Filipinos Fit and Trim - A feasible and efficacious DPP-based intervention trial.

INTRODUCTION: Filipino Americans are at higher risk for obesity and related Type 2 diabetes (T2D) compared to other Asian subgroups and non-Hispanic whites. Yet, there are limited research studies to reduce health disparities and improve health outcomes for Filipinos. Weight loss lifestyle intervention trials such as the Diabetes Prevention Program (DPP) can reduce obesity and T2D risks through physical activity and healthy eating.
METHODS: Thus, we conducted a pilot Fit&Trim (DPP-based) intervention study - a randomized controlled trial 3-month intervention augmented with mobile technology + 3-month maintenance follow-up with a waitlist control. The objective assessed the intervention feasibility and potential efficacy to reduce T2D risks in Filipino Americans with overweight/obesity. The overall study goal was a mean 5% weight reduction.
RESULTS: Sixty-seven eligible Filipino men and women were enrolled and randomized to either an intervention or waitlist control group. Participant retention was 91%. In Phase 1 (baseline to 3-months), the intervention group had greater weight reduction compared to the waitlist control (-4.3% vs. -0.88%; cross-level interaction = -0.85 (-1.4, -0.35). In Phase 2 (3- to 6-months), after receiving the Fit&Trim intervention, the waitlist group also had similar significant weight reduction [-4.8% (- 0.75 (-0.92, -0.58)]. A majority of intervention group (57%) also maintained their weight loss. Overall, 41% of study participants achieved a 5% weight loss.
CONCLUSION: The Fit&Trim intervention demonstrated feasibility and potential efficacy for Filipino Americans. Findings warrant a further larger, longer trial to test the Fit&Trim feasibility and effectiveness in a real-world Filipino community setting. CLINICALTRIALSGOV REGISTRATION NUMBER: NCT02278939.

Introduction

Over one third of US adults (84–92 million) [1] are affected by obesity related chronic diseases such as cardiovascular disease (CVD) and type 2 diabetes (T2D), resulting in an estimated $800 billion annual healthcare costs [2,3]. Asian Americans (Asians) have a higher risk for developing obesity and T2D at lower body mass index (BMI) and younger ages compared non-Hispanic Whites [4,5]. Among Asians, Filipino Americans (Filipinos) have higher body fat and central adiposity - a predictor for T2D, compared to non-Hispanic Whites [6,7]. Moreover, studies found Filipinos had one of the highest risks for overweight/obesity, hypertension, and T2D compared to other Asian subgroups [[8], [9], [10], [11]]. Despite their health disparities, Filipinos remain understudied and underrepresented in preventive health research [12,13]. Rectifying this major knowledge gap in preventive health research is essential for reducing Filipino health disparities and improving health equity in this vulnerable population.

Most lifestyle related chronic diseases are preventable with only a modest weight reduction of 5% [14]. The 6-month Diabetes Prevention Program (DPP) clinical trial that promoted weight loss through increased physical activity (PA) and health eating showed weight loss of 5–7%, and an effective reduction in the incidence of obesity related chronic diseases (e.g., CVD and T2D) [15,16]. A recent report identified effective strategies promoting weight loss in DPP-based lifestyle interventions including: 1) culturally tailoring interventions for diverse populations, 2) providing social support for healthy behaviors, and 3) incorporating mobile technology to enhance engagement and adoption [17].

To address the paucity in preventive health research focused on Filipinos, we conducted the Filipinos Fit and Trim (Fit&Trim) – a pilot randomized controlled trial (RCT) that assessed the impact of a DPP-based culturally adapted mobile phone-based weight loss lifestyle intervention including virtual social support (Facebook) to reduce obesity and T2D risks in Filipino Americans with overweight/obesity. The purpose of this paper is to: 1) provide a brief overview of the Fit&Trim intervention study, 2) describe baseline socio-demographic data, and 3) present the primary and initial secondary outcomes.

Methods

Design

This was a pilot RCT of a 3-month culturally adapted weight loss lifestyle intervention (Fit&Trim) followed by a 3-month maintenance period to reduce T2D risks in Filipino Americans who are overweight, with an active waitlist control. The Fit&Trim, a DPP-based intervention, was augmented with mobile technology using a Fitbit Zip, mobile app/diary, and private Facebook group. The study objective was to assess the Fit&Trim intervention for feasibility and potential efficacy for weight loss. This study was conducted in the San Francisco Bay Area from July 2014 to Feb 2017. Institutional review board approval was obtained prior to implementing the study. Participant informed consent was obtained prior to study participation.

