Other models currently exist to aid in the prediction of life expectancy. In this study, Dr Murphy and colleagues compared their claims-based model of noncancer survival against the Charlson comorbidity index (CCI) and the Elixhauser comorbidity index (ECI). The CCI uses a composite of 19 diseases to generate an individual score for patients; this score represents the risk of noncancer mortality. The ECI is similar to the CCI, but uses a composite of 30 comorbidity measures.
The noncancer survival model included 143 covariates and improved survival prediction compared with the CCI and the ECI.
The cancer-specific model was compared with the Memorial Sloan Kettering Cancer Center (MSKCC) Prostate Cancer nomogram, which estimates risk of cancer mortality using tumor characteristics such as clinical T stage, Gleason score, and PSA level. The claims-based, cancer-specific survival model included 9 covariates, but had survival prediction that was almost identical to the MSKCC prediction model.
“Given the added complexity of using claims in the prediction model, there would be little impetus to use this new approach in a clinical or research setting when evaluating cancer-specific mortality risks,” the researchers wrote in the study.
Commenting on the study, Parth Modi, MD, of the department of urology, The University of Michigan, Ann Arbor, explained that this and other types of claims-based research are attempts to use large amounts of data to answer various research questions that are of interest to the health care community. However, instead of relying on carefully selected patients like randomized controlled trials, claims-based research uses all-comers and better reflects what is happening in reality.
“This study is a very early step on the way to something that could be more relevant for clinical care,” Dr Modi told Cancer Therapy Advisor. “The main thing the researchers have done is create a method to use claims to estimate comorbidity that is better than some of the older methods used now.”
Dr Murphy agreed that this claims-based approach would require validation in another group of patients before widespread implementation were considered. However, in theory, there is a path forward to clinical implementation.
“The risk score can help provide a patient with an estimate of their risk of death – though in some respects this may not always provide clarity,” Dr Murphy said. “Relaying risk or probability to patients is complex, and may not always help with the decision-making process. With clinical implementation, we would need more research on how to clinically implement this information, and how to best communicate this information to patients.”
- Riviere P, Tokeshi C, Hou J, et al. Claims-based approach to predict cause-specific survival in men with prostate cancer. JCO Clin Cancer Inform. 2019;3:1-7.