Mark A. Berman
Research Subjects
Mark A. Berman has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Kevin Appelbaum
Better Therapeutics, San Francisco, CA, USA
7 shared publications - Nicole L. Guthrie
Better Therapeutics, Inc
6 shared publications - David L. Katz
True Health Initiative
4 shared publications - David Miles Eisenberg
Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, MA 02115, USA
4 shared publications - Katherine Edwards
Better Therapeutics, San Francisco, California
4 shared publications - Jason Carpenter
Manifold, Inc, Oakland, CA, United States
2 shared publications
Latest Publications
- Open Access
- 5 August 2025
- Published by Wolters Kluwer Health in Journal of the American Heart Association
- Vol. 14 (15) , e038737
Digital Therapeutic Based Randomized Investigation to Improve Glycemic Control in Patients With Type 2 Diabetes and Residual Hyperglycemia on Stable Medical Therapy: Rationale, Design, and Baseline Characteristics of the BRIGHT Trial
Background: The prevalence of type 2 diabetes continues to rise in the United States and worldwide. Although novel pharmacotherapies have shown benefits, not all patients have access to these medications and of those who do, some are unable to tolerate treatment. Behavior change is long known to be foundational for patients with cardiometabolic disease due to its direct positive impact on health outcomes. Cognitive behavioral therapy delivered by a digital therapeutic has the potential to address access to behavioral therapy for patients with type 2 diabetes. Methods: The BRIGHT (Digital Therapeutic Based Investigation to Improve Glycemic Control in Patients With Type 2 Diabetes and Residual Hyperglycemia on Stable Medical Therapy) Trial (NCT05266625) is a randomized, controlled, multicenter, pragmatic, pivotal trial with open‐label extension evaluating the hypothesis that AspyreRx improves glycemic control and cardiometabolic measures in patients with type 2 diabetes compared with a control application added to usual care across a broad range of patients in a real‐world setting. The current study objectives are to evaluate the efficacy and safety of AspyreRx at 90 and 180 days compared with a control group and to assess the long‐term effectiveness, safety, and impact on health care and medication use over long‐term follow up. Conclusions: The AspyreRx digital therapeutic is a Food and Drug Administration approved therapy for patients with type 2 diabetes. Although its efficacy for lowering hemoglobin A1C at 90 days was demonstrated, opportunities remain to better understand the long‐term effects and the effectiveness and safety in a broader and more diverse real‐world environment. BRIGHT is designed to evaluate these questions using a virtual, decentralized, model and include a diverse site mix designed to better understand real‐world effectiveness and generalizability.
- 5 August 2025
- Published by Wolters Kluwer Health in Journal of the American Heart Association
- Vol. 14 (15) , e038737
Open AccessDigital Therapeutic Based Randomized Investigation to Improve Glycemic Control in Patients With Type 2 Diabetes and Residual Hyperglycemia on Stable Medical Therapy: Rationale, Design, and Baseline Characteristics of the BRIGHT Trial
Background: The prevalence of type 2 diabetes continues to rise in the United States and worldwide. Although novel pharmacotherapies have shown benefits, not all patients have access to these medications and of those who do, some are unable to tolerate treatment. Behavior change is long known to be foundational for patients with cardiometabolic disease due to its direct positive impact on health outcomes. Cognitive behavioral therapy delivered by a digital therapeutic has the potential to address access to behavioral therapy for patients with type 2 diabetes. Methods: The BRIGHT (Digital Therapeutic Based Investigation to Improve Glycemic Control in Patients With Type 2 Diabetes and Residual Hyperglycemia on Stable Medical Therapy) Trial (NCT05266625) is a randomized, controlled, multicenter, pragmatic, pivotal trial with open‐label extension evaluating the hypothesis that AspyreRx improves glycemic control and cardiometabolic measures in patients with type 2 diabetes compared with a control application added to usual care across a broad range of patients in a real‐world setting. The current study objectives are to evaluate the efficacy and safety of AspyreRx at 90 and 180 days compared with a control group and to assess the long‐term effectiveness, safety, and impact on health care and medication use over long‐term follow up. Conclusions: The AspyreRx digital therapeutic is a Food and Drug Administration approved therapy for patients with type 2 diabetes. Although its efficacy for lowering hemoglobin A1C at 90 days was demonstrated, opportunities remain to better understand the long‐term effects and the effectiveness and safety in a broader and more diverse real‐world environment. BRIGHT is designed to evaluate these questions using a virtual, decentralized, model and include a diverse site mix designed to better understand real‐world effectiveness and generalizability.
