Video Abstract
African American adolescents appear to be the most at risk for asthma morbidity and mortality even compared with other minority groups, yet there are few successful interventions for this population that are used to target poorly controlled asthma.
African American adolescents (age 12–16 years) with moderate-to-severe persistent asthma and ≥1 inpatient hospitalization or ≥2 emergency department visits in 12 months were randomly assigned to Multisystemic Therapy–Health Care or an attention control group (N = 167). Multisystemic Therapy–Health Care is a 6-month home- and community-based treatment that has been shown to improve illness management and health outcomes in high-risk adolescents by addressing the unique barriers for each individual family with cognitive behavioral interventions. The attention control condition was weekly family supportive counseling, which was also provided for 6 months in the home. The primary outcome was lung function (forced expiratory volume in 1 second [FEV1]) measured over 12 months of follow-up.
Linear mixed-effects models revealed that compared with adolescents in the comparison group, adolescents in the treatment group had significantly greater improvements in FEV1 secondary outcomes of adherence to controller medication, and the frequency of asthma symptoms. Adolescents in the treatment group had greater reductions in hospitalizations, but there were no differences in reductions in emergency department visits.
A comprehensive family- and community-based treatment significantly improved FEV1, medication adherence, asthma symptom frequency, and inpatient hospitalizations in African American adolescents with poorly controlled asthma. Further evaluation in effectiveness and implementation trials is warranted.
African American adolescents are at high risk for asthma morbidity and mortality. Most interventions for high-risk adolescents with poor asthma outcomes are focused on education and case management and do not reveal improvements in objective health outcomes.
A home-based social-ecological behavioral treatment improved lung function, inpatient hospitalizations, adherence to daily controller medications, and asthma symptoms over 12 months of follow-up.
Pediatric asthma causes high rates of functional morbidity, such as school absences, frequent doctor visits, and hospitalizations,1 and inner-city adolescents are at the highest risk for asthma morbidity and mortality.2,3 African American youth have greater odds of an asthma diagnosis even compared with other minority groups, and the odds increase with each decrease in socioeconomic status.4 African American adolescents have higher emergency department (ED) visits and hospitalizations for asthma than white adolescents,3,5,–8 and recent data reveal that other ethnic minorities, such as Hispanic children, do not have such disparities.9 Differences in illness management may account for some of these disparities.10,11 There is a critical need to develop successful, culturally tailored illness management interventions for African American adolescents with moderate-to-severe persistent asthma.12 Yet, studies in which researchers test interventions with African American adolescents with asthma have either included nonminority adolescents,13 included youth with mild or intermittent asthma,14 or were typically focused on younger, school-aged children. Significant concerns have been expressed regarding the underrepresentation of African American participants in clinical studies.15 A recent review16 revealed that although there are several school-based interventions available for minority adolescents, there are only a few interventions for African American adolescents with moderate-to-severe persistent asthma, and those few interventions had minimal effects on health outcomes.
Researchers in meta-analyses of pediatric illness management interventions have concluded that intensive interventions combining behavioral and educational strategies are necessary to improve health.16,17 One such intervention is Multisystemic Therapy (MST),18 an intensive, family- and community-based intervention approach that originated in mental health settings and is grounded in the social-ecological model.19 This model depicts problem behavior, such as poor illness management, as a reciprocal interchange between the individual and nested concentric system structures that mutually influence one another (eg, individual child, family dynamics, family–health care provider relations, and family–school relations). Multisystemic Therapy–Health Care (MST-HC) is used to address barriers within and between these systems (Fig 1). As such, MST-HC therapists deliver behavioral interventions for the multisystem factors that are common in high-risk youth, such as family disorganization, a lack of routine for asthma care, low parental involvement in and monitoring of asthma care, inadequate social support from extended family members and school personnel, and poor engagement with asthma care and asthma treatment providers in addition to inadequate asthma knowledge and deficits in medication device skills. Critical service delivery characteristics of MST-HC include the following: (1) low caseloads (6–8 families per clinician) that allow intensive services to be provided to each family (2–6 hours per week), (2) the delivery of services in community settings (eg, home, school, neighborhood center, or physician office), (3) time-limited duration of treatment (6 months), (4) availability of therapists 24 hours per day and 7 days per week, and (5) the provision of comprehensive services. MST-HC, when delivered over 6 months, has been shown to improve illness management, health outcomes, and health care use among urban adolescents with other chronic conditions.20,–26
Our purpose in the current study is to test the efficacy of MST-HC compared with in-home family support (FS) (comparison condition) on forced expiratory volume in 1 second (FEV1) (lung function as the primary outcome) and on medication adherence, symptom severity, and health care use (secondary outcomes) over 12 months of follow-up in a prospective randomized controlled trial with African American adolescents with poorly controlled moderate-to-severe persistent asthma. We hypothesized that high-risk African American adolescents with poorly controlled asthma would have greater improvements in lung function and the secondary outcomes of medication adherence, asthma symptoms, inpatient hospitalizations, and ED visits when receiving MST-HC compared with FS over 12 months of follow-up.
