Accepted Manuscript Title: Effects of a structural intervention and implementation on physical activity among youth in residential children’s homes Author: Gregory M. Dominick Ruth P. Saunders Marsha Dowda Kelli Kenison Alexandra E. Evans PII: DOI: Reference:

S0149-7189(14)00067-6 http://dx.doi.org/doi:10.1016/j.evalprogplan.2014.05.011 EPP 1126

To appear in: Received date: Revised date: Accepted date:

16-12-2013 21-5-2014 26-5-2014

Please cite this article as: Dominick, G. M., Saunders, R. P., Dowda, M., Kenison, K., and Evans, A. E.,Effects of a structural intervention and implementation on physical activity among youth in residential children’s homes, Evaluation and Program Planning (2014), http://dx.doi.org/10.1016/j.evalprogplan.2014.05.011 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Physical activity in Residential Children’s Homes

1

Effects of a structural intervention and implementation on physical activity among youth in residential children’s homes

ip t

Gregory M. Dominicka*, Ruth P. Saundersb, Marsha Dowdac, Kelli Kenisond, Alexandra E. Evanse a

an

us

cr

Department of Behavioral Health and Nutrition, University of Delaware, 26 North College Avenue, Newark, DE 19716, USA b Department of Health Promotion, Education, and Behavior, Arnold School of Public Health, University of South Carolina, 915 Greene St. Columbia, SC 29208, USA c Department of Exercise Science, Arnold School of Public Health, University of South Carolina, 921 Assembly St. Columbia, SC 29208, USA d Center for Health Services and Policy Research, Arnold School of Public Health, University of South Carolina, 730 Devine St. Columbia, SC 29208, USA e Division of Health Promotion and Behavioral Sciences, University of Texas School of Public Health, 313 E. 12th St. Austin, TX 78701, USA

Ac ce p

te

d

M

*Corresponding author. Tel.: 13028313672 Email addresses: [email protected] (G.M. Dominick), [email protected] (R.P. Saunders), [email protected] (M. Dowda), [email protected] (K. Kenison), [email protected] (A.E. Evans)

Funding: This work was supported by The Duke Endowment. Any opinions, findings, conclusions, or recommendations contained herein are those of the authors and do not necessarily reflect the views of The Duke Endowment.

Page 1 of 38

Physical activity in Residential Children’s Homes

2

Ac ce p

te

d

M

an

us

cr

ip t

Effects of a structural intervention and implementation on physical activity among youth in residential children’s homes

Funding: This work was supported by The Duke Endowment. Any opinions, findings, conclusions, or recommendations contained herein are those of the authors and do not necessarily reflect the views of The Duke Endowment.

Page 2 of 38

Physical activity in Residential Children’s Homes

3

Abstract: This study reports the effects of a structural intervention, ENRICH (Environmental Interventions in Children’s Homes) which targeted the physical and social environment

ip t

within residential children’s homes (RCHs) to increase physical activity (PA) among residents (N=799). Participating RCHs (N=29) were randomized to Early (n=17) or

cr

Delayed (n=12) groups from 2004-2006 and 2006-2008, respectively. Children’s PA

us

was measured at three time periods (2004, 2006, 2008). Intent-to-treat analysis revealed no intervention impact on PA. Subsequent analyses used process evaluation

an

data to group organizations into high and low PA-promoting RCHs to compare PA level, controlling for assignment to condition. Organizations with high PA-promoting

M

environments were found to have more active youth. Utility of a comprehensive

d

implementation monitoring plan and the need for formative assessment of

Ac ce p

te

organizational capacity is discussed.

Key Words: Physical activity, foster youth, implementation monitoring, process evaluation

Page 3 of 38

Physical activity in Residential Children’s Homes

4

1. Introduction Health behaviors including physical activity (PA) are typically reinforced within settings that provide supportive structures and opportunities (Booth, Sallis, Ritenbaugh,

ip t

Birch, Frank, et al., 2001; Cradock, Melly, Allen, Morris, & Gortmaker, 2007; Elder,

Lytle, Sallis, Young, Steckler, et al., 2007; Authorc et al., 2011; Giles-Corti & Donovan,

cr

2002; Pate, Ward, Saunders, Felton, Dishman, et al., 2005; Roemmich, Epstein, Raja,

us

Yin, Robinson, et al., 2006; Sallis, McKenzie, Conway, Elder, Prochaska, et al., 2003). Structural interventions target influencing factors within physical and social

an

environments that are beyond individual control (Blankenship, Friedman, Dworkin, & Mantell, 2006; Booth et al., 2001; Cohen, Scribner, & Farley, 2000; Koplan, Liverman, &

M

Kraak, 2005; Matson-Koffman, Brownstein, Neiner, & Greaney, 2005; McLeroy, Bibeau,

d

Steckler, & Glanz, 1988). The structural ecologic model (SEM) identifies four structural

te

factors of environmental influences (availability of products and services, characteristics of available opportunities, social structures and policies, and media/cultural messages)

2000).

Ac ce p

which have the capacity to impact population-level health outcomes (Cohen et al.,

Structural interventions are situated in “real world” settings, entail working extensively with stakeholders, focus on change in higher order units (e.g., organization or community versus individual) with inherent variability, and are subject to strong contextual influences (Chen, 2004; Shadish, Cook, & Campbell, 2001). It is important to consider natural variations within the implementation setting as part of structural intervention design (Poland, Krupa, & McCall, 2009). This may be achieved in part by conceptualizing these approaches as “complex” interventions, defined as interventions

Page 4 of 38

Physical activity in Residential Children’s Homes

5

that interact with the setting during implementation (Chen, 2004; Cohen et al., 2000; Foster-Fishman, Nowell, & Yang, 2007; Hawe, Shiell, & Riley, 2004). Structural interventions that aim to change social and physical environments with the aid of local

ip t

change agents present implementation challenges that necessitate a comprehensive approach to program evaluation and implementation monitoring (Medical Research

cr

Council, 2008; Durlak & DuPre, 2008, Saunders, Ward, Felton, Dowda, & Pate, 2006).

us

Programs lacking implementation assessment cannot ascertain the effectiveness of program implementation and how outcome data should be interpreted (Durlak & DuPre,

an

2008). Furthermore, inadequate assessment of implementation coupled with the assumption that an intervention is implemented homogeneously across diverse

M

organizational settings can result in a “Type III error” where a program is found to be

d

ineffective, when actually it was not completely and/or properly implemented (Durlak &

te

DuPre, 2008; Harachi, Abbott, Catalano, Haggerty, & Fleming, 1999; Linnan & Steckler, 2002; McGraw, Sellers, Johnson, Stone, Bachman, et al., 1996; Pate, Saunders,

