METHODS: NICE diagnostics guidance were checked by a reviewer. For each piece of guidance on prognostics-related or predictive technologies identified, information was extracted by 1 reviewer, and checked by a second reviewer, with differences resolved by discussion.

RESULTS: A linked evidence approach was the predominant method used in economic models, in which evidence for test performance was linked to decisions about treatment and the impact of this on longer term health. Sources of data for long term prognosis were varied, including trials and real-world evidence. Issues raised at committee meetings included the impact on people of knowing the prognosis of their condition and how knowledge of prognosis impacts on treatment decisions. Uncertainties about test performance and long-term outcomes are also frequently seen and addressed through sensitivity analyses.

CONCLUSIONS: Decision modelling is frequently used in NICE DAP assessments to estimate how the performance of prognostic and predictive tests can be extrapolated onto downstream clinical outcomes. This raises a number of challenges which are described here, potentially identifying areas for methodological research.