IMPROVING THE MODEL FOR THE UK MS DISEASE MODIFYING TREATMENT RISK SHARING SCHEME ANALYSIS: A NEW NATURAL HISTORY DATASET

Background Although in 2002 the National Institute for Clinical Excellence, in 2002, concluded that the disease modifying treatments (DMTs) for MS–interferon–b and glatiramer acetate–were not cost effective over the short–term, it was recognised that longer–term benefits were possible. The ‘UK risk sharing scheme’ was initiated in order to deliver these drug cost–effectively by monitoring a cohort of MS patients over a 10 year period after starting a DMT, and if necessary adjusting the cost to meet a 20 year target of £36,000 per Quality Adjusted Life Year. The first (2 year) analysis,1 used a natural history dataset from London, Ontario, Canada as the comparator cohort to estimate the transition probabilities. However, the model proved too susceptible to change in the sensitivity analysis, mainly related to the artificial ‘smoothing’ of key disability–related (Expanded Disability Status Scale (EDSS)) data which prevented scores from being recorded as improving. Thus the scientific advisory committee advised that an alternative data set should be sought where the actual EDSS scores were accessible. It was agreed that access to the dataset to allow validation of different models was important and that the original Discrete Markov model used would be compared to a Continuous Model to allow potential covariates and out of window EDSS scores to be used. Methods A review of MS databases was performed and the British Columbia MS, Canada (BCMS) dataset was considered the most suitable. A subgroup of patients who fulfilled the 2001 ABN criteria for eligibility for DMTs were selected to act as a ‘natural history‘ comparator for the UK cohort. Only EDSS scores prior to the availability of DMTs in BC were included (1980–1995). Discrete and continuous Markov models with and without baseline covariates (onset age, disease duration, MS severity scale, gender) were tested. Probabilities of changes in EDSS (i.e. transition probabilities) were used to predict disability (EDSS) at year 10, relative to baseline EDSS (taken at the first date the patients fulfilled the ABN eligibility criteria). The predicted EDSS was then compared to the actual outcome. Having identified the most accurate mathematical model from the entire eligible BCMS dataset, this was verified by using data from a randomly selected half of the cohort to predict the 10 year progress of the other half. Results 978 BCC patients were selected as suitable for the comparator data set and were similar in baseline characteristics to the UK RSS cohort: 74% were female, with mean; onset age 29.14 yrs, age at eligibility to receive DMTs 37.3 yrs, disease duration 8.16 yrs and a 2.85 relapses in the prior 2 years. The best model at predicting outcome was the continuous Markov model with age at onset as the single, binary covariate, split by the median (27.9 years). Conclusion The use of the BCMS dataset as a ‘natural history’ comparator cohort has allowed us to develop a more reliable model, to analyse the cost effectiveness of the DMTs in the UK risk sharing scheme. This BCMS dataset and model will be used in the price adjustment analysis for the 4 and 6 year results. We gratefully acknowledge the BC MS Clinic neurologists who contributed to the BCC data base, and to the UK neurologists, nurses and administrational staff who have been key in collecting the RSS data.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.