About

Gary G. Bennett is the Bishop-MacDermott Family Professor of Psychology & Neuroscience, Global Health, and Medicine at Duke University. He directs the Duke Global Digital Health Science Center and has a research program focused on designing, testing, and disseminating digital obesity treatments. Dr. Bennett is the founding director of the Duke Digital Health Science Center and has served as Past-President of the Society of Behavioral Medicine, the nation's largest organization of behavioral change scientists. His work emphasizes integrating digital health treatments into primary care settings, especially for medically vulnerable patients, and he developed the interactive obesity treatment approach (iOTA). His recent research demonstrates the effectiveness of digital strategies in treating obesity and improving cardiometabolic functioning. Dr. Bennett has authored over 170 scientific papers, with his research supported continuously by the National Institutes of Health. He is a fellow of the Association for Psychological Science and an elected member of the Academy of Behavioral Medicine Research and the Behavioral Medicine Research Council. Prior to Duke, he served on faculties at Harvard School of Public Health and the Dana-Farber Cancer Institute. At Duke, he is a member of the Bass Society of Fellows, the founding director of Duke's undergraduate major in global health, and has been recognized as a top instructor.

Research topics

• Internal medicine
• Medicine
• Physical therapy
• Gerontology
• Virology
• Nursing
• Psychiatry
• Emergency medicine
• Psychology
• Social psychology

Selected publications

• Identifying DASH Adherence Patterns and Their Association With Dietary Behaviors Among Adults With Hypertension

  Research in Nursing & Health · 2026-05-04

  articleOpen access

  The Dietary Approaches to Stop Hypertension (DASH) diet is effective in lowering blood pressure yet adherence to DASH remains low. Intuitive eating, a behavior that emphasizes responsiveness to hunger and satiety cues, may influence DASH adherence but has not been well studied among adults with hypertension. This study examined DASH adherence subgroups and their association with dietary behaviors using data from the Nourish U.S. based randomized controlled trial. A cross-sectional secondary analysis of baseline data from 301 participants was conducted. DASH adherence was assessed using the Mellen Index. Dietary behaviors were measured with the Intuitive Eating Scale-2 (IES-2). Latent class analysis (LCA) was used to identify DASH adherence subgroups, and subgroup differences in dietary behaviors and clinical outcomes were analyzed. Overall, DASH adherence was low (mean = 3.13 ± 1.35). LCA identified three DASH adherence subgroups: low adherence (44.2%), high adherence (41.2%), and mixed-pattern adherence (14.6%). Subgroups significantly differed in IES-2 subscales, including unconditional permission to eat (p < 0.0001), reliance on hunger and satiety cues (p = 0.0487), and body-food choice congruence (p < 0.0001). The low-adherence group scored higher on hunger cues, while the high-adherence group scored higher on body-food choice congruence. These findings highlight the value of examining overall dietary patterns rather than a single DASH score, as patterns better capture interactions among foods and nutrients that influence diet quality and health. Each subgroup's distinct demographic and behavioral characteristics further emphasize tailored DASH diet guidance which could lead to better cardiovascular outcomes. PATIENT OR PUBLIC CONTRIBUTION: 301 adults with hypertension contributed to this study by providing baseline data as part of the Nourish randomized controlled trial (ClinicalTrials.gov: NCT03875).

  PublisherDOI

• Optimizing Self-Monitoring in a Digital Weight Loss Intervention (Spark): Protocol for a Factorial Randomized Trial

