Skip to main content

Abstract

Objective:

Because of pervasive poor general medical and mental health status among patients receiving Medicaid, there has been substantial debate about whether Medicaid, as currently financed and delivered, is better than no insurance. The study aimed to address whether insurance status is associated with the subsequent incidence and persistence of common mental disorders.

Methods:

Data came from a nationally representative U.S. population–based longitudinal survey that assessed mental disorders at two time points three years apart. Propensity score methods were used to adjust for potential confounding and to assess the association between three mutually exclusive insurance status groups (no insurance, private insurance only, and Medicaid only) and the subsequent incidence and persistence of mood, anxiety, and substance use disorders for persons ages 18–65 (N=26,410).

Results:

Compared with private insurance, lack of insurance was associated with higher odds of both the incidence and persistence of substance use disorders and with higher odds of persistence of any mood or anxiety disorder. Compared with having private insurance, having Medicaid insurance was associated with increased odds of persistent mood and anxiety disorders during follow-up. Overall, findings did not significantly differ between the uninsured and Medicaid groups.

Conclusions:

The findings do not support prior reports that U.S. adults with Medicaid have worse mental health outcomes than uninsured adults. Lacking insurance may put individuals at higher risk of developing substance use disorders, and uninsured individuals with preexisting mental conditions were more likely to have mood, anxiety, and substance use problems that persist over time.
Although the Affordable Care Act (ACA) extends eligibility for health insurance to millions of Americans (1), there has been substantial controversy as to whether providing Medicaid as currently financed and delivered to those who have no or minimal insurance will positively affect health status (2). Currently, approximately 50% of U.S. states have expanded Medicaid (3). Results from several studies indicated that the general health of individuals with Medicaid is worse or similar to those without it (4,5). On the basis of these findings, some have suggested that Medicaid is worse than having no insurance (6). Others have argued that the poorer mental and general health of Medicaid recipients is due to selection bias because these individuals tend to have lower socioeconomic status and worse baseline health status than those with private or no insurance (2). Most studies examining health status and insurance type have been limited by lack of adjustment for potential confounding variables (2). Studies that have used sophisticated analytic methods and quasi-experimental designs have found a reduction in mortality associated with Medicaid (1,7). Furthermore, much research examining the association of insurance type with health status has focused on general medical conditions (4,5) or those conditions in combination with mental health conditions (8). Fewer studies have examined the association of insurance type with mental health status. Given that Medicaid finances mental health care more heavily than general medical care, any potential public health impact of Medicaid coverage should be more pronounced in the former group.
A small number of studies have examined associations between insurance status, mental disorders, and adequacy of care for mental health problems. Findings from cross-sectional studies have demonstrated that people without insurance and those with Medicaid have a higher prevalence of mental disorders than those with private insurance (911). Zatzick and colleagues (12) found that lack of insurance was associated with increased odds of posttraumatic stress disorder (PTSD) in a large cross-sectional sample of physically injured individuals requiring hospitalization. A similar finding was noted among postpartum women, where lack of insurance was a risk factor for PTSD symptoms (13). Lack of health insurance might increase the risk of later mental disorder stemming from the financial stresses of medical expenses incurred by the individual or the individual’s family members. A recent study of Oregon residents found that individuals who received Medicaid coverage by randomization (that is, through a lottery) had no improvements in general health status but had lower levels of financial strain and depressive symptoms compared with those without Medicaid (14). However, the study was criticized by some who suggested that reductions in financial strain and depressive symptoms could have occurred by giving cash prizes to lottery winners rather than providing Medicaid specifically (15). The Oregon study was also limited because the assessment of depressive symptoms was based on a screening instrument rather than a diagnostic interview of depression, and other mental disorders such as anxiety and substance use disorders were not assessed (15).
There are several limitations in the state of knowledge regarding the relationship between health insurance status and mental disorders. Most studies have been cross-sectional and thus preclude causal inferences. The few longitudinal studies assessing this association have not adequately addressed potential confounding variables. Thus it remains unclear whether insurance status is associated with mental disorder outcomes. Finally, there is little in the way of empirical data on the impact of insurance status on persistence of mental disorders.
To address these important issues, we examined whether insurance type is associated with an increased risk of occurrence or persistence of mental health problems. We used a unique, nationally representative, longitudinal, U.S. population based survey to investigate the following questions: Is health insurance status associated with subsequent development of mental disorders? Is health insurance status associated with persistence of mental disorders? We assessed these queries, using propensity score–based analyses, to address concerns regarding potential confounding factors.

Methods

Sample

Data from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) were used for this study (16,17,18). The NESARC is a nationally representative longitudinal survey of noninstitutionalized adults in the United States and is conducted by the National Institute on Alcohol Abuse and Alcoholism. With the consent of the respondents, the interviews were conducted face to face by trained lay interviewers. The survey consists of two waves. Wave 1 was conducted in 2001–2002 with a sample of 43,093 (ages ≥18). Of these, 34,653 individuals participated in wave 2 (2004–2005). The remaining participants from wave 1 did not participate in wave 2 as a result of death, deportation, impairment, active military duty, and other, less common reasons. The overall response rate was 70.9%.

