OBJECTIVES: New drugs launching into Germany undergo benefit assessment by G-BA, followed by a price negotiation with GKV-Spitzenverband (GKV-SV). In this paper we analyze the determinants of German rebate outcomes for oncology drugs. METHODS: We used the (Coveragedecisions} payer decision database as the basis of this analysis. This database consists of systematically abstracted data from publicly available payer decisions on pricing, reimbursement, HTA and formulary coverage decisions from 21 countries worldwide, covering all new molecular entities (NMEs) approved in EU or US since 2011. We extracted data from the (Coveragedecisions} payer decision database on G-BA benefit assessment and GKV-SV pricing outcomes of all oncology NMEs with a known GKV-SV pricing outcome. The data extracted included G-BA benefit rating at subgroup level, relative sizes of subgroups, intervention and comparator costs by subgroups, time limitations on benefit determination and magnitude of rebate. We then conducted a stepwise multiple linear regression with backward elimination, with magnitude of post-AMNOG rebate as the dependent variable, and benefit category, difference between intervention and comparator costs and presence of time limit on benefit recommendation as the independent variables. RESULTS: Of the 27 benefit assessments analyzed, rebates varied from 10%-54%. The final model (adjusted R square 0.75; F statistic 9.35; p<0.001) showed that the difference between intervention and comparator costs was a significant predictor of rebate (p=0.004). Compared to 'indication of considerable benefit', ‘no benefit’ (p=0.001) and ‘hint of minor benefit’ (p<0.001) were associated with significantly higher rebates. Though other benefit ratings did not achieve statistical significance, directionally, 'proof of minor benefit' and 'indication of lower benefit' were associated with higher rebates compared to 'indication of considerable benefit'.  CONCLUSIONS: Based on our analysis, GKV rebate for oncology drugs can be expressed using a multiple regression equation, as a function of G-BA benefit rating category and incremental cost vs. comparator.