A Review of Technology-Based Interventions to Maintain Weight Loss

Abstract

Background: For many decades, healthcare providers and researchers have developed weight-loss interventions to help people achieve weight loss. Unfortunately, it is typical for people to lose weight quickly during the intervention period but then slowly regain weight until they return to their approximate baseline. Technology-based maintenance interventions are among the newest approaches to long-term weight loss. Several advantages make technology helpful for maintaining weight loss. The purpose of this article was to review and critique the randomized controlled trials of technology-based weight-loss maintenance interventions (WLMIs) for adults. Materials and Methods: A systematic search through electronic databases and a manual citation search were conducted. Limited numbers of controlled trials published since 2000 that included randomization, and technology-based WLMIs were identified. Results: The characteristics of the eight studies were diverse. The average score of study design quality was moderate. The results of the effectiveness of technology-based WLMIs were mixed. Technology-based WLMIs are more likely to be effective than usual care but not more effective than personal contact. Conclusions: Based on the review, guidelines were established for the selection and potential success of technology-based WLMIs. The effectiveness of technology-based maintenance interventions for weight loss varied, and potential strategies and approaches are discussed to improve their effectiveness. Further studies are needed to better evaluate and refine the efficacy of technology-based WLMIs.

Introduction

Obesity is problematic in American society and worldwide. In 2011 and 2012, more than two-thirds of adults were reported to be obese or overweight, more than one-third of adults were obese, and about 6.4% were extremely obese.¹ The World Health Organization reported that being overweight or obese is the fifth leading risk factor for global death and leads to at least 2.8 million adult deaths every year.^2,3 Overweight and obese individuals have a higher risk for diseases such as coronary heart disease, stroke, hypertension, type 2 diabetes, and certain types of cancer.⁴ Obese individuals may also suffer from psychosocial problems such as depression, anxiety, low self-esteem, low quality of life, social stigmatization, and discrimination.^3,4 In addition, these negative outcomes result in higher healthcare costs. Recent reports estimate that weight-related healthcare costs total almost $147 billion in out-of-pocket, private, and insurance expenses.^4,5 Clearly, obesity is a significant problem in American society, one with adverse physical, social, psychological, and financial consequences.

An important fact is that obesity is largely preventable² and remediable.⁵ For decades, healthcare providers and researchers have developed and studied the effects of weight-loss interventions (WLIs). They have found that WLIs reduced blood pressure, serum triglycerides, blood glucose, and serum cholesterol levels and increased high-density lipoprotein levels.³ Indeed, weight loss can make individuals healthier and reduce their risk factors for chronic diseases.

Although overweight and obese individuals do lose weight successfully through these kinds of WLIs, previous studies have reported that participants quickly regain the weight after completing the interventions.⁶ A recent report asserted that most participants regained about one-third of the weight lost during the year following the completion of an intervention program and tended to return to their baseline weight within 3–5 years.⁷ This confirms that weight-loss maintenance is extremely challenging for participants who lose weight through short-term interventions. Researchers have developed and examined several strategies to maintain weight loss, such as ongoing therapist contact, training in relapse prevention, problem-solving therapy, providing prepackaged foods, and incorporating peer support.⁶

Another weight-loss maintenance strategy—technology-based interventions—is one of the newest approaches and has several advantages. First, technology makes the information easier and quicker to access.⁸ People can easily access Internet- or mobile phone-based programs to provide reminders to maintain their healthier lifestyles despite their other commitments. Second, technology-based interventions are less expensive than in-person interventions over the long term. Many individuals can access technology-based programs anytime and anywhere. Computers, tablet PCs, smartphones, smart TVs, and wireless Internet access have quickly grown in popularity. Companies are developing new fitness devices to monitor individuals' physical activities. The availability of such technology means there is great potential to augment or improve interventions and to offer participants powerful tools. In addition, technology-based interventions can provide a private, varied, and rapid contact using computers, tablet PCs, or smartphones that can access WLIs through the Internet or downloadable applications.⁹ Secure, personalized access can be used to provide enhanced cultural and personal sensitivity.⁸ Therefore, particular programs can be personalized for specific users.

A previous meta-analysis revealed significant evidence that the use of Web-based interventions improved behavioral change results and increased exercise time, knowledge of nutritional status, and knowledge of healthcare, among other concerns.¹⁰ Also, several reviews have recently examined the effectiveness of technology-based interventions.^4,11,12 However, there are relatively few weight-loss maintenance interventions (WLMIs), and few studies have discussed technology-based WLMIs.

The purpose of this article is to review the randomized controlled trials of technology-based WLMIs for adults published since 2000. The following research questions were the focus of this review:

• Q1: What are the characteristics of the primary studies of technology-based WLMIs?

• Q2: What is the quality of study designs of the primary studies?

• Q3: Are the technology-based maintenance interventions effective in maintaining weight loss?

Materials and Methods

Article Identification Search Strategy

A systematic literature review of scientific journal articles published since 2000 was conducted using the following five electronic databases: OVID Medline, CINAHL, PsycINFO, Compendex, and PubMed. Key words were selected in accordance with the review purpose to include three main concepts of interest: purpose and length of intervention, outcomes, and technologies used. Therefore, key words were “long term*,” “maintain*,” “sustain*,” “weight loss,” “weight gain,” “body weight change,” “telecommunication,” “telephone,” “cellular phone,” “Internet,” “telemedicine,” “technology,” and “wireless technology.” The search was limited to journal articles published in English since 2000. In addition, a manual citation search was performed of the reference lists of the selected articles.

Selection Criteria

Table 1 shows inclusion and exclusion criteria. The criteria were developed to narrow the results and to focus on the rigorous research studies conducted on technology-based WLMIs. Selection criteria were established according to population, study design, intervention, publication type, and outcome. Only journal articles published after 2000 were included because technology-based interventions are a relatively recent development. Articles were included when they were published in a scientific journal, reported quantitative body weight changes, and conducted a technology-based intervention for maintenance of weight loss that performed random assignment.