Ac ce p

Dishman, Addy, Dowda, et al., 2007; Saunders, Evans, & Joshi, 2005). Process evaluation data can be useful in determining several facets of implementation including, the extent to which a program is implemented as planned (i.e. fidelity), level of satisfaction and/or confidence of program change-agents to deliver program components (dose-received), and the extent to which all program components were completed (i.e. dose-delivered) (Saunders, Evans, & Joshi, 2005). For example, McGraw and colleagues (1996) used process data to examine the relationship between implementation of classroom-based health education and diet self-efficacy (i.e. confidence) for The Child and Adolescent Trial for Cardiovascular Health (CATCH)

Page 5 of 38

Physical activity in Residential Children’s Homes

6

intervention. Implementation of classroom activities was monitored through random teaching observations and teacher self-report surveys, including self-efficacy to deliver the CATCH education curriculum. Results determined that greater implementation of

ip t

CATCH health education significantly predicted student outcomes (McGraw, Sellers, Johnson, Stone, Bachman, et al., 1996).

cr

Because physical activity is performed in specific settings and is influenced by

us

contextual factors within the physical and social environment (Watts, Phillips, Petticrew, Harden, & Renton, 2011) structural interventions should examine the complex

an

ecological influences within organizational settings which provide services to vulnerable youth. Residential children’s homes (RCHs), or residential group homes, are facilities

M

that provide onsite residential care and treatment services to children who have been

d

removed from their biological families. Although the mission of RCHs are similar

te

regarding the well-being of foster children, individual organizations can differ in the number of residential foster youth served, organization size (e.g. number of RCH staff),

Ac ce p

and types of services provided (Curtis, Alexander,& Lunghofer, 2001). In 2010, over 400,000 American children were placed into foster care

(Children’s Defense Fund, 2013) due to abuse, neglect, or child’s behavioral or emotional problems (U.S. Department of Health and Human Services, 2005). In North and South Carolina, nearly 18,000 children were placed into either foster family care or residential group homes in 2010 (Children’s Defense Fund, 2013). Though a child’s length of stay in foster care is typically less than two years, the long-term health implications from this experience increases a child’s odds for a myriad of negative social, mental, and physical health outcomes (James, Landsverk, Slymen, & Leslie,

Page 6 of 38

Physical activity in Residential Children’s Homes

7

2004; Viner & Taylor, 2005, Zlotnick, Tam, & Soman, 2012). A recent study found adults placed into foster care as children were more likely to have extended gaps in employment, suffer from asthma, Type 2 diabetes, hypertension, stroke, and heart

ip t

disease compared to adults who had never been placed into foster care (Zlotnick et al., 2012). This finding warrants preemptive intervention efforts that promote healthy

cr

lifestyles that can track into adulthood. Structural interventions targeting environmental

us

change may be an ideal approach for improving health among vulnerable youth who are at greater risk for multiple negative social and health outcomes. To date, this is the first

an

study to report on youth physical activity outcomes and implementation from a structural intervention targeting RCHs in North and South Carolina.

M

The purpose of this paper is to report the effect of a group randomized

d

intervention, Environmental Interventions in Children’s Homes (ENRICH), on youth PA.

te

It was hypothesized that a greater percentage of children in the early-intervention group would report 2+ blocks of moderate-to-vigorous physical activity (MVPA) and Total

Ac ce p

MET-weighted blocks at post-test (2006), compared to those in the delayed-intervention group (i.e., control-group). Furthermore, it was hypothesized that intervention effects would be greater in RCHs with documented higher levels of intervention implementation. 2. Methods 2.1.

Participants There were 63 potential participant RCH organizations in North and South

Carolina affiliated with The Duke Endowment at the beginning of the study. Nine of these declined participation (most commonly because the program was “not needed”).

Page 7 of 38

Physical activity in Residential Children’s Homes

8

Of the remaining 54 RCHs, 29 were eligible to participate based on the criteria of having a relatively stable population of children requiring low-to-moderate management with no restrictions on PA.

ip t

Assignment to condition was done at the organizational level. Eligible RCHs were matched on organizational characteristics and randomly assigned to Early (treatment)

cr

or Delayed (control) intervention groups; using a randomized crossover design, the

us

Early and Delayed groups received the intervention from 2004-2006 and 2006-2008, respectively. Matching criteria included location (SC or NC), complex versus simple

an

organizational structure based on number of locations and services provided, participation in National Breakfast and Lunch Program, state accreditation, and existing

M

PA programs. At baseline, it was discovered that one North Carolina RCH organization

d

randomly assigned to the Early intervention group was situated across five regional

te

locations. It was ultimately considered to be six separate RCHs, resulting in 18 Early intervention (or treatment) RCHs. One Early intervention RCH dropped out during the

Ac ce p

first year, leaving 17 treatment and 12 Delayed RCHs in the final organizational sample. Individual participants were 799 children residing in 24 of the 29 RCHs that had at least 10 children in residence at the time of measurement; eight and 15 RCHs were located in North Carolina and South Carolina, respectively. Children were recruited if they were 11 to 18 years of age and could complete questionnaires with minimal assistance. Due to the transient nature of this population (average length of stay was less than one year), we used a cross-sectional study design to assess intervention impact on PA at the individual level; data were collected across three measurement waves in 2004, 2006, and 2008. Because of the wide distribution of RCHs across two

Page 8 of 38

Physical activity in Residential Children’s Homes

9

states, RCH staff helped recruit children using materials that were developed by the research team. Children who participated received a small item valued at $1 (e.g., cologne, lotion, airplane glider, cards). This study was approved by the Institutional

ip t

Review Board of the University of South Carolina. Before data collection, written informed consent from a parent/guardian or authorized case worker from the

cr

department of social services and signed assent forms from children were obtained.

us

Trained data collectors measured children’s height and weight with a stadiometer (Shorr Productions, Olney, MD) and digital scale (Seca 880/881, Seca Corporation, Hanover,

an

MD), respectively and administered self-report questionnaires to children at each participating organization. Descriptive variables included age (Mean ± SD); years in

M

current RCH residence (Mean ± SD); gender (male/female); race (White/Black/Other);

ENRICH Intervention

te

2.2.

d

weight (normal/overweight/obese); and body mass index (BMI) score (Mean ± SD).