- Open Access
- 1 April 2020
- Published by The Endocrine Society in Journal of the Endocrine Society
- Vol. 4 (Supplement)
SAT-LB120 A Software Application Delivering Behavioral Therapy Improved Glycemic Control in Adults With Type 2 Diabetes
Despite advances in pharmacological treatment, half of U.S. patients with type 2 diabetes are not achieving glycemic control (1). Even when glycemic control is achieved via pharmacotherapy, a substantial residual risk to all-cause mortality still exists (2). An overlooked contributor to both poor glycemic control and mortality risk is the need to address the behavioral determinants of type 2 diabetes. We tested a novel software application (app) designed to change behaviors that are root causes of type 2 diabetes and improve glycemic control. If behavioral therapy is effective and can be delivered using software installed on a patient’s smartphone, it may provide a cost-effective means of improving outcomes in this patient population. To study this hypothesis, we examined data from app users with suboptimal glycemic control, as defined by having a 3-day average of self-reported fasting blood glucose (FBG) values ≥ 152 mg/dL, who used the app for 12 weeks. 74 participants residing in 32 U.S. states met inclusion criteria. Their baseline characteristics (means±SD) were an age of 55.2±7.0 years, body mass index 34.7±7.9 kg/m2, 71.6% (53/74) female, current use of 2.2±0.9 antihyperglycemic medications, FBG 199.4±45.1 mg/dL and 10.6±7.1 years since diagnosis of type 2 diabetes. Change in FBG was calculated as the difference between 3-day averages anchored by the first and last values reported. Mean change in FBG was -22.9±42.0 mg/dL (P=0.001, using paired t-test) with an average of 69.0±29.9 days between the first and last value reported. Participants reported an average of 4.4±1.7 FBG values each week over 12-weeks. Participants used the app on average 4.6±1.8 days each week and spent 5.6±8.3 minutes each use. We performed a tertile analysis to examine the relationship between degree of behavioral therapy use and change in glycemic status. The mean changes in FBG in the lowest to highest tertile of behavioral therapy use were -14.4±39.2, -20.4±47.6 and -32.9±38.4 mg/dL (P=0.06, comparing lowest to highest tertile, using a pairwise comparison of the least squares means accounting for baseline FBG and years since diagnosis). These results suggest use of this novel software application is associated with meaningful improvements in glycemic control in a widely distributed treatment population and warrants further study. Refs: (1) Edelman et al., Diabetes Care 2017 Nov; 40(11): 1425-1432. (2) Rodríguez-Gutiérrez et al., Circ Cardiovasc Qual Outcomes. 2016; 9:504-512.
