The price of precision: trade-offs between usability and validity in the World Health Organization Health Economic Assessment Tool for walking and cycling.

Abstract Objectives The widely used World Health Organization (WHO) Health Economic Assessment Tool (HEAT) for walking and cycling quantifies health impacts in terms of premature deaths avoided or caused as a result of changes in active transport. This article attempts to assess the effect of incorporating ‘life-years’ as an impact measure to increase the precision of the model and assess the effect on the tool's usability. Study design This article is a methods paper, using simulation to estimate the effect of a methodological change to the HEAT 4.2 physical activity module. Methods We use the widely used WHO HEAT for walking and cycling as a case study. HEAT currently quantifies health impacts in terms of premature deaths avoided or caused as a result of changes in active transport. We assess the effect of incorporating “duration of life gained” as an impact measure to increase the precision of the model without substantially affecting usability or increasing data requirements. Results Compared with the existing tool (HEAT version 4.2), which values premature deaths avoided, estimates derived by valuing life-years gained are more sensitive to the age of the population affected by an intervention, with results for older and younger age groups being markedly different between the two methods. This is likely to improve the precision of the tool, especially where it is applied to interventions that affect age groups differentially. The life-years method requires additional background data (obtained and used in this analysis) and minimal additional user inputs; however, this may also make the tool harder to explain to users. Conclusions Methodological improvements in the precision of widely used tools, such as the HEAT, may also inadvertently reduce their practical usability. It is therefore important to consider the overall impact on the tool's value to stakeholders and explore ways of mitigating potential reductions in usability.