273(1), 59-65.