- 1 April 2020
- Published by The Endocrine Society in Journal of the Endocrine Society
- Vol. 4 (Supplement)
Open AccessSAT-LB120 A Software Application Delivering Behavioral Therapy Improved Glycemic Control in Adults With Type 2 Diabetes
Despite advances in pharmacological treatment, half of U.S. patients with type 2 diabetes are not achieving glycemic control (1). Even when glycemic control is achieved via pharmacotherapy, a substantial residual risk to all-cause mortality still exists (2). An overlooked contributor to both poor glycemic control and mortality risk is the need to address the behavioral determinants of type 2 diabetes. We tested a novel software application (app) designed to change behaviors that are root causes of type 2 diabetes and improve glycemic control. If behavioral therapy is effective and can be delivered using software installed on a patient’s smartphone, it may provide a cost-effective means of improving outcomes in this patient population. To study this hypothesis, we examined data from app users with suboptimal glycemic control, as defined by having a 3-day average of self-reported fasting blood glucose (FBG) values ≥ 152 mg/dL, who used the app for 12 weeks. 74 participants residing in 32 U.S. states met inclusion criteria. Their baseline characteristics (means±SD) were an age of 55.2±7.0 years, body mass index 34.7±7.9 kg/m2, 71.6% (53/74) female, current use of 2.2±0.9 antihyperglycemic medications, FBG 199.4±45.1 mg/dL and 10.6±7.1 years since diagnosis of type 2 diabetes. Change in FBG was calculated as the difference between 3-day averages anchored by the first and last values reported. Mean change in FBG was -22.9±42.0 mg/dL (P=0.001, using paired t-test) with an average of 69.0±29.9 days between the first and last value reported. Participants reported an average of 4.4±1.7 FBG values each week over 12-weeks. Participants used the app on average 4.6±1.8 days each week and spent 5.6±8.3 minutes each use. We performed a tertile analysis to examine the relationship between degree of behavioral therapy use and change in glycemic status. The mean changes in FBG in the lowest to highest tertile of behavioral therapy use were -14.4±39.2, -20.4±47.6 and -32.9±38.4 mg/dL (P=0.06, comparing lowest to highest tertile, using a pairwise comparison of the least squares means accounting for baseline FBG and years since diagnosis). These results suggest use of this novel software application is associated with meaningful improvements in glycemic control in a widely distributed treatment population and warrants further study. Refs: (1) Edelman et al., Diabetes Care 2017 Nov; 40(11): 1425-1432. (2) Rodríguez-Gutiérrez et al., Circ Cardiovasc Qual Outcomes. 2016; 9:504-512.
- Open Access
- 9 October 2019
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 21 (10) , e15814
Estimating the Impact of Novel Digital Therapeutics in Type 2 Diabetes and Hypertension: Health Economic Analysis
Background: Behavioral interventions can meaningfully improve cardiometabolic conditions. Digital therapeutics (DTxs) delivering these interventions may provide benefits comparable to pharmacologic therapies, displacing medications for some patients. Objective: Our objective was to estimate the economic impact of a digital behavioral intervention in type 2 diabetes mellitus (T2DM) and hypertension (HTN) and estimate the impact of clinical inertia on deprescribing medications. Methods: Decision analytic models estimated health resource savings and cost effectiveness from a US commercial payer perspective. A 3-year time horizon was most relevant to the intervention and payer. Effectiveness of the DTx in improving clinical outcomes was based on cohort studies and published literature. Health resource utilization (HRU), health state utilities, and costs were drawn from the literature with costs adjusted to 2018 dollars. Future costs and quality-adjusted life years (QALYs) were discounted at 3%. Sensitivity analyses assessed uncertainty. Results: Average HRU savings ranged from $97 to $145 per patient per month, with higher potential benefits in T2DM. Cost-effectiveness acceptability analyses using a willingness-to-pay of $50,000/QALY indicated that the intervention would be cost effective at total 3-year program costs of $6468 and $6620 for T2DM and HTN, respectively. Sensitivity analyses showed that reduced medication costs are a primary driver of potential HRU savings, and the results were robust within values tested. A resistance to deprescribe medications when a patient’s clinical outcomes improve can substantially reduce the estimated economic benefits. Our models rely on estimates of clinical effectiveness drawn from limited cohort studies with DTxs and cannot account for other disease management programs that may be implemented. Performance of DTxs in real-world settings is required to further validate their economic benefits. Conclusions: The DTxs studied may provide substantial cost savings, in part by reducing the use of conventional medications. Clinical inertia may limit the full cost savings of DTxs.
