A Faith-Integrated Physical Activity Intervention and Cardiometabolic Risk in African American Women

Abstract

PURPOSE:

To determine the effects of a 10-month secular (SEC) versus faith-integrated (FI) community-based physical activity (PA) intervention on cardiometabolic risk factors in low active, African-American women.

METHODS:

Participants (age: 55.4±11.6 years; body mass index (BMI): 36.0±7.9 kg/m²; average baseline steps/day: 3,807±1,250) from a larger study (n=418) participated in a sub-study to measure cardiometabolic disease indicators (primary outcomes) and PA (secondary outcomes) pre- and post-intervention (SEC: n=42; FI: n=43). Height, weight, waist and hip circumferences, resting blood pressure, hemoglobin A1c, average steps/day, sedentary behavior, light-intensity physical activity (LPA), and moderate-to-vigorous-intensity physical activity (MVPA) were acquired at baseline and 10 months. Multivariate generalized linear mixed models that included churches as a random effect were used to compare mean changes in outcomes at 10 months between the two study groups (α=0.05).

RESULTS:

The FI group showed significant time effects for weight (93.4±2.4 to 92.2±2.3 kg), BMI (35.7±1.0 to 35.3±1.0 kg/m²), and waist circumference (106.9±2.2 to 103.8±2.5 cm), while the SEC group had a significant time effect for hip circumference (121.6±1.9 to 119.9±1.7 cm). There were no time effects in either group for blood pressure, hemoglobin A1c, steps/day, sedentary time, or MVPA. FI significantly decreased LPA in both 1-minute activity bouts (641±13 to 588±16 minutes/day) and 10-minute bouts (536±11 to 479±15 minutes/day). There were no significant differences between SEC and FI for any variable.

CONCLUSIONS:

The improvements in body weight, BMI, and waist circumference shown after the FI intervention could have long-term implications on cardiometabolic health, particularly if exercise is continued. Further research is needed to examine the effects of culturally-relevant interventions on chronic disease indicators in African-American women, particularly those established as high risk for cardiometabolic disease.

INTRODUCTION

Physical inactivity and obesity, as well as the comorbidities related to both of these issues continue to present health concerns for U.S. adults in disparate proportions. Data have shown that only 46% of African-Americans (AAs) meet federal PA guidelines, compared to 56% of whites (1). It has also been shown that AA women engage in lower levels of total physical activity (PA) and less moderate-to-vigorous physical activity (MVPA) than AA men, white women, and white men (16), putting them at increased risk for obesity and related comorbidities. In addition, AA women have considerably higher body mass index (BMI) values when compared to women of other ethnic/racial backgrounds (12, 25). Recent National Health and Nutrition Examination Survey (NHANES) statistics show that 57% of non-Hispanic black women are obese (≥30 kg/m²) compared to the next highest groups of women (Hispanic 47%; white 39%) (12). Further, 29% of AA women were categorized as class II or III obese (BMI ≥35 kg/m²) compared to 20% and 16% of Hispanic and white women, respectively (26). It has been established that when overweight and obese individuals adopt PA and lose weight, improvements in multiple risk factors related to cardiovascular disease and type II diabetes result (10). Also notable is evidence that AAs, in general, have higher incidence of chronic diseases such as hypertension and type II diabetes (5, 24). Given this compilation of health concerns that persists in AA women, identifying successful and sustainable exercise and PA interventions is important for the longevity of this group.

Faith-based PA interventions launched in the AA community have gained attention in recent years as they may be more culturally relevant and effective in changing health-related behaviors in this population compared to traditional PA interventions. This is suggested because the church provides a critical foundation for social guidance in AA communities (20). As such, positive health messaging and programming that is rooted in the church environment may elicit more long-term behavior changes and, in turn, assist in reducing some of the health disparities mentioned above. Data from previous faith-based interventions have been successful in improving body weight, body composition, physical fitness, and PA in the AA community (27). However, there are currently few studies that examine the effects of faith-based PA interventions on risk factors associated with cardiometabolic disease in AAs (11, 29, 32, 41–43). Intervention characteristics vary widely in this area of study. Several studies use shorter intervention periods (8 weeks to 6 months) (11, 32, 42, 43), include both women and men (29, 32, 41, 42), and measure a range of variables that influence cardiometabolic disease (BMI, body composition, blood pressure, blood biomarkers, etc.). The current study adds to this body of literature by examining the effects of a long-term intervention on dependent variables indicative of critical health issues (hypertension and diabetes) specifically in AA women.

The purpose of the current study was to determine the effects of a 10-month secular (SEC) versus faith-integrated (FI) PA intervention on cardiometabolic risk factors in inactive AA women. The cardiometabolic risk factors being examined include BMI, waist circumference, resting blood pressure, and hemoglobin A1c. Secondarily, this study also examined measures of PA in sub-study participants over 10 months. When considering the translational science continuum, this study provides clinical research to test our intervention’s effect on selected cardiometabolic outcomes. The current intervention could influence the design and implementation of culturally relevant community-based PA interventions in AA women. Importantly, the current study’s use of community health workers and delivering the intervention sessions in a centralized community organization increased the potential for future translation and broader dissemination.