Table 1.

Search Criteria

INCLUSION	EXCLUSION
Adult population (18 years of age and older)	Targeted specific illnesses
Published after year 2000	Participating in eating disorder groups or post–gastric bypass surgery groups
Published in a scientific journal	Secondary analysis
Reported quantitative body weight change	Review articles
Conducted technology-based intervention for maintenance of weight loss	Performed short-term weight-loss intervention only
Performed random assignment	Not in English language

Articles that targeted specific illnesses were excluded, except for those that focused on cardiovascular disease risk factors, such as hypertension, diabetes, and dyslipidemia. Because many studies' long-term goals were to prevent cardiovascular diseases or other chronic diseases, cardiovascular disease prevention studies were also considered relevant to the present review. Eating disorder groups and post–gastric bypass surgery groups were excluded because the specific nature of their disease or surgery serves as a guide to different intervention strategies. Secondary analysis studies, such as cost-effective analyses and review articles, were excluded. Those articles that reported only WLI and did not report WLMI were also excluded (see Appendix 1).

Data Synthesis

To examine the characteristics of primary studies, data were extracted from the original studies. The study quality was measured according to the nine methodological characteristics tool developed by Norman et al.⁸ (see Appendix 2). Originally, Norman et al.⁸ established the study quality coding criteria to rate methodological characteristics for e-health interventions. In this article, the study quality coding criteria were adapted for this review to assess the quality of the studies identified. The score given to each study was calculated as a percentage of the maximum score. In addition, each study was categorized by the level of effectiveness of the technology-based WLMI to facilitate the maintenance of weight loss. Meta-analysis was not conducted because of the heterogeneity of the studies and the lack of data.

Results

Study Selection

Eight original studies were identified. Figure 1 shows the flow diagram of study selection that was used for identifying the studies. The 166 records were identified through the electronic search databases OVID Medline, CINAHL, PsycINFO, Compendex, and PubMed. An additional eight records were identified through a manual search of article reference citations. After duplicates were removed, 82 studies remained. Then, 41 articles were excluded after articles were screened by title and abstract. Twenty articles that did not meet the inclusion criteria were excluded, and 21 WLI studies using technology were identified.^{7,9,13

–31}

Fig. 1.

Flow diagram of study selection.

Many technology-based WLIs have been conducted over the last decade using the telephone,^13

–16 the Internet,^{7,9,14,17

–21} remote support,²² computer,²³ e-mail,²⁴ Web sites,^13,25
–27 mobile phones,²⁸ text messaging,²⁹ a tailored expert system,³⁰ or interactive technology.³¹ This article focused on technology-based WLMIs. This focus led to 13 studies being excluded because they examined only WLIs (during primary program implementation), not maintenance interventions. Appendix 1 provides a summary of the 13 studies and the reasons they were excluded. Finally, eight studies were identified that used randomized controlled trials and conducted technology-based WLMIs; these were subjected to description and evaluation. As an exception, the study of Wing et al.⁷ added value because it focused on the maintenance intervention for participants who lost weight over long periods.

Characteristics of Included Studies

All eight studies conducted two-arm^9,17 or three-arm^{7,15,19,20,21,31} randomized controlled trials. Table 2 shows the summary of these eight primary studies according to population, intervention, comparison, outcomes, setting, and publication year (time). The most recent study was published in 2012.¹⁷ In each of these studies, more women were recruited than men. The majority of the studies' participants were in their forties^{9,17,19,20,21} and fifties.^7,15,31 The participants were either overweight or obese (mean body mass index range, 27.5–37.1 kg/m²). Their health characteristics were diverse. Participants were generally healthy adults (employed),¹⁷ perimenopausal women,⁹ obese women in rural areas,¹⁵ women with hypertension, dyslipidemia, or both,³¹ or healthy, overweight, and obese adults.^19
–21

Table 2.

