The present study examined the association between PDDS and EDSS scores and the associations between PDDS and EDSS scores with FS scores, ambulatory outcomes, cognitive processing speed, and clinical and demographic factors in 96 persons with MS. Overall, the PDDS had a strong, albeit not perfect, correlation with the EDSS, supporting criterion aspects of validity. The pattern and magnitude of correlations with FS scores, ambulatory outcomes, cognitive processing speed, and clinical and demographic variables further did not differ between the PDDS and EDSS, and supported construct aspects of validity. The magnitude and pattern of correlations between PDDS and EDSS scores was consistent between persons with mild and moderate-to-severe disability. Such results provide evidence for the validity of PDDS scores as a PRO of disability in persons with MS. The findings and limitations of this study do not suggest that the PDDS should replace the EDSS in clinical research, but rather that researchers and clinicians might consider the PDDS as an alternative assessment of disability, particularly when the EDSS is impractical (e.g., non-face-to-face research, lack of a clinician or other trained personnel available for administration), too costly, or inconvenient (e.g., time constraints of data collection, community-based research). Such recommendations and results are consistent with the original intention of the PDDS serving as a surrogate measure for the EDSS in clinical research involving MS.
This study examined the pattern of associations between FS scores with both the EDSS and PDDS in a cross-sectional analysis. Such an analysis is important for understanding the main components of neurological functioning that correlate with PDDS scores, and if the pattern is consistent between the PDDS and EDSS. The EDSS and PDDS had strong correlations with Pyramidal and Cerebellar FS scores, and moderate correlations with Sensory and Bowel/Bladder FS scores. The correlations were generally small between EDSS and PDDS with Visual, Mental, and Brainstem FS scores. Importantly, there were no differences in the magnitude of correlations between the EDSS and PDDS scores with FS scores. This pattern of correlations is generally consistent with previous research examining the associations between FS and EDSS scores in persons with MS [32
] and suggests that motor involvement is a primary contributor to PDDS and EDSS scores. This supports the convergent aspects of construct validity for the PDDS.
The present study further examined the association between PDDS scores and measures of ambulation as well as demographic/clinical variables, cognition and functional limitations. This was warranted as the PDDS was developed based on the DS and this latter measure was primarily designed based on ambulation and motor functioning being the main determinants of disability in MS [3
]. To that end, analyses indicated that the PDDS was most strongly correlated with ambulatory assessments (i.e., 6 MW, T25FW, TUG, Steps/day, & MSWS-12) and assessments of basic and advanced lower extremity functional limitations (i.e., LL-FDI scores). Other researchers too have reported associations between PDDS scores and MSWS-12 scores (ρ=.847)
, 6 MW performance (ρ=-.427)
, and free-living accelerometry (ρ=.519)
in persons with MS [18
]. Collectively, such results along with the pattern of associations with FS scores from the EDSS imply good convergent aspects of construct validity for the PDDS.
By comparison, the PDDS correlated moderately with age, upper extremity functional limitations on the LL-FDI, and performance on the SDMT, and weakly with MS disease duration, education, income, and performance on the PASAT. Importantly, the EDSS and PDDS correlated similarly with each of those outcomes. Such evidence alongside the weaker correlations between PDDS and EDSS with Visual, Mental, and Brainstem FS scores supports the divergent aspects of construct validity for the PDDS in persons with MS. Of further interest herein is the moderate relationship between cognitive processing speed (SDMT) and the EDSS and PDDS. Future studies might consider examining if impairment of cognitive processing speed influences the validity of PDDS and EDSS scores.
The primary benefit of the evidence provided in the current study is the provision of information on the PDDS as a valid PRO of disability in persons with MS. This is essential for building a stronger body of evidence regarding the actual validity of PDDS scores and clarifying misreporting of validity evidence in the literature. Indeed, we identified a correlation of .783 (95% CI
.691-.850) between PDDS and EDSS scores in the present study, and this is substantially and significantly less than the values of .93 and .958 often cited in the literature for the validity of PDDS scale; we believe, as noted in the introduction, that these large values may be unsubstantiated or reflect the association between EDSS and DS. Our correlation is stronger than the value of .64 reported in a validation study of the Performance Scales, another self-report of disability in MS [7
]. Importantly, participants in the earlier study [7
] were less disabled (median EDSS
3.5) compared with the current sample (median EDSS
4.5) and this may have accounted for the difference in correlations between studies. Indeed, analyses within disability subgroups in the present study support this as a likely explanation (i.e., the correlations were weaker, albeit still significant and strong, in the subgroups who would have a truncated range of scores). Overall, this study provides the first comprehensive assessment of the validity of PDDS scores as a PRO of disability in persons with MS.
There are multiple limitations of the present study. The first limitation is that we only provide validity evidence from a cross-sectional analysis, rather than data on the correspondence between changes in EDSS and PDDS over time. It will be important to perform such a longitudinal study, and doing so will allow for determination of the test-retest reliability of the PDDS over time. Although persons with a wide range of disability were included in this study, the results are limited in that our sample did not cover the full disability range present in persons with MS; there were no persons in the present analysis with an EDSS score of less than 2. We further did not include the DS in this study for an overall comparison of associations among EDSS, DS, and PDDS scores together and with other outcomes. This would have provided more definitive evidence on the source of misreporting regarding the validity of PDDS scores. This should be done as part of a longitudinal study. Lastly, the sample consisted mostly of Caucasian women with RRMS and a short disease-duration, and our results might not be generalizable broadly amongst those with MS. We only examined the validity of an English, print version of the PDDS, and future research should validate the PDDS in different languages and cultures as well as using electronic media (e.g. applications on cellular phones, the Internet, or tablets in the office or at home).