==== Front Emerg Infect Dis Emerg Infect Dis EID Emerging Infectious Diseases 1080-6040 1080-6059 Centers for Disease Control and Prevention 23-0489 10.3201/eid2907.230489 Letters to the Editor Letters to the Editor Challenges in Forecasting Antimicrobial Resistance Challenges in Forecasting Antimicrobial Resistance Challenges in Forecasting Antimicrobial Resistance Aldeyab Mamoon A. Lattyak William J. University of Huddersfield, Huddersfield, UK (M.A. Aldeyab); Scientific Computing Associates Corp., River Forest, Illinois, USA (W.J. Lattyak) Address for correspondence: Mamoon A Aldeyab, Department of Pharmacy, School of Applied Sciences, University of Huddersfield, Huddersfield HD1 3DH, UK; email: m.aldeyab@hud.ac.uk 7 2023 29 7 14961496 2023 https://creativecommons.org/licenses/by/4.0/ Emerging Infectious Diseases is a publication of the U.S. Government. This publication is in the public domain and is therefore without copyright. All text from this work may be reprinted freely. Use of these materials should be properly cited. PeiS , BlumbergS , VegaJC , RobinT , ZhangY , MedfordRJ , et al. ; CDC MIND-Healthcare Program. CDC MIND-Healthcare Program. Challenges in forecasting antimicrobial resistance. Emerg Infect Dis. 2023;29 :679–85. 10.3201/eid2904.221552Keywords: predictive modeling infections antimicrobial resistance bacteria United Kingdom United States ==== Body pmcTo the Editor: We read with interest the article by Pei et al. (1), which discussed challenges in forecasting antimicrobial resistance (AMR) and emphasized the need for improved AMR predictive intelligence. We complement the authors on their findings and share our experience with a threshold-logistic modeling concept that we recently introduced to improve understanding of the relationship between antimicrobial drug use thresholds and incidence of resistant pathogens and a threshold transfer function model that can be used to project AMR prevalence (2,3). AMR modeling and forecasting are challenging because of the evolutionary ability of pathogens to adapt to changing environmental conditions. To be useful, a model should address time-varying effects of explanatory variables on response function and changes in the structural relationship between predictor and target variables. Concepts such as threshold transfer function modeling that includes autoregressive moving average components can partially address those issues. Model recalibration is essential and dictated by breakdown in predictive ability. Model parameters can be reestimated after recording new observations. Model structure does not change, but parameter estimates are updated to reflect new information. If statistical significance of a parameter falls below a defined confidence level, model structure might need modification. In practice, model estimation and assessment can be automated. Forecasting model accuracy can also be automated by comparing ongoing performance to baseline accuracy. When model or predictive performance degradation is flagged, a more comprehensive model recalibration is dictated. The time between model calibrations is unknown and depends on the stability of identified relationships in the model and degree to which evolutionary changes in AMR are observed. Further real-world testing is required to determine other factors that can explain resistance and define thresholds, find optimal interventions to reduce antimicrobial drug use to identified thresholds, and assess feasibility of implementing those interventions in daily clinical practice. Suggested citation for this article: Aldeyab MA, Lattyak WJ. Challenges in forecasting antimicrobial resistance. Emerg Infect Dis. 2023 Jul [date cited]. https://doi.org/10.3201/eid2907.230489 ==== Refs References 1. Pei S, Blumberg S, Vega JC, Robin T, Zhang Y, Medford RJ, et al. ; CDC MIND-Healthcare Program. CDC MIND-Healthcare Program. Challenges in forecasting antimicrobial resistance. Emerg Infect Dis. 2023;29 :679–85. 10.3201/eid2904.221552 36958029 2. Aldeyab MA, Bond SE, Conway BR, Lee-Milner J, Sarma JB, Lattyak WJ. Identifying antibiotic use targets for the management of antibiotic resistance using an extended-spectrum β-lactamase-producing Escherichia coli case: a threshold logistic modeling approach. Antibiotics (Basel). 2022;11 :1116. 10.3390/antibiotics11081116 36009985 3. Aldeyab MA, Bond SE, Conway BR, Lee-Milner J, Sarma JB, Lattyak WJ. A threshold logistic modelling approach for identifying thresholds between antibiotic use and methicillin-resistant Staphylococcus aureus incidence rates in hospitals. Antibiotics (Basel). 2022;11 :1250. 10.3390/antibiotics11091250 36140029