Validation of a 5-Year Mortality Prediction Model among U.S. Medicare Beneficiaries.
Ontology highlight
ABSTRACT: Background/objectives
A claims-based model predicting 5-year mortality (Lund-Lewis) was developed in a 2008 cohort of North Carolina (NC) Medicare beneficiaries and included indicators of comorbid conditions, frailty, disability, and functional impairment. The objective of this study was to validate the Lund-Lewis model externally within a nationwide sample of Medicare beneficiaries.Design
Retrospective validation study.Setting
U.S. Medicare population.Participants
From a random sample of Medicare beneficiaries, we created four annual cohorts from 2008 to 2011 of individuals aged 66 and older with an office visit in that year. The annual cohorts ranged from 1.13 to 1.18 million beneficiaries.Measurements
The outcome was 5-year all-cause mortality. We assessed clinical indicators in the 12 months before the qualifying office visit and estimated predicted 5-year mortality for each beneficiary in the nationwide sample by applying estimates derived in the original NC cohort. Model performance was assessed by quantifying discrimination, calibration, and reclassification metrics compared with a model fit on a comorbidity score.Results
Across the annual cohorts, 5-year mortality ranged from 24.4% to 25.5%. The model had strong discrimination (C-statistics ranged across cohorts from .823 to .826). Reclassification measures showed improvement over a comorbidity score model for beneficiaries who died but reduced performance among beneficiaries who survived. The calibration slope ranged from .83 to .86; the model generally predicted a higher risk than observed.Conclusion
The Lund-Lewis model showed strong and consistent discrimination in a national U.S. Medicare sample, although calibration indicated slight overfitting. Future work should investigate methods for improving model calibration and evaluating performance within specific disease settings.
SUBMITTER: Ross RK
PROVIDER: S-EPMC7781164 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:PurposeDe-implementation of low-value services among patients with limited life expectancy is challenging. Robust mortality prediction models using routinely collected health care data can enhance health care stakeholders' ability to identify populations with limited life expectancy. We developed and validated a claims-based prediction model for 5-year mortality using regularized regression methods.MethodsMedicare beneficiaries age 66 or older with an office visit and at least 12 months of pre-visit continuous Medicare A/B enrollment were identified in 2008. Five-year mortality was assessed through 2013. Secondary outcomes included 30-, 90-, and 180-day and 1-year mortality. Claims-based predictors, including comorbidities and indicators of disability, frailty, and functional impairment, were selected using regularized logistic regression, applying the least absolute shrinkage and selection operator (LASSO) in a random 80% training sample. Model performance was assessed and compared with the Gagne comorbidity score in the 20% validation sample.ResultsOverall, 183 204 (24%) individuals died. In addition to demographics, 161 indicators of comorbidity and function were included in the final model. In the validation sample, the c-statistic was 0.825 (0.823-0.828). Median-predicted probability of 5-year mortality was 14%; almost 4% of the cohort had a predicted probability greater than 80%. Compared with the Gagne score, the LASSO model led to improved 5-year mortality classification (net reclassification index = 9.9%; integrated discrimination index = 5.2%).ConclusionsOur claims-based model predicting 5-year mortality showed excellent discrimination and calibration, similar to the Gagne score model, but resulted in improved mortality classification. Regularized regression is a feasible approach for developing prediction tools that could enhance health care research and evaluation of care quality.