- 9 October 2019
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 21 (10) , e15814
Open AccessEstimating the Impact of Novel Digital Therapeutics in Type 2 Diabetes and Hypertension: Health Economic Analysis
Background: Behavioral interventions can meaningfully improve cardiometabolic conditions. Digital therapeutics (DTxs) delivering these interventions may provide benefits comparable to pharmacologic therapies, displacing medications for some patients. Objective: Our objective was to estimate the economic impact of a digital behavioral intervention in type 2 diabetes mellitus (T2DM) and hypertension (HTN) and estimate the impact of clinical inertia on deprescribing medications. Methods: Decision analytic models estimated health resource savings and cost effectiveness from a US commercial payer perspective. A 3-year time horizon was most relevant to the intervention and payer. Effectiveness of the DTx in improving clinical outcomes was based on cohort studies and published literature. Health resource utilization (HRU), health state utilities, and costs were drawn from the literature with costs adjusted to 2018 dollars. Future costs and quality-adjusted life years (QALYs) were discounted at 3%. Sensitivity analyses assessed uncertainty. Results: Average HRU savings ranged from $97 to $145 per patient per month, with higher potential benefits in T2DM. Cost-effectiveness acceptability analyses using a willingness-to-pay of $50,000/QALY indicated that the intervention would be cost effective at total 3-year program costs of $6468 and $6620 for T2DM and HTN, respectively. Sensitivity analyses showed that reduced medication costs are a primary driver of potential HRU savings, and the results were robust within values tested. A resistance to deprescribe medications when a patient’s clinical outcomes improve can substantially reduce the estimated economic benefits. Our models rely on estimates of clinical effectiveness drawn from limited cohort studies with DTxs and cannot account for other disease management programs that may be implemented. Performance of DTxs in real-world settings is required to further validate their economic benefits. Conclusions: The DTxs studied may provide substantial cost savings, in part by reducing the use of conventional medications. Clinical inertia may limit the full cost savings of DTxs.
- Open Access
- 8 August 2019
- Published by JMIR Publications Inc. in JMIR Preprints
Estimating the Impact of Novel Digital Therapeutics in Type 2 Diabetes and Hypertension: Health Economic Analysis (Preprint)
BACKGROUND Behavioral interventions can meaningfully improve cardiometabolic conditions. Digital therapeutics (DTxs) delivering these interventions may provide benefits comparable to pharmacologic therapies, displacing medications for some patients. OBJECTIVE Our objective was to estimate the economic impact of a digital behavioral intervention in type 2 diabetes mellitus (T2DM) and hypertension (HTN) and estimate the impact of clinical inertia on deprescribing medications. METHODS Decision analytic models estimated health resource savings and cost effectiveness from a US commercial payer perspective. A 3-year time horizon was most relevant to the intervention and payer. Effectiveness of the DTx in improving clinical outcomes was based on cohort studies and published literature. Health resource utilization (HRU), health state utilities, and costs were drawn from the literature with costs adjusted to 2018 dollars. Future costs and quality-adjusted life years (QALYs) were discounted at 3%. Sensitivity analyses assessed uncertainty. RESULTS Average HRU savings ranged from $97 to $145 per patient per month, with higher potential benefits in T2DM. Cost-effectiveness acceptability analyses using a willingness-to-pay of $50,000/QALY indicated that the intervention would be cost effective at total 3-year program costs of $6468 and $6620 for T2DM and HTN, respectively. Sensitivity analyses showed that reduced medication costs are a primary driver of potential HRU savings, and the results were robust within values tested. A resistance to deprescribe medications when a patient’s clinical outcomes improve can substantially reduce the estimated economic benefits. Our models rely on estimates of clinical effectiveness drawn from limited cohort studies with DTxs and cannot account for other disease management programs that may be implemented. Performance of DTxs in real-world settings is required to further validate their economic benefits. CONCLUSIONS The DTxs studied may provide substantial cost savings, in part by reducing the use of conventional medications. Clinical inertia may limit the full cost savings of DTxs.
- 8 August 2019
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessEstimating the Impact of Novel Digital Therapeutics in Type 2 Diabetes and Hypertension: Health Economic Analysis (Preprint)
BACKGROUND Behavioral interventions can meaningfully improve cardiometabolic conditions. Digital therapeutics (DTxs) delivering these interventions may provide benefits comparable to pharmacologic therapies, displacing medications for some patients. OBJECTIVE Our objective was to estimate the economic impact of a digital behavioral intervention in type 2 diabetes mellitus (T2DM) and hypertension (HTN) and estimate the impact of clinical inertia on deprescribing medications. METHODS Decision analytic models estimated health resource savings and cost effectiveness from a US commercial payer perspective. A 3-year time horizon was most relevant to the intervention and payer. Effectiveness of the DTx in improving clinical outcomes was based on cohort studies and published literature. Health resource utilization (HRU), health state utilities, and costs were drawn from the literature with costs adjusted to 2018 dollars. Future costs and quality-adjusted life years (QALYs) were discounted at 3%. Sensitivity analyses assessed uncertainty. RESULTS Average HRU savings ranged from $97 to $145 per patient per month, with higher potential benefits in T2DM. Cost-effectiveness acceptability analyses using a willingness-to-pay of $50,000/QALY indicated that the intervention would be cost effective at total 3-year program costs of $6468 and $6620 for T2DM and HTN, respectively. Sensitivity analyses showed that reduced medication costs are a primary driver of potential HRU savings, and the results were robust within values tested. A resistance to deprescribe medications when a patient’s clinical outcomes improve can substantially reduce the estimated economic benefits. Our models rely on estimates of clinical effectiveness drawn from limited cohort studies with DTxs and cannot account for other disease management programs that may be implemented. Performance of DTxs in real-world settings is required to further validate their economic benefits. CONCLUSIONS The DTxs studied may provide substantial cost savings, in part by reducing the use of conventional medications. Clinical inertia may limit the full cost savings of DTxs.