Summary of Characteristics of the Primary Studies

	BASELINE CHARACTERISTICS OF PARTICIPANTS
REFERENCE (YEAR)	N ^a	MEN:WOMEN RATIO	AGE (YEARS)	BMI (KG/m²)	PARTICIPANTS	OUTCOMES	INTERVENTIONS	MEASURE TIME	STUDY DESIGN; AMERICAN STATE
Thorndike et al.¹⁷ (2012)	330	158:172^b	EG: 44.2 (11.8) CG: 41.6 (13.6)^c	EG: 28.0 (5.8) CG: 27.5 (5.9)	Employees	Weight, height, WC, BP, FL, FSG, PA, NB, overall health	10-week WLI: nutrition and physical activity program (n=506) 9-month WLMI: EG: Internet and personal contact (n=174, 157 analyzed); CG, usual care (n=156, 145 analyzed)	Baseline, 10 weeks, 12 months	RCT; MA
Cussler et al.⁹ (2008)	135	0:135	EG: 48.0 (4.6) CG: 48.4 (4.3)	EG: 31.0 (3.9) CG: 30.4 (3.3)	Perimenopausal women	Weight, height, PA (EEE), DI (EI), characteristics of Internet use	4-month WLI: cognitive-behavioral therapy-based program (n=161) 12-month WLMI: EG, Internet (n=66); CG, self-directed (n=69)	Baseline, 4 months, 16 months	RCT; AZ
Perri et al.¹⁵ (2008)	234	0:234	EG1: 59.8 (6.2) EG2: 59.2 (6.2) CG: 58.6 (6.0)	EG1: 36.9 (5.7) EG2: 37.1 (4.5) CG: 36.2 (4.3)	Obese women in rural area	Weight, height, BP, FL, FSG, adherence, cost	6-month WLI: lifestyle modification program (n=298) 12-month WLMI: EG1, telephone counseling (n=72); EG2, face-to-face counseling (n=83); CG, education (n=79)	Baseline, 6 months, 18 months	RCT; FL
Svetkey et al.³¹ (2008)	1,032	378:654	EG1: 55.7 (8.5) EG2: 55.4 (9.1) CG: 55.8 (8.5)	EG1: 34.2 (4.9) EG2: 34.2 (4.8) CG: 34.0 (4.8)	Overweight or obese and had risk factors	Weight, PA, CI	6-month WLI: group-based behavioral intervention (n=1,685) 30-month WLMI: EG1, interactive technology (n=348, 347 analyzed); EG2, personal contact (n=342, 341 analyzed); CG, self-directed (n=342, 341 analyzed)	Baseline and 6, 12, 18, 24, 30, and 36 months	RCT; NC
Wing et al.⁷ (2006)	314	59:255^b	EG1: 50.9 (9.3) EG2: 51.0 (10.3) CG: 52.0 (10.8)	EG1: 28.1 (4.6) EG2: 28.7 (4.7) CG: 29.1 (5.0)	Those who had lost at least 10% of their body weight during the prior 2 years	Weight, PA, EI	Orientation (n=521) 18-month WLMI: EG1, Internet intervention (n=104); EG2, face-to-face intervention (n=105); CG, control (n=105)	Baseline and 6, 12, and 18 months	RCT; RI
Harvey-Berino et al.¹⁹ (2004)	255	46:209	EG1: 46.5 (9.8) EG2: 45.2 (8.9) EG3: 46.5 (7.7)	EG1: 29.3 (5.2) EG2: 28.9 (3.8) EG3: 29.0 (4.3)	Healthy overweight and obese adults	Weight, PA (EEE), DI, adherence, attendance, acceptability, peer support	6-month WLI: interactive television weight-loss program (n=255) 12-month WLMI: EG1, Internet support (n=77); EG2, frequent in-person contact (n=77); EG3, minimal in-person contact (n=78)	Baseline and 6, 12, and 18 months	RCT; across the United States
Harvey-Berino et al.²⁰ (2002)	122	18:104	EG1: 46.3 (11.1) EG2: 49.8 (8.4) EG3: 49.1 (9.1)	EG1: 32.2 (4.0) EG2: 31.5 (4.8) EG3: 32.8(4.6)	Healthy overweight and obese adults	Weight, PA (EEE), DI, adherence, attendance, acceptability, peer support	6-month WLI: in-person behavioral obesity treatment (n=122) 12-month WLMI: EG1, Internet support (n=40); EG2, frequent in-person support (n=41); EG3, minimal in-person support (n=41)	Baseline and 6, 12, and 18 months	RCT; VT
Harvey-Berino et al.²¹ (2002)	46	9:37	46.3 (7.4)^d	33.7 (4.6)^d	Healthy overweight and obese adults	Weight, PA (EEE), DI, adherence, attendance, acceptability, peer support	15-week WLI: behavioral weight control program (n=46) 22-week WLMI: EG1, Internet (n=15); EG2, in-person, therapist-led (n=14); CG, no treatment (n=15)	Baseline, 15 weeks, 37 weeks	RCT; VT

Data for age and body mass index (BMI) are mean (standard deviation) values.

Total number in study (at randomization).

Recalculated from the original data.

p<0.05, for differences between groups.

Not reported by groups.

BP, blood pressure; CG, control group; CI, caloric intake; DI, dietary intake; EEE, exercise energy expenditure; EG, experimental group; EI, energy intake; FL, fasting lipids; FSG, fasting serum glucose; NB, nutrition behavior; PA, physical activity; RCT, randomized controlled trial; WC, waist circumference; WLI, weight-loss intervention; WLMI, weight-loss maintenance intervention.

The studies focused on the changes in the participants' weight, body mass index (BMI), blood chemistry, cardiovascular disease risk factors, physical activity, and dietary intake. Additionally, one study measured the participants' self-reported overall health as “excellent, very good, good, fair, or poor.”¹⁷ Another study analyzed the Spearman correlation coefficients of the Internet log entries with changes in weight, exercise energy expenditures, and energy intake during the maintenance period.⁹ Several investigators analyzed the adherence^15,19
–21 to and cost-effectiveness¹⁵ of weight-loss programs. The studies included maintenance interventions immediately following the same short-term interventions and randomization. The WLIs were 10 weeks–6 months long, and the WLMIs were 22 weeks–30 months long. Svetkey et al.³¹ performed the longest intervention, which lasted 30 months.

Table 3 summarizes the primary studies' measurements and results. In terms of retention rates, the rates for Experimental Group 1 (technology-based WLMIs) were lower than those of the control group or Experimental Group 2 or 3 (non–technology-based WLMIs). During the WLI period, participants lost significant amounts of weight (−2.7 to −9.4 kg) except in the study of Wing et al.⁷ Among the studies reviewed, the participants in the study of Perri et al.¹⁵ lost the largest proportion of body weight. However, all participants regained the weight during the maintenance period (0.4–5.5 kg) except those in the study of Harvey-Berino et al. The longer WLMI studies tended to report more weight regain.³¹

Table 3.

