Mental Health Service Use by Medicaid-Enrolled Children and Adolescents in Primary Care Safety-Net Clinics
Abstract
Objective:
Little is known about the role of primary care safety-net clinics, including federally qualified health centers and rural health clinics, in providing mental health services to youths. This study examined correlates and quality of mental health care for youths treated in these settings.
Methods:
Medicaid claims data (2008–2010) from nine states were used to identify youths initiating medication for attention-deficit hyperactivity disorder (ADHD) (N=6,433) and youths with an incident depression diagnosis (N=13,209). The authors identified youths who received no ADHD or depression-related visits in a primary care safety-net clinic, some (but less than most) visits in these clinics, and most visits in these clinics. Using bivariate and regression analyses, they examined correlates of mental health treatment in these settings and whether mental health visits in these settings were associated with quality measures.
Results:
Only 13.5% of the ADHD cohort and 7.2% of the depression cohort sought any ADHD- or depression-related visits in primary care safety-net clinics. Residence in a county with a higher (versus lower) percentage of residents living in an urban area was negatively associated with receiving the majority of mental health visits in these settings (p<0.05). Compared with youths with no visits in these settings, youths who received most of their mental health treatment in these settings received lower-quality care on five of six measures (p<0.01).
Conclusions:
As investment in the expansion of mental health services in primary care safety-net clinics grows, future research should assess whether these resources translate into improved mental health care access and quality for Medicaid-enrolled youths.
HIGHLIGHTS
• Correlates and quality of mental health treatment for Medicaid-enrolled youths in nine states (2008–2010) who were treated in primary care safety-net settings (including federally qualified health centers and rural health clinics) were examined.
• Only 13.5% of 6,433 youths initiating ADHD medication and 7.2% of 13,209 youths with an index depression diagnosis sought mental health treatment in primary care safety-net clinics.
• Compared with youths who were not treated in primary care safety-net settings, those who received most of their mental health treatment in these settings generally received lower-quality care.
Mental disorders are common and undertreated among youths (1, 2). Medicaid is the largest insurer of youths (3), and research has identified a number of access-related barriers to mental health treatment for Medicaid-enrolled youths (4–8). Researchers and policy makers have highlighted the potential of federally qualified health centers (FQHCs) and rural health clinics (RHCs)—collectively referred to as primary care safety-net clinics—to address access-related barriers to mental health treatment (9, 10).
FQHCs and RHCs are safety-net facilities that provide primary care to underserved populations, including Medicaid enrollees. These clinics are located in federally designated Health Professional Shortage Areas, receive favorable reimbursement rates from Medicaid, and are eligible for participation in federal initiatives, such as loan repayment programs (11, 12). These two programs differ from one another on certain key dimensions; FQHCs have greater requirements than do RHCs for staffing and service offerings, and RHCs must be located in nonurbanized areas (11, 12). Over 10,000 FQHC sites and 4,100 RHCs deliver primary care to communities across the country (13).
Primary care safety-net clinics have the potential to improve access-related barriers to mental health treatment for Medicaid-enrolled youths. First, these clinics can address geographic barriers to care, because more than three-fourths of counties that lack any specialty mental health treatment facility have at least one primary care safety-net clinic (7). Second, for some families, these clinics may help reduce stigma associated with seeking services in a separate mental health specialty setting (10). Finally, FQHCs are required by law to offer enabling services to address access-related barriers, such as transportation, translation and interpretation, and insurance enrollment (14, 15).
Research has found that the percentage of FQHCs that offer specialty mental health services on site has increased substantially in the past 2 decades (16, 17). There is, however, little information about the role of primary care safety-net clinics in providing mental health services to Medicaid-enrolled youths. We addressed this gap by using Medicaid claims data to identify two cohorts of youths—those with attention-deficit hyperactivity disorder (ADHD) and those with depression, two of the most common mental disorders in the child and adolescent population (1, 2). In each cohort, we described the percentage that received mental health care in a primary care safety-net clinic and examined the correlates associated with receipt of mental health treatment in these settings. For each cohort, we also examined several measures of care quality in primary care safety-net settings.
Methods
Data
Data came from the 2008–2010 Medicaid Analytic eXtract (MAX) Files for nine states (Alabama, Georgia, Kentucky, Louisiana, Missouri, North Carolina, Tennessee, Texas, and Virginia). The MAX Files include information on Medicaid eligibility, health care utilization, and enrollee demographic characteristics. Researchers have evaluated the completeness and accuracy of managed care data in the MAX files for each state (18–20), and the states included in this study have sufficiently complete managed care claims for use in data analysis.
The MAX Files with enrollee information were merged with three additional files to obtain taxonomy codes that could be used to identify visits in primary care safety-net clinics. These files included the Centers for Medicare and Medicaid Services (CMS) MAX Provider Characteristics (MAXPC) file (21), the National Provider Identifier (NPI) Data File (22), and the CMS Provider of Services (POS) file (23). We also merged measures from the Area Health Resources File (24).
Cohorts
We used specifications from the Healthcare Effectiveness Data and Information Set (HEDIS) guidelines (25) and prior literature (26) to derive a cohort of youths (ages 6–12) with a diagnosis of ADHD (i.e., at least two claims with an ADHD diagnosis code [ICD-9-CM codes 314.00 and 314.01]) who initiated ADHD medication for the first time between January 1, 2010, and February 28, 2010 (N=6,433). We identified those with continuous Medicaid enrollment from the time they were first observed in the database through the end of the treatment initiation period (with an allowable administrative gap up to 30 days) and without a fill for an ADHD medication for at least 120 days prior to medication initiation (i.e., the HEDIS-defined exclusion period) (25).