Methods
Study methods and the interim analysis have been previously reported,27,–29 as have descriptive baseline results.29,–31
Eligibility criteria included the following: age 12 to 16 years, medical record confirmation of diagnosis of moderate-to-severe persistent asthma (to ensure that participants would be prescribed a daily asthma controller medication based on national standards of care32), self-identified as African American, residing in a home setting with a caregiver who was willing to participate in treatment, and living within a 30-mile radius of the university. Adolescents had to be at high risk for poorly controlled asthma as evidenced by having at least 1 asthma-related hospitalization or at least 2 asthma-related ED visits in the last 12 months that were confirmed with medical records. Exclusion criteria included thought disorder or cognitive limitations that would interfere with assent or data collection, having a chronic health condition other than primary obesity, or being unable to complete assessments or interventions in English. The local university institutional review board approved all study procedures (063908B3F).
Enrollment occurred from January 2009 to June 2012, and the recruitment is described in the Consolidated Standards of Reporting Trials diagram (Fig 2). Study researchers contacted 196 participants and completed home-based consent and data collections with 170 families (7% refused, and 12 families could not be reached). Randomization was stratified on the basis of (1) severity of urgent care use as indicated by the number of recent hospitalizations or ED visits (3 or more hospitalizations or ED visits in the previous 12 months versus 0–2 hospitalizations and/or ED visits) and (2) receipt of asthma specialty care at a hospital-based multidisciplinary asthma specialty clinic in the last 12 months. Families were randomly assigned to MST-HC (N = 84) or FS (N = 86). Two families that were randomly assigned to FS were removed from the study because of safety concerns that developed during treatment that interfered with the delivery of a home-based intervention; and another family was removed when it was discovered that they did not meet study eligibility criteria. Thus, the final analyzed sample was 167 (n = 84 in MST-HC; n = 83 in FS).
Data Collection and Study Outcomes
Trained and blinded data collectors completed baseline and 7- and 12-months postbaseline assessments in the families’ homes. Families were compensated $50 for participating in each data collection assessment. The primary outcome was FEV1 (baseline, 7 months, and 12 months). Pulmonary function was assessed by using forced expiratory maneuvers that were obtained by using a portable calibrated recording spirometer (KoKo) at the time of the research interview. Secondary outcomes were medication adherence and the frequency and/or severity of asthma symptoms (baseline, 7 months, and 12 months), inpatient hospitalizations, and ED visits from 12 months before enrollment to 12 months after enrollment.
Adherence to daily controller medication was assessed by using 2 validated assessment tools: (1) the Medication Adherence subscale of the Family Asthma Management System Scale (FAMSS)33 and (2) the Daily Phone Diary (DPD).34 The FAMSS is a clinical interview that is completed conjointly with caregivers and teenagers; the Medication Adherence subscale has been shown to be related to objective measures of medication adherence and asthma morbidity.33 This scale is scored on the basis of both adolescent and caregiver responses, and scores range from 1 to 9. Scale developers trained study raters, and interrater reliability was high (intraclass correlation coefficient = 0.933). The DPD,34 a cued recall procedure, is used to collect information about participants’ activities, companions, and moods during the previous 24 hours. For the current study, information from 2 DPD assessments within a 2-week period was combined to determine the use of controller medication in either of the 24-hour periods on a 3-point scale (2 = participant took controller medication on both days, 1 = participant took controller medication on at least 1 of the 2 days, and 0 = participant did not take controller medication on either day). Adolescents completed a standard measure of asthma symptom frequency and severity.13
The number of asthma-related ED visits and inpatient hospitalizations was obtained from medical records for the 12 months before baseline data collection to 12 months after baseline. In addition to searching records from the local institution, we queried caregivers at data collection visits, by phone at 3 and 9 months, and at treatment sessions. When a hospitalization or ED visit outside the local institution was reported via any of these sources, the research team obtained a release and collected medical records. Only reports that were corroborated by medical records were included in the analysis.