| S-EPMC6519458 | biostudies-literature
Project description:BackgroundParkinson disease is a neurodegenerative disease that affects gait and postural stability, resulting in an increased risk of falling. The purpose of this study was to estimate mortality associated with demographic factors after hip or pelvic (hip/pelvic) fracture in people with Parkinson disease. A secondary goal was to compare the mortality associated with Parkinson disease to that associated with other common medical conditions in patients with hip/pelvic fracture.MethodsThis was a retrospective observational cohort study of 1,980,401 elderly Medicare beneficiaries diagnosed with hip/pelvic fracture from 2000 to 2005 who were identified with use of the Beneficiary Annual Summary File. The race/ethnicity distribution of the sample was white (93.2%), black (3.8%), Hispanic (1.2%), and Asian (0.6%). Individuals with Parkinson disease (131,215) were identified with use of outpatient and carrier claims. Cox proportional hazards models were used to estimate the risk of death associated with demographic and clinical variables and to compare mortality after hip/pelvic fracture between patients with Parkinson disease and those with other medical conditions associated with high mortality after hip/pelvic fracture, after adjustment for race/ethnicity, sex, age, and modified Charlson comorbidity score.ResultsAmong those with Parkinson disease, women had lower mortality after hip/pelvic fracture than men (adjusted hazard ratio [HR] = 0.63, 95% confidence interval [CI]) = 0.62 to 0.64), after adjustment for covariates. Compared with whites, blacks had a higher (HR = 1.12, 95% CI = 1.09 to 1.16) and Hispanics had a lower (HR = 0.87, 95% CI = 0.81 to 0.95) mortality, after adjustment for covariates. Overall, the adjusted mortality rate after hip/pelvic fracture in individuals with Parkinson disease (HR = 2.41, 95% CI = 2.37 to 2.46) was substantially elevated compared with those without the disease, a finding similar to the increased mortality associated with a diagnosis of dementia (HR = 2.73, 95% CI = 2.68 to 2.79), kidney disease (HR = 2.66, 95% CI = 2.60 to 2.72), and chronic obstructive pulmonary disease (HR = 2.48, 95% CI = 2.43 to 2.53).ConclusionsMortality after hip/pelvic fracture in Parkinson disease varies according to demographic factors. Mortality after hip/pelvic fracture is substantially increased among those with Parkinson disease.
| S-EPMC3918936 | biostudies-literature
Project description:ObjectiveTo determine the prevalence and incidence of epilepsy among U.S. Medicare beneficiaries aged 65 years old and over, and to compare rates across demographic groups.MethodsWe performed a retrospective analysis of Medicare administrative claims for 2001-2005, defining prevalent cases as persons with ≥1 claim with diagnosis code 345.xx (epilepsy) or 2 or more with diagnosis code 780.3x (convulsion) ≥1 month apart, and incident cases as prevalent cases with 2 years immediately before diagnosis without such claims. Prevalence and incidence rates were calculated for the years 2003-2005 using denominators estimated from a 5% random sample of Medicare beneficiaries. Results were correlated with gender, age, and race.ResultsWe identified 282,661 per year on average during 2001-2005 (a total of 704,243 unique cases overall), and 62,182 incident cases per year on average during 2003-2005. Average annual prevalence and incidence rates were 10.8/1,000 and 2.4/1,000. Overall, rates were higher for black beneficiaries (prevalence 18.7/1,000, incidence 4.1/1,000), and lower for Asians (5.5/1,000, 1.6/1,000) and Native Americans (7.7/1,000, 1.1/1,000) than for white beneficiaries (10.2/1,000, 2.3/1,000). Incidence rates were slightly higher for women than for men, and increased with age for all gender and race groups.ConclusionsEpilepsy is a significant public health problem among Medicare beneficiaries. Efforts are necessary to target groups at higher risk, such as minorities or the very old, and to provide the care necessary to reduce the negative effects of epilepsy on quality of life.
| S-EPMC3280050 | biostudies-literature
Project description:RationalePulmonary nontuberculous mycobacteria (PNTM) are an important cause of morbidity among older adults in the United States, but national prevalence estimates are lacking.ObjectivesTo describe the prevalence and trends of PNTM disease among adults aged 65 years or older throughout the United States.MethodsA nationally representative 5% sample of Medicare Part B beneficiaries was analyzed from 1997 to 2007. Demographic and medical claims data were compiled and prevalence estimates for PNTM and selected comorbidities were calculated and trends over time evaluated. Logistic regression was used to identify demographic and geographic factors associated with PNTM.Measurements and main resultsFrom 1997 to 2007, the annual prevalence significantly increased from 20 to 47 cases/100,000 persons, or 8.2% per year. The period prevalence was 112 cases/100,000 persons, although prevalence was twofold higher among Asians/Pacific Islanders than among whites (228 vs. 116 cases/100,000 persons). Western states had the highest period prevalence at 149 cases/100,000 persons, with Hawaii having the highest prevalence at 396 cases/100,000 persons, followed by southeastern states, which had a period prevalence of 131 cases/100,000 persons. PNTM cases had more comorbid conditions than noncases and were 40% more likely to die than noncases. Women were 1.4 times more likely to be a PNTM case than men. Relative to whites, Asians/Pacific Islanders were twice as likely to be a case, whereas blacks were half as likely.ConclusionsThe prevalence of PNTM is increasing across all regions of the United States and among both men and women. Significant racial/ethnic and geographic differences suggest important gene-environment interactions.