Next, we derived a cohort of youths (ages 5–17) with an incident diagnosis of depression between January 1, 2010, and August 8, 2010 (N=13,209). Our cohort included those with at least two claims with a depression diagnosis code (ICD-9-CM codes 296.2, 296.3, 300.4, and 311) on different days in 2010. We identified those with continuous Medicaid enrollment from the time they were first observed in the database through the end of the study period (with an allowable administrative gap of up to 30 days) and without any encounters with a depression diagnosis code or a fill for an antidepressant medication for at least 90 days prior to the index diagnosis (i.e., the exclusion period used in prior literature [27]). In both cohorts, we excluded those with dual Medicare eligibility, an inpatient claim for mental health or substance abuse treatment, multiple county codes, or missing information on control variables.
Safety-Net Measure
To derive measures of mental health treatment in primary care safety-net clinics, we used codes from the MAX files, including place-of-service codes (03, 50, 53, and 72), type-of-program codes (03, 04), revenue codes (521, 522, 524, 525, 527, and 528), and procedure code (T1015); taxonomy codes from the MAXPC file and NPI Data File (261QF0400x and 261QR1300x); and provider category codes (12, 21) from the CMS POS file. Using these codes, we created two categorical variables for ADHD- and depression-related visits (i.e., visits with a primary or secondary diagnosis of one of these conditions). The first measure identified those who did not receive any ADHD- or depression-related visits in a primary care safety-net clinic (FQHC or RHC), those who received some but not a majority of ADHD- or depression-related visits in a primary care safety-net clinic, and those who received the majority of ADHD- or depression-related visits in a primary care safety-net clinic. Next, we classified youths in each cohort into those who did not receive any ADHD- or depression-related visits in a primary care safety-net clinic, those who received any ADHD- or depression-related visits in an FQHC, and those who received ADHD- or depression-related visits in a primary care safety-net clinic exclusively from an RHC (i.e., no visit in an FQHC). (Bivariate and regression analyses examining correlates associated with mental health treatment in primary care safety-net settings using this second measure are available in an online supplement . For those who received any ADHD- or depression-related visits in a primary care safety-net facility, we also provide information in the online supplement about psychotherapy visits received inside and outside these settings.)
Quality of Care
On the basis of HEDIS specifications and prior literature (25, 26), we derived three measures to assess adequate follow-up care and medication continuity after the child initiated ADHD medication. The first measure assessed adequate follow-up care in the initiation phase of ADHD medication treatment (i.e., the first 30 days after initiating medication), which was defined as at least one visit with a health care provider during this period. The second measure assessed continuous medication treatment, defined by HEDIS as those who filled medication for 210 of the 300-day continuation and maintenance (C&M) phase following the 30-day medication initiation period (28). We analyzed this outcome measure for a subgroup with continuous Medicaid enrollment in the C&M phase (N=5,968). The third measure assessed adequate follow-up care in the C&M phase, defined as receiving at least two additional health care visits in the 300-day C&M phase. This outcome measure was assessed for those with continuous enrollment and continuous medication in the C&M phase (N=2,370).
In the cohort with an index depression diagnosis, we used specifications from prior research (27, 29) to create indicators for youths who received minimally adequate psychotherapy (four or more individual, family, or group psychotherapy sessions outside an inpatient setting in the 12 weeks following initiation of treatment), minimally adequate medication treatment (antidepressant medication for ≥84 of the 144 days following initiation), and minimally adequate treatment (receipt of minimally adequate psychotherapy or minimally adequate medication treatment).
Covariates
Individual-level measures.
We assessed predisposing (age, gender, and race-ethnicity), enabling (Medicaid health plan type [29, 30]), and need-related characteristics (basis of Medicaid eligibility and comorbidities) that may be correlated with receipt of mental health treatment in a primary care safety-net clinic or the quality of care received (31) (see table in online supplement for details).
County-level measures.
Contextual-level enabling characteristics (31) included the percentage of county residents living in an urban area (2000) (32) and living in poverty (2008). We also examined the per capita (100,000 population) number of primary care safety-net clinics (FQHCs and RHCs) (2008), primary care physicians (2010), and psychologists (2009).
Analysis
We conducted bivariate analyses using Wald tests and multivariate analyses using generalized ordered logistic regressions to examine the correlates of mental health treatment in a primary care safety-net clinic. Next, we conducted bivariate analyses using Wald tests and multiple logistic regression analyses to examine whether the receipt of mental health care in one of these settings was correlated with quality measures. Regression models controlled for covariates described above; these analyses also included state indicators, and standard errors were clustered at the county level.
Results
Mental Health Visits in a Primary Care Safety-Net Setting
In the cohort that initiated ADHD medication, 4.0% received some ADHD-related visits (but not the majority of such visits) in a primary care safety-net clinic and 9.5% received the majority of ADHD-related visits in one of these settings (Table 1). Most of those who received any treatment in one of these clinics sought care exclusively in an RHC.
| Youths initiating ADHD medication (N=6,433) | Youths with an index depression diagnosis (N=13,209) | |||
|---|---|---|---|---|
| ADHD- or depression-related visit | N | % | N | % |
| No visits | 5,566 | 86.5 | 12,266 | 92.9 |
| Some visits (but fewer than the majority) | 256 | 4.0 | 365 | 2.8 |
| Majority of visits | 611 | 9.5 | 578 | 4.4 |
| Any visits to an FQHC | 354 | 5.5 | 482 | 3.6 |
| Any visits to an RHC (no FQHC visit) | 513 | 8.0 | 461 | 3.5 |
a
Includes federally qualified health centers (FQHCs) and rural health clinics (RHCs).
