The SmartLoss lifestyle intervention approach is grounded in the ability to (1) determine the weight maintenance energy requirement for an individual, (2) establish realistic weight loss and dietary intake goals, (3) objectively quantify the adherence of an individual user to this weight loss and dietary prescription goal, and (4) from the individual data being inputted (ie, body weight) provide immediate and personalized feedback and treatment recommendations. Using data from adults in the National Health and Nutrition Examination Survey and adults who have completed highly controlled clinical trials that led to weight loss [31,32] or weight gain [33,34], we developed and validated dynamic differential equations based on the energy balance model [35-37]. We have demonstrated that these equations allow energy requirements of individuals to be accurately predicted from basic demographic and anthropometric data and, most importantly, the equations generate a prediction for the change in weight over the course of weight loss interventions that include an energy restricted diet and moderate levels of exercise. The model-predicted weight changes are then displayed graphically, which provides a guide for weight change throughout a given program. We further extended these mathematical models to provide accurate estimates of energy intake during weight change by inputting body weights over time [38]. Finally, we developed and validated a model that predicts weight change secondary to high levels of exercise only, and we developed and validated separate models that predict weight change during programs that include both an energy restricted diet and high levels of exercise [39]. The models form the cornerstone of the SmartLoss virtual weight management suite since, based on observed body weights, they provide the ability to quantify adherence to energy balance prescriptions induced by changing dietary intake and/or exercise behaviors.

The functionality of the mathematical models is demonstrated on the Weight Loss Calculator website (Figure 1) [40]. As shown in Figure 1, to use the models the user is required to enter sex, age, weight, and height. The user then can elect to either reduce or increase their calorie intake to view the effect of the energy balance prescription on their body weight. The example in Figure 1 is for a 50-year-old female who is 65 inches tall, weighs 200 pounds, and wishes to reduce her intake by 500 kcal/day. This example is for a client wishing to diet and not dramatically increase levels of exercise, yet exercise-only interventions and interventions that include both energy restriction and high levels of exercise can be accommodated, as detailed in the following section. In such cases, the weight graphs provide a proxy measure of adherence to the energy balance prescription, which promotes an energy imbalance of a specified size through changes to diet and exercise behaviors. As demonstrated in the output, this patient’s weight will decrease by 17.4 pounds or 8.7% over 12 months if adherent to her new energy intake level. This website is free to use, and the SmartLoss intervention relies on these mathematical models to generate a SmartGraph, which shows the predicted course of weight change for the individual as well as the acceptable upper and lower limits of the projected weight change for the chosen energy intake target. As demonstrated in Figure 2, the upper and lower bounds of the weight loss projection present a proxy for adherence to the energy balance prescription, which, in our example, is induced by reducing energy intake. The individual is considered adherent if, throughout the course of the weight management program, the change in body weight falls between the upper and lower bounds, which is termed the “weight zone” or “zone of adherence.” Intervention strategies, including personalized feedback, are delivered based on the actual weight of the individual in relation to this zone [21]. One of the strengths of this approach is the ability to objectively quantify adherence to diet and exercise recommendations based on observed body weight. As described herein, the SmartGraph is automatically created by SmartLoss and updated with new weight data in near real time, and it is viewable by the client and health care or research professional on the SmartLoss clinician dashboard and the app.

SmartLoss relies on weight data that can be automatically imported from commercially available devices. For example, SmartLoss is presently programmed with the Body Trace bathroom scale application program interface (API) to directly import body weight data. The bathroom scale synchronizes with the Internet via a cellular card, and body weight information is automatically and wirelessly transmitted to a server. The SmartLoss system retrieves weight data from the server in near real time and weight is plotted on the SmartGraph (Figure 2), which is available immediately via the SmartLoss app or the SmartLoss clinician dashboard. Additionally, weight data can be hand entered by clients via the mobile phone application or dashboard, or a health care or research professional can enter weight data into the SmartLoss clinician dashboard. Should a client weigh multiple times during the day or enter multiple daily weights, SmartLoss is programmed to plot only the first recorded weight of the day. It is also programmed to eliminate weights not obtained from the client (ie, weights that deviate by +/− 5% from the last available weight). The system is adaptable and the programming logic can be manipulated should a health care or research professional wish to evaluate weight fluctuations throughout the day, for example.