Summary of Measurements and Results

		SIGNIFICANT MAIN OUTCOMES
REFERENCE (YEAR)	RETENTION (%)^a	WEIGHT (KG)^b	PA	DIETARY (ENERGY, CALORIC) INTAKE	OTHERS
Thorndike et al.¹⁷ (2012)	506 406 (80) 330 (81) EG: 120 (76) CG: 118 (81)	0–10 weeks: EG, −2.17^c; CG, −1.95^c 0–12 months: EG, −1.54^c; CG, −1.17^c 10 weeks–12 months: EG, +0.63^c; CG, +0.78^c No group differences	At 12 months: time spent in moderate PA increased by ≥2 h/week for 30%; climbing ≥5 flights of stairs/day increased from 42% to 49%^d	Baseline versus 12 months: eating fruits and vegetables ≥3 times/day, 38% versus 57%^e; never eating fatty foods and snacks, 51% versus 70%^e; never eating sugary foods and beverages, 26% versus 42%^e	BMI, WC, and total and low-density lipoprotein cholesterol were lower at 10 weeks and 1 year (but no group differences) Overall health^e: baseline versus 12 months as “excellent,” 13% versus 25%
Measurement conditions		With clothes and shoes off, using Salter lithium electronic scale model 9037	By asking participants to estimate (1) time spent per week during the past 3 months in exercise, (2) average flights of stairs climbed daily, and (3) normal walking pace	Estimating the number of time a day over the last month that a participant had eaten the following categories of food: fruits and vegetables, fatty goods and snacks, and sugary foods and drinks
Cussler et al.⁹ (2008)	300161 (54)135 (84)EG 52 (79)CG 59 (86)	0–4 months: EG, −5.3 (3.6); CG, −5.2 (3.8)0–16 months: EG, −4.9^c; CG, −4.6^c 4–16 months: EG, +0.4 (5.0); CG, +0.6 (4.0)No group differences	EEE (kcal/day): for 0–4 months, EG of 151 (196), CG of 144 (151); for 4–16 months, EG of 55 (301), CG of 62 (279)	EI (kcal/day): for 0–4 months, EG of −442 (545), CG of −370 (471); for 4–16 months, EG of 123 (390), CG of 171 (399)	Correlation coefficients with 4–16-month change in (Internet use): weight, diet log^d, negatively; EEE, diet log^e; weekly weight log^e; PA log^e; participants contacted^d; nonparticipants contacted^d; articles posted^d
Measurement conditions		With light clothes and without shoes; using electronic scale (model 770; SECA, Hamburg, Germany)	The Seven-Day Physical Activity Recall interview	Nine randomly assigned days of diet records
Perri et al.¹⁵ (2008)	559298 (53)234 (79)EG1 70 (97)EG2 75 (90)CG 75 (94)	0–6 months: EG1, −9.4 (0.6); EG2, −10.1 (0.6); CG, −10.5 (0.6)0–18 months: EG1, −8.2^c; EG2, −8.9^c; CG, −6.8^c 6–18 months: EG1, +1.2 (0.7); EG2, +1.2 (0.6); CG, +3.7 (0.7)EG1^d, EG2^d <CG			Adherence: self-monitoring, EG1^d, EG2^d >CG; less adherence, more weight gain^e Cost: telephone, $2,125; FTF, $2,555; education control, $1,824
Measurement conditions		Using certified scale
Svetkey et al.³¹ (2008)	3,1781,685 (53)1,032 (61)EG1 323 (93)EG2 321 (94)CG 320 (94)	0–6 months: EG1, −8.6 (4.5); EG2, −8.3 (4.2); CG: −8.5 (4.0)0–36 months: EG1, −3.3 (0.4); EG2, −4.2 (0.4); CG, −2.9 (0.4)6–36 months: EG1, +5.2 (0.3); EG2, +4.0 (0.3); CG, +5.5 (0.3)Until 30 months: EG1 <CG^d, EG2 <CG^d At 30 and 36 months: EG1 <EG2^d, EG2 <CG^d	Minutes of MVPA/week: 0–6 months, all −48; 6–36 months, EG1 of 35, EG2 of 33, CG of 32	EI (kcal/day): for 0–6 months, all −325; for 6–36 months, EG1 of 16, EG2 of 55, CG of 88; for 0–36 months, EG1 of −326, EG2 of −272, CG of −231
Measurement conditions		Wearing light in-door clothes without shoes; using high-quality, calibrated digital scale; values measured on 2 separate days were averaged	Measured by accelerometry (calibrated, triaxial accelerometer [RT3; Stayhealthy Inc., Monrovia, CA])	Block Food Frequency Questionnaire
Wing et al.⁷ (2006)	648521 (80)314 (60)EG1 92 (88)EG2 101 (96)CG 98 (93)	0–18 months: EG1, +4.7 (8.6); EG2, +2.5 (6.7); CG, +4.9 (6.5)EG2^d <CG	PA: no group differences	Caloric intake: no group differences	Adherence: attendance at the FTF and chat room decreased over time^e but better in the FTF group than in the Internet group^d Self-weighing: daily self-weighing associated with a decreased risk of regaining^e
Measurement conditions		Wearing light street clothes; using calibrated scale	Paffenbarger Physical Activity Questionnaire	Block Food Frequency Questionnaire
Harvey-Berino et al.¹⁹ (2004)	255232 (91)232 EG1 52(68)EG2 61(79)EG3 63(81)	0–6 months: EG1, −8.4 (6.1); EG2, −7.6 (5.0); EG3, −7.6 (4.9)0–18 months: EG1, −7.6 (7.3); EG2, −5.1 (6.5); EG3, −5.5 (8.9)6–18 months: EG1, +0.8^f; EG2, +2.5^f; EG3, +2.1^f No group differences	All groups: significantly increased reported PA in time effect^e, no group effect (during the treatment)	All groups: significantly decreased reported EI in time effect^e, no group effect	Adherence: attendance, EG2 of 10 (5.1) versus EG1 of 7.7 (5.3)^d; self-monitoring, EG2 of 18.6 (13.2) versus EG1 of 11.6 (13.2)^e; peer support contact, EG2 of 4.9 (17.4) versus EG1 of 27.1 (58.2)^e
Measurement conditions		Wearing street clothes and without shoes, using beam-balance scale	Paffenbarger Physical Activity Questionnaire	Block Food Frequency Questionnaire
Harvey-Berino et al.²⁰ (2002)	Unclear122122 (100)EG1 30 (75)EG2 32 (78)EG3 28 (68)	0–6 months: EG1, −8.0 (5); EG2, −9.8 (5.9); EG3, −11.0 (6.5)0–18 months: EG1, −5.7 (5.9); EG2, −10.4 (6.3); EG3, −10.4 (9.3)6–18 months: EG1, +2.3^f; EG2, −0.6^f; EG3, +0.6^f EG1^d >EG2, EG3	All groups: significantly increased reported PA during treatment (EG2 higher than baseline^e)	All groups: significantly decreased reported EI during treatment (EG2 below than baseline^e)	Adherence (6–18 months)^d: attendance, EG2 of 54% versus EG1 of 39%; self-monitoring, EG2 of 22% versus EG1 of 19%; peer support, EG2 of 0.04 versus EG1 of 6.8^d
Measurement conditions		Wearing street clothes and without shoes Using beam-balance scale	Paffenbarger Physical Activity Questionnaire	Block Food Frequency Questionnaire
Harvey-Berino et al.²¹ (2002)	9451 (54)44 (86)EG1 (87)^f EG2 (100)^f CG (93)^f	0–15 weeks: all −6.4 (5.6), no group difference0–37 weeks: EG1, −7.0 (10.6)^c; EG2, −8.9 (7.8)^c; CG, −7.8 (6.9)^c 15–37 weeks: all −1.6, no group differences	15–37 weeks: increased calories expended in exercise by 419/week	15–37 weeks: decreased calorie and fat intake by 20 calories and 0.33 fat g/day	Attendance: EG2 of 58% versus EG1 of 33%^d
Measurement conditions		Wearing street clothes and without shoes; using beam-balance scale	Paffenbarger Physical Activity Questionnaire	Complete 3-day food records and analyzed using the Food Intake Analysis System version 2.1