  JMIR Research Protocols · 2025-09-23

  articleOpen access

  BACKGROUND: Self-monitoring is a vital component of behavioral obesity treatment. It often involves tracking dietary intake, physical activity, and body weight. However, the optimal combination of self-monitoring strategies that maximizes weight loss is unknown. To address this gap, we leverage a framework called the multiphase optimization strategy, which facilitates the identification of an intervention's "active ingredients" that promote weight loss and its "inactive ingredients" that have little impact, thus adding unnecessary patient effort and time demands. OBJECTIVE: This study aims to examine the unique and combined weight loss effects of 3 popular self-monitoring strategies (tracking dietary intake, steps, and body weight). METHODS: Spark was an optimization-randomized clinical trial that used a 2 × 2 × 2 full factorial design with 8 experimental conditions. Participants, US adults with overweight or obesity (N=176), were randomized to receive 0-3 self-monitoring strategies in a 6-month fully digital weight loss intervention. For each assigned strategy, participants were instructed to self-monitor daily via commercially available digital tools (a mobile app, wearable activity tracker, and smart scale) and received a corresponding goal (eg, a daily calorie goal) and weekly automated feedback. All participants received core intervention components, including weekly lessons and action plans informed by Social Cognitive Theory, to promote healthy eating and physical activity. Assessments occurred at baseline and at 1, 3, and 6 months. Weight was assessed objectively via a smart scale. The primary aim is to test the main effects of the 3 self-monitoring components and their interactions on weight change from baseline to 6 months. Secondary outcomes include change in BMI, caloric intake, diet quality, physical activity, and health-related quality of life, as well as 1- and 3-month weight change and the relation between self-monitoring engagement and weight change. Patterns of engagement will be operationalized as the percentage of days of self-monitoring during the 6-month intervention. Moderators of weight loss success will be explored to understand whether certain subgroups of individuals benefit more from specific self-monitoring strategies. We also conducted a separate embedded experiment to test the impact of a self-directed web-based orientation session on 6-month trial retention. After the intervention, semistructured qualitative interviews were conducted with a subset of participants to elucidate factors that impact engagement and its link to weight loss. RESULTS: Recruitment occurred from September 2023 to November 2024. Data collection was completed in June 2025. Data analysis is ongoing. CONCLUSIONS: This trial will provide evidence as to which self-monitoring strategies are the "active ingredients" in a fully digital weight loss intervention and begin to explore which subgroups may do best with which strategies. These results have potential for public health impact by maximizing weight loss while minimizing patient burden. TRIAL REGISTRATION: ClinicalTrials.gov NCT05249465, https://clinicaltrials.gov/study/NCT05249465. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/75629.

  PublisherDOI

• Recruitment of diverse community health center patients in a pragmatic weight gain prevention trial

  UNC Libraries · 2025-04-22

  articleOpen access1st authorCorresponding

  Introduction: Pragmatic trials are needed to establish evidence-based obesity treatment in primary care settings, particularly in community health centers (CHCs) that serve populations at heightened risk of obesity. Recruiting a representative trial sample is a critical first step to informing care for diverse communities. We described recruitment strategies utilized in a pragmatic obesity trial and assessed the sociodemographic characteristics and odds of enrollment by recruitment strategy. Methods: We analyzed data from Balance, a pragmatic trial implemented within a network of CHCs. We recruited participants via health center-based and electronic health record (EHR)-informed mail recruitment. We analyzed associations between sociodemographic characteristics and the return rate of patient authorization forms (required for participation) from EHR-informed mail recruitment. We also compared sociodemographic characteristics and randomization odds by recruitment strategy after returning authorization forms. Results: Of the individuals recruited through EHR-informed mail recruitment, females were more likely than males to return authorization forms; however, there were no differences in rates of return by preferred language (English/Spanish) or age. Females; underrepresented racial and ethnic groups; Spanish speakers; younger adults; and those with lower education levels were recruited more successfully in the health center. In contrast, their counterparts were more responsive to mail recruitment. Once authorization forms were returned, the odds of being randomized did not significantly differ by recruitment method. Conclusion: Health center-based recruitment was essential to meeting recruitment targets in a pragmatic weight gain prevention trial, specifically for Hispanic and Spanish-speaking communities. Future pragmatic trials should consider leveraging in-person recruitment for underrepresented groups in research.

