This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mhealth and uhealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, as well as this copyright and license information must be included.
The prevalence of obesity and associated metabolic conditions continue to be challenging and costly to address for health care systems; 71% of American adults were overweight, with 35% of men and 40% of women diagnosed with obesity in 2014. Digital health coaching is an innovative approach to decreasing the barriers of cost and accessibility of receiving health coaching for the prevention and management of chronic disease in overweight or obese individuals.
To evaluate the early impact of a mobile phone-based health coaching service on weight loss and blood pressure management in a commercially insured population.
This was a retrospective study using existing registry data from a pilot commercial collaboration between Vida Health and a large national insurance provider, which enrolled adult members who were overweight (body mass index >25 kg/m2) and able to engage in a mobile phone-based coaching intervention. Participants received 4 months of intensive health coaching via live video, phone, and text message through the Vida Health app. Participants were also provided with a wireless scale, pedometer, and blood pressure cuff. Of the 1012 enrolled, 763 (75.40%) participants had an initial weight upon enrollment and final weight between 3 and 5 months from enrollment; they served as our intervention group. There were 73 participants out of the 1012 (7.21%) who had weight data 4 months prior to and after Vida coaching, who served as the matched-pair control group.
Participants in the intervention group lost an average of 3.23% total body weight (TBW) at 4 months of coaching and 28.6% (218/763) intervention participants achieved a clinically significant weight loss of 5% or more of TBW, with an average of 9.46% weight loss in this cohort. In the matched-pair control group, participants gained on average 1.81% TBW in 4 months without Vida coaching and lost, on average, 2.47% TBW after 4 months of Vida coaching, demonstrating a statistically significant difference of 4.28% in mean percentage weight change (P<.001). Among 151 intervention participants with blood pressure data, 112 (74.2%) had a baseline blood pressure that was above the goal (systolic blood pressure >120 mmHg); 55 out of 112 (49.1%) participants improved their blood pressure at 4 months by an entire hypertensive stage—as defined by the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure.
Mobile phone app-based health coaching interventions can be an acceptable and effective means to promote weight loss and improve blood pressure management in overweight or obese individuals. Given the ubiquity of mobile phones, digital health coaching may be an innovative solution to decreasing barriers of access to much-needed weight management interventions for obesity.
The prevalence of obesity and associated metabolic conditions continue to be challenging and costly to address for health care systems; 71% of American adults were overweight, with 35% of men and 40% of women diagnosed with obesity in 2014 [
With primary care clinicians in increasingly short supply and overwhelmed by burdens of the preventative and chronic care needs of their patient panel [
In addition to decreasing the barriers of cost and accessibility for receiving health coaching, digital health coaching provided remotely through an Internet-connected device has been shown to be an effective approach for the prevention and management of cardiometabolic risk factors in overweight and diabetic individuals [
While studies of Web- [
This was a retrospective study using existing registry data from a pilot commercial collaboration between Vida Health and a large insurance provider. Interested members under this insurance provider were enrolled starting in November 2015 in a commercial program using a mobile phone-based, digital health-coaching app, Vida Health (described below). The coaching platform was used to manage cardiometabolic conditions such as obesity, hypertension, hyperlipidemia, and prediabetes. This study has received Institutional Review Board (IRB) exemption from the University of California, San Francisco (UCSF), IRB (IRB No. 16-19903).
The registry included adults over the age of 18 years who were fully insured members in the US states of Wisconsin, Georgia, and Colorado and who were overweight—BMI >25 kg/m2. Participants had to be English speaking and own an advanced-feature mobile phone—iPhone or Android—to ensure they were able to engage in the Vida Health coaching program. Participants were recruited from an email campaign by their insurance provider offering the free program for any current member who had a BMI >25 kg/m2. In order not to bias the sample with individuals who would have achieved weight loss and other health goals regardless of whether or not they worked with a coach, the barrier to entry into the program was kept at a minimum—simply replying to a Web form in an email invite—with minimal exclusion criteria, including not owning a mobile phone and having type 1 diabetes. Clinical data elements used for enrollment, including weight, height, and medical history, were based on self-reported responses. Invitations to the program were sent randomly to members in eligible geographies and the first 1000 participants were included in this pilot program.