Data are numbers for the recruitment, enrollment, allocation, and completion stages (by groups), respectively. The percentage of retention rate is calculated from the original participant numbers.

Weight is given as mean or mean (standard deviation) value.

Recalculated from the original data.

p<0.05, ^e p<0.001, for differences between groups.

Not reported by groups.

BMI, body mass index; CG, control group; EEE, exercise energy expenditure; EG, experimental group; EI, energy intake; FTF, face-to-face; MVPA, moderate to vigorous physical activity; PA, physical activity; WC, waist circumference.

Most of the studies used reliable measures to assess outcomes. The majority used a calibrated weight scale with the participant wearing light clothes and no shoes. To measure physical activity and dietary intake, various questionnaires were used. Of the eight studies, four used the Paffenbarger Physical Activity Questionnaire³² to measure physical activity, and three studies used the Block Food Frequency Questionnaire^33,34 to measure dietary intake. The majority of the studies stated their measurements clearly, using tools that had been tested for their validity and reliability through previous research studies. (For more specific details on this result, see Table 3.)

Quality of Study Design

The range of study design quality scores across the studies was from 55.6% to 100%. The average score was 80.9%, which is moderate. Table 4 shows the results of the study design quality assessment and characterized results of effectiveness in maintaining weight loss. All studies met the individual randomization, control group, and pre- and posttest design standards. Three studies reported that they conducted paired, blocked, or stratified randomization,^7,17,31 whereas other studies mentioned randomization briefly. Six studies had a non-technological comparison group, such as face-to-face counseling,^7,15 personal contact,^21,31 frequent in-person contact,^19,20 and minimal in-person contact.^19,20 All studies tested the effectiveness of designed technology-based interventions except the study of Thorndike et al.¹⁷; its maintenance intervention consisted of two components: Internet and personal contact. The average retention rate for all of the studies was 84.7%, but three studies' retention rates did not exceed 80%.^17,19,20 Three studies reported significant differences among groups at baseline on some variables,^9,15,17 and one study did not mention any group differences at baseline.²¹ A majority of the studies conducted analyses with consideration of missing data that maintained the fidelity of the randomization such as the baseline-observation-carried-forward method,^9,17 intent-to-treat analysis,^19,20 missing-at-random assumption,¹⁵ or imputation.³¹ All studies mentioned their measurement validity information, except the study by Thorndike et al.¹⁷; these investigators mentioned their measurement methods but did not fully describe their validity or make references about nutrition behavior and overall health.

Table 4.

Quality of Study Design Scores and Effectiveness of Maintaining Weight Loss

REFERENCE (YEAR)	INDIVIDUAL RANDOMIZATION	CONTROL GROUP	ISOLATE TECHNOLOGY	PRE-/POSTTEST DESIGN	RETENTION ≥80%	BASELINE GROUPS EQUIVALENT	MISSING DATA	SAMPLE SIZE CALCULATION	VALIDITY OF MEASURES	SCORE (% OF MAXIMUM)^a	EFFECTIVENESS OF MAINTAINING^b
Thorndike et al.¹⁷ (2012)	Y	Y	N	Y	N	N	Y	Y	UK	55.6%	—
Cussler et al.⁹ (2008)	Y	Y	Y	Y	Y	N	Y	Y	Y	88.9%	—
Perri et al.¹⁵ (2008)	Y	Y	Y	Y	Y	N	Y	Y	Y	88.9%	++/−
Svetkey et al.³¹ (2008)	Y	Y	Y	Y	Y	Y	Y	Y	Y	100%	++/−
Wing et al.⁷ (2006)	Y	Y	Y	Y	Y	Y	N	Y	Y	88.9%	++/−
Harvey-Berino et al.¹⁹ (2004)	Y	Y	Y	Y	N	Y	Y	Y	Y	88.9%	++/+
Harvey-Berino et al.²⁰ (2002)	Y	Y	Y	Y	N	Y	Y	N	Y	77.8%	—
Harvey-Berino et al.²¹ (2002)	Y	Y	Y	Y	Y	UK	N	N	Y	66.7%	−/−

The quality of study design scoring was adapted from that of Norman et al.⁸

Mean score of total percentage, 81.96%.