Insurance Groups

At wave 1, all participants were asked about their insurance status with the following question: “Are you currently covered by either Medicare; Medicaid or (local name); CHAMPUS, CHAMPVA, the VA, or other military health care; or health insurance obtained privately or through a current or former employer or union?” The study sample was limited to individuals under the age of 65, because older individuals would be eligible for Medicare. We included in the study individuals without insurance, Medicaid only, or private insurance only, whereas we excluded from the sample those with other types of insurance, such as through the U.S. Department of Veterans Affairs (VA), and those with more than one type of insurance (VA and Medicaid) to test our specific hypotheses regarding the uninsured and Medicaid. This resulted in a sample of 26,410.

Mood, Anxiety, and Substance Use Disorders

DSM-IV axis I mental disorders were assessed at both waves with the Alcohol Use Disorder and Associated Disability Interview Schedule (AUDADIS-IV) (19), a structured diagnostic interview that has been shown to be a reliable and valid measure of mental disorders (16). Baseline lifetime mood, anxiety, and substance use disorders were available. Mood disorders included major depression, dysthymia, mania, and hypomania. Anxiety disorders included panic disorder with or without agoraphobia, agoraphobia without panic disorder, social phobia, specific phobia, and generalized anxiety disorder. Substance use disorders included abuse and dependence for both alcohol and illicit drugs. For our analyses, four groups of mental disorders were examined: any mood disorder; any anxiety disorder; any substance use disorder; and any mood, anxiety, or substance use disorder. Similar to previous approaches in defining persistence of mental disorders (20,21), in this study respondents were categorized as having persistence of a baseline disorder if they met criteria for a disorder in the year prior to wave 1 and continued to meet criteria for the same category of disorder during the year prior to wave 2. Incident mood, anxiety, and substance use disorders were defined as the presence of one of the disorders of interest at wave 2 among individuals who did not meet criteria for lifetime or current disorder at baseline.

Potentially Confounding Variables

Seventeen potentially confounding sociodemographic, psychiatric, and general health variables were included in the statistical analyses examining associations between insurance coverage and psychiatric outcomes via propensity score models: sex, age, marital status, education level, household income, urbanicity, race-ethnicity, number of individuals in the household, having children under age 17, current pregnancy, census division, employment status, presence of any adverse childhood event, history of mental health treatment seeking, baseline psychotic disorder or schizophrenia assessed via self-report (22), personality disorder (including all ten DSM-IV personality disorders), and number of past-year (baseline) general health conditions diagnosed by a health professional (including arteriosclerosis, hypertension, liver cirrhosis, other form of liver disease, angina pectoris, tachycardia, myocardial infarction, other form of heart disease, stomach ulcer, gastritis, arthritis, and obesity) (23). Furthermore, past-year health-related quality of life (HRQoL) was included; it was assessed with the 12-Item Medical Outcomes Study–Short Form (SF-12) (24); this assessment produces mental and physical component scores that range from 0 to 100, with a standardized mean of 50; higher scores indicate better quality of life. Information for each of the potential confounding variables was gathered during wave 1, with the exception of some personality disorders and adverse childhood experiences, which were assessed at wave 2. Most of these variables were categorized on the basis of frequency analyses or conceptual categories commonly used in the literature, with the exception of sex, urbanicity, and census division, which were grouped on the basis of preexisting categories in the data set (25).

Analyses

Sampling weights provided by the NESARC were applied to achieve representativeness of the U.S. population (16). Taylor series linearization was performed as a variance estimation technique to account for the complex sampling design. A total of 12 sets of propensity scores (four baseline mental disorder groups—any mood disorder; any anxiety disorder; any substance-use disorder; and any mood, anxiety, and substance-use disorder—by three insurance comparisons—no insurance versus private insurance, Medicaid insurance versus private insurance, and Medicaid insurance versus no insurance) were calculated with logistic regression models. For any mood disorder, any anxiety disorder, and any substance use disorder, the other two baseline disorder groups (that were not the disorder group of interest) were added to the propensity score model in addition to the other variables to further account for possible confounding effects. After the propensity score had been generated, each score was then divided into deciles for inclusion in the final model. To assess whether the propensity score served its intended purpose of balancing the comparison groups at baseline, we checked for the balancing composition of the comparison groups by examining the groups’ characteristics. The observed characteristics among the comparison groups after adjustment for the propensity scores were very similar, suggesting that the application of the propensity score was successful.
Multivariable logistic regression analysis tested the association between insurance status and the presence of lifetime mental disorders at baseline, as well as the persistence and incidence of mental disorders at follow-up. Deciles of propensity score were included in all models. All analyses were conducted with SUDAAN, version 11.0 (26), with the exception of the propensity score calculation, which was conducted with SPSS, version 22.0. Because of the large number of inferential analyses that were conducted, we used a conservative nominal alpha level of .01 to assess statistical significance.