Intervention Condition: MST-HC
Adolescents who were assigned to the intervention condition received MST-HC as adapted for the treatment of poor self-management in adolescents with asthma.26,27 Although MST is well operationalized with fidelity to specific, measurable treatment principles,35 it is not a typical, manualized 1-size-fits-all intervention in which the therapist follows a set of prearranged tasks in a time-limited sequence. Instead, MST therapists begin with an initial motivational session to promote engagement followed by a functional assessment of asthma management based on the social-ecological model (Fig 1). This assessment includes interviews and in vivo observations in the home and community to identify setting events, antecedents, and consequences of inadequate asthma management within and between the child, family, school, community, and health care systems. On the basis of this assessment, the MST therapist chooses from a menu of evidence-based interventions (eg, skills training, behavioral therapy, family therapy, and communication training with school and medical staff) that can be used to best treat the identified problem behaviors (eg, underuse of preventive medications, poor identification of asthma triggers, not carrying rescue medications at all times, and low parental monitoring) and their particular causes in each family (Table 1). Goal achievement (eg, adherence to asthma action plan, improved medication adherence, and symptom reduction) and barriers are continually assessed, and intervention plans are modified accordingly. Sessions can take place several times per week initially and then only weekly once an adolescent’s asthma management has improved. Based on our previous MST-HC trials, treatment was planned to last for up to 6 months after baseline. The mean length of treatment from the first session to termination, excluding dropouts, was 5.14 months (SD = 1.25), and the mean number of sessions was 27.09 (SD = 12.03; range 4–62). MST-HC was provided by 4 masters-level therapists with varied backgrounds (1 psychologist and 3 social workers; 1 white and 3 African American). There were no study-related adverse events in the treatment group.
Menu of Evidence-Based MST-HC Interventions
| Individual Child | Family | Community |
|---|---|---|
| Education and asthma device skills training | Education and asthma device skills training | Education of school personnel |
| Emotion regulation | Organization and planning skills | Family–school communication training |
| Stress management | Problem-solving skills | Connecting with school personnel for support |
| Motivational interventions | Parent training | Linkage to medical care and prescriptions |
| Problem-solving skills | Marriage counseling | Family–provider communication training |
| Coping skills | Stress management | Collaborative development of asthma action plan |
| Motivational interventions | Parent–child communication training | Management of environmental triggers |
| Mental health treatment | Social support interventions | Connection with prosocial peers |
| Substance abuse treatment or smoking cessation | Mental health or substance abuse treatment | Case management for resource access |
| Individual Child | Family | Community |
|---|---|---|
| Education and asthma device skills training | Education and asthma device skills training | Education of school personnel |
| Emotion regulation | Organization and planning skills | Family–school communication training |
| Stress management | Problem-solving skills | Connecting with school personnel for support |
| Motivational interventions | Parent training | Linkage to medical care and prescriptions |
| Problem-solving skills | Marriage counseling | Family–provider communication training |
| Coping skills | Stress management | Collaborative development of asthma action plan |
| Motivational interventions | Parent–child communication training | Management of environmental triggers |
| Mental health treatment | Social support interventions | Connection with prosocial peers |
| Substance abuse treatment or smoking cessation | Mental health or substance abuse treatment | Case management for resource access |
Attention Control Condition: FS
Families that were randomly assigned to the comparison condition received weekly, home-based, client-centered, nondirective supportive family counseling to match the intervention condition for access to services and to control for non–MST-specific treatment factors of attention by allowing the family to discuss any asthma symptoms and providing empathic support and referrals for any barriers that the families identified. The FS intervention was 6 months in length and, hence, was matched to MST-HC for length of treatment. Because the MST-HC session dose is flexible, matching the comparison condition for dose was not possible. However, an approximate dose of a weekly 45-minute session that is consistent with traditional outpatient therapy approaches (and therefore with what would be provided in a real-world setting) was chosen. The mean length of treatment, excluding dropouts, was 4.20 months (SD = 1.78), and the mean number of sessions was 11.03 (SD = 5.74; range 3–24). FS was provided by 6 masters-level psychologists and social workers and 1 bachelor’s-level psychology graduate student (5 African American and 2 white). There were no study-related adverse events in the comparison group.
