0
We're unable to sign you in at this time. Please try again in a few minutes.
Retry
We were able to sign you in, but your subscription(s) could not be found. Please try again in a few minutes.
Retry
There may be a problem with your account. Please contact the AMA Service Center to resolve this issue.
Contact the AMA Service Center:
Telephone: 1 (800) 262-2350 or 1 (312) 670-7827  *   Email: subscriptions@jamanetwork.com
Error Message ......
Original Contribution |

Insurance Coverage, Medical Care Use, and Short-term Health Changes Following an Unintentional Injury or the Onset of a Chronic Condition FREE

Jack Hadley, PhD
[+] Author Affiliations

Author Affiliation: The Urban Institute, Washington, DC.

More Author Information
JAMA. 2007;297(10):1073-1084. doi:10.1001/jama.297.10.1073.
Text Size: A A A
Published online

Context Given the large and increasing number of uninsured US individuals, identifying the health consequences of being uninsured has assumed increased importance.

Objective To compare medical care use and short-term health changes among US uninsured individuals and insured nonelderly individuals following a health shock caused by either an unintentional injury or the onset of a chronic condition.

Design, Setting, and Participants Multivariate logistic regression analysis of longitudinal data from Medical Expenditure Panel Surveys (1997-2004) limited to nonelderly individuals whose insurance status was established for 2 months prior to 1 or more unintentional injuries (20 783 cases among 15 866 individuals) and onset of 1 or more chronic conditions (10 485 cases among 7954 individuals).

Main Outcome Measures Self-reported medical care use and change in short-term general health status following the health shock.

Results After experiencing a health shock, uninsured individuals were less likely to obtain any medical care (unintentional injury [UI] group: 78.8% uninsured vs 88.7% insured [adjusted odds ratio {AOR}, 0.47; 95% confidence interval {CI}, 0.43-0.51]; new chronic condition [NCC] group: 81.7% uninsured vs 91.5% insured [AOR, 0.45; 95% CI, 0.40-0.50]) and more likely not to have received any recommended follow-up care (UI group: 19.3% uninsured vs 9.2% insured [AOR, 2.59; 95% CI, 2.15-3.11]; NCC group: 9.4% uninsured vs 4.4% insured [AOR, 1.65; 95% CI, 1.32-2.06]). Based on the AORs, uninsured individuals with UIs had fewer outpatient visits (6.1% uninsured vs 9.0% insured; AOR, 0.71 [95% CI, 0.63-0.80]), office-based visits (41.8% uninsured vs 57.3% insured; AOR, 0.59 [95% CI, 0.56-0.62]), and prescription medicines (35.5% uninsured vs 35.6% insured; AOR, 0.71 [95% CI, 0.67-0.75]). Uninsured individuals with an NCC had fewer office-based visits (58.9% uninsured vs 68.3% insured; AOR, 0.77 [95% CI, 0.72-0.82]) and prescription medicines (52.7% uninsured vs 61.7% insured; AOR, 0.66 [95% CI, 0.57-0.76]). Higher proportions of uninsured individuals reported a decrease in health status (classified as much worse) approximately 3.5 months after the health shock (UI group: 9.8% uninsured vs 6.7% insured; AOR, 0.86 [95% CI, 0.75-0.98]; NCC group: 12.3% uninsured vs 10.1% insured; AOR, 0.74 [95% CI, 0.68-0.80]). Uninsured individuals with UIs were more likely to report not being fully recovered and no longer receiving treatment. At approximately 7 months after the health shock, uninsured individuals with NCCs still reported worse health status.

Conclusions Among individuals who experienced a health shock caused by an unintentional injury or a new chronic condition, uninsured individuals reported receiving less medical care and poorer short-term changes in health than those with insurance.

Although prior research has found that uninsured individuals with chronic conditions receive less care than insured individuals,1,2 these studies have neither focused on those who experience the onset of a chronic condition nor analyzed the effect of not having insurance on health. Other studies have analyzed the effects of insurance status on individuals who experienced an injury but typically only in a single setting or for a narrowly defined type of injury.35

A substantial literature suggests that not having insurance has adverse effects on medical care use and health.6,7 However, much of this research is subject to methodological limitations because individuals were not randomly assigned to not having insurance. Analyses of observational data may be biased if unobserved differences between uninsured and insured individuals, rather than the difference in insurance coverage, explain both insurance status and the differences in medical care use and health outcomes.8

The RAND Health Insurance Experiment,9 conducted more than 30 years ago, randomly assigned individuals to different health insurance plans. However, its highest cost-sharing plan was essentially a catastrophic coverage plan and cannot be used to make inferences about those without insurance coverage.