At the initial 2004 planning meeting, ENRICH staff and advisory board committee

Ac ce p

members facilitated small discussion groups with participating RCH CEOs and staff members, and community partners and stakeholders, to elicit input regarding the intervention design, feasibility and acceptability of proposed intervention strategies and measures to assess children’s PA. Summaries from these group discussions revealed several RCHs had already identified health initiatives targeting childhood obesity and if randomized to the delayed condition would likely pursue these initiatives rather than wait to receive the intervention. The ENRICH intervention entailed developing working relationships with RCH staff to achieve the mutually agreed-upon goal of enhancing RCH social and physical

Page 9 of 38

Physical activity in Residential Children’s Homes

10

environments to promote PA for children residing in the homes. Adult staff formed small working groups, Wellness Teams (WT) which served as organizational change agents (Commers, Gottlieb, & Kok, 2007) by developing and carrying out plans to create RCH

ip t

environments that supported PA. Project staff provided two annual regional WT

trainings between 2004-2006 (Early Group) and 2006-2008 Delayed Group. Individual

cr

WT consultations and personal/technical support were also provided upon request.

us

ENRICH was designed as a flexible intervention which allowed for WTs to modify the intervention as it suited their organizational context. This process allowed for

an

organizational adaptation in which to customize to their specific setting while keeping with the overall ENRICH intervention elements based on a pre-defined conceptual

M

framework (Authord, et al., 2013). As described previously (Authorb et al., 2009; Authorc

d

et al., 2011; Authord et al., 2013), elements of the health-promoting environment

te

included: 1) opportunities for enjoyable PA; 2) RCH policies and practices that support PA; 3) media (promotion) and cultural (adult modeling) messages promoting PA; and 4)

Ac ce p

adult support and encouragement for PA. The WTs implemented a variety of strategies based on their assessments. Most promoted on and offsite PA opportunities through flyers and announcements, and provided training to staff members on the importance of encouraging the children to be active and participating in activities with the children. Some WTs implemented “girls only” workout sessions or added activities selected by the children. Others modified schedules to preclude other activities (e.g. homework) during time scheduled for PA.

3. Measures

Page 10 of 38

Physical activity in Residential Children’s Homes

11

3-Day Physical Activity Recall (3-DPAR): Children’s PA was assessed using the 3DPAR which includes self-reported physical and sedentary activities (McMurray, Ring, Treuth, Welk, Pate, et al., 2004; Pate, Ross, Dowda, Trost, & Sirard, 2003). The 3-

ip t

DPAR is administered by trained research assistants and has been found to be valid for adolescent boys and girls (McMurray et al., 2004, Pate et al., 2003). A detailed

cr

description of the 3-DPAR used for this study has been previously reported (Authora et

us

al., 2009, Authorc et al., 2011). Briefly, each child referred to a list of 62 activities indicating the predominant activity performed within 30-minute block periods (7:00 a.m.

an

to midnight) for each of the three previous days. The 3-DPAR was administered on a Tuesday, Wednesday, or Thursday to capture weekday and weekend activities.

M

Physical and sedentary activities were grouped into the following categories: eating,

d

work, after school/spare-time/hobbies, transportation, sleep/bathing, school, physical

te

activities and sports. Two open ended responses were offered for children to include additional activities not provided in the list. An age-appropriate script with graphic

Ac ce p

figures was used to help children identify the intensity of each reported activity (light, moderate, hard, very hard intensity) (Authora et al., 2009). Self-reported activities corresponded to metabolic equivalent (MET) values based on the Compendium of Physical Activities (Ainsworth, Haskell, Whitt, Irwin, Swartz, et al., 2000). Data were summarized by adding the number of 30-minute blocks for which the reported activity was rated at an intensity of three or more METs consistent with the 2008 PA guidelines for moderate-to-vigorous physical activity (MVPA) (U.S. Department of Health and Human Services, 2008). To assess program

Page 11 of 38

Physical activity in Residential Children’s Homes

12

effectiveness, Early and Delayed homes were compared on total METS and on the percentage of children achieving two or more blocks of MVPA. 3.1.

Description of Environmental Measures

ip t

Environmental characteristics of participating RCHs were assessed annually with two surveys specifically developed for this study: 1) the Physical Activity and Dietary

cr

Environmental Assessment questionnaire (PADEA) completed by the Assistant Chief

us

Executive Officer (CEO) or designated organizational representative and 2) the Organizational Assessment Survey (OA) completed by the CEO or designee. Baseline

an

response rates were high for the PADEA (N=25, 86%) and the OA (N=29, 100%). PADEA: The PADEA is a 69-item paper-and pencil questionnaire based on the

M

Structural Ecologic Model and assesses the physical activity and food environments of

d

RCHs. Pilot-testing for the PADEA was done with Assistant CEOs from non-

te

participating RCHs. The PADEA includes 22-items measuring eight physical activity environmental variables: availability of physical activity opportunities (1-item, e.g.

Ac ce p

free time to play after school), physical activity structures (4-items, e.g. availability of playground equipment) and characteristics (3-items, e.g. children provide input for the types of physical activities offered), social environment related to physical activity (7-items, e.g. children observe staff being physically active), physical activity policies (2-items,e.g. RCH provides transportation for children to participate in sports or activities off-site), organizational support for physical activity (3-items, e.g. time children can be active does not conflict with other RCH scheduled activities or appointments), staff wellness programs (1-item), and collaboration with external agencies (1-item, e.g. partnering with local YMCA). Response options ranged from 0-3,

Page 12 of 38

Physical activity in Residential Children’s Homes

13

with 0 indicating “does not exist” or “Never” and 3 indicating “Fully in place” or “Always”. A physical activity environment score was determined for each index by summing and averaging response to items in the index; higher index scores indicate the RCH

ip t

environment is more conducive to physical activity. A two-week test-retest was

conducted with 21 RCHs and internal reliability (from between 41-43 RCHs) for the

cr

eight PADEA variables indicated acceptable to good reliability among the indices (ICC

us

scores ranged from 0.38 - 0.98 and standardized Cronbach alpha coefficients ranging from 0.62 - 0.90).

an

Organizational Assessment (OA) Survey: The OA was developed to ascertain descriptive organizational information about each RCH and was updated annually. The

M

OA Survey included 30 items which assessed whether or not PA programs were

d

provided, presence of an on-site school, presence of a recreation director, number of

te

sites, populations served, and types of services provided. Organization’s that had multiple locations, served multiple populations, and/or provided multiple types of care

Ac ce p

were classifies as “complex”; the remaining homes were classified as “simple”. As previously reported (Authorc et al., 2011), urban versus rural setting was determined based on 2000 census data using the Rural Urban Commuting Area (RUCA) system (Washington State Department of Health, 2013). Additional information obtained included details about the children’s schedule and contact persons as well as information relevant for nutrition (not reported in this paper). 3.2.

