Changes in health care expenditure associated with gaining or losing health insurance

Department of Family and Community Medicine, University of California, San Francisco, San Francisco, California 94110, USA.
Annals of internal medicine (Impact Factor: 16.1). 07/2007; 146(11):768-74. DOI: 10.7326/0003-4819-146-11-200706050-00005
Source: PubMed

ABSTRACT Cross-sectional data suggest that changes in health insurance status are associated with expenditures. No national longitudinal analysis has examined this relationship.
To evaluate the association between changes in health insurance status and expenditures.
Cohort analyses using the 2000 to 2003 Medical Expenditure Panel Surveys.
U.S. civilian noninstitutionalized population.
Three 2-year cohorts that included 20,848 adults age 21 to 64 years who were stratified by insurance type (private, public, military, or none): 17,130 participants were insured in both years, 342 participants were insured in year 1 and were uninsured in year 2, 385 participants were uninsured in year 1 and were insured in year 2, and 2991 participants were uninsured in both years. Persons who were insured for longer than 2 months but less than 10 months or who switched insurance type were excluded (n = 4039).
Annual health care expenditures (any or none; amount, contingent on any expenditure; and the difference between year 1 and year 2).
Adjusted expenditure probabilities were similar among all participant groups while insured and were higher than those for all participant groups while uninsured: 92.1% (95% CI, 91.4% to 92.7%) in year 1 and 91.8% (CI, 90.9% to 92.5%) in year 2 for persons insured in both years, 74.2% (CI, 71.7% to 76.5%) in year 1 and 74.8% (CI, 72.1% to 77.4%) in year 2 for persons uninsured in both years, and 90.7% (CI, 87.1% to 93.4%) for persons insured in year 1 and 74.6% (CI, 69.4% to 79.2%) for persons uninsured in year 2. The pattern was also consistent for the group that was uninsured in year 1 but insured in year 2. Adjusted annual expenditures among all participant groups with insurance were similar; expenditures among participant groups without insurance were similar but were lower than those among participants with insurance. Consistent differences in expenditures between year 1 and year 2 were observed for all groups.
Few participants changed insurance status.
Changing insurance status is associated with changes in expenditures to levels that are similar to those for persons who are continuously insured or uninsured.