  PublisherDOI

• Peer support: Current status and future opportunities for college mental health promotion

  Journal of American College Health · 2025-07-14

  article

  The past decade has seen a 60% increase in the number of programs providing peer mental health support. The availability, nature, and training requirements of programs varied widely, with implications for quality of support and stress on trainees. Greater collaboration may improve the impact of these programs.

  PublisherDOI

• Immunomodulatory therapy in children with paediatric inflammatory multisystem syndrome temporally associated with SARS-CoV-2 (PIMS-TS, MIS-C; RECOVERY): a randomised, controlled, open-label, platform trial

  The Lancet Child & Adolescent Health · 2024 · 34 citations

  • Medicine
  • Internal medicine
  • Virology

  BACKGROUND: Paediatric multisystem inflammatory syndrome temporally associated with SARS-CoV-2 (PIMS-TS), also known as multisystem inflammatory syndrome in children (MIS-C) emerged in April, 2020. The paediatric comparisons within the RECOVERY trial aimed to assess the effect of intravenous immunoglobulin or corticosteroids compared with usual care on duration of hospital stay for children with PIMS-TS and to compare tocilizumab (anti-IL-6 receptor monoclonal antibody) or anakinra (anti-IL-1 receptor antagonist) with usual care for those with inflammation refractory to initial treatment. METHODS: We did this randomised, controlled, open-label, platform trial in 51 hospitals in the UK. Eligible patients were younger than 18 years and had been admitted to hospital for PIMS-TS. In the first randomisation, patients were randomly assigned (1:1:1) to usual care (no additional treatments), usual care plus methylprednisolone (10mg/kg per day for 3 consecutive days), or usual care plus intravenous immunoglobulin (a single dose of 2 g/kg). If further anti-inflammatory treatment was considered necessary, children aged at least 1 year could be considered for a second randomisation, in which patients were randomly assigned (1:2:2) to usual care, intravenous tocilizumab (12 mg/kg in patients <30 kg; 8mg/kg in patients ≥30 kg, up to a maximum dose of 800 mg), or subcutaneous anakinra (2 mg/kg once per day in patients ≥10 kg). Randomisation was by use of a web-based simple (unstratified) randomisation with allocation concealment. The primary outcome was duration of hospital stay. Analysis was by intention to treat. For treatments assessed in each randomisation, a single Bayesian framework assuming uninformative priors for treatment was used to jointly assess the efficacy of each intervention compared with usual care. The trial was registered with ISRCTN (50189673) and ClinicalTrials.gov (NCT04381936). FINDINGS: Between May 18, 2020, and Jan 20, 2022, 237 children with PIMS-TS were enrolled and included in the intention-to-treat analysis. Of the 214 patients who entered the first randomisation, 73 were assigned to receive intravenous immunoglobulin, 61 methylprednisolone, and 80 usual care. Of the 70 children who entered the second randomisation (including 23 who did not enter the first randomisation), 28 were assigned to receive tocilizumab, 14 anakinra, and 28 usual care. Mean age was 9·5 years (SD 3·8) in the randomisation and 9·6 years (3·6) in the second randomisation. 118 (55%) of 214 patients in the first randomisation and 39 (56%) of 70 patients in the second randomisation were male. 130 (55%) of 237 patients were Black, Asian, or minority ethnic, and 105 (44%) were White. Mean duration of hospital stay was 7·4 days (SD 0·4) in children assigned to intravenous immunoglobulin and 7·6 days (0·4) in children assigned to usual care (difference -0·1 days, 95% credible interval [CrI] -1·3 to 1·0; posterior probability 59%). Mean duration of hospital stay was 6·9 days (SD 0·5) in children assigned to methylprednisolone (difference from usual care -0·7 days, 95% CrI -1·9 to 0·6; posterior probability 87%). Mean duration of hospital stay was 6·6 days (SD 0·7) in children assigned to second-line tocilizumab and 9·9 days (0·9) in children assigned to usual care (difference -3·3 days, 95% CrI -5·6 to -1·0; posterior probability >99%). Mean duration of hospital stay was 8·5 days (SD 1·2) in children assigned to anakinra (difference from usual care -1·4 days, 95% CrI -4·3 to 1·8; posterior probability 84%). Two persistent coronary artery aneurysms were reported among patients assigned to usual care in the first randomisation. There were few cardiac arrythmias, bleeding, or thrombotic events in any group. Two children died; neither was considered related to study treatment. INTERPRETATION: Moderate evidence suggests that, compared with usual care, first-line intravenous methylprednisolone reduces duration of hospital stay for children with PIMS-TS. Good evidence suggests that second-line tocilizumab reduces duration of hospital stay for children with inflammation refractory to initial treatment. Neither intravenous immunoglobulin nor anakinra had any effect on duration of hospital stay compared with usual care. FUNDING: Medical Research Council and National Institute of Health Research.

