TY - JOUR

T1 - Assessing the value of data for prediction policies

T2 - The case of antibiotic prescribing

AU - Huang, Shan

AU - Ribers, Michael Allan

AU - Ullrich, Hannes

N1 - Funding Information: We thank Tomaso Duso, Christian Peukert, Maximilian Sch?fer, and seminar participants at DIW Berlin, the University of Copenhagen, and the University of Kassel for helpful comments and Herlev/Hvidovre hospitals for generously sharing their data. We are indebted to Lars Bjerrum and Gloria Cristina Cordoba Currea for providing expertise on diagnostics and antibiotic prescribing in Danish primary care and to Jenny Dahl Knudsen, Sidsel Kyst, and Rolf Magnus Arpi for enabling us to work with laboratory data. We thank Adam Lederer for proofreading. This work was funded by the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (grant agreement no. 802450). Funding Information: We thank Tomaso Duso, Christian Peukert, Maximilian Schäfer, and seminar participants at DIW Berlin, the University of Copenhagen, and the University of Kassel for helpful comments and Herlev/Hvidovre hospitals for generously sharing their data. We are indebted to Lars Bjerrum and Gloria Cristina Cordoba Currea for providing expertise on diagnostics and antibiotic prescribing in Danish primary care and to Jenny Dahl Knudsen, Sidsel Kyst, and Rolf Magnus Arpi for enabling us to work with laboratory data. We thank Adam Lederer for proofreading. This work was funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement no. 802450 ). Publisher Copyright: © 2022 The Author(s)

PY - 2022/4

Y1 - 2022/4

N2 - We quantify the value of data for the prediction policy problem of reducing antibiotic prescribing to curb antibiotic resistance. Using varying combinations of administrative data, we evaluate machine learning predictions for diagnosing bacterial urinary tract infections and the outcomes of prescription rules based on these predictions. Simple patient demographics improve prediction quality substantially but larger reductions in prescribing can be achieved by making use of rich health data. Our results suggest decreasing returns to data for prediction quality and increasing returns for policy outcomes. Hence, data needs for prediction policy problems must be assessed based on the policy objective and not only on prediction quality.

AB - We quantify the value of data for the prediction policy problem of reducing antibiotic prescribing to curb antibiotic resistance. Using varying combinations of administrative data, we evaluate machine learning predictions for diagnosing bacterial urinary tract infections and the outcomes of prescription rules based on these predictions. Simple patient demographics improve prediction quality substantially but larger reductions in prescribing can be achieved by making use of rich health data. Our results suggest decreasing returns to data for prediction quality and increasing returns for policy outcomes. Hence, data needs for prediction policy problems must be assessed based on the policy objective and not only on prediction quality.

KW - Administrative data

KW - Antibiotic prescribing

KW - Machine learning

KW - Prediction policy problem

KW - Value of data

U2 - 10.1016/j.econlet.2022.110360

DO - 10.1016/j.econlet.2022.110360

M3 - Journal article

AN - SCOPUS:85125664848

VL - 213

JO - Economics Letters

JF - Economics Letters

SN - 0165-1765

M1 - 110360

ER -