In the cohort with depression, a smaller percentage received care in a primary care safety-net clinic (Table 1). Specifically, 2.8% received some of their depression-related visits (but not the majority of such visits) in a primary care safety-net clinic, and 4.4% received the majority of their depression-related visits in one of these settings. Just under half of those who sought care in a primary care safety-net clinic received treatment exclusively in an RHC.
Correlates Associated With Mental Health Treatment in a Primary Care Safety-Net Clinic
Child-level correlates.
In the cohort that initiated ADHD medication, bivariate analyses indicated that diagnoses for comorbid conditions—including oppositional defiant disorder or conduct disorder (p<0.001), other mental disorders (i.e., anxiety, bipolar disorder, schizophrenia/psychoses, and other mental health conditions; p<0.001), and asthma (p<0.001)—were negatively correlated with receiving the majority of ADHD-related visits in a primary care safety-net clinic (Table 2). In the depression cohort, bivariate analyses indicated that having any diagnosis of major depression (p<0.001) or dysthymia (no major depression, p<0.05) was negatively correlated with receiving most depression-related visits in a primary care safety-net setting (Table 2). In each cohort, these negative correlations remained significant in the regression analysis that controlled for other child- and county-level correlates (Table 3).
| ADHD-related visits (N=6,433)b | Depression-related visits (N=13,209)b | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| None (N=5,566) | Some (but fewer than majority) (N=256) | Majority (N=611) | None (N=12,266) | Some (but fewer than majority) (N=365) | Majority (N=578) | |||||||
| Characteristic | N | % | N | % | N | % | N | % | N | % | N | % |
| Demographic | ||||||||||||
| Race-ethnicity | ||||||||||||
| Non-Hispanic white | 2,560 | 46.0 | 129 | 50.4 | 354 | 57.9*** | 5,770 | 47.0 | 219 | 60.0*** | 325 | 56.2*** |
| Black | 1,703 | 30.6 | 75 | 29.3 | 134 | 21.9*** | 3,863 | 31.5 | 78 | 21.4*** | 131 | 22.7*** |
| Hispanic | 827 | 14.9 | 35 | 13.7 | 75 | 12.3 | 1,828 | 14.9 | 54 | 14.8 | 89 | 15.4 |
| Other or unknown | 476 | 8.6 | 17 | 6.6 | 48 | 7.9 | 805 | 6.6 | 14 | 3.8** | 33 | 5.7 |
| Age (M±SD) | 8.2±1.6 | 8.0±1.6* | 8.2±1.6 | 12.8±3.0 | 13.8±2.6*** | 13.3±2.8*** | ||||||
| Female (reference: male) | 1,786 | 32.1 | 79 | 30.9 | 203 | 33.2 | 6,685 | 54.5 | 245 | 67.1*** | 355 | 61.4*** |
| Medicaid plan type | ||||||||||||
| Fee for service (no carve-out) | 362 | 6.5 | 31 | 12.1** | 127 | 20.8*** | 1,346 | 11.0 | 68 | 18.6*** | 116 | 20.1*** |
| Primary care case management (no carve-out) | 1,011 | 18.2 | 35 | 13.7* | 186 | 30.4*** | 1,656 | 13.5 | 60 | 16.4 | 174 | 30.1*** |
| Comprehensive managed care plan (no carve-out)c | 2,517 | 45.2 | 114 | 44.5 | 158 | 25.9*** | 7,311 | 59.6 | 195 | 53.4* | 202 | 34.9*** |
| Mixed pland | 1,676 | 30.1 | 76 | 29.7 | 140 | 22.9*** | 1,953 | 15.9 | 42 | 11.5** | 86 | 14.9 |
| Medicaid eligibility type: blind, disabled, or foster care | 967 | 17.4 | 65 | 25.4** | 91 | 14.9 | 2,843 | 23.2 | 65 | 17.8** | 77 | 13.3*** |
| Mental illness comorbidity | ||||||||||||
| ADHD | — | — | — | — | — | — | 3,947 | 32.2 | 90 | 24.7** | 182 | 31.5 |
| Any depressive disorder | 240 | 4.3 | 24 | 9.4** | 28 | 4.6 | — | — | — | — | — | — |
| Depression not otherwise specified only | — | — | — | — | — | — | 5,670 | 46.2 | 103 | 28.2*** | 351 | 60.7*** |
| Dysthymia (no major depression) | — | — | — | — | — | — | 1,443 | 11.8 | 25 | 6.9*** | 51 | 8.8* |
| Any major depression diagnosis | — | — | — | — | — | — | 5,153 | 42.0 | 237 | 64.9*** | 176 | 30.5*** |
| Oppositional defiant disorder or conduct disorder | 962 | 17.3 | 56 | 21.9 | 59 | 9.7*** | 3,034 | 24.7 | 81 | 22.2 | 99 | 17.1*** |
| Other mental disorder | 1,852 | 33.3 | 104 | 40.6* | 159 | 26.0*** | 6,218 | 50.7 | 225 | 61.6*** | 259 | 44.8** |
| Physical health comorbidity: asthma | 915 | 16.4 | 35 | 13.7 | 70 | 11.5*** | 1,615 | 13.2 | 53 | 14.5 | 73 | 12.6 |
| M | SD | M | SD | M | SD | M | SD | M | SD | M | SD | |
| County-level characteristic | ||||||||||||
| % living in urban area | 68.5 | 28.8 | 58.4*** | 33.3 | 49.7*** | 30.3 | 67.4 | 30.5 | 54.3*** | 33.4 | 53.9*** | 32.4 |
| % living in poverty | 16.9 | 6.1 | 17.4 | 5.5 | 18.5*** | 5.4 | 16.5 | 6.3 | 17.5*** | 5.7 | 18.5*** | 6.1 |
| Primary care safety-net clinics per 100,000 population | 2.9 | 5.4 | 7.4*** | 9.6 | 11.0*** | 11.8 | 3.7 | 7.4 | 9.5*** | 13.5 | 10.4*** | 13.4 |
| Primary care physicians per 100,000 population | 61.5 | 28.2 | 55.1*** | 28.3 | 48.7*** | 25.8 | 62.4 | 29.6 | 55.1*** | 32.0 | 51.8*** | 27.7 |
| Psychologists per 100,000 population | 16.7 | 18.0 | 15.3 | 17.7 | 10.7*** | 15.2 | 18.7 | 19.1 | 14.8*** | 16.0 | 14.1*** | 16.7 |
a
Includes federally qualified health centers (FQHCs) and rural health clinics (RHCs).