Treatment Fidelity
Quality assurance protocols included an initial 5-day training, a review of the MST-HC manual for asthma,26 weekly on-site clinical supervision from a doctorate-level supervisor, a weekly phone consultation with an MST expert, and quarterly booster trainings. Quality assurance protocols also included feedback on measures of therapist and supervisor fidelity to MST-HC procedures.36,37 All sessions were audio recorded, and independent coders rated 1 randomly selected session per month per therapist using the MST fidelity code scheme38 adapted for MST-HC.25 For the FS condition, quality assurance protocols included a detailed manual, an initial 3-day training, and a minimum of biweekly on-site clinical supervision from a doctorate-level supervisor, who reviewed 1 audio-recorded session per month. Fifteen FS recordings were randomly selected and coded to ensure that MST-HC elements were not present.
Sample Size
The sample size for the primary outcome was based on (1) the expected effect size for FEV1 (0.80 SD) and a small effect of the comparison condition (0.10 SD), (2) a 0.60 correlation coefficient in the repeated measurements, (3) 3 repeated measures during the course of the study, and (4) a 2-sided test (α = .05) with a power of 0.80. The estimation based on the formula for repeated measures described by Twisk39 generates a sample size of 57 to 72 subjects per group and a more conservative effect size of 0.40 to 0.45 SD. With an assumed attrition rate of 15%, we selected a sample size of 85 subjects per group.
Statistical Analyses
By using t tests for continuous variables and χ2 tests for categorical variables, demographic and other baseline variables were compared between the treatment and comparison groups to check for randomization success and identify necessary covariates and between those who completed the 12-month follow-up assessment and those who did not. Analyses were intent to treat. We used the linear mixed-effects models procedure to test differences between groups in trajectories in FEV1, FAMSS adherence, DPD adherence (treated as continuous scales), and the frequency and severity of asthma symptoms. In this analysis, the treatment-by-time interaction reveals the significance of changes in outcomes between the intervention and control groups over time. The regression coefficient of parameters (β), SE, and 95% confidence intervals (CIs) are presented. To account for overdispersion of count data, multiple negative binomial regression was used to test the number of hospitalizations and ED visits in the 12 months after baseline between the treatment and control groups, adjusting for the same covariates as the previous models. For the primary outcome at 12 months, 14.3% in the intervention group and 12% of the control group were missing FEV1. Missing data were accounted for in the mixed-effects models by using multiple imputation methods within the trajectory analysis.40 The proportional reduction in variance, an accepted local effect size estimate for mixed-effects models, is reported.39
Results
At the 7-month follow-up, 91% of families remained in the study, and at the 12-month follow-up, 89% of families remained in the study. There were no differences in retention between the MST-HC and FS groups at the 12-month assessment (x2 = 0.001; P = .98; Fig 2). In the MST-HC group, 85% of families received a minimum dose (3 sessions41) of the allocated intervention. In the FS group, 71% received the allocated intervention.
Table 2 reveals characteristics of the sample at baseline. MST-HC and FS participants differed in single-parent household (x2 = 4.655; P = .031) and income (x2 = 14.507; P = .043); thus, both were included as covariates in subsequent models. There were no significant differences on any outcome measures between MST-HC and FS participants at baseline.