ENRICH Implementation Monitoring and Results ENRICH used a comprehensive approach to assess “fidelity and completeness”

of implementation, reflected by the activities of the Wellness Team. The specific

Page 13 of 38

Physical activity in Residential Children’s Homes

14

process evaluation purposes and methods used to assess fidelity and completeness for physical activity implementation have been reported previously (Authord et al., 2013) and are summarized here. Physical activity implementation addressed PA objectives

ip t

and PA plan implementation, with an emphasis on making PA opportunities available to RCH residents, and creating a social and media environment that was supportive of PA.

cr

Data from five sources, including the WT contact (the implementer), environmental

us

observations, and ratings from measurement and intervention staff, were triangulated to assess PA implementation. Between 2004-2006, 9 RCHs in the Early and 4 in the

an

Delayed groups met the criteria for physical activity implementation; 4 RCHs in the Early and 4 in the Delayed groups met the criteria for physical activity implementation in

M

2006-2008; and 5 RCHs of 29 (3 in Early and 2 in Delayed) met the criteria for PA

d

implementation from 2004-2008 (Authord et al., 2013).

te

We also measured global implementation using five data sources to assess perceptions of overall (versus PA-specific) implementation from the change agents’,

Ac ce p

evaluator’s, interventionist’s, and other staff perspectives using the same methods reported by Saunders and colleagues (Authord et al., 2013). From 2004-2006, 9 RCHs assigned to the Early group met the criteria for global implementation (not assessed in delayed); from 2006-2008, 10 Early and 5 Delayed RCHs met the criteria); and between 2004 and 2008, 12 of 29 organizations (7 in Early and 5 in Delayed) met the criteria for global implementation. 3.3.

Data Analysis To determine intervention impact on child PA (intent-to-treat analysis), Early and

Delayed intervention groups were compared on study outcomes using a mixed-model

Page 14 of 38

Physical activity in Residential Children’s Homes

15

ANOVA with nested cross-sectional design over three measurement waves (2004, 2006, and 2008), controlling for age, sex, BMI, years in home, race, presence of a recreation director, and on-site school. Statistical significance was set at alpha of 0.05

ip t

using SAS statistical software 9.1.

Subsequent analyses examined the effects of condition adjusted for level of

cr

physical activity implementation and global implementation. Given the variability in

us

implementation among early intervention homes and early “implementation” in homes serving as controls, we also examined the effects of grouping RCHs by level of PA and

an

Global implementation for PA (versus grouping by condition) from 2004-2008 without and with adjusting for the effect of assignment to condition. A final exploratory

M

descriptive analysis was also done to examine differences between RCHs with more

d

versus less active children.

Participant Characteristics

Ac ce p

4.1.

te

4. Results

As presented in Table 1, a total of 799 youth between 11-18 years of age were included in Delayed or Early intervention groups. Across the three measurement waves, approximately half of the children were male. Distributions for race remained relatively constant with a greater percentage of White children compared to Black/Other. BMI scores were calculated from children’s height/weight measures and were applied using the CDC BMI cut-points (CDC, 2014). BMI scores indicate participants overall, were within normal weight range (24.0-25.2). Weight status, stratified by condition across measurement years suggest prevalence of overweight and obesity in Early and Delayed groups was approximately 40% in 2004 and 2006, and 44% in 2008. Mean age ranged

Page 15 of 38

Physical activity in Residential Children’s Homes

16

between 14.5 and 15.6 years across the three measurement waves for both Early and Delayed groups. Mean years living in the RCHs ranged from 0.9 to 1.2 years.

4.2.

Intervention Effects based on Assignment to Condition

ip t

[Table 1]

As shown in Table 2, there were no intervention effects on total METs or 2+

cr

blocks of MVPA (Model 1) in the intent-to-treat analysis. Adjusting for level of physical

the results.

4.3.

an

[Table 2]

us

activity implementation (Model 2) and global implementation (Model 3) did not change

Physical Activity Outcomes based on Grouping by Level of Implementation

M

Because of the variability in implementation of health-promoting environments

d

among organizations in both conditions, post-hoc analyses were done grouping homes

te

by PA implementation from 2004-2008, ignoring assignment to condition (Table 3). Children in “high” (N=273 children in 5 RCHs) compared to “low” (N=562 children in 19

Ac ce p

RCHs) implementing organizations reported significantly more MET-weighted blocks and a higher percent of > 2 blocks of MVPA in Year 2 (2006) only. When organizations were grouped by global implementation, children in “high” (N=366 children in 10 RCHs) compared to “low” (N=433 children in 14 RCHs) implementing organizations reported significantly more MET-weighted blocks and a higher percent of > 2 blocks of MVPA in Years 1 (2004) and 2 (2006) but not in year 3 (2008). Controlling for assignment to condition did not affect these results. [Table 3] 4.4.

Results of Baseline Characteristics of Higher PA and Global 4-Year Implementers

Page 16 of 38

Physical activity in Residential Children’s Homes

17

Baseline (2004) characteristics of RCHs described as “high implementers” across all study years were examined for organizational features existing within the social and physical environments to explore baseline characteristics associated with

ip t

“high” versus “low” PA-implementation (or PA-promoting environments, since these characteristics were present at baseline) within RCHs, characteristics that were not

cr

influenced by the ENRICH intervention. Baseline characteristics of “high” versus “low”

us

PA-promoting environments are described in Table 4 and suggest that RCHs with high PA-promoting environments can be characterized by urban versus rural settings (62%

an

versus 25% in “high” versus “low”) and as having recreation directors (75% versus 25% in “high” versus “low”); they also had higher PADEA scores (>.5) for availability of

M

PA opportunities, favorable characteristics of PA opportunities, PA policy, Wellness for

d

adult staff, and collaboration with community on PA.

te

[Table 4]

Ac ce p

When global implementation was examined at baseline (Table 5), “high” compared to “low” global RCHs can be characterized as being simple rather than complex organizations (47% versus 22% for “high” and “low”) and “low” compared to “high” as having an on-site school (78% versus 53% for “low” and “high”). “High” compared to “low” global homes also had higher mean PADEA scores (>.5) for availability of PA opportunities.

