Continuity of care following psychiatric hospitalization is crucial to successful outcomes for children and adolescents. Studies of both adult and pediatric populations have demonstrated the benefits of timely aftercare, including decreased suicidal ideation (
1), reduced readmissions (
2,
3), and improved medication adherence (
4). As the length of hospital stays has markedly decreased, discharge planning and linkage to timely and appropriate aftercare are increasingly necessary to monitor treatment response, ensure continued stabilization, maintain and extend health gains, and prevent relapse or readmission after psychiatric hospitalization.
Despite the importance of aftercare, little is known about rates and predictors of follow-up within seven and 30 days of discharge after psychiatric hospitalization, which is considered an indicator of quality of care and which states employ as a performance measure (
5). Most studies of follow-up after psychiatric hospitalization have focused on follow-up periods ranging from two months to six years rather than the week or month immediately following hospitalization (
6). Moreover, existing studies are difficult to compare and do not provide definitive conclusions because of varying definitions of aftercare, study populations, and predictor variables. Previous studies have also been limited by the fact that the study samples were often diagnostically heterogeneous and were derived from a single psychiatric hospital and the studies focused solely on individual demographic and clinical variables rather than on system and community characteristics.
Methodological challenges aside, existing studies found several factors to be associated with receipt of aftercare, including younger age, higher socioeconomic status, presence of a biological parent or grandparent in the home, prior outpatient service use before hospitalization, and psychiatric comorbidity (
6–
8). In contrast, youths who are members of specific racial and ethnic minority groups, who live in rural areas, and who have a history of self-harm, suicide attempts, substance use disorders, and multiple previous hospitalizations are less likely to receive care after psychiatric hospitalization (
6,
7,
9).
This study examined factors associated with timely aftercare services for Medicaid-enrolled children and adolescents with mood disorders in order to identify patient subgroups at high risk of discontinuity of care and inform quality improvement initiatives. Mood disorders, such as depressive and bipolar disorders, are prominently associated with pediatric psychiatric admissions, particularly in community hospital settings (
10). Admission of children with mood disorders is often associated with suicide risk, which is especially high in the period immediately after hospitalization (
11). To our knowledge, no studies have examined predictors of timely follow-up after inpatient psychiatric hospitalization for a Medicaid-enrolled pediatric population, and most prior studies have focused solely on the demographic and clinical characteristics of individual patients as predictors of aftercare. Because associations have been observed among hospital- and community-level factors, access to care, and mental health outcomes (
12–
14), we examined multiple patient-, hospital-, and community-level factors by using population-based data that have been merged from multiple sources to enhance validity.
Methods
Study Design and Data Sources
A retrospective, longitudinal cohort design was used to examine associations between patient-, hospital-, and community-level factors and timely follow-up care. Data were drawn from three sources: Medicaid Analytic eXtract (MAX) data (
15), the American Hospital Association (AHA) annual survey (
16), and the Area Resource File (ARF) (
17). Medicaid data obtained from the Centers for Medicare and Medicaid Services were the primary source of patient-level data, including demographic characteristics, diagnoses, dates of services, procedures, and providers (
15). The AHA annual survey provided information on hospital-level data, including organizational structure, facilities and services, and utilization (
16). The Medicaid national provider identifier on the hospital discharge claims was used to link the Medicaid patient-level data with the hospital data. The 2010 ARF provided information on county-level data, including sociodemographic, economic, and health care system characteristics (
17). County of residence was used to link the Medicaid patient-level data with community-level data from the ARF. All study procedures were approved by the Ohio State Institutional Review Board.
Study Population
We identified youths (ages six to 17) from four states (California, Florida, Maryland, and Ohio) who were admitted to psychiatric hospitals with a primary diagnosis of mood disorder (ICD-9-CM codes 296xx and 311) between July 2009 and November 2010 and who were continuously enrolled in Medicaid for the 30-day period after hospitalization and the 180-day period prior to hospitalization (N=10,095). Youths who were readmitted within the 30-day follow-up period (N=984) or those who had a hospital stay of greater than 30 days (N=35) were excluded. Also excluded were youths with missing hospital- (N=1,243) and county-level data (N=7). The final analytic sample was 7,826 youths.
Outcome Variables
The study outcome measures of one or more mental health visits within seven days and 30 days of hospital discharge were based on the Health Plan Employer Data and Information Set quality-of-care guidelines for follow-up after psychiatric hospitalization (
5). An outpatient follow-up visit was defined as any Medicaid-reimbursed behavioral health visit with a primary mental health diagnosis (
ICD-9-CM codes 290–319), including visits for psychotherapy or pharmacotherapy, partial hospitalization, rehabilitation, and other community-based services, such as case management.