Descriptive Statistics for Variables at Baseline by Intervention and Control Groups
| Variables | MST, No. (%) | Control, No. (%) | P |
|---|---|---|---|
| Age, y, mean (SD) | 13.32 (1.28) | 13.64 (1.41) | .13 |
| Sex | |||
| Female | 33 (39) | 32 (39) | .92 |
| Male | 51 (61) | 51 (61) | — |
| Annual family income, $ (USD) | |||
| 1 = <1000 | 5 (7) | 12 (15) | .04 |
| 2 = 1000–9999 | 20 (26) | 15 (19) | — |
| 3 = 10 000–12 999 | 13 (17) | 4 (5) | — |
| 4 = 13 000–15 999 | 6 (8) | 10 (13) | — |
| 5 = 16 000–21 999 | 6 (8) | 16 (20) | — |
| 6 = 22 000–31 999 | 9 (12) | 7 (9) | — |
| 7 = 32 000–49 999 | 11 (15) | 12 (15) | — |
| 8 = 50 000–99 999 | 6 (8) | 4 (5) | — |
| Single-parent household | 41 (50) | 54 (67) | .03 |
| Outcome measures at baseline | |||
| Frequency of asthma symptoms, mean (SD) | 2.75 (3.20) | 2.67 (2.34) | .84 |
| Severity of asthma symptoms, mean (SD) | 1.13 (0.95) | 1.20 (0.84) | .61 |
| FEV1, mean (SD) | 2.05 (0.56) | 2.21 (0.60) | .08 |
| FAMSS adherence to medication, mean (SD) | 4.19 (2.02) | 4.61 (2.21) | .20 |
| Adherence to DPD of controller use, mean (SD) | 0.22 (0.50) | 0.33 (0.55) | .17 |
| Frequency of hospitalizations in the previous 12 mo, mean (SD) | 0.87 (0.82) | 0.66 (0.74) | .09 |
| Frequency of ED visits in the previous 12 mo, mean (SD) | 1.15 (1.44) | 1.19 (1.52) | .86 |
| Variables | MST, No. (%) | Control, No. (%) | P |
|---|---|---|---|
| Age, y, mean (SD) | 13.32 (1.28) | 13.64 (1.41) | .13 |
| Sex | |||
| Female | 33 (39) | 32 (39) | .92 |
| Male | 51 (61) | 51 (61) | — |
| Annual family income, $ (USD) | |||
| 1 = <1000 | 5 (7) | 12 (15) | .04 |
| 2 = 1000–9999 | 20 (26) | 15 (19) | — |
| 3 = 10 000–12 999 | 13 (17) | 4 (5) | — |
| 4 = 13 000–15 999 | 6 (8) | 10 (13) | — |
| 5 = 16 000–21 999 | 6 (8) | 16 (20) | — |
| 6 = 22 000–31 999 | 9 (12) | 7 (9) | — |
| 7 = 32 000–49 999 | 11 (15) | 12 (15) | — |
| 8 = 50 000–99 999 | 6 (8) | 4 (5) | — |
| Single-parent household | 41 (50) | 54 (67) | .03 |
| Outcome measures at baseline | |||
| Frequency of asthma symptoms, mean (SD) | 2.75 (3.20) | 2.67 (2.34) | .84 |
| Severity of asthma symptoms, mean (SD) | 1.13 (0.95) | 1.20 (0.84) | .61 |
| FEV1, mean (SD) | 2.05 (0.56) | 2.21 (0.60) | .08 |
| FAMSS adherence to medication, mean (SD) | 4.19 (2.02) | 4.61 (2.21) | .20 |
| Adherence to DPD of controller use, mean (SD) | 0.22 (0.50) | 0.33 (0.55) | .17 |
| Frequency of hospitalizations in the previous 12 mo, mean (SD) | 0.87 (0.82) | 0.66 (0.74) | .09 |
| Frequency of ED visits in the previous 12 mo, mean (SD) | 1.15 (1.44) | 1.19 (1.52) | .86 |
USD, United States dollars; —, not applicable.
Primary Outcome: FEV1
Consistent with hypotheses, adolescents who were randomly assigned to MST-HC had greater improvement in FEV1 over time than adolescents who were randomly assigned to FS (β = 0.097; t[164.27] = 2.52; P = .01; Table 3). The MST-HC group improved from 2.05 at baseline to 2.25 at 7 months (a 9.8% improvement), and the comparison group changed from 2.21 to 2.31 from baseline to 7 months (a 4.5% improvement). From baseline to 12 months, the MST-HC group improved from 2.05 to 2.37 (a 15.6% improvement), whereas the FS group changed from 2.21 to 2.33 (a 5.4% improvement). The level-1 residual variance decreased by 28.6% between the unconditional model and the FEV1 predictor model.
