BACKGROUND AND OBJECTIVES: We translated family-based behavioral treatment (FBT) to treat children with overweight and obesity and their parents in the patient-centered medical home. We reported greater reductions in child and parent weight at 6 and 24 months compared with an attention-controlled information control (IC) group. This article reports the cost-effectiveness of long-term weight change for FBT compared with IC.
METHODS: Ninety-six children 2 to 5 years of age with overweight or obesity and with parents who had a BMI ≥25 were randomly assigned to FBT or IC, and both received diet and activity education (12-month treatment and 12-month follow-up). Weight loss and cost-effectiveness were assessed at 24 months. Intention-to-treat, completers, and sensitivity analyses were performed.
RESULTS: The average societal cost per family was $1629 for the FBT and $886 for the IC groups at 24 months. At 24 months, child percent over BMI (%OBMI) change decreased by 2.0 U in the FBT group versus an increase of 4.4 U in the IC group. Parents lost 6.0 vs 0.2 kg at 24 months in the FBT and IC groups, respectively. The incremental cost-effectiveness ratios (ICERs) for children and parents’ %OBMI were $116.1 and $83.5 per U of %OBMI, respectively. Parental ICERs were also calculated for body weight and BMI and were $128.1 per 1, and $353.8/ per kilogram, respectively. ICER values for child %OBMI were similar in the intention-to-treat group ($116.1/1 U decrease) compared with completers ($114.3).
CONCLUSIONS: For families consisting of children and parents with overweight, FBT presents a more cost-effective alternative than an IC group.
- %OBMI —
- percent over BMI
- CI —
- confidence interval
- FBT —
- family-based behavioral treatment
- IC —
- information control
- ICER —
- incremental cost-effectiveness ratio
- ITT —
- intention to treat
- PCMH —
- patient-centered medical home
- PEA —
- practice enhancement assistants
What’s Known on This Subject:
Data on efficacy and cost-effectiveness of family-based behavioral treatment of obesity are available in the specialized setting. Family-based behavioral treatment has shown efficacy in treating young children with overweight and their parents in primary care, but data on cost-effectiveness are lacking.
What This Study Adds:
In this study, we demonstrate the long-term cost-effectiveness of a family-based weight control program implemented in the pediatricians’ office. This model of treatment may have further cost-effectiveness impact because it can potentially positively affect the entire family.
Family-based behavioral treatment (FBT) is an evidence-based treatment of children with obesity and their parents.1 The majority of FBT has targeted 8- to 12-year-old children with obesity in specialty clinics,1,2 but FBT is effective for treating 2- to 5-year-old children.3 We developed a simplified FBT for implementation in the patient-centered medical home (PCMH) that used practice enhancement assistants (PEAs).4,5 FBT was more effective than an information control (IC) group treatment at 6, 12, and 24 months, with a positive correlation between child percent over BMI (%OBMI) change and parent BMI change from baseline to 12, 18, and 24 months.4,5 These data confirm that FBT provides the advantage of simultaneous treatment of the parent and child1,2 and a strong correlation in weight change among parents and their children.6,7
Considering the great prevalence of multigenerational obesity8 and rising health care costs associated with obesity,9,10 it is important to document the costs of effective treatments and the cost-effectiveness of simultaneously treating children and their parents with versus separately treating children and parents.11 Group FBT is more cost-effective than individual treatment,12 but data are lacking on the cost-effectiveness of implementing FBT in the PCMH and treating younger children. In this report, we quantify the cost-effectiveness of FBT from the societal perspective and evaluate whether FBT is more cost-effective in terms of child and parent weight change than an IC group.4,5
Participants and Design
A summary of the protocol previously described is given herein.4,5 The study took place in 4 pediatric PCMHs, which managed children from diverse socioeconomic and ethnic backgrounds. Children who had a BMI over the 85th percentile for their age and sex with a parent who had a BMI ≥25 were included. The main exclusion criteria were as follows: small for gestational age, short stature, and child and/or parent inability to perform physical activity. Families were recruited in cohorts of 12 child-parent dyads and then stratified by sex of the targeted child and randomly assigned to intervention or the IC group by using a random number generator.