| S-EPMC3360574 | biostudies-literature
Project description:Hospital readmission within thirty days is common among Medicare beneficiaries, but the relationship between rehospitalization and subsequent mortality in older adults is not known.To compare one-year mortality rates among community-dwelling elderly hospitalized Medicare beneficiaries who did and did not experience early hospital readmission (within 30 days), and to estimate the odds of one-year mortality associated with early hospital readmission and with other patient characteristics.A cohort study of 2133 hospitalized community-dwelling Medicare beneficiaries older than 64 years, who participated in the nationally representative Cost and Use Medicare Current Beneficiary Survey between 2001 and 2004, with follow-up through 2006.One-year mortality after index hospitalization discharge.Three hundred and four (13.7 %) hospitalized beneficiaries had an early hospital readmission. Those with early readmission had higher one-year mortality (38.7 %) than patients who were not readmitted (12.1 %; p<0.001). Early readmission remained independently associated with mortality after adjustment for sociodemographic factors, health and functional status, medical comorbidity, and index hospitalization-related characteristics [HR (95 % CI) 2.97 (2.24-3.92)]. Other patient characteristics independently associated with mortality included age [1.03 (1.02-1.05) per year], low income [1.39 (1.04-1.86)], limited self-rated health [1.60 (1.20-2.14)], two or more recent hospitalizations [1.47 (1.01-2.15)], mobility difficulty [1.51 (1.03-2.20)], being underweight [1.62 (1.14-2.31)], and several comorbid conditions, including chronic lung disease, cancer, renal failure, and weight loss. Hospitalization-related factors independently associated with mortality included longer length of stay, discharge to a skilled nursing facility for post-acute care, and primary diagnoses of infections, cancer, acute myocardial infarction, and heart failure.Among community-dwelling older adults, early hospital readmission is a marker for notably increased risk of one-year mortality. Providers, patients, and families all might respond profitably to an early readmission by reviewing treatment plans and goals of care.
| S-EPMC3475824 | biostudies-other
Project description:OBJECTIVES:To assess the variability in short-term sepsis mortality by hospital among Centers for Medicare and Medicaid Services beneficiaries in the United States during 2013-2014. DESIGN:A retrospective cohort design. SETTING:Hospitalizations from 3,068 acute care hospitals that participated in the Centers for Medicare and Medicaid Services inpatient prospective payment system in 2013 and 2014. PATIENTS:Medicare fee-for-service beneficiaries greater than or equal to 65 years old who had an inpatient hospitalization coded with present at admission severe sepsis or septic shock. INTERVENTIONS:None. MEASUREMENTS AND MAIN RESULTS:Individual level mortality was assessed as death at or within 7 days of hospital discharge and aggregated to calculate hospital-level mortality rates. We used a logistic hierarchal linear model to calculate mortality risk-adjusted for patient characteristics. We quantified variability among hospitals using the median odds ratio and calculated risk-standardized mortality rates for each hospital. The overall crude mortality rate was 34.7%. We found significant variability in mortality by hospital (p < 0.001). The middle 50% of hospitals had similar risk-standardized mortality rates (32.7-36.9%), whereas the decile of hospitals with the highest risk-standardized mortality rates had a median mortality rate of 40.7%, compared with a median of 29.2% for hospitals in the decile with the lowest risk-standardized mortality rates. The median odds ratio (1.29) was lower than the adjusted odds ratios for several measures of patient comorbidities and severity of illness, including present at admission organ dysfunction, no identified source of infection, and age. CONCLUSIONS:In a large study of present at admission sepsis among Medicare beneficiaries, we showed that mortality was most strongly associated with underlying comorbidities and measures of illness on arrival. However, after adjusting for patient characteristics, mortality also modestly depended on where a patient with sepsis received care, suggesting that efforts to improve sepsis outcomes in lower performing hospitals could impact sepsis survival.
