Most states have passed insurance mandates requiring health plans to cover services for children with autism spectrum disorder (ASD). Research reveals that these mandates increased treated prevalence, service use, and spending on ASD-related care. As employer-sponsored insurance shifts toward high-deductible health plans (HDHPs), it is important to understand how mandates affect children with ASD in HDHPs relative to traditional, low-deductible plans.
Insurance claims for 2008–2012 for children covered by 3 large US insurers (United Healthcare, Aetna, and Humana) available through the Health Care Cost Institute were used to compare the effects of mandates on ASD-related spending for children in HDHPs and traditional health plans.
Relative to children in traditional plans, mandates were associated with higher average monthly spending increases for children in HDHPs. Mandate-attributable spending differences between children enrolled in HDHPs relative to traditional plans were $77 for ASD-specific services (95% confidence interval [CI]: $10 to $144), $125 for outpatient health services (95% CI: $26 to $223), and $144 for all health services (95% CI: $36 to $253). These spending differentials were driven by differences in plan spending and not out-of-pocket (OOP) spending.
Spending on ASD-related services attributable to autism mandates was higher among children in HDHPs, but higher spending did not translate into a greater OOP burden. For families with consistently high health care expenditures on ASD-related services, high-deductible products may be worth considering in the context of mandate laws. Families in mandate states with children with ASD enrolled in HDHPs were able to increase service use without paying more OOP.
Forty-seven states and the District of Columbia have enacted autism insurance mandates to improve access to health care among commercially insured children with autism spectrum disorder (ASD). These laws increased diagnosed ASD prevalence and, among children with ASD, health care service use and spending.
High-deductible health plans are increasingly common in the United States, and results reveal that autism mandates were associated with higher average monthly spending increases for children in high-deductible health plans relative to children in traditional health plans.
Autism spectrum disorder (ASD) is a neurobehavioral condition involving impaired social communication, restricted interests, and repetitive behaviors. Children with ASD often need behavioral, speech-language, occupational, and physical therapy1 and educational interventions delivered for up to 40 hours a week.2 ASD frequently is accompanied by other conditions requiring treatment, including seizures, hearing impairments, sleep and gastrointestinal problems, and psychiatric comorbidities.3 The costs of care can be substantial,4,5 and families report challenges paying for services.6 Yet, historically, commercial insurers have excluded or provided minimal coverage for ASD services.
In response, 47 states and the District of Columbia enacted insurance mandates to broaden coverage of these services under commercial insurance. These laws require insurers to cover many ASD-related health services, such as diagnostic and assessment services, and behavioral and functional therapies.
Previous research suggests that mandates expanded access to services for ASD under commercial insurance. Mandates led to increases in the prevalence of diagnosed ASD and use of health services among children with ASD.7,8 Mandell et al7 estimated a 10% increase in the prevalence of diagnosed ASD attributable to mandates in the first year after implementation, with increases in subsequent years. Barry et al8 found that mandate implementation was associated with a 3.4–percentage point increase in the probability of using ASD-related services and a $77 increase in average monthly spending on ASD-related services among children with ASD.
Increased spending associated with mandates may depend on the type of insurance product a family has. Enrollment in high-deductible health plans (HDHPs) (plans with high deductibles often paired with tax-advantaged savings accounts) is increasing rapidly. In 2017, 28% of workers with employer-sponsored health insurance were enrolled in HDHPs,9 an increase of 9 percentage points in the last 10 years. As employers shift toward HDHPs, it is important to understand how children with ASD in HDHPs benefit from mandates compared with those in traditional plans.
Although HDHPs have been shown to reduce health spending,10 high deductibles come with higher out-of-pocket (OOP) cost exposure. Proponents of HDHPs argue that these plans encourage enrollees to be better consumers of health care by increasing value-based choices relative to consumers in low-deductible plan options (eg, preferred provider organizations [PPOs] or health maintenance organizations [HMOs]). Critics contend that price-shopping tools available to enrollees in HDHPs do not encourage value-based shopping,11 and HDHP enrollees indiscriminately use less care in the short-term,12 including beneficial services, leading to poor health outcomes. Furthermore, increased cost exposure associated with HDHPs may lead to acute financial strain,13 especially among individuals with high-cost chronic conditions such as ASD.