Other studies have found that obtaining Medicare coverage at age 65 years closed the gap in medical care use between previously uninsured and insured individuals,10 and that elderly individuals' mortality rates declined faster than expected following the enactment of Medicare in 1965.11 Several studies used longitudinal data to compare populations grouped by their baseline insurance status and also concluded that not having insurance adversely affects health.1217 However, these studies may still be subject to observational data bias because prior health may have affected both the general measures of the subsequent health change and the baseline health insurance status.

This study investigates whether uninsured individuals who experienced a health shock caused by either an unintentional injury or the onset of a chronic condition received the same amount of medical care and had similar short-term health outcomes as insured individuals.

The data are from the household component of the Medical Expenditure Panel Surveys conducted between 1997 and 2004.18 The Medical Expenditure Panel Survey is a continuous, nationally representative, in-person survey of the civilian, noninstitutionalized population. It uses a rotating panel design that replenishes half of the total sample each year and collects information during 5 interview rounds spaced 4 to 5 months apart over 2 years. Individuals' data can be linked across rounds to build a longitudinal round-by-round data file spanning the 2 years of survey participation. The combined response rates for the 1997-2004 Medical Expenditure Panel Survey range from 64.5% to 67.4% for the full-year files.19 The Urban Institute's institutional review board process determined that further review was not required because all of the data are publicly available and do not include any personal identifying information.

The survey collects data on individuals' sociodemographic characteristics, monthly insurance coverage, general health status, and medical care use. A sample of medical care providers identified by respondents are contacted to verify use and obtain additional information about diagnosis, charges, payments, and specific services provided. Insurance coverage is verified by checking insurance cards.20

The initial interview inventories each person's current health conditions. Subsequent interviews ask whether the person experienced any new health conditions since the previous interview. For each new condition identified, interviewers ascertain whether the condition is included on the Department of Health and Human Services' list of “priority” conditions,21 and whether the condition was due to an injury. Chronic conditions were identified by applying the classification algorithm developed by the Agency for Healthcare Research and Quality22 to the survey's International Classification of Diseases, Ninth Revision, condition codes.23

New chronic conditions identified by this process may be either reoccurrences of a prior chronic condition, such as back pain or depression, that had been resolved before the baseline interview and was not included in the person's initial condition inventory, or first occurrences of a condition the person had never experienced before. Both types of situations are referred to as new chronic conditions even though some of the cases may be reoccurrences of previously resolved conditions.

Priority-list conditions and unintentional injuries trigger the additional questions of (1) when the unintentional injury occurred and its type and location; (2) when the new condition was first noticed or diagnosed; (3) whether the person sought medical care; (4) the types and amounts of care received; and (5) whether the person is still receiving care at the time of the interview.

Analysis Samples

The research design assumes that prior insurance status in the 2 months before the unintentional injury or new chronic condition can be treated as exogenous (ie, predetermined and not caused by subsequent medical care use or changes in health associated with the injury or chronic condition). Therefore, the analysis samples are limited to individuals with uniform insurance coverage (uninsured or insured) for the 2 months before, the month of, and the month after the health shock. Those with multiple unintentional injuries or new chronic conditions are included as separate observations for each occurrence. (An additional analysis limiting the sample to 1 occurrence per person also was conducted.) Individuals aged 65 years or older are excluded because Medicare coverage is virtually universal.

Outcome Measures

There are 8 measures of medical care use: 2 dichotomous indicators of whether the person received any medical care; if he/she did receive care, whether any further care was recommended; if further care was recommended, a multichotomous indicator of whether the person is still receiving treatment, or received all, some, or none of the recommended follow-up care; and 5 measures of the quantities of care received for inpatient hospital stays, outpatient visits, office-based visits, emergency department visits, and number of prescription medications.

The distributions of the quantity variables are skewed and clustered at a limited number of discrete values, especially 0, 1, and 2. To accommodate this type of distribution in the statistical analysis, these measures were coded as ordered categorical variables. Inpatient hospital stays has 4 categories for 0, 1, 2, or 3 or more stays. The other use variables have 6 categories representing 0, 1, 2, 3 to 5, 6 to 10, or 11 or more visits (or prescriptions).

Self-reported general health status has been shown to be a good predictor of both medical care use and mortality both over time and across different populations.24,25 Diehr et al26 found that large short-term declines in older adults' self-reported health measured at 6-month intervals were good predictors of mortality and other major adverse health events. In this study, the short-term change in health was constructed from individuals' self-reported general health status, defined by a Likert scale with 5 categories (excellent, very good, good, fair, and poor health), in the interview round before the health shock, and the first and second (if available) interview rounds after the health shock. The change in health was measured as the arithmetic difference between the preevent (unintentional injury or new chronic condition) and postevent scales. The measure has 9 possible categories (values) ranging from plus 4 (from poor to excellent [the largest possible health improvement]) to negative 4 (from excellent to poor [the largest possible health decline]). A value of 0 represents no change in health.