Predictor Variables
Patient-level factors included age at hospital discharge (six to 11 versus 12–17), sex, race-ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, and other), Medicaid-eligibility category (poverty, disability, and foster care), length of stay (one to seven versus more than seven days), and primary diagnosis (depressive disorder, bipolar disorder, and other mood disorder). The following variables were abstracted from Medicaid claims from the six months prior to hospital admission: presence or absence of prior mental health visits, substance use disorders (ICD-9-CM codes 291–292 and 303–305), chronic medical conditions, and number of psychiatric comorbidities (zero, one, and two or more).
Hospital-level factors included the total number of beds (small, <200; medium, 200–399; and large, ≥400), ownership (public, private nonprofit, and private for profit), percentage of the total annual discharges involving patients enrolled in Medicaid (low, 0%−18.9%; medium, 19%−25.9%; and high, ≥26%), medical resident teaching status (major teaching, minor teaching, and nonteaching), and type of hospital (general and psychiatric).
Community-level factors included health care resources, such as the number of providers per 100,000 youths (child psychiatrists [none; low, 1.0–6.9; and high, ≥7.0], psychologists [low, 0.0–29.9; medium, 30.0–84.9; and high, ≥85.0], and social workers [none; low, <650; and high, ≥650]) and presence or absence of community mental health centers; economic factors, such as the annual per capita income (low, <$30,000; medium, $30,000–$36,999; and high, ≥$37,000), percentage of county population living in poverty (low, <14.0%; medium, 14.0%−17.9%; and high, ≥18.0%), percentage of county population unemployed (low, <10%; medium, 10.0%−11.9%; and high, ≥12.0%); and the area of residence (metropolitan and nonmetropolitan).
Statistical Analyses
Rates of follow-up at seven and 30 days were calculated across all patients and stratified by each independent variable. Multivariable random-effects logistic regression models were used to examine the association between patient-level, hospital-level, and community-level factors and timely follow-up care. Separate models were performed for each outcome measure. Random-effects logistic regression is the appropriate analyses for multilevel or hierarchical data because it takes into account the nesting of individuals within hospitals and generates unbiased estimates as well as correct standard errors. We considered variables to be statistically significant at p=.05 (two-tailed) and of clinical and policy significance if odds ratios were above 1.20 or below .80. All analyses were performed using STATA, version 13 (
18).
Discussion
In this multistate study of publicly insured youths, roughly half received any outpatient mental health care during the first week following discharge from an inpatient psychiatric hospital and about two-thirds received mental health care during the first month. Although these findings mirror national rates of follow-up for Medicaid populations (45.9% for seven days and 65.4% for 30 days) and trends suggest that rates of follow-up have increased (
19), there is considerable room for improvement, particularly in light of the increased risk of suicide during the posthospitalization transition period.
Our findings confirm that receipt of timely follow-up care is influenced by multiple patient, hospital, and community characteristics. Consistent with previous research (
8,
12,
20), we found that prior outpatient treatment was the strongest predictor of linkage to timely outpatient care. The odds of receiving outpatient mental health care in the first month following discharge were over five times higher for youths who were previously engaged in treatment. Although not surprising, these findings underscore the importance of having a usual source of care as well as the likely value of established relationships in improving transition from one level of care to another. Patients lacking a previous connection with a mental health provider may require more proactive clinical efforts to explore and address structural and attitudinal barriers to care after a hospitalization, engender hope and positive expectations, and better coordinate the transition from inpatient to outpatient care. The value of personal introductions to providers of the next level of care, peer-support models, and other bridging interventions deserves additional application and study.
Psychiatric comorbidity was positively associated with receipt of timely follow-up care, consistent with studies that associated psychiatric comorbidity with greater use and more persistent use of mental health services (
8,
21). Psychiatric comorbidity is associated with greater symptom severity, functional impairment, and perceived illness burden, and such markers of clinical need have been shown to be potent predictors of help seeking among both adults and children (
22). Notably, although consistent with prior research (
7,
20), the added morbidity of co-occurring substance use disorders actually decreased the likelihood of timely follow-up care for psychiatrically hospitalized youths with mood disorders. Substance use disorders are among the most common comorbidities among patients with mood disorders, complicate treatment efforts, and are associated with negative outcomes (
23), so robust efforts to engage and retain such patients in active treatment are of particular relevance. Psychoeducation efforts combined with evidence-based methods of patient and family engagement are likely applicable to efforts to engage patients with substance use disorders. Like children classified as disabled, children in foster care are more likely to be perceived as experiencing a greater overall burden of illness. In addition, they are likely to receive better monitoring and supervision by caseworkers.
