Prevalence of Suicidal Ideation,Suicide Planning,and Suicide Attempt and Type of Health Insurance Coverage

Abstract

The aim of this study was to examine prevalence of suicidal ideation, suicide planning and suicide attempt and type of health insurance, to identify potential behavioural health needs of populations covered by different health plans and issuers of health insurance. For each health insurance and suicide category, states were assigned a numeric rank, based on descending percentages of coverage and suicide-related concerns. Using these rankings, correlations between the prevalence of suicide outcomes and the type of health insurance were explored using Kendall’s tau-b. There were negative correlations between employer-based insurance and adolescent suicide outcomes. There were positive correlations between direct purchase insurance and adult suicide planning, TRICARE insurance and adolescent suicidal ideation and suicide attempt, VA Care and adolescent and adult suicide outcomes, and being uninsured and adolescent suicidal ideation and suicide attempt. There were no correlations between Medicaid and Medicare and adolescent and adult suicide outcomes. Taken together, individuals with public health coverage and uninsured individuals appear to experience a higher prevalence of mental health concerns, which may be attributable in part to geographic region and socioeconomic/sociodemographic characteristics. Prevention initiatives should incorporate complementary resources from multiple sectors to address both behavioural health and non-health needs of publicly insured and uninsured individuals.

Introduction

In the USA, suicide represents a top 10 leading cause of death among individuals aged 5–64 (National Center for Health Statistics, 2023). Evidence suggests that many suicide decedents had a diagnosable mental disorder, with mental health problems and a previous suicide attempt representing major risk factors for suicide (Arsenault-Lapierre et al., 2004; Institute of Medicine [IOM], 2002). Centers for Disease Control and Prevention (CDC) surveillance evidence suggests that among suicide decedents, a mental health problem was the most frequently identified circumstance preceding death, with 48.3% having received a mental health diagnosis at time of death and 33% experiencing depressed mood at time of death (Wilson et al., 2022). However, CDC surveillance evidence suggests that among suicide decedents, only 25.4% were receiving mental health treatment at time of death (Wilson et al., 2022). Psychological autopsies further suggest that among suicide decedents, only 6%–14% of individuals with depressive symptoms were adequately treated, and only 8%–17% of individuals received psychiatric medications (IOM, 2002).

Evidence also suggests that substance use disorders follow mood disorders as the most common risk factor for suicide, particularly among youth and emerging adults (IOM, 2002; US Department of Health and Human Services [HHS], Office of the Surgeon General [OSG], and National Action Alliance for Suicide Prevention [NAASP], 2012). Substance use disorder is associated with suicidal ideation, suicide attempt, and death by suicide (Poorolajal et al., 2016). Alcohol use disorder in particular is correlated with population-level suicide rates, in part, due to the disinhibiting effects of alcohol (IOM, 2002). CDC surveillance evidence suggests that among suicide decedents, 19.2% experienced alcohol use problems at time of death and 17.4% experienced other substance use problems at time of death (Wilson et al., 2022). However, substance use disorder is frequently under-treated in suicidal individuals (IOM, 2002).

Approximately 90% of suicides are associated with mental health problems and/or alcohol and substance use disorders (IOM, 2002). A co-occurring mental health problem and substance use disorder increase risk for suicide more than each of these disorders separately (HHS, OSG, & NAASP, 2012; IOM, 2002; Wilson et al., 2022). Evidence suggests that about half of individuals who attempt suicide experience co-occurring mood disorder and substance use disorder (IOM, 2002).