Results

Table 1 shows the frequency distribution and bivariate regression analyses for insurance type and the univariable logistic regression analyses for sociodemographic factors, childhood adversity, pregnancy status, self-reported psychotic disorders, personality disorders, and HRQoL. Lack of insurance was strongly associated with low household income, unemployment, and nonwhite race (odds ratios above 3). Relative to private insurance, lack of insurance was also associated with a history of childhood adversity, personality disorders, self-reported psychotic disorders, and lower HRQoL. Medicaid insurance was strongly associated with being in a racial-ethnic minority group, having self-reported psychotic disorders, being pregnant at baseline, and having lower HRQoL.
TABLE 1. Baseline characteristics associated with baseline insurance status in the National Epidemiologic Survey on Alcohol and Related Conditions (N=26,410)
 Mutually exclusive insurance groups at baselineBivariate association 
 Private insuranceNo insuranceMedicaid insuranceNone versus privateMedicaid versus privateMedicaid versus none 
CharacteristicN%N%N%OR99% CIpOR99% CIpOR99% CIp
Sex               
 Male (reference)8,26848.32,71749.530629.61.00  1.00  1.00  
 Female10,53151.73,44150.51,14770.4.93.84–1.04.0922.151.71–2.70<.0012.301.79–2.97<.001
Age               
 18–29 (reference)3,71821.12,03536.153138.31.00  1.00  1.00  
 30–447,57738.32,19734.359539.3.51.46–.57<.001.55.43–.71<.0011.07.84–1.38.447
 45–657,50440.61,92629.732721.9.42.37–.47<.001.29.21–.40<.001.70.50–.98.007
Marital status               
 Married, common law (reference)11,38269.52,80752.044040.91.00  1.00  1.00  
 Separated, divorced, or widowed3,28811.41,21314.041123.41.601.38–1.86<.0013.392.58–4.45<.0012.111.57–2.84<.001
 Never married4,12919.22,13834.060235.62.342.08–2.64<.0013.102.35–4.10<.0011.33.99–1.77.012
Education               
 Less than high school (reference)1,4257.01,49822.651634.31.00  1.00  1.00  
 High school4,84725.91,97232.251936.0.38.32–.45<.001.29.23–.37<.001.77.59–1.00.009
 Some postsecondary education12,52766.12,68845.241829.7.21.18–.24<.001.09.07–.12<.001.44.34–.58<.001
Household income               
 $0–$19,999 (reference)2,18110.32,38933.81,01762.71.00  1.00  1.00  
 $20,000–$39,9994,95422.61,91829.532226.4.40.35–.46<.001.19.15–.25<.001.48.37–.63<.001
 $40,000–$69,9996,12232.41,18521.2858.1.20.17–.24<.001.04.03–.07<.001.21.13–.34<.001
 ≥$70,0005,54234.773315.62929.2.14.11–.16<.001.01.01–.03<.001.12.06–.19<.001
Urbanicity               
 Urban5,86326.12,34233.372242.01.00  1.00  1.00  
 Rural12,93673.93,81666.773158.0.69.60–.81<.001.48.35–.65<.001.69.51–.92<.001
Race-ethnicity               
 White (reference)11,77376.72,67058.347548.31.00  1.00  1.00  
 Black3,1459.21,25413.150325.01.851.61–2.13<.0014.333.33–5.63<.0012.331.76–3.09<.001
 Hispanic/Latino2,9968.11,90420.942120.63.482.97–4.07<.0014.072.63–6.30<.0011.17.75–1.83.360
 Other8856.13307.8546.11.681.33–2.11<.0011.56.98–2.50.014.93.57–1.51.693
N individuals in the household               
 1 (reference)4,38614.21,37313.924712.01.00  1.00  1.00  
 25,66531.51,56325.029819.4.81.71–.92<.001.75.55–1.02.014.93.66–1.31.556
 33,59821.41,17520.931620.51.01.87–1.16.9241.15.82–1.60.2721.14.79–1.64.336
 ≥45,15033.02,04740.259248.11.271.09–1.48<.0011.751.27–2.43<.0011.38.96–2.00.023
Child <17               
 No (reference)10,48255.33,20152.341029.91.00  1.00  1.00  
 Yes8,31744.72,95747.81,04370.11.141.02–1.28.0022.892.23–3.75<.0012.531.91–3.35<.001
Currently pregnant               
 No (reference)8,66042.92,94043.996359.01.00  1.00  1.00  
 Yes2071.160.9665.5.82.51–1.32.2723.622.25–5.82<.0014.412.40–8.08<.001
 Other (males and women over 55 who were not asked the question)9,74456.03,10055.239535.6.99.89–1.10.728.48.39–.59<.001.48.39–.60<.001
Census division               
 New England (reference)9925.617513.11197.11.00  1.00  1.00  
 Middle Atlantic2,59614.483614.722213.01.871.25–2.79<.001.68.24–1.98.346.37.13–1.00.010
 East North Central3,18618.469012.223517.41.18.88–1.57.134.69.36–1.30.126.58.26–1.31.081
 West North Central1,3578.32585.4696.11.14.83–1.58.273.53.23–1.23.050.46.17–1.24.042
 South Atlantic3,49317.51,32120.619912.32.121.55–2.90<.001.51.26–.99.009.24.11–.54<.001
 East South Central1,1106.23425.915612.41.691.20–2.90<.0011.43.72–2.86.172.85.36–1.98.605
 West South Central1,8018.11,03714.2765.33.192.30–4.43<.001.47.23–.95.006.15.06–.35<.001
 Mountain1,3776.94887.8834.72.031.39–2.98<.001.49.23–1.05.156.24.10–.58<.001
 Pacific2,88714.51,01116.129421.92.021.42–2.86<.0011.13.56–2.28.653.56.25–1.27.064
Employment status               
 Employed (reference)15,89283.24,12168.058842.41.00  1.00  1.00  
 Unemployed4912.5675511.343028.25.344.40–6.49<.00121.4116.35–28.03<.0014.013.05–5.27<.001
 Other2,41614.21,28220.843529.41.761.54–2.00<.0013.933.16–4.88<.0012.241.78–2.81<.001
Any adverse childhood event               
 No (reference)8,89949.12,67445.147632.61.00  1.00  1.00  
 Yes9,77550.93,39954.996067.41.171.04–1.32<.0011.971.64–2.37<.0011.691.37–2.08<.001
Any treatment seeking               
 No (reference)1,36756.641963.512558.61.00  1.00  1.00  
 Yes1,04943.426236.57441.5.75.55–1.01.013.94.57–1.53.7261.26.71–2.21.284
Any baseline psychotic disorders               
 No (reference)18,40399.75,98299.41,35195.61.00  1.00  1.00  
 Yes65.345.6664.41.83.94–3.53.01813.977.81–24.96<.0017.653.74–15.65<.001
Any baseline personality disorders               
 No (reference)14,69479.04,49273.190162.11.00  1.00  1.00  
 Yes4,10521.01,66626.955237.91.361.18–1.57<.0012.251.84–2.77<.0011.661.35–2.04<.001
N of general medical conditions (M±SE).6±.0 1.0±.6 .6±.0 .93.88–.99.0021.341.24–1.45<.0011.391.28–1.51<.001
SF-12 mental component score (M±SE)a53.4±.1 52.1±.2 47.4±.4 .99.98–.99<.001.94.93–.95<.001.97.96–.97<.001
SF-12 physical component score (M±SE)a53.7±.1 52.2±.2 46.1±.5 .98.98–.99<.001.94.93–.94<.001.96.95–.96<.001
a
From the Medical Outcomes Study 12-Item Short Form (SF-12). Possible mental component or physical component scores range from 0 to 100, with higher scores indicating better quality of life.