METHODS

Parent Study Overview

The current study is a pilot sub-study of the Learning and Developing Individual Exercise Skills (L.A.D.I.E.S.) for a Better Life study. The L.A.D.I.E.S. study is a cluster randomized, controlled trial that tested three strategies to increase PA for inactive active AA women. All parent study and sub-study procedures were reviewed and approved by Copernicus Group Independent Review Board. A detailed account of the L.A.D.I.E.S. parent study methodology was published previously (36, 39). However, for the readers’ ease and to ensure accessibility, an overview of parent study is warranted.

Self-identified, predominantly AA churches were recruited to enter this 22-month PA intervention study. The intervention included a 10-month intervention period followed by a 12-month maintenance period. Each church identified one female congregation member as the study liaison, who was not a study participant, but assisted study staff with disseminating study information, recruiting 12–15 participants, and scheduling intervention sessions. Participant inclusion criteria included self-described regular attendees of the church, self-identified AA women, at least 18 years of age, and not currently meeting national PA recommendations (≥150 minutes/week of MVPA, or ≥75 minutes/week of vigorous activity). Participants were excluded for physical limitations or medical conditions that would prevent participation in daily PA, or if they planned to relocate within the next two years.

Per the cluster randomized design, each church was assigned to one of three intervention groups. The SEC and FI groups received 24 intervention sessions that were up to 90 minutes each, which included 60 minutes of group discussion/education and 15–30 minutes of group PA. Over the 10-month intervention, the first 16 sessions were conducted weekly, sessions 17–20 were conducted bi-weekly, and sessions 21–24 were conducted monthly, in an effort to assist participants with their transition from the structured intervention to independent PA. All L.A.D.I.E.S. educational and exercise sessions took place in the recreational center of one designated, centrally located church in the local community. The SEC intervention used general motivational readings to emphasize concepts, while the FI intervention included biblical content and faith-based principles in its educational components. The self-guided group served as the control condition, where participants received standardized written materials promoting PA. The educational components of the SEC and FI sessions were led by female community health workers recruited from the local community and hired as study staff. These community health workers were familiar with the culture in AA churches and had previous experience leading groups. The staff who led the SEC group sessions had educational background in health promotion or health-related behavior changes. The staff hired to lead the FI group sessions were familiar with biblical ideologies and were comfortable interpreting and teaching biblical passages. The exercise components of each intervention session were led by separate female group exercise leaders, also hired as study staff from the local community. Activities included a variety of group exercise formats such as dance aerobic exercise and kickboxing offered at varied intensity levels. While supervised exercise only took place during the 24 intervention sessions, group leaders continually encouraged participants to strive for the overarching goal of increasing daily walking by ≥2,000 steps/day. If 10,000 steps/day were obtained, participants were encouraged to maintain this level of PA or to continue increasing.

Parent study intervention data were collected at baseline (pre-intervention), 10 months (immediately post-intervention), and 22 months (immediately following the 12-month maintenance period). The main outcome of the parent study was objectively-measured PA.

Sub-Study Recruitment & Overview

The current sub-study examined intervention effects on participants in the SEC and FI groups only. As the parent project did not propose to collect data on health markers directly associated with cardiometabolic disease, the sub-study was designed to provide additional information regarding the effectiveness of the SEC and FI curricula on cardiometabolic risk factors among participants assigned to the active intervention. A subsample of churches taken from the L.A.D.I.E.S. parent study was used to recruit participants for the sub-study. This subsample of churches included the first twelve churches that entered the parent study and were randomized to one of the active intervention groups (SEC or FI). Upon entry into the parent study, all participants who were members of the identified subsample churches were approached at their site’s study introduction meeting, initial data collection, or via telephone, and were presented with information about the sub-study. In addition to the parent study’s inclusion criteria, sub-study participation required that individuals be non-smokers. When recruiting, investigators communicated that sub-study participation was completely optional and had no bearing on the women’s parent study participation.

The measurements required for participation in the sub-study took place immediately pre- (baseline) and post-intervention (10 months). No sub-study data were collected at the 22-month follow-up. Participants who volunteered to enter the sub-study completed a separate informed consent form (relevant to the additional measurements) prior to moving forward.

Cardiometabolic Risk Factors

Participants then made an appointment to meet study staff at a designated patient service center in Winston-Salem, NC. This appointment took place in the morning after a 12-hour fast (except water). Participants were also asked to refrain from taking any prescribed blood pressure medications the day of the appointment (12–24 hours abstinence, depending on when they typically took their medication), but were encouraged to resume normal dosage immediately after the appointment. After at least 5 minutes of seated rest, an indirect resting blood pressure measurement using a digital blood pressure monitor was taken for participants using standard guidelines for accuracy (28). Body weight in indoor clothing was then measured on a portable scale (SECA; Hamburg, Germany). BMI was calculated using this weight measurement and the height measurement that was previously acquired for the parent study using a portable stadiometer (SECA; Hamburg, Germany). Waist and hip circumferences were then measured using a fiberglass measuring tape with a tension handle, per guidelines outlined by the American College of Sports Medicine (28). Finally, a venous blood sample was collected by a trained phlebotomist and processed at the patient service center, then transported to a central laboratory for analysis the same day. Participants received cash incentives up to $20 for sub-study participation ($10 after completing the pre-intervention blood draw and $10 after completing the post-intervention blood draw). This incentive was earned in addition to any incentives received for participation in the parent study (39).