Among individuals in the unintentional injury sample, 26 died and 7 were institutionalized (ie, were resident in a hospital or nursing home for ≥30 days at the time of the interview). Among those in the new chronic condition sample, 87 died and 3 were institutionalized subsequent to the first follow-up interview. These individuals were assigned to the category for the maximum decrease in health. Higher proportions of the deceased or institutionalized cases did not have insurance in the unintentional injury group (21.1% vs 16.2%) and in the new chronic condition group (25% vs 13%). The largest difference between the deceased or institutionalized cases and other cases was in their baseline health status. Among deceased or institutionalized cases, 25.2% in the unintentional injury group and 21.5% in the new chronic condition group reported poor baseline health compared with 2.0% and 4.9% of other cases. (Excluding these cases did not change any of the substantive results reported below.)

An additional health outcome for the unintentional injury sample was constructed from questions asking whether the person had fully recovered and if not fully recovered whether they were still being treated. The resulting variable was a trichotomous indicator of whether the person was fully recovered, still receiving treatment, or not recovered and not receiving care.

Control Variables

To reduce potential confounding, the statistical models control for the specific health condition that led to inclusion in the sample; the type and location of the unintentional injury; baseline health and sociodemographic characteristics (from the interview before the health shock); the person's census region and metropolitan residence status; and a set of survey indicators for the month and year of the event, the survey round of the follow-up interview, the time interval between the event and the follow-up interview, and whether the follow-up information is self-reported or obtained from a proxy.

Statistical Analysis

The null hypotheses of no differences between uninsured and insured individuals were tested by estimating multivariate logistic regression models: binomial models for the dichotomous measures; multinomial models for the categorical measures of the amount of follow-up care received and injury outcome; and ordered for the measures of medical care use and change in health status. All cases were weighted using the MEPS person-weights and all models were estimated using the survey logistic regression programs27 by StataCorp (College Station, Tex). All models include the control variables summarized in the previous section.

Although insurance status was observed 2 months before the health shock, it is possible that persistent unobserved differences between insured and uninsured individuals may influence both the likelihood of an unintentional injury and the distribution and severity of new health conditions. For example, insured and uninsured individuals may differ in terms of risky behaviors (ie, individuals without insurance may be more likely to drive without a seat belt or overuse alcohol whereas individuals with insurance may be more likely to ski or develop tennis elbow). Prior insurance status also can affect the likelihood of reporting a condition that typically requires a medical test to be detected, such as hypertension or early stage cancer. It also is possible that an uninsured person who experiences disease symptoms will conceal this information to obtain insurance coverage before seeking medical care.

While there is no formal statistical test of the assumption that an independent variable (insurance status) is exogenous,28 sensitivity tests can indicate whether the estimated effects of insurance are robust with respect to alternative variable specifications and alternative sample specifications that may reflect possible unobserved associations between insurance status and the distributions and the severity of the unintentional injuries and reported chronic conditions.

Therefore, the following sensitivity analyses also were conducted: (1) estimating the models without the control variables for the specific health condition to test whether they are correlated with insurance coverage and their omission results in altered estimates; (2) adding a self-reported measure of the condition's severity (perceived influence on health); (3) excluding cases if the person reported that the health shock had no effect on their health; (4) eliminating cases with specific conditions that appear to be associated with insurance status, based on supplementary regression models of the relationship between insurance status and the indicator variables for specific conditions; (5) limiting the sample to individuals with family incomes that are less than 400% of the federal poverty level because family income is strongly associated with insurance status and also may reflect unobserved personal differences between insured and uninsured individuals; (6) estimating models limited to 1 occurrence per person to test for possible bias arising from multiple observations of some individuals; and (7) extending the look-back period for determining insurance status to 12 months before the health shock.

In additional analysis designed to estimate the magnitude of the association with insurance represented by the odds ratios in the primary analyses, the logistic regression models also were used to predict the distributions of the dependent variables first using the actual values of the insurance status variable and then changing the value for the uninsured from 1 to 0 so that it simulates having insurance. The differences between the predicted distributions in these hypothetical models illustrate the magnitudes of the associations with not having insurance after controlling for differences in health conditions and other characteristics.

Sample Characteristics

The new chronic condition sample includes 10 485 cases (7954 individuals) and the unintentional injury sample includes 20 783 cases (15 866 individuals). The baseline health and sociodemographic characteristics (control variables) by insurance status for each sample are shown in Table 1.

Table Graphic Jump LocationTable 1. Baseline Health and Sociodemographic Characteristics*

Uninsured individuals make up 18.5% of the unintentional injury sample and 16.1% of the chronic condition sample. More uninsured individuals reported being in fair or poor health at baseline and had family incomes below 100% of the federal poverty level. Uninsured individuals also were more likely to be racial and ethnic minorities and to have less education.