This RCT incorporates a bi-phasic design: Phase 1(baseline to 3-months) and Phase 2 (3-month to 6-month) with a waitlist control. It was NOT a cross-over design. Based on this study design, potential efficacy of the Fit&Trim intervention can only be assessed at Phase 1 (outcomes of the intervention group compared with a control). In the intervention received the Fit&Trim intervention. The active wait-list control group received only a Fitbit accelerometer to use on their own without education related to the Fit&Trim intervention or physical activity. In the intervention group transitioned to a 3-month maintenance with in-person weight checks and personal coaching, and were asked to continue to track their health behaviors on their own. Meanwhile the waitlist control group received the Fit&Trim intervention.

Participants were randomized per a computer-generated randomized sequence by a statistician in a 1:1 ratio. Randomization was stratified by fasting plasma glucose: normal <100 mg/dL or prediabetes ≥100 mg/dL < 125 mg/dL, in permuted randomly selected block sizes of 2 and 4 to an intervention group or active waitlist control group. Due to the nature of a lifestyle intervention study, participants, investigators and research staff were not blinded to randomization allocations. Biostatisticians and laboratory staff who conducted blood draws were blinded to the randomization allocations.

Theoretical framework

The social cognitive learning theory was used to guide this study [18]. This theory posits that behaviors are often learned via significant role models (e.g., parents, mentors, or heroes); and learned behaviors are influenced by one's culture and social environment. Moreover, self-efficacy (confidence of achieving behaviors) and positive (negative) social support from one's environment encourages (discourages) engagement and reproduction of learned behaviors.

Sample and setting

Filipino American adults who were overweight or obese were recruited from the SF Bay Area from November 2014 through July 2016. A convenience sample of 67 eligible participants was enrolled in this study.

Inclusion/exclusion criteria - based on the 2002 DPP trial design [19]

Inclusion criteria: a) men and women, b) Age ≥ 18 years; c) self-identified as Filipino; d) BMI for Asians > 23 kg/m2, e) Diabetes Risk score > 5 points [20], fasting plasma glucose test between 100 and 125 mg/dL, A1c > 5.6%, or oral glucose tolerance test (OGTT) between 140 and 200 mg/dL, f) able to walk at least 20 min without assistance, and g) able to speak and read English.

Exclusion criteria: a) physician diagnosed diabetes and confirmed by clinical data (e.g. documentation of fasting blood glucose >126 mg/dL or 2-h OGTT>140 mg/dL, A1C > 7.0%; b) Glucose metabolism disorder associated diseases precluding weight management (e.g., Cushing's syndrome, and polycystic ovary disease); and c) known medical conditions requiring a special exercise program (e.g., prior myocardial infarction and chronic obstructive pulmonary disease). Only one member per household was allowed to participate in the study.

Recruitment

Participant recruitment strategies included: 1) posting flyers: at local ethnic grocery stores and academic/clinic settings, and on the study's Facebook website and Craigslist.com, a website for classified ads; 2) select mailings to San Francisco and San Mateo County communities, 3) in-person presentations at Filipino community meetings and cultural events, and 4) snowball sampling and word-of-mouth referrals. Previous study participants who expressed interest in future studies were also contacted by email or phone.

Pre-screening and run-in

Potential participants completed a telephone screening. Those who met the eligibility criteria and provided informed consent were invited to a screening baseline visit (SBV). Potential participants received a brief physical exam to collect anthropometric measures (weight, height, waist circumference, and blood pressure) and completed a set of baseline surveys. Based on the SBV, those who met the inclusions/exclusion criteria were invited to participate in a 2-week run-in, asked to wear a Fitbit Zip at least 10 h/day, and email photos to research staff of all food and drink intake for three consecutive days. Prior to the run-in period they received a Fitbit Zip with mobile app and were trained on the use of the app on their smartphone, tablet, and/or laptop.

Potential participants were scheduled for laboratory tests at the institutional research lab and for a subsequent in-person enrollment visit in 2-weeks. At this enrollment visit, those who completed at least 70% of the 2-week run-in requirements and met the laboratory test eligibility criteria (see Methods section for inclusion and exclusion criteria) were invited to enroll in the study. Fig. 1 presents the results of the participant recruitment, enrollment, randomization, and completion.

Fig. 1

Consort diagram.

