Abstract
Patient-reported outcome measures (PROMs) are essential to evidence-based practice but their implementation is a significant challenge in low- and middle-income countries. In this thesis, I aimed to understand how the use and acceptability of PROMs can be improved in Nepal.
First, I sought to understand context-specific barriers and facilitators to outcome measure use in Nepal. I conducted focus group discussions with 24 physiotherapists and an online survey of 125 physiotherapists. Major barriers reported were lack of time, inadequate outcome measures in local languages, and inability to follow-up with patients. Participants expressed a need for external regulation and data transparency.
Second, I explored if the Patient-Specific Functional Scale (PSFS) might be useful in Nepal. This is assessed using two main criteria: (a) is the measure acceptable and feasible for use; and (b) does it have sufficient measurement properties. The PSFS is a PROM where patients nominate activities they have difficulty performing and rate them on a scale of 0-10. However, in a previous study from Nepal I found up to 64% of participants made errors in a 0-10 pain scale; verbal and faces scales were better understood. Hence, I hypothesised that an alternative verbal response scale might make the Nepali version of the PSFS (i.e. PSFS-NP) more acceptable. Therefore, I developed two verbal response scales for the PSFS-NP (i.e. v-PSFS) after interviewing 42 individuals. I then pre-tested the v-PSFS and PSFS-NP on 119 individuals. I found that although the verbal scales were more preferred (50% versus 12%), there was no difference in error rates between numeric (34%) and verbal scales (32% and 36%). Higher error rates were associated with greater age, fewer years of education, and inexperience with numeric scales.
Next, I sought to assess the measurement properties of PSFS in non-musculoskeletal conditions in two studies. So for the third study, I analysed the content validity of the PSFS in neurological and cardiopulmonary conditions by comparing data from the PSFS in study two with the International Classification of Functioning, Disability and Health (ICF), the ICF core sets and validated condition-specific measures. I found that the PSFS predominantly assesses the Activities and Participation component of the ICF, 31% to 79% of the responses in the PSFS overlap with condition-specific outcome measures, and the Brief Core Sets covered 40%-71% of participants' responses whereas the Comprehensive Core Sets covered 67%-100% of responses.
Lastly, I updated the previous systematic review on measurement properties and current uses of the PSFS using the COSMIN guideline. After searching 11 databases, two independent reviewers screened all records, extracted data, and performed risk of bias assessments and GRADE assessments. Of 985 articles screened, 57 articles on measurement properties and 255 articles on the use of PSFS were included. The PSFS demonstrated ‘sufficient’ test-retest reliability in musculoskeletal (22 studies, 845 participants, low to moderate-quality) and non-musculoskeletal conditions (6 studies, 197 participants, very low-quality), ‘insufficient’ construct validity as a measure of physical function (21 studies, 2 945 participants, low to moderate-quality evidence), and ‘sufficient’ responsiveness (32 studies, 13 770 participants, moderate to high-quality evidence). The PSFS was used in 87 unique health conditions, some without prior evidence of validity.
The findings in this thesis highlight the context-specific challenges and future needs for the implementation of PROMs in Nepal, especially the need for guidelines. Additionally, one out of three participants in the study made errors in using the PSFS despite a verbal scale and even with an interview format; error rates were higher among participants with low literacy. The findings call for actions to identify ways to improve the validity of using PROMs in people with low literacy. I discuss alternatives that might be useful based on previous literature.