Table 2 shows the cross-sectional associations of baseline mental health status with insurance status. In adjusted models, persons with Medicaid insurance were not significantly different from those with no insurance or private insurance with regard to lifetime mood, anxiety, or substance use disorders. However, people without insurance had a significantly lower likelihood of any mood, anxiety, or substance use disorder compared with those with private insurance.
TABLE 2. Cross-sectional analysis of baseline insurance status and lifetime history of mental disorders in the National Epidemiologic Survey on Alcohol and Related Conditions at baseline
Baseline mental health statusBaseline insurance groupaBivariate and propensity-based adjusted models comparing three insurance groups
PrivateNo insuranceMedicaidNo insurance versus privatebMedicaid versus privatebMedicaid versus no insuranceb
N%N%N%OR99% CIpAORc99% CIpOR99% CIpAORc99% CIpOR99% CIpAORc99% CIp
Lifetime mood, anxiety, or substance use disorder10,35056.73,05451.783760.1.80.71–.91<.001.86.75–.99.0051.13.90–1.42.144.95.74–1.24.6361.411.13–1.76<.0011.05.81–1.36.638
Lifetime mood disorder3,98921.01,32821.946633.21.03.90–1.18.5721.01.87–1.17.8101.841.51–2.26<.0011.06.81–1.38.5581.791.43–2.24<.0011.09.84–1.42.379
Lifetime anxiety disorder3,47119.01,06217.637025.5.90.78–1.03.049.89.76–1.05.0661.441.19–1.74<.001.92.71–1.19.3911.601.30–1.97<.001.97.79–1.27.986
Lifetime substance use disorder7,69043.22,23939.557843.5.85.74–.97.002.92.79–1.06.116.99.80–1.24.9371.05.82–1.34.6201.17.95–1.45.0481.08.84–1.38.434
a
Groups were mutually exclusive.
b
The reference group is the latter category of the pair.
c
Adjusted odds ratio, adjusted for propensity score deciles. Twelve pairs of propensity scores were generated for AORs. All propensity scores were calculated with the covariates included in Table 1.
Table 3 shows the associations between baseline insurance status and risk of persistent mental disorders at wave 2. In adjusted models, lack of insurance at baseline was associated with an increased likelihood of a persistent mood, anxiety, or substance use disorder compared with having private insurance. Compared with private insurance, Medicaid insurance was significantly associated with an increased likelihood of persistent mood and anxiety disorders but not compared with no insurance.
TABLE 3. Baseline insurance status and persistence of baseline past-year mental disorders during the third year of the National Epidemiologic Survey on Alcohol and Related Conditions follow-up
Disorder persistenceBaseline insurance groupaBivariate and propensity-based adjusted models comparing three insurance groups
PrivateNo insuranceMedicaidNo insurance versus privatebMedicaid versus privatebMedicaid versus no insuranceb
N%N%N%ORp99% CIAORcp99% CIORp99% CIAORcp99% CIORp99% CIAORcp99% CI
Any mood, anxiety, or substance use disorder3,36658.51,36867.844871.41.47<.0011.21–1.801.27.0051.02–1.591.71<.0011.30–2.261.23.112.87–1.741.16.201.85–1.59.96.741.67–1.37
Any lifetime mood disorder56933.326536.214650.01.13.280.84–1.54.89.365.64–1.241.91<.0011.24–2.932.66<.0011.43–4.921.68.0041.06–2.661.09.692.63–1.87
Any lifetime anxiety disorder72832.927137.713656.01.22.071.91–1.641.16.231.84–1.592.58<.0011.62–4.112.66<.0011.45–4.882.11<.0011.27–3.511.74.010.95–3.20
Any lifetime substance use disorder2,08060.790969.125068.21.44<.0011.15–1.801.35.0011.06–1.711.35.043.92–1.981.00.994.63–1.59.94.669.63–1.40.83.324.50–1.37
a
Groups were mutually exclusive.
b
The reference group is the latter category of the pair.
c
Adjusted odds ratio, adjusted for propensity score deciles. Twelve pairs of propensity scores were generated for AORs. All propensity scores were calculated with the covariates included in Table 1.
Table 4 displays the results of the multivariable logistic regression analyses examining the associations between insurance status and incident mental disorders. Respondents without insurance had significantly increased odds of developing an incident substance use disorder in both unadjusted and adjusted models compared with those with private insurance. In comparison with people without insurance and with private insurance at baseline, those with Medicaid had increased odds of many of the outcomes in unadjusted models. However, these associations became nonsignificant when we controlled for baseline factors in the propensity score–adjusted models.
TABLE 4. Baseline insurance status and risk of incident mood, anxiety, or substance use disorders during the National Epidemiologic Survey on Alcohol and Related Conditions three-year follow-up period
Incident disorderBaseline insurance groupaBivariate and propensity-based adjusted models comparing three insurance groups
PrivateNo insuranceMedicaidNo insurance versus privatebMedicaid versus privatebMedicaid versus no insuranceb
N%N%N%ORp99% CIAORcp99% CIORp99% CIAORcp99% CIORp99% CIAORcp99% CI
Any incident mood, anxiety, or substance use disorder1,79821.181026.620031.71.34<.0011.14–1.001.17.023.98–1.401.75<.0011.17–2.641.28.294.69–2.351.30.123.83–2.051.21.367.69–2.11
Any incident lifetime mood disorder1,1147.2249910.117017.21.43<.0011.16–1.761.13.179.89–1.422.69<.0011.99–3.631.45.015.98–2.161.88<.0011.32–2.671.41.042.91–2.18
Any incident lifetime anxiety disorder1,5869.9262212.420417.21.26.0011.05–1.511.13.123.92–1.381.84<.0011.38–2.461.26.128.85–1.891.46.0021.08–1.991.05.726.71–1.57
Any incident lifetime substance use disorder1,18111.358315.812717.01.46<.0011.22–1.751.24.0081.01–1.541.61.0011.11–2.331.40.098.82–2.391.10.538.73–1.661.21.307.74–1.98
a
Groups were mutually exclusive.
b
The reference group is the latter category of the pair.
c
Adjusted odds ratio, adjusted for propensity score deciles. Twelve pairs of propensity scores were generated for AORs. All propensity scores were calculated with the covariates included in Table 1.
In a supplementary analysis, we reanalyzed the models, replacing census division with a state variable, and we tested for effect modification of state by insurance status on the outcomes. We did not find evidence of effect modification by state (data available on request).