- Open Access
Emergence of digital biomarkers to predict and modify treatment efficacy: machine learning study
Objectives Development of digital biomarkers to predict treatment response to a digital behavioural intervention. Design Machine learning using random forest classifiers on data generated through the use of a digital therapeutic which delivers behavioural therapy to treat cardiometabolic disease. Data from 13 explanatory variables (biometric and engagement in nature) generated in the first 28 days of a 12-week intervention were used to train models. Two levels of response to treatment were predicted: (1) systolic change ≥10 mm Hg (SC model), and (2) shift down to a blood pressure category of elevated or better (ER model). Models were validated using leave-one-out cross validation and evaluated using area under the curve receiver operating characteristics (AUROC) and specificity- sensitivity. Ability to predict treatment response with a subset of nine variables, including app use and baseline blood pressure, was also tested (models SC-APP and ER-APP). Setting Data generated through ad libitum use of a digital therapeutic in the USA. Participants Deidentified data from 135 adults with a starting blood pressure ≥130/80, who tracked blood pressure for at least 7 weeks using the digital therapeutic. Results The SC model had an AUROC of 0.82 and a sensitivity of 58% at a specificity of 90%. The ER model had an AUROC of 0.69 and a sensitivity of 32% at a specificity at 91%. Dropping explanatory variables related to blood pressure resulted in an AUROC of 0.72 with a sensitivity of 42% at a specificity of 90% for the SC-APP model and an AUROC of 0.53 for the ER-APP model. Conclusions Machine learning was used to transform data from a digital therapeutic into digital biomarkers that predicted treatment response in individual participants. Digital biomarkers have potential to improve treatment outcomes in a digital behavioural intervention.
Open AccessEmergence of digital biomarkers to predict and modify treatment efficacy: machine learning study
Objectives Development of digital biomarkers to predict treatment response to a digital behavioural intervention. Design Machine learning using random forest classifiers on data generated through the use of a digital therapeutic which delivers behavioural therapy to treat cardiometabolic disease. Data from 13 explanatory variables (biometric and engagement in nature) generated in the first 28 days of a 12-week intervention were used to train models. Two levels of response to treatment were predicted: (1) systolic change ≥10 mm Hg (SC model), and (2) shift down to a blood pressure category of elevated or better (ER model). Models were validated using leave-one-out cross validation and evaluated using area under the curve receiver operating characteristics (AUROC) and specificity- sensitivity. Ability to predict treatment response with a subset of nine variables, including app use and baseline blood pressure, was also tested (models SC-APP and ER-APP). Setting Data generated through ad libitum use of a digital therapeutic in the USA. Participants Deidentified data from 135 adults with a starting blood pressure ≥130/80, who tracked blood pressure for at least 7 weeks using the digital therapeutic. Results The SC model had an AUROC of 0.82 and a sensitivity of 58% at a specificity of 90%. The ER model had an AUROC of 0.69 and a sensitivity of 32% at a specificity at 91%. Dropping explanatory variables related to blood pressure resulted in an AUROC of 0.72 with a sensitivity of 42% at a specificity of 90% for the SC-APP model and an AUROC of 0.53 for the ER-APP model. Conclusions Machine learning was used to transform data from a digital therapeutic into digital biomarkers that predicted treatment response in individual participants. Digital biomarkers have potential to improve treatment outcomes in a digital behavioural intervention.