Pharmacotherapy and psychotherapy reduce the risk of suicide, although there are well-established barriers to mental health and substance use treatment (e.g., lack of adequate health insurance; IOM, 2002). Fragmented organisation of mental health services and cost of care represent commonly reported barriers, particularly among individuals with co-occurring mental health and substance use problems (IOM, 2002; Sterling et al., 2010). Economic analyses and national surveys suggest mental health service use is sensitive to price, with use decreasing as costs rise (e.g., high co-payments), and use increasing with more comprehensive health insurance coverage (IOM, 2002). Health insurance that provides benefits for integrated mental health and substance use services is important in improving access to cost-effective coordinated treatment (Clark et al., 2008; IOM, 2002; Mojtabai et al., 2014). Integrated services are important in improving the reach and effect of complementary suicide prevention initiatives (HHS, OSG, & NAASP, 2012).

To address access and cost, the Mental Health Parity and Addiction Equity Act (MHPAEA) of 2008 states that if large group health plans or health insurance issuers include mental health or substance use disorder benefits, the financial requirements and treatment limits that apply to those benefits must not be more restrictive than substantially all financial requirements and treatment limits that apply to medical/surgical benefits (Centers for Medicare & Medicaid Services [CMS], 2023). The MHPAEA applies to non-federal governmental plans with more than 50 employees, group plans of private employers with more than 50 employees, and health insurance coverage in the individual health insurance market (CMS, 2023). The MHPAEA does not directly apply to small group health plans, although its requirements are indirectly applied through the Affordable Care Act’s essential health benefit statutes (CMS, 2023). Similarly, while Medicaid and the Children’s Health Insurance Program (CHIP) are not group plans or issuers of health insurance, the Social Security Act requires compliance with certain parts of the MHPAEA (CMS, 2023). While Medicare is also not a group plan or issuer of health insurance, the Medicare Improvements for Patients and Providers Act of 2008 lowered cost-sharing for outpatient mental health services, although MHPAEA rules do not apply to fee-for-service Medicare or Medicare Advantage (Kaiser Family Foundation [KFF], 2022).

Further, MHPAEA requirements do not apply to self-insured non-federal governmental plans with 50 or fewer employees, self-insured small private employers with 50 or fewer employees, group health plans and health insurance issuers exempt due to increased costs of at least 2% in year one or 1% in any subsequent year, and large self-funded non-federal governmental employers that opt out of MHPAEA requirements (CMS, 2023). Importantly, the MHPAEA does not require coverage of mental health and substance use disorder benefits, and the law applies only to plans and issuers that voluntarily include these benefits (CMS, 2023).

Study Purpose

The purpose of this study is to conduct a state-level evaluation to examine prevalence of suicide ideation, planning and attempt, and type of health insurance coverage. Examining prevalence of suicide outcomes provides opportunities to identify potential behavioural health needs of populations covered by different health plans and issuers of health insurance. Mental health and substance use problems increase risk for suicide, particularly if these disorders co-occur, and affordable access to mental health and substance use disorder services is important in reducing suicide risk. Identifying health plans and issuers of health insurance that provide coverage to large proportions of the population with mental health and/or substance use concerns may provide opportunities for issuers to increasingly integrate behavioural health benefits, which may help address suicide risk among those they insure.

Methods

State-level data on suicide ideation, planning, and attempt among adolescents were derived from 2019 Youth Risk Behaviour Surveillance System (YRBSS) surveys. State-level data on suicide ideation, planning, and attempt among adults aged 18 or older were derived from 2018–2019 National Survey on Drug Use and Health (NSDUH) surveys. State-level data on health insurance coverage were derived from the 2019 American Community Survey (ACS).

American Community Survey

The ACS is an ongoing survey conducted by the US Census Bureau (2022b). The ACS utilises a series of monthly samples to produce annual estimates for the same census tracts and block groups previously surveyed during the decennial census long-form sample (US Census Bureau, 2022b). Initial 5-year estimates were published in 2010, and new small-area data are now produced annually (US Census Bureau, 2022b). The US Census Bureau (2022b) also publishes 1-year data for larger geographic regions with populations greater than 65,000, and 1-year supplemental estimates for regions with populations of 20,000 or more.