  DOI

• Effects of a Digital Intervention to Improve DASH and Blood Pressure Among US Adults

  Hypertension · 2024-12-23 · 8 citations

  articleOpen accessSenior author

  BACKGROUND: Dietary Approaches to Stop Hypertension (DASH) is a recommended first-line treatment for adults with hypertension, yet adherence to DASH is low. To evaluate the efficacy of a digital health intervention (DHI), compared with attention control, on changes in DASH adherence and blood pressure among adults with hypertension. METHODS: Nourish was a 12-month, parallel, 2-arm, randomized controlled trial of a virtually delivered DHI. Participants had a previous diagnosis of hypertension. The primary outcome was a 6-month change in DASH adherence. The secondary outcome was a change in blood pressure. We used linear mixed models to compare 6 and 12-month changes in DASH adherence, systolic blood pressure, and diastolic blood pressure. RESULTS: Nourish randomized 301 adults who averaged 54.4 (SD, 13.4) years and predominately identified as female (65%), White (53%), or Black (31%). Adjusted mean baseline DASH score was 2.30 (95% CI, 2.03–2.58). The adjusted mean baseline systolic blood pressure and diastolic blood pressure were 123.2 (95% CI, 119.5–126.9) and 77.1 (95% CI, 74.6–79.6) mm Hg. DASH score change was not significantly different between arms at 6 months ( M diff , 0.02 [95% CI, −0.37 to 0.40]). Yet, DHI participants had significantly greater 12-month changes in DASH score, relative to control ( M diff , 0.62 [95% CI, 0.16–1.08]). Between-group differences in 6-month changes were insignificant for systolic blood pressure and marginally significant for diastolic blood pressure, despite the DHI group showing significant blood pressure reductions from baseline. CONCLUSIONS: A DHI led to modest improvements in DASH and blood pressure among adults with hypertension but did not outperform the attention control. Further research is needed to understand the utility of DHIs to promote DASH and identify intervention components that support long-term behavior change.

  PublisherOA PDFDOI

• Weight Gain Prevention Outcomes From a Pragmatic Digital Health Intervention With Community Health Center Patients: Randomized Controlled Trial