In terms of demographic characteristics, race and age appeared to influence the receipt of aftercare following psychiatric hospitalization. African-American youths were less likely to receive follow-up care after psychiatric hospitalization compared with other racial or ethnic groups. Racial and ethnic disparities in access to and quality of mental health care are well documented in the research literature, with potential explanatory factors including structural barriers, such as geographic location of treatment resources and the availability of transportation, as well as attitudinal barriers, such as mistrust of the health care system in general, heightened sense of mental health–related stigma, and other cultural beliefs, including a belief that mental health is more related to spiritual health than to biology (
24,
25). Finally, members of racial-ethnic minority groups may also have more negative perceptions of treatment and believe that mental health treatment is ineffective (
26). Consequently, culturally informed educational and engagement efforts deserve consideration in vulnerable populations.
The finding that adolescents had significantly lower rates of follow-up compared with younger children is consistent with the results of Goldston and others (
8). In contrast to younger children, who are typically more comfortable being reliant on parents, adolescents often strive for greater independence and input into the decision to seek care. Moreover, adolescents tend to have more negative attitudes about treatment compared with younger children and may perceive help seeking as a threat to their autonomy and sense of self-reliance (
27).
Treatment in a specialized psychiatric hospital increased the odds of receiving timely follow-up care. Specialized psychiatric hospitals may be more likely to offer outpatient psychiatric services compared with general hospitals, and a subanalysis from the AHA survey suggests that the specialized psychiatric hospitals included in this study were more likely than the general hospitals surveyed to offer outpatient psychiatric services (results not shown). Admission to hospitals with fewer beds was also positively associated with receipt of timely follow-up care, and these findings are consistent with some (
12), but not all, previous research (
28,
29). Smaller hospitals may have more time to devote to discharge planning and cultivate relationships with their patients and families, but small hospital size might also reflect specialty psychiatric status as well.
Consistent with prior medical research that suggests that teaching hospitals are associated with better quality of care (
28,
30), we found that treatment in a teaching hospital was associated with a higher likelihood of timely follow-up care. This finding may have to do with greater resources and more intensive continuing medical education for physicians, nurses, and social workers at teaching hospitals.
In contrast, treatment in hospitals with a higher percentage of Medicaid patients was negatively associated with receipt of timely follow-up care. Our results generally concur with those of Goldman and others (
31), who found that nonteaching hospitals with high percentages of Medicaid patients have lower adherence to quality measures. In comparison with other types of insurance, Medicaid limits the types of outpatient services available to beneficiaries, who often have longer wait times for appointments because of fewer provider options.
Our results with regard to the availability of specialty behavioral health professionals were mixed. Consistent with findings of Hendryx and colleagues (
32), we found that youths who lived in counties with more child and adolescent psychiatrists per capita were more likely to receive outpatient follow-up care in seven and 30 days, but, surprisingly, the number of per capita psychologists and social workers did not appear to be associated with the likelihood of receiving aftercare. This may reflect greater shortages of child and adolescent psychiatrists relative to other behavioral health professionals but may also suggest that psychiatrists play a special role in ensuring appropriate access to and coordination of care.
Areas with high rates of unemployment were also associated with a lower likelihood of receiving timely follow-up care. These areas tend to have high corresponding rates of poverty, homelessness, substance use, and residential turnover and fewer mental health resources (
25). In these communities, low-income populations insured through Medicaid often must rely on an inadequate supply of safety-net providers.
This study had several potential limitations. First, analyses were limited to Medicaid-enrolled children and adolescents from four states and may not be generalizable to other state Medicaid programs and privately insured or uninsured populations. However, the rates of follow-up are strikingly similar to rates found in the United States Medicaid population, suggesting that our results may be broadly relevant for youths enrolled in Medicaid anywhere in the United States (
19). Second, our use of claims data precluded an examination of other important factors that may affect receipt of follow-up care among children and adolescents, such as use of psychotropic medication, family functioning and support, caretakers’ perception of burden of care, and the intervention strategies employed by hospital staff to link patients to outpatient care. Third, diagnoses were based on clinical judgment and were derived from claims, and they were not subject to expert validation through standardized or structured assessment. Fourth, nonresponse on the AHA survey resulted in excluding 13% of youths for missing hospital data, which may potentially lead to selection bias. Finally, the results may not generalize to the 10% of youths who were readmitted within 30 days or had extended lengths of stay.