No research has examined whether the effects of state mandates differ on the basis of the insurance product in which a child with ASD is enrolled. On the one hand, after the mandate, spending among children with ASD in HDHPs may grow faster than spending among similar children in traditional plans. Given the frequent need for services, forward-thinking parents who are relatively savvy consumers might anticipate spending through a high deductible relatively early in the year. To the extent that these families have lower premium contributions from enrolling in an HDHP, they may be able to achieve lower total spending outlays (ie, OOP spending plus premium contribution) by the end of the year relative to a traditional plan. However, if families enrolled in HDHPs are deterred from using services by large deductibles rather than calculating the long-term shadow price of spending through the deductible,12 then the introduction of mandates could have a smaller effect among enrollees in HDHPs relative to traditional plans. If so, families enrolled in HDHPs would not benefit fully from mandates. Understanding how families with different insurance plans react to mandates is crucial given increased HDHP prevalence.
Studies of the effects of HDHPs in the context of other chronic conditions with similar cost characteristics might offer clues about their likely effects for children with ASD. Evidence for chronic conditions with predictably high year-to-year costs might be particularly informative. A number of studies have revealed that HDHPs generate modest reductions in continuation or adherence in prescription drug use, increased emergency department visits, and delayed or forgone care among patients with chronic conditions.14,–19 Thus, previous research suggests that enrollees with chronic conditions fare worse under HDHPs relative to traditional plans.
We used a natural experimental design, taking advantage of geographic and temporal variation in the implementation of state mandates, to compare eligible and ineligible children within states implementing and not implementing mandates. Then, we examined whether there were differential effects of mandates among children enrolled in HDHPs versus traditional health plans.
Methods
Data
In this analysis, we used inpatient, outpatient, and pharmaceutical claims data from the Health Care Cost Institute (HCCI) for 2008–2012. HCCI data include claims from 3 national insurers (United Healthcare, Aetna, and Humana) representing >50 million individuals per year in all 50 states and the District of Columbia. The analytic sample included children 0 to 21 years old with at least 2 claims on different days during the 5-year study with the International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis code for ASD: 299.xx. The unit of analysis was the child-month. The analytic sample included only enrollees with behavioral health claims that were available in the data set (89% of all child-months); if a child’s behavioral health services were managed through a carve-out vendor, the child was excluded from our analysis because we cannot observe their complete claims. We excluded months when children were covered by individually insured plans (<0.2% of all child-months), and we excluded children who switched between HDHPs and traditional plans during the study period. We did not require continuous enrollment. The final study sample included 98 639 unique children and 3 340 349 child-months.
Measures
We examined total (OOP and insurer) monthly spending, OOP monthly spending, and insurer monthly spending within the following 5 categories: (1) ASD-specific outpatient services, (2) ASD-specific outpatient behavioral and functional therapies, (3) ASD-specific (outpatient and inpatient) services, (4) total outpatient health care services, and (5) total health care services and pharmaceuticals. Outpatient ASD-specific services consisted of claims from the outpatient file with a 299.xx diagnosis code. Outpatient ASD-specific behavioral and functional therapies consisted of claims with a 299.xx diagnosis code and a procedure code indicating receipt of an outpatient behavioral, speech, occupational, or physical therapy service. These procedure codes are detailed elsewhere.20 ASD-specific services consisted of inpatient and outpatient services with claims with a 299.xx diagnosis code. Health care services and pharmaceuticals included all inpatient and outpatient services (not limited to those with a 299.xx diagnosis code on the claim) and prescription medications. When constructing these measures, we excluded claims in the top 0.01% of the spending distribution from the analysis to reduce the potential for undue influence of outliers on effect estimates.
We calculated total monthly spending as the sum of the amounts paid by the insurer and enrollee OOP for services during the month. Then, we examined the disaggregated amounts paid by the insurer and the enrollee OOP for each of the 5 categories of services. The total costs of claims spanning multiple months were apportioned on the basis of the number of claim-days in each month. Spending was inflated to 2012 US dollars by using the relevant components of the Personal Health Care Index from the Centers for Medicare and Medicaid Services Office of the Actuary.21
We used information collected by Autism Speaks to identify the month and year that each state implemented a mandate and the age range of children eligible for mandated coverage.22 We then verified this information by reviewing the original statutes (Supplemental Table 3). Four states (IN, IL, SC, and TX) were excluded because mandates were implemented before our study period. During the study period, 2008–2012, 6 states implemented mandates in 2009 (AZ, FL, LA, NM, PA, and WI), 4 states in 2010 (CO, CT, MT, and NJ), 8 states in 2011 (AR, KY, MA, ME, MO, NH, NV, and VT), and 7 states in 2012 (CA, DE, MI, NY, RI, VA, and WV). All mandates applied to fully insured firms with >50 employees; 22 of the 29 mandates also applied to fully insured firms with ≤50 employees.