The comparisons of the distributions of the control variables representing the types and locations of the injuries in the unintentional injury sample are shown in Table 2 and the variables representing the health conditions in the new chronic condition sample are shown in Table 3.

Table Graphic Jump LocationTable 2. Types and Locations of Injuries in the Unintentional Injury Sample*
Table Graphic Jump LocationTable 3. Health Conditions in the New Chronic Condition Sample*
Association of Insurance Status and Health Outcomes

The distributions of the dependent variables and the adjusted odds ratios from the logistic regression models for the insurance status variable in each sample are shown in Table 4. Uninsured individuals were significantly less likely to see a clinician following an unintentional injury or a new chronic condition. However, if an individual did see a clinician, he/she was equally likely to have further care recommended regardless of insurance status. Despite the equal recommendations for further care, uninsured individuals in the unintentional injury sample were significantly more likely to have received none of the recommended follow-up care and significantly less likely to have received all of the recommended follow-up care. In the new chronic condition sample, uninsured individuals also were significantly more likely to have received none of the recommended follow-up care and were significantly less likely to still be receiving treatment for their chronic condition.

Table Graphic Jump LocationTable 4. Distributions of Dependent Variables and Odds Ratios*

The models for the quantities of medical care indicate that uninsured individuals had significantly fewer office-based visits and prescription medicines in both samples. (The relative odds represent the proportional odds ratio of being in the next higher quantity category.) In the new chronic condition sample, uninsured individuals had significantly more emergency department visits and similar numbers of inpatient hospital stays as those with insurance. The number of emergency department visits did not vary significantly by insurance status in the unintentional injury sample.

Table 4 also shows that uninsured individuals in the chronic condition sample reported significantly worse short-term health changes at the first follow-up interview (approximately 3.5 months after the health shock), and that uninsured individuals in the unintentional injury sample were significantly more likely to report that they were not fully recovered and were no longer being treated. The difference in health change was not statistically significant between insured and uninsured persons in the unintentional injury sample at the second follow-up interview (approximately 7 months after the health shock). In contrast, the difference in health change persisted and remained statistically significant for those in the new chronic condition sample.

The results of the sensitivity analyses are reported in Table 5 and Table 6. The range of estimates from these alternative analyses cluster in a narrow range around those reported in Table 4, and the results generally remain statistically significant when the variation reduces the sample size.

Table Graphic Jump LocationTable 5. Results of Sensitivity Analyses for Unintentional Injury Sample*
Table Graphic Jump LocationTable 6. Results of Sensitivity Analyses for New Chronic Condition Sample*

Using the models to estimate the hypothetical change in outcomes that may occur if those who did not have insurance acquired insurance (Table 7), having insurance would be associated with an increase in the proportions seeing a clinician by 8.9% in the unintentional injury sample and by 7.8% in the new chronic condition sample. If insured, the proportion who received none of the recommended follow-up care would decrease by more than 40% in both samples. In the unintentional injury sample, most of the decrease would be offset by an increase in the proportion receiving all of the recommended care. In the new chronic condition sample, the proportions receiving some or all of the recommended care also would decline, with the percentage still receiving treatment absorbing the decreases in the other categories.

Table Graphic Jump LocationTable 7. Adjusted Differences Associated With Uninsured Individuals Hypothetically Having Insurance*

In these hypothetical models, the percentages of uninsured individuals receiving care would increase for outpatient visits, office-based visits, and number of prescription medications, whereas uninsured individuals would have fewer emergency department visits in the chronic condition sample, implying that the locus of care may shift from emergency departments to other ambulatory care settings.

In these hypothetical analyses, smaller proportions of uninsured individuals would experience substantial short-term health deteriorations (a change of 2 or more health states: from excellent health to good, fair, or poor; from very good health to fair or poor; or from good health to poor health) and correspondingly larger proportions would experience improved health. In the unintentional injury sample, the proportion of uninsured individuals who stop receiving treatment before they were fully recovered would decrease by 5.8% from the baseline value of 22.1%.

The results of these analyses suggest that an uninsured person who experiences an unintentional injury or a new chronic condition has greater difficulty obtaining recommended medical care and takes longer to return to full health, if at all. These findings are consistent with and reinforce earlier research showing that uninsured individuals receive significantly less care than those with insurance and have poorer health outcomes.6,7

The finding that there were no differences in the likelihood of having further care recommended if a clinician was seen suggests that either there were no differences in severity or, more probably, that the severity threshold for recommending care is higher for an uninsured person than a person with insurance. If the latter, the major and significant differences in receiving follow-up care reinforce the inference that differences in the amounts of care received are at least partially related to poorer health outcomes in uninsured individuals.