  Journal of Medical Internet Research · 2024-02-26 · 6 citations

  articleOpen accessSenior author

  BACKGROUND: The prevalence of obesity and its associated comorbidities continue to rise in the United States. Populations who are uninsured and from racial and ethnic minority groups continue to be disproportionately affected. These populations also experience fewer clinically meaningful outcomes in most weight loss trials. Weight gain prevention presents a useful strategy for individuals who experience barriers to weight loss. Given the often-limited weight management resources available to patients in primary care settings serving vulnerable patients, evaluating interventions with pragmatic designs may help inform the design of comprehensive obesity care delivered in primary care. OBJECTIVE: This study aims to evaluate the effectiveness of Balance, a 2-arm, 12-month pragmatic randomized controlled trial of a digital weight gain prevention intervention, delivered to patients receiving primary care within federally qualified community health centers. METHODS: , spoke English or Spanish, and were receiving primary care within a network of federally qualified community health centers in North Carolina. The Balance intervention was designed to encourage behavioral changes that result in a slight energy deficit. Intervention participants received tailored goal setting and tracking, skills training, self-monitoring, and responsive health coaching from registered dietitians. Weight was measured at regular primary care visits and documented in the electronic health record. We compared the percentage of ≤3% weight gain in each arm at 24 months after randomization-our primary outcome-using individual empirical best linear unbiased predictors from the linear mixed-effects model. We used individual empirical best linear unbiased predictors from participants with at least 1 electronic health record weight documented within a 6-month window centered on the 24-month time point. RESULTS: . Most participants were Latino or Hispanic (n=200, 45.1%) or non-Latino or Hispanic White (n=115, 26%). In total, 53% (n=235) of participants had at least 1 visit with weight measured in the primary time window. The intervention group had a higher proportion with ≤3% weight gain at 6 months (risk ratio=1.12, 95% CI 0.94-1.28; risk difference=9.5, 95% CI -4.5 to 16.4 percentage points). This difference attenuated to the null by 24 months (risk ratio=1.00, 95% CI 0.82-1.20; risk difference=0.2, 95% CI -12.1 to 11.0 percentage points). CONCLUSIONS: In adults with overweight or obesity receiving primary care at a community health center, we did not find long-term evidence to support the dissemination of a digital health intervention for weight gain prevention. TRIAL REGISTRATION: ClinicalTrials.gov NCT03003403; https://clinicaltrials.gov/study/NCT03003403. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): RR2-10.1186/s12889-019-6926-7.

  PublisherDOI

• Weight Gain Prevention Outcomes From a Pragmatic Digital Health Intervention With Community Health Center Patients: Randomized Controlled Trial

  UNC Libraries · 2024-10-17

  articleOpen access

  BACKGROUND: The prevalence of obesity and its associated comorbidities continue to rise in the United States. Populations who are uninsured and from racial and ethnic minority groups continue to be disproportionately affected. These populations also experience fewer clinically meaningful outcomes in most weight loss trials. Weight gain prevention presents a useful strategy for individuals who experience barriers to weight loss. Given the often-limited weight management resources available to patients in primary care settings serving vulnerable patients, evaluating interventions with pragmatic designs may help inform the design of comprehensive obesity care delivered in primary care. OBJECTIVE: This study aims to evaluate the effectiveness of Balance, a 2-arm, 12-month pragmatic randomized controlled trial of a digital weight gain prevention intervention, delivered to patients receiving primary care within federally qualified community health centers. METHODS: Balance was a 2-arm, 12-month pragmatic randomized controlled trial of a digital weight gain prevention intervention delivered to individuals who had a BMI of 25-40 kg/m², spoke English or Spanish, and were receiving primary care within a network of federally qualified community health centers in North Carolina. The Balance intervention was designed to encourage behavioral changes that result in a slight energy deficit. Intervention participants received tailored goal setting and tracking, skills training, self-monitoring, and responsive health coaching from registered dietitians. Weight was measured at regular primary care visits and documented in the electronic health record. We compared the percentage of ≤3% weight gain in each arm at 24 months after randomization-our primary outcome-using individual empirical best linear unbiased predictors from the linear mixed-effects model. We used individual empirical best linear unbiased predictors from participants with at least 1 electronic health record weight documented within a 6-month window centered on the 24-month time point. RESULTS: We randomized 443 participants, of which 223 (50.3%) participants were allocated to the intervention arm. At baseline, participants had a mean BMI of 32.6 kg/m². Most participants were Latino or Hispanic (n=200, 45.1%) or non-Latino or Hispanic White (n=115, 26%). In total, 53% (n=235) of participants had at least 1 visit with weight measured in the primary time window. The intervention group had a higher proportion with ≤3% weight gain at 6 months (risk ratio=1.12, 95% CI 0.94-1.28; risk difference=9.5, 95% CI -4.5 to 16.4 percentage points). This difference attenuated to the null by 24 months (risk ratio=1.00, 95% CI 0.82-1.20; risk difference=0.2, 95% CI -12.1 to 11.0 percentage points). CONCLUSIONS: In adults with overweight or obesity receiving primary care at a community health center, we did not find long-term evidence to support the dissemination of a digital health intervention for weight gain prevention. TRIAL REGISTRATION: ClinicalTrials.gov NCT03003403; https://clinicaltrials.gov/study/NCT03003403. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): RR2-10.1186/s12889-019-6926-7.