Enrollment for the Vida Health program started November 2015. Prior to starting the Vida Health program, accepted members were sent a package containing a Bluetooth-connected pedometer and wireless scale. Those members with known hypertension also received a Bluetooth-enabled blood pressure cuff. Participants received instructions on how to synchronize their wireless devices for passive data collection through the Vida app.
After installing the Vida app, participants were asked to complete an onboarding survey regarding their baseline health behaviors, past medical history, self-selected personality preferences (ie, if they might benefit more from a “cheerleader”- or “drill sergeant”-like coach), and general availability. Based on these attributes, participants were matched with a short list of Vida-recommended coaches from which they could select their own ongoing health coach. Health coaches were professional licensed nutritionists, physical therapists, and social workers who are certified to provide health coaching and were additionally trained by Vida once employed in the Vida network.
The first 4 months of the Vida program consisted of an intensive
In addition to data passively collected by the wireless scales, pedometers, and blood pressure cuffs provided to participants, members were also asked to enter their activity (ie, steps, exercise, food intake, sleep, and stress levels). Participants who had difficulty using or synchronizing the provided devices had the option of self-entering biometric data (ie, weight and blood pressure) directly into the Vida app. For the purposes of accurate analyses, we excluded metric points that were deemed to be unrealistic outliers that fell outside of the scope of clinical weight loss or the trend of the participant. Members were not paid for their participation. Program costs, including both the coaching service and associated hardware devices, were covered by their insurance provider.
The primary outcome of the study was weight loss at 4 months as defined by percent change in total body weight (TBW) to normalize against a range of starting body weights. Participants were encouraged to weigh themselves weekly during the 4 months of the intensive coaching program. Weight data was primarily collected through the direct transmission of values via a Bluetooth scale provided as part of the intervention. Some participants chose to self-enter data due to technical difficulties setting up and using the Bluetooth scale. Change in TBW was calculated as the difference between the first weight since program enrollment and the last weight entered closest to the end of 4 months of coaching, with the caveat that the provided weight was recorded between 3 and 5 months after program enrollment.
Secondary outcomes included change in systolic blood pressure (SBP) after 4 months of intensive health coaching, as well as the change in number of participants in each hypertensive category (ie, normal, prehypertensive, type 1 hypertension, and type 2 hypertension) from the beginning of enrollment to after 4 months of coaching.
Satisfaction with the intervention was summarized using participant-reported ratings of the Vida Health app. Throughout the program, participants were asked, “How are we doing? Please help us improve your experience by taking a minute to leave feedback for your coach.” Participants could input text feedback and rate their experience on a scale of 0-10, with 10 being the highest rating.
Given that this study was based off of registry data from an existing pilot program, a designated control arm that did not receive coaching was not defined. However, there was a subset of individuals (n=73) enrolled in the program who had historic weight data from owning Bluetooth scales prior to starting the program. Using the Validic application programming interface, these historic weight data points were added to the Vida app database when individuals synchronized their devices as part of the program. These individuals’ historic weight data 4 months prior to starting the Vida program were used as a proxy for a matched-pair control group, as these individuals successfully enrolled and received coaching through the program, thereby eliminating any bias toward motivation.
Summary statistics describing the demographic characteristics are provided in
Participant data were statistically analyzed using R version 3.3.3 (The R Foundation) using descriptive analysis and two-sided
Similarly, baseline blood pressure was designated as the first blood pressure reading since enrollment. Final blood pressure was designated as the blood pressure reading taken between 12 and 20 weeks that was closest to 4 months from enrollment. The mean change in SBP was calculated and assessed for significance using a two-sided
Baseline demographic summary statistics.