[Table 5] 5. Discussion The ENRICH intervention had no impact on physical activity of youth residing within RCHs, even after level of intervention implementation was considered. An

Page 17 of 38

Physical activity in Residential Children’s Homes

18

extensive process evaluation documented varied implementation among RCHs assigned to the intervention as well as “implementation” of physical activity-promoting practices among RCHs assigned to the Delayed condition. To explore this further we

ip t

grouped organizations by “high” and “low” physical activity-promoting environments and found that some organizations were promoting physical activity, with more active

cr

residents at baseline. These organizations can be characterized as having more PA

us

opportunities, more supportive PA policies and practices at baseline, and having dedicated recreation staff.

an

Experts recommend multi-level interventions assess for changes within the physical and social environments where children are typically situated and through

M

which health behaviors can be influenced (Boon & Clydesdale, 2005; Brown &

d

Summerbell, 2008; Elder et al., 2007; Kriemler, Meyer, Martin, Van Sluijs, Anderson, et

te

al., 2011; Pate et al., 2005, Sallis et al., 2003). ENRICH was a structural intervention that worked with adult staff as change agents to create a more health-promoting

Ac ce p

environment in the RCH, which was hypothesized to result in increased PA among children. Based on theory and expert recommendation, we anticipated that working with organizational change agents to create PA-promoting environments that could be adapted to setting and population needs would increase PA among the children residing there, but this was not the case. There are several issues to consider with these results. A lack of intervention effect can be attributed to theory failure, implementation failure or simply an ineffective intervention. ENRICH aimed to change PA behavior by targeting structures within the physical and social environment and was guided by a conceptual framework based on the structural ecologic model. The lack of significant

Page 18 of 38

Physical activity in Residential Children’s Homes

19

findings may suggest that the theory was inadequate (theory failure) for this organizational setting. However, given that effects were observed when RCHs were grouped by implementation of a PA-promoting environment, based on theory, theory

ip t

failure seems unlikely. It is possible that delivery of the intervention was not sufficient for producing individual behavior change for this population, especially since

cr

implementation varied greatly between homes. However, implementation was

us

documented by a comprehensive process evaluation and subsequently controlled for in the analysis and did not change the results. Lastly, these results could be a reflection of

an

the transient nature of the population who were exposed the RCH environments for varying periods of time. We adjusted for length of residence in analyses which did not

M

affect results. It seems likely, then, that the intervention was wholly ineffective or, more

d

likely, that different intervention approaches may be more effective for organizations in

te

“high” versus “low” PA-promoting environments at baseline. This approach would require conducting an organizational “readiness” or “capacity” assessment at baseline

Ac ce p

to inform intervention development. Ehlers and colleagues recently reported that a priori assessment of school community readiness was useful in explaining differences in PA outcomes following a school-recess physical activity intervention (Ehlers, Huberty, & Beseler, 2013). However, homogenous measures of organizational readiness or capacity are not likely generalizable to all organizational settings. At baseline, 62% of foster youth in Early and 68% in Delayed groups reported engaging in MVPA which remained relatively stable across subsequent measurement years. This is much higher than national prevalence estimates of 42% and 50% for children of similar age, race, and gender (Centers for Disease Control, 2011; Troiano,

Page 19 of 38

Physical activity in Residential Children’s Homes

20

Berrigan, Dodd, Masse, Tilert, et al., 2008). Thus it may have been difficult to observe an increase in activity over time, even if environments changed significantly (Hortz & Petosa, 2006; Shadish et al., 2001).

ip t

Strengths of this study include the comprehensive process evaluation, extensive implementation and contextual monitoring, use of a valid instrument for self-reported

cr

PA, and four-year time frame of this study. Examining outcome data with baseline

us

PADEA and OA data creates a more complete, contextually-based understanding of study outcomes. Given that foster youth are predisposed for many physical, mental, and

an

social health problems (Curtis, Alexander, & Lunghofer, 2001; James et al., 2004; Zlotnick et al., 2012), health promotion efforts designed to impact current and life-long

M

health behaviors is needed, especially longitudinal research (Curtis et al., 2001). The

d

results from this study are applicable to other organizations that provide services to

te

disadvantaged and vulnerable youth including therapeutic foster care, short-and-longterm care facilities, as well as correctional facilities. Based on previous

Ac ce p

recommendations (Curtis et al., 2001), this is the first study to have used an ecologic framework to address this population within the organizational context of the RCH, targeting the physical and social environments in which to impact health outcomes of youth served.

This study has limitations. Although ENRICH began with an optimal group randomized design, we were unable to maintain this design as reflected by widespread “implementation” of the PA-promoting environment independent of assignment to condition. In retrospect, a stratified design (for “high” and “low” PA-promoting environments) would have been a stronger approach, although the organizational

Page 20 of 38

Physical activity in Residential Children’s Homes

21

sample size (N=29) was already small. Due to the inherent contextual differences within individual RCH organizations, these findings are not likely representative to group homes situated in other geographic regions. Other limitations include the use of cross-

ip t

sectional, self-reported physical activity at the individual level; however, the 3-DPAR has been validated with accelerometers (McMurray et al., 2004, Pate et al., 2003). Due

cr

to the transient nature and complexity of the population and RCH setting, a more

us

extensive approach using objective data collection methods (i.e. accelerometry) was not feasible or appropriate in this setting. Moreover, the use of accelerometers was

an

explicitly discouraged by RCH staff during the initial ENRICH planning meeting. Although the instruments used to assess the organizational climate and characteristics

M

(OA and PADEA) yielded adequate reliability, future validation studies are needed.

d

Furthermore, additional work should examine methods to assess organizational

5.1.

Lessons Learned

te

capacity or readiness (Ehlers et al., 2013).

Ac ce p

Within the group home setting, it is clear that implementation does not occur in an ordered and uniform process and is influenced by many contextual factors with organizational environments (Cradock et al., 2007, Durlak & DuPre, 2008, Ehlers et al., 2013, Elder et al., 2007, Foster-Fishman et al., 2007, Sallis et al., 2003, Authord et al., 2013). The ENRICH intervention had no measureable effect on MVPA at the individual level; however, we found that some organizations were promoting PA based on our framework at baseline and that these RCHs had more children reporting MVPA. Interventions targeting organizational changes to impact population health behaviors should devote careful attention and provide adequate resources to effectively monitor

Page 21 of 38

Physical activity in Residential Children’s Homes

22

and document implementation. Moreover, contextual factors within organizational settings including the PA environment and organizational capacity to implement health promotion programs are important to consider and should be assessed prior to

ip t

implementation.

