OBJECTIVES: Comprehensive systematic reviews of clinical and cost-effectiveness evidence underpin company submissions to the UK National Institute for Health and Care Excellence (NICE) Single Technology Appraisal (STA) process for drug reimbursement. NICE mandates up-to-date, well-conducted and robust methods, however STA approaches vary between submissions. The objective of this study was to survey the sources searched and currency of clinical effectiveness evidence identification in NICE STA submissions.

METHODS: We conducted a survey of 50 randomly selected STA submissions included in committee papers published between 10.1.18 and 6.12.23. As NICE limit the content of committee papers in the public domain, data were extracted from company submissions, clarification responses and Evidence Assessment Group reports. For each STA submission, characteristics relating to the disease area, ERG/EAG, databases and grey literature searched, date of last search and submission date, and language restrictions, were extracted. An initial pilot sample of 10 STAs was assessed.

RESULTS: The pilot sample (n=10) included indications in oncology (7), cardiovascular (2) and gynecology (1). The mean number of databases searched per topic was 4.8 (range: 2-9).

The top databases searched were CENTRAL (n=8), Embase (n=6) and PubMed (n=5) and Medline (n=4). Four submissions ran combined Embase/Medline searches, and the same number also searched at least one trials register. 60% of submissions included conferences (range: 2-7). Additional grey literature searching was reported in only 30% of the sample STAs. Currency of searching was an issue; the mean time elapsed between search date and submission was 8 months (254 days; range: 54-483 days). One submission was updated 52 days post-submission, during the clarification process.

CONCLUSIONS: Due to the lack of completeness in STA committee papers in the public domain, many of the clinical and cost-effectiveness methods informing NICE reimbursement decisions lack detail, reproducibility and transparency. Currency of evidence searches could be improved.