Intervention

Six-month Program

During the 6-month Fit&Trim program (3-month Fit&Trim intervention and 3-month maintenance period), participants: 1) attended 5 in-person intervention office visits, 2) tracked real-time steps by wearing a Fitbit Zip on their torso at least 10 h/day, 3) logged daily food/drink intake and weekly home weights on a mobile app/diary, and 4) received weekly postings of discussion topics (related to weight loss, PA and healthy eating) on the study's private Facebook group site to reinforce healthy behaviors learned during in-person intervention sessions. Table 1 presents baseline to 3-month intervention sessions and subject matter provided via Facebook virtual social support and targeted health behavior coaching. Participants were scheduled for a total of 6 in-person office visits, including the baseline enrollment visit.

Table 1

Fit&Trim schedule (baseline to 3 months) by group.

Schedule	In-Person Intervention Sessions	Active Wait-List Control Sessions
Baseline Visit	• Lifestyle balance and social networking • Initiating physical activity and healthy diet • Set individual short- and long-term goals • Training tracking weight, physical activity and food/drinks using mobile technology	• Hepatitis A education and handout • Track physical activity via Fitbit Zip
1 Month Visit	• Benefits and ways to be physically active • Filipino dancing – Zumba, cha cha • Benefits and ways to eati healthy • Limiting fat and healthy Filipino food options	• Hepatitis B Education and handout • Track physical activity via Fitbit Zip
2 Month Visit	• Social support for physical activity and diet • Problem solving: family events & eating out • Benefits of tracking health behaviors	Track physical activity via Fitbit Zip
3 Month Visit	• Slippery slope and relapse prevention • Staying motivated • Social support for maintaining behaviors Transition to 3-month maintenance phase	Start 3-month Fit&Trim Intervention• Individual short- + long-term goals • Train on tracking weight, physical activity food/drinks with mobile app
Facebook Group Virtual Social Network
3 months Fit&Trim	• Monitored & moderated by research staff • Peer sharing messages and photos • Weekly posts: Healthy eating and physical activity education and discussion topics	Facebook group only @ 3–6 months

Intervention participant training

Prior to receiving the intervention, participants were trained on how to: 1) use the Fitbit Zip; download the mobile app/diary; and 2) access and join the private Facebook group. Participants' Fitbit data was monitored by research staff and uploaded in real-time to the study's online account and secure data servers. Participants experiencing technological problems received remote assistance via phone.

Intervention group

In (baseline to 3-months), intervention group set individual tailored goals for weight, diet, and PA step. Participants returned for three monthly office visits and completed the intervention at the 3-month visit. In (3- to 6-months), the intervention group transitioned to a 3-month maintenance period with in-person weight checks and personal coaching, completing the study at the 6-month visit.

Active waitlist control

In (baseline to 3-months), the waitlist control group were only asked to wear the Fitbit Zip on their own with no education or coaching. They received education on Hepatitis A and B at the office visit. The waitlist group completed the control period at the 3-month visit. In (3- to 6-months) the waitlist group received the Fit&Trim intervention, completing the study at the 6-month visit. Table 1 presents the baseline to 3-month schedule for both study arms.

Outcome measures

Primary outcome

Feasibility was measured by participant recruitment, engagement and retention. The recruitment goal was to enroll and randomize 50 participants. The engagement goal required participants to attend at least 4 of the 6 intervention office visits, thus ensuring participants received at least 9 out of 16 DPP-based education sessions to meet the intervention completion guidelines established by the Centers for Disease Control and Prevention (CDC) [21]. Retention was defined by study completion as measured by the number of enrolled participants who attended their final 6-month visit. We hypothesized that at least 80% of enrolled participants would complete the study.

Secondary outcome

The Fit&Trim overall weight loss goal was 5%. This is the weight loss threshold for clinical significance in reducing T2D and cardiometabolic disease risks [14]. The overall mean weight change was measured in percent (%) and kilograms (kg) over time in (baseline to 3-months) and (3- to 6-months) for both study arms. Due to the RCT bi-phasic design (as opposed to a cross-over design), potential efficacy of the Fit&Trim intervention could only be assessed during Phase 1 (baseline to 3-months). Other secondary outcomes included change in waist circumference, BMI, fasting blood glucose, and A1c.

Data collection

Socio-demographics and survey data were collected at the baseline visit. Anthropometric (BMI, waist/hip circumference, and blood pressures), laboratory tests (fasting blood glucose, and lipid panel, insulin, A1c), and survey data were collected at baseline, 3- and 6- month visits. All participants’ weights were collected at each office visit throughout the 6-month study.