Physical Activity Assessment

PA data collected during the parent study were used for participants in the sub-study. These data were collected using two PA monitoring devices worn concurrently for 14 consecutive days at baseline and immediately after the 10-month intervention. Average steps/day were measured using a New Lifestyles Lifecorder 2160 (New Lifestyles; Lees Summit, MO). The Lifecorder was worn on the hip/waistband and positioned along the midline of the thigh. Participants were instructed to wear this device during all waking hours, except when participating in water activities (showering, swimming, etc.). The Lifecorder was chosen for its combination of validated accuracy in estimating step counts (2) and its data storage capacity (up to 60 days). The duration and intensity of PA bouts were measured using an Actical® accelerometer (Phillips Respironics, Inc.; Bend, OR), worn on the dorsal side of the wrist chosen by participants. The Actical® was chosen, in-part, due to it being one of few research-grade devices of this nature that is waterproof. As such, investigators were able to maximize wearing compliance by mounting this device with a locking wrist strap so participants could wear it 24 hours/day for the entire 14 days (including when showering, swimming, etc.). A complete description of the procedures used to screen and process the Actical® data was published previously (39). In summary, data were downloaded using Actical® software (Actical® Version 2.12) then screened for wear compliance (at least 23 hours/day of continuous wear time) and proper device function. A previously published algorithm (14) was used to convert the raw data (15-second epochs) into minute-by-minute values of activity energy expenditure, which were then reanalyzed for the duration of time (both 1-min and 10-min activity bouts) spent engaged in light (<0.0385 kcal/kg/min and >50 counts/min), moderate (0.0385 – 0.0895 kcal/kg/min) and vigorous-intensity PA (>0.0895 kcal/kg/min) using cut -points established from a sample with similar characteristics as the current study (15).

Statistical Analysis

Baseline sample characteristics were summarized using descriptive statistics (mean ± standard deviation [SD], frequency [%]). Baseline characteristics were compared between respondents and non-respondents (dropouts) to determine if they differed systematically on values of baseline variables using chi-square tests and t-tests adjusted for clustering within churches. Analyses of outcomes were conducted using multivariable generalized linear mixed models (GLMM) that included churches as a random effect. These GLMMs compared mean changes in outcomes at 10 months between the two study groups to assess the impact of intervention. In addition, a separate set of GLMMs that included only baseline values of dependent variables as a fixed covariate were used to compare adjusted differences in mean changes in outcomes between groups. Additionally, an intention-to-treat analysis was used to evaluate the effects of the intervention on all participants who initially entered the sub-study. The principle of last observation carried forward was used, where missing data points were filled using data collected closest to the time of study dropout. Analyses took into account clustering within the churches. Analyses were conducted using SAS® analytics software, Version 9.4 (SAS Institute, Inc.; Cary, NC) and SPSS Statistics software, Version 22.0 (IBM; Armonk, NY). Significance was accepted at the 0.05 alpha level.

RESULTS

Figure 1 provides a flow diagram of participant progression through the sub-study. Of 199 parent study participants in the subset of sub-study churches, 104 women or 52% (SEC: 58; FI: 46) volunteered to undergo the additional measurements for sub-study participation. Of these, 16 (28% dropout) and 3 (7% dropout) discontinued participation in the parent study or sub-study in the SEC and FI groups, respectively (i.e. 72% adherence and 93% adherence for the SEC and FI groups, respectively). Compared to those who completed the sub-study (n=85), the women who did not complete follow-up data collection for the sub-study (n=19) were significantly younger (51.0 versus 56.4 years, respectively; P=0.032) and acquired more MVPA (40 versus 20 minutes/day in 1-min bouts, and 18 versus 8 minutes/day for 10-min bouts; P=0.002). There were no significant differences in study completers versus dropouts for any other baseline variable.

Figure 1:

Flow diagram of participant progression through the study. SEC=secular intervention churches; FI=faith-integrated intervention.

Data presented in Table 1 show descriptive characteristics of the entire sample of L.A.D.I.E.S sub-study participants at baseline. On average, participants were middle-aged (mean ± SD: 55.4 ± 11.6 years), sedentary (accumulated <5,000 steps per day) (33), class II obese (BMI 35.0–39.9 kg/m²), categorized as “high risk” for chronic disease based on waist circumferences (90–110 cm) (28), hypertensive (systolic and/or diastolic blood pressures ≥130/80 mmHg, respectively) (35), and had hemoglobin A1c levels indicative of diabetes (≥6.5%) (3). Of the variables that were also measured in the parent study (age, BMI, blood pressure, PA), baseline characteristics of sub-study participants were comparable (37).

Table 1:

L.A.D.I.E.S. for a Better Life sub-study participant characteristics at baseline (N=104).

VARIABLES	MEAN ± SD	RANGE
Age (years)	55.4 ± 11.6	19 – 85
Height (m)	1.62 ± 0.07	1.40 – 1.88
Weight (kg)	95.1 ± 22.4	52.0 – 156.6
Body mass index (kg/m2)	36.0 ± 7.9	20.2 – 58.8
Waist circumference (cm)	105.3 ± 15.9	70.0 – 145.8
Hip circumference (cm)	122.3 ± 15.8	87.8 – 160.0
Systolic blood pressure (mmHg)	132 ± 17	103 – 175
Diastolic blood pressure (mmHg)	83 ± 11	55 – 116
Hemoglobin A1c (%)	7.4 ± 1.6	5.6 – 14.9
Average steps/day^	3,807 ± 1,250	1,369 – 6,868
Sedentary time (minutes/day)#
1-minute bouts	778 ± 174	289 – 1066
10-minute bouts	692 ± 187	521 – 1000
LPA (minutes/day)#
1-minute bouts	639 ± 162	374 – 1648
10-minute bouts	528 ± 158	251 – 1411
MVPA (minutes/day)#
1-minute bouts	23 ± 28	0 – 140
10-minute bouts	10 ± 16	0 – 93

LPA=light-intensity physical activity; MVPA=moderate-to-vigorous physical activity.