Out of 171 potential participating families, 105 families were found to be eligible and were randomly assigned to 1 of the 2 intervention groups (Fig 1). The study was approved by the institutional review board of the Women and Children’s Hospital of Buffalo and was conducted in concordance with the Declaration of Helsinki.
Health care providers identified children with overweight, approached the parents, provided a brief description of the program, and introduced the family to a PEA who further described the study, received consent from a parent or guardian, conducted a screening, and participated in FBT delivery.
Table 1 describes the difference in treatment implementation among groups.
Protocol Common to Both Groups
Parents attended 16 treatment meetings. During the 12-month treatment period there were 13 group sessions (4 weekly, 2 biweekly, 4 monthly, and 3 at 8- to 10-week intervals), and during the follow-up period there were 3 group sessions (weeks 65, 78, and 91). These were followed by a final visit for assessment only at week 104, so that there were a total of 17 visits to the pediatrician’s office. The sessions were held on weekday evenings at 6:00 pm. During each session, a group meeting was led by a consistent group leader (medical examiner or 1 PEA) who delivered dietary, physical, and sedentary activity guidelines as previously described.5 A PEA, assigned to each family and referred to as a health coach, telephoned the parent between scheduled meetings (10 times during treatment and 3 times during follow-up) to remind him or her of the upcoming meeting in the IC group and to counsel and problem solve in the FBT group. The child’s weight goal was 0.5 to 1 lb per week of weight loss. This goal, in keeping with the stage 3 program goals of the “Expert Committee Recommendations for Prevention and Treatment of Pediatric Obesity”13 (0.25 lb of weight loss per week for 2–5-year-old children and 1–2 lb per week in 6–12-year-old children) was chosen because many children were severely overweight and some would turn 6 years old shortly after entering the study. The rate of change in weight and %OBMI was carefully monitored. The child’s pediatrician provided feedback to the parents regarding the child’s %OBMI changes every 6 months during an in-person visit and by sending a letter in between the 6-month visits.
Protocol in the FBT Intervention
Behavior modification and education on parenting techniques were delivered by the group leader during group sessions and by the PEA, assigned as a health coach to each family, during brief individual sessions. Parents were instructed to monitor their child and their own weight twice a week and received dietary (1500 and 1800 kcal/day for mothers and fathers, respectively), physical, and sedentary activity guidelines with the goal of a minimum of 1 lb per week of weight loss. Parents recorded intake and activity for their child and themselves in a diary by crossing off icons detailing food groups and physical and sedentary activity.
Children’s weight was measured by a trained PEA every session in both groups using a Tanita electronic scale, model BWB-800A. Because growth velocity is high in young children, height was measured at baseline and every 3 months throughout the study by using a Seca 216 Accu-Hite stadiometer. Parents’ weight was measured every session in the intervention and at baseline and 6, 18, and 24 months in the IC group. Height and weight were used to calculate BMI. Child weight changes are expressed as %OBMI, which is defined as ([child’s actual BMI − 50th percentile BMI] ÷ 50th percentile BMI × 100). %OBMI is a standardized measure indicating how much a child is overweight or underweight relative to the 50th percentile BMI reference standard specific to his or her age and sex.14,15
The Hollingshead 4-Factor Index16 was used to measure socioeconomic status (Table 2). This index uses 4 factors to determine socioeconomic status: sex, marital status, educational level, and occupation. Educational level and occupation status are scored on a predetermined scale, and these values are multiplied by a weight of 3 and 5, respectively. The resulting values are then summed for the head of household. If there is >1 head of household, their individual scores are averaged. An SES of 46 through 48 is equivalent to medium-size business owners, minor professionals, and technical jobs, such as computer programmers, real estate agents, sales managers, social workers, and teachers. The study population included 26% minority children.
In this study, we focused on the cost-effectiveness at 24 months from baseline. Treatment costs (payers) and opportunity costs (participating families) were included to represent societal cost.