Data analysis

Descriptive statistics (mean, median, and standard deviations for continuous variables; frequencies and percents for categorical variables) were used to describe the socio-demographic characteristics. Between-group differences in percent weight change categories over time were analyzed using a bootstrap chi-square test, including the Mantel-Haenszel test of trend. Statistical analysis was performed out using SPSS 23 (IBM, SPSS Statistics 23).

The primary feasibility recruitment outcome was based upon achieving the target sample size. The simple proportion (%) of participants within each randomized group was reported for multiple target behavior threshold criteria used to measure engagement and retention during (baseline to 3-months) and (3-months to 6-months) endpoints.

Potential efficacy was analyzed only in Phase 1 to assess whether change was greater for the intervention group compared to the control. Other outcomes were analyzed to assess whether the change was greater for the group receiving the Fit&Trim intervention (during or ) compared to the group not receiving the Fit&Trim. A multilevel regression (i.e., linear mixed models or hierarchical linear models) was used to test for differences between the intervention vs. waitlist control groups at baseline and for change in trajectories over time. The intention-to-treat principle was used with analysis for those variables with missing data for each study arm. Differences in the change trajectories are called the “cross-level interaction” or “effect” between time and group [22,23]. The simple slopes were also assessed to determine if there was a significant change within each group for and . This approach accounted for missing data.

Multilevel regression analysis allowed for the use of bootstrapping when the assumption of a normal distribution of the data was not met. Bootstrapped full information maximum likelihood models were estimated to obtain nonparametric, bias-corrected bootstrapped Confidence Intervals (BC CI) for estimation and inference about the study hypotheses [[24], [25], [26]]. Intention to treat analyses were used for the primary and secondary outcomes. These analyses were performed using Stata/SE version 14 [27,28]. Significance was set at a 2-sided alpha of 0.05.

Results

Baseline sociodemographic and clinical measure characteristics

Sixty-seven eligible Filipino American adults were enrolled and randomized in the study (Fig. 1). Mean age was 41.7 years, 52.2% were women, 76.1% completed college, 47.8% were married/cohabitating, and 91% were full/part-time employed. Although 54% were immigrants and lived ≤ 5 years in the United States, overall, the participants were highly acculturated (Marin Acculturation Scale, mean score = 4 out of 5 points) to the Western culture. However, a majority (96%) claimed to have inadequate health literacy, defined as “… the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions” [29], A majority (61%) had never used a pedometer.

There were no significant differences between genders for: weight (p = 0.42), BMI (p = 0.96) or waist circumference (p = 0.59). Mean fasting blood sugar was 91 mg/dL, and the mean A1c was 5.8% with no significant difference between sex for either variable (p = 0.31; p = 0.42, respectively). Overall, there were no between-group differences for sociodemographic characteristics or laboratory tests. See Table 2 for baseline socio-demographic and clinical characteristics.

Table 2

Baseline and sociodemographic characteristics.

Variables	All Participants (N = 67)	Control (n = 34)	Intervention (n = 33)	p-value
	Mean (±SD) % (n)	Mean (±SD) % (n)	Mean (±SD) % (n)
Age years	41.7(±12.0)	42.1 (±12.2)	41.3 (±12.1)	.79
Gender
Male	47.8 (32)	44.1 (15)	51.5 (17)	.63
Female	52.2 (35)	55.9 (19)	48.5 (16)
Education
Completed high school + some college
	23.9 (16)	32.4 (11)	15.2 (5)	.15
Completed college	76.1 (51)	67.7 (23)	84.9 (28)
Marital status
Never married, divorced, or widowed	52.2 (35)	50 (17)	54.6 (18)	.81
Married/cohabitating	47.8 (32)	50 (17)	45.5 (15)
Employment (full/part time	91 (61)	88.2 (30)	93.9 (31)	.67
Years lived in US
Native born	46.3 (31)	52.9 (18)	39.4 (13)	.33
≤5 years	53.7 (36)	47.1 (16)	60.6 (20)
Previous pedometer use
Yes	38.8 (26)	38.2 (13)	39.4 (13)	1
No	61.2 (41)	61.8 (21)	60.6 (20)
Health Literacy	5 (±1.9)	5.1 (±2.3)	4.9 (±1.4)	.61
Adequate	4.5 (3)	9.1 (3)	0	.11
Not adequate	95.5 (64)	91.9 (30)	100 (34)
Acculturation score	4 (±0.5)	4 (±0.5)	4 (±0.5)	1
Baseline Outcome Measures
Weight (kg)	81.5(±15.2)	80.0(±14.9)	83.1(±15.6)	.42
BMI (kg/m2)	30.5(±4.4)	30.5(±4.9)	30.5(±3.9)	.96
Waist circumference (cm)**	99.0(±10.2)	98.3(±10.5)	99.7(±10.0)	.59
BP systolic mm Hg	128.4 (±12.7)	130.1(±14.7)	126.7(±10.2)	.27
Men≥130 mm Hg
Women≥130 mm Hg	132.1(±10.6)
125.0(±13.6)			.022
BP diastolic mm Hg	80.4(±10.3)	81.4(±9.3)	79.3(±11.3)	.41
Fasting Glucose > 100 mg/dL	90.9(±10.0)	92.1(±9.3)	89.5(±10.7)	.31
A1c	5.8(±0.31)	5.9(±.33)	5.8(±0.29)	.42