The ACS includes persons living in housing units and group quarters, and it collects data on population, housing, and socioeconomic characteristics (US Census Bureau, 2022b). The US Census Bureau (2022b) uses a ratio estimation procedure to assign a set of weights to each sample person record, household and group quarters persons, and each sample housing unit record. To bring sample characteristics into congruence with those of the full population, weights compensate for differences in sampling rates across regions, differences between the full and interviewed samples, and differences between the sample and independent estimates of demographic characteristics (US Census Bureau, 2022b). This method corrects for over- or under-coverage by geography and demographic detail, producing estimates that are consistent with population estimates for each county (US Census Bureau, 2022b).

Youth Risk Behavior Surveillance System

The YRBSS includes ongoing biennial school-based national, state, territorial, tribal, and large urban school district surveys conducted in collaboration with the CDC (Brener et al., 2013). School-based surveys monitor health risk-behaviours and health experiences among adolescents (Brener et al., 2013).

Each state, territorial, tribal, and large urban school district survey utilise a two-stage cluster sample design to produce representative samples of US high school students in each jurisdiction (Brener et al., 2013). Surveys are weighted if they contain a sample selected according to CDC protocols, and if they contain an overall response rate of 60% or greater (Brener et al., 2013). To create representative samples in each jurisdiction, a weight is applied to each student record to adjust for non-response and distribution of students by grade, sex, and race/ethnicity in each jurisdiction (Brener et al., 2013).

The national YRBSS utilises a three-stage cluster sample design to produce nationally representative samples of US high school students (Brener et al., 2013). A weight based on student sex, race/ethnicity, and grade is applied to each student record to adjust for non-response and oversampling of Black and Hispanic students (Brener et al., 2013). Final weights are scaled to ensure weighted estimates are congruent with population projections for each survey year (Brener et al., 2013).

National Survey on Drug Use and Health

The NSDUH is an ongoing annual survey conducted by the Substance Abuse and Mental Health Services Administration (SAMHSA, 2023). The NSDUH monitors substance use disorders, substance use treatment, mental health concerns, and use of mental health services (SAMHSA, 2023). The NSDUH design includes 50% overlap in area segments within each consecutive 2-year period to produce estimates regarding year-to-year trends (SAMHSA, 2021). A two-stage interview sample procedure is utilised, and the first interview stage includes a large number of screener dwelling units (SAMHSA, 2021). The second interview stage includes the selection of zero, one, or two people from each screener dwelling unit, with fixed sampling rates for the five designated age categories in each state (SAMHSA, 2021). The two-stage design permits estimation of calibration controls for post-stratification of household-level weights and person pair-level weights (SAMHSA, 2021).

In estimation for pair domains, the NSDUH considers multiplicities and extreme weights (SAMHSA, 2021). Pair weights are divided by the person-level multiplicity factors for each category of interest, with multiplicity factors produced for 12 categories (SAMHSA, 2021). A modified Hajek method calibrates to controls for the number of people in households belonging to a category, with repeat post-stratification creating distinct bound restrictions on extreme weights (SAMHSA, 2021). Cumulatively, the generalised exponential model method is employed to calibrate dwelling unit and pair-level design weights through numerous phases of adjustment (SAMHSA, 2021).

Statistical Analysis

The US Census Bureau (2022a), Centers for Disease Control and Prevention (CDC) (2020) and SAMHSA (2023) produce weighted data for the ACS, YRBSS and NSDUH, respectively. Publicly available weighted data were used in the current analysis. Statistical tests were conducted using SPSS version 28.

Data from the ACS were organised into a table (see supplementary file). For each individual health insurance category, states were assigned a numeric rank based on descending percentages of coverage. Data from the YRBSS and NSDUH were organised into another table (see supplementary file). For each individual suicide category, states were provided a numeric rank based on descending percentages of population experiencing each suicide-related concern.