Mixed-Effects Models of the Trajectory of Baseline, 7, and 12 Months
| Parameter | Model 1 | Model 2 | Model 3 | Model 4 | Model 5 |
|---|---|---|---|---|---|
| FEV1 | FAMSS Adherence to Medication | DPD Adherence | Frequency of Symptoms, Child Report | Severity of Symptoms, Child Report | |
| β (95% CI) | β (95% CI) | β (95% CI) | β (95% CI) | β (95% CI) | |
| Fixed effect | |||||
| Annual family income | −0.020 (−0.060 to 0.020) | 0.034 (−0.083 to 0.151) | 0.004 (−0.027 to 0.035) | −0.017 (−0.158 to 0.124) | −0.014 (−0.063 to 0.035) |
| Single-parent household | 0.018 (−0.169 to 0.206) | 0.054 (−0.486 to 0.595) | 0.111 (−0.033 to 0.255) | −0.156 (−0.811 to 0.498) | 0.117 (−0.111 to 0.345) |
| Hospitalization in past 12 mo | −0.076 (−0.193 to 0.040) | 0.199 (−0.136 to 0.535) | 0.011 (−0.078 to 0.100) | 0.164 (−0.242 to 0.570) | −0.008 (−0.150 to 0.133) |
| Specialty clinic | −0.052 (−0.243 to 0.139) | 0.027 (−0.524 to 0.578) | 0.083 (−0.063 to 0.230) | 0.104 (−0.563 to 0.772) | 0.008 (−0.225 to 0.240) |
| Intervention group | −0.141 (−0.337 to 0.055) | −0.203 (−0.872 to 0.466) | −0.051 (−0.228 to 0.127) | −0.224 (−1.044 to 0.596) | −0.147 (−0.415 to 0.120) |
| Time | 0.060* (0.007 to 0.114) | 0.062 (−0.233 to 0.357) | 0.014 (−0.077 to 0.105) | −0.075 (−0.443 to 0.292) | −0.083 (−0.190 to 0.024) |
| Time intervention group* | 0.097* (0.021 to 0.173) | 0.485* (0.062 to 0.908) | 0.193** (0.062 to 0.323) | −0.601* (−1.127 to −0.074) | −0.106 (−0.260 to 0.048) |
| Random effect | 0.250*** (0.185 to 0.338) | 0.985 (0.356 to 2.727) | 0.061* (0.030 to 0.127) | 1.495* (0.576 to 3.88) | 0.328*** (0.231 to 0.465) |
| Parameter | Model 1 | Model 2 | Model 3 | Model 4 | Model 5 |
|---|---|---|---|---|---|
| FEV1 | FAMSS Adherence to Medication | DPD Adherence | Frequency of Symptoms, Child Report | Severity of Symptoms, Child Report | |
| β (95% CI) | β (95% CI) | β (95% CI) | β (95% CI) | β (95% CI) | |
| Fixed effect | |||||
| Annual family income | −0.020 (−0.060 to 0.020) | 0.034 (−0.083 to 0.151) | 0.004 (−0.027 to 0.035) | −0.017 (−0.158 to 0.124) | −0.014 (−0.063 to 0.035) |
| Single-parent household | 0.018 (−0.169 to 0.206) | 0.054 (−0.486 to 0.595) | 0.111 (−0.033 to 0.255) | −0.156 (−0.811 to 0.498) | 0.117 (−0.111 to 0.345) |
| Hospitalization in past 12 mo | −0.076 (−0.193 to 0.040) | 0.199 (−0.136 to 0.535) | 0.011 (−0.078 to 0.100) | 0.164 (−0.242 to 0.570) | −0.008 (−0.150 to 0.133) |
| Specialty clinic | −0.052 (−0.243 to 0.139) | 0.027 (−0.524 to 0.578) | 0.083 (−0.063 to 0.230) | 0.104 (−0.563 to 0.772) | 0.008 (−0.225 to 0.240) |
| Intervention group | −0.141 (−0.337 to 0.055) | −0.203 (−0.872 to 0.466) | −0.051 (−0.228 to 0.127) | −0.224 (−1.044 to 0.596) | −0.147 (−0.415 to 0.120) |
| Time | 0.060* (0.007 to 0.114) | 0.062 (−0.233 to 0.357) | 0.014 (−0.077 to 0.105) | −0.075 (−0.443 to 0.292) | −0.083 (−0.190 to 0.024) |
| Time intervention group* | 0.097* (0.021 to 0.173) | 0.485* (0.062 to 0.908) | 0.193** (0.062 to 0.323) | −0.601* (−1.127 to −0.074) | −0.106 (−0.260 to 0.048) |
| Random effect | 0.250*** (0.185 to 0.338) | 0.985 (0.356 to 2.727) | 0.061* (0.030 to 0.127) | 1.495* (0.576 to 3.88) | 0.328*** (0.231 to 0.465) |
See Table 1 for income brackets. The single-parent category was no = 0 and yes = 1. Hospitalizations included the number of inpatient hospitalizations recorded in the medical record during the 12 mo before baseline. The specialty clinic category included whether an adolescent was seen in a specialty asthma clinic during the 12 mo before baseline (no = 0 and yes = 1). The intervention group was FS = 0 and MST = 1. The time category was baseline = 0, 7 mo = 1, and 12 mo = 2. β is the regression coefficient.
