Primary Care Follow-Up After Behavioral Health ED Visits

Key Takeaways

• In this cohort study of Washington State Medicaid beneficiaries, condition-concordant primary care follow-up within 30 days after emergency department visits for mental health conditions, substance use disorders, or alcohol use disorder was uncommon, occurring in fewer than 15% of visits.
• Follow-up occurred after 14.2% of mental health–related visits, 11.2% of substance use disorder–related visits, and 11.1% of alcohol use disorder–related visits.
• Lower probability of follow-up was observed among non-Hispanic Black beneficiaries and people experiencing homelessness, suggesting important racial and social differences in post-ED care continuity.

The clinical focus was whether beneficiaries received primary care follow-up within 30 days of an ED visit for one of these conditions, and which characteristics were associated with a higher or lower probability of such follow-up. Rather than measuring any outpatient visit after emergency care, the study specifically evaluated follow-up in primary care that matched the condition associated with the ED visit.

The analysis included 859,043 Medicaid ED visit claims from 367,245 unique beneficiaries in Washington during 2022. Mean age at the time of the ED claim was 41.7 years, and 57.8% of claims were for women. The study evaluated ED visits related to mental health conditions, all substance use disorders, and, in a subgroup analysis, alcohol use disorder. Condition-concordant follow-up was defined as a primary care visit within 30 days in which the primary diagnosis code matched the primary diagnosis code for the index ED visit. Both telehealth and in-person primary care visits were included if they met the study’s claims-based definition of primary care.

Follow-up rates were low across all three categories. Among 131,704 mental health–related ED visits, 18,722, or 14.2%, were followed by 30-day condition-concordant primary care follow-up. Among 101,684 substance use disorder–related ED visits, 11,353, or 11.2%, had such follow-up. In the alcohol use disorder subgroup, 3,675 of 33,196 ED visits, or 11.1%, were followed by condition-concordant primary care within 30 days. The authors therefore characterized follow-up after these ED visits as infrequent.

The study also examined beneficiary characteristics associated with follow-up using multivariable logistic regression with estimated marginal effects. Across all three condition groups, non-Hispanic Black beneficiaries had a lower probability of receiving condition-concordant primary care follow-up than other racial groups. Compared with non-Hispanic Black beneficiaries, non-Hispanic White beneficiaries had higher estimated probabilities of follow-up by 4.47 percentage points for mental health–related visits, 4.70 percentage points for substance use disorder–related visits, and 4.00 percentage points for alcohol use disorder–related visits. Other racial groups, including Alaska Native or American Indian and Asian or Pacific Islander beneficiaries, also had higher follow-up probabilities than non-Hispanic Black beneficiaries in these models.

Experiencing homelessness was consistently associated with lower probability of follow-up. Estimated marginal effects were −2.74 percentage points for mental health–related visits, −1.88 percentage points for substance use disorder–related visits, and −1.86 percentage points for alcohol use disorder–related visits, compared with beneficiaries not experiencing homelessness. These findings were highlighted by the authors as evidence of social differences in access to timely primary care after behavioral health–related ED use.

Other associations varied by condition group. For mental health–related ED visits, female sex, rural residence, and higher comorbidity burden were associated with higher probability of follow-up, while older age was associated with lower probability. For substance use disorder–related visits, older age and higher comorbidity burden were associated with greater follow-up probability, while homelessness remained the main factor associated with lower follow-up. For alcohol use disorder–related visits, older age, female sex, and higher comorbidity burden were associated with higher follow-up probability, whereas rural residence and homelessness were associated with lower follow-up probability.

The authors framed condition-concordant primary care follow-up as a continuity measure rather than a catch-all marker of any care after the ED visit. They also noted that their analysis did not capture follow-up in specialty behavioral health, addiction medicine, residential treatment, or detoxification settings. As a result, the findings should be interpreted as showing gaps in primary care follow-up specifically, not the absence of all post-ED care.

Overall, the study describes low rates of 30-day condition-concordant primary care follow-up after ED visits for mental health conditions, substance use disorders, and alcohol use disorder among Washington Medicaid beneficiaries. The findings point to substantial racial and social differences, particularly for non-Hispanic Black beneficiaries and people experiencing homelessness, and suggest that tailored outreach and care coordination may be needed to improve continuity into primary care after ED use.