Direction of the effectiveness of maintain was coded as follows: “++”=technology-based intervention resulted in significant maintenance in weight loss compared with pure control group; “—”=technology-based intervention resulted in no significant maintenance in weight loss compared with pure control group; “+”=technology-based intervention resulted in significant maintenance in weight loss compared with non-technology comparison group; “−”=technology-based intervention resulted in no significant maintenance in weight loss compared with non-technology comparison group.

N, no; UK, unclear/unknown; Y, yes.

Effectiveness in Maintaining Weight Loss

There were various results in terms of effectiveness in maintaining weight loss. The researchers in three recent studies concluded that technology-based maintenance interventions did not improve outcomes¹⁷ or work better than self-direction in sustaining weight loss.^9,21 However, several studies reported that technology-based maintenance interventions allowed for sustained weight loss better than that of the control group^7,15,19,31 but not better than non–technology-based maintenance interventions.^7,15,31 Therefore, the results of the effectiveness of technology-based WLMIs were mixed in the primary research studies.

Discussion

In a search of the literature, eight studies were identified. Those selected for review were randomized controlled trials to evaluate the effectiveness of technology-based WLMIs in sustaining weight loss. Neve et al.¹¹ also conducted a systematic review with meta-analysis to evaluate the effectiveness of Web-based interventions, published in 2010. They concluded that Web-based interventions have the potential to achieve outcomes to change other lifestyle behaviors; however, they also questioned the effectiveness of Web-based interventions for weight loss and weight-loss maintenance.¹¹

Khaylis et al.⁴ also qualitatively reviewed the efficacy of technology-based WLIs and identified the five components necessary to succeed in enabling a successful weight-loss program: “self-monitoring, counselor feedback and communication, social support, use of a structured program, and use of an individually tailored program.” They concluded that long-term results are varied, although the short-term goals of technology-based intervention programs have been accomplished.

This review article supports previous review studies that report that long-term WLMI studies using technology have mixed findings. The article adds to the literature beyond previous reviews of long-term outcomes of WLI with its specific focus on the use/testing of technology-based interventions that were implemented during the maintenance phase (not just follow-up after a technology-based intervention). From this review, it is evident that there is a strong need to find an effective way to improve study interventions and protocols that will affect the weight-related outcome.

Suggested Guidelines for Technology-Based Maintenance Interventions

From the above review of eight studies, six guidelines have been derived that suggest ways to use technology-based WLMI methods effectively:

• Tailor the technology-based maintenance intervention for specific subjects.^17,31

• To improve retention rates, develop additional sessions for special concerns.¹⁵

• Encourage participants to utilize the technology.^9,17

• Consider having participants keep weekly self-weighing and diet logs.^7,9,15

• Check participants' technical abilities.²¹

• Consider an occasional personal contact component to complement the technology-based maintenance intervention.^20,31

Svetkey et al.³¹ elaborated that it is very important to provide more tailored interventions to achieve greater benefits because different individuals may prefer different types of interventions. They also reported that social support, reengagement, and motivation may need to be strengthened for long-term success. Thorndike et al.¹⁷ conducted a wellness program to modify employees' diet and exercise behaviors. The researchers focused on a specific population (employees) in a specific environment (a workplace). They reported that the employees became less motivated to participate in the less intensive 9-month maintenance intervention and that utilization of the Web site was relatively low in comparison with the 10-week intervention. These results showed that long-term maintenance interventions may need more specific intervention strategies focused on the workplace and on the employees. Therefore, more tailored interventions are needed on long-term aspects³¹ for specific subjects and environments.¹⁷

Generally, 20–25% attrition rates were observed in behavioral weight-loss programs.³⁴ The retention rates of the eight primary studies were in this range, but investigators could improve their retention rates by using additional sessions to satisfy participants. Perri et al.¹⁵ designed additional sessions for women in rural areas, such as low-calorie cooking techniques and performance for a healthy eating plan while away from home, to cope with the lack of family or friend support while losing weight. It is interesting that these researchers reported the highest retention rate among eight studies. Therefore, it is possible to conclude that the special additions make participants more likely to stay in the study.

Cussler et al.⁹ stated that there was considerable variation in weight changes during the 12-month weight maintenance period within the Internet group. The weight change range was 12.2–20.9 kg, so the authors assumed that the amount of weight change might reflect differences in the degree of Internet usage during the maintenance period. In addition, Harvey-Berino et al.¹⁹ tested participants' computer experience and comfort, attendance, and adherence. It was surprising that the attendance rate was better for the in-person group than for the Internet group. Moreover, the technology group's retention rate was relatively low in the majority of studies. These results indicate that investigators should simplify the interface, make it more convenient to use, and encourage participants to use technology more frequently.

Cussler et al.⁹ also reported that the diet log was correlated with a change in weight (p<0.05). The authors mentioned that the logs helped participants feel at ease and might have encouraged more mindful eating habits. Perri et al.¹⁵ also mentioned that completing the food logs mediated between extended-care counseling and long-term weight-loss success had a direct effect on the self-regulation theory (self-monitoring). Wing et al.⁷ also reported that daily self-weighing was associated with effective weight maintenance.

Harvey-Berino et al.^19
–21 tested the acceptability for participants of their group assignment. In early studies they found that some participants did not have a computer equipped with technology such as a sound card and speakers.²¹ Use of the Internet and mobile phones became more popular with middle-aged or elderly people. However, there can still be barriers to the effective use of these technologies. Investigators should check the participants' abilities and their devices to make sure participants can use the technology properly.

Svetkey et al.³¹ concluded that personal contact interventions were effective across all subgroups. Although the intervention contact was only a 10–15-min telephone conversation, it still increased the effectiveness of weight-loss maintenance. Researchers suggested that occasional personal contact may be needed to augment technology-based interventions. Other studies reported that face-to-face contact produced the best outcomes.⁷ Therefore, personal contact can be added to technology-based interventions to increase the likelihood of weight-loss maintenance.

