Table 3.

Limitations, caveats and improvements for clinical outcome measures

Limitations and caveats	Improvements
Expanded Disability Status Scale (EDSS)
High intra- and inter-observer variability  Non-linearity (bimodal distribution)  Limited responsiveness  Necessity to use non-parametric statistics (ordinal scale)  Uneven distribution of relapsing–remitting and progressive patients  Several functional domains not assessed	Accounting for baseline score when determining change (e.g. change ≥1.0 with baseline score 0–5.5, and ≥0.5 for higher baseline scores) Determining disability worsening with confirmation of the EDSS progression after at least 6 months Using standardized scripts for questioning patients (improving reliability and decreasing risk of unblinding) Simplification of scoring rules (decreasing variability) Streamlining by stripping components of the functional systems that are less informative Modification to improve linearity and facilitate statistical analysis
Relapses
Strong subjectivity  Recovery of signs or symptoms before confirmation of relapse  Recall bias of patient and observer bias of examiner  Newly reported symptoms not always clearly depicted in change of the EDSS  Identification largely depends on patient reporting it  Higher relapse rate prior to inclusion: over-reporting to fulfil inclusion criteria, high relapse rate inclusion criterion leading to decrease of relapse rate because of regression to the mean, placebo effect, decrease of relapse due to natural course of MS	Confirming a relapse by another examiner Increasing number of visits to identify more relapses
Multiple Sclerosis Functional Composite (MSFC)
Moderate reliability, sensitivity and responsiveness of the PASAT  The PASAT often disliked by patients, requirement of mathematical ability and ceiling effect  Several important functional domains are not assessed  Lack of a clear dimension of the overall score (resulting in difficult interpretability)  Z scores are influenced by results of the reference population and obscure the meaning of crude scores	Replacing the PASAT with the symbol digit modalities test Adding the low-contrast letter acuity test (covering visual domain) Adding other functional domains Determining minimal clinically relevant changes of the Z scores and confirming change after 6 months Determining clinical relevance Keeping elements separated instead of combining them into a single score
Patient-reported outcome measures (PROM)
Unblinded nature  Potential expectance bias  Assessment of quality of life may be influenced by multiple factors  Possible response shift over time	Weighing of individual questions appropriately Using (computer) adaptive testing to reduce test length and improve tolerability

MS multiple sclerosis, PASAT paced auditory serial addition task