Primary outcome

As outlined in the methods section, the primary outcome, feasibility, was measured by meeting goals for recruitment, engagement, completion of intervention, and retention. For study recruitment, 67 eligible participants were enrolled and randomized. Out of enrolled participants 60 (90%) completed the Fit&Trim intervention per the CDC DPP completion guidelines [21]. For retention, 61 (91%) completed the final 6-month study visit (see Fig. 1).

Secondary outcomes

Potential efficacy - weight loss

Potential efficacy was measured for both the intervention and waitlist control groups by mean weight (% and kg) change over time. Table 3 presents the estimated means and simple slopes for difference within-group and between-group for and . Bold type indicates all statistically significant within-group estimated simple slopes and between-group cross-level interactions. The 3-month estimated means were slightly different because seven monthly data points were used to estimate independent linear simple slopes for both and groups. Thus, in Table 3, we reported the 3-month estimated means separately for both phases. The cross-level interactions describe the between-group differences in weight loss with the “estimated simple slopes.”

Table 3

Multilevel Regression Outcomes: (Baseline to 3 months) and (3- to 6- months).

N = 67		Intervention n = 33 (Received Fit&Trim @ P1)		Waitlist Control n = 34 (Received Fit&Trim @ P2)		Cross-level Interactions (95% CI)	Cohen d
		Estimated		Estimated
Outcome measures		Mean	Simple Slope (95%CI)c	Mean	Simple Slope (95% CI)c
Weight (%)	P1 - BL	−0.32		−0.31
	P1-3Mod	−4.62	−1.4 (-1.9, -.92)a	−1.19	- 0.29 (−.63, .05) a	−1.1 (-1.7, -0.53)
	P2-3Mod	−4.40		−1.31
	P2—6Mo	−5.55	−0.38 (-.71, -.06)b	−6.08	- .75 (-.92, -.58)b	−1.6 (-2.10, -1.09)
Weight (kg)	P1 - BL	82.78		79.81
	P1-3Mod	79.39	−1.1 (-1.56, -.70)a	78.99	-.28 (-.54, -0.01)a	−0.85 (-1.4, -0.35)	0.93
	P2-3Mod	78.85		78.90
	P2—6Mo	77.87	−0.33 (-.62, -.03)b	75.24	−1.2 (-1.6, -.81)b	−0.90 (-1.4, -0.39)
Body Mass Index (kg/m2)	BL	30.53		30.46
	P1b	29.27	−1.2 (-1.76, -.81)	30.15	- 0.32 (- .66, -.02)	−0.93 (-1.5, -0.40)
	P2b	28.91	- .37 (-.69, -.07)	28.75	−1.4 (-1.9, - .93)	−1.0 (-1.6, -0.47)
Waist Circumference (cm)	BL	99.65		98.29
	P1b	93.53	−6.1 (-8.4, -4.2)	97.06	−1.2 (−2.5, .01)	−4.9 (-7.5, -2.6)
	P2	93.30	- .22 (−1.4, 1.1)	92.52	−4.5 (-6.8, -2.3)	−4.3 (-6.8, -1.8)
Fasting Plasma Glucose (mg/dl)	BL	89.58		92.08
	P1b	88.68	-.90 (−4.5, 3.2)	92.54	.45 (−2.5, 3.1)	−1.4 (−5.9, 3.6)
	P2b	91.39	2.7 (−1.0, 7.7)	91.34	−1.2 (−3.3, .88)	3.9 (−0.3, 9.3)
A1c (%)	BL	5.815		5.876
	P1b	5.738	- .08 (- .17, .003)	5.900	.02 (- .04, .08)	-.10 (−.21, .002)
	P2b	5.739	.001 (−.07, .07)	5.811	-.09 (- .16, .02)	-.09 (−.19, .008)