  • Source
    • "Applications of imputation by chained equations have now appeared in quite diverse fields: addiction (Schnoll et al. 2006; MacLeod et al. 2008; Adamczyk and Palmer 2008; Caria et al. 2009; Morgenstern et al. 2009), arthritis and rheumatology (Wolfe et al. 2006; Rahman et al. 2008; van den Hout et al. 2009), atherosclerosis (Tiemeier et al. 2004; van Oijen et al. 2007; McClelland et al. 2008), cardiovascular system (Ambler et al. 2005; van Buuren et al. 2006a; Chase et al. 2008; Byrne et al. 2009; Klein et al. 2009), cancer (Clark et al. 2001, 2003; Clark and Altman 2003; Royston et al. 2004; Barosi et al. 2007; Fernandes et al. 2008; Sharma et al. 2008; McCaul et al. 2008; Huo et al. 2008; Gerestein et al. 2009), epidemiology (Cummings et al. 2006; Hindorff et al. 2008; Mueller et al. 2008; Ton et al. 2009), endocrinology (Rouxel et al. 2004; Prompers et al. 2008), infectious diseases (Cottrell et al. 2005; Walker et al. 2006; Cottrell et al. 2007; Kekitiinwa et al. 2008; Nash et al. 2008; Sabin et al. 2008; Thein et al. 2008; Garabed et al. 2008; Michel et al. 2009), genetics (Souverein et al. 2006), health economics (Briggs et al. 2003; Burton et al. 2007; Klein et al. 2008; Marshall et al. 2009), obesity and physical activity (Orsini et al. 2008a; Wiles et al. 2008; Orsini et al. 2008b; van Vlierberghe et al. 2009), pediatrics and child development (Hill et al. 2004; Mumtaz et al. 2007; Deave et al. 2008; Samant et al. 2008; Butler and Heron 2008; Ramchandani et al. 2008; van Wouwe et al. 2009), rehabilitation (van der Hulst et al. 2008), behavior (Veenstra et al. 2005; Melhem et al. 2007; Horwood et al. 2008; Rubin et al. 2008), quality of care (Sisk et al. 2006; Roudsari et al. 2007; Ward and Franks 2007; Grote et al. 2007; Roudsari et al. 2008; Grote et al. 2008; Sommer et al. 2009), human reproduction (Smith et al. 2004a,b; Hille et al. 2005; Alati et al. 2006; O'Callaghan et al. 2006; Hille et al. 2007; Hartog et al. 2008), management sciences (Jensen and Roy 2008), occupational health (Heymans et al. 2007; Brunner et al. 2007; Chamberlain et al. 2008), politics (Tanasoiu and Colonescu 2008), psychology (Sundell et al. 2008) and sociology (Finke and Adamczyk 2008). All authors use some form of chained equations to handle the missing data, but the details vary considerably. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Multivariate Imputation by Chained Equations (MICE) is the name of software for imputing incomplete multivariate data by Fully Conditional Speci cation (FCS). MICE V1.0 appeared in the year 2000 as an S-PLUS library, and in 2001 as an R package. MICE V1.0 introduced predictor selection, passive imputation and automatic pooling. This article presents MICE V2.0, which extends the functionality of MICE V1.0 in several ways. In MICE V2.0, the analysis of imputed data is made completely general, whereas the range of models under which pooling works is substantially extended. MICE V2.0 adds new functionality for imputing multilevel data, automatic predictor selection, data handling, post-processing imputed values, specialized pooling and model selection. Imputation of categorical data is improved in order to bypass problems caused by perfect prediction. Special attention to transformations, sum scores, indices and interactions using passive imputation, and to the proper setup of the predictor matrix. MICE V2.0 is freely available from CRAN as an R package mice. This article provides a hands-on, stepwise approach to using mice for solving incomplete data problems in real data.
    Journal of statistical software 12/2011; 45(3). · 3.80 Impact Factor
  • Source
    • "physiotherapy. Studies conducted in the United States have shown that people with health insurance incurred more healthcare expenditure and utilised more healthcare resources (Duru et al 2007, Hadley 2007, Nelson et al 2005, Ward and Franks 2007). In contrast to the situation in the United States, residents of Australia have access to a comprehensive range of healthcare services via Medicare, Australia's public healthcare system, which is free at the point of delivery. "
    [Show abstract] [Hide abstract]
    ABSTRACT: What are the costs and utilisation of healthcare resources, their determinants, and quality of life for people attending outpatient physiotherapy after ankle fracture? Longitudinal observational study. Ninety-four adults (2 dropouts) following cast removal after isolated ankle fracture attending outpatient physiotherapy at three hospitals in Sydney, Australia. Costs incurred (direct healthcare costs and out-of-pocket costs) and utilisation of healthcare system resources were measured at 4, 8, 12, 16, 20, and 24 weeks. Quality of life was measured shortly after cast removal and at 4, 12, and 24 weeks. Factors known to influence costs and utilisation in other conditions (private health insurance, income level, gender, and pain) were also measured. The total cost per person was AUD 735 (SD 876) over 24 weeks. Outpatient physiotherapy accounted for the highest costs in both direct healthcare (39%) and out-of-pocket (42%) costs. Less than 20% of participants sought private non-medical care in addition to receiving outpatient physiotherapy. None of the factors investigated had a significant influence on costs and utilisation. Quality of life score improved over the 24 weeks by a mean of 6.1 points out of 45 (95% CI 5.2 to 6.9), with most of the improvement occurring in the domain of independent living. Information on costs and utilisation of healthcare resources can be used to plan health services, eg, the number of physiotherapy sessions required after ankle fracture. Private health insurance, income level, gender, or pain did not influence the costs or the decision behind seeking care over and above publicly-provided physiotherapy.
    The Australian journal of physiotherapy 02/2008; 54(3):201-8. DOI:10.1016/S0004-9514(08)70027-8 · 3.48 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Obese workers incur greater health care costs than normal weight workers. Possibly viewed by employers as an increased financial risk, they may be at a disadvantage in procuring employment that provides health insurance. This study aims to evaluate the association between body mass index [BMI, weight in kilograms divided by the square of height in meters] of employees and their likelihood of holding jobs that include employment-based health insurance [EBHI]. We used the 2004 Household Components of the nationally representative Medical Expenditure Panel Survey. We utilized logistic regression models with provision of EBHI as the dependent variable in this descriptive analysis. The key independent variable was BMI, with adjustments for the domains of demographics, social-economic status, workplace/job characteristics, and health behavior/status. BMI was classified as normal weight (18.5-24.9), overweight (25.0-29.9), or obese (> or = 30.0). There were 11,833 eligible respondents in the analysis. Among employed adults, obese workers [adjusted probability (AP) = 0.62, (0.60, 0.65)] (P = 0.005) were more likely to be employed in jobs with EBHI than their normal weight counterparts [AP = 0.57, (0.55, 0.60)]. Overweight workers were also more likely to hold jobs with EBHI than normal weight workers, but the difference did not reach statistical significance [AP = 0.61 (0.58, 0.63)] (P = 0.052). There were no interaction effects between BMI and gender or age. In this nationally representative sample, we detected an association between workers' increasing BMI and their likelihood of being employed in positions that include EBHI. These findings suggest that obese workers are more likely to have EBHI than other workers.
    BMC Health Services Research 02/2008; 8(1):101. DOI:10.1186/1472-6963-8-101 · 1.66 Impact Factor
Show more