If health remains compromised, it could make it more difficult for an uninsured person to obtain health insurance in the future, and could also contribute to reduced labor force participation, lower productivity, and lower future income. Although the change in health is measured over only a relatively short period, other research suggests that the progression from good to poor health resulting from lack of health insurance is cumulative and gradually leads to higher mortality rates for uninsured individuals over time.13

The findings for uninsured individuals with a new chronic condition are of particular concern. Because chronic conditions generally require care over an extended period, the finding that the uninsured are more likely to report no longer being treated at the first follow-up interview suggests that their care may have been inadequate. Moreover, their greater dependence on emergency departments for care probably increases the likelihood that their care will be episodic and lack continuity. For individuals who require treatment for chronic conditions, a continuing relationship with a clinician and appropriate medication use may be especially important for improving or maintaining health status.

Although the differences between the proportions of uninsured and insured individuals who experienced much poorer health were not large in absolute terms, these patients often require costly care and are at increased risk for death and disability. Thus, it is possible that savings in future medical costs may be larger than implied by the small absolute percentage changes between insured and uninsured individuals.

The problem of an increasing proportion of individuals without health insurance is likely to worsen if health insurance premium growth continues to outpace income and wage growth, especially for those with lower incomes.29,30 The results of this analysis imply that the failure to address the problem of no insurance for US individuals will have adverse health consequences. Moreover, the fact that these consequences apply to uninsured individuals who experienced unintentional injuries or new chronic conditions runs counter to the perception that the uninsured receive care, either through the safety net or their own resources, when they really need it (eg, when they experience a health shock from an unintentional injury or develop a new chronic condition).

This study has several important limitations. Despite the robustness of the results in the sensitivity analyses, bias due to unobserved differences between uninsured and insured individuals is still possible. Two recent studies of insurance and health status used statistical methods to adjust for possible confounding from unobserved factors in longitudinal samples and found that the adverse effects of not having insurance on health status were understated.31,32 Thus, even if the results of the sensitivity analyses also are biased, other research suggests that the bias is conservative and the qualitative conclusion that uninsured individuals have adverse outcomes in medical care use and health status would still be true.

Second, the health change measure is constructed from subjective self-reported assessments of general health status. A more detailed and more objective health measure, such as the Short Form 3633 or the Health and Limitations Index34 would provide a more accurate measure of the overall change in health as well as changes in particular dimensions of health status.

Third, the conditions in the new chronic condition sample are neither representative of all chronic conditions nor were they consistently coded from medical records. However, the broad similarity of the results across the 2 samples, which had varying mixes of underlying conditions, suggests that the results are not specific to the particular conditions represented in the samples.

Fourth, measures of the quantities of medical care received may not capture variations in either the content or quality of care. Other research documents that uninsured individuals receive lower quality care,35,36 which implies that the differences in quantities found in this analysis may understate the difference in the quality-adjusted amounts of care received.

And fifth, the follow-up observation period is too short to determine whether the differences in health outcomes are permanent or transitory, especially in the unintentional injury sample. Longer observation periods are necessary to investigate this issue but data with the requisite information on insurance status and medical care use do not currently exist.

Among individuals who experienced a health shock caused by an unintentional injury or a new chronic condition, uninsured individuals reported receiving less medical care and poorer short-term changes in health than individuals with insurance. If the proportion of individuals without insurance continues to increase, more and more persons may experience preventable deteriorations in their health.

Corresponding Author: Jack Hadley, PhD, The Urban Institute, 2100 M St NW, Washington, DC 20037 (jhadley@ui.urban.org).

Author Contributions: Dr Hadley had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Dr Hadley was responsible for the conception and design; data acquisition; analysis and interpretation of the data; drafting and revising of the manuscript; statistical analysis; obtaining funding; and administrative management.

Financial Disclosures: None reported.

Funding/Support: The research for this article was supported by contracts 04-1102-700 and 01-1282-800 from the Kaiser Family Foundation, Washington, DC.

Role of the Sponsor: The sponsor had no role in the design and conduct of the study; or the collection, management, analysis, and interpretation of the data. The sponsor's staff commented on a preliminary draft of the manuscript but had no role in the preparation, final review, or approval of the article.

Acknowledgment: Joel Ruhter and Matthew Cravens provided computer programming and research assistance at The Urban Institute and were compensated as part of their salaries.