Characteristics | Enrolled participants (N=1012) | Intervention group (n=763) | Matched-pair control group (n=73) | Enrolled versus intervention, P | Intervention versus control, P | |
Age (years), mean (SD) | 44.63 (11.25) | 44.78 (11.18) | 42.36 (10.28) | .41a | .41a | |
Sex (male), n (%) | 337 (33.30) | 261 (34.2) | 17 (23) | .73b | .08b | |
Starting BMIc (kg/m2), mean (SE) | 33.50 (0.21) | 33.34 (0.24) | 34.33 (0.69) | .59a | .19a | |
.96b | .31b | |||||
25.0-29.9 | 368 (36.36) | 283 (37.1) | 19 (26) | |||
30.0-34.9 | 305 (30.14) | 234 (30.7) | 26 (36) | |||
35.0-39.9 | 189 (18.67) | 138 (18.1) | 15 (21) | |||
≥40.0 | 150 (14.82) | 108 (14.2) | 13 (18) | |||
.78b | .02b | |||||
Colorado | 195 (19.27) | 155 (20.3) | 26 (36) | |||
Wisconsin | 626 (61.86) | 457 (59.9) | 36 (49) | |||
Georgia | 168 (16.60) | 136 (17.8) | 11 (15) | |||
Other | 23 (2.27) | 15 (2.0) | 0 (0) | |||
.87b | .54b | |||||
Prediabetes | 57 (5.63) | 44 (5.8) | 4 (5) | |||
Diabetes | 41 (4.05) | 25 (3.3) | 0 (0) | |||
Hypertension | 157 (15.51) | 109 (14.3) | 7 (10) | |||
Hyperlipidemia | 124 (12.25) | 93 (12.2) | 7 (10) |
a
b
cBMI: body mass index.
Participant enrollment and retention in the program is displayed
Study population including enrollment and retention.
Baseline demographics and chronic disease profiles of all enrolled participants, intervention participants, and matched-pair controls with weight data are reported in
As shown in
The level of engagement impacted the amount of weight loss among study participants. There were 306 participants in the high-engagement cohort, defined as participants who were at the top quartile of messages sent per month or number of coaching consults in the 4-month coaching period. There were 74 participants in the low-engagement cohort, defined as participants at the bottom quartile of messages and video consults. There were 383 participants in the medium-engagement cohort. The high-engagement cohort lost the most weight, followed by the medium- and then low-engagement cohorts (see
Percentage change from total body weight in intervention group.
Weight and TBWa measures | Weight, mean (SE) or |
Number of participants |
Baseline weight (kg), mean (SE) | 96.23 (0.78) | 763 (100) |
Weight at 4 months (kg), mean (SE) | 92.99 (0.76) | 763 (100) |
Mean weight loss at 4 months, |
-3.23 (0.22)b | 763 (100) |
≥5% weight loss at 4 months, |
-9.46 (0.41)b | 218 (28.6) |
≥2% to <5% weight loss at 4 months, |
-3.41 (0.06)b | 210 (27.5) |
≥-2% to <2% weight change at 4 months, |
-0.30 (0.07) | 255 (33.4) |
>2% weight gain at 4 months, |
+4.90 (0.81)b |
80 (10.5) |
aTBW: total body weight.
b
Percentage change from total body weight for matched-pair control group before and after Vida program.
Outcome | Weight (n=73), mean (SE) or |
Weight 4 months before Vida program (kg), mean (SE) | 95.44 (2.26) |
Weight at enrollment (kg), mean (SE) | 97.10 (2.29) |
Weight 4 months after Vida program (kg), mean (SE) | 94.65 (2.26) |
4 months before Vida program compared to enrollment, mean % TBW change (SE) | 1.81 (0.41)b |
4 months after Vida program compared to enrollment, mean % TBW change (SE) | -2.47 (0.48)b |
aTBW: total body weight.
b
Relationship between engagement and weight loss.
Level of cohort engagement | % total body weight loss, |
Number of messages per month sent to coach, |
Number of consults over 4 months, |
Low (n=74) | -1.37 (0.63) | 4.83 (2.73) | 1.51 (0.20) |
Medium (n=383) | -2.84 (0.29) | 42.65 (7.14) | 5.70 (0.33) |
High (n=306) | -3.86 (0.34) | 120.59 (12.55) | 10.71 (0.26) |
Of 151 participants with blood pressure data in Vida, the baseline average SBP was 131 mmHg and the mean change in SBP was a 6 mmHg decrease after 4 months (see
Mean percentage change in weight in matched-pair controls during the 4 months before coaching and 4 months after coaching.
