Methods of Model Calibration

Taylor, Douglas C.A.; Pawar, Vivek; Kruzikas, Denise; Gilmore, Kristen E.; Pandya, Ankur; Iskandar, Rowan; Weinstein, Milton C.

doi:10.2165/11538660-000000000-00000

Cited by 34 publications

(29 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The current analysis relied on visual assessment of trend, acceptance window (95% confidence interval of patient-level data), and percentage deviation from the mean, all of which are consistent with practical applications in the literature and can be implemented in settings with limited computing capacity and statistical support. 25,32,33,38,39 …”

Section: Discussionmentioning

confidence: 99%

“…In addition, no guidelines exist regarding determining how to best search the parameter space and when it has been adequately searched. While more efficient algorithms (e.g., Nelder-Mead) could have been used, 33,47 we preferred to use methods that were straightforward, were reproducible, and used fewer computer resources, but also that performed comparably to more sophisticated methods. 33 …”

Section: Discussionmentioning

confidence: 99%

“…32 We also calculated the mean of the percentage deviation between the model output and the target data at each available time point. 33 …”

Section: Methodsmentioning

confidence: 99%

“…Among these good-fitting sets, we identified the top sets of input values that minimized the percentage deviation between model predictions and the empiric data of the number on ART between 2004 and 2009. 33,34 We explored alternative thresholds for the number of good-fitting sets by evaluating the top 5, top 10, top 50, and top 75 sets minimizing the percentage deviation. We also varied the threshold for a good-fitting set between 5% and 30%.…”

Section: Methodsmentioning

confidence: 99%

See 3 more Smart Citations

Performance of a Mathematical Model to Forecast Lives Saved from HIV Treatment Expansion in Resource-Limited Settings

et al. 2014

View full text Add to dashboard Cite

Background International guidelines recommend HIV treatment expansion in resource-limited settings, but funding availability is uncertain. We evaluated performance of a model that forecasts lives saved through continued HIV treatment expansion in Haiti. Methods We developed a computer-based, mathematical model of HIV disease and used incidence density analysis of patient-level Haitian data to derive model parameters for HIV disease progression. We assessed internal validity of model predictions and internally calibrated model inputs when model predictions did not fit the patient-level data. We then derived uncertain model inputs related to diagnosis and linkage to care, pre-treatment retention, and enrollment on HIV treatment through an external calibration process that selected input values by comparing model predictions to Haitian population-level data. Model performance was measured by fit to event-free survival (patient-level) and number receiving HIV treatment over time (population-level). Results For a cohort of newly HIV-infected individuals with no access to HIV treatment, the model predicts median AIDS-free survival of 9.0 years pre-calibration and 6.6 years post-calibration versus 5.8 years (95% CI 5.1, 7.0) from the patient-level data. After internal validation and calibration, 16 of 17 event-free survival measures (94%) had a mean percentage deviation between model predictions and the empiric data of <6%. After external calibration, the percentage deviation between model predictions and population-level data on the number on HIV treatment was <1% over time. Conclusions Validation and calibration resulted in a good-fitting model appropriate for health policy decision making. Using local data in a policy model-building process is feasible in resource-limited settings.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

“…32 We also calculated the mean of the percentage deviation between the model output and the target data at each available time point. 33 …”

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 2 more Smart Citations

Performance of a Mathematical Model to Forecast Lives Saved from HIV Treatment Expansion in Resource-Limited Settings

et al. 2014

View full text Add to dashboard Cite

show abstract

“…The commonly used methods for automatic model calibration are mainly based on the gradient-based methods [27], Nelder-Mead [28], and linear regression [29]. These deterministic search methods can provide solutions quickly, but they may get trapped into local optima easily [30].…”

Section: Related Workmentioning

confidence: 99%

Density-based evolutionary framework for crowd model calibration

Zhong

Cai

et al. 2015

Journal of Computational Science

View full text Add to dashboard Cite

Impact of Expanding Eligibility Criteria for Cochlear Implantation – Dynamic Modeling Study

et al. 2022

View full text Add to dashboard Cite

Objectives Eligibility criteria for cochlear implantation (CI) are shifting due to technological and surgical improvements. The aim of this study was to explore the impact of further expanding unilateral CI criteria in those with severe hearing loss (HL) (61–80 dBHL) in terms of number of CI recipients, costs, quality of life, and cost‐effectiveness. Methods A dynamic population‐based Markov model was constructed mimicking the Dutch population in three age categories over a period of 20 years. Health states included severe HL (61–80 dBHL), profound HL (>81 dBHL), CI recipients, and no‐CI recipients. Model parameters were based on published literature, (national) databases, expert opinion, and model calibration. Results If persons with severe HL would qualify and opt for CI similar to those with profound HL now, this would lead to a 6–7 times increase of new CI recipients and an associated increase in costs (€550 million) and QALYs (54.000) over a 20‐year period (incremental cost utility ratio: 10.771 euros/QALY [2.5–97.5 percentiles: 1.252–23.171]). One‐way‐sensitivity analysis indicated that model outcomes were most sensitive to regaining employment, utility associated with having a CI, and costs of surgery and testing. Conclusion Our findings suggest that expanding eligibility for CI to persons with severe HL could be a cost‐effective use of resources. Clearly, however, it would require a significant increase in diagnostic, operative, and rehabilitative capacity. Our quantitative estimates can serve as a basis for a wider societal deliberation on the question whether such an increase can and should be pursued. Level of Evidence NA Laryngoscope, 133:924–932, 2023

show abstract

Methods of Model Calibration

Abstract: The Nelder-Mead and manual techniques may be preferable calibration methods based on both performance and efficiency, provided that sufficient resources are available.

Cited by 34 publications

References 13 publications

Performance of a Mathematical Model to Forecast Lives Saved from HIV Treatment Expansion in Resource-Limited Settings

Performance of a Mathematical Model to Forecast Lives Saved from HIV Treatment Expansion in Resource-Limited Settings

Density-based evolutionary framework for crowd model calibration

Impact of Expanding Eligibility Criteria for Cochlear Implantation – Dynamic Modeling Study

Contact Info

Product

Resources

About