Using this ranking method, a state’s ranking in the ACS table was compared to this same state’s ranking in the YRBSS/NSDUH table for each individual health insurance category and each individual suicide category. Kendall’s tau-b was used to explore potential correlations between the type of health insurance and suicide ideation, planning, and attempt. There were 42 correlational analyses conducted across tables. States with equal percentages were assigned a mean numeric rank. Kendall’s tau-b corrects for these ‘tied’ ranked observations. For the YRBSS, states that did not have representative data were excluded. For the YRBSS, pairwise deletion was used when states had missing data for variables of interest. Data were analysed at the ordinal level of measurement. Two-tailed tests were used to assess significance. A 95% confidence interval was used to determine significance.

Results

The three states with the highest percentage of population reporting they had employment-based health insurance were Utah (64.6%), Hawaii (63.9%), Maryland (63.1%, tie) and Massachusetts (63.1%, tie). The three states with the highest percentage of population reporting they had direct purchase health insurance were North Dakota (19%), Montana (18.5%) and South Dakota (17.7%). The three states with the highest percentage of population reporting they had TRICARE health insurance were Alaska (9.8%), Hawaii (7.8%) and Virginia (7.7%). The three states with the highest percentage of population reporting they were covered by Medicaid were New Mexico (33.4%), Louisiana (29.1%), Arkansas (26.6%, tie) and West Virginia (26.6%, tie). The three states with the highest percentage of population reporting they were covered by Medicare were West Virginia (24.3%), Maine (23.7%) and Florida (22.6%). The three states with the highest percentage of population reporting they were covered by VA Care were Alaska (4.2%, tie), Montana (4.2%, tie), West Virginia (3.8%) and South Dakota (3.7%). The three states with the highest percentage of population reporting they were uninsured were Texas (18.4%), Florida (16.3%) and Oklahoma (14.3%).

The three states with the highest percentage of adolescents reporting suicidal ideation were California (26.6%), Alaska (25.3%) and Montana (23.4%). The three states with the highest percentage of adults aged 18 and older reporting suicidal ideation were Utah (6.19%), Alaska (6.11%) and Ohio (6.09%). The three states with the highest percentage of adolescents reporting suicide planning were California (23.7%), Alaska (21.6%) and Montana (19.5%). The three states with the highest percentage of adults aged 18 and older reporting suicide planning were Utah (2.01%), Idaho (1.92%, tie), Indiana (1.92%, tie), Nebraska (1.92%, tie) and Wyoming (1.85%). The three states with the highest percentage of adolescents reporting attempting suicide one or more times were Alaska (19.7%), Louisiana (15.2%) and Rhode Island (14.7%). The three states with the highest percentage of adults aged 18 and older reporting attempting suicide were Utah (0.92%), West Virginia (0.79%) and North Dakota (0.78%).

There were no significant correlations between employer-based health insurance and adult suicide outcomes. However, there were significant negative correlations between employer-based health insurance and adolescent suicidal ideation (tau-b = –0.277; p = .010; CI –0.456 to –0.077), adolescent suicide planning (tau-b = –0.272; p = .012; CI –0.454 to –0.069), and adolescent suicide attempt(s) (tau-b = –0.295; p = .006; CI –0.471 to –0.096).

There were no significant correlations between direct purchase health insurance and adolescent suicide outcomes. However, there was a significant positive correlation between direct purchase health insurance and suicide planning among adults (tau-b = 0.236; p = .017; CI 0.050–0.407).

There were no significant correlations between TRICARE health insurance and adult suicide outcomes. However, there were significant positive correlations between TRICARE health insurance and adolescent suicidal ideation (tau-b = 0.260; p = .016; CI 0.058–0.441) and adolescent suicide attempt(s) (tau-b = 0.377; p < .001; CI 0.187–0.540).

There were no significant correlations between Medicaid health coverage and adolescent and adult suicide outcomes. There were no significant correlations between Medicare health coverage and adolescent and adult suicide outcomes.