The use of multiple evaluation data sources across several organizational levels

cr

was vital for understanding the contextual influences that impacted ENRICH and

us

facilitated the examination of implementation differences among RCH organizations and resulting effect on PA outcomes. Although speculative, it is feasible that the items used

an

to measure global and PA implementation reflects different organizational characteristics which directly or indirectly support PA behavior. Here, global

M

implementation may capture perceived organizational structures and supports that were

d

indicative to a more general health promoting environment (e.g. “How would you rate

te

your progress implementing ENRICH?”) versus specific PA characteristics. Future work is needed to examine the different organizational climates and cultures which influence

Ac ce p

implementation of health promotion programs and changes to the physical and social PA environment.

Page 22 of 38

Physical activity in Residential Children’s Homes

23

Acknowledgements: We thank the RCHs and RCH staff and residents who worked with us on the ENRICH

Ac ce p

te

d

M

an

us

cr

ip t

project.

Page 23 of 38

Physical activity in Residential Children’s Homes

24

References Ainsworth, B. E., Haskell, W. L., Whitt, M. C., Irwin, M. L., Swartz, A. M., Strath, S. J., et al. (2000). Compendium of physical activities: an update of activity codes and

ip t

MET intensities. Medicine and Science in Sports and Exercise, 32, S498-504. Blankenship, K. M., Friedman, S. R., Dworkin, S., & Mantell, J. E. (2006). Structural

cr

interventions: concepts, challenges and opportunities for research. Journal of

us

Urban Health, 83, 59-72. DOI: 10.1007/s11524-005-9007-4.

Boon, C. S., & Clydesdale, F. M. (2005). A review of childhood and adolescent obesity

an

interventions. Critical Reviews in Food Science and Nutrition, 45, 511-525. DOI: 10.1080/10408690590957160.

M

Booth, S. L., Sallis, J. F., Ritenbaugh, C., Hill, J.O., Birch, L.L., Frank, L.D., et al. (2001).

d

Environmental and societal factors affect food choice and physical activity:

te

rationale, influences, and leverage points. Nutrition Reviews, 59, S21-39. Brown, T., & Summerbell, C. (2008). Systematic review of school-based interventions

Ac ce p

that focus on changing dietary intake and physical activity levels to prevent childhood obesity: an update to the obesity guidance produced by the National Institute for Health and Clinical Excellence. Obesity Reviews, 10, 110-141. DOI: 10.1111/j.1467-789X.2008.00515.x. Centers for Disease Control and Prevention. (2011). Youth Risk Behavior Surveillance System: 2011 National Overview. Available: http://www.cdc.gov/healthyyouth/yrbs/pdf/us_overview_yrbs.pdf. [Accessed June 10, 2013].

Page 24 of 38

Physical activity in Residential Children’s Homes

25

Centers for Disease Control and Prevention. (2011). Healthy Weight – it’s not a diet, it’s a lifestyle! About BMI for Children and Teens. Available: http://www.cdc.gov/healthyweight/assessing/bmi/childrens_bmi/about_childrens_

ip t

bmi.html. [Accessed May 20, 2014].

Chen, H.T. (2004). Practical program evaluation: Assessing and improving planning,

cr

implementation, and effectiveness. Thousand Oaks: SAGE Publications.

us

Children’s Defense Fund. (2013). Policy Priorities: Foster Care. Available:

http://www.childrensdefense.org/policy-priorities/child-welfare/foster-care/

an

[Accessed June 12 2012].

Cohen, D. A., Scribner, R. A., & Farley, T. A. (2000). A structural model of health

M

behavior: a pragmatic approach to explain and influence health behaviors at the

d

population level. Preventive Medicine, 30, 146-154. DOI:

te

10.1016/j.bbr.2011.03.031.

Commers, M. J., Gottlieb, N., & Kok, G. (2007). How to change environmental

Ac ce p

conditions for health. Health Promotion International, 22, 80-87. DOI: 10.1093/heapro/dal038.

Cradock, A. L., Melly, S. J., Allen, J. G., Morris, J. S., & Gortmaker, S. L. (2007). Characteristics of school campuses and physical activity among youth. American Journal of Preventive Medicine, 33, 106-113. DOI: 10.1016/j.amepre.2007.04.009. Curtis, P. A., Alexander, G., & Lunghofer, L. A. (2001). A Literature Review Comparing the Outcomes of Residential Group Care and Therapeutic Foster Care. Child and Adolescent Social Work Journal, 18, 377-392. DOI: 10.1023/A:1012507407702.

Page 25 of 38

Physical activity in Residential Children’s Homes

26

Authora et al. (2009). Durlak, J. A., & Dupre, E. P. (2008). Implementation matters: a review of research on the influence of implementation on program outcomes and the factors affecting

ip t

implementation. American Journal of Community Psychology, 41, 327-350. DOI: 10.1007/s10464-008-9165-0.

cr

Ehlers, D. K., Huberty, J. L., & Beseler, C. L. (2013). Is school community readiness

us

related to physical activity before and after the Ready for Recess intervention? Health Education Research, 28, 192-204. DOI: 10.1093/Her/Cys102.

an

Elder, J. P., Lytle, L., Sallis, J. F., Young, D. R., Steckler, A., Simons-Morton, D., et al. (2007). A description of the social-ecological framework used in the trial of

M

activity for adolescent girls (TAAG). Health Education Research, 22, 155-165.

d

DOI: 10.1093/her/cyl059.

te

Authorb et al. (2009).

Foster-Fishman, P. G., Nowell, B., & Yang, H. L. (2007). Putting the system back into

Ac ce p

systems change: a framework for understanding and changing organizational and community systems. American Journal of Community Psychology, 39, 197215. DOI: 10.1007/s10464-007-9109-0. Authorc et al. (2011).

Giles-Corti, B., & Donovan, R. J. (2002). The relative influence of individual, social and physical environment determinants of physical activity. Social Science and Medicine, 54, 1793-1812. Harachi, T. W., Abbott, R. D., Catalano, R. F., Haggerty, K. P., & Fleming, C. B. (1999). Opening the black box: Using process evaluation measures to assess

Page 26 of 38

Physical activity in Residential Children’s Homes

27

implementation and theory building. American Journal of Community Psychology, 27, 711-731. Hawe, P., Shiell, A., & Riley, T. (2004). Complex interventions: how "out of control" can

ip t

a randomised controlled trial be? British Medical Journal, 328, 1561-1563. DOI: 10.1136/bmj.328.7455.1561.

cr

Hortz, B., & Petosa, R. (2006). Impact of the "planning to be active" leisure time

us

physical exercise program on rural high school students. Journal of Adolescent Health, 39, 530-535. DOI: 10.1016/j.jadohealth.2006.03.015.

an

James, S., Landsverk, J., Slymen, D. J., Leslie, L. K. (2004). Predictors of outpatient mental health service use: the role of foster care placement change. Mental

M

Health Services Research, 6, 127-141. DOI: 10.1016/j.bbr.2011.03.031.

d

Koplan, J., Liverman, C. T., & Kraak, V. I. (2005). Preventing childhood obesity: health

te

in the balance, Washington D.C.: National Academy Press. Kriemler, S., Meyer, U., Martin, E., Van Sluijs, E. M., Andersen, L. B., & Martin, B. W.

