Prediction impact curve is a new measure integrating intervention effects in the evaluation of risk models |
| |
| Institution: | 1. Department of Epidemiology, Rollins School of Public Health, Emory University, 1518 Clifton Road NE, Atlanta, GA, USA;2. Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, PO Box 281, SE-171 77 Stockholm, Sweden;3. Department of Medical Sciences, Molecular Epidemiology and Science for Life Laboratory, Uppsala University, BOX 1115, 751 41 Uppsala, Sweden;4. Department of Clinical Genetics, Section Community Genetics, EMGO Institute for Health and Care Research, VU University Medical Center, PO Box 7057, 1007 MB Amsterdam, The Netherlands;1. Women’s Health Research Unit, Barts and The London School of Medicine, Queen Mary, University of London, Turner Street, London E1 2AB, United Kingdom;2. Barts and The London School of Medicine and Dentistry, London E1 2AB, United Kingdom;3. Barts Health NHS Trust, The Royal London Hospital, Whitechapel Road, London E1 1BB, United Kingdom;1. Centre for Clinical Epidemiology & Biostatistics, School of Medicine & Public Health, University of Newcastle, Newcastle, Australia;2. Deakin Biostatistics Unit, Faculty of Health, Deakin University, Melbourne, Australia;3. Department of Medical and Health Sciences, Linköping University, Linköping, Sweden;4. Department of Health Sciences, University of York, York, England;1. Department of Internal Medicine, Federal Armed Forces Central Hospital, Koblenz, Germany;2. Department of General and Interventional Cardiology, University Heart Center Hamburg, Hamburg, Germany;3. Department of Medicine II, GPR Rüsselsheim, Rüsselsheim, Germany;4. Institute of Clinical Chemistry and Laboratory Medicine, University Medical Center of the Johannes Gutenberg-University Mainz, Mainz, Germany;5. ICAN Institute for Cardiometabolism and Nutrition, Paris, France;6. Sorbonne Universités, UPMC Univ. Paris 06, INSERM, UMR_S 1166, Team Genomics & Pathophysiology of Cardiovascular Diseases, Paris, France;1. Department of Medicine, University of Ottawa, 85 Primrose Ave, Ottawa, ON, Canada K1N 6M1;2. Department of Clinical Epidemiology and Biostatistics, McMaster University, 1280 Main Street West, Hamilton, ON L8S 4K1;3. Republican Helthcare Development Centre, 4 Orynbor St., Astana, Kazakhstan, 010000;4. Canadian Society for International Health Project Manager, Astana, Kazakhstan;5. Canadian Society for International Health, One Nicholas, Suite 726, Ottawa, ON K1N 7B7;6. Clinical Department Head for Emergency Medicine, Alberta Health Services, Calgary Zone, Academic Department Head for Emergency Medicine, Associate Professor of Emergency Medicine, Cumming School of Medicine, University of Calgary, RGH, Holy Cross Ambulatory Care, 7007 - 14 St. SW - Room 5A105, Calgary, AB Canada T2V 1P9 |
| |
| Abstract: | ObjectiveWe propose a new measure of assessing the performance of risk models, the area under the prediction impact curve (auPIC), which quantifies the performance of risk models in terms of their average health impact in the population.Study Design and SettingUsing simulated data, we explain how the prediction impact curve (PIC) estimates the percentage of events prevented when a risk model is used to assign high-risk individuals to an intervention. We apply the PIC to the Atherosclerosis Risk in Communities (ARIC) Study to illustrate its application toward prevention of coronary heart disease.ResultsWe estimated that if the ARIC cohort received statins at baseline, 5% of events would be prevented when the risk model was evaluated at a cutoff threshold of 20% predicted risk compared to 1% when individuals were assigned to the intervention without the use of a model. By calculating the auPIC, we estimated that an average of 15% of events would be prevented when considering performance across the entire interval.ConclusionWe conclude that the PIC is a clinically meaningful measure for quantifying the expected health impact of risk models that supplements existing measures of model performance. |
| |
| Keywords: | Prediction impact curve AUC Risk model Predictive model Coronary heart disease Predictive ability |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