SmartLoss can transmit automated feedback to clients at prespecified intervals or based on weight change characteristics. For example, if the weight of a client is changing as expected (ie, is within the zone of adherence), the feedback includes a congratulatory message and a tip to encourage continued success. Via the clinician dashboard, the health care or research professional can have further interaction with clients by initiating additional personalized feedback messages and treatment advice. The collection of objective weight data (and exercise data, as described below) by clients, the integration of those data into personalized behavioral goals, and the delivery of both automated feedback and customized feedback generated by health care or research professional are illustrated in Figure 3. SmartLoss also provides instantaneous feedback to clients regarding the change in weight relative to the zone of adherence. The program automatically generates color-coded flags indicating if the weight of the client is: (1) within the zone of adherence (green flag); (2) within the zone but plateauing or approaching the upper edge of the zone (green/yellow flag); (3) out of the zone (red flag); or (4) above of the zone but decreasing at a rate that reflects adherence to the energy intake target (red/green flag) and, consequently, if continued will result in the client being back in the zone on the weight graph. If weight exceeds the target and is out of the zone for a given period of time (eg, 3 of 5 days), the client is provided with supportive treatment recommendations to modify energy intake and/or physical activity and foster adherence based on a toolbox approach, which is described below.

SmartLoss is programmed to automatically receive activity information from the FitBit API twice per day. Additionally, step data can be hand entered by participants via the SmartLoss app or dashboard, or a health care or research professional can manually enter weight data into the dashboard. After receiving activity data, the SmartLoss program summarizes the activity goal of the client and current steps per day on a SmartSteps graph for immediate viewing within the mobile phone app and clinician dashboard (Figure 4).

Adoption of a regular program of physical activity is important for weight management and exercise was considered when developing the energy balance models that predict weight loss; hence, the models include a term for activity [35-37]. The model currently used by SmartLoss produces weight loss predictions that are valid for individuals engaging in light to moderate levels of physical activity at the federally recommended level of 150 minutes per week [41]. Should an intervention rely only on higher levels of exercise, a different model that relies on a modified Forbes curve would be used to predict weight change [39]. Similarly, if an intervention promoted a combination of higher levels of exercise and an energy-restricted diet, an alternative model would be used to predict weight change that includes inputs for exercise and restriction of energy intake.

The physical activity component of SmartLoss is designed to encourage individuals to adopt a regular program of physical activity with the goal of increasing their daily steps by 3000 to 4000 steps per day above baseline [42], which results in 7000 to 8000 steps per day [41,43] and is consistent with federal guidelines of 150 minutes per week of moderate intensity activity. Since very few activity monitors provide valid estimates of caloric expenditure, the focus of the exercise goals in SmartLoss is steps per day. Similar to feedback of weight data, individuals receive automated feedback messages about daily adherence to the physical activity goals. Long-term weight loss maintenance is promoted by higher levels of exercise [41,44], and elevated levels of exercise can be promoted and monitored over the long-term through SmartLoss, particularly after weight is lost and clients enter a weight maintenance phase.

SmartLoss can deliver a comprehensive set of health materials called SmartTips to educate clients on strategies for effective behavior change and weight management. A key feature of intensive in-person sessions is the receipt of health information to motivate and inform lifestyle change. During in-person interventions, these materials are generally printed and a health care or research professional reviews this information with clients in group or individual sessions and works with clients to apply the material to their own situation in an attempt to facilitate and maintain behavior change. The content of the SmartTips can be customized for different groups of individuals (eg, post-menopausal women, young men, lactating women) or with a focus toward a particular type of diet (ie, DASH diet, Mediterranean diet, etc). The interval at which the SmartTips are delivered can be customized for individual clients or programs, but participation in intensive intervention programs generally warrants receipt of health information weekly at the onset of the program and can taper off to biweekly and then monthly as the program progresses.