P < .05;
P < .01;
P < .001.
Secondary Outcomes: FAMSS Adherence, DPD Adherence, Asthma Symptoms Frequency, and Asthma Symptoms Severity
There were significant treatment-by-time interactions for FAMSS adherence (β = 0.485; P = .03) and DPD adherence (β = 0.193; P = .004), revealing that the MST-HC group had a greater improvement over time relative to the FS group (Table 3). From baseline to 12 months, FAMSS adherence scores improved from 4.19 to 5.24 in the MST-HC group and from 4.61 to 4.72 in the control group. The proportional reduction in variance was 10.0%. DPD adherence scores improved from a mean of 0.33 at baseline to 0.69 for the MST-HC group but only from 0.43 to 0.46 in the comparison condition. The level-1 residual variance decreased by 8.6%. Adolescents who were randomly assigned to MST-HC had a greater reduction in the frequency of asthma symptoms over time relative to adolescents who were randomly assigned to FS (β = −0.601; P = .03). The mean frequency of symptoms declined in the MST-HC group from 2.75 at baseline to 1.43 at 12 months relative to a decline of 2.67 to 2.58 in the FS group. The level-1 residual variance decreased by 10.5%. Changes in asthma symptom severity over time were not significantly different between groups.
Secondary Outcomes: Inpatient Hospitalizations and ED Visits
Multiple negative binomial regression analysis of hospitalization data in the 12 months after the intervention revealed that the MST-HC group had a lower frequency of hospitalizations in the 12 months postbaseline compared with the FS group (β = −0.882; P = .04; incidence rate ratio = 0.414; 95% CI = 0.175 to 0.978) when controlling for hospitalizations prebaseline and covariates. The mean number of hospitalizations for the MST-HC group in the 12 months before baseline was 0.87 and dropped to 0.24 in the 12 months postbaseline. The FS group had smaller declines, from 0.66 prebaseline to 0.34 postbaseline. When controlling for ED visits prebaseline and covariates, multiple negative binomial regression analysis of ED visits by MST-HC versus FS group was not significant (β = −0.126; P = .63). The MST-HC group declined from 1.15 to 0.83, and the FS group declined from 1.19 to 0.87.
Discussion
A home- and community-based, individualized behavioral treatment significantly improved FEV1 as a measure of lung function in African American adolescents with poorly controlled asthma. These findings are especially noteworthy because African American adolescents experience greater morbidity and mortality from asthma than white adolescents even when controlling for socioeconomic variables.5,42,43 This is consistent with other studies in which researchers use a home-based approach with younger, urban children,35,43,44 although this is the first rigorous randomized clinical trial (in an intent-to-treat analysis with a strong comparison condition) to reveal such improvements in objective health outcomes in African American adolescents with poorly controlled asthma. The study had high rates of retention with a low-socioeconomic status sample, possibly because of the home-based data collections.
