Abstract
Youth treated in a residential treatment center (RTC) receive an array of services to address their mental health and behavioral issues, including psychotherapy and medication. Prior studies have shown that youth continue to receive a considerable number of services after discharge from a RTC. This study describes the characteristics of children and youth served in Florida’s Medicaid-funded residential treatment facilities, examines whether family involvement in treatment is associated with discharge to a family-like setting, and assesses continuity of treatment after discharge. Data for this study were obtained from the SIPP Provider Monthly Report Database from January 2005 through December 2011. These data contain information on family involvement during treatment. Post-discharge data were from Medicaid enrollment/claims files and Prepaid Mental Health Plan encounters. Results of this study suggest that family contacts for interventions and treatment planning are associated with a greater likelihood of discharge to a family-like setting. Among youth discharged to a family-like setting, family involvement during treatment was associated with receiving follow-up outpatient care. Results support the continued emphasis on including family members in the treatment of youth with mental health needs to ensure better outcomes and continuity of treatment.
Keywords
Youth treated in a residential treatment center (RTC) receive an array of services to address their mental health and behavioral issues, including individual and family therapy and medication. The role of family in the treatment process has been emphasized in recent years, and youth in programs that engage parents in service planning and delivery tend to have more favorable outcomes (Affronti & Levison-Johnson, 2009; Blader, 2004; Frensch & Cameron, 2002; Hair, 2005; Lakin, Brambila, & Sigda, 2004; Stage, 1999; Walter & Petr, 2008; Wells, Wyatt, & Hobfoll, 1991). For example, parental participation in family therapy with the youth during inpatient treatment improved the odds of discharge to a less restrictive setting (Stage, 1999) and the youth’s ability to maintain treatment gains post discharge (Frensch & Cameron, 2002; Hair, 2005; Walter & Petr, 2008). In addition, when parents were more involved in the treatment process, youth were more likely to complete treatment and had a lower risk of being rehospitalized (Blader, 2004; Sunseri, 2001). Youth in residential treatment whose parents participated in a program combining parent education and non-therapeutic visits were more likely to be reunified with their families after discharge (Carlo, 1993). By contrast, maladaptive behaviors were more likely to return after discharge when parents were not involved in family therapy with the youth during treatment (Jenson & Whittaker, 1987). Lee (2011) found that parental as well as non-family visits during residential treatment were associated with greater permanency following treatment. Thus, extended family and friends may have a role in the treatment process as well.
For many youth, increased family involvement during residential care was associated with improved outcomes following discharge. However, given the chronic nature of mental illness and the severity of symptoms among youth admitted to residential facilities, additional services are usually necessary after the youth returns to the community. Post-discharge services may include educational and vocational services, psychiatric services, and community support services. Numerous studies have found that continuing care is important to improved outcomes (e.g., Jee & Cabana, 2006; Joyce, Fontanella, & Phillips, 2013). For example, connection to community services in a timely way after a RTC stay contributes to a reduced likelihood of returning to a RTC facility (e.g., Armstrong et al., 2008; Robst, Dollard, Rohrer, & Armstrong, 2012; Yampolskaya et al., 2010). Similarly, Axelrod and Wetzler (1989) found that reducing the amount of time between psychiatric hospitalization and aftercare services results in better outcomes. Continuity of care also contributes to lower health care costs, better symptom control, maintenance of treatment gains, fewer hospitalizations, and better quality of life (Adair et al., 2005; Chien, Steinwachs, Lehman, Fahey, & Skinner, 2000; Larzelere et al., 2001; Olfson, Mechanic, Boyer, & Hansell, 1998).
Research suggests, however, that outpatient services after discharge are often underutilized (Asarnow, Aoki, & Elson, 1996) due to barriers related to service availability, transportation, insurance, stigma, and perceptions of treatment benefit and effectiveness (Horwitz et al., 2012; Sareen et al., 2007; Turner & Liew, 2010). Thus, it is important for parents to be involved in the treatment of their children to emphasize the significance of following up with treatment providers as soon as possible after leaving the RTC. However, little is known about the linkage between family involvement during residential treatment and continuity of care after discharge.
This article explores the continuity of outpatient treatment after discharge from a RTC setting and the relationship between family involvement while the youth received RTC services and follow-up care after discharge. In particular, we examined whether family involvement during the RTC episode was associated with the likelihood of being discharged to a family-like setting (e.g., home or a foster home) versus another treatment setting (RTC, therapeutic group home, or therapeutic foster care), and the continuity of outpatient treatment post discharge.
