Carolina Moreira
Carolina Moreira has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Anabela C. Areias
Sword Health, Inc, Draper, UT, USA
6 shared publications - Dora Janela
Sword Health, Inc, Draper, UT, USA
6 shared publications - Fabíola Costa
Sword Health
6 shared publications - Fernando Dias Correia
SWORD Health, Porto, Portugal
6 shared publications - Maria Molinos
SWORD Health, Inc., Draper, UT 84043, USA
6 shared publications - Steven Paul Cohen
Northwestern University Feinberg School of Medicine