b
Bivariate analyses were conducted by using Wald tests to compare those who received some or the majority of their ADHD- or depression-related visits in a primary care safety-net clinic with those who did not receive any visits in these settings (reference).
c
Monthly information about plan enrollment was used to measure health plan type. This category includes youths who were enrolled in a comprehensive managed care plan for their entire observation period.
d
Includes youths who were enrolled in more than one plan type during the observation period (including behavioral health plans).
*
p<.05, **p<.01, *** p<.001.
Plan type was also significantly associated with whether youths had mental health-related visits in a primary care safety-net setting in bivariate and regression analyses. In each cohort, regression results indicated that those enrolled in comprehensive managed care plans were less likely than those enrolled in fee for service to receive the majority of their ADHD- or depression-related visits in a primary care safety-net setting (p<0.001) (Table 3).
| ADHD-related visitsb | Depression-related visitsb | |||||||
|---|---|---|---|---|---|---|---|---|
| Some (but fewer than majority) (intercept=3.9%) | Majority (intercept=9.6%) | Some (but fewer than majority) (intercept=2.8%) | Majority (intercept=4.3%) | |||||
| Characteristic | Adjusted percentage point difference | 95% CI | Adjusted percentage point difference | 95% CI | Adjusted percentage point difference | 95% CI | Adjusted percentage point difference | 95% CI |
| Demographic | ||||||||
| Race-ethnicity (reference: non-Hispanic white) | ||||||||
| Black | .3 | –1.0, 1.5 | .2 | –1.8, 2.2 | −.4 | –1.4, .5 | −.2 | –1.6, 1.3 |
| Hispanic | −.8 | –3.1, 1.5 | .1 | –3.6, 3.8 | 1.1 | −.5, 2.6 | 1.5 | −.6, 3.6 |
| Other or unknown | –1.2 | –3.1, .7 | 1.8 | –1.3, 4.9 | −.6 | –2.2, 1.0 | 1.4 | −.6, 3.4 |
| Age | −.3 | −.6, .0002 | .02 | −.4, .5 | .2 | .1, .3 | .3 | .1, .4 |
| Female (reference: male) | −.3 | –1.1, .6 | .1 | –1.2, 1.4 | 1.1 | .5, 1.7 | .6 | −.2, 1.4 |
| Medicaid plan type (reference: fee for service [no carve-out]) | ||||||||
| Primary care case management (no carve-out) | 1.5 | −.5, 3.6 | –3.1 | –6.3, .04 | .3 | –1.3, 2.0 | –1.5 | –3.9, .8 |
| Comprehensive managed care plan (no carve-out)c | 2.9 | .6, 5.3 | –9.0 | –11.9, –6.1 | .5 | −.5, 1.6 | –1.9 | –3.7, –.1 |
| Mixed pland | 1.9 | −.2, 4.0 | –5.8 | –8.7, –3.0 | .6 | −.5, 1.8 | −.9 | –2.8, 1.0 |
| Medicaid eligibility type: blind, disabled, or foster care (reference: other type)e | 2.9 | 1.2, 4.6 | –2.0 | –4.8, .8 | −.8 | –2.0, .4 | –2.3 | –3.7, –.8 |
| Mental health comorbidity | ||||||||
| ADHD | — | — | — | — | .5 | −.2, 1.1 | .8 | −.04, 1.6 |
| Any depressive disorder (reference: depression not otherwise specified only) | 4.1 | 1.3, 6.9 | −.1 | –3.2, 3.0 | — | — | — | — |
| Dysthymia (no major depression) | — | — | — | — | .1 | –1.5, 1.6 | –2.4 | –4.3, –.5 |
| Any major depression diagnosis | — | — | — | — | 1.7 | 1.1, 2.4 | –2.7 | –3.7, –1.6 |
| Oppositional defiant disorder or conduct disorder | 1.6 | −.1, 3.4 | –3.7 | –6.2, –1.3 | .03 | −.7, .8 | −.7 | –1.9, .5 |
| Other mental disorder | 1.6 | .5, 2.7 | –1.9 | –3.6, –.2 | .9 | .3, 1.6 | .04 | −.8, .9 |
| Physical health comorbidity: asthma | −.8 | –2.2, .7 | –2.1 | –4.1, –.03 | .2 | −.7, 1.2 | .3 | −.7, 1.3 |
| County-level characteristicf | ||||||||
| % living in urban area | −.3 | –1.1, .5 | –2.3 | –3.8, –.9 | −.6 | –1.1, –.1 | –1.2 | –2.0, –.4 |
| % living in poverty | −.4 | –1.0, .3 | −.4 | –1.8, .9 | .001 | −.5, .5 | .2 | −.7, 1.0 |
| Primary care safety-net clinic per 100,000 population | 1.8 | 1.1, 2.5 | 3.6 | 2.4, 4.8 | .5 | .2, .9 | 1.0 | .5, 1.5 |
| Primary care physicians per 100,000 population | .2 | −.6, 1.0 | −.9 | –2.4, .6 | .01 | −.5, .5 | −.3 | –1.1, .4 |
| Psychologists per 100,000 population | −.4 | –1.1, .3 | .8 | −.9, 2.4 | −.1 | −.6, .3 | .8 | .03, 1.6 |
a
Includes federally qualified health centers (FQHCs) and rural health clinics (RHCs).