^{^}Pedometer-assessed (N=100).

^#Accelerometer-assessed (N=97).

Table 2 shows pre- and post-intervention descriptive characteristics and variable comparisons of participants who completed the sub-study. There were no significant differences between the SEC and FI groups at baseline. FI showed significant time effects for body weight (94.4 ± 3.3 to 93.2 ± 3.2 kg; P=0.017; partial η²=0.04), BMI (35.7 ± 1.0 to 35.3 ± 1.0 kg/m²; P=0.017; partial η²=0.03), waist circumference (106.9 ± 2.2 to 103.8 ± 2.5 cm; P=0.030; partial η²=0.14), and light-intensity physical activity (LPA) for both 1-minute activity bouts (641 ± 13 to 588 ± 16 minutes/day; P=0.033; partial η²=0.04) and 10-minute bouts (536 ± 11 to 479 ± 15 minutes/day; P=0.016; partial η²=0.04). SEC had a significant time effect for hip circumference (121.6 ± 1.9 to 119.9 ± 1.7 cm; P=0.039; partial η²=0.08). There were no significant time effects for blood pressure, hemoglobin A1c, steps/day, sedentary time, or MVPA in either group. However, the FI group showed a positive trend in steps/day over time (3,884 ± 195 to 4,973 ± 459 steps/day; P=0.086), that was not as notable in the SEC group (3,727 ± 80 to 4,591 ± 416 steps/day; P=0.106). There were no significant differences between SEC and FI for any variable. When the analyses were conducted after adjusting for age and baseline PA, no results were significantly altered compared to the unadjusted analyses.

Table 2:

Pre- and post-intervention descriptive characteristics and variable comparisons of participants who completed the L.A.D.I.E.S. for a Better Life sub-study (N=85).

	SEC (n=42)			FI (n=43)			Difference in Mean Change (95% CI)
VARIABLES	PRE	POST	Change in Mean (95% CI)	PRE	POST	Change in Mean (95% CI)
Descriptives & Cardiometabolic Risk Factors
Weight (kg)¥	95.2 ± 3.2	95.1 ± 3.4	−0.07 (−0.88 – 0.75)	93.4 ± 2.4	92.2 ± 2.3 *	−1.26 (−2.15 – 0.37)	1.19 (−2.63 – 0.25)
Body mass index (kg/m2)	35.8 ± 1.2	35.8 ± 1.3	−0.01 (−0.32 – 0.29)	35.7 ± 1.0	35.3 ± 1.0 *	−0.45 (−0.76 – 0.13)	0.44 (−0.99 – 0.11)
Waist circumference (cm)	103.9 ± 2.1	101.4 ± 1.6	−2.50 (−5.65 – 0.65)	106.9 ± 2.2	103.8 ± 2.5 *	−3.14 (−5.81 – 0.48)	0.64 (−4.01 – 2.48)
Hip circumference (cm)	121.6 ± 1.9	119.9 ± 1.7 *	−1.74 (−3.35 – 0.12)	122.6 ± 1.2	121.3 ± 2.1	−1.33 (−5.13 – 2.47)	−0.40 (−3.21 – 2.99)
Systolic BP (mmHg)	132 ± 2	133 ± 3	1.49 (−2.49 – 5.47)	134 ± 4	138 ± 1	4.00 (−6.37 – 14.37)	−2.51 (−5.17 – 10.07)
Diastolic BP (mmHg)	83 ± 3	83 ± 3	0.72 (−2.35 – 3.79)	82 ± 2	84 ± 1	1.50 (−1.97 – 4.96)	−0.77 (−3.06 – 4.61)
Hemoglobin A1c (%)	7.5 ± 0.2	7.6 ± 0.2	0.13 (−0.16 – 0.42)	7.2 ± 0.2	7.4 ± 0.4	0.17 (−0.21 – 0.54)	−0.04 (−0.30 – 0.44)
Physical Activity
Average steps/day^	3,727 ± 80	4,591 ± 416	864 (−248 – 1,976)	3,884 ± 195	4,973 ± 459	1,089 (−246 – 2,424)	−225 (−1,020 – 1,417)
Sedentary time (min/day)#
1-minute bouts	820 ± 25	835 ± 21	15.43 (−56.5 – 87.37)	779 ± 14	822 ± 22	42.88 (−24.1 – 109.9)	−27.5 (−52.7 – 107.6)
10-minute bouts	736 ± 25	758 ± 22	22.03 (−54.3 – 98.37)	695 ± 16	743 ± 25	48.62 (−27.4 – 124.6)	−26.6 (−59.9 – 113.1)
LPA (min/day)#
1-minute bouts	599 ± 22	584 ± 23	−14.70 (−89.10 – 59.67)	641 ± 13	588 ± 16 *	−53.0 (−99.0 – 6.95)	38.27 (−114 – 37.03)
10-minute bouts	487 ± 22	482 ± 23	−5.70 (−76.90 – 65.48)	536 ± 11	479 ± 15 *	−56.40 (−95.70 – 17.0)	50.65 (−124 – 22.54)
MVPA (min/day)#
1-minute bouts	22 ± 5	21 ± 6	−0.73 (−11.90 – 10.47)	21 ± 4	31 ± 17	10.12 (−35.40 – 55.61)	−10.8 (−25.0 – 66.29)
10-minute bouts	8 ± 2	9 ± 4	0.78 (−7.41 – 8.96)	8 ± 2	19 ± 14	10.91 (−28.10 – 49.96)	−10.1 (−21.3 – 62.12)

