Abstract Introduction The Cancer Drugs Fund (CDF) was reformed in 2016 with an ambition to generate new evidence on effectiveness, and to review existing drugs in the fund. We set out to evaluate: whether drugs transitioning from the old CDF were re-reviewed as planned; whether new drugs have a “data collection arrangement” (DCA) as planned; and whether evidence generated under the DCA using routine data from the “Systemic Anti-Cancer Treatment” (SACT) database was of high quality. Methods We accessed documents from NHS England, Public Health England and NICE at August 2018. We calculated the proportion of old CDF drugs re-reviewed, and of new drugs and indications with a DCA. We described key features of the DCAs. For all SACT studies we set out to obtain a protocol in order to analyse the quality of the planned methods. Results 47 old drugs and indications transitioned to the new CDF. For 14 there was no evidence of a re-review; 9 of these remain under CDF at August 2019 (all off-label uses). 33 had marketing authorisations: 22 of these had a review completed by September 2017 as planned (67%). 20 new drugs and indications entered the CDF by August 2018: 19 had a DCA; one (off-label) had no DCA or equivalent. All DCAs identify uncertainty about overall survival; all express an intent to conduct observational analysis using SACT data; SACT data was central to decision-making for 6 (32%). We were able to find 0 protocols of the 19 planned SACT studies (0). Following Freedom of Information requests we were told these protocols are prepared after the data are collected, and posted with the reappraisal: however we could not locate any protocol for either of the two published re-appraisals. We were therefore unable to assess the quality of the methods in any of the proposed SACT studies. Conclusions The revised CDF has not been implemented as planned. Reporting of observational analyses in SACT data fall substantially short of best practice, and the full methods used cannot be established. There is very little information in the public domain around evaluation of off-label uses. Lastly, SACT data itself does not appear to be able to support clinical decision-making in the manner suggested by the CDF policy documents. NHS England should review the conduct of the fund, but also the planning, as unrealistic commitments may have been made.