b
Generalized ordered logistic regression was estimated with state indicators; standard errors were clustered at the county level. The adjusted percentage-point difference refers to the marginal effect, or the change in the predicted percentage of youths in a given outcome category associated with a 1-unit change in an independent variable of interest. In generalized ordered logistic regressions, the comparison group for a specific outcome category includes youths in all other categories. For example, the reference group for youths who received the majority of visits in a primary care safety-net setting included those who received no visits or some (but not the majority) of visits in these settings.
c
Monthly information about plan enrollment was used to measure health plan type. This category includes youths who were enrolled in a comprehensive managed care plan for their entire observation period.
d
Includes youths who were enrolled in more than one plan type during the observation period (including behavioral health plans).
e
Reference category includes youths eligible for Medicaid based on household income, classification as “medically needy,” or other criteria specified in each state’s Section 1115 waiver.
f
Variables were standardized so that a 1-unit increase corresponds to an increase of 1 standard deviation in the measure above its mean value.
County-level correlates.
In both cohorts, bivariate analyses indicated that county-level measures of the percentage of residents who lived in an urban area, primary care physicians per capita, and psychologists per capita were negatively associated with receiving the majority of ADHD- or depression-related visits in a primary care safety-net setting (p<0.001 for all) (Table 2). Conversely, county-level measures of the percentage of residents living in poverty and primary care safety-net facilities per capita were positively associated with receiving the majority of ADHD- or depression-related visits in a primary care safety-net setting (p<0.001 for both). In regression analyses, the findings associated with percentage of county residents living in an urban area remained negative and significant, and the finding associated with supply of primary care safety-net clinics remained positive and significant (Table 3).
Quality Measures Across Safety-Net Settings
The findings of bivariate and multivariate analyses examining measures of care quality across settings were mixed for the ADHD cohort. Results of regression analysis controlling for individual- and community-level characteristics are shown in Table 4. (The full regression results for Table 4 and Table 5 are available in an online supplement ). Compared with youths who did not receive any ADHD-related visits in a primary care safety-net clinic, those who had some (but not most) of their visits in one of these clinics were 7.9 percentage points more likely to receive adequate follow-up care in the initiation phase and 7.6 percentage points more likely to continue medication. Conversely, compared with youths who did not have any ADHD-related visits in a primary care safety-net clinic, those who received the majority of their visits in these clinics were 27.7 percentage points less likely to receive adequate follow-up care in the initiation phase. They were also 24.3 percentage points less likely to receive adequate follow-up care in the C&M phase of medication treatment and 6.6 percentage points less likely to continue medication.
| ≥1 follow-up visit in initiation phase (N=6,433)b | Continued medication (N=5,968)c | Adequate follow-up care in C&M phase (N=2,370)d | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Visits in safety-net clinic | Total N | N | %e | Adjusted percentage point difference (intercept=63.8)f | 95% CI | Total N | N | %e | Adjusted percentage point difference (intercept=39.7)f | 95% CI | Total N | N | %e | Adjusted percentage point difference (intercept=69.3)f | 95% CI |
| None (reference) | 5,566 | 3,705 | 66.6 | — | — | 5,147 | 2,030 | 39.4 | — | — | 2,030 | 1,464 | 72.1 | — | — |
| Some (but fewer than majority) | 256 | 192 | 75.0** | 7.9* | 1.5, 14.2 | 244 | 128 | 52.5*** | 7.6* | 1.4, 13.8 | 128 | 99 | 77.3 | 5.6 | –3.1, 14.4 |
| Majority | 611 | 207 | 33.9*** | –27.7*** | –32.5, –23.0 | 577 | 212 | 36.7 | –6.6** | –11.2, –2.1 | 212 | 79 | 37.3*** | –24.3*** | –31.4, –17.3 |
a
Includes federally qualified health centers (FQCHs) and rural health clinics (RHCs).
b
Follow-up visit in the first 30 days after initiating medication.
c
Filled a medication prescription for 210 of the 300-day continuation and maintenance (C&M) phase following the 30-day medication initiation period.
d
Received at least two additional visits in the 300-day C&M phase.
e
Bivariate comparisons using Wald tests were conducted to compare outcomes among those who received some or the majority of ADHD visits in a primary care safety-net clinic versus those who received no ADHD visits in a primary care safety-net clinic (reference).
f
Logistic regressions were estimated with state indicators, and standard errors were clustered at the county level. All models controlled for individual-level age, gender, race-ethnicity, health plan type, basis of eligibility, comorbidities, and county-level demographic characteristics and health care resources. Youths who received no ADHD visits in a primary care safety-net clinic were the reference group.