There were significant positive correlations between VA Care and adolescent suicidal ideation (tau-b = 0.394; p < .001; CI 0.206–0.554), adolescent suicide planning (tau-b = 0.260; p = .017; CI 0.056–0.444), and adolescent suicide attempt(s) (tau-b = 0.323; p = .003; CI 0.127–0.495). There were significant positive correlations between VA Care and adult suicidal ideation (tau-b = 0.240; p = .016; CI 0.054–0.410), adult suicide planning (tau-b = 0.327; p = .001; CI 0.148–0.486), and adult suicide attempt (tau-b = 0.233; p = .020; CI 0.046–0.404).

There were no significant correlations between being uninsured and adult suicide outcomes. However, there were significant positive correlations between being uninsured and adolescent suicidal ideation (tau-b = 0.313; p = .003; CI 0.116–0.487) and adolescent suicide attempt(s) (tau-b = 0.302; p = .005; CI 0.104–0.477).

Discussion

National surveillance data and historic studies that include national samples frequently utilise scoping ‘private’ and ‘public’ health insurance categories when examining associations with mental health, substance use, and suicide. Consequently, current findings were compared to evidence that utilise scoping ‘private’ and ‘public’ insurance categories.

The ACS (US Census Bureau, 2021) includes the following in the ‘private’ health insurance category: (a) employer-based health insurance through one’s own employment or a relative’s; (b) own employment-based health insurance offered through one’s own employment, where only the policyholder is covered; (c) direct-purchase health insurance, which is purchased directly from a private company or an exchange by an individual or a relative; and (d) the TRICARE military health care programme for active duty and retired members of the uniformed services, their families, and survivors. The ACS (US Census Bureau, 2021) includes the following in the ‘public’ health insurance category: (a) Medicare for individuals aged 65 and older, and for individuals under age 65 with long-term disabilities; (b) Medicaid for individuals with low incomes or a disability; (c) CHIP for low-income children whose parents do not qualify for Medicaid; (d) state-specific plans that provide health insurance for low-income or high-risk uninsured individuals; (e) the Department of Veterans Affairs VA health care programme, which provides medical assistance to eligible veterans; and (f) the Department of Veterans Affairs CHAMPVA medical programme, which pays medical service costs for eligible veterans, veteran’s dependents, and survivors of veterans. Because the Indian Health Service (IHS) provides limited coverage, individuals are considered uninsured if their only health coverage comes from the IHS (US Census Bureau, 2021). Further, individuals without any type of public or private health insurance are considered uninsured (US Census Bureau, 2021).

With respect to private insurance, in the current analysis, there were negative correlations between employer-based health insurance and adolescent suicide ideation, planning, and attempt. This may suggest employer-based coverage includes behavioural health provisions that benefit adolescents with a high prevalence of mental health concerns. Consistent with this notion, CDC surveillance evidence suggests children with private health insurance had a higher prevalence of positive mental health indicators when compared to children with public health insurance (Bitsko et al., 2022). Somewhat consistent with this notion, a national evaluation of employer-based health insurance suggests that from 2007 to 2017, use of in-network child psychotherapy increased, and out-of-network child psychotherapy also increased modestly (Benson & Song, 2020). Psychological interventions that address suicidal ideation are important in reducing suicide attempts and death by suicide (Brown & Jager-Hyman, 2014; IOM, 2002; Winter et al., 2013), and employer-based insurance may improve access, particularly if it is affordable. This national evaluation of employer-based health insurance additionally suggests that from 2007 to 2017, in-network prices and cost-sharing for child psychotherapy declined, although out-of-network prices and cost-sharing for child psychotherapy increased (Benson & Song, 2020).

In the current analysis, there was a positive correlation between direct purchase health insurance and suicide planning among adults. This may suggest that adults covered by direct purchase insurance have a high prevalence of mental health concerns and might benefit from additional behavioural health provisions that address suicide planning and suicidal behaviours. Related to this notion, national survey data suggest the suicide rate among young adults with private health insurance increased between 2009 and 2015 (Han et al., 2018).