  PublisherDOI

• Design and pilot test of an implicit bias mitigation curriculum for clinicians

  Frontiers in Medicine · 2024-06-06 · 3 citations

  articleOpen access

  Introduction: Clinician implicit racial bias (IB) may lead to lower quality care and adverse health outcomes for Black patients. Educational efforts to train clinicians to mitigate IB vary widely and have insufficient evidence of impact. We developed and pilot-tested an evidence-based clinician IB curriculum, "REACHing Equity." Methods: To assess acceptability and feasibility, we conducted an uncontrolled one-arm pilot trial with post-intervention assessments. REACHing Equity is designed for clinicians to: (1) acquire knowledge about IB and its impact on healthcare, (2) increase awareness of one's own capacity for IB, and (3) develop skills to mitigate IB in the clinical encounter. We delivered REACHing Equity virtually in three facilitated, interactive sessions over 7-9 weeks. Participants were health care providers who completed baseline and end-of-study evaluation surveys. Results: Of approximately 1,592 clinicians invited, 37 participated, of whom 29 self-identified as women and 24 as non-Hispanic White. Attendance averaged 90% per session; 78% attended all 3 sessions. Response rate for evaluation surveys was 67%. Most respondents agreed or strongly agreed that the curriculum objectives were met, and that REACHing Equity equipped them to mitigate the impact of implicit bias in clinical care. Participants consistently reported higher self-efficacy for mitigating IB after compared to before completing the curriculum. Conclusions: Despite apparent barriers to clinician participation, we demonstrated feasibility and acceptability of the REACHing Equity intervention. Further research is needed to develop objective measures of uptake and clinician skill, test the impact of REACHing Equity on clinically relevant outcomes, and refine the curriculum for uptake and dissemination.ClinicalTrials.gov ID: NCT03415308.

  PublisherOA PDFDOI

• Correction: Weight Gain Prevention Outcomes From a Pragmatic Digital Health Intervention With Community Health Center Patients: Randomized Controlled Trial

  Journal of Medical Internet Research · 2024-05-13 · 1 citations

  erratumOpen accessSenior author

  [This corrects the article DOI: 10.2196/50330.].

  PublisherOA PDFDOI

Recent grants

• NIH Grant K22CA126992

  NIH · $529k · 2011

• A Pragmatic Trial of a Digital Health Intervention to Prevent Weight Gain in Primary Care

  NIH · $3.3M · 2016–2023

• New Media Obesity Treatment in Community Health Centers

  NIH · $3.2M · 2012–2018

• NIH Grant R01DK078798

  NIH · $3.0M · 2014

• Optimizing a standalone text messaging-based weight loss intervention

  NIH · $2.0M · 2017–2023

Frequent coauthors

• Sandy Askew

  Duke University

  106 shared

• Dori Steinberg

  State Street (United States)

  104 shared

• Laura P. Svetkey

  83 shared

• Karina W. Davidson

  Feinstein Institute for Medical Research

  67 shared

• Simon Bacon

  Concordia University

  58 shared

• John Ruiz

  University of Arizona

  58 shared

• Tracey A. Revenson

  The Graduate Center, CUNY

  58 shared

• Elizabeth Brondolo

  58 shared

Awards & honors

• Fellow of the Association for Psychological Science
• Elected member of the Academy of Behavioral Medicine Researc…
• Elected member of the Behavioral Medicine Research Council
• Member of Duke's Bass Society of Fellows