Discussion

There are four key findings that are important to consider in the context of the controversy over the ACA. First, our findings did not support the notion that Medicaid was associated with poorer mental health status at baseline or poorer mental health outcomes three years later compared with no insurance. Second, we found that adults with Medicaid, compared with privately insured adults, had a higher likelihood of a persistent mood or anxiety disorder. Third, uninsured adults had better mental health status at baseline compared with privately insured adults. Finally, uninsured adults were more likely than privately insured adults to have persistence of baseline mood, anxiety, or substance use disorders during follow-up.
There has been substantial concern that Medicaid, as currently financed and delivered, may not be associated with better access to mental health care than no insurance. In our study, compared with having no insurance, having Medicaid at baseline was associated in bivariate associations with increased odds of mental disorders, but these associations became nonsignificant after adjustment for a comprehensive list of confounding variables. Thus the finding of worse mental health status of Medicaid recipients in prior studies may be due to residual confounding unaddressed in analyses. On the other hand, we did not find that Medicaid was associated with better outcomes than no insurance at follow-up. This is in contrast to the findings of the Oregon experiment in which adults receiving coverage under expansions in Medicaid had lower rates of depressive symptoms (27). Furthermore, the finding that the Medicaid group had a higher likelihood of persistent mood and anxiety disorders than the privately insured group at follow-up may reflect limited access and availability of evidence-based treatments among Medicaid enrollees. A recent study found that psychiatrists were less likely than other physicians to accept all types of insurance, including Medicaid (28), which may especially affect the availability of appropriate services among Medicaid enrollees, who are unlikely to be able to pay out of pocket. Clearly, provision of insurance may be only one of several steps needed to improve access and successful treatment of psychiatric and addictive disorders.
Recent work has demonstrated a lack of infrastructure of substance use facilities for Medicaid enrollees in the United States, especially in counties with high percentages of uninsured adults, racial-ethnic minorities, and people living in rural areas (29). Expansion of infrastructure for residential treatment for substance use disorders, in addition to Medicaid expansion, may be necessary to improve outcomes for people with addictive disorders (29). Finally, there are between-state differences in Medicaid coverage (30). Although we did not detect any differences in the association of insurance with incidence and persistence of mental disorders across states, these analyses were limited because of small sample sizes from each state.
The findings for the uninsured group have potential policy implications. Contrary to previous cross-sectional studies that found a higher prevalence of mental disorders among uninsured compared with privately insured adults (11), we found that, after we adjusted for potential confounding factors, uninsured people had better mental health than the privately insured. People in good mental health may opt to not acquire insurance because of a lack of concern for potential health problems in the future. Our longitudinal findings of a higher incidence and persistence of mental health problems of the uninsured compared with the privately insured are novel. There has been little if any prior assessment of whether mental disorders are associated with an increased likelihood of becoming uninsured or whether having no insurance affects the occurrence of mental disorders. Our findings support the latter explanation. With a lack of detailed information about the use of mental health services during the follow-up period, we can only speculate regarding the mechanisms that may explain our findings. One possibility is that individuals without insurance are less likely than those with private insurance to receive mental health services that would effectively treat their disorders. Alternatively, having a mental disorder may prevent uninsured adults from acquiring health insurance because of a preexisting condition.
Our study should be considered in light of the following limitations. First, insurance status may change over time. The NESARC did not collect information regarding the temporal relationship between changes in insurance status and mental disorders. Nonetheless, we were able to examine whether baseline insurance status was associated with risk of poor mental health outcomes. Second, although we were able to take into account a large number of potentially confounding characteristics, residual confounding cannot be ruled out. Third, the observational nature of the study limits causal inference. Future studies will need to examine the impact of Medicaid expansion on mental health outcomes. Fourth, the data used from this survey are nearly ten years old. However, the NESARC remains the largest and most comprehensive prospective U.S. mental health survey. In addition, data gathered prior to the introduction of the ACA provide a useful baseline for future research. Fifth, the measurement of psychotic disorders was limited to cross-sectional assessment. Thus we could not examine psychotic disorders as an outcome but adjusted for them in the analysis. Sixth, the NESARC did not assess quality of services. Finally, we did not examine insurance status in relation to specific mental disorders or mental disorder severity, with the goal being to limit the number of comparisons and because of a lack of theoretical rationale for differential impact across disorders.