Values are mean ± SE.

SEC=secular intervention group; FI=faith-integrated intervention group; PRE=pre-intervention/baseline; POST=post-intervention, 10 months after baseline; BP=blood pressure; LPA=light-intensity physical activity; MVPA=moderate-to-vigorous physical activity.

^¥N=42 in FI group.

^{^}Pedometer-assessed (SEC: N=40; FI: N=34).

^#Accelerometer-assessed (SEC: N=40; FI: N=34).

^*Statistically significant time effect from pre- to post-intervention (P≤0.05).

There were no significant differences between groups at baseline for any variable.

There were no significant interactions between groups.

Table 3 presents the results of the intention-to-treat analyses for all participants who entered the sub-study. When comparing these results to Table 2 (sub-study completers), findings were similar. FI showed significant time effects for body weight (93.9 ± 2.3 to 92.8 ± 2.3 kg; P=0.016; partial η²=0.04), BMI (35.8 ± 1.0 to 35.4 ± 1.1 kg/m²; P=0.016; partial η²=0.04), waist circumference (106.4 ± 2.2 to 103.5 ± 2.6 cm; P=0.030; partial η²=0.12), and LPA for both 1-minute activity bouts (644 ± 16 to 604 ± 18 minutes/day; P=0.045; partial η²=0.03) and 10-minute bouts (538 ± 15 to 495 ± 18 minutes/day; P=0.019; partial η²=0.03). A small difference in the intention-to-treat analyses compared to completers was that the time effect for hip circumference did not reach statistical significance by a narrow margin (122.2 ± 1.1 to 120.9 ± 2.0 cm; P=0.056; partial η²=0.06). Adjusting for age and baseline PA did not significantly change any results compared to the unadjusted analyses.

Table 3:

Pre- and post-intervention descriptive characteristics and variable comparisons using intention-to-treat analyses for all participants of the L.A.D.I.E.S. for a Better Life sub-study (N=104).

	SEC (n=58)			FI (n=46)			Difference in Mean Change (95% CI)
VARIABLES	PRE	POST	Change in Mean (95% CI)	PRE	POST	Change in Mean (95% CI)
Descriptives & Cardiometabolic Risk Factors
Weight (kg)	96.0 ± 2.2	96.0 ± 2.3	−0.05 (−0.64 – 0.54)	93.9 ± 2.3	92.8 ± 2.3 *	−1.15 (−1.94 – 0.36)	1.10 (−2.27 – 0.07)
Body mass index (kg/m2)	36.2 ± 0.7	36.2 ± 0.8	−0.01 (−0.23 – 0.21)	35.8 ± 1.0	35.4 ± 1.1 *	−0.41 (−0.69 – 0.13)	0.40 (−0.85 – 0.04)
Waist circumference (cm)	104.4 ± 1.7	102.6 ± 1.2	−1.18 (−4.31 – 0.69)	106.4 ± 2.2	103.5 ± 2.6 *	−2.94 (−5.41 – 0.47)	1.13 (−3.99 – 1.51)
Hip circumference (cm)	122.5 ± 1.1	121.2 ± 1.0	−1.26 (−2.56 – 0.04)	122.2 ± 1.1	120.9 ± 2.0	−1.24 (−4.75 – 2.26)	−0.01 (−3.14 – 2.22)
Systolic BP (mmHg)	131 ± 2	132 ± 2	1.08 (−1.82 – 3.98)	133 ± 4	137 ± 1	3.74 (−6.10 – 13.58)	−2.66 (−3.71 – 9.02)
Diastolic BP (mmHg)	83 ± 2	83 ± 2	0.52 (−1.67 – 2.72)	82 ± 2	84 ± 1	1.40 (−1.88 – 4.68)	−0.88 (−2.27 – 4.02)
Hemoglobin A1c (%)	7.6 ± 0.2	7.7 ± 0.3	0.09 (−0.12 – 0.30)	7.2 ± 0.2	7.3 ± 0.3	0.15 (−0.20 – 0.51)	−0.06 (−0.20 – 0.40)
Physical Activity
Average steps/day^	3,789 ± 122	4,380 ± 333	581 (−55 – 1,317)	3,806 ± 156	4,722 ± 431	876 (−153 – 1,905)	−295 (−596 – 1,191)
Sedentary time (min/day)#
1-minute bouts	781 ± 33	793 ± 30	10.64 (−36.0 – 57.28)	776 ± 17	809 ± 16	31.70 (−21.8 – 85.20)	−21.1 (−35.9 – 77.98)
10-minute bouts	694 ± 36	710 ± 33	15.19 (−34.1 – 64.47)	691 ± 18	728 ± 19	35.93 (−24.8 – 96.67)	−20.7 (−40.7 – 82.19)
LPA (min/day)#
1-minute bouts	634 ± 30	623 ± 30	−10.1 (−58.2 – 37.92)	644 ± 16	604 ± 18 *	−39.2 (−76.8 – 1.50)	29.01 (−82.6 – 24.59)
10-minute bouts	521 ± 28	517 ± 27	−3.93 (−50.0 – 42.18)	538 ± 15	495 ± 18 *	−41.7 (−72.2 – 11.1)	37.72 (−89.9 – 14.42)
MVPA (min/day)#
1-minute bouts	25 ± 6	24 ± 8	−0.50 (−7.79 – 6.79)	20 ± 4	28 ± 13	7.48 (−25.4 – 40.39)	−7.98 (−14.2 – 35.68)
10-minute bouts	10 ± 3	11 ± 5	0.53 (−4.71 – 5.78)	9 ± 3	17 ± 10	8.07 (−20.0 – 36.17)	−7.53 (−11.3 – 33.50)