Limitations

There are several limitations in this article. First, publication bias should be considered. Only electronic databases were searched. At RCT.org, many studies are listed as currently being implemented throughout the United States. Because of time-consuming data collection and publication processes, it takes significant time to prepare articles for publication. In addition, some studies are not published, including those with nonsignificant results or negative results. Second, only English-language journal articles were included. Issues related to being overweight and obese exist all over the world. More evidence might be published in languages other than English. Third, only a few studies were included. Previous studies included large ranges of intervention periods for weight-loss programs that were implemented. In this article, the focus was on long-term, technology-based WLMIs. It was difficult to find articles that satisfied both criteria (i.e., use of technology and application over the long term). In a few years, more studies will be published that will meet the inclusion requirements. Fourth, there was a lack of weight change data in the original studies. One study did not report body weight changes in each group, making it difficult to compare its results with those of other studies.²¹ In addition, some studies provided either a standard deviation or a confidence interval for weight change, but other studies provided only weight change. If there were more data, a meta-analysis could have been conducted. Fifth, there is considerable risk of bias in individual studies and across studies. To enhance the validity of study review, pairs of reviewers or peer reviewers can work independently. In this article, no assessment of risk of bias was performed.

Conclusions

The current obesity epidemic is problematic in the United States and worldwide as well. Overall, investigators of most of the studies agreed that technology could be a more effective tool in preventing weight gain than self-directed maintenance alone. However, some questions remain about the efficacy of technology-based WLMIs. In this review, results of the effectiveness of technology-based WLMIs were mixed. Therefore, this article has discussed potential strategies and approaches to improve the effectiveness of technology-based WLMIs.

This review provides important information to both patients and healthcare providers. Healthcare providers can suggest technology-based WLMIs to their patients and can include information about the strengths and weaknesses of each of the interventions. Moreover, suggested guidelines in this article can aid in planning further research design and implementation.

Further studies are needed to better evaluate the efficacy of technology-based WLMIs. Even longer maintenance intervention periods are needed to better evaluate the effectiveness of technology-based maintenance interventions. Moreover, this review included only eight studies; more primary studies are needed to evaluate reliable results. Meta-analysis also can be useful to examine actual body weight changes during the long-term period.

Footnotes

Acknowledgments

We would like to thank Dr. Diane Treat-Jacobson, Dr. Casey Hook, and Dr. Ulf Bronas for their generous support, feedback, and comments on this article.

Disclosure Statement

No competing financial interests exist.

Appendix 2.

Study Quality Coding Criteria

TABLE HEADING	SCORING CRITERIA
Individual randomization	Were participants randomized to study conditions? If so, was randomization at the individual level? Stratified and blocked randomization is acceptable. Studies that used individual randomization combined with a small proportion of randomized matched pairs are considered Y. Appropriately designed and powered group randomization would also be acceptable if the group was also the unit of analysis. Individual randomization is N when the authors fail to mention randomization, specify that another method of assigning group status was used, or randomize at the group level and analyze at the individual level.
Control group	Did the study include a comparison group? The comparison group could be a no treatment as usual or an alternate treatment group.
Isolate technology	Did the study design allow for test of effectiveness of the technology (e.g., Web-based delivery versus no treatment)? To isolate the technology, the authors had to test the technology alone and compare with a group with no technology (Y). Packaged interventions where the technological components can't be parsed out are coded as not isolating the technology (N).
Pre-/posttest design	Was assessment of behavior completed preintervention and postintervention?
Retention	Was study retention at least 80% of the subjects who initially agreed to participate the study? Retention is calculated for the entire sample and not by group. For studies that did not report retention or dropout rates, retention can be calculated using the sample size used for analyses (e.g., 300 were randomized, but only 250 were included in analyses=83.3% retention).
Baseline group equivalent	Were tests conducted to determine if groups were equivalent at baseline in important variables (e.g., gender, age, weight)? If no tests mentioned, then UK. If subset of test indicated any group differences at baseline, then N.
Missing data	Were analyses conducted with consideration of missing data that maintain the fidelity of the randomization (e.g., intent-to-treat, imputation)? Listwise case deletion (completer analysis)=N if this was the only analysis conducted. If 100% retention, then completer analysis is appropriate=Y. If authors compared the “dropped subgroup” with the selected or randomized sample, but did not consider the impact of the dropped subgroup on randomization (e.g., intent-to-treat or imputation), then code as N.
Sample size calculation	Was power analysis reported to determine study sample size?
Validity of measures	Did description of measures include reliability and validity information? If reference or coefficients, then Y. If a well-established measure that is known to be validated was used, then Y. For objective measures without validity evidence, if the objective measure is used as a proxy (e.g., food receipt for nutrition intake), then N. If the objective measure is used as a direct measure of behavior (e.g., food receipts for food purchase), then Y. If validity not reported and measure unknown, then UK.
Total	Percentage of Y values (%)

The study quality coding criteria were adapted from those of Norman et al.⁸

N, no; UK, unclear/unknown; Y, yes.

References

Ogden

, Carroll

, Kit

, Flegal

. Prevalence of childhood and adult obesity in the United States, 2011–2012. JAMA, 2014; 311:806–814.

World Health Organization. Obesity and overweight. Available at www.who.int/mediacentre/factsheets/fs311/en/ (last accessed February 26, 2014 ).

National Institute of Health, National Heart, Lung, and Blood Institute. Clinical guidelines on the identification, evaluation and treatment of overweight and obesity in adults; the evidence report. Obes Res, 1998; 6(Suppl 2):51S–209S.

Khaylis

, Yiaslas

, Bergstrom

, Gore-Felton

. A review of efficacious technology-based weight-loss interventions: Five key components. Telemed J E Health, 2010; 16:931–938.

Weinstein

. A review of weight loss programs delivered via the Internet. J Cardiovasc Nurs, 2006; 21:251–258.

Turk

, Yang

, Hravnak

, Sereika

, Ewing

, Burke

. Randomized clinical trials of weight loss maintenance: A review. J Cardiovasc Nurs, 2009; 24:58–80.