- Open Access
- 12 March 2019
- Published by JMIR Publications Inc. in JMIR Cardio
- Vol. 3 (1) , e13030
Achieving Rapid Blood Pressure Control With Digital Therapeutics: Retrospective Cohort and Machine Learning Study
Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet #Preprint #PeerReviewMe: Warning: This is a unreviewed preprint. Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn. Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period. Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author). Background: Behavioral therapies, such as e-counseling and self-monitoring dispensed through mobile apps, have been shown to improve blood pressure but results vary and long-term engagement is a challenge. Machine learning is a rapidly advancing discipline that can be used to generate predictive and responsive models for the management and treatment of chronic conditions and shows potential for meaningfully improving outcomes. Objective: This retrospective analysis of data from the digital therapeutic, Better, examines its effect on blood pressure in hypertensive adults and explores the use of machine learning methods to predict intervention completion. Methods: The Better database was queried for participants with hypertension, who engaged with the intervention for at least 2 weeks and had paired blood pressure (BP) values. Participants were required to be ≥ 18 years old, reside in the United States and own a smartphone. The digital intervention offers personalized behavior therapy, including goal setting, skill building and self-monitoring. Participants reported BP at will. Changes in BP were calculated using averages of baseline and ending BP for each participant. Machine learning was used to generate a model of participants who would complete the intervention. Random forest models were trained at days 1, 3 and 7 of the intervention and the generalizability of the models was assessed using leave-one-out cross validation. Results: The primary analysis cohort was comprised of 172 hypertensive participants with paired BP values who were engaged with the intervention. Participants were 86.1% female, with a mean age of 55.0 years (95% CI 53.7, 56.2), baseline systolic blood pressure (SBP) of 138.9 mmHg (95% CI 136.6, 141.3) and diastolic (DBP) of 86.2 mmHg (95% CI 84.8, 87.7). Mean BP change was -11.5 mmHg for SBP and -5.9 mmHg for DBP over a mean of 62.6 days (P < .001). Amongst participants with Stage 2 Hypertension, mean BP change was -17.6 mmHg for SBP and -8.8 mmHg for DBP. Changes in BP remained significant in a mixed effects model accounting for baseline SBP, age, gender and BMI (P < .001). 43% of participants tracking BP at 12 weeks achieved the 2017 ACC/AHA definition of blood pressure control. The 7-day predictive model for intervention completion was trained on 429 participants and the AUC of the receiver operating characteristic was .72 with a false positive rate of 25%. Conclusions: The Better digital therapeutic was effective in lowering blood pressure among hypertensive adults. The degree of BP reduction was clinically meaningful and achieved rapidly by a majority of studied participants. Greater improvement was observed in participants with more severe hypertension at baseline. A successful proof of concept for using machine learning to predict intervention completion was presented
- 12 March 2019
- Published by JMIR Publications Inc. in JMIR Cardio
- Vol. 3 (1) , e13030
Open AccessAchieving Rapid Blood Pressure Control With Digital Therapeutics: Retrospective Cohort and Machine Learning Study
Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet #Preprint #PeerReviewMe: Warning: This is a unreviewed preprint. Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn. Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period. Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author). Background: Behavioral therapies, such as e-counseling and self-monitoring dispensed through mobile apps, have been shown to improve blood pressure but results vary and long-term engagement is a challenge. Machine learning is a rapidly advancing discipline that can be used to generate predictive and responsive models for the management and treatment of chronic conditions and shows potential for meaningfully improving outcomes. Objective: This retrospective analysis of data from the digital therapeutic, Better, examines its effect on blood pressure in hypertensive adults and explores the use of machine learning methods to predict intervention completion. Methods: The Better database was queried for participants with hypertension, who engaged with the intervention for at least 2 weeks and had paired blood pressure (BP) values. Participants were required to be ≥ 18 years old, reside in the United States and own a smartphone. The digital intervention offers personalized behavior therapy, including goal setting, skill building and self-monitoring. Participants reported BP at will. Changes in BP were calculated using averages of baseline and ending