*
p<.05, **p<.01, ***p<.001.
| Minimally adequate psychotherapyb | Minimally adequate pharmacotherapyc | Minimally adequate treatmentd | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Visits in safety-net clinic | Total N | N | %e | Adjusted percentage point difference (intercept=32.9)f | 95% CI | N | %e | Adjusted percentage point difference (intercept=16.1)f | 95% CI | N | %e | Adjusted percentage point difference (intercept=43.8)f | 95% CI |
| None (reference) | 12,266 | 4,128 | 33.7 | — | — | 1,878 | 15.3 | — | — | 5,396 | 44.0 | — | — |
| Some (but fewer than majority) | 365 | 126 | 34.5 | –3.1 | –8.3, 2.0 | 123 | 33.7*** | 9.3*** | 5.8, 12.9 | 199 | 54.5*** | 4.0 | –1.6, 9.7 |
| Majority | 578 | 88 | 15.2*** | –19.9*** | –25.3, –14.5 | 124 | 21.5*** | 4.0* | .9, 7.1 | 196 | 33.9*** | –9.5*** | –14.6, –4.4 |
a
Includes federally qualified health centers (FQHCs) and rural health clinics (RHCs).
b
Four or more individual, family, or group psychotherapy sessions outside an inpatient setting in the 12 weeks following initiation of treatment.
c
Antidepressant medication for ≥84 of the 144 days following initiation.
d
Received minimally adequate psychotherapy or medication treatment.
e
Bivariate comparisons using Wald tests were conducted to compare outcomes among those who received some or the majority of depression visits in a primary care safety-net clinic versus those who received no depression visits in a primary care safety-net clinic (reference).
f
Logistic regressions were estimated with state indicators, and standard errors were clustered at the county level. All models controlled for individual-level age, gender, race-ethnicity, health plan type, basis of eligibility, comorbidities, and county-level demographic characteristics and health care resources. Youths who received no depression visits in a primary care safety-net clinic were the reference group.
*
p<.05, **p<.01, ***p<.001.
Findings from the bivariate and multivariate comparisons were also mixed for the depression cohort (Table 5). Compared with youths with no depression-related visits in a primary care safety-net clinic, those who received some (but not most) of their visits in these settings were 9.3 percentage points more likely to receive minimally adequate pharmacotherapy. Those with most of their visits in these settings were 4.0 percentage points more likely than youths with no such visits to receive minimally adequate pharmacotherapy. On the other hand, compared with youths with no visits in these settings, youths who received most of their depression-related visits in these settings were 19.9 percentage points less likely to receive minimally adequate psychotherapy and 9.5 percentage points less likely to receive any minimally adequate treatment (psychotherapy or pharmacotherapy).
Discussion
Children living in counties with a higher percentage of residents living in an urban area were less likely to receive mental health treatment in a primary care safety-net clinic than those in counties with a lower percentage of residents in an urban area, which adds to prior literature highlighting the potential of primary care safety-net clinics to fill gaps in the mental health treatment system outside urban areas (7). Our findings also shed light on the role of RHCs in this infrastructure, because more than half of youths initiating ADHD medication in a primary care safety-net clinic sought treatment exclusively from an RHC (versus an FQHC). Because RHCs are required to be located in nonurbanized areas (11), many RHCs serve populations living in communities with extremely limited (if any) mental health care resources (5).
In the cohort that initiated ADHD medication, there was a negative association between having diagnosed comorbid mental disorders and receiving the majority of ADHD-related visits in a primary care safety-net setting. Similarly, in the cohort with an index depression diagnosis, there was a negative association between having any diagnosis of major depression and receiving the majority of visits in a primary care safety-net clinic. Together, these findings add to prior literature indicating that primary care safety-net clinics may serve those with less severe mental health needs, compared with individuals treated in other settings (16). Another possible explanation, however, may involve coding practices. If providers in primary care safety-net clinics are less likely to enter secondary diagnosis codes for mental disorders into the medical record (regardless of the underlying severity of mental health needs), this may also account for some of the differences in diagnosed comorbidities in the ADHD cohort
Our results also indicate that enrollment in comprehensive managed care plans (compared with enrollment in fee-for-service plans) was negatively associated with the receipt of most ADHD- or depression-related visits in primary care safety-net clinics. This association may be explained by multiple mechanisms, including greater enrollment in comprehensive managed care plans in urban areas (33) (where RHCs are not located) and more complete coding practices by providers in areas and states served by comprehensive managed care plans (34).
Compared with youths who received no visits in primary care safety-net clinics, youths who received most of their ADHD- or depression-related visits in these clinics received lower quality care on five of the six outcome measures examined. These findings diverge from prior literature reporting that the quality of care for patients treated in primary care safety-net clinics (FQHCs in particular) is comparable to national averages or to the care received by those treated in other physician offices (35, 36). Our findings may represent unmeasured differences in child- or family-level characteristics (such as need or preferences for services) between those seeking care in different settings. It is also possible that primary care safety-net clinics have fewer staff with specialty training needed to serve youths with mental health needs. Nevertheless, it is worth noting that the outcome measures used in this study were based either on specifications from the HEDIS performance measurement database (28, 37) or on clinical guidelines (38, 39). Thus these measures represent important targets that any health care setting or provider should aim to achieve.