Wing

, Tate

, Gorin

, Raynor

, Fava

. A self-regulation program for maintenance of weight loss. N Engl J Med, 2006; 355:1563–1571.

Norman

, Zabinski

, Adams

, Rosenberg

, Yaroch

, Atienza

. A review of eHealth interventions for physical activity and dietary behavior change. Am J Prev Med, 2007; 33:336–345.

Cussler

, Teixeira

, Going

, et al. Maintenance of weight loss in overweight middle-aged women through the Internet. Obesity (Silver Spring), 2008; 16:1052–1060.

10.

Wantland

, Portillo

, Holzemer

, Slaughter

, McGhee

. The effectiveness of web-based vs. non web-based interventions: A meta-analysis of behavioral change outcomes. J Med Internet Res, 2004; 6:e40.

11.

Neve

, Morgan

, Jones

, Collins

. Effectiveness of web-based interventions in achieving weight loss and weight loss maintenance in overweight and obese adults: A systematic review with meta-analysis. Obes Rev, 2010; 11:306–321.

12.

LaPlante

, Peng

. A systematic review of e-Health Interventions for physical activity: An analysis of study design, intervention characteristics, and outcomes. Telemed J E-Health, 2011; 27:509–523.

13.

Cadmus-Bertram

, Wang

, Patterson

, Newman

, Parker

, Pierce

. Web-based self-monitoring for weight loss among overweight/obese women at increased risk for breast cancer: The HELP pilot study. Psychooncology, 2012; 22:1821–1828.

14.

Van Wier

, Dekkers

, Hendriksen

IJM

, et al. Effectiveness of phone and e-mail lifestyle counseling for long term weight control among overweight employees. J Occup Environ Med, 2011; 53:680–686.

15.

Perri

, Limacher

, Durning

, et al. Extended-care programs for weight management in rural communities: The Treatment of Obesity in Underserved Rural Settings (TOURS) randomized trial. Arch Intern Med, 2008; 168:2347–2354.

16.

Sherwood

, Jeffery

, Pronk

, et al. Mail and phone interventions for weight loss in a managed-care setting: Weight-to-be 2-year outcomes. Int J Obesity, 2006; 30:1565–1573.

17.

Thorndike

, Sonnenberg

, Healey

, Myint-U

, Kvedar

, Regan

. Prevention of weight gain following a worksite nutrition and exercise program: A randomized controlled trial. Am J Prev Med, 2012; 43:27–33.

18.

Morgan

, Lubans

, Collins

, Warren

, Callister

. 12-month outcomes and process evaluation of the SHED-IT RCT: An Internet-based weight loss program targeting men. Obesity (Silver Spring), 2011; 19:142–151.

19.

Harvey-Berino

, Pintauro

, Buzzell

, Gold

. Effect of Internet support on the long-term maintenance of weight loss. Obes Res, 2004; 12:320–329.

20.

Harvey-Berino

, Pintauro

, Bussell

, DiGiulio

, Gold

, Moldovan

, Ramirez

. Does using the Internet facilitate the maintenance of weight loss?. Int J Obes, 2002; 26:1254–1260.

21.

Harvey-Berino

, Pintauro

, Gold

. The feasibility of using Internet support for the maintenance of weight loss. Behav Modif, 2002; 26:103–116.

22.

Appel

, Clark

, Yeh

, et al. Comparative effectiveness of weight-loss interventions in clinical practice. N Engl J Med, 2011; 365:1959–1968.

23.

Van Geungten

, Van Empelen

, Boon

, Borsboom

, Visscher

, Oenema

. Results from an online computer-tailored weight management intervention for overweight adults: Randomized controlled trial. J Med Internet Res, 2012; 14:e44.

24.

Thomas

, Vydelingum

, Lawrence

. E-mail contact as an effective strategy in the maintenance of weight loss in adults. J Hum Nutr Diet, 2011; 24:32–38.

25.

Bennett

, Herring

, Puleo

, Stein

EK.

, Emmons

, Gillman

. Web-based weight loss in primary care: A randomized controlled trial. Obesity, 2010; 18:308–313.

26.

Womble

, Wadden

, McGuckin

, Sargent

, Rothman

, Krauthamer-Ewing

. A randomized controlled trail of a commercial Internet weight loss program. Obes Res, 2004; 12:1011–1018.

27.

Gold

, Burk

, Pintauro

, Buzzell

, Harvey-Berino

. Weight loss on the web: A pilot study comparing a structured behavioral intervention to a commercial program. Obesity (Silver Spring), 2007; 15:155–164.

28.

Haapala

, Barengo

, Biggs

, Surakka

, Manninen

. Weight loss by mobile phone: A 1-year effectiveness study. Public Health Nutr, 2009; 12:2382–2391.

29.

Patrick

, Raab

, Adams

, et al. A text message-based intervention for weight loss: Randomized controlled trial. J Med Internet Res, 2009; 11:e1.

30.

Rothert

, Strecher

, Doyle

, et al. Web-based weight management programs in an integrated health care setting: A randomized, controlled trial. Obesity (Silver Spring), 2006; 14:266–272.

31.

Svetkey

, Stevens

, Brantley

, et al. Comparison of strategies for sustaining weight loss: The weight loss maintenance randomized controlled trial. JAMA, 2008; 299:1139–1148.

32.

Paffenbarger

Jr , Wing

, Hyde

. Physical activity as an index of heart attack risk in college alumni. Am J Epidemiol, 1978; 108:161–175.

33.

Harlan

, Block

. Use of adjustment factors with a brief food frequency questionnaire to obtain nutrient values. Epidemiology, 1990; 1:224–231.

34.

Block

, Hartman

, Dresser

, Carroll

, Gannon

, Gardner

. A database approach to diet questionnaire design and testing. Am J Epidemiol, 1986; 124:453–469.