Families USA.  Getting Less Care: The Uninsured With Chronic Health Conditions. Washington, DC: Families USA; 2001
Stagnitti M. Medical care and treatment for chronic conditions, 2000. http://www.meps.ahrq.gov/mepsweb/data_files/publications/st5/stat05.pdf. Accessed February 19, 2007
Doyle J. Health insurance, treatments and outcomes: using auto accidents as health shocks.  Rev Econ Stat. 2005;87:256-270
Link to Article
Haas JS, Goldman L. Acutely injured patients with trauma in Massachusetts: differences in care and mortality, by insurance status.  Am J Public Health. 1994;84:1605-1608
PubMed   |  Link to Article
Tilford JM, Simpson PM, Yeh TS.  et al.  Variation in therapy and outcome for pediatric head trauma patients.  Crit Care Med. 2001;29:1056-1061
PubMed   |  Link to Article
Hadley J. Sicker and poorer: the consequences of being uninsured.  Med Care Res Rev. 2003;60:(2 suppl)  3S-75S
PubMed   |  Link to Article
Institute of Medicine.  Care Without Coverage. Washington, DC: National Academy Press; 2002
Levy H, Meltzer D. What do we really know about whether health insurance affects health? In: McLaughlin C, ed. Health Policy and the Uninsured. Washington, DC: Urban Institute Press; 2005
Newhouse J.Insurance Experiment Group.  Free for All? Lessons From the RAND Health Insurance Experiment. Cambridge, Mass: Harvard University Press; 1993
McWilliams JM, Zaslavsky AM, Meara E, Ayanian JZ. Impact of Medicare coverage on basic clinical services for previously uninsured adults.  JAMA. 2003;290:757-764
PubMed   |  Link to Article
Lichtenberg F. The effects of Medicare on health care utilization and outcomes. In: Garber A, ed. Frontiers in Health Policy Research. Vol 5. Cambridge, Mass: MIT Press; 2002
McWilliams JM, Zaslavsy A, Meara E, Ayanian J. Health insurance coverage and mortality among the near-elderly.  Health Aff (Millwood). 2004;23:223-233
PubMed   |  Link to Article
Baker DW, Sudano JJ, Durazo-Arvizu R.  et al.  Health insurance coverage and the risk of decline in overall health and death among the near elderly, 1992-2002.  Med Care. 2006;44:277-282
PubMed   |  Link to Article
Baker DW, Sudano JJ, Albert JM.  et al.  Loss of health insurance and the risk for a decline in self-reported health and physical functioning.  Med Care. 2002;40:1126-1131
PubMed   |  Link to Article
Baker DW, Sudano JJ, Albert JM.  et al.  Lack of health insurance and decline in overall health in late middle age.  N Engl J Med. 2001;345:1106-1112
PubMed   |  Link to Article
Sorlie PD, Johnson NJ, Backlund E, Bradham D. Mortality in the uninsured compared with that in persons with public and private health insurance.  Arch Intern Med. 1994;154:2409-2416
PubMed   |  Link to Article
Franks P, Clancy C, Gold M. Health insurance and mortality: evidence from a national cohort.  JAMA. 1993;270:737-741
PubMed   |  Link to Article
Agency for Healthcare Research and Quality.  Detailed information about the Medical Expenditure Panel Survey. http://www.meps.ahrq.gov/. Accessed September 27, 2006
Agency for Healthcare Research and Quality.  MEPS-HC response rates by panel. http://www.meps.ahrq.gov/mepsweb/survey_comp/hc_response_rate.jsp. Accessed February 19, 2007
Kreider B, Hill S. What do we know about the effect of insurance when insurance status is sometimes misreported? Paper presented at: Informing Policy on Health Care Costs and Insurance Coverage Conference; May 3, 2006; Washington, DC
Agency for Healthcare Research and Quality.  Medical Expenditure Panel Survey HC-078: medical conditions, Appendix 4. See website. Accessed February 19, 2007
Hwang W, Weller W, Ireys H, Anderson G. Out-of-pocket medical spending for care of chronic conditions.  Health Aff (Millwood). 2001;20:267-278
PubMed   |  Link to Article
World Health Organization.  International Classification of Diseases, Ninth Revision (ICD-9). Geneva, Switzerland: World Health Organization; 1977
DeSalvo K, Fan V, McDonell M, Fihn S. Predicting mortality and healthcare utilization with a single question.  Health Serv Res. 2005;40:1234-1246
PubMed   |  Link to Article
Idler EL, Benyamini Y. Self-rated health and mortality: a review of twenty-seven community studies.  J Health Soc Behav. 1997;38:21-37
PubMed   |  Link to Article
Diehr P, Williamson J, Patrick D.  et al.  Patterns of self-rated health in older adults before and after sentinel health events.  J Am Geriatr Soc. 2001;49:36-44
PubMed   |  Link to Article
StataCorp.  Binary, count, and limited dependent variables. http://www.stata.com/capabilities/binary.html. Accessed September 26, 2006
Engle R, Hendry D, Richard J. Exogeneity.  Econometrica. 1983;51:277-304
Link to Article
Gabel J, Claxton G, Gill I.  et al.  Health benefits in 2005: premium increases slow down, coverage continues to erode.  Health Aff (Millwood). 2005;24:1273-1280
PubMed   |  Link to Article
Hadley J. The effects of recent employment changes and premium increases on adults' insurance coverage.  Med Care Res Rev. 2006;63:447-476
PubMed   |  Link to Article
Hadley J, Waidmann T. Health insurance and health at age 65: implications for Medicare spending on new beneficiaries.  Health Serv Res. 2006;41:429-451
PubMed   |  Link to Article
Dor A, Sudano J, Baker D. The effects of private insurance on the health of older, working age adults: evidence from the health and retirement study.  Health Serv Res. 2006;41:(3 part I)  759-787
PubMed   |  Link to Article
 Short Form 36 Web site. http://www.sf-36.org/. Accessed September 26, 2006
Ericson P, Wilson R, Shannon I. Years of healthy life. In: Statistical Notes No. 7. Hyattsville, Md: National Center for Health Statistics; 1995
Schoen C, DesRoches C. Uninsured and Unstably Insured: the Importance of Continuous Insurance Coverage.  Health Serv Res. 2000;35:187-206
PubMed
Schoen C, Puleo E. Low–income working families at risk: uninsured and underserved.  J Urban Health. 1998;75:30-49
PubMed   |  Link to Article