The results also suggest that primary care safety-net clinics played a relatively small role in the provision of mental health services to youths during the study period (2008–2010). The percentage of primary care safety-net clinics offering specialty mental health services has increased in recent years (18), and the federal government has invested considerable resources to help primary care safety-net clinics expand their capacity to offer mental health services (40). In fiscal year 2017, the Health Resources and Services Administration awarded more than $200 million for behavioral health expansion grants to 1,178 health centers and 13 rural health organizations to increase access to substance abuse and mental health services (40–42). Future studies should assess whether this investment has translated into an expansion of behavioral health services in primary care safety-net settings for the child and adolescent population.
Several limitations of this study should be acknowledged. First, the data are several years old, and the findings from these states may not generalize to other states. Second, there were unmeasured organization-level characteristics, including the demographic composition of the practice or clinic (e.g., age composition) and whether the practice or clinic had any collaborative care relationships with mental health providers outside the practice. Third, coding errors in administrative claims databases may result in measurement error (43). Finally, because the data are cross-sectional, causality cannot be inferred from these analyses.
Conclusions
This study examined the role of primary care safety-net clinics in the provision of mental health services to Medicaid-enrolled youths. Children in less urbanized areas were more likely to receive their mental health care in these settings. Nevertheless, these facilities served a relatively small percentage of Medicaid-enrolled youths seeking mental health treatment. As investment in the expansion of mental health services in the primary care safety-net grows, it will be critical to assess whether additional resources translate into improved mental health care access and quality for this population.
Acknowledgments
The authors gratefully acknowledge the helpful comments of Adam Wilk, Ph.D.
Footnotes
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
The authors report no financial relationships with commercial interests.
Supplementary Material
File (appi.ps.201800540.ds001.pdf)
- View/Download
- 211.48 KB
References
1.
Merikangas KR, He J-P, Brody D, et al: Prevalence and treatment of mental disorders among US children in the 2001–2004 NHANES. Pediatrics 2010; 125:75–81
2.
Merikangas KR, He J-P, Burstein M, et al: Lifetime prevalence of mental disorders in US adolescents: results from the National Comorbidity Survey Replication–Adolescent Supplement (NCS-A). J Am Acad Child Adolesc Psychiatry 2010; 49:980–989
3.
Medicaid Facts: United States. Itasca, IL, American Academy of Pediatrics, 2017. https://www.aap.org/en-us/Documents/federaladvocacy_medicaidfactsheet_all_states.pdf
4.
Bishop TF, Press MJ, Keyhani S, et al: Acceptance of insurance by psychiatrists and the implications for access to mental health care. JAMA Psychiatry 2014; 71:176–181
5.
Cummings JR, Allen L, Clennon J, et al: Geographic access to specialty mental health care across high- and low-income US communities. JAMA Psychiatry 2017; 74:476–484
6.
Cummings JR, Case BG, Ji X, et al: Availability of youth services in US mental health treatment facilities. Adm Policy Ment Health Ment Health Serv Res 2016; 43:717–727
7.
Cummings JR, Wen H, Ko M, et al: Geography and the Medicaid mental health care infrastructure: implications for health care reform. JAMA Psychiatry 2013; 70:1084–1090
8.
Syed ST, Gerber BS, Sharp LK: Traveling towards disease: transportation barriers to health care access. J Community Health 2013; 38:976–993
9.
Community Health Centers: Opportunities and Challenges of Health Reform. Washington, DC, Kaiser Family Foundation, 2010. https://publichealth.gwu.edu/departments/healthpolicy/DHP_Publications/pub_uploads/dhpPublication_F72DEE65-5056-9D20-3DF07943F1DA545B.pdf
10.
Proser M, Cox L: Health Centers' Role in Addressing the Behavioral Health Needs of the Medically Underserved. Special Topics Issue Brief 8. Washington, DC, National Association of Community Health Centers, 2004. http://www.nachc.org/wp-content/uploads/2015/06/BHReport04.pdf. Accessed Oct 18, 2018
11.
RHC Overview: What Is a Rural Health Clinic? Fremont, MI, National Association of Rural Health Clinics, 2018. https://narhc.org/about-us/rhc-overview/. Accessed Nov 1, 2018
12.
Taylor J: The Primary Care Safety Net: Strained, Transitioning, Critical. Background Paper 79. Washington, DC, National Health Policy Forum, 2010. https://www.nhpf.org/uploads/announcements/BP79_PrimaryCareSafetyNet_09-28-10.pdf. Accessed Oct 18, 2018
13.
HRSA Health Center Program. Rockville, MD, Health Resources and Services Administration, 2017; https://bphc.hrsa.gov/about/healthcenterfactsheet.pdf. Accessed Aug 8, 2018
14.
Improving Commercial Reimbursement for Community Health Centers: Case Studies and Recommendations for New York. Los Angeles, Manatt Health Solutions, 2007. http://www.nachc.org/wp-content/uploads/2015/06/CHCReimbursement.pdf. Accessed Aug 7, 2018
15.
42 US Code § 254b–Health Centers. https://www.law.cornell.edu/uscode/text/42/254b
16.
Wells R, Morrissey JP, Lee IH, et al: Trends in behavioral health care service provision by community health centers, 1998–2007. Psychiatr Serv 2010; 61:759–764
17.
Shin P, Msc JS, Mauery DR: The role of community health centers in providing behavioral health care. J Behav Health Serv Res 2013; 40:488–496
18.
Byrd VLH, Dodd AH, Malsberger R, et al: Assessing the Usability of MAX 2008 Encounter Data for Enrollees in Comprehensive Managed Care. Washington, DC, Mathematica Policy Research, 2012
19.
Byrd VLH, Dodd AH: Assessing the Usability of Encounter Data for Enrollees in Comprehensive Managed Care Across MAX 2007–2009. Washington, DC, Mathematica Policy Research, 2012
20.