BP for each participant. Machine learning was used to generate a model of participants who would complete the intervention. Random forest models were trained at days 1, 3 and 7 of the intervention and the generalizability of the models was assessed using leave-one-out cross validation. Results: The primary analysis cohort was comprised of 172 hypertensive participants with paired BP values who were engaged with the intervention. Participants were 86.1% female, with a mean age of 55.0 years (95% CI 53.7, 56.2), baseline systolic blood pressure (SBP) of 138.9 mmHg (95% CI 136.6, 141.3) and diastolic (DBP) of 86.2 mmHg (95% CI 84.8, 87.7). Mean BP change was -11.5 mmHg for SBP and -5.9 mmHg for DBP over a mean of 62.6 days (P < .001). Amongst participants with Stage 2 Hypertension, mean BP change was -17.6 mmHg for SBP and -8.8 mmHg for DBP. Changes in BP remained significant in a mixed effects model accounting for baseline SBP, age, gender and BMI (P < .001). 43% of participants tracking BP at 12 weeks achieved the 2017 ACC/AHA definition of blood pressure control. The 7-day predictive model for intervention completion was trained on 429 participants and the AUC of the receiver operating characteristic was .72 with a false positive rate of 25%. Conclusions: The Better digital therapeutic was effective in lowering blood pressure among hypertensive adults. The degree of BP reduction was clinically meaningful and achieved rapidly by a majority of studied participants. Greater improvement was observed in participants with more severe hypertension at baseline. A successful proof of concept for using machine learning to predict intervention completion was presented
- Open Access
- 6 December 2018
- Published by JMIR Publications Inc. in JMIR Preprints
Achieving Rapid Blood Pressure Control With Digital Therapeutics: Retrospective Cohort and Machine Learning Study (Preprint)
BACKGROUND Behavioral therapies, such as electronic counseling and self-monitoring dispensed through mobile apps, have been shown to improve blood pressure, but the results vary and long-term engagement is a challenge. Machine learning is a rapidly advancing discipline that can be used to generate predictive and responsive models for the management and treatment of chronic conditions and shows potential for meaningfully improving outcomes. OBJECTIVE The objectives of this retrospective analysis were to examine the effect of a novel digital therapeutic on blood pressure in adults with hypertension and to explore the ability of machine learning to predict participant completion of the intervention. METHODS Participants with hypertension, who engaged with the digital intervention for at least 2 weeks and had paired blood pressure values, were identified from the intervention database. Participants were required to be ≥18 years old, reside in the United States, and own a smartphone. The digital intervention offers personalized behavior therapy, including goal setting, skill building, and self-monitoring. Participants reported blood pressure values at will, and changes were calculated using averages of baseline and final values for each participant. Machine learning was used to generate a model of participants who would complete the intervention. Random forest models were trained at days 1, 3, and 7 of the intervention, and the generalizability of the models was assessed using leave-one-out cross-validation. RESULTS The primary cohort comprised 172 participants with hypertension, having paired blood pressure values, who were engaged with the intervention. Of the total, 86.1% participants were women, the mean age was 55.0 years (95% CI 53.7-56.2), baseline systolic blood pressure was 138.9 mmHg (95% CI 136.6-141.3), and diastolic was 86.2 mmHg (95% CI 84.8-87.7). Mean change was –11.5 mmHg for systolic blood pressure and –5.9 mmHg for diastolic blood pressure over a mean of 62.6 days (P<.001). Among participants with stage 2 hypertension, mean change was –17.6 mmHg for systolic blood pressure and –8.8 mmHg for diastolic blood pressure. Changes in blood pressure remained significant in a mixed-effects model accounting for the baseline systolic blood pressure, age, gender, and body mass index (P<.001). A total of 43% of the participants tracking their blood pressure at 12 weeks achieved the 2017 American College of Cardiology/American Heart Association definition of blood pressure control. The 7-day predictive model for intervention completion was trained on 427 participants, and the area under the receiver operating characteristic curve was .78. CONCLUSIONS Reductions in blood pressure were observed in adults with hypertension who used the digital therapeutic. The degree of blood pressure reduction was clinically meaningful and achieved rapidly by a majority of the studied participants. Greater improvement was observed in participants with more severe hypertension at baseline. A successful proof of concept for using machine learning to predict intervention completion was presented.