Ac ce p

(2011). Effect of school-based interventions on physical activity and fitness in children and adolescents: a review of reviews and systematic update. British Journal of Sports Medicine, 45, 923-930. DOI: 10.1136/bjsports-2011-090186. Linnan, L. S., & Steckler A. (2002). Process Evaluation for Public Health Interventions and Research: An Overview. In A. Steckler, & L. Linnan (Eds.), Process Evaluation for Public Health Interventions and Research (pp. 1-23). San Francisco: Jossey-Bass. Matson-Koffman, D. M., Brownstein, J. N., Neiner, J. A., & Greaney, M. L. (2005). A site-specific literature review of policy and environmental interventions that

Page 27 of 38

Physical activity in Residential Children’s Homes

28

promote physical activity and nutrition for cardiovascular health: What works? American Journal of Health Promotion, 19, 167-193. McGraw, S. A., Sellers, D. E., Johnson, C. C., Stone, E. J., Bachman, K. J., Bebchuk,

ip t

J., et al. (1996). Using Process Data To Explain Outcomes An Illustration From the Child and Adolescent Trial for Cardiovascular Health (CATCH). Evaluation

cr

Review, 20, 291-312. DOI: 10.1177/0193841X9602000304.

us

McLeroy, K. R., Bibeau, D., Steckler, A., & Glanz, K. (1988). An ecological perspective on health promotion programs. Health Education Quarterly, 15, 351-377.

an

McMurray, R. G., Ring, K. B., Treuth, M. S., Welk, G. J., Pate, R. R., Schmitz, K. H., et el. (2004). Comparison of two approaches to structured physical activity surveys

M

for adolescents. Medicine and Science in Sports and Exercise, 36, 2135-2143.

d

Medical Research Council. (2008). Developing and evaluating complex interventions:

te

new guidance. Available: http://www.mrc.ac.uk/complexinterventionsguidance. [Accessed May 20, 2013].

Ac ce p

Pate, R. R., Ross, R., Dowda, M., Trost, S.G., & Sirard, J. (2003). Validation of a threeday physical activity recall instrument in female youth. Pediatric Exercise Science, 15, 257-265.

Pate, R. R., Ward, D. S., Saunders, R. P., Felton, G., Dishman, R. K., & Dowda, M. (2005). Promotion of Physical Activity Among High-School Girls: A Randomized Controlled Trial. American Journal of Public Health, 95, 1582-1587. DOI: 10.2105/AJPH.2004.045807. Pate, R. R., Saunders, R., Dishman, R. K., Addy, C., Dowda, M., & Ward, D. S. (2007). Long-term effects of a physical activity intervention in high school girls. American

Page 28 of 38

Physical activity in Residential Children’s Homes

29

Journal of Preventive Medicine, 33, 276-280. DOI: 10.1016/j.amepre.2007.06.005. Poland, B., Krupa, G., & McCall, D. (2009). Settings for health promotion: an analytic

Practice, 10, 505-516. DOI: 10.1177/1524839909341025.

ip t

framework to guide intervention design and implementation. Health Promotion

cr

Roemmich, J. N., Epstein, L. H., Raja, S., Yin, L., Robinson, J., & Winiewicz, D. (2006).

us

Association of access to parks and recreational facilities with the physical activity of young children. Preventive Medicine, 43, 437-441. DOI:

an

10.1016/j.ypmed.2006.07.007.

Sallis, J. F., McKenzie, T. L., Conway, T. L., Elder, J. P., Prochaska, J. J., Brown, M., et

M

al. (2003). Environmental interventions for eating and physical activity: a

te

Medicine, 24, 209-217.

d

randomized controlled trial in middle schools. American Journal of Preventive

Saunders, R. P., Evans, M. H., & Joshi, P. (2005). Developing a process-evaluation

Ac ce p

plan for assessing health promotion program implementation: a how-to guide. Health Promotion Practice, 6, 134-147. DOI: 10.1177/1524839904273387. Saunders, R. P., Ward, D., Felton, G. M., Dowda, M., & Pate, R. R. (2006). Examining the link between program implementation and behavior outcomes in the lifestyle education for activity program (LEAP). Evaluation and Program Planning, 29, 352-364. DOI: 10.1016/j.evalprogplan.2006.08.006 Authord et al. (2013).

Page 29 of 38

Physical activity in Residential Children’s Homes

30

Shadish, W. R., Cook, T. D., & Campbell, D. T. (2001). Experimental and Quasiexperimental Designs for Generalized Causal Inference, Boston: Houghton Mifflin.

ip t

Troiano, R. P., Berrigan, D., Dodd, K. W., Masse, L. C., Tilert, T., & McDowell, M. (2008). Physical activity in the United States measured by accelerometer.

cr

Medicine and Science in Sports and Exercise, 40, 181-188. DOI:

us

10.1249/mss.0b013e31815a51b3.

United States Department of Health and Human Services. (2005). Administration for

an

Children and Families National Survey of Child and Adolescent Well-Being (NSCAW): CPA Sample Component Wave 1 Data Analysis Report. Available:

d

[Accessed June 12, 2012].

M

http://www.acf.hhs.gov/programs/opre/abuse_neglect/nscaw/index.html.

te

United States Department of Health and Human Services. (2008). 2008 physical activity guidelines for Americans. Available: http://www.health.gov/paguidelines/.

Ac ce p

[Accessed April 21, 2013].

Viner, R. M., Taylor, B. (2005). Adult health and social outcomes of children who have been in public care: population-based study. Pediatrics, 115, 894-899. DOI: 10.1542/peds.2004-1311.

Washington State Department of Health. (2013). Health Data: Guidelines for Using Rural-Urban Classification Systems for Public Health Assessment. Available: http://www.doh.wa.gov/DataandStatisticalReports/DataGuidelines.aspx 2005. [Accessed June 3, 2013].