Bolded numbers = Statistical significance indicated by NO “0” in 95% CI; BL=Baseline; P1 = Phase 1; P2 = Phase 2  = Estimated Simple Slope BL to 3Mo;  = Estimated Simple Slope 3Mo to 6Mo;  = Nonparametric bias corrected bootstrapped 95% CI Cross-level interaction = between group difference in simple slopes;  = slightly different means were used for P1- and P2—3Mo Simple Slope weight (kg and %) regression analysis to accommodate all seven monthly data points from baseline to 6 Mo.

In , all between-group differences indicated by the bolded cross-level interactions were statistically significantly different. For example, the estimated simple slope for the intervention group's weight loss was significant (−1.1, 95% CI: −1.6, −0.70). In the cross-level test for interaction, the estimated weight (kg) shows a greater decrease of −0.85 kg/month in the intervention group compared to the waitlist control group, which was statistically significant. The confidence interval for the population difference can be as great as −1.4 kg/month or as small as −0.35 kg/month. The difference in the weight (kg) change between the groups showed a large effect of 0.93 (Cohen ) indicating that the Fit&Trim intervention had potential efficacy.

In , weight loss was greater in the waitlist group after receipt of the Fit&Trim intervention compared to the intervention group that was in the maintenance period. In Phase 2, the intervention group exhibited only a small weight loss [estimated simple slope = −0.33 kg/month (95% CI: −0.62, −0.03)]. The cross-level interaction for weight loss for the waitlist control was significantly greater compared to the intervention group [-0.90 kg/month (95% CI: −1.4, −0.35)]. Cohen's d was not analyzed in Phase 2 because this was not a true cross-over design (see Section 2.1 above).

weight loss by group

The Fit&Trim weight loss goal was to achieve at least an 5% overall weight reduction. As shown in Fig. 2, when each arm received the 3-month intervention, both groups closely approached the 5% mean weight loss goal (intervention group = 4.3% in Phase 1; waitlist control = 4.8% in Phase 2). By 6 months, both groups achieved the 5% mean weight loss goal (intervention group = 5.2%; waitlist control = 5.8%).

Fig. 2

Percent weight change by group: Phase 1 (baseline to 3-months) and Phase 2 (3- to 6-months).

For the Intervention group in , a majority (68%) lost weight compared to only 36% in the waitlist group. Among the intervention participants, 36% achieved the 5% weight loss goal, while 32% lost between 2% and 5% of their starting weight. Only 4% of the participants in the intervention group gained 2%–5% weight. During , while in the maintenance follow-up, a majority (57%) of intervention participants were able to maintain their weight, while 32% continued to lose more weight.

Table 4 presents the percent weight loss categories achieved by those who completed the study in each group (28 intervention and 31 control). For the Waitlist Control group in , almost half (49%) participants maintained their weight, while 6% achieved the 5% weight loss goal. In , after completing the Fit&Trim, almost 70% of waitlist participants lost weight; and almost half (47%) achieved the 5% weight loss goal. Only 3% gained 2%–5% of their starting weight.

Table 4

Percentage weight change achieved by group N = 61.a

Phases	Gained ≥2% to <5%, % (n)	Stable lost <2% % (n)	Lost ≥2% to <5%, % (n)	Lost ≥5% to <10%, % (n)	Overall p-valueb
Phase 1 (base to 3 months)					.002
Interventionc (n = 28)	4 (1)	28 (8)	32 (9)	36 (10)
Waitlist control (n = 33)	15 (5)	49 (16)	30 (10)	6 (2)
Phase 2 (3- to 6-months)					.006
Intervention d (n = 28)	11 (3)	57 (16)	14 (4)	18 (5)
Waitlist control c (n = 32)	3 (1)	28 (9)	22 (7)	47 (15)

Actual data for those completing the study (N = 61) rather than imputed intention to treat data (N = 67) was used for this table.

= Mantel-Haenszel chi-square test for trend.

Received Fit and Trim intervention.

 = in maintenance follow-up.

Other outcomes

In , similar to the between group weight change, BMI and waist circumference exhibited significantly greater reductions in the intervention group compared to the waitlist control, and vice-versa in . Findings for group differences in both phases were statistically significant (Table 3). There was also an initial trend toward lower fasting blood glucose and A1c levels when the intervention and waitlist groups received the Fit&Trim intervention. However, within-group differences were not consistent or statistically significant.

