Cognitive decline may not be adequately captured in economic evaluations of multiple sclerosis: are new treatments being undervalued?

Introduction

Cognitive impairment interferes with a patient’s attention span, decision-making ability, understanding and concentration, impacting on their day-to-day life¹. Nearly half of patients with multiple sclerosis (MS) have cognitive impairment and rates of impairment are particularly high in patients with secondary progressive MS (SPMS; 79.4%) or primary progressive MS (91.3%)². The most commonly impaired aspect of cognition is cognitive processing speed, impacting around 75% of patients with SPMS³. Deterioration in cognitive processing speed is associated with unemployment in patients with MS, highlighting the importance of cognition to peoples’ ability to work⁴. Limiting cognitive decline should therefore be an important treatment goal in MS. Siponimod and alemtuzumab have been shown to preserve cognitive functioning in SPMS and relapsing–remitting MS, respectively^5,6. However, these newer treatments are more expensive than older treatments, such as glatiramer acetate and interferons.

Economic models of treatments for MS are usually constructed using health states defined by Expanded Disability Status Scale (EDSS) scores, an important endpoint in clinical trials for evaluating disease progression. Within the model, patients move between health states according to disease progression (reflecting treatment response). Each health state is associated with a value representing patients’ health-related quality of life (HRQL), often derived using the EQ-5D^7,8. However, because the EDSS and EQ-5D scores may not be influenced by changes in cognitive function, cost-effectiveness estimates from economic models built using the EDSS and EQ-5D may fail to reflect the effect of treatment on cognitive function and its value to patients.

This article discusses whether the EDSS and EQ-5D are adequately influenced by changes in cognitive function, and how this might affect the apparent cost-effectiveness of new treatments.

Limitations of the EDSS in evaluating cognitive decline

The EDSS assessment is completed by a neurologist and evaluates physical functioning, muscle weakness, ataxia, bladder or bowel dysfunction, visual problems, swallowing and speech difficulties, numbness and impairments in cognition⁹. Cognition is scored as normal, mood alteration only, mild, moderate or marked decrease in mentation, or dementia; however, mentation is not defined and consequently the evaluation is highly subjective. In addition, fatigue and depression are not evaluated by the EDSS but these may affect cognition.

Regulatory authorities recognize that the EDSS is heavily influenced by physical functioning and does not fully capture the lives of patients with MS¹⁰. EDSS scores have been shown to correlate with measures of physical function, but not with neuropsychological assessments¹¹, suggesting that the EDSS is influenced most heavily by the assessment of physical disability. Additionally, EDSS scores were found to only weakly correlate with cognitive function in a clinical trial of siponimod in patients with SPMS¹². Therefore, cognitive function might not have much influence on EDSS scores.

Limitations of the EQ-5D in evaluating cognitive decline

The EQ-5D is a generic, preference-based instrument used to measure health status¹³. Patients complete the questionnaire themselves, scoring their health on five dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression; this profile is then used to derive a single summary index¹⁴. Although the EQ-5D is designed to be used across a range of conditions, there is evidence that the EQ-5D may not be influenced by changes in cognitive function¹⁵.

In a study evaluating HRQL in older adults with mobility impairment, none of the domains of the EQ-5D were found to correlate to the Mini-Mental State Examination score, a measure of cognitive function¹⁵. This may be because the dimensions of the EQ-5D do not cover cognition; the only dimension likely to be affected is the dimension describing usual activities. Additionally, the EQ-5D is subjective and self-reported. This is important to capture the impact of a condition on a patient but relies on their ability to identify and recognize their feelings, an ability often impaired (known as alexithymia) in MS¹⁶, and patients’ recall, which may be limited by cognitive dysfunction. Self-report may therefore be inappropriate in MS because if patients cannot recognize their own limitations they are unable to report them¹⁷. Indeed, one study found that participants with moderate cognitive impairment reported better EQ-5D scores than those with mild or no impairment¹⁸. Additionally, in a systematic review of patients with dementia, EQ-5D scores were consistently higher when collected by self-report than those collected by a proxy. These findings may reflect patients’ lack of insight and awareness¹⁹.

Alternatives to the EDSS and EQ-5D

If the EDSS and EQ-5D do not adequately capture changes in cognitive function, changes in cognitive decline may not be reflected in cost-effectiveness evaluations using these measures. Consequently, treatments that slow or stop cognitive decline may be undervalued, potentially limiting reimbursement and reducing patient access to these treatments.

There are various tests to evaluate cognitive functioning, the most common being the Symbol Digit Modalities Test (SDMT), which evaluates processing speed and is included in most neuropsychological batteries (e.g. the Brief International Cognitive Assessment for MS)²⁰. The SDMT has high reliability, good sensitivity, is quick to administer and doesn’t require a trained psychologist. Consequently, the SDMT may be easy to incorporate into clinical trials to capture cognition. To ensure cognition is considered alongside key outcomes, measures of cognition could be incorporated into composite endpoints. One composite endpoint that has been developed combines EDSS score with the SDMT²¹.

In addition, members of the EuroQol group are developing “bolt-on” versions of the EQ-5D that include additional areas of functioning or symptoms, such as the EQ-5D + C, which attempts to measure loss of cognitive function²². There is evidence that the EQ-5D + C performs slightly better than the EQ-5D in patients with traumatic brain injury²³. Other generic measures of HRQL (e.g. the SF-6D and the Health Utilities Index-3) may be more sensitive to changes in cognitive function than the EQ-5D and could help to address this measurement issue. However, all self-reported instruments may still be affected by a lack of patient insight, which may leave us reliant on the use of proxy HRQL assessments. Finally, one approach that may avoid these issues is to evaluate the impact of cognitive decline on people’s capabilities. For example, the ICECAP measure captures capabilities and has been shown to be more sensitive than the EQ-5D in reflecting changes in cognition in older adults¹⁵.

Conclusion

The EDSS and EQ-5D, which are commonly used in economic evaluations in MS, are influenced by physical disability, an important outcome in MS, but may not sufficiently reflect changes in cognitive function. Consequently, the effectiveness of new treatments at limiting cognitive decline may not be adequately reflected or valued in economic evaluations, potentially limiting reimbursement and patient access.

Transparency

Declaration of funding

Acaster Lloyd Consulting Ltd and Oxford PharmaGenesis received funding from Novartis Pharma AG.

Author contributions:

A.L., H.S. and N.A. drafted the manuscript, critically reviewed all versions and approved the final version.

Declaration of financial/other relationships

A.L. has disclosed that he is an employee of Acaster Lloyd Consulting Ltd, London, UK. H.S. has disclosed that she is a paid employee of Oxford PharmaGenesis, Oxford, UK. N.A. has disclosed that he is a paid employee of Novartis Pharma AG, Basel, Switzerland. CMRO peer reviewers on this manuscript have received an honorarium from CMRO for their review work but have no relevant financial or other relationships to disclose.

Acknowledgements

The authors thank John Findlay at Oxford PharmaGenesis, Oxford, UK for medical writing support.