Page 30 of 38

Physical activity in Residential Children’s Homes

31

Watts, P., Phillips, G., Petticrew, M., Harden, A., & Renton, A. (2011). The influence of environmental factors on the generalisability of public health research evidence: physical activity as a worked example. International Journal of Behavioral

ip t

Nutrition and Physical Activity, 8, 128. DOI: 10.1186/1479-5868-8-128.

Zlotnick, C., Tam, T. W., Soman, L. A. (2012). Life course outcomes on mental and

cr

physical health: the impact of foster care on adulthood. American Journal of

Ac ce p

te

d

M

an

us

Public Health, 102, 534-540. DOI: 10.2105/AJPH.2011.300285.

Page 31 of 38

Physical activity in Residential Children’s Homes

32

ip t

Gregory M. Dominick is an Assistant Professor in the Department of Behavioral Health and Nutrition at the University of Delaware. He was a research assistant for ENRICH and worked on several aspects of the intervention, including process evaluation. He is currently collecting process evaluation data for a statewide health promotion initiative in Delaware. Ruth P. Saunders is an Associate Professor in the Arnold School of Public Health at the University of South Carolina. She was PI for ENRICH and has conducted process evaluation in seven large-scale intervention trials, and oversaw process evaluation in ENRICH.

us

cr

Alexandra E. Evans is an Associate Professor in the Division of Health Promotion and Behavioral Sciences at the University of Texas, School of Public Health. She was Co-PI for ENRICH and has recently secured USDA funding to investigate the effects of school garden and/or physical activity strategies for childhood obesity prevention.

an

Marsha Dowda is a Biostatistician in the Arnold School of Public Health at the University of South Carolina and conducted the statistical analysis for ENRICH. She has worked on several large-scale physical activity interventions.

Ac ce p

te

d

M

Kelli Kenison is an Assistant Research Professor in the Arnold School of Public Health at the University of South Carolina. She served as the Intervention Coordinator for ENRICH.

Page 32 of 38

Table

Table 1. Characteristics of children measured in RCHs in 2004, 2006, and 2008 Delayed

N=799

2004 N=120

2006 N=115

2008 N=158

2004 N=124

2006 N=146

Males

52.5%

55.7%

52.5%

49.2%

Race Black White Other

26.7% 50.0% 23.3%

27.0% 44.4% 28.7%

34.8% 39.2% 26.0%

30.7% 54.0% 15.3%

ip t

Early

Age

14.9 (1.7)

15.0 (1.9)

14.7 (1.9)

24.5 (6.3)

24.0 (6.8)

24.5 (5.7)

72 (60.0) 20 (16.6) 28 (23.3)

69 (60.0) 22 (19.1) 24 (20.8)

88 (55.7) 30 (18.9) 40 (25.3)

0.9 (1.1)

0.9 (0.9)

1.0 (1.1)

a

th

th

M

Weight Status Healthy weight Overweight Obese Years in current RCH

26.7% 50.7% 22.6%

23.5% 52.9% 23.5%

15.6 (1.6)

14.5 (1.8)

15.3 (1.9)

24.7 (6.2)

24.3 (5.1)

25.1 (7.0)

71 (57.2) 28 (22.5) 25 (20.1)

87 (59.5) 31 (21.2) 28 (19.1)

76 (55.8) 23 (16.9) 37 (27.2)

1.2 (2.0)

1.2 (1.8)

1.0 (1.2)

cr

48.5%

an

a

54.8%

us

BMI

th

2008 N=136

th

Ac ce p

te

d

Healthy weight = < 85 %, overweight = > 85 % but < 95 %, obese = >95 %

Page 33 of 38

Table 2. LS Means (SE) of Physical Activity at 3 Time Points by Condition by Time, Condition, and Condition and Time (Model 1); Adjusting for Physical Activity Implementation (Model 2); and Adjusting for Global Implementation (Model 3) INTERACTION Model/Variable

Early

F-value

2006 64.0 (1.7)

2008 62.5 (1.7)

2004 61.9 (1.6)

2006 61.1 (1.6)

63.9%

70.8%

71.6%

64.9%

61.3%

Model 2

2004

2006

2008

2004

2006

2008

Total METs (±SD)

62.2 (2.1)

63.6 (1.7)

61.8 (1.8)

62.2 (1.9)

60.4 (1.7)

63.7%

70.5%

70.5%

64.2%

Model 3

2004

2006

2008

Total METs (±SD)

62.3 (2.1)

64.3 (2.0)

63.7%

70.5%

2+ Blocks of MVPA (%)

2+ Blocks of MVPA (%) a

.48

0.75

.47

62.4 (1.6)

1.19

.32

59.1%

70.7%

1.22

.31

2004

2006

2008

61.6 (1.9)

62.4 (1.9)

60.7 (1.7)

62.4 (1.6)

1.42

.26

64.2%

59.1%

70.7%

1.23

.31

70.5%

70.1%

cr

us

2+ Blocks of MVPA (%)

0.76

an

Total METs (±SD)

2008 62.2 (1.5)

ip t

2004 62.0 (1.8)

p-value

M

Model 1

Delayed

Ac ce p

te

d

Model 1. Controlling for age, sex, BMI, race, school onsite, time in home, recreation director, with home nested in group b Model 2. Also adjusting for physical activity implementation c Model 3. Adjusting for age, sex, BMI, race, time in home and global implementation

Page 34 of 38

Table 3. LS Means (SE) for PA and Global Implementation across 3 measurement years

Global Implementation Total MET blocks (±SD) 2+ blocks MVPA (%) a

Implementation by

(2004)

(2006)

(2008)

Time Interaction

Low

High

Low

High

Low

High

n=174

n=70

n=185

n=76

n=203

n=91

60.9 (1.2)

65.0 (2.2)

61.2 (1.2)

67.4 (2.2)

61.1%

74.0%

61.4%

82.5%

Low

High

Low

n=137

n=107

59.9 (1.3)

64.6 (1.6)

55.7%

75.6%

†

†

†

62.8 (1.2)

F-value

ip t

2+ blocks MVPA (%)

Year 3

p-value

61.7 (2.1)

5.12

.01

72.1%

69.9%

4.12

.03

High

Low

High

F-value

p-value

n=142

n=119

n=154

n=140

60.9 (1.3)0

65.3 (1.5)

63.6 (1.2)

61.2 (1.4)

6.53

.01

69.4%

6.06

.01

†

cr

Total MET blocks (±SD)

Year 2

us

PA Implementation

Year 1

†

†

an

Implementation

58.2%

78.6%

73.1%

Controlling for age, sex, BMI, race and years in home, with home as a random variable †Low and High implementing homes differ, p