Evaluating Encounter Data Completeness: For Researchers Using the Centers for Medicare & Medicaid Services Chronic Condition Data Warehouse (CCW). Falls Church, VA, LewinGroup, 2012
21.
MAX Provider Characteristics. Baltimore, Centers for Medicare and Medicaid Services. https://www.cms.gov/Research-Statistics-Data-and-Systems/Computer-Data-and-Systems/MedicaidDataSourcesGenInfo/MAXPC.html. Accessed Aug 17, 2018
22.
NPI Data Fields. HIPAASpace. https://www.hipaaspace.com/documentation/Databases/NPI.html. Accessed Oct 1, 2018
23.
POS File. Baltimore, Centers for Medicare and Medicaid Services, 2010. https://www.cms.gov/Research-Statistics-Data-and-Systems/Downloadable-Public-Use-Files/Provider-of-Services/POS2010.html. Accessed Feb 1, 2017
24.
Area Health Resources Files. Rockville, MD, Health Resources and Services Administration, Bureau of Health Professions, 2011–2012
25.
HEDIS: Summary Table of Measures, Product Lines and Changes. Washington, DC, National Committee for Quality Assurance, 2016. https://www.ncqa.org/wp-content/uploads/2018/08/20150701_HEDIS_2016_List-_of_Measures.pdf. Accessed Oct 22, 2018
26.
Cummings JR, Ji X, Allen L, et al: Racial and ethnic differences in ADHD treatment quality among Medicaid-enrolled youth. Pediatrics 2017; 139:e20162444
27.
Stein BD, Sorbero MJ, Dalton E, et al: Predictors of adequate depression treatment among Medicaid-enrolled youth. Soc Psychiatry Psychiatr Epidemiol 2013; 48:757–765
28.
HEDIS Measures and Technical Resources. Washington, DC, National Committee for Quality Assurance, 2014. http://www.ncqa.org/hedis-quality-measurement/hedis-measures/hedis-archives. Accessed Nov 1, 2016
29.
Cummings JR, Ji X, Lally C, et al: Racial and ethnic differences in minimally adequate depression care among Medicaid-enrolled youth. J Am Acad Child Adolesc Psychiatry 2019; 58:128–138
30.
Wells KB, Sherbourne C, Schoenbaum M, et al: Impact of disseminating quality improvement programs for depression in managed primary care: a randomized controlled trial. JAMA 2000; 283:212–220
31.
Andersen RM: Revisiting the behavioral model and access to medical care: does it matter? J Health Soc Behav 1995; 36:1–10
32.
Geography Program: Geographic Terms and Concepts. Washington, DC, US Census Bureau. http://www.census.gov/geo/reference/pdfs/glossry2.pdf. Accessed May 20, 2015
33.
Silberman P, Poley S, James K, et al: Tracking Medicaid managed care in rural communities: a fifty-state follow-up. Health Aff 2002; 21:255–263
34.
Geruso M, Layton T: Upcoding: evidence from Medicare on squishy risk adjustment. J Polit Econ (Epub ahead of print June 4, 2019)
35.
Shi L, Lebrun-Harris LA, Chen LR, et al: Preventive counseling services during primary care visits: a comparison of health centers versus other physician offices. J Health Care Poor Underserved 2015; 26:519–535
36.
Shi L, Lebrun LA, Zhu J, et al: Clinical quality performance in US health centers. Health Serv Res 2012; 47:2225–2249
37.
Antidepressant Medication Management. Rockville, MD, Agency for Healthcare Research and Quality, National Quality Measures Clearinghouse, 2013
38.
Birmaher B, Brent D, Bernet W, et al: Practice parameter for the assessment and treatment of children and adolescents with depressive disorders. J Am Acad Child Adolesc Psychiatry 2007; 46:1503–1526
39.
Wolraich M, Brown L, Brown RT, et al: ADHD: clinical practice guideline for the diagnosis, evaluation, and treatment of attention-deficit/hyperactivity disorder in children and adolescents. Pediatrics 2011; 128:1007–1022
40.
Paradise J, Rosenbaum S, Markus A, et al: Community Health Centers: Recent Growth and the Role of the ACA. Washington, DC, Kaiser Family Foundation, 2017. https://www.kff.org/report-section/community-health-centers-recent-growth-and-the-role-of-the-aca-issue-brief. Accessed Aug 8, 2018
41.
FY 2017 Access Increases in Mental Health and Substance Abuse Services (AIMS) Supplemental Funding Technical Assistance. Rockville, MD, Health Resources and Services Administration, 2017. https://bphc.hrsa.gov/programopportunities/fundingopportunities/supplement/
42.
HRSA awards $200 million to health centers nationwide to tackle mental health and fight the opioid overdose crisis. Rockville, MD, Health Resources and Services Administration, 2017; https://www.hhs.gov/about/news/2017/09/14/hrsa-awards-200-million-to-health-centers-nationwide.html. Accessed Aug 8, 2018.
43.
O'Malley KJ, Cook KF, Price MD, et al: Measuring diagnoses: ICD code accuracy. Health Serv Res 2005; 40:1620–1639
Information & Authors
Information
Published In
History
Received: 27 November 2018
Revision received: 10 August 2019
Accepted: 29 October 2019
Published online: 21 January 2020
Published in print: April 01, 2020
Keywords
Authors
Funding Information
This work was supported by grant K01MH095823 from the National Institute of Mental Health.
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.
View Options
View options
PDF/EPUB
View PDF/EPUBLogin 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 loginNot a subscriber?
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.).