Conclusions

This study provided novel information on the relationship between insurance status and the risk of persistence and incidence of mental disorders. Compared with being uninsured, Medicaid insurance status was not associated with worse mental health status at baseline or worse mental health outcomes at follow-up. Privately insured U.S. adults had better mental health outcomes than uninsured and Medicaid-insured adults for several outcomes. Improving access to health insurance may have an important impact on reducing the development of substance use disorders and persistence of common mental disorders.

References

1.
Sommers BD, Baicker K, Epstein AM: Mortality and access to care among adults after state Medicaid expansions. New England Journal of Medicine 367:1025–1034, 2012
2.
Frakt A, Carroll AE, Pollack HA, et al: Our flawed but beneficial Medicaid program. New England Journal of Medicine 364:e31, 2011
3.
Status of State Action on Medicaid Expansion Decision: 2014. Menlo Park, Calif, Kaiser Family Foundation, 2014. Available at kff.org/health-reform/state-indicator/state-activity-around-expanding-medicaid-under-the-affordable-care-act
4.
Gaglia MA Jr, Torguson R, Xue Z, et al: Effect of insurance type on adverse cardiac events after percutaneous coronary intervention. American Journal of Cardiology 107:675–680, 2011
5.
LaPar DJ, Bhamidipati CM, Mery CM, et al: Primary payer status affects mortality for major surgical operations. Annals of Surgery 252:544–551, 2010
6.
Gottlieb S: Medicaid is worse than no coverage at all. Wall Street Journal, March 10, 2011
7.
Bhattacharya J, Goldman D, Sood N: The link between public and private insurance and HIV-related mortality. Journal of Health Economics 22:1105–1122, 2003
8.
Smolen JR, Thorpe RJ Jr, Bowie JV, et al: Health insurance and chronic conditions in low-income urban whites. Journal of Urban Health 91:637–647, 2014
9.
Garfield RL, Zuvekas SH, Lave JR, et al: The impact of national health care reform on adults with severe mental disorders. American Journal of Psychiatry 168:486–494, 2011
10.
Norquist G, Wells K: Mental health needs of the uninsured. Archives of General Psychiatry 48:475–478, 1991
11.
Wells KB, Sherbourne CD, Sturm R, et al: Alcohol, drug abuse, and mental health care for uninsured and insured adults. Health Services Research 37:1055–1066, 2002
12.
Zatzick D, Jurkovich GJ, Rivara FP, et al: A national US study of posttraumatic stress disorder, depression, and work and functional outcomes after hospitalization for traumatic injury. Annals of Surgery 248:429–437, 2008
13.
Beck CT, Gable RK, Sakala C, et al: Posttraumatic stress disorder in new mothers: results from a two-stage US national survey. Birth 38:216–227, 2011
14.
Baicker K, Finkelstein AN: Effects of Medicaid on clinical outcomes [commentary]. New England Journal of Medicine 369:583, 2013
15.
Hay JW: Effects of Medicaid on clinical outcomes [commentary]. New England Journal of Medicine 369:581–582, 2013
16.
Ruan WJ, Goldstein RB, Chou SP, et al: The Alcohol Use Disorder and Associated Disabilities Interview Schedule–IV (AUDADIS-IV): reliability of new psychiatric diagnostic modules and risk factors in a general population sample. Drug and Alcohol Dependence 92:27–36, 2008
17.
Grant BF, Kaplan K, Shepard J, et al: Source and accuracy statement for wave 1 of the 2001–2002 National Epidemiologic Survey on Alcohol and Related Conditions. Rockville, Md, National Institute on Alcohol Abuse and Alcoholism, 2003
18.
Grant BF, Kaplan KD: Source and accuracy statement for the wave 2 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC). Rockville, Md, National Institute on Alcohol Abuse and Alcoholism, 2005
19.
Hasin D, Carpenter KM, McCloud S, et al: The Alcohol Use Disorder and Associated Disabilities Interview Schedule (AUDADIS): reliability of alcohol and drug modules in a clinical sample. Drug and Alcohol Dependence 44:133–141, 1997
20.
Crum RM, La Flair L, Storr CL, et al: Reports of drinking to self-medicate anxiety symptoms: longitudinal assessment for subgroups of individuals with alcohol dependence. Depression and Anxiety 30:174–183, 2013
21.
Sareen J, Henriksen CA, Stein MB, et al: Common mental disorder diagnosis and need for treatment are not the same: findings from a population-based longitudinal survey. Psychological Medicine 43:1941–1951, 2013
22.
McMillan KA, Enns MW, Cox BJ, et al: Comorbidity of axis I and II mental disorders with schizophrenia and psychotic disorders: findings from the National Epidemiologic Survey on Alcohol and Related Conditions. Canadian Journal of Psychiatry 54:477–486, 2009
23.
El-Gabalawy R, Katz LY, Sareen J: Comorbidity and associated severity of borderline personality disorder and physical health conditions in a nationally representative sample. Psychosomatic Medicine 72:641–647, 2010
24.
Ware J Jr, Kosinski M, Keller SD: A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Medical Care 34:220–233, 1996
25.
Crum RM, Mojtabai R, Lazareck S, et al: A prospective assessment of reports of drinking to self-medicate mood symptoms with the incidence and persistence of alcohol dependence. JAMA Psychiatry 70:718–726, 2013
26.
Software for Survey Data Analysis (SUDAAN): Release 9.0.1. Research Triangle Park, NC, Research Triangle Institute, 2005
27.
Baicker K, Taubman SL, Allen HL, et al: The Oregon experiment: effects of Medicaid on clinical outcomes. New England Journal of Medicine 368:1713–1722, 2013
28.
Bishop TF, Press MJ, Keyhani S, et al: Acceptance of insurance by psychiatrists and the implications for access to mental health care. JAMA Psychiatry 71:176–181, 2014
29.
Cummings JR, Wen H, Ko M, et al: Race/ethnicity and geographic access to Medicaid substance use disorder treatment facilities in the United States. JAMA Psychiatry 71:190–196, 2014
30.
Blumenthal D, Collins SR: Health care coverage under the Affordable Care Act: a progress report. New England Journal of Medicine 371:275–281, 2014