Payer Costs Components
Costs for teaching, weighing, caregiving, and counseling were estimated from hourly wages and time spent by study personnel discharging these duties during each session. Sixteen sessions were administered for each cohort in both the FBT and IC group, consisting of 1 90-minute session, 3 60-minute sessions, and 13 45-minute sessions administered by the group leader. In the FBT intervention, the PEA assigned to the family also acted as a health coach, conducting an in-person, 1:1 coaching lasting on average 20 minutes before or after each group meeting. Both the FBT and IC group used PEAs to provide child care while parents received education. The number of PEAs and time spent by PEAs fluctuated, depending on the number of families attending and whether they needed 1:1 counseling (FBT only). In addition to the costs of regularly scheduled sessions, we included costs for telephone calls performed by coaches to counsel between meetings (counseling, Table 3) and costs for calls made to set make-up sessions (make-up sessions, Table 3). The average number of make-up sessions was 1.8 and 1.3 per child in the FBT and IC group, respectively. FBT make-up sessions (n = 56) took 60 minutes to complete, including the 1:1 session, whereas IC group sessions (n = 64) took 45 minutes. One PEA provided care for the child while the PEA assigned to the family as a health coach conducted the 1:1 session and covered the didactic materials from the group session. The average hourly wage without benefits was $12 for the PEAs and $25.47 for group leaders.
Staff Travel and Fuel Costs
Staff were paid during travel time to each PCMH and reimbursed for fuel cost (1-way) by using the business mileage rate (56.5 cents per mile) of the Internal Revenue Service Standard Mileage Rates for 2013.17 Mileage to each site (Amherst, Buffalo, Williamsville, and Tonawanda, New York) was used to estimate fuel cost and travel time by using MapQuest for scheduled and make-up sessions (Table 3).
Senior staff (M.A.E., J.N.R., and T.Q.) and 5 PEAs participated in 47 team meetings, during which PEAs received training and reviewed the child’s progress and parents’ progress in the FBT intervention group and parent’s progress in the IC group when parent weight was measured (6-, 12-, 18-, and 24-month visits). J.N.R. and T.Q. led the meeting and coached the PEAs on areas in which they could improve and offered strategies for problem solving with parents. At each meeting, J.N.R. and T.Q. would also review parenting skills and techniques to promote behavior change. Expenses for staffing were estimated on the basis of time spent on each activity with each treatment group (Table 4).
Material Preparation and Copies
Staff costs for material preparation by PEA were aggregated over the study and combined with expenses for copies, paper, and supplies related to the preparation of study materials. Indirect costs (clinic rental, heating, electricity, etc) were not considered because these costs were incurred by those clinics and not our intervention.
Participant Costs Components
Parents completed an expense questionnaire during baseline assessments, in which they reported round-trip mileage and travel time as well as what they would be doing if they were not traveling and attending meetings. Fuel cost was computed on the basis of mileage and 2013 Internal Revenue Service standard rates for all participant-attended commutes to the clinic.17 Opportunity costs for lost wages or lost time doing household activities were considered for participants reporting that the appointment displaced “work at a job outside the home.”
Opportunity costs for treatment and transportation time were monetized by using an average hourly wage rate of US adults from the Bureau of Labor Statistics the year each participant completed the questionnaire and adjusted for inflation to 2013 wages.18
Fuel cost and travel time for make-ups were estimated similarly and are included in reported opportunity and fuel costs. Opportunity costs for time spent in make-up sessions are reported separately.
Payer and participant costs were used to compute cost per participant. %OBMI was used as part of the incremental cost-effectiveness ratio (ICER) because it was the primary outcome of the study; absolute BMI changes do not appropriately track weight changes in childhood.14,19 For parents, besides calculating the cost per unit of BMI and weight changes, %OBMI was also used to have a dependent measure that could be used to compare program cost-effectiveness (costs per unit of %OBMI change) among children and parents. ICER, defined as the difference in incremental cost (cost difference between the FBT and IC group) divided by the incremental effect observed (difference in %OBMI between the FBT and IC group), was computed to assess additional costs for obtaining additional benefits gained from this new intervention (FBT).20 ICERs comparing FBT to IC were calculated for both parents and children in which the ICER equals the cost difference between interventions divided by the difference in health outcomes (ie, child %OBMI reduction or parent weight reduction). Weight or %OBMI were treated so that reductions in these outcomes were shown as positive values to indicate positive clinical changes and preserve conventions of interpretation of ICER values.20,21 Analyses were completed by using SYSTAT and SAS.22,23
The outcome data that were used to calculate cost-effectiveness are presented on the basis of intention-to-treat (ITT) analyses (N = 96) as well as for “completers” (N = 70). ICER scores were computed accordingly for both ITT and completers for comparison. In addition, a sensitivity analysis was performed on the effect side (ie, health outcomes). Ninety-five percent confidence intervals (CIs) were calculated for the effects and their ICER, respectively. To estimate uncertainty in the ITT sample, we handled missing data by performing multiple imputations for child and parent data using NORM Version 2.0324 and following the approach of imputing 10 data sets.25 Imputation models included age, sex, treatment group, and baseline value of child %OBMI, or parent weight (kilograms), BMI, or %OBMI, respectively.
