Predicting later life health status and mortality using state-level socioeconomic characteristics in early life

Studies extending across multiple life stages promote an understanding of factors influencing health across the life span. Existing work has largely focused on individual-level rather than area-level early life determinants of health. In this study, we linked multiple data sets to examine whether early life state-level characteristics were predictive of health and mortality decades later. The sample included 143,755 U.S. employees, for whom work life claims and administrative data were linked with early life state-of-residence and mortality. We first created a "state health risk score" (SHRS) and "state mortality risk score" (SMRS) by modeling state-level contextual characteristics with health status and mortality in a randomly selected 30% of the sample (the "training set"). We then examined the association of these scores with objective health status and mortality in later life in the remaining 70% of the sample (the "test set") using multivariate linear and Cox regressions, respectively. The association between the SHRS and adult health status was β=0.14 (95%CI: 0.084, 0.20), while the hazard ratio for the SMRS was 0.96 (95%CI: 0.93, 1.00). The association between the SHRS and health was not statistically significant in older age groups at a p-level of 0.05, and there was a statistically significantly different association for health status among movers compared to stayers. This study uses a life course perspective and supports the idea of "sensitive periods" in early life that have enduring impacts on health. It adds to the literature examining populations in the U.S. where large linked data sets are infrequently available.

Integrating Well-Being Information and the Multidimensional Adaptive Prediction Process to Estimate Individual-Level Future Health Care Expenditure Levels

Decades of research exist focusing on the utility of self-reported health risk and status data in health care cost predictive models. However, in many of these studies a limited number of self-reported measures were considered. Compounding this issue, prior research evaluated models specified with a single covariate vector and distribution. In this study, the authors incorporate well-being data into the Multidimensional Adaptive Prediction Process (MAPP) and then use a simulation analysis to highlight the value of these findings for future cost mitigation. Data were collected on employees and dependents of a nationally based employer over 36 months beginning in January 2010. The first 2 years of data (2010, 2011) were utilized in model development and selection; 51239 and 54085 members were included in 2010 and 2011, respectively. The final results were based on prospective prediction of 2012 cost levels using 2011 data. The well-being-augmented MAPP results showed a 5.7% and 13% improvement in accurate cost capture relative to a reference modeling approach and the first study of MAPP, respectively. The simulation analysis results demonstrated that reduced well-being risk across a population can help mitigate the expected upward cost trend. This research advances health care cost predictive modeling by incorporating well-being information within MAPP and then leveraging the results in a simulation analysis of well-being improvement.

Using the SF-12 health status measure to improve predictions of medical expenditures

BACKGROUND

Relatively few studies have used self-reported health status in models to predict medical expenditures, and many of these have used the SF-36.

OBJECTIVES

We sought to examine the ability of the briefer SF-12 measure of health status to predict medical expenditures in a nationally representative sample.

METHODS

We used data from the 2000-2001 panel of the Medical Expenditure Panel Study. Respondents (n = 5542) completed the SF-12 in a questionnaire. Interviews obtained data on demographics and selected chronic conditions. Data on expenditures incurred subsequent to the interview were obtained in part from provider records. We examined different regression model specifications and compared different statistical estimation techniques.

RESULTS

Adding the SF-12 to a regression model improved the prediction of subsequent medical expenditures. In a model with only age and gender, adding the SF-12 increased R from 0.06 to 0.13. The coefficients for the Physical Component Summary (PCS) and the Mental Component Summary (MCS) of the SF-12 for this model were -0.045 (P < 0.01) and -0.012 (P < 0.01), respectively. In a model including demographic characteristics, chronic conditions, and previous expenditures, adding the SF-12 increased the R from 0.26 to 0.29. The coefficients for the PCS and the MCS for this model were -0.025 (P < 0.001) and -0.005 (P = 0.15), respectively. A single general health status question performed almost as well as the full SF-12. Models estimated using ordinary least squares had undesirable properties. In terms of R, a generalized linear model (GLM) with a Poisson variance function was consistently superior to a GLM with a gamma variance function.

CONCLUSIONS

Information on self-reported health status is useful in predicting medical expenditures. The extent to which the SF-12 adds predictive power over a comprehensive array of diagnostic data remains to be examined.

The performance of administrative and self-reported measures for risk adjustment of Veterans Affairs expenditures

OBJECTIVE

To evaluate the performance of different prospective risk adjustment models of outpatient, inpatient, and total expenditures of veterans who regularly use Veterans Affairs (VA) primary care.

DATA SOURCES

We utilized administrative, survey and expenditure data on 14,449 VA patients enrolled in a randomized trial that gave providers regular patient health assessments.

STUDY DESIGN

This cohort study compared five administrative data-based, two self-report risk adjusters, and base year expenditures in prospective models.

DATA EXTRACTION METHODS

VA outpatient care and nonacute inpatient care expenditures were based on unit expenditures and utilization, while VA expenditures for acute inpatient care were calculated from a Medicare-based inpatient cost function. Risk adjusters for this sample were constructed from diagnosis, medication and self-report data collected during a clinical trial. Model performance was compared using adjusted R2 and predictive ratios.

PRINCIPAL FINDINGS

In all expenditure models, administrative-based measures performed better than self-reported measures, which performed better than age and gender. The Diagnosis Cost Groups (DCG) model explained total expenditure variation (R2=7.2 percent) better than other models. Prior outpatient expenditures predicted outpatient expenditures best by far (R2=42 percent). Models with multiple measures improved overall prediction, reduced over-prediction of low expenditure quintiles, and reduced under-prediction in the highest quintile of expenditures.

CONCLUSIONS

Prediction of VA total expenditures was poor because expenditure variation reflected utilization variation, but not patient severity. Base year expenditures were the best predictor of outpatient expenditures and nearly the best for total expenditures. Models that combined two or more risk adjusters predicted expenditures better than single-measure models, but are more difficult and expensive to apply.

Predicting Healthcare Costs in a Population of Veterans Affairs Beneficiaries Using Diagnosis-Based Risk Adjustment and Self-Reported Health Status

BACKGROUND

Many healthcare organizations use diagnosis-based risk adjustment systems for predicting costs. Health self-report may add information not contained in a diagnosis-based system but is subject to incomplete response.

OBJECTIVE

The objective of this study was to evaluate the added predictive power of health self-report in combination with a diagnosis-based risk adjustment system in concurrent and prospective models of healthcare cost.

RESEARCH DESIGN

This was a cohort study using Department of Veterans Affairs (VA) administrative databases. We tested the predictive ability of the Adjusted Clinical Group (ACG) methodology and the added value of SF-36V (short form functional status for veterans) results. Linear regression models were compared using R(2), mean absolute prediction error (MAPE), and predictive ratio.

SUBJECTS

Subjects were 35,337 VA beneficiaries at 8 VA medical centers during fiscal year (FY) 1998 who voluntarily completed an SF-36V survey.

MEASURES

Outcomes were total FY 1998 and FY 1999 cost. Demographics and ACG-based Adjusted Diagnostic Groups (ADGs) with and without 8 SF-36V multiitem scales and the Physical Component Score and Mental Component Score were compared.

RESULTS

The survey response rate was 45%. Adding the 8 scales to ADGs and demographics increased the crossvalidated R by 0.007 in the prospective model. The 8 scales reduced the MAPE by 236 US dollars among patients in the upper 10% of FY 1999 cost.

CONCLUSIONS

The limited added predictive power of health self-report to a diagnosis-based risk adjustment system should be weighed against the cost of collecting these data. Adding health self-report data may increase predictive accuracy in high-cost patients.