Figures

Tables

Table Graphic Jump LocationTable 1. Baseline Health and Sociodemographic Characteristics*
Table Graphic Jump LocationTable 2. Types and Locations of Injuries in the Unintentional Injury Sample*
Table Graphic Jump LocationTable 3. Health Conditions in the New Chronic Condition Sample*
Table Graphic Jump LocationTable 4. Distributions of Dependent Variables and Odds Ratios*
Table Graphic Jump LocationTable 5. Results of Sensitivity Analyses for Unintentional Injury Sample*
Table Graphic Jump LocationTable 6. Results of Sensitivity Analyses for New Chronic Condition Sample*
Table Graphic Jump LocationTable 7. Adjusted Differences Associated With Uninsured Individuals Hypothetically Having Insurance*

References

Families USA.  Getting Less Care: The Uninsured With Chronic Health Conditions. Washington, DC: Families USA; 2001
Stagnitti M. Medical care and treatment for chronic conditions, 2000. http://www.meps.ahrq.gov/mepsweb/data_files/publications/st5/stat05.pdf. Accessed February 19, 2007
Doyle J. Health insurance, treatments and outcomes: using auto accidents as health shocks.  Rev Econ Stat. 2005;87:256-270
Link to Article
Haas JS, Goldman L. Acutely injured patients with trauma in Massachusetts: differences in care and mortality, by insurance status.  Am J Public Health. 1994;84:1605-1608
PubMed   |  Link to Article
Tilford JM, Simpson PM, Yeh TS.  et al.  Variation in therapy and outcome for pediatric head trauma patients.  Crit Care Med. 2001;29:1056-1061
PubMed   |  Link to Article
Hadley J. Sicker and poorer: the consequences of being uninsured.  Med Care Res Rev. 2003;60:(2 suppl)  3S-75S
PubMed   |  Link to Article
Institute of Medicine.  Care Without Coverage. Washington, DC: National Academy Press; 2002
Levy H, Meltzer D. What do we really know about whether health insurance affects health? In: McLaughlin C, ed. Health Policy and the Uninsured. Washington, DC: Urban Institute Press; 2005
Newhouse J.Insurance Experiment Group.  Free for All? Lessons From the RAND Health Insurance Experiment. Cambridge, Mass: Harvard University Press; 1993
McWilliams JM, Zaslavsky AM, Meara E, Ayanian JZ. Impact of Medicare coverage on basic clinical services for previously uninsured adults.  JAMA. 2003;290:757-764
PubMed   |  Link to Article
Lichtenberg F. The effects of Medicare on health care utilization and outcomes. In: Garber A, ed. Frontiers in Health Policy Research. Vol 5. Cambridge, Mass: MIT Press; 2002
McWilliams JM, Zaslavsy A, Meara E, Ayanian J. Health insurance coverage and mortality among the near-elderly.  Health Aff (Millwood). 2004;23:223-233
PubMed   |  Link to Article
Baker DW, Sudano JJ, Durazo-Arvizu R.  et al.  Health insurance coverage and the risk of decline in overall health and death among the near elderly, 1992-2002.  Med Care. 2006;44:277-282
PubMed   |  Link to Article
Baker DW, Sudano JJ, Albert JM.  et al.  Loss of health insurance and the risk for a decline in self-reported health and physical functioning.  Med Care. 2002;40:1126-1131
PubMed   |  Link to Article
Baker DW, Sudano JJ, Albert JM.  et al.  Lack of health insurance and decline in overall health in late middle age.  N Engl J Med. 2001;345:1106-1112
PubMed   |  Link to Article
Sorlie PD, Johnson NJ, Backlund E, Bradham D. Mortality in the uninsured compared with that in persons with public and private health insurance.  Arch Intern Med. 1994;154:2409-2416
PubMed   |  Link to Article
Franks P, Clancy C, Gold M. Health insurance and mortality: evidence from a national cohort.  JAMA. 1993;270:737-741
PubMed   |  Link to Article
Agency for Healthcare Research and Quality.  Detailed information about the Medical Expenditure Panel Survey. http://www.meps.ahrq.gov/. Accessed September 27, 2006
Agency for Healthcare Research and Quality.  MEPS-HC response rates by panel. http://www.meps.ahrq.gov/mepsweb/survey_comp/hc_response_rate.jsp. Accessed February 19, 2007
Kreider B, Hill S. What do we know about the effect of insurance when insurance status is sometimes misreported? Paper presented at: Informing Policy on Health Care Costs and Insurance Coverage Conference; May 3, 2006; Washington, DC