Values are mean ± SE.

SEC=secular intervention group; FI=faith-integrated intervention group; PRE=pre-intervention/baseline; POST=post-intervention, 10 months after baseline; BP=blood pressure; LPA=light-intensity physical activity; MVPA=moderate-to-vigorous physical activity.

^{^}Pedometer-assessed.

^#Accelerometer-assessed.

^*Statistically significant time effect from pre- to post-intervention (P≤0.05).

There were no significant differences between groups at baseline for any variable.

There were no significant interactions between groups.

When comparing effect sizes between the study completers (Table 2) and all participants included in the intention-to-treat analyses (Table 3), partial η² values were similar but reflected small effects for weight and BMI (0.03–0.04) and a moderately large to large effect for waist circumference (0.12–0.14). Partial η² values were also similar between the two analyses groups for LPA (0.03–0.04) and hip circumference (0.06–0.08), where both variables showed small and moderate effects, respectively.

DISCUSSION

This sub-study of the parent L.A.D.I.E.S. for a Better Life study was beneficial in that it provided additional data related to the effects of the two active intervention curricula on multiple hard endpoints associated with cardiometabolic disease risk. Given the health disparities that persist in AA women with regard to obesity, hypertension, and type II diabetes, this sub-study provides a critical supplement to the parent study, as it speaks more specifically to the health status of this sample. A notable and perhaps disturbing aspect of this study is the threatening cluster of cardiometabolic risk factors present in the women examined. The combination of class II obesity and waist circumference >88 cm alone places participants in the current study at “very high” risk of hypertension, type II diabetes, and cardiovascular disease (28). This point is substantiated by the fact that, on average, study participants exhibited measured values indicative of both hypertension (35) and diabetes (3). In addition to the presence of these issues themselves, the magnitude of the elevation for several of these markers compared to normal values was also remarkable. Specifically, values in the overall sample for BMI (study sample: 36.0 kg/m² versus normal: <25 kg/m²), waist circumference (study sample: 105.3 cm versus recommended: ≤ 88 cm), blood pressure (study sample: 132/83 mmHg versus normal: <120/80 mmHg), and hemoglobin A1c (study sample: 7.4% versus normal: <5.7%) were all measured at alarmingly high levels. This speaks to the gravity of these issues in the current study participants. The descriptive data collected in the sub-study appear to be in line with national estimates. For example, NHANES data from 2007–2012 showed that 68% of AA women had waist circumferences ≥88 cm, 50% had elevated blood pressure (defined at that time as ≥130/85 mmHg), and 25% had elevated fasting blood glucose (≥100 mg/dL or on medication for the treatment of elevated blood glucose) (22). The current sample’s average hemoglobin A1c of 7.4% translates to an estimated average fasting blood glucose of 169 mg/dL (23), suggesting that the current sample may be abnormally high compared to other AA women in the U.S. National data and data from the current sub-study highlight critical issues with chronic disease risk factors and chronic disease prevalence that need to be addressed in high risk populations like AA women.

The authors are encouraged by the significant improvements in body weight (and consequently, BMI) and waist circumference in the FI group during the 10-month intervention period. It has been well-established that body weight, BMI, and abdominal adiposity are strong predictors of cardiometabolic health and that improvements in these variables have been associated with reduced cardiometabolic risk (9, 18). Visceral adipose tissue, which is stored in the abdomen, has been shown to release pro-inflammatory adipokines (a precursor for vascular damage, creating vulnerability to the development of atherosclerotic plaque) and to be less sensitive to insulin’s antilipolytic effects. Further, acute and chronic increases in plasma free fatty acids can promote hyperinsulinemia and pancreatic β-cell apoptosis, respectively (9). Collectively, these events contribute to hypertension, cardiovascular disease, and type II diabetes Although neither intervention improved blood pressure or hemoglobin A1c in this sample, continued reductions in body weight and waist circumference could have important, positive long-term implications on these and other risk factors. Further, we note that the average baseline blood pressure in the sub-study sample was <140/90 mmHg (i.e. not indicative of hypertension) at the time the study was conducted. Thus, women in the sub-study were not specifically warned about abnormal blood pressure values unless values were ≥140/90 mmHg and may not have recognized their pre-hypertensive state or the increased risk at which their blood pressure values placed them for developing hypertension, cardiovascular disease, and stroke. Updated definitions for elevated/pre-hypertension and hypertension were not adopted by the American Heart Association until after the L.A.D.I.E.S. study concluded (35). Although sub-study participants were provided with their own data regarding weight, waist and hip circumferences, blood pressure, and hemoglobin A1c immediately after they were measured at baseline, future interventions that share baseline results with participants should perhaps emphasize any abnormal and pre-disease values, as a motivator for behavior change.