Discussion

The pilot Fit&Trim RCT of a culturally tailored DPP-based weight loss lifestyle intervention augmented with mobile technology was feasible and showed potential efficacy to reduce the risk for T2D in Filipino Americans who were overweight or obese. These findings add to the body of knowledge that supports culturally adapted lifestyle interventions focused on Asian American subgroups [[30], [31], [32]].

The Fit&Trim intervention met the goals for recruitment, engagement and retention. To enhance feasibility, investigators followed recommended strategies for culturally adapting the Fit&Trim intervention to promote acceptability and relevance for the target population [33,34]. Three focus groups (10 participants/group) were conducted with Filipino community members including: the Philippine Nurses Association of Northern California, Filipino Lions Clubs, and 15 individual interviews with community stakeholder (organizational leaders, members, and healthcare providers serving this community). Community feedback informed: the intervention study design, culturally relevant intervention strategies, and potential barriers and facilitators for recruitment, engagement, and retention. Based on stakeholder input, Filipinos were less likely to join or stay in the study if they did not receive the intervention. Therefore, to incentivize participant recruitment and retention, the study design was modified to include a waitlist control group allowing all participants the opportunity to receive the Fit&Trim intervention. Several barriers included family and social pressures to eat high fat, sugar-ladened foods and drinks; and limited knowledge of healthier food/drink options. Therefore, research staff developed a Filipino food pamphlet with healthier options as substitutes for commonly eaten unhealthy Filipino foods and drinks. More details are provided in a prior publication [35].

Recommended recruitment strategies included targeting potential participants from Filipino communities [[36], [37], [38]]. The most effective strategies for identifying potential participants included: select mailings to specified census tract zip codes for Filipino communities (n = 92); and flyers in the community, clinics and academic/research institutions (n = 64). The next most successful strategy was snowball family/friend referrals (n = 39), followed by online ads, such as Craiglist and Facebook (n = 21). The remaining balance of recruited potential participants were from local community events and civic organization presentations.

Engagement

Filipino cultural norms have a strong influence on lifestyle behaviors including: close family/community ties or “Kapwa” (a collectivist social support mindset), frequent community gatherings, a preference for sedentary behaviors, and an aversion to being outside in the sun [39,40]. Moreover, Filipino cultural foods are high in fats (e.g., sausage and pork), carbohydrates (e.g., white rice), and sugary drinks (e.g., soda, Baba and specialty coffee drinks), with limited fruits and vegetable consumption [8,41,42].

Based on stakeholder feedback, cultural norms, and “Kapwa”, we focused on social support for PA and healthier food and drink choices [39,40]. First, family members were welcome to attend intervention sessions to garner social support for healthy weight loss behaviors. To offset cultural influences for unhealthy eating behaviors, participants learned how to modify common Filipino dishes to reduce fat, and add more vegetables and fruit in their diets. They were coached to cope with the social pressures of unhealthy eating during family meals by bringing vegetable dishes to share and gradually substituting healthier brown rice for white rice. A Filipino food pamphlet was developed by Filipino research staff showcasing photos of commonly eaten Filipino foods and drinks including nutritional information (calories, fats, protein, carbohydrates, and salt) that participants could use to log daily food/drink intake.

Filipino cultural norms include an aversion to being outside in the sun. Therefore, the intervention educational slide shows included colored photos of Filipino family members exercising and walking together outside. Participants were also encouraged to engage in common Filipino activities such as, walking, dancing, Zumba, basketball and bowling. Moreover, they learned how to practice these activities indoors at: gyms, indoor malls, and at home on stairways.

Including mobile technologies (e.g., Fitbit tracker with mobile app and Facebook) to monitor health behaviors and provide virtual social support may have motivated participant engagement with healthy lifestyle behaviors, thus promoting increased weight reduction. Future studies should assess the relative contributions of various adaptation strategies used (e.g., mobile technologies, social media, and family participation) that may have influenced intervention outcomes.

Retention

Culturally adapted lifestyle interventions promoting weight loss through increased physical activity and healthy eating have been effective in improving recruitment, engagement, retention, and health outcomes, particularly in at-risk minority populations [17,[43], [44], [45]]. Utilizing community health workers who are familiar with the cultural, language and beliefs has been effective for intervention delivery and promoting recruitment and retention [41,46]. Therefore, we selected and trained our research staff from Filipino communities that were familiar with the local cultural beliefs, norms, and language. This cultural tailoring strategy may have enhanced the intervention program feasibility (recruitment) and improved participant adherence (engagement and retention) for adopting target behaviors, thus promoting the intervention efficacy (increased weight reduction).