Agency for Healthcare Research and Quality.  Medical Expenditure Panel Survey HC-078: medical conditions, Appendix 4. See website. Accessed February 19, 2007
Hwang W, Weller W, Ireys H, Anderson G. Out-of-pocket medical spending for care of chronic conditions.  Health Aff (Millwood). 2001;20:267-278
PubMed   |  Link to Article
World Health Organization.  International Classification of Diseases, Ninth Revision (ICD-9). Geneva, Switzerland: World Health Organization; 1977
DeSalvo K, Fan V, McDonell M, Fihn S. Predicting mortality and healthcare utilization with a single question.  Health Serv Res. 2005;40:1234-1246
PubMed   |  Link to Article
Idler EL, Benyamini Y. Self-rated health and mortality: a review of twenty-seven community studies.  J Health Soc Behav. 1997;38:21-37
PubMed   |  Link to Article
Diehr P, Williamson J, Patrick D.  et al.  Patterns of self-rated health in older adults before and after sentinel health events.  J Am Geriatr Soc. 2001;49:36-44
PubMed   |  Link to Article
StataCorp.  Binary, count, and limited dependent variables. http://www.stata.com/capabilities/binary.html. Accessed September 26, 2006
Engle R, Hendry D, Richard J. Exogeneity.  Econometrica. 1983;51:277-304
Link to Article
Gabel J, Claxton G, Gill I.  et al.  Health benefits in 2005: premium increases slow down, coverage continues to erode.  Health Aff (Millwood). 2005;24:1273-1280
PubMed   |  Link to Article
Hadley J. The effects of recent employment changes and premium increases on adults' insurance coverage.  Med Care Res Rev. 2006;63:447-476
PubMed   |  Link to Article
Hadley J, Waidmann T. Health insurance and health at age 65: implications for Medicare spending on new beneficiaries.  Health Serv Res. 2006;41:429-451
PubMed   |  Link to Article
Dor A, Sudano J, Baker D. The effects of private insurance on the health of older, working age adults: evidence from the health and retirement study.  Health Serv Res. 2006;41:(3 part I)  759-787
PubMed   |  Link to Article
 Short Form 36 Web site. http://www.sf-36.org/. Accessed September 26, 2006
Ericson P, Wilson R, Shannon I. Years of healthy life. In: Statistical Notes No. 7. Hyattsville, Md: National Center for Health Statistics; 1995
Schoen C, DesRoches C. Uninsured and Unstably Insured: the Importance of Continuous Insurance Coverage.  Health Serv Res. 2000;35:187-206
PubMed
Schoen C, Puleo E. Low–income working families at risk: uninsured and underserved.  J Urban Health. 1998;75:30-49
PubMed   |  Link to Article
CME
Also Meets CME requirements for:
Browse CME for all U.S. States
Accreditation Information
The American Medical Association is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. The AMA designates this journal-based CME activity for a maximum of 1 AMA PRA Category 1 CreditTM per course. Physicians should claim only the credit commensurate with the extent of their participation in the activity. Physicians who complete the CME course and score at least 80% correct on the quiz are eligible for AMA PRA Category 1 CreditTM.
Note: You must get at least of the answers correct to pass this quiz.
Please click the checkbox indicating that you have read the full article in order to submit your answers.
Your answers have been saved for later.
You have not filled in all the answers to complete this quiz
The following questions were not answered:
Sorry, you have unsuccessfully completed this CME quiz with a score of
The following questions were not answered correctly:
Commitment to Change (optional):
Indicate what change(s) you will implement in your practice, if any, based on this CME course.
Your quiz results:
The filled radio buttons indicate your responses. The preferred responses are highlighted
For CME Course: A Proposed Model for Initial Assessment and Management of Acute Heart Failure Syndromes
Indicate what changes(s) you will implement in your practice, if any, based on this CME course.

Multimedia

Some tools below are only available to our subscribers or users with an online account.

Web of Science® Times Cited: 66

Related Content

Customize your page view by dragging & repositioning the boxes below.

Articles Related By Topic
Related Collections
PubMed Articles
JAMAevidence.com