key: cord-0937974-jtz5abc6
authors: Gravesteijn, Benjamin; Krijkamp, Eline; van Alphen, Anouk; Busschbach, Jan; Geleijnse, Geert; Helmrich, Isabel Retel; Bruinsma, Sophie; van Lint, Céline; van Veen, Ernest; Steyerberg, Ewout; Verhoef, Kees; van Saase, Jan; Lingsma, Hester; Baatenburg de Jong, Rob
title: Update on a Model to Minimize Population Health Loss in Times of Scarce Surgical Capacity During the COVID-19 Crisis and Beyond
date: 2022-03-30
journal: Value Health
DOI: 10.1016/j.jval.2022.01.027
sha: 923375db07d4b42d7a29ed9e046b552f4372096e
doc_id: 937974
cord_uid: jtz5abc6

nan

Previously, we published a decision model to guide prioritization of semielective surgeries in times of operation room scarcity across disciplines from a utilitarian perspective. 1 The model estimates expected health loss because of delay of surgery for these procedures, for instance because of high incidence of patients with COVID-19.

During further model development, we identified inconsistencies in model output. These inconsistencies could eventually be explained by a coding error in the original model. The error involves the following mistake: Instead of dividing the weeks spent in perfect quality of life (the primary outcome of the model) by the number of weeks in a year (52 weeks), we divided the weeks spent in perfect quality of life by the number of simulated years for the cohort (eg, time between average age at start of the simulation until 100 years of age, interquartile range 34.9-44.0). Therefore, we did not calculate disability-adjusted lifeyears correctly.

The consequences of this error are that the urgency for younger patients was slightly underestimated, whereas the urgency for older patients was slightly overestimated. Therefore, the urgency ranking of the diseases changes somewhat and Figure 1 of the publication was updated. Bypass surgery for Fontaine III/IV peripheral arterial disease is still the highest ranked procedure, but total nephrectomy for renal cancer ranks second instead of transaortic valve implantation (which dropped to rank 9). Placing a shunt for dialysis in end-stage renal disease still ranks last, but bullectomy for chronic obstructive pulmonary disease now ranks second last instead of resection of thyroid cancer. The updated model output can be accessed through our user-friendly web application: https://tinyurl.com/ y2yzudgw.

We would like to emphasize that the article as presented in Value in Health was a proof-of-concept description of the model. The message conveyed by the article does not change by our discovery of the coding error. Therefore, we will continue to develop the model and facilitate and evaluate its implementation. Because coding errors are inherent to model development, we want to once more raise awareness for systematic inconsistency checks to remove the most significant errors. 

Minimizing population health loss in times of scarce surgical capacity during the coronavirus disease 2019 crisis and beyond: a modeling study