Method
Sample
Florida has three types of Medicaid-funded out-of-home mental health treatment programs: the Statewide Inpatient Psychiatric Program (SIPP), Therapeutic Group Care (TGC), and Therapeutic Foster Care. This study focused on youth in SIPP RTCs that provide short-term inpatient psychiatric services such as crisis intervention; biopsychological and/or psychiatric evaluation; close monitoring by staff; medication management; individual, family, and group therapy; and connections to community services after discharge. In 2009, there were 14 RTC facilities with 414 beds in Florida. The duration of the RTC stay is expected to be 6 months or less. Youth entering a RTC must have a primary Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994) diagnosis other than substance abuse, mental retardation, or autism. The most prevalent diagnoses are mood and affective disorders, disruptive behavior disorders, attention deficit hyperactivity disorder, and anxiety and stress disorders (Armstrong et al., 2010). Before admission, qualified examiners must certify that youth referred for residential treatment is expected to benefit from residential treatment and that no other appropriate treatment is available in less restrictive settings.
Data for this study were obtained from the SIPP Provider Monthly Report Database (SPMRD) from January 2005 through December 2011. Treatment episodes were created for each child based on their admission and discharge dates. These data contain demographic information (age, gender, race), as well as information on family involvement, youth functioning, the county in which the youth resided at the time of admission, and the name of the RTC provider where the youth was treated. Youth in protective custody were excluded from the sample because their family setting is likely to be a foster home, and foster parent (and biological parent) involvement in RTC treatment was very low among youth in protective custody.
Post-discharge data for outpatient services were extracted from Medicaid enrollment and claims files, and encounters from Prepaid Mental Health Plans (PMHPs), the Medicaid mental health carve-out. To be included in the study, youth were required to have Medicaid enrollment for the 30 days after discharge. Youth with Health Maintenance Organization (HMO) enrollment for both physical and mental health care in the 30 days following the RTC episode were excluded due to a lack of encounter data. Eight percent of RTC episodes in the Medicaid claims were for youth with HMO coverage after treatment.
Merging of the three data sets was performed using deterministic matching. Medicaid IDs were not available in the SPMRD in 2005, and data entry errors also resulted in problems for matching data sets. Thus, we used a matching algorithm based on a combination of Medicaid ID, RTC provider, date of birth, admission date, discharge date, gender, and county of residence. Overall, we were able to match 1,505 RTC episodes.
Measures
The SPMRD data contain information on each contact made with a family member, including the relationship of each visitor to the child (mother, father, grandparent, aunt/uncle/cousin, siblings), the purpose of each visit (phone family therapy, in-person family therapy, treatment planning, campus visit, home visit), and the date that each visit occurred. Visits by non-family members only, such as friends or professionals (e.g., guardian ad litem, case manager, therapist), were excluded from the analysis. Contacts that included both family and professionals were retained.
To examine the association between family involvement and youth outcomes, the sample was divided based on the number of family contacts during the treatment episode. For example, to examine the association between maternal contacts and youth outcomes, the sample was divided into episodes that had maternal contacts greater than the sample median and episodes that had maternal contacts less than or equal to the median. Thus, the analysis compared youth outcomes for families with “higher” and “lower” involvement in treatment. This approach was used because family contacts are an imprecise measure of the family’s involvement in treatment. Families with the same number of contacts can differ in how active they were in the treatment process. Some may have been highly involved whereas others attended a meeting because they were told to do so by the RTC staff. Thus, it was decided that family contacts were not sufficiently precise to treat as a continuous measure of family involvement. However, families with contacts above the median were assumed to be more involved in the youth’s treatment than families with contacts below the median.
Discharge to a family-like setting was defined as a dichotomous variable with youth discharged to biological or adoptive parents, or to other family members coded 1. Youth discharged to another treatment setting such as a different RTC, a therapeutic group home, or a specialized therapeutic foster care home were coded 0.
Continuity of treatment was measured based on whether the youth received outpatient services within 30 days of discharge. We utilized a short time frame to focus on continuing treatment promptly after being discharged from residential treatment. A limited set of services was assessed that represented outpatient mental health services provided to youth [day treatment, outpatient, Mental Health–Targeted Case Management (MH-TCM), community-based services, and school-based services]. Inpatient and emergency treatment were excluded because they likely represent a response to a crisis and not a planned continuation of care. Youth discharged from the RTC to another treatment setting (e.g., another RTC, TGC, Specialized Therapeutic Foster Care [STFC], or Group Care with Behavioral Health Overlay Services [BHOS]) were excluded from the analysis of treatment continuity because all such youth continued treatment.
Analytical Method
The primary outcomes were whether the youth was discharged to a family-like setting and among youth discharged to a family-like setting, whether they received outpatient services in the first 30 days after discharge. Logistic regressions were estimated to examine characteristics related to post-discharge placement and continuity of treatment:
where i denotes youth. Both dependent variables were dichotomous variables that required a logistic distribution.