Information & Authors

Information

Published In

Go to Psychiatric Services
Go to Psychiatric Services

Cover: Landscape, by Louis Comfort Tiffany, ca. 1892. Stained glass. Private Collection. Photo credit: Art Resource, New York City.

Psychiatric Services
Pages: 62 - 70
PubMed: 26567928

History

Received: 15 July 2014
Revision received: 2 December 2014
Revision received: 22 March 2015
Accepted: 11 May 2015
Published online: 16 November 2015
Published in print: January 01, 2016

Authors

Details

Jitender Sareen, M.D.
Dr. Sareen and Dr. Bolton are with the Department of Psychiatry, Ms. Wang and Ms. Henriksen are with the Department of Psychology, Dr. Mota is with the Department of Clinical Health Psychology, and Dr. Lix and Dr. Afifi are with the Department of Community Health Sciences, all at the University of Manitoba, Winnipeg, Manitoba, Canada. Dr. Mojtabai is with the Department of Mental Health, Johns Hopkins School of Public Health, Dr. Bienvenu is with the Department of Psychiatry and Behavioral Sciences, and Dr. Crum is with the Department of Epidemiology, Johns Hopkins University, Baltimore. Send correspondence to Dr. Sareen (e-mail: [email protected]).
Yunqiao Wang, M.A.
Dr. Sareen and Dr. Bolton are with the Department of Psychiatry, Ms. Wang and Ms. Henriksen are with the Department of Psychology, Dr. Mota is with the Department of Clinical Health Psychology, and Dr. Lix and Dr. Afifi are with the Department of Community Health Sciences, all at the University of Manitoba, Winnipeg, Manitoba, Canada. Dr. Mojtabai is with the Department of Mental Health, Johns Hopkins School of Public Health, Dr. Bienvenu is with the Department of Psychiatry and Behavioral Sciences, and Dr. Crum is with the Department of Epidemiology, Johns Hopkins University, Baltimore. Send correspondence to Dr. Sareen (e-mail: [email protected]).
Natalie Mota, Ph.D.
Dr. Sareen and Dr. Bolton are with the Department of Psychiatry, Ms. Wang and Ms. Henriksen are with the Department of Psychology, Dr. Mota is with the Department of Clinical Health Psychology, and Dr. Lix and Dr. Afifi are with the Department of Community Health Sciences, all at the University of Manitoba, Winnipeg, Manitoba, Canada. Dr. Mojtabai is with the Department of Mental Health, Johns Hopkins School of Public Health, Dr. Bienvenu is with the Department of Psychiatry and Behavioral Sciences, and Dr. Crum is with the Department of Epidemiology, Johns Hopkins University, Baltimore. Send correspondence to Dr. Sareen (e-mail: [email protected]).
Christine A. Henriksen, M.A.
Dr. Sareen and Dr. Bolton are with the Department of Psychiatry, Ms. Wang and Ms. Henriksen are with the Department of Psychology, Dr. Mota is with the Department of Clinical Health Psychology, and Dr. Lix and Dr. Afifi are with the Department of Community Health Sciences, all at the University of Manitoba, Winnipeg, Manitoba, Canada. Dr. Mojtabai is with the Department of Mental Health, Johns Hopkins School of Public Health, Dr. Bienvenu is with the Department of Psychiatry and Behavioral Sciences, and Dr. Crum is with the Department of Epidemiology, Johns Hopkins University, Baltimore. Send correspondence to Dr. Sareen (e-mail: [email protected]).
James Bolton, M.D.
Dr. Sareen and Dr. Bolton are with the Department of Psychiatry, Ms. Wang and Ms. Henriksen are with the Department of Psychology, Dr. Mota is with the Department of Clinical Health Psychology, and Dr. Lix and Dr. Afifi are with the Department of Community Health Sciences, all at the University of Manitoba, Winnipeg, Manitoba, Canada. Dr. Mojtabai is with the Department of Mental Health, Johns Hopkins School of Public Health, Dr. Bienvenu is with the Department of Psychiatry and Behavioral Sciences, and Dr. Crum is with the Department of Epidemiology, Johns Hopkins University, Baltimore. Send correspondence to Dr. Sareen (e-mail: [email protected]).
Lisa M. Lix, Ph.D.
Dr. Sareen and Dr. Bolton are with the Department of Psychiatry, Ms. Wang and Ms. Henriksen are with the Department of Psychology, Dr. Mota is with the Department of Clinical Health Psychology, and Dr. Lix and Dr. Afifi are with the Department of Community Health Sciences, all at the University of Manitoba, Winnipeg, Manitoba, Canada. Dr. Mojtabai is with the Department of Mental Health, Johns Hopkins School of Public Health, Dr. Bienvenu is with the Department of Psychiatry and Behavioral Sciences, and Dr. Crum is with the Department of Epidemiology, Johns Hopkins University, Baltimore. Send correspondence to Dr. Sareen (e-mail: [email protected]).
