VP39 The Alphabet Lottery? How NICE Outcomes Vary By Appraisal Committee

Copyright © Cambridge University Press 20192019Cambridge University PressIntroductionNICE (National Institute for Health and Care Excellence) makes recommendations on the public reimbursement of medicines based on their clinical- and cost-effectiveness. The recommendation is made by an Appraisal Committee (comprising a multi-disciplinary group of independent experts) as part of a technology appraisal. There are four Appraisal Committees (A,B,C,D); this research investigates whether appraisal outcomes vary by committee.MethodsAll publicly-available Final Appraisal Determinations from NICE Single Technology Appraisals (STA) were screened (01/10/2009-14/11/2018) and key data were extracted. Homogeneity in rates of acceptance or rejection across the committees was assessed using Chi-squared tests.ResultsThe Appraisal Committee was identified for 298 technologies, 56% (168/298) of which were ‘recommended’. The number of technologies assessed by each committee was similar (A:79, B:62, C:91, D:66). However, STAs conducted by Committee D were significantly less likely to receive ‘recommended’ outcomes (A:68% [54/79], B:65% [40/62], C:53% [48/91], D:39% [26/66]; p < 0.01). STAs for oncology indications had higher ’not recommended’ outcomes than those for non-oncology indications (25% vs. 9%). The lower ‘recommendation’ rates for committee D persisted across oncology (A:60%, B:83%, C:50%, D:38%; p = 0.01) and non-oncology indications (A:73%, B:53%, C:55%, D:40%; p < 0.01). However, STAs conducted by Committee D were significantly more likely to receive ‘optimized’ recommendations (A:16%, B:21%, C:33%, D: 36%; p < 0.01) and when considering the rates of ‘recommended’ and ‘optimized’ outcomes compared to ‘only in research’ and ‘not recommended’ outcomes, no significant differences were found (A:85%, B: 85%, C:86%, D:76%; p = 0.27).ConclusionsSTAs undertaken by NICE Appraisal Committee D was associated with a significantly lower rate of ‘recommended’ outcomes but tended to an ‘optimized’ recommendation significantly more than the other committees. Further research is needed to determine if this reflects any deviation in uniform implementation of NICE methodology between Committees.