The independent variables (X) included dichotomous variables denoting family involvement above the sample median (reference: involvement ≤ median), age (≥14; reference: <14), gender (female; reference: male), race (Black, Asian, Hispanic, multi-racial; reference: White), whether the RTC is in the same county in which the youth resides (reference: RTC located in a different county), and variables denoting each RTC. In addition, a continuous variable denoting the duration of the treatment episode (in months) was included in the specification.
This study was approved by the Institutional Review Board at the University of South Florida.
Results
Table 1 presents descriptive statistics for the number of family contacts with the child as well as child characteristics. The youth on average were 13 years old and 59% were male. The sample was composed of 59% White, 26% Black, 9% Hispanic, and 6% Other racial groups. The mother was the primary family member involved in the treatment of the youth. Mothers averaged more than 17 contacts with the youth during an episode, in comparison with 5 contacts by fathers with the youth. The majority of contacts were for therapeutic interventions (e.g., family therapy) and treatment planning. More than 80% of all family contacts were for intervention or treatment planning reasons, and 13% involved therapeutic home passes. Forty-two percent of intervention contacts (7.71 out of 18.23) occurred over the phone.
Child Characteristics and Average Contacts per Episode.
Note. RTC = residential treatment center.
Discharge rates to family-like settings are provided in Table 2 with the sample divided based on whether family contact was above or below the median. Statistical significance is assessed based on the logistic regression results. Only involvement by the mother was associated with a greater likelihood that a youth is discharged to a family-like setting (odds ratio [OR] = 2.43; 95% confidence interval [CI] = [1.79, 3.29]). The reasons for the contacts were more strongly associated with post-discharge placement. The number of treatment-related contacts was positively associated with discharge to a family-like setting, including in-person interventions (OR = 2.23, 95% CI = [1.64, 3.05]), phone interventions (OR = 1.58, 95% CI = [1.15, 2.18]), and treatment planning contacts (OR = 1.76, 95% CI = [1.28, 2.42]). Therapeutic home passes were associated with discharge to the family setting (OR = 2.42, 95% CI = [1.78, 3.29]). Non-treatment-related visits by family members (e.g., campus visits) were not associated with family placement after discharge.
Proportion Discharged to Family-Like Setting by Level of Family Involvement.
Note. The p values are from logistic regressions where being discharged to a family-like setting is regressed on a variable denoting whether contacts were above or below the median. The regression controlled for youth age, race, gender, whether the youth resided in the same county as the residential treatment center (RTC), the duration of the treatment episode, and the provider.
Table 3 examines the relationship between family involvement and follow-up outpatient mental health treatment among youth discharged to a family-like setting. Outpatient mental health treatment was more common among youth with greater involvement in treatment by the mother (OR = 1.33, 95% CI = [1.00, 1.77]). Treatment-related contacts during residential treatment were associated with receiving outpatient services within 30 days of discharge. In particular, therapeutic home passes (OR = 1.50, 95% CI = [1.11, 2.03]) and treatment planning contacts (OR = 1.60, 95% CI = [1.19, 2.15]) were associated with follow-up outpatient treatment.
Proportion Receiving Mental Health Services Within 30 Days by Level of Family Involvement Among Youth Discharged to Family-Like Setting.
Note. Mental health services include day treatment, outpatient, targeted case management, and community-based and school-based services. Youth receiving out-of-home services (residential treatment center [RTC], Therapeutic Group Care [TGC], Specialized Therapeutic Foster Care [STFC], or Behavioral Health Overlay Services [BHOS]) were excluded. The p values are from logistic regressions where the receipt of outpatient-based services is regressed on a variable denoting whether contacts were above or below the median. The regression controlled for youth age, race, gender, whether the youth resided in the same county as the RTC, the duration of the treatment episode, and the provider.
Discussion
This study examined whether family involvement during residential treatment was associated with a greater likelihood of being discharged to a family-like setting and continuity of outpatient mental health treatment after a RTC episode. Family involvement was associated with an increased probability of discharge to a family-like setting instead of another treatment setting, such as a different RTC. Such results are consistent with prior research that found parental involvement was associated with discharge to less restrictive settings (Carlo, 1993; Stage, 1999). However, the results were specific to involvement by the mother and we did not find that involvement by extended family was important to discharge placement.
Contacts due to therapeutic home passes, in person or phone intervention, and treatment planning were associated with a greater likelihood of the youth being discharged to a family-like setting. However, despite being statistically significant, the effects are relatively larger for therapeutic home passes (76% vs. 84%) and in-person interventions (76% vs. 84%) compared with phone interventions (79% vs. 81%) and treatment planning contacts (79% vs. 81%).