| S-EPMC6774245 | biostudies-literature
Project description:OBJECTIVE:We sought to develop a risk scoring system for predicting functional deterioration, institutionalization, and mortality. Identifying predictors of poor health outcomes informs clinical decision-making, service provision, and policy development to address the needs of persons at greatest risk for poor health outcomes. DESIGN:This is a cohort study with 21,257 community-dwelling Medicare beneficiaries 65 yrs and older who participated in the 2001-2008 Medicare Current Beneficiary Survey.?Derivation of the model was conducted in 60% of the sample and validated in the remaining 40%. Multinomial logistic regression model generated ? coefficients, which were used to create a risk scoring system. Our outcome was instrumental activity of daily living stage transitions (stable/improved function and functional deterioration), institutionalization, or mortality for 2 yrs of follow-up. RESULTS:A total of 18 factors were identified for functional deterioration (P < 0.05). In the derivation cohort, the likelihood of functional deterioration ranged from 6.27% to 33.51%, risk of institutionalization from 0.07% to 12.13%, and risk of mortality from 2.13% to 31.83%, in comparison with stable/improved function. CONCLUSIONS:A risk scoring system predicting Medicare beneficiaries' risk of functional deterioration, institutionalization, and mortality based on demographic and clinical indicators may feasibly be developed with implications for healthcare delivery.
| S-EPMC6148394 | biostudies-literature
Project description:To estimate the effect of pre-intensive care unit (ICU) health categories on mortality during and after critical illness, focusing specifically on the effect of pre-ICU frailty on short- and long-term mortality.Retrospective cohort study.Medicare claims data from 2004 to 2008.A nationally representative sample of elderly Medicare beneficiaries admitted to an ICU in 2005.Participants were classified into four pre-ICU health categories (robust, cancer, chronic organ failure, frailty) using claims data from the year before admission, allowing for assignment to multiple categories. The association between pre-ICU health category and hospital and 3-year mortality was assessed using multivariable logistic regression and Cox proportional hazards models.Of 47,427 elderly individuals in the ICU, 18.8% were robust, 28.6% had cancer, 68.1% had chronic organ failure, and 34.0% were frail; 41.3% qualified for multiple categories. Overall hospital mortality was 12.6%, with the lowest mortality for robust participants (9.7%). Participants with pre-ICU frailty had higher hospital mortality than those with the same pre-ICU health categories without frailty. (Adjusted odds ratios ranged from 1.27 (95% confidence interval (CI) 1.10-1.47) to 1.52 (95% CI = 1.35-1.63).) Robust hospital survivors had the lowest 3-year mortality (24.6%). Pre-ICU frailty conferred higher 3-year mortality than pre-ICU categories without frailty. (Adjusted hazard ratios ranged from 1.54 (95% CI = 1.45-1.64) to 1.84 (95% CI = 1.70-1.99).)Critically ill elderly adults can be categorized according to pre-ICU health categories. These categories, particularly pre-ICU frailty, may be important for understanding risk of death during and after critical illness.