A binary variable was used to capture whether a child was living in a mandate state, and another binary variable was used to flag whether a child was eligible to be covered by that mandate in a given month. Eligibility for mandated coverage was determined for a specific child in a given month on the basis of whether the child was enrolled in employer-based insurance that was fully insured (versus self-insured) and whether the child met the mandate’s age criteria. Mandates only apply to a subset of a state’s commercially insured population. The Employee Retirement Income Security Act exempts self-insured firms (those contracting with health plans to administer employee health benefits only and not to manage their insurance risk pool) from state insurance regulations. The remainder of fully insured firms are subject to the law. However, in practice, this means that, over the entire study period, approximately half of child-months are for children subject to state mandates in that month.23
Other covariates included sex, age in the month (estimated on the basis of July 1 of birth year), calendar month, calendar year, residence state, and insurance product type. Models included an indicator for prescription drug coverage (42% of children with ASD in our sample did not have pharmacy benefits observable in our data).
Another binary variable was used to identify whether the child was enrolled in an HDHP or traditional plan. We identified enrollees in HDHPs using an indicator provided in the insurance enrollment file. Traditional plan enrollees were all children not enrolled in HDHPs and included enrollees in a variety of employer-sponsored insurance plans, including HMO, PPO, point of service (POS), and indemnity plans and other types of plans (ie, private fee-for-service, special needs, exclusive provider organization [EPO], short-term, and other or unknown).
Statistical Approach
We examined whether the effects of mandates differed among those enrolled in HDHPs versus traditional plans using a difference-in-difference-in-differences approach with state, year, and calendar-month fixed effects (see Supplemental Information and Supplemental Table 4 for details on study design and statistical approach). We compared changes in outcomes within states before and after mandate implementation and between groups of children eligible and ineligible for the mandates. Then, we estimated how these mandate effects differed across children in HDHPs versus those in traditional plans. For states that had not implemented a mandate during the study period, we identified children who would have been mandate eligible (had a mandate been implemented in their state) on the basis of enrollment in a fully insured plan and age between 0 and 21 years, the modal age range covered under mandates (10 states).
The treatment group for the study included children who lived in states with active mandates and were eligible for the mandate. We used 3 comparison groups: (1) children in states with an active mandate who were not subject to the mandate, (2) children in states without a mandate who would have been subject to a mandate if one were active, and (3) children in states without a mandate who would not have been subject to a mandate if one were active. The difference-in-difference-in-differences design was used to account for secular trends in outcomes unrelated to mandates.
We calculated descriptive statistics for children in the sample overall and among those in traditional plans and HDHPs. For each group, we included demographic information on those in mandate and nonmandate states who were eligible and ineligible for mandate-level insurance coverage. To strengthen our confidence in the natural experimental design, we compared unadjusted trends for outcomes in the years before and after implementation of the mandates among enrollees eligible and ineligible for the mandates for HDHPs and traditional plans (Supplemental Figs 2 and 3).
To examine whether the effects of mandates on outcomes differed between children in HDHPs or traditional plans, we estimated adjusted regression models. Because a high proportion of child-month observations had no service use, we analyzed spending using 2-part models.24 The first part was a logistic regression used to predict any service use in the month, and the second part was a generalized linear model used to predict nonzero mean monthly spending with a log link and the best-fitting error distribution, as identified by modified Park tests.25 Adjusted analyses included child-level control variables and state, year, and calendar-month fixed effects. Confidence intervals (CIs) were adjusted to account for the clustering of observations within states. To ease interpretation, we converted results to predictive margins on the dollar scale.26 In a sensitivity analysis, we reestimated models, excluding insurance product type (ie, PPO, HMO, POS, indemnity, or other type) as a covariate because insurance product types differ between HDHP and traditional plan enrollees, so mandate effects by these 2 groups might be masked by controlling for product type; however, results were qualitatively similar (Supplemental Table 5). In addition, we re-estimated the models of all health care spending using only the 58% of child-months with complete pharmacy claims (see Supplemental Table 6).
The study was determined exempt by the Johns Hopkins Bloomberg School of Public Health and University of Pennsylvania Institutional Review Boards.