As noted by Birmaher and Brent (2007), parental involvement in treatment enables providers to educate parents about the importance of continuing treatment. Consistent with this hypothesis, greater involvement through therapeutic home passes and treatment planning was associated with a greater likelihood of follow-up outpatient mental health treatment. However, the results were mixed as intervention contacts (in-person and phone) were not associated with treatment continuity. The findings suggest that providers and case managers might place greater emphasis on the importance of treatment continuity during intervention contacts.
Therapeutic home passes are associated with discharge to a family-like setting and the receipt of follow-up care. Home passes are partly treatment-related as one of the goals is to begin the transition process for the youth to return home on a permanent basis. Such contacts are also likely to have some characteristics consistent with non-treatment-related campus visits. Assessing the relationship between home passes and discharge placement is also more challenging as home passes are more likely when the youth is doing well and as a result is more likely to be discharged home.
These findings have several implications for public policy. First, as found in other studies, family involvement in treatment is important to youth outcomes. RTC providers should encourage family involvement in treatment and treatment planning. Informed parents are more likely to understand the importance of follow-up care for the youth to retain gains made while in the RTC and hopefully promote continued improvement. RTC providers in Florida currently require that families participate in therapy, although based on the results, compliance is far from complete. Thus, an important question is how family involvement of those youth with lower than median involvement can be increased. The median number of contacts by fathers was 0, and given the average 6-month treatment duration, the median intervention (nine in-person and four phone) contacts were modest.
There are several challenges to and opportunities for increasing family involvement. From the parent’s perspective, facilities needed to implement a number of changes to encourage family involvement. Residential facilities need to include parents when developing the child’s goals, increase communication, provide frequent updates on the child’s status, make meeting times more convenient, increase the focus on family issues in treatment, organize support groups for parents, and work to build relationships with parents (Demmitt & Joanning, 1998; Kruzich, Jivanjee, Robinson, & Friesen, 2003). Parents (79.4%) also report that residential facilities impose restrictions on contacts with their children that included limiting telephone calls, facility visits by parents, and therapeutic home passes for youth (Robinson, Kruzich, Friesen, Jivanjee, & Pullmann, 2005). Surveys with family members also indicated that transportation issues, such as distance from providers, transportation costs, and lack of transportation, were an impediment to family involvement (Kruzich et al., 2003). Phone contacts may be an important vehicle for families who live at a distance from the RTC where their child is receiving treatment. In addition to face-to-face contacts, efforts to expand technological approaches to family involvement should be explored.
Another important question is to identify why there are still some youth who are not making connections to Medicaid-funded community services after RTC discharge. RTC services are merely a step in the treatment process, and we would expect to see continued outpatient treatment after the youth is discharged. Indeed, previous studies (e.g., Armstrong et al., 2008; Yampolskaya et al., 2010) found that failure to get connected to community services in a timely way increases the likelihood of return to RTC services. Prompt connections are optimal because the first month post discharge is when youth have the highest return rates to RTC treatment (Robst, Armstrong, & Dollard, 2012).
As with any study that uses administrative data, there are numerous limitations to this analysis. First, the measure of family involvement is based on contacts. Contacts are an approximation of involvement as engagement in treatment can differ substantially among parents with the same number of contacts. Second, we do not know the family status for the youth in this study. Thus, we do not know whether the youth lives in a single parent household, whether the youth has extended family members who reside near the RTC or the parents, or the level of involvement by the mother or father in the youth’s life. Third, it is important to note that the analysis was descriptive and did not necessarily indicate a causal relationship. The causal relationship between family involvement and discharge placement would be difficult to examine using administrative data because it is unclear why families were or were not involved. For example, families may have been more involved in treatment when the child was more likely to return home after discharge. In addition, it is not clear whether family involvement caused the discharge placement or continuity of treatment, or whether there was an omitted variable correlated with both family involvement and the outcomes.
Future work can address a number of remaining questions. In addition to addressing some of the limitations in this study, future work could examine differences in the importance of family involvement across gender, race, age, and diagnostic groups. Replicating the study with a single data set that did not require deterministic matching would be useful. Exploration of the quality of family involvement is a key factor to explore further.
Conclusion
This study examined the role of family involvement in the treatment of youth who receive residential treatment services. Family involvement was associated with an increased likelihood of being discharged to a family-like setting, and among such youth, a higher likelihood of receiving follow-up outpatient mental health treatment. Overall, the results support the continued emphasis on including family in the treatment of youth with mental health needs.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Financial support for this project was provided by the Florida Agency for Health Care Administration under contract MED134.