Private Equity in Autism Services

The prevalence of autism spectrum disorder (ASD) in the US nearly tripled between 2011 and 2022 (from 2.3 to 6.3 per 1,000) with the greatest increases occurring among 5-to-8-year-olds.¹ This increase, supported by robust state and federal insurance mandates,² fueled a commensurate expansion in demand for therapeutic services, particularly Applied Behavior Analysis (ABA), and created an investment opportunity for private equity (PE) firms. Although PE has rapidly entered many health care sectors, little is known about its role in autism services.

Methods

We searched for any PE acquisition in PitchBook from January 1, 2015 to December 31, 2024 where the industry was “healthcare” and the acquisition contained the key words “autism”, “autism treatment facilities”, or “ABA therapy services” and manually verified and expanded this list in mid-2025, using publicly available data, press releases, and company websites.

We mapped the locations of all ASD service delivery sites owned by PE as of December 31, 2024 and overlaid them with autism prevalence and state autism insurance generosity scores. Autism prevalence was measured using the “Special Education Child Count” dataset hosted by the Centers for Disease Control and Prevention (CDC)³ and state autism insurance generosity scores ranged from 0 to 8 with 8 being the most generous. States were awarded a higher score if they were less restrictive on benefit age limits and spending caps (eMethods in Supplement 1).²

We estimated a linear probability model with state-year level observations. The dependent variable was an indicator variable that equaled 1 in the year a state experienced PE entry and for all subsequent years. The independent variables were an indicator variable that equaled 1 if a state’s autism prevalence was in the top tercile (14.9 children per 1,000) and an indicator variable that equaled 1 if a state has a generous autism insurance mandate (i.e., generosity score> 4). The coefficients of interest represent an absolute probability difference rather than a relative measure (e.g., odds ratio). The model included state and year fixed effects and a control variable for whether the state had expanded Medicaid. Standard errors were clustered at the state level. The study period for the model was 2015 to 2022 given data on autism prevalence and insurance generosity are available through 2022 (eMethods in Supplement 1).

Results

Between 2015 and 2024, we identified PE-acquired 574 ASD service delivery sites, stemming from 147 acquisitions. The majority of PE acquisitions occurred between 2018 and 2022, comprising 79.6% (117 out of 147) of all acquisitions (Figure 1).

Figure 1. PE acquisitions of ASD service delivery sites, 2015–2024.

Source: Authors’ analysis of PitchBook data.

PE-acquired ASD service delivery sites span 42 states, with the largest number in California (97), Texas (81), Colorado (38), Illinois (36), and Florida (36). 16 states had one or no PE-owned ASD service delivery sites.

Visually it appears that PE is more likely to enter states with a higher prevalence of ASD and more generous state autism insurance mandates (Figure 2). In regression analysis, we find that being in the top tercile of ASD prevalence is associated with a 24% increase (0.241, 95% CI: 0.078 to 0.405, P<0.01) in the likelihood of PE entry. Sensitivity analyses using quartiles and different sets of control variables are consistent with this result (eMethods in Supplement 1).

Figure 2. Locations of PE-owned ASD service delivery sites, 2024.

Source: Authors’ analysis of Pitchbook data, autism prevalence as reported in the CDC’s “Special Education Child Count” dataset, and state autism insurance generosity scores developed in McBain et al (2020).²

Notes: Figure 2 Panels A and B overlay the ASD service delivery sites owned by PE as of December 31, 2024 with autism prevalence and state autism insurance generosity scores, respectively.

To construct the maps, we identified all private equity–backed acquisitions dating back to 2010 that remain under current PE ownership. These earlier searches ensured that legacy transactions still held by PE firms were captured in our analysis. However, for reporting and descriptive purposes, we restricted the main sample to acquisitions occurring in 2015 or later, given the very limited number of deals identified between 2010 and 2014. State-level insurance generosity scores were developed in McBain et al (2020)² and range from 0 to 8 with 8 being the most generous. States are awarded a higher score if they were less restrictive on benefit age limits and spending caps.

Discussion

We document the rapid increase in PE investments in the ASD therapy market, coinciding with the large increase in childhood ASD diagnoses. PE entry is associated with ASD prevalence, but it is unclear if this is leading to increases in availability and accessibility to ABA services. Given concerns about PE involvement in ASD services and other healthcare sectors,^4,5 further study is needed to determine potential implications for children with ASD.

Two notable limitations of our study are that we are unable to show PE’s percentage of all ASD service delivery sites, and we are likely undercounting PE acquisitions, the latter of which is generally true of studies that attempt to fully-track PE activity. Additionally, the precision of our coefficient estimates could be affected by limited state-year observations and potential undercounting of acquisitions and there could be important state time-varying confounders missing from our model that would bias our results.