| S-EPMC4752202 | biostudies-other
Project description:ImportanceA growing proportion of the population is enrolling in Medicare Advantage (MA), which typically offers additional benefits compared with traditional Medicare (TM).ObjectiveTo determine whether frailty and frailty trajectories differ between MA enrollees and TM enrollees.Design, setting, and participantsThis retrospective cohort study used data from the National Health and Aging Trends Study (2015-2016). Analyses were conducted from August 2023 to March 2024. Participants were community-dwelling Medicare beneficiaries aged 65 years and older.ExposureEnrollment in MA vs TM.Main outcomes and measuresFrailty was calculated by a frailty index (FI) (range, 0-1, with higher values indicating greater frailty) and the Fried Frailty Phenotype (FFP) score (range, 0-5, with higher values indicating greater frailty). Physical performance, including Short Physical Performance Battery (SPPB) score (range, 0-12, with higher values indicating better performance), and gait speed (meters per second) were measured. The primary outcome was the difference in FI and FFP scores from the 2015 baseline assessment to the 2016 follow-up assessment. Secondary outcomes include the 1-year changes in SPPB and gait speed.ResultsThe final cohort consisted of 7063 participants (2775 [23.1%] aged >80 years; 4040 [54.7%] female), representing a sample of the 38.8 million beneficiaries. There were 2583 (35.0%) MA enrollees (13.6 million) and 4480 (65.0%) TM enrollees (25.2 million). At baseline, the FI score was similar between MA and TM enrollees (mean [SD], 0.22 [0.15] vs 0.21 [0.14]), although MA enrollees had worse phenotypic frailty (496 participants [15.2%] vs 811 participants [13.7%] considered frail by FFP score), SPPB scores (mean [SD], 6.91 [3.34] vs 7.21 [3.27]), and gait speed (0.79 [0.24] m/s vs 0.82 [0.23] m/s) than TM enrollees. One year later, there were no differences between MA and TM enrollees in the 1-year change in FI score (mean [SD], 0.016 [0.071] vs 0.014 [0.066]; adjusted mean difference, 0.001 [95% CI, -0.004 to 0.005]), FFP score (mean [SD], 0.017 [1.004] vs 0.007 [0.958]; adjusted mean difference, -0.009 [95% CI, -0.067 to 0.049]), SPPB score (mean [SD], -0.144 [2.064] vs -0.211 [1.968]; adjusted mean difference, 0.068 [95% CI, -0.076 to 0.212]), and gait speed (mean [SD], -0.0160 [0.148] m/s vs -0.007 [0.148] m/s; adjusted mean difference, -0.010 m/s [95% CI, -0.067 to 0.049 m/s]).Conclusions and relevanceIn this cohort study of Medicare beneficiaries from 2015, MA enrollees experienced similar declines in frailty over 1 year compared with TM enrollees. Future work should examine whether the specific types of services covered by health insurance can impact frailty and health trajectories for older adults.
| S-EPMC11365002 | biostudies-literature
Project description:5.8 M living Americans have experienced a stroke at some time in their lives, 780K had either their first or a recurrent stroke this year, and 150K died from strokes this year. Stroke costs about $66B annually in the US, and also results in serious, long-term disability. Therefore, it is prudent to identify all possible risk factors and their effects so that appropriate intervention points may be targeted.Baseline (1993-1994) interview data from the nationally representative Survey on Assets and Health Dynamics among the Oldest Old (AHEAD) were linked to 1993-2005 Medicare claims. Participants were 5,511 self-respondents >or= 70 years old. Two ICD9-CM case-identification approaches were used. Two approaches to stroke case-identification based on ICD9-CM codes were used, one emphasized sensitivity and the other emphasized specificity. Participants were censored at death or enrollment into managed Medicare. Baseline risk factors included sociodemographic, socioeconomic, place of residence, health behavior, disease history, and functional and cognitive status measures. A time-dependent marker reflecting post-baseline non-stroke hospitalizations was included to reflect health shocks, and sensitivity analyses were conducted to identify its peak effect. Competing risk, proportional hazards regression was used.Post-baseline strokes occurred for 545 (9.9%; high sensitivity approach) and 374 (6.8%; high specificity approach) participants. The greatest static risks involved increased age, being widowed or never married, living in multi-story buildings, reporting a baseline history of diabetes, hypertension, or stroke, and reporting difficulty picking up a dime, refusing to answer the delayed word recall test, or having poor cognition. Risks were similar for both case-identification approaches and for recurrent and first-ever vs. only first-ever strokes. The time-dependent health shock (recent hospitalization) marker did not alter the static model effect estimates, but increased stroke risk by 200% or more.The effect of our health shock marker (a time-dependent recent hospitalization indicator) was large and did not mediate the effects of the traditional risk factors. This suggests an especially vulnerable post-hospital transition period from adverse effects associated with both their underlying health shock (the reasons for the recent hospital admission) and the consequences of their treatments.
| S-EPMC2683849 | biostudies-other