The SmartTips also provide a platform to reinforce and facilitate the use of strategies that effectively promote weight management. For example, use of portioned-controlled foods is an effective strategy in weight management programs and links can be embedded in the SmartTips to direct clients to credible websites where information on a variety of portioned-controlled foods can be obtained. The SmartTips are also interactive, requiring input from the client that acknowledges comprehension of the content. SmartTips also reference the SmartGraphs and recent weight change of the client. SmartLoss is therefore adaptable to many different intervention approaches, from purely automated delivery of SmartTips to the delivery of SmartTips with a counselor remotely reviewing the material with the client via a mobile phone’s multimedia capabilities. Thus, similar to in-person intensive lifestyle interventions, SmartLoss provides a platform for health care or research professionals to customize intervention delivery to the needs of individual clients and facilitate behavior change on an individual level. Further, SmartLoss provides a platform for researcher professionals to test the effects of remotely delivered interventions that include counselor support versus interventions that are fully automated but otherwise identical. Such research is needed, as SmartLoss and similar interventions that have been found to be efficacious were also fairly intense [21-23], while less intense and less directive interventions have produced less weight loss [45,46].

When body weight is repeatedly outside the zone of adherence, it serves as an objective indicator to the client and health care or research professional that more intensive treatment strategies are needed. Different treatment strategies to increase the intensity of the weight management program are maintained in the SmartToolBox. Examples of these strategies include the use of portion controlled foods, increased frequency of contact with the assigned health care or research professional, increased activity or exercise, adoption of a plan to self-monitor food intake, etc. This toolbox approach is similar to the strategy used in the Comprehensive Assessment of Long-term Effects of Reducing Intake of Energy (eg, the CALERIE study) [30,47] and other studies (eg, LookAhead) [28], and it provides a systematic and algorithmic method to improve adherence to diet and weight goals.

When using this approach, the client and health care or research professional are prompted by the SmartLoss app and dashboard to first select less intensive strategies from the toolbox, followed by more intensive and frequently more expensive strategies if the less intense strategies fail to result in the desired outcome (usually weight loss) over a given period of time (eg, 2 weeks). Treatment strategies provided within the SmartToolBox are specified by the clinic and customized for the type of weight management program being delivered and to meet the needs of individual clients (Figure 5). Use of the toolbox is tracked and quantified by the clinician dashboard. One aim of the toolbox approach is to build self-efficacy as selection of toolbox strategies is a collaborative process between the client and their healthcare or research professional, particularly early in treatment. Options are personalized to meet the individual client’s needs based on their current circumstances and abilities. As treatment progresses, the client learns to identify potential problems, evaluate and select a strategy to overcome the problem, and track the success of the strategy. This framework follows an active problem-solving approach and builds accountability and self-efficacy since the client is actively involved in choosing methods to improve adherence, evaluating the effectiveness of the chosen solution, and selecting an alternative solution if needed.

The clinician dashboard allows users with administrator and healthcare or research professional rights to generate reports that summarize data on usage of the SmartLoss app, as well as reports showing weight and exercise data at the individual and group level. Usage data includes: (1) the number of days that body weight and step data were recorded, (2) the number of days clients were enrolled in the program, (3) the percent of days with weight and step data, (4) the number of times participants viewed the weight and step graphs, and (5) the number of times and which SmartTips were viewed. Reports can also be generated that summarize the following body weight data: (1) client body weight over time, (2) the target/goal weight and the upper and lower bounds of the zone of adherence, (3) client weight change (pounds and percent) over time, and (4) the number of days and percent of time that clients were in, below, or above the zone of adherence. Finally, reports can be generated that summarize SmartToolBox use and weight change and step counts during the time that different toolbox options were utilized. All reports can be viewed within the clinician dashboard and the raw data can be downloaded into other software packages for analysis.