In the current analysis, there were positive correlations between TRICARE health insurance and adolescent suicidal ideation and suicide attempt. This may suggest that adolescents with TRICARE coverage have a high prevalence of mental health concerns and might benefit from additional behavioural health provisions that address suicide risk. Related to this notion, an analysis of national data suggests privately insured individuals were more likely to receive a mental health evaluation during emergency department visits and were more likely to be admitted for self-harm, when compared to Medicaid-insured individuals (Marcus et al., 2012). However, establishing continuity of care after a suicide attempt is important (HHS, OSG, & NAASP, 2012), and this national analysis suggests individuals with Medicaid were more likely to receive outpatient mental health services when compared to privately insured individuals (Marcus et al., 2012).

With respect to public health insurance, in the current analysis, there were positive correlations between VA Care and adolescent suicide ideation, planning, and attempt. While only a very small percentage of adolescents are covered by VA Care nationally, military-dependent youth may have a high prevalence of mental health concerns. In the current study, there were also positive correlations between VA Care and adult suicide ideation, planning, and attempt. The Veterans Health Administration (VHA) represents the nation’s largest provider of mental health services (National Academies [NAS], 2018), and adults covered by VA Care may also have a high prevalence of mental health concerns. However, these associations should be interpreted conservatively, as correlations between VA Care and adolescent and adult suicide outcomes may also be associated with geographic region (i.e., rurality) and socioeconomic/sociodemographic characteristics of populations with elevated risk for suicide (i.e., military-dependent youth, veterans).

Because several largely rural states have high proportions of population covered by VA Care, correlations may also be associated with geographic region, as rurality is a well-established risk factor for suicide (Ehlman et al., 2022; IOM, 2002; Ivey-Stephenson et al., 2021, 2022; Kegler et al., 2017; Stone et al., 2021). Sociodemographic characteristics that influence suicide risk among youth are also important to consider, as an analysis of national data suggests youth who are dependents of military personnel, and who are covered by TRICARE, CHAMPUS or CHAMPVA, experience more inpatient hospital admissions for attempted suicide than non-military-dependent youth (Pressley et al., 2012). In this national evaluation of military-dependent youth with military health benefits, odds of a mental health diagnosis at time of admission or during the hospital stay were higher in military-dependent children than in non-military children, and military-dependent children had higher substance and alcohol related diagnoses than non-military-dependent children (Pressley et al., 2012).

Similarly, VA Care covers high proportions of adult population with elevated risk for suicide and correlations may also be associated with socioeconomic/sociodemographic characteristics that influence suicide risk, as veterans who utilise VA Care were more likely than veterans who do not utilise VA Care to identify as non-Hispanic Black (Dursa et al., 2016; Meffert et al., 2019; Nelson et al., 2007) or Hispanic (Nelson et al., 2007), be unmarried (Dursa et al., 2016; Meffert et al., 2019), have lower educational attainment (Meffert et al., 2019; Nelson et al., 2007), have lower incomes (Dursa et al., 2016, Farmer et al., 2016; Meffert et al., 2019; Nelson et al., 2007), reside in rural areas (Farmer et al., 2016), experience deployment (Dursa et al., 2016, Farmer et al., 2016), experience more combat exposure (Meffert et al., 2019), and have limited access to health services from other sectors (Farmer et al., 2016). More generally, the age- and sex-adjusted suicide rate among veterans was 57.3% higher than that of non-veterans in 2020 (US Department of Veterans Affairs, 2022). In every year between 2001 and 2020, age- and sex-adjusted suicide rates were higher and rose faster among veterans than non-veterans (US Department of Veterans Affairs, 2022). A national cross-sectional study suggests increases in the percentage of veterans per county were associated with higher county-level suicide rates (Steelesmith et al., 2019).

However, while age-adjusted suicide rates among veterans not using the VHA increased 38% from 2001 to 2014, age-adjusted suicide rates among veterans using the VHA increased 5% during this same period (Lemle, 2018). Suicide rates among veterans using VHA services increased less than suicide rates among veterans not using VHA services, despite VHA users experiencing more trauma, lifetime drug use disorder, lifetime suicide attempt, current suicidal ideation, current mental health problems, prior mental health treatment (Meffert et al., 2019); lifetime psychopathology (Farmer et al., 2016; Meffert et al., 2019); chronic medical conditions (Dursa et al., 2016; Farmer et al., 2016; Meffert et al., 2019); disability resulting from physical or mental health problems (Meffert et al., 2019; Nelson et al., 2007); and poorer overall health (Farmer et al., 2016; Nelson et al., 2007). Taken together, the Action Alliance reports the VHA’s comprehensive suicide prevention model has been associated with reductions in suicide among high-risk subgroups (HHS, OSG, & NAASP, 2012). Further, the National Academies (2018) report that mental health services provided by the VHA are equal to or of better quality than mental health services provided in private and non-VA public sectors, recommending the VHA’s integrated mental health care service model serve as a guide for the delivery of mental health services across the USA.

With respect to being uninsured, in the current analysis, there were positive correlations between being uninsured and adolescent suicidal ideation and suicide attempt. This may suggest that uninsured adolescents have a high prevalence of mental health concerns and might benefit from behavioural health insurance that addresses suicide risk. As with the public VA Care programme, associations should be interpreted conservatively, as correlations may also be associated with geographic region (i.e., rurality) and socioeconomic/sociodemographic characteristics. Specifically, between 2010 and 2019, uninsured rates in rural areas were consistently higher than in urban areas, and when compared to Medicaid expansion states, rural uninsured rates were almost twice as high in non-Medicaid expansion states (Turrini et al., 2021). Suicide risk among the uninsured may also be associated with socioeconomic/sociodemographic characteristics that influence suicide risk, as uninsured individuals were more likely to identify as American Indian or Alaska Native, Hispanic or Latino, or non-Hispanic Black; have lower incomes; and reside in a state that did not expand Medicaid (Finegold et al., 2021).

While the NHANES and National Health Interview Survey suggest use of mental health services in children with no health insurance did not differ from children with private health insurance, remaining CDC surveillance evidence suggests use of mental health services was lower among children with no health insurance than among children with health insurance (Bitsko et al., 2022). Similarly, national longitudinal survey data suggest uninsured adolescents had fewer mental health visits than insured adolescents (Heboyan et al., 2021). National survey data also suggest receipt of mental health care among suicidal young adults without health insurance decreased between 2009 and 2015 (Han et al., 2018).

While there were no correlations between being uninsured and adult suicide outcomes in the current analysis, national survey data suggest adults with moderate mental health problems were more likely to be uninsured than adults with no mental health concerns (Rowan et al., 2013). National ecological and cross-sectional studies suggest states with higher proportions of uninsured individuals have higher state-level suicide rates (Rockett et al., 2022; Tondo et al., 2006). Another national cross-sectional study similarly suggests increases in percentages of individuals without health insurance per county were associated with higher county-level suicide rates (Steelesmith et al., 2019). Somewhat similarly, CDC surveillance evidence suggests adults without health insurance were less likely to receive medical attention for suicide attempts than adults with health insurance (Ivey-Stephenson et al., 2022). Consequently, the Action Alliance (HHS, OSG, & NAASP, 2012) recommends integrating synergistic suicide prevention initiatives across diverse community settings to increase the dissemination of suicide prevention-related resources that reach individuals without access to health care (e.g., uninsured).

In the current study, there were no significant correlations between Medicaid and Medicare and suicide outcomes among youth and adults. While an ecological study of national data suggests more comprehensive Medicaid benefits were associated with decreased suicide rates (Choi et al., 2020), CDC surveillance evidence suggests adults with Medicaid report a higher prevalence of suicide ideation, planning, and attempt than adults in other insurance categories and uninsured adults (Ivey-Stephenson et al., 2022). Additional national survey data suggest prevalence of mental health problems was higher in adults with public insurance than adults in other insurance categories (Rowan et al., 2013). Further, an analysis of national survey data suggests rates of psychiatric-related emergency department visits among adults with Medicaid were two times higher than uninsured adults and eight times higher than privately insured adults (Hazlett et al., 2008). Nationwide emergency department data similarly suggest that more than half of emergency visits for suicidal ideation or suicide attempt were paid by Medicaid or disability-related Medicare for individuals under age 65 (Owens et al., 2020). As with uninsured status and the public VA Care programme, higher prevalence of mental health concerns may be attributable to socioeconomic/sociodemographic characteristics that may disproportionately impact Medicaid-enrolled adults (Ivey-Stephenson et al., 2022). Prevention initiatives should incorporate complementary resources from multiple sectors to address both the behavioural health and non-health needs of Medicaid-enrolled adults (e.g., federal nutrition and housing assistance, employment supports).

While data on health insurance statuses did not differ when examining prevalence of depression among children in the NSDUH and National Health and Nutrition Examination Survey (NHANES), the National Survey of Children’s Health suggests prevalence of depression in children with public insurance was higher than in children with private health insurance or no health insurance (Bitsko et al., 2022). CDC youth surveillance evidence additionally suggests youth with public health insurance had higher prevalence of attention-deficit/hyperactivity disorder, behaviour problems, anxiety, autism spectrum disorder, and illicit drug use disorder (Bitsko et al., 2022). This CDC surveillance evidence also suggests assessment, diagnosis, and treatment of mental disorders were higher among children with public insurance than children with private health insurance (Bitsko et al., 2022). However, national re-admissions data suggest public health insurance was associated with higher odds of unplanned, all-cause re-admission among youth hospitalised for suicidal ideation or suicide attempt (Doupnik et al., 2018). As with uninsured status and the public VA Care programme, higher prevalence of mental health concerns may partly be attributable to socioeconomic/sociodemographic characteristics that may disproportionately impact Medicaid-enrolled youth (e.g., parental income at or below the federal poverty threshold, residing in low-resource neighbourhoods; Bitsko et al., 2022; Cree et al., 2018; Robinson et al., 2017). Prevention initiatives should include appropriate resources from multiple sectors to address both the behavioural health and non-health needs of Medicaid-enrolled youth.

Limitations

The ACS, YRBSS, and NSDUH collect data by self-report, and individuals may over- or under-report health insurance coverage and health behaviours. Further, Minnesota, Oregon, Washington, and Wyoming did not conduct a state YRBSS in 2019; Delaware and Indiana did not have representative data; and data on suicide-related variables were sometimes unavailable within select states (CDC, 2020).

The ACS, YRBSS, and NSDUH are cross-sectional and determining causal relationships is difficult. Similarly, this analysis examined correlation and does not imply causation. Correlations may be associated with a higher prevalence of mental health concerns, although correlations may also be associated with geographic region (i.e., rurality) and socioeconomic/sociodemographic characteristics of populations with elevated risk for suicide (e.g., military-dependent youth, veterans, uninsured).

Conclusion

Comparisons between current findings and historic evidence were sometimes challenging, as historic studies frequently utilised scoping ‘private’ and ‘public’ health insurance categories. Broad categorisations make it challenging to draw comparisons by specific type of health insurance coverage. There is opportunity for future research to examine private health insurance coverage to facilitate the identification of plans and issuers that cover large proportions of populations with high behavioural health needs.

Individuals with public health insurance and uninsured individuals appear to experience higher prevalence of mental health concerns, which may be partly attributable to geographic region and socioeconomic/sociodemographic characteristics. Prevention initiatives should incorporate complementary resources from multiple sectors to address both the behavioural health and non-health needs of publicly insured and uninsured individuals.

Data Availability

The data analysed during the current study were derived from tables that are included in the supplementary file.