The fact that the improvements in body weight, BMI, and waist circumference occurred in the FI group may indicate that this curriculum could be more effective on health measures over time. This theory is supported given the long-term success of the L.A.D.I.E.S. parent study in maintaining PA over 22 months in the FI group compared to SEC (38). In addition, the parent study showed that the FI group completed significantly more sessions compared to the SEC group (P<0.01), which may also speak to the long-term promise of using a FI curriculum to gradually improve health and PA, and sustain these improvements (38). As all of the L.A.D.I.E.S. participants were regular church attendees, it is possible that they perceived the FI curriculum as more personally relevant, of greater holistic value (i.e. both physically and spiritually beneficial), and/or simply more enjoyable compared to participants in the SEC groups. These factors could have contributed to greater sub-study adherence and in turn, more improvements in the variables examined in this study. Our results also reiterate the importance of considering cultural context when translating PA promotion interventions.

In contrast to the parent study which showed statistically significant increases in steps/day in the SEC and FI groups (+1,107 and +1,451 steps/day, respectively), neither intervention group showed a significant increase in PA for this sub-study. Although not statistically significant, the sub-study SEC and FI groups did increase average daily steps by 864 and 1,089, respectively. It is unclear why the sub-study results did not mirror the parent study results, but the small sample size was a likely contributor.

A significant decrease in weight, BMI, and waist circumference was shown despite only modest improvements in steps/day and a significant decrease in LPA in the FI group. This could speak to the value in maintaining small PA improvements over time. The recently released Physical Activity Guidelines for Americans, 2^nd edition indicated the important benefits of even small improvements in PA, particularly among sedentary individuals (34). That said, the authors are unable to comment on the caloric consumption of sub-study participants (as diet data were not collected) and therefore, make this speculation cautiously given the context. More research inclusive of both PA and diet evidence is needed to disentangle the influence of each when modest PA improvements are achieved and maintained over longer periods of time. While previous research has shown improvements in blood pressure and glucose tolerance with increases in steps/day (8, 31), the average step increase (2,004–4,241 steps/day) was higher than what was observed in the present study. It may also be notable that the current participants entered the study with baseline step counts indicative of sedentary lifestyles (<5,000 steps/day), and did not achieve a step/day increase that advanced them to the next highest activity category (4,591 steps/day and 4,973 steps/day post-intervention in SEC and FI, respectively), based upon step indices (low-active: 5,000–7,499 steps/day) (33). A review by Tudor-Locke et al. identified evidence of poor health profiles in individuals achieving <5,000 steps/day including higher BMIs, less favorable body composition, and a higher prevalence of metabolic syndrome (defined as the presence of three or more cardiometabolic risk factors including high waist circumference, resting blood pressure, fasting glucose, and fasting triglycerides, as well as low high-density lipoprotein cholesterol levels) (33). This may suggest that for health benefits, achieving 5,000 steps/day should be the minimum goal.

Neither the SEC nor the FI group showed significant improvements in overall accelerometer-measured PA, which was consistent with the parent study. The significant LPA decrease in the FI group would have been easier to accept if a concurrent significant increase in MVPA was observed. However, FI had non-significant increases in MVPA for both 1-minute (+10 minutes/day) and 10-minute bout durations (+11 minutes/day). Consistent with the parent study, both the SEC and FI groups in the sub-study sample also showed non-significant increases in sedentary behavior. Most notable of the two groups were the increases shown in the FI group (1-minute [+43 minutes/day] and 10-minute bout durations [+49 minutes/day)]). The notion of perceived prioritized importance of rest and sleep in the AA community, particularly as a necessary precursor to engaging in PA/exercise, has been previously noted (38). Further, participants may have chosen to partake in more sedentary activity after feeling accomplished or deserving of a “reward” for completing more PA (i.e. steps/day), if not simply tired after activity. It is possible that the increases in both MVPA and sedentary behavior negated each other such that neither increase was statistically significant. Quantifiable data on lifestyle habits (diet, rest/sleep), along with supplemental qualitative information on attitudes toward these factors while attempting to adopt more PA may help provide a more comprehensive interpretation of these type of data in future studies.

As higher levels of MVPA have been associated with more favorable cardiometabolic risk profiles (19, 30), even the small improvements in MVPA shown in the current sub-study may have contributed to the observed significant improvements in weight, BMI, and waist circumference. Despite the unexpected fluctuations in activity intensities, the authors are encouraged by the modest MVPA and PA improvements that occurred over the 10-month period. Continued small increases in MVPA could have clinical significance for further body weight/body composition improvements and may have a positive impact in other cardiometabolic risk factors in this at-risk population over time. Also relevant to this line of discussion is the possibility that the modest increase in MVPA in the FI group may have blunted our ability to see changes in blood pressure and hemoglobin A1c, despite the improvements in weight, BMI, and waist circumference. The fact that the sub-study dropouts were significantly younger and acquired more MVPA is also notable, as this could have had an influence on the small step/day and MVPA increases that were observed. Future studies designed for AA women that aim to improve cardiometabolic risk factors may benefit from a focus on more aggressive PA prescriptions with higher average daily step goals and/or placing emphasis on MVPA.