Secondary outcomes for overall weight reduction was large, similar, and significant after each study group received the Fit&Trim intervention. In , the intervention group's estimated mean weight loss was −3.4 kg (4.3%), while in , the waitlist group's mean weight loss was −3.8 kg (4.8%). The Fit&Trim study also had greater overall mean weight loss compared to the similar culturally adapted 3-month DPP-based “PILI Ohana” Hawaiian study that reported only a 1.7 kg mean weight loss [31]. In just 3-months, Fit&Trim participants also had a greater mean weight loss (4.3%) compared to the Chinese immigrants who participated in another 6-month DPP-based intervention trial (3.5%) [32].

The PilAm Go4Health study [47], which assessed the same Fit&Trim intervention, but for Filipinos with diagnosed T2D, showed a 2.9% overall weight loss among intervention participants. Participants reported that the cultural tailoring strategies (e.g., using Filipino research staff to facilitate the intervention and mobile technology to track health behaviors) were relevant and acceptable. According to participants, these strategies helped promote self-efficacy and adherence to the target health behaviors [48]. Thus, the cultural tailoring strategies may have influenced participants in the Fit&Trim study to achieve their weight loss goals.

Both Fit & Trim study groups lost similar amounts of weight when they received the intervention (Phase 1 intervention group = 4.3%; Phase 2 waitlist group = 4.8%). In , a majority (68%) of intervention participants lost weight when they received the Fit&Trim intervention. Similarly, in 69% of waitlist participants lost weight after receiving the Fit&Trim intervention. Further evidence for the Fit&Trim's feasibility and sustainability is found in , when over half (57%) the intervention participants sustained their weight loss. In contrast, in a culturally tailored 3-month DPP-based trial to reduce T2D risk in a Gujarati Asian Indian community [30], there was no difference in weight loss between the intervention and control groups at 3–6 months. However, at 6 months, about 35% of the intervention group achieved a 5% weight loss.

During each phase, significant reductions in BMI and waist circumference followed weight loss trends for each study group. However, improvements in fasting plasma glucose and A1c levels post-intervention did not follow weight loss trends, possibly due to the short duration of the Fit&Trim intervention. The effect of a DPP-based intervention on these outcomes is not well understood due to the limited number of prior studies focused on Asian subgroups and wide variability in the design and culturally adapted strategies.

Strengths and limitations

The Fit&Trim trial exhibited several similarities to other Asian American DPP-based interventions. Specifically, the Fit&Trim trial: 1) was an RCT DPP-based lifestyle intervention with an active waitlist control [31], 2) used published guidelines and recommended strategies to culturally tailor the intervention for the target Filipino population [34,38], 3) assessed its feasibility based on recruitment, engagement, intervention completion and retention [30,31], 4) clearly defined the Fit&Trim's cultural adaptations and modifications, and 5) conducted post-program evaluations via individual semi-structure interviews [[49], [50], [51]].

The Fit&Trim sample size was small, limiting the generalizability to other Filipino and general populations. The sample size also precluded analysis of the relative contributions to feasibility and potential efficacy from the various Fit&Trim cultural adaptations strategies. Also, the intervention duration was short (3-month intervention + 3-month maintenance), limiting the possibility of assessing for sustained weight loss and long-term effectiveness. Nevertheless, this culturally tailored Fit&Trim intervention provided substantial evidence for the potential efficacy of a DPP-based intervention for Filipinos. Of the participants who completed the study, 41% achieved the 5% weight loss goal during the 3-month period they received the intervention. Overall, a majority (80%) of Filipino participants lost weight when they received the intervention, thus supporting the Fit&Trim's potential efficacy for reducing obesity related T2D risks.

Conclusion

The Fit & Trim demonstrated that a culturally tailored DPP-based mHealth intervention was feasible with potential efficacy for the understudied population of Filipino Americans who are overweight or obese and at high risk for T2D. Findings from this study add to the body of knowledge about culturally adapted lifestyle interventions focused on Asian American subgroups to reduce health disparities and improve health equity. A study with a larger sample over a longer period of time is needed to assess the Fit&Trim intervention feasibility, effectiveness, and sustainability in a real-world community setting for Filipino Americans.

Conflicts of interest

Authors have none to declare.