Baseline characteristics of children and parents are presented in Table 2. No differences were observed in any of the child or parent characteristics. As previously reported, differential group changes were observed in child %OBMI across time (P = .002), with the FBT group showing greater decreases compared with the IC group from baseline to 6 (P < .001), 12 (P = .005), 18 (P = .005), and 24 months (P = .001).5 Similar effects were seen for parent weight, with changes across time (P < .001) and greater decreases for FBT at all follow-up time points (P < .001).
Societal costs at 24 months (Table 3) for FBT ($74 928) were greater than IC ($44 289). The extra expenses in FBT arose from payer costs in FBT ($56 112) versus IC ($24 210) rather than participant costs, which were $18 816 in FBT versus $20 079 in IC.
It cost ∼$886 for a 1-U decrease in %OBMI in the FBT group at 24 months, whereas children in the IC group showed increased %OBMI at 24 months, so funds spent on IC did not provide benefit (Table 5). At 24 months, the cost for each kilogram of weight loss for FBT parents was ∼$271 versus $4430 per kilogram of weight loss for IC parents, and the cost for a 1 U decrease in %OBMI in FBT parents was $179 compared with $4430 in IC parents. The ICER for %OBMI is $116.1 and $83.5 per 1% decrease in %OBMI in the child and parent, respectively. The ICER for parent weight and BMI is $128.1 per kilogram and $353.8, respectively. The cost per unit of %OBMI reduction did not correlate with baseline %OBMI in children (r = −0.039) or parents (r = 0.076).
In addition to these point estimators, 95% CIs are also presented in Table 5 for the ITT group to account for uncertainties on the effect side. In an additional sensitivity analysis, ICERs were computed for families that completed the treatment (Table 6, N = 70). ICER values for child %OBMI were similar ($114.3/1 U decrease) compared with ITT ($116.1). For the parents, ICERs were also similar for %OBMI, BMI, and kilogram body weight in completers ($76.8/1-U %OBMI, $353.8, and $116.1/kilogram respectively) versus ITT ($83.5/1-U %OBMI, 353.2, and 128.1/kilogram respectively).
In this study, we revealed that FBT for parent and child obesity implemented in the PCMH is not only more clinically effective, as demonstrated by significant between-group differences in %OBMI or body weight at all measurement points, but also more cost-effective than an attention-matched IC group. At 24 months, FBT was more cost-effective than IC for both parents and children, with $1629 spent to decrease the child %OBMI by almost 2 U, whereas a cost of $886 resulted in an increase in %OBMI by 4.4 U. ICER for parents’ %OBMI reduction is ∼$84 per U compared with the child at $117 per U. Given the great prevalence of obesity in children whose parents also have obesity26,27 and the rising cost of health care for patients with obesity,9,10 cost-effective treatments are needed. There are estimates for the cost of health care for adults with obesity9,10 and projected costs for children with obesity,28 but it is important to recognize that obesity usually affects several family members,29 and the costs can be multiplied if several family members within or across generations have obesity. Preliminary data suggest that there is a generalization of treatment effects within families,30 so that children in the family who have obesity but are not targeted for treatment also can benefit from their siblings being treated, as an additional benefit.
