Benjamin P Chapman1, Alexander Weiss, Kevin Fiscella, Peter Muennig, Ichiro Kawachi, Paul Duberstein. 1. *Department of Psychiatry, University of Rochester Medical Center, Rochester, NY †Department of Psychology, School of Philosophy, Psychology and Language Sciences, University of Edinburgh, South Bridge, Edinburgh, UK ‡Departments of Family Medicine and Community and Preventive Medicine, Family Medicine Research Programs, University of Rochester Medical Center, Rochester §Department of Health Policy and Management, Mailman School of Public Health, Columbia University, New York, NY ∥Harvard School of Public Health, Division of Social Epidemiology, Boston, MA.
Abstract
BACKGROUND: Predicting risk of premature death is one of the most basic tasks in medicine and public health, but has proven to be difficult over the long term even with the best prognostic models. One popular strategy has been to improve prognostic models with candidate genes and other novel biomarkers. However, the gains in predictive power have been modest and the costs have been high, leading to a demand for cost-effective alternatives. We conducted a proof-of-principle investigation to examine whether simple, cheap, and noninvasive paper-and-pencil measures of social class and personality phenotype could improve the performance of one of the most widely used prediction models for all-cause mortality, the Charlson Comorbidity Index (CCI). METHODS: We used data from baseline and 25-year mortality follow-up of the UK Health and Lifestyle Study cohort. In a subset of the cohort, we first identified 5 psychosocial factors highly predictive of mortality: income, education, type A personality, communalism (preference for the company of others), and "lie" scale (a measure of denial, putatively associated with ill health). We then examined the predictive performance of the CCI with and without these measures in a validation subsample. RESULTS: Across 5-, 10-, 15-, 20-, and 25-year time horizons, the psychosocially augmented CCI showed substantially better discrimination [area under the receiver-operating curves (95% confidence interval) from 0.83 (0.81-0.85) to 0.84 (0.83-0.86)] than the CCI [area under the receiver-operating curves from 0.74 (0.71-0.76) to 0.77 (0.76-0.79)]. These translated into net reclassification improvements from 27% (23%-31%) to 35% (32%-38%) of survivors and from 23% (17%-30%) to 34% (17%-30%) of decedents; and 23%-42% reductions in the Number Needed to Screen. Calibration improved at all time horizons except 25 years, where it was decreased. CONCLUSION: Widespread attempts to improve prognostic models might consider not only novel biomarkers, but also psychosocial questionnaire measures.
BACKGROUND: Predicting risk of premature death is one of the most basic tasks in medicine and public health, but has proven to be difficult over the long term even with the best prognostic models. One popular strategy has been to improve prognostic models with candidate genes and other novel biomarkers. However, the gains in predictive power have been modest and the costs have been high, leading to a demand for cost-effective alternatives. We conducted a proof-of-principle investigation to examine whether simple, cheap, and noninvasive paper-and-pencil measures of social class and personality phenotype could improve the performance of one of the most widely used prediction models for all-cause mortality, the Charlson Comorbidity Index (CCI). METHODS: We used data from baseline and 25-year mortality follow-up of the UK Health and Lifestyle Study cohort. In a subset of the cohort, we first identified 5 psychosocial factors highly predictive of mortality: income, education, type A personality, communalism (preference for the company of others), and "lie" scale (a measure of denial, putatively associated with ill health). We then examined the predictive performance of the CCI with and without these measures in a validation subsample. RESULTS: Across 5-, 10-, 15-, 20-, and 25-year time horizons, the psychosocially augmented CCI showed substantially better discrimination [area under the receiver-operating curves (95% confidence interval) from 0.83 (0.81-0.85) to 0.84 (0.83-0.86)] than the CCI [area under the receiver-operating curves from 0.74 (0.71-0.76) to 0.77 (0.76-0.79)]. These translated into net reclassification improvements from 27% (23%-31%) to 35% (32%-38%) of survivors and from 23% (17%-30%) to 34% (17%-30%) of decedents; and 23%-42% reductions in the Number Needed to Screen. Calibration improved at all time horizons except 25 years, where it was decreased. CONCLUSION: Widespread attempts to improve prognostic models might consider not only novel biomarkers, but also psychosocial questionnaire measures.
Authors: Richard D Riley; Greta Ridley; Katrina Williams; Douglas G Altman; Jill Hayden; Henrica C W de Vet Journal: J Clin Epidemiol Date: 2007-06-06 Impact factor: 6.437
Authors: Graeme Rocker; Deborah Cook; Peter Sjokvist; Bruce Weaver; Simon Finfer; Ellen McDonald; John Marshall; Anne Kirby; Mitchell Levy; Peter Dodek; Daren Heyland; Gordon Guyatt Journal: Crit Care Med Date: 2004-05 Impact factor: 7.598
Authors: David Cella; Susan Yount; Nan Rothrock; Richard Gershon; Karon Cook; Bryce Reeve; Deborah Ader; James F Fries; Bonnie Bruce; Mattias Rose Journal: Med Care Date: 2007-05 Impact factor: 2.983
Authors: Wouter de Ruijter; Rudi G J Westendorp; Willem J J Assendelft; Wendy P J den Elzen; Anton J M de Craen; Saskia le Cessie; Jacobijn Gussekloo Journal: BMJ Date: 2009-01-08
Authors: Maria Rosaria A Muscatello; Antonio Bruno; Carmela Mento; Gianluca Pandolfo; Rocco A Zoccali Journal: World J Gastroenterol Date: 2016-07-28 Impact factor: 5.742
Authors: Eng Sing Lee; Hui Li Koh; Elaine Qiao-Ying Ho; Sok Huang Teo; Fang Yan Wong; Bridget L Ryan; Martin Fortin; Moira Stewart Journal: BMJ Open Date: 2021-05-05 Impact factor: 2.692
Authors: Michael Hoerger; Maria Coletta; Silvia Sörensen; Benjamin P Chapman; Kim Kaukeinen; Xin Tu; Paul R Duberstein Journal: J Aging Res Date: 2016-04-07