Increasing MVPA could be achieved by suggesting more specific exercise and PA prescriptions to address intensity, duration, and/or modality (i.e. incorporate intervals of jogging into your daily walks, exercise for progressively longer durations, etc.). Previous studies in obese AA women have had success increasing the dose of MVPA by introducing intervals of moderate- and vigorous-intensities into PA regimens (4, 17). Although introducing intervals of higher intensities may be perceived as undesirable, intimidating, or even inappropriate for the current study population (i.e., obese, sedentary, at increased cardiometabolic risk), this training method has been successfully implemented (often in conjunction with a baseline training period prior to incorporating intervals) in low active, overweight/obese, and subclinical populations without adverse events (6, 17, 19). The incorporation of moderate- and vigorous-intensity intervals could translate into community interventions using strategies such as coaching participants to perform brief intervals of increased intensity during group exercise activities, or to walk faster for a given distance while walking for exercise outdoors. Planning an appropriate progression strategy is key for safe and successful transitions.

There are study limitations worth noting. First, this sub-study was designed to supplement the parent study by providing preliminary data on the cardiometabolic effects of the active interventions. As such, the sub-study did not compare them to the control condition. The rationale was that, given the limited time and resources compared to the parent study, focus should be placed on understanding the impact of the two active interventions on cardiometabolic risk factors. The focus on volunteers from the first 12 active intervention studies enrolled in the study lent to a small sample size, which could have had a negative impact our ability to detect significant differences. Also, the parent study was powered to detect differences between either active intervention and the control group and not between the two active interventions, which likely impacted our ability to detect significance between group in the sub-study sample. It is also recognized that differences may have been more difficult to detect since both groups received a 10-month PA intervention. The purpose of the sub-study was to collect pilot data in preparation for a larger, fully powered study focused on reducing cardiometabolic risk factors in AA women. While the factors mentioned may have presented a sampling bias within the L.A.D.I.E.S. study, the authors mentioned previously that the overall sample characteristics for the parent study and the sub-study were similar. This also speaks to the diminished likelihood that participants who elected to participate in the sub-study may have had different health profiles than those who declined. That said, the authors recognize that the current sample is not representative of all AA women, and thus the current results cannot be generalized to other AA women or populations other than AA women.

As no data were collected on existing chronic disease diagnoses or medication use/dosage, one could also speculate that participants could have reduced medication use, yet shown no changes in blood pressure and/or hemoglobin A1c. However, investigators obtained no anecdotal evidence of this notion from participants or intervention group leaders, who were knowledgeable of the participants’ health goals and accomplishments. Further, in an effort to mitigate this type of issue and obtain a somewhat “raw” blood pressure measurement, participants were asked to refrain from blood pressure medication on the day their data were collected. This was 12–24 hours (depending on when they typically took their medication) prior to the pre- and post-intervention data collection appointments.

The authors accept that when examining human subjects in a free-living setting over a 10-month period, there are inherent variations in factors such as diet, sleep, acute illnesses, and other lifestyle activities that are difficult to control. Diet data were not collected from the current participants, which could have potentially helped to explain the lack of improvement in blood pressure and hemoglobin A1c that occurred in both groups (through examination of chronic sodium and carbohydrate intake) (13, 21). Further, the modest and non-significant improvements in steps/day and MVPA simply may not have been potent enough to elicit changes in these variables, even if diet remained consistent.

Despite the noted limitations, the authors feel that this study is pertinent as it adds to the sparse literature that examines the effects of culturally relevant, faith-based PA interventions on cardiometabolic risk in AA women. In addition, the intervention was conducted over a period of 10 months and employed methods to objectively monitor PA. Although research in this vulnerable population has increased in recent years (7, 40), AA women are still a population in great need of sustainable interventions focused on decreasing indicators of chronic disease, as they remain at a greater risk compared to women of other racial and ethnic groups. The current study suggests that faith-based interventions may be a strategy to assist in this effort and to achieve long-term success in improving variables related to cardiometabolic disease. This study falls under early translation on the translational science continuum, as it will assist with the refining the methods for development of a larger trial. Further, this study provides preliminary clinical evidence of our intervention’s effects and can be used to estimate effect sizes of the primary outcomes (cardiometabolic risk factors). If proven efficacious, our community-based intervention design increases the potential for intervention dissemination and implementation more broadly.

Although, further research is needed to examine the effects of other FI and culturally relevant interventions on obesity and its related comorbidities in AA women, the authors gained valuable information from this sub-study. Key lessons that surfaced from this sub-study include: 1) cultural relevance in the intervention curriculum (i.e. the FI in the current study) resulted in higher study adherence, which could have contributed to the positive health changes observed compared to SEC; and 2) emphasis on MVPA should be targeted if improvements in cardiometabolic risk factors are the programmatic focus. These lessons could have important implications for future researchers and community programs. These authors feel that future studies should aim to identify intervention strategies that are culturally and/or personally relevant to the study population in order to be effective in sustaining PA improvements and have a positive long-term impact on health and physiological variables.