Secondary outcomes of asthma medication adherence, symptom frequency, and reduced hospitalizations also revealed improvement in this high-risk population. This is especially promising because we managed participants for 12 months, accounting for seasonal variability in asthma symptoms and assessing whether intervention effects were sustained 6 months after treatment termination. MST-HC effects were stronger for hospitalizations than for ED visits, possibly because the ED is often used for primary care in inner-city populations. As ED visits potentially decrease overall with the improvements in primary care use that are expected with the Patient Protection and Affordable Care Act, it may be possible to more carefully determine MST-HC effects on asthma-related ED visits. Furthermore, adding interventions to link and engage families to primary care may have a greater impact on this outcome.
As expected, MST-HC families received a higher dose of treatment despite offering home-based services in both conditions. The MST-HC model of intervention on multiple factors across multiple systems (instead of only providing FS) improved intervention retention. The dose of MST-HC was somewhat lower than in our previous studies of youth with poorly controlled diabetes (averaging once per week instead of twice per week).20,45 It is possible that families did not perceive asthma, which is a common childhood illness, to be as severe as diabetes and therefore did not participate in as many sessions. A lower dose of MST-HC may have been sufficient to improve asthma management relative to our work with diabetes, for which dietary management is also necessary. However, keeping caseloads low to ensure flexibility of scheduling and allowing for multiple cancellations and rescheduling is important to achieve intervention retention in this high-risk population.
Limitations include asthma management measurement being focused on adherence to controller medications, and other areas of asthma management, such as environmental control, could be added to strengthen the study. Also, an objective measure of medication adherence was not included. Second, secondary outcomes analyses were not powered for adjustments for multiple comparisons. Third, economic analysis is necessary to determine cost offsets in terms of reduced health care use both during the trial and in the future. These analyses are underway, and researchers have demonstrated cost offsets for adolescents with diabetes.45 Finally, although the multimethod approach to the assessment of asthma management is a strength, an additional limitation is the lack of an objective assessment of asthma management. However, MST-HC improved objective health outcomes, such as FEV1. Furthermore, the MST-HC group improved FEV1 by 14%, which is considered a clinically significant change.46 In fact, with each 10% increase in FEV1, there is a progressive decrease in asthma attacks.47
Additional research is also necessary to support the transportability of MST-HC to community settings as has been done with MST for mental health.48 Future studies can test whether using community health workers (CHWs) can maintain efficacy and reduce costs.35,43,44 In 1 version of MST-HC for adolescents with HIV in rural settings, CHWs were used to augment services,22 and CHWs have been shown to successfully provide intensive home-based behavioral treatment of obesity.49 We are currently conducting a trial using CHWs from community-based organizations to deliver MST-HC intervention strategies consistent with the Centers for Disease Control call for expanded use of CHWs in chronic disease management50 with careful attention to implementation and training.51
Conclusions
There are few interventions that have been shown to successfully improve asthma management and health outcomes in African American adolescents with poorly controlled asthma, who are at the highest risk for morbidity and mortality. MST-HC, a home-based social-ecological behavioral treatment, improved adherence to daily controller medications, symptoms, lung function, and use over 12 months of follow-up. MST-HC is ready to be studied in effectiveness and implementation trials.52
- CHW
community health worker
- CI
confidence interval
- DPD
Daily Phone Diary
- ED
emergency department
- FAMSS
Family Asthma Management System Scale
- FEV1
forced expiratory volume in 1 second
- FS
family support
- MST
Multisystemic Therapy
- MST-HC
Multisystemic Therapy–Health Care
Dr Naar was the principle investigator of the study and was responsible for the conception and design, interpretation of data, and drafting of the article; Dr Ellis participated in the conception and design, interpretation of data, and drafting of the article; Dr Bruzzese was involved in the conception and design, interpretation of data, and revision of the article; Dr Cunningham contributed to the study design, intervention specification, and revision of the article; Dr Lam was responsible for the acquisition of data and revision of the article; Drs Pennar and Brownstein provided statistical analyses, interpreted data, drafted the data analysis sections, and revised the article; and all authors approved the final manuscript as submitted and agreed to be accountable for all aspects of the work.
This trial has been registered at www.clinicaltrials.gov (identifier NCT00916240).
FUNDING: Supported by National Institutes of Health research grant 1R01HL087272-01A1. The findings and conclusions are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health. Funded by the National Institutes of Health (NIH).
Acknowledgment
We thank Dr Elizabeth Secord for supporting the study and referring her patients to participate.
Comments