Results
Table 1 summarizes the characteristics of children with ASD in our study sample in the first month they entered the data set. For example, a child whose first month appearing in the data set occurred before mandate implementation or at an age ineligible for his or her state’s mandate age requirement would be categorized as ineligible. Approximately 80% of children were boys. The mean overall age during the first month in which the child entered the analytic sample was 8 years; 78% were between 0 and 12 years of age. Most children (88%) were enrolled in traditional plans rather than HDHPs. Of the total sample of children with ASD in our study population (N = 98 639), ∼5% were eligible for mandates, with 4479 eligible children in traditional plans and 682 eligible children in HDHPs.
Descriptive Statistics for Children With ASD in the Study Sample, 2008–2012
| Children With ASD, N = 98 639 | Traditional Health Plans, N = 86 691 | HDHPs, N = 11 948 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Children in Nonmandate States, N = 22 States | Children in Mandate States, N = 29 States | Children in Nonmandate States, N = 22 States | Children in Mandate States, N = 29 States | ||||||
| Ineligible, N = 47 763 | Eligible, N = 16 859 | Ineligible, N = 17 600 | Eligible, N = 4479 | Ineligible, N = 6648 | Eligible, N = 1508 | Ineligible, N = 3110 | Eligible, N = 682 | ||
| Male sex, % (n) | 81 (80 221) | 82 (38 935) | 82 (13 763) | 81 (14 228) | 81 (3609) | 81 (5365) | 82 (1232) | 81 (2523) | 83 (567) |
| Age at study entry, mean (SD) | 8.14 (5.03) | 7.97 (5.07) | 8.16 (4.93) | 8.54 (5.07) | 7.47 (4.57) | 8.23 (5.06) | 8.29 (5.05) | 8.90 (5.07) | 8.26 (4.84) |
| Age group, y, % (n) | |||||||||
| 0–5 | 35 (34 518) | 36 (17 376) | 35 (5900) | 31 (5472) | 42 (1860) | 34 (2266) | 34 (520) | 28 (886) | 35 (238) |
| 6–12 | 43 (42 802) | 43 (20 310) | 43 (7305) | 46 (8010) | 42 (1890) | 44 (2896) | 43 (652) | 46 (1445) | 43 (294) |
| 13–17 | 18 (17 490) | 17 (8346) | 18 (3051) | 19 (3263) | 14 (618) | 18 (1215) | 19 (280) | 19 (597) | 18 (120) |
| 18–21 | 4 (3829) | 4 (1731) | 4 (593) | 5 (855) | 3 (111) | 4 (271) | 4 (56) | 6 (182) | Blindeda |
| Type of benefits plan, % (n) | |||||||||
| HMOb | 14 (14 059) | 8.6 (4094) | 41 (6865) | 8 (1468) | 29 (1311) | Blindeda | 9 (130) | Blindeda | 19 (130) |
| POSc | 64 (63 209) | 65 (30 874) | 44 (7411) | 71 (12 500) | 52 (2336) | 89 (5928) | 62 (942) | 90 (2811) | 60 (407) |
| PPOd | 14 (13 939) | 16 (7608) | 13 (2117) | 12 (2137) | 15 (661) | 10 (652) | 28 (427) | 7 (202) | 20 (135) |
| EPO, IND, or OTHe | 8 (7432) | 11 (5187) | 3 (456) | 9 (1495) | 4 (171) | 1 (55) | Blindeda | Blindeda | Blindeda |
| Children With ASD, N = 98 639 | Traditional Health Plans, N = 86 691 | HDHPs, N = 11 948 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Children in Nonmandate States, N = 22 States | Children in Mandate States, N = 29 States | Children in Nonmandate States, N = 22 States | Children in Mandate States, N = 29 States | ||||||
| Ineligible, N = 47 763 | Eligible, N = 16 859 | Ineligible, N = 17 600 | Eligible, N = 4479 | Ineligible, N = 6648 | Eligible, N = 1508 | Ineligible, N = 3110 | Eligible, N = 682 | ||
| Male sex, % (n) | 81 (80 221) | 82 (38 935) | 82 (13 763) | 81 (14 228) | 81 (3609) | 81 (5365) | 82 (1232) | 81 (2523) | 83 (567) |