In considering the results of our study, it is important to acknowledge that the practices engaged in the study were all motivated to take advantage of the PEAs embedded in the practice. Because of the fact that the overwhelming majority of sessions took place during office hours, which extended in the evenings and weekends, some minor costs such as extra electricity and/or space costs of the conference room used for group sessions were not charged to our program. Moreover, it would be important to test the cost-effectiveness of programs implemented in primary care that were not dependent on group meetings. This approach may be more costly and would not allow for developing support among participant parents, but it would be more in keeping with the reality of difficult family and pediatricians’ office schedules.
Sensitivity analysis is useful to account for potential uncertainties that may arise during the estimation procedure for costs and effectiveness. In this study, we calculated the 95% CIs of ICER ratios as the potential upper and lower bounds of the estimation, which took into account the uncertainty on the effect side (ie, health outcomes). Although some of these intervals show a wide range because of the uncertainties of health outcomes, all the estimates indicate that FBT is more cost-effective than IC. Lastly, the economic value of the weight reductions may also depend on the %OBMI starting point; however, in our study, the cost per unit of %OBMI reduction did not correlate with baseline %OBMI.
With our data, we show that FBT can be cost-effectively implemented in the PCMH. Interventions implemented in the PCMH cannot only provide an evidence-based treatment to more children than possible in specialty clinics but also easily reach out to the entire family so that the pediatrician’s office becomes a family-centered medical home. Recent research data suggest that FBT is a cost-effective alternative to treating children and parents separately.11,31 Our success with FBT in the PCMH enhances the possibility of dissemination of cost-effective FBTs. Research is needed to strengthen weight-change maintenance, which would lead to better long-term success, resulting in long-term economic benefits. Embedding psychologists in the PCMH to improve behavioral health is an innovative model,32,33 which may allow behavioral treatments to be adapted for treatment in areas other than obesity and potentially affect the entire family.
We thank the children and families who participated in this study. We thank the pediatricians, allied health personnel, and office staff in the practices who actively partnered with us in this study. We thank Thomas Songer, PhD, for his assistance with the initial plan and tool development for CEA. We thank our PEAs Amanda House, MPH; Elissa Ortolani, BA, medical student; Megan Russell, MS, CAS; Lisa Rychlicki, RD; and Jessica Sickau, BA. We thank Sherry Ortiz for her assistance with manuscript preparation.
- Accepted June 27, 2017.
- Address correspondence to Teresa Quattrin, MD, Department of Pediatrics, Women and Children’s Hospital of Buffalo, 239 Bryant St, Buffalo, NY 14222. E-mail:
Drs Quattrin and Roemmich contributed to study design, trial organization and execution, collection, analysis, and interpretation of the data, and manuscript preparation; Dr Cao and Mr Paluch contributed to study design, data management and analysis, interpretation of the data, and manuscript preparation; Dr Epstein contributed to study design, data analysis, interpretation of the data, and manuscript preparation; Ms Ecker contributed to trial organization, study design, and execution of both the pilot study and the study published herein, study protocol development, manuscript preparation (with particular attention to data collection, management, and interpretation); and all authors gave final approval of the version to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The National Institutes of Health was not involved in analysis or interpretation of the data in this article.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: This research was funded by the National Institutes of Health grant 1R01HAD053773-01, Principal Investigator, Teresa Quattrin. Funded by the National Institutes of Health (NIH).
POTENTIAL CONFLICT OF INTEREST: Dr Epstein was a consultant to Kurbo (Datri Health, Inc) and had equity in the company when the study was implemented, but he has since resigned from the company; the other authors have indicated no potential conflicts of interest to disclose.
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2017-1916.
- Quattrin T,
- Roemmich JN,
- Paluch R,
- Yu J,
- Epstein LH,
- Ecker MA
- Quattrin T,
- Roemmich JN,
- Paluch R,
- Yu J,
- Epstein LH,
- Ecker MA
- Garn SM,
- Clark DC
- Barlow SE; Expert Committee
- Hollingshead AB
- Internal Revenue Service
- Bureau of Labor Statistics
- Muennig P
- Phillips C,
- Thompson G.
- ↵SYSTAT [computer program]. Version 11.0 Richmond, CA: SYSTAT Software, Inc; 2004
- ↵SAS [computer program]. Version 9.1.3. Cary, NC: SAS Institute Inc; 2004
- Schafer J
- Schafer JL
- Copyright © 2017 by the American Academy of Pediatrics