Ramin Mojtabai, M.D., Ph.D.
Dr. Sareen and Dr. Bolton are with the Department of Psychiatry, Ms. Wang and Ms. Henriksen are with the Department of Psychology, Dr. Mota is with the Department of Clinical Health Psychology, and Dr. Lix and Dr. Afifi are with the Department of Community Health Sciences, all at the University of Manitoba, Winnipeg, Manitoba, Canada. Dr. Mojtabai is with the Department of Mental Health, Johns Hopkins School of Public Health, Dr. Bienvenu is with the Department of Psychiatry and Behavioral Sciences, and Dr. Crum is with the Department of Epidemiology, Johns Hopkins University, Baltimore. Send correspondence to Dr. Sareen (e-mail: [email protected]).
O. Joseph Bienvenu, M.D., Ph.D.
Dr. Sareen and Dr. Bolton are with the Department of Psychiatry, Ms. Wang and Ms. Henriksen are with the Department of Psychology, Dr. Mota is with the Department of Clinical Health Psychology, and Dr. Lix and Dr. Afifi are with the Department of Community Health Sciences, all at the University of Manitoba, Winnipeg, Manitoba, Canada. Dr. Mojtabai is with the Department of Mental Health, Johns Hopkins School of Public Health, Dr. Bienvenu is with the Department of Psychiatry and Behavioral Sciences, and Dr. Crum is with the Department of Epidemiology, Johns Hopkins University, Baltimore. Send correspondence to Dr. Sareen (e-mail: [email protected]).
Rosa M. Crum, M.D., M.H.S.
Dr. Sareen and Dr. Bolton are with the Department of Psychiatry, Ms. Wang and Ms. Henriksen are with the Department of Psychology, Dr. Mota is with the Department of Clinical Health Psychology, and Dr. Lix and Dr. Afifi are with the Department of Community Health Sciences, all at the University of Manitoba, Winnipeg, Manitoba, Canada. Dr. Mojtabai is with the Department of Mental Health, Johns Hopkins School of Public Health, Dr. Bienvenu is with the Department of Psychiatry and Behavioral Sciences, and Dr. Crum is with the Department of Epidemiology, Johns Hopkins University, Baltimore. Send correspondence to Dr. Sareen (e-mail: [email protected]).
Tracie O. Afifi, Ph.D.
Dr. Sareen and Dr. Bolton are with the Department of Psychiatry, Ms. Wang and Ms. Henriksen are with the Department of Psychology, Dr. Mota is with the Department of Clinical Health Psychology, and Dr. Lix and Dr. Afifi are with the Department of Community Health Sciences, all at the University of Manitoba, Winnipeg, Manitoba, Canada. Dr. Mojtabai is with the Department of Mental Health, Johns Hopkins School of Public Health, Dr. Bienvenu is with the Department of Psychiatry and Behavioral Sciences, and Dr. Crum is with the Department of Epidemiology, Johns Hopkins University, Baltimore. Send correspondence to Dr. Sareen (e-mail: [email protected]).

Competing Interests

Dr. Mojtabai reports receiving consulting fees from Lundbeck Pharmaceuticals. The other authors report no financial relationships with commercial interests.

Funding Information

Preparation of this article was supported by operating grant 273657 from the Canadian Institutes of Health Research, by a Manitoba Health Research Council chair award (Dr. Sareen), and by grant AA016346 from the National Institute on Alcohol Abuse and Alcoholism (Dr. Crum).

Metrics & Citations

Metrics

Citations

Export Citations

If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

For more information or tips please see 'Downloading to a citation manager' in the Help menu.

Format
Citation style
Style
Copy to clipboard

View Options

View options

PDF/EPUB

View PDF/EPUB

Login options

Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.

Personal login Institutional Login Open Athens login
Purchase Options

Purchase this article to access the full text.

PPV Articles - Psychiatric Services

PPV Articles - Psychiatric Services

Not a subscriber?

Subscribe Now / Learn More

PsychiatryOnline subscription options offer access to the DSM-5-TR® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.

Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).

Media

Figures

Other

Tables

Share

Share

Share article link

Share