Kenneth L Kehl1, Elizabeth B Lamont2, Barbara J McNeil3, Samuel R Bozeman4, Michael J Kelley5, Nancy L Keating6. 1. MD Anderson Cancer Center, Houston, TX, USA. 2. Department of Health Care Policy, Harvard Medical School, Boston, MA, USA; Massachusetts General Hospital Cancer Center, Boston, MA, USA. 3. Department of Health Care Policy, Harvard Medical School, Boston, MA, USA; Department of Radiology, Brigham and Women's Hospital, Boston, MA, USA. 4. Abt Associates, Cambridge, MA, USA. 5. Durham Veterans Affairs Medical Center, Durham, ND, USA. 6. Department of Health Care Policy, Harvard Medical School, Boston, MA, USA; Division of General Internal Medicine, Brigham and Women's Hospital, Boston, MA, USA. Electronic address: keating@hcp.med.harvard.edu.
Abstract
OBJECTIVE: Ascertaining comorbid conditions in cancer patients is important for research and clinical quality measurement, and is particularly important for understanding care and outcomes for older patients and those with multi-morbidity. We compared the medical records-based ACE-27 index and the claims-based Charlson index in predicting receipt of therapy and survival for lung and colon cancer patients. MATERIALS AND METHODS: We calculated the Charlson index using administrative data and the ACE-27 score using medical records for Veterans Affairs patients diagnosed with stage I/II non-small cell lung or stage III colon cancer from January 2003 to December 2004. We compared the proportion of patients identified by each index as having any comorbidity. We used multivariable logistic regression to ascertain the predictive power of each index regarding delivery of guideline-recommended therapies and two-year survival, comparing the c-statistic and the Akaike information criterion (AIC). RESULTS: Overall, 97.2% of lung and 90.9% of colon cancer patients had any comorbidity according to the ACE-27 index, versus 59.5% and 49.7%, respectively, according to the Charlson. Multivariable models including the ACE-27 index outperformed Charlson-based models when assessing receipt of guideline-recommended therapies, with higher c-statistics and lower AICs. Neither index was clearly superior in prediction of two-year survival. CONCLUSIONS: The ACE-27 index measured using medical records captured more comorbidity and outperformed the Charlson index measured using administrative data for predicting receipt of guideline-recommended therapies, demonstrating the potential value of more detailed comorbidity data. However, the two indices had relatively similar performance when predicting survival.
OBJECTIVE: Ascertaining comorbid conditions in cancerpatients is important for research and clinical quality measurement, and is particularly important for understanding care and outcomes for older patients and those with multi-morbidity. We compared the medical records-based ACE-27 index and the claims-based Charlson index in predicting receipt of therapy and survival for lung and colon cancerpatients. MATERIALS AND METHODS: We calculated the Charlson index using administrative data and the ACE-27 score using medical records for Veterans Affairs patients diagnosed with stage I/II non-small cell lung or stage III colon cancer from January 2003 to December 2004. We compared the proportion of patients identified by each index as having any comorbidity. We used multivariable logistic regression to ascertain the predictive power of each index regarding delivery of guideline-recommended therapies and two-year survival, comparing the c-statistic and the Akaike information criterion (AIC). RESULTS: Overall, 97.2% of lung and 90.9% of colon cancerpatients had any comorbidity according to the ACE-27 index, versus 59.5% and 49.7%, respectively, according to the Charlson. Multivariable models including the ACE-27 index outperformed Charlson-based models when assessing receipt of guideline-recommended therapies, with higher c-statistics and lower AICs. Neither index was clearly superior in prediction of two-year survival. CONCLUSIONS: The ACE-27 index measured using medical records captured more comorbidity and outperformed the Charlson index measured using administrative data for predicting receipt of guideline-recommended therapies, demonstrating the potential value of more detailed comorbidity data. However, the two indices had relatively similar performance when predicting survival.
Authors: Melisa L Wong; Timothy L McMurry; George J Stukenborg; Amanda B Francescatti; Carla Amato-Martz; Jessica R Schumacher; George J Chang; Caprice C Greenberg; David P Winchester; Daniel P McKellar; Louise C Walter; Benjamin D Kozower Journal: Lung Cancer Date: 2016-11-09 Impact factor: 5.705
Authors: Nadine J McCleary; Sui Zhang; Chao Ma; Fang-Shu Ou; Tiffany M Bainter; Alan P Venook; Donna Niedzwiecki; Heinz-Josef Lenz; Federico Innocenti; Bert H O'Neil; Blase N Polite; Howard S Hochster; James N Atkins; Richard M Goldberg; Kimmie Ng; Robert J Mayer; Charles D Blanke; Eileen M O'Reilly; Charles S Fuchs; Jeffrey A Meyerhardt Journal: J Geriatr Oncol Date: 2022-01-31 Impact factor: 3.929
Authors: Philippe Jean Bousquet; Delphine Lefeuvre; Philippe Tuppin; Marc Karim BenDiane; Mathieu Rocchi; Elsa Bouée-Benhamiche; Jérôme Viguier; Christine Le Bihan-Benjamin Journal: PLoS One Date: 2018-10-31 Impact factor: 3.240