- 6 December 2018
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessAchieving Rapid Blood Pressure Control With Digital Therapeutics: Retrospective Cohort and Machine Learning Study (Preprint)
BACKGROUND Behavioral therapies, such as electronic counseling and self-monitoring dispensed through mobile apps, have been shown to improve blood pressure, but the results vary and long-term engagement is a challenge. Machine learning is a rapidly advancing discipline that can be used to generate predictive and responsive models for the management and treatment of chronic conditions and shows potential for meaningfully improving outcomes. OBJECTIVE The objectives of this retrospective analysis were to examine the effect of a novel digital therapeutic on blood pressure in adults with hypertension and to explore the ability of machine learning to predict participant completion of the intervention. METHODS Participants with hypertension, who engaged with the digital intervention for at least 2 weeks and had paired blood pressure values, were identified from the intervention database. Participants were required to be ≥18 years old, reside in the United States, and own a smartphone. The digital intervention offers personalized behavior therapy, including goal setting, skill building, and self-monitoring. Participants reported blood pressure values at will, and changes were calculated using averages of baseline and final values for each participant. Machine learning was used to generate a model of participants who would complete the intervention. Random forest models were trained at days 1, 3, and 7 of the intervention, and the generalizability of the models was assessed using leave-one-out cross-validation. RESULTS The primary cohort comprised 172 participants with hypertension, having paired blood pressure values, who were engaged with the intervention. Of the total, 86.1% participants were women, the mean age was 55.0 years (95% CI 53.7-56.2), baseline systolic blood pressure was 138.9 mmHg (95% CI 136.6-141.3), and diastolic was 86.2 mmHg (95% CI 84.8-87.7). Mean change was –11.5 mmHg for systolic blood pressure and –5.9 mmHg for diastolic blood pressure over a mean of 62.6 days (P<.001). Among participants with stage 2 hypertension, mean change was –17.6 mmHg for systolic blood pressure and –8.8 mmHg for diastolic blood pressure. Changes in blood pressure remained significant in a mixed-effects model accounting for the baseline systolic blood pressure, age, gender, and body mass index (P<.001). A total of 43% of the participants tracking their blood pressure at 12 weeks achieved the 2017 American College of Cardiology/American Heart Association definition of blood pressure control. The 7-day predictive model for intervention completion was trained on 427 participants, and the area under the receiver operating characteristic curve was .78. CONCLUSIONS Reductions in blood pressure were observed in adults with hypertension who used the digital therapeutic. The degree of blood pressure reduction was clinically meaningful and achieved rapidly by a majority of the studied participants. Greater improvement was observed in participants with more severe hypertension at baseline. A successful proof of concept for using machine learning to predict intervention completion was presented.
- Open Access
- 14 February 2018
- Published by JMIR Publications Inc. in JMIR Diabetes
- Vol. 3 (1) , e4
Change in Glycemic Control With Use of a Digital Therapeutic in Adults With Type 2 Diabetes: Cohort Study
Emerging Technologies, Medical Devices, Apps, Sensors, and Informatics to Help People with Diabetes
- 14 February 2018
- Published by JMIR Publications Inc. in JMIR Diabetes
- Vol. 3 (1) , e4
Open AccessChange in Glycemic Control With Use of a Digital Therapeutic in Adults With Type 2 Diabetes: Cohort Study
Emerging Technologies, Medical Devices, Apps, Sensors, and Informatics to Help People with Diabetes
- Open Access
- 7 December 2017
- Published by JMIR Publications Inc.
- Vol. 3 (1) , e4
Change in Glycemic Control With Use of a Digital Therapeutic in Adults With Type 2 Diabetes: Cohort Study
Emerging Technologies, Medical Devices, Apps, Sensors, and Informatics to Help People with Diabetes
- 7 December 2017
- Published by JMIR Publications Inc.
- Vol. 3 (1) , e4
Open AccessChange in Glycemic Control With Use of a Digital Therapeutic in Adults With Type 2 Diabetes: Cohort Study
Emerging Technologies, Medical Devices, Apps, Sensors, and Informatics to Help People with Diabetes