Mean change in systolic blood pressure after 4 months of coaching (n=151).
Outcome | Blood pressure (mmHg), |
Baseline SBPa | 131.27 (1.52) |
SBP after 4 months | 125.31 (1.18) |
Mean change in SBP after 4 months | -5.96 (1.64)b |
aSBP: systolic blood pressure.
b
Change in hypertensive category after 4 months of coaching.
Of 1012 enrolled participants, 386 (38.14%) submitted ratings of the app, with an average rating of 9.77 out of 10, with 10 being most satisfied (see
Satisfaction data.
Participant cohort | Rating of the app (out of 10), |
Participants who provided a rating, |
All enrolled participants (N=1012) | 9.77 (0.92) | 386 (38.14) |
Participants in intervention group (n=763) | 9.81 (0.73) | 333 (43.6) |
In this study, a mobile phone-based health-coaching intervention was found to be effective in reducing weight and blood pressure in overweight and obese adults over a 4-month period, with participants in the matched-pair control group demonstrating a statistically significant difference of 4.28% in mean percentage weight change.
Unlike other interventions targeting more complex conditions [
Key aspects of the intervention included the ability for participants to self-monitor progress and be accountable for goals set with their health coach. Self-monitoring is the most common feature of mobile apps for health intervention [
There was notable weight loss in the intervention group at 4 months, with an average total body weight loss of 3.23%. In addition, 28.6% of participants achieved a clinically significant weight loss of 5% TBW or more, with an average of 9.5% TBW lost in this cohort. The change in weight loss is even more significant when taking into account the fact that the intervention began in late November through the winter holiday season when people have the hardest time losing weight [
Unique in this analysis was the availability of 4-month pre- and postcoaching data for a subset of participants that could act as their own matched-pair controls, thereby controlling for the role of intrinsic motivation and allowing the study to isolate the impact of Vida coaching on weight loss. Given that participants in this cohort gained 1.81% TBW over the 4 months prior to using Vida and lost 2.47% TBW over the 4 months after starting Vida suggests the true impact of Vida coaching to be even greater than measured, which is due to the likely continued weight gain in the absence of additional coaching intervention. The dose-dependent relationship between engagement and weight loss further demonstrates that weight loss was driven by engagement with the coach.
In the cohort of 151 participants who had blood pressure data, Vida coaching had a positive impact in reducing average SBP by 5.96 mmHg and by moving significant numbers of participants within each blood pressure stage to a better blood pressure stage after 4 months of coaching. Overall, half of the participants with blood pressure data improved their blood pressure by an entire hypertensive stage after 4 months of coaching, with greater proportions of participants improving at higher stages of baseline hypertension. There is limited data in existing literature regarding the impact of digital health interventions on blood pressure in a comparable cohort. One meta-analysis found that digital health interventions did not have statistically significant impact on blood pressure [
There were limitations to this study. Criteria for enrollment were largely self-reported. Participants skewed toward having more female than male participants and toward being a relatively young cohort, making generalization to older individuals more difficult. While participants were geographically diverse, it is challenging to draw conclusions around the impact of the intervention on minority populations or populations of varying education and income brackets, given the lack of demographic data around ethnicity and socioeconomic status. All participants were mobile phone owners, which is associated with higher socioeconomic status and, therefore, possibly less prevalence of obesity and chronic disease [
Despite these limitations, this study demonstrates that a mobile app-based coaching intervention can be an acceptable and effective means to achieving desired clinical results for overweight and obese individuals. As mobile phones continue to penetrate the consumer market, digital health coaching may serve as a promising model to increase access to evidence-based behavioral coaching for obesity and related cardiovascular conditions.
body mass index
Institutional Review Board
Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of
systolic blood pressure
short message service
total body weight
University of California, San Francisco
The authors would like to acknowledge the UCSF Center for Excellence for their support in this work through the Primary Care Leadership Academy Fellowship.
AYM and CC were formerly employed at Vida Health, a company that provides digital health coaching, which is the subject of this study. AYM is a paid independent consultant and KCB is a current employee at Vida Health. JNO and CM have no conflicts of interest.