| Age at study entry, mean (SD) | 8.14 (5.03) | 7.97 (5.07) | 8.16 (4.93) | 8.54 (5.07) | 7.47 (4.57) | 8.23 (5.06) | 8.29 (5.05) | 8.90 (5.07) | 8.26 (4.84) |
| Age group, y, % (n) | |||||||||
| 0–5 | 35 (34 518) | 36 (17 376) | 35 (5900) | 31 (5472) | 42 (1860) | 34 (2266) | 34 (520) | 28 (886) | 35 (238) |
| 6–12 | 43 (42 802) | 43 (20 310) | 43 (7305) | 46 (8010) | 42 (1890) | 44 (2896) | 43 (652) | 46 (1445) | 43 (294) |
| 13–17 | 18 (17 490) | 17 (8346) | 18 (3051) | 19 (3263) | 14 (618) | 18 (1215) | 19 (280) | 19 (597) | 18 (120) |
| 18–21 | 4 (3829) | 4 (1731) | 4 (593) | 5 (855) | 3 (111) | 4 (271) | 4 (56) | 6 (182) | Blindeda |
| Type of benefits plan, % (n) | |||||||||
| HMOb | 14 (14 059) | 8.6 (4094) | 41 (6865) | 8 (1468) | 29 (1311) | Blindeda | 9 (130) | Blindeda | 19 (130) |
| POSc | 64 (63 209) | 65 (30 874) | 44 (7411) | 71 (12 500) | 52 (2336) | 89 (5928) | 62 (942) | 90 (2811) | 60 (407) |
| PPOd | 14 (13 939) | 16 (7608) | 13 (2117) | 12 (2137) | 15 (661) | 10 (652) | 28 (427) | 7 (202) | 20 (135) |
| EPO, IND, or OTHe | 8 (7432) | 11 (5187) | 3 (456) | 9 (1495) | 4 (171) | 1 (55) | Blindeda | Blindeda | Blindeda |
Information for each child with an ASD diagnosis in the study sample is based on their characteristics in the first month they entered the data set. IND, xxx; OTH, xxx.
Numbers are blinded because n < 50 observations.
Members have a wide range of health care services through a network of providers who agree to supply services. Most HMOs have arrangements for using providers that are out of network for an additional cost.
Like an HMO plan, members may be required to designate a primary care physician, who then makes referrals to network specialists when needed. Members may receive care from nonnetwork providers but with greater OOP costs.
Members are allowed to visit whichever in-network physician or health care provider they wish without first requiring a referral from a primary care physician.
This category includes indemnity plans (in which members can visit any provider, and the insurer pays an established portion of charges), EPO plans (in which members can use the health care providers in the EPO network, but there are no out-of-network benefits), and other types of health plan not already described.
Table 2 displays the differences in the estimated effects of mandates on mean monthly spending among traditional and HDHP enrollees. Mandates were associated with significant increases in average monthly spending across all service categories among HDHP enrollees but not among traditional plan enrollees. Mandate effects differed significantly for average monthly spending on ASD-specific outpatient health services, total outpatient health services, and overall health services. Specifically, relative to traditional plan enrollees, mandates were associated with a $77 (95% CI: $10 to $144) greater increase for HDHP enrollees in average monthly spending on ASD-specific services. In addition, relative to traditional plan enrollees, mandates were associated with a $125 (95% CI: $26 to $223) greater increase in average monthly spending on outpatient health services and a $144 (95% CI: $36 to $253) greater increase in average monthly spending on all health services for HDHP enrollees.
Adjusted Estimates of Effects of State Mandates on Mean Spending in 2012 US Dollars at the Month Level Among Children With ASD in Traditional Health Plans and HDHPs, 2008–2012
| Traditional Health Plans | HDHPs | Difference in Mandate Effects for HDHPs Versus Traditional Plans (95% CI) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Mandate Eligible | Mandate Ineligible | Changes in Value Attributable to Autism Mandates (95% CI) | Mandate Eligible | Mandate Ineligible | Changes in Value Attributable to Autism Mandates (95% CI) | ||||||
| Mandate States | Nonmandate States | Mandate States | Nonmandate States | Mandate States | Nonmandate States | Mandate States | Nonmandate States | ||||
| ASD-specific outpatient health care services | 252.76 | 193.73 | 128.43 | 129.22 | 59.81 (−6.31 to 125.93) | 305.64 | 175.75 | 124.17 | 130.79 | 136.51 (43.34 to 229.67)* | 76.70 (9.82 to 143.57)* |
| ASD-specific behavioral and functional therapy services | 170.37 | 129.26 | 77.38 | 77.69 | 41.42 (−6.13 to 88.97) | 202.43 | 117.96 | 74.70 | 77.78 | 87.57 (29.01 to 146.12)* | 46.14 (−5.19 to 97.48) |
| ASD-specific health care services | 290.06 | 228.81 | 155.19 | 158.64 | 64.70 (−3.42 to 132.81) | 338.00 | 213.34 | 154.46 | 160.83 | 131.04 (29.30 to 232.78)* | 66.34 (−8.62 to 141.31) |
| All outpatient health care services | 517.34 | 456.15 | 383.90 | 391.87 | 69.15 (−33.25 to 171.56) | 592.01 | 407.79 | 379.59 | 389.07 | 193.70 (38.73 to 348.67)* | 124.54 (25.98 to 223.11)* |
| All health care services | 669.00 | 606.00 | 519.94 | 530.37 | 73.43 (−32.60 to 179.45) | 749.94 | 534.28 | 515.15 | 517.09 | 217.60 (48.21 to 386.98)* | 144.17 (35.60 to 252.74)* |
| Traditional Health Plans | HDHPs | Difference in Mandate Effects for HDHPs Versus Traditional Plans (95% CI) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Mandate Eligible | Mandate Ineligible | Changes in Value Attributable to Autism Mandates (95% CI) | Mandate Eligible | Mandate Ineligible | Changes in Value Attributable to Autism Mandates (95% CI) | ||||||
| Mandate States | Nonmandate States | Mandate States | Nonmandate States | Mandate States | Nonmandate States | Mandate States | Nonmandate States | ||||
| ASD-specific outpatient health care services | 252.76 | 193.73 | 128.43 | 129.22 | 59.81 (−6.31 to 125.93) | 305.64 | 175.75 | 124.17 | 130.79 | 136.51 (43.34 to 229.67)* | 76.70 (9.82 to 143.57)* |
| ASD-specific behavioral and functional therapy services | 170.37 | 129.26 | 77.38 | 77.69 | 41.42 (−6.13 to 88.97) | 202.43 | 117.96 | 74.70 | 77.78 | 87.57 (29.01 to 146.12)* | 46.14 (−5.19 to 97.48) |
| ASD-specific health care services | 290.06 | 228.81 | 155.19 | 158.64 | 64.70 (−3.42 to 132.81) | 338.00 | 213.34 | 154.46 | 160.83 | 131.04 (29.30 to 232.78)* | 66.34 (−8.62 to 141.31) |
| All outpatient health care services | 517.34 | 456.15 | 383.90 | 391.87 | 69.15 (−33.25 to 171.56) | 592.01 | 407.79 | 379.59 | 389.07 | 193.70 (38.73 to 348.67)* | 124.54 (25.98 to 223.11)* |
| All health care services | 669.00 | 606.00 | 519.94 | 530.37 | 73.43 (−32.60 to 179.45) | 749.94 | 534.28 | 515.15 | 517.09 | 217.60 (48.21 to 386.98)* | 144.17 (35.60 to 252.74)* |
Adjusted models included child sex, age in the given month (estimated on the basis of July 1 in their year of birth), insurance product type (HMO, POS, PPO, EPO, and indemnity or other), and calendar month. The difference-in-differences estimate of change in outcomes attributable to the mandates was calculated as (eligible children in mandate states – eligible children in nonmandate states) – (ineligible children in mandate state – ineligible children in nonmandate states). Differences may be present because of rounding.
P < .05.
We disaggregated total spending to examine the costs borne by the insurer versus the enrollee. Most of the differential effects of the mandates among HDHP and traditional plan enrollees were attributable to changes in insurer-covered health spending. Figure 1 displays the estimated changes in insurer-covered and OOP spending among HDHP and traditional plan enrollees. Relative to traditional plan enrollees, among HDHP enrollees, mandate effects were associated with a $98 (95% CI: $24 to $172) greater increase in average monthly insurer spending on ASD-specific outpatient services, a $142 (95% CI: $34 to $250) greater increase in average monthly insurer spending on all outpatient services, and a $142 (95% CI: $31 to $253) greater increase in average monthly insurer spending on all health services (Fig 1). Differences in mandate effects on OOP spending by consumers were not significantly different between HDHP and traditional plan enrollees.
Differences in estimated effects of autism mandates on monthly OOP and insurer-covered health spending by health plan type, 2008–2012. a Statistically significant difference (P < .05) between HDHP and traditional health plan enrollees in estimated autism mandate effects. Detailed model estimates are available in Supplemental Tables 7 and 8.
Differences in estimated effects of autism mandates on monthly OOP and insurer-covered health spending by health plan type, 2008–2012. a Statistically significant difference (P < .05) between HDHP and traditional health plan enrollees in estimated autism mandate effects. Detailed model estimates are available in Supplemental Tables 7 and 8.
Discussion
In this study, we examined whether the effects of state autism mandates differed on the basis of whether a child with ASD was enrolled in an HDHP. We identified larger increases in spending attributable to mandates in HDHPs relative to traditional plans. Spending differentials after implementation of mandates were driven by higher insurer spending (and not OOP spending) among enrollees in HDHPs.
Our findings suggest that mandates increased service spending among enrollees in HDHPs more than for those in traditional plans. Higher ASD-related spending in mandate states among HDHP enrollees suggests that HDHP enrollees may be able to navigate a complicated benefit design and increase ASD service use. Both HDHP and traditional plan families spend large sums on ASD-related services in mandate states, and these sums likely exceed HDHP deductibles, leading any increases to fall squarely in the insurer-covered spending.
Several limitations are worth noting. First, ASD diagnoses captured in insurance claims were not verified through clinical interview, although they have been shown in other research to have high specificity.27 Second, our data do not capture services delivered in schools or paid completely OOP by families. Likewise, because we only observe claims for children who are using services, if a family’s choice of health plan affects the likelihood of using services at all, that differential selection into our study sample would be unobserved in our analysis. Finally, although the data indicate when a health product is an HDHP, no information is available on the size of the deductible, the presence of a medical savings account (eg, health savings account), or the premiums associated with these plans. There is considerable variation in deductibles across the United States, with 51% of workers having a deductible of ≥$1000, 22% having a deductible of ≥$2000, and 11% having a deductible of ≥$3000 in 2017.9 Although these differences almost certainly influence care-use decisions and we are unable to examine heterogeneity across plan type, we use a large, heterogenous sample of individuals that is roughly generalizable to the commercial large-group insured population as a whole.
Findings from this study are relevant to clinicians, who play an increasingly important role in helping individuals with ASD and their families navigate cost and insurance-related issues. As commercial insurance becomes a more important source for financing health care services for ASD, especially in states that have enacted insurance mandates, clinicians will need to guide their patients on managing the costs of care and serve in an intermediary role with insurers. Across medical specialties, studies have documented patient interest in discussing the cost implications of care with their clinicians,28 and communication about costs has been connected to patient satisfaction rates.29 Children with ASD in our sample used substantial health care services covered by commercial insurance. Although consumers with HDHPs appeared to be sensitive to costs, our study findings suggest that clinicians of those in traditional health plans can play a greater role in helping families with children needing services take better advantage of their commercial insurance benefits to allow for access to beneficial ASD-related services.
The reasons behind the spending differences between those in HDHPs and traditional health plans are not entirely clear. It may be that enrollees in HDHPs are better able to avail themselves of decision support tools that allow them to be more cost savvy consumers relative to those in traditional health plans. According to 1 estimate, as of 2017, 82% of health insurers offered health care cost consumer decision support tools to their HDHP enrollees.30
As the cost of medical care continues to rise, HDHP enrollment is likely to increase. Our findings suggest that families with children with ASD enrolled in HDHPs were not deterred by the high deductible; instead, they increased their health care use without increasing their OOP spending. Future research is needed to better understand how features of HDHPs, such as deductible size and health savings account structure, influence the ability of families to make wise choices in obtaining care and to examine plan premiums and financial strain associated with these plans, particularly for families with children with ASD and high expenditures.
Dr Barry conceptualized the design of the study, the statistical analyses, the interpretation of the data, the drafting of the manuscript, and the review and revision of the manuscript; Dr Kennedy-Hendricks contributed to the conceptualization and design of the study, conducted statistical analyses, contributed to drafting the manuscript, and critically reviewed the manuscript; Drs Eisenberg, Mandell, Epstein, and Candon contributed to the conceptualization and design of the study, contributed to drafting the manuscript, and critically reviewed the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
FUNDING: Supported by National Institute of Mental Health grant R01MH096848. Funded by the National Institutes of Health (NIH).
Comments