Dataset Information

Calibrating predictive model estimates to support personalized medicine.

ABSTRACT: Predictive models that generate individualized estimates for medically relevant outcomes are playing increasing roles in clinical care and translational research. However, current methods for calibrating these estimates lose valuable information. Our goal is to develop a new calibration method to conserve as much information as possible, and would compare favorably to existing methods in terms of important performance measures: discrimination and calibration.We propose an adaptive technique that utilizes individualized confidence intervals (CIs) to calibrate predictions. We evaluate this new method, adaptive calibration of predictions (ACP), in artificial and real-world medical classification problems, in terms of areas under the ROC curves, the Hosmer-Lemeshow goodness-of-fit test, mean squared error, and computational complexity.ACP compared favorably to other calibration methods such as binning, Platt scaling, and isotonic regression. In several experiments, binning, isotonic regression, and Platt scaling failed to improve the calibration of a logistic regression model, whereas ACP consistently improved the calibration while maintaining the same discrimination or even improving it in some experiments. In addition, the ACP algorithm is not computationally expensive.The calculation of CIs for individual predictions may be cumbersome for certain predictive models. ACP is not completely parameter-free: the length of the CI employed may affect its results.ACP can generate estimates that may be more suitable for individualized predictions than estimates that are calibrated using existing methods. Further studies are necessary to explore the limitations of ACP.

SUBMITTER: Jiang X

PROVIDER: S-EPMC3277613 | biostudies-literature | 2012 Mar-Apr

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Calibrating predictive model estimates to support personalized medicine.

Jiang Xiaoqian X Osl Melanie M Kim Jihoon J Ohno-Machado Lucila L

Journal of the American Medical Informatics Association : JAMIA 20111007 2

<h4>Objective</h4>Predictive models that generate individualized estimates for medically relevant outcomes are playing increasing roles in clinical care and translational research. However, current methods for calibrating these estimates lose valuable information. Our goal is to develop a new calibration method to conserve as much information as possible, and would compare favorably to existing methods in terms of important performance measures: discrimination and calibration.<h4>Material and me ...[more]

PMID: 21984587

Similar Datasets

Project description:Regenerative medicine is a multidisciplinary field that aims to determine different factors and develop various methods to regenerate impaired tissues, organs, and cells in the disease and impairment conditions. When treatment procedures are specified according to the individual's information, the leading role of personalized regenerative medicine will be revealed in developing more effective therapies. In this concept, endocrine disorders can be considered as potential candidates for regenerative medicine application. Diabetes mellitus as a worldwide prevalent endocrine disease causes different damages such as blood vessel damages, pancreatic damages, and impaired wound healing. Therefore, a global effort has been devoted to diabetes mellitus investigations. Hereupon, the preclinical study is a fundamental step. Up to now, several species of animals have been modeled to identify the mechanism of multiple diseases. However, more recent researches have been demonstrated that animal models with the ability of tissue regeneration are more suitable choices for regenerative medicine studies in endocrine disorders, typically diabetes mellitus. Accordingly, zebrafish has been introduced as a model that possesses the capacity to regenerate different organs and tissues. Especially, fine regeneration in zebrafish has been broadly investigated in the regenerative medicine field. In addition, zebrafish is a suitable model for studying a variety of different situations. For instance, it has been used for developmental studies because of the special characteristics of its larva. In this review, we discuss the features of zebrafish that make it a desirable animal model, the advantages of zebrafish and recent research that shows zebrafish is a promising animal model for personalized regenerative diseases. Ultimately, we conclude that as a newly introduced model, zebrafish can have a leading role in regeneration studies of endocrine diseases and provide a good perception of underlying mechanisms.

Project description:BackgroundQuantification of motor performance has a promising role in personalized medicine by diagnosing and monitoring, e.g. neurodegenerative diseases or health problems related to aging. New motion assessment technologies can evolve into patient-centered eHealth applications on a global scale to support personalized healthcare as well as treatment of disease. However, uncertainty remains on the limits of generalizability of such data, which is relevant specifically for preventive or predictive applications, using normative datasets to screen for incipient disease manifestations or indicators of individual risks.ObjectiveThis study explored differences between healthy German and Japanese adults in the performance of a short set of six motor tests.MethodsSix motor tasks related to gait and balance were recorded with a validated 3D camera system. Twenty-five healthy adults from Chiba, Japan, participated in this study and were matched for age, sex, and BMI to a sample of 25 healthy adults from Berlin, Germany. Recordings used the same technical setup and standard instructions and were supervised by the same experienced operator. Differences in motor performance were analyzed using multiple linear regressions models, adjusted for differences in body stature.ResultsFrom 23 presented parameters, five showed group-related differences after adjustment for height and weight (R 2 between .19 and .46, p<.05). Japanese adults transitioned faster between sitting and standing and used a smaller range of hand motion. In stepping-in-place, cadence was similar in both groups, but Japanese adults showed higher knee movement amplitudes. Body height was identified as relevant confounder (standardized beta >.5) for performance of short comfortable and maximum speed walks. For results of posturography, regression models did not reveal effects of group or body stature.ConclusionsOur results support the existence of a population-specific bias in motor function patterns in young healthy adults. This needs to be considered when motor function is assessed and used for clinical decisions, especially for personalized predictive and preventive medical purposes. The bias affected only the performance of specific items and parameters and is not fully explained by population-specific ethnic differences in body stature. It may be partially explained as cultural bias related to motor habits. Observed effects were small but are expected to be larger in a non-controlled cross-cultural application of motion assessment technologies with relevance for related algorithms that are being developed and used for data processing. In sum, the interpretation of individual data should be related to appropriate population-specific or even better personalized normative values to yield its full potential and avoid misinterpretation.Supplementary informationThe online version contains supplementary material available at 10.1007/s13167-021-00236-3.

Dataset Information

Calibrating predictive model estimates to support personalized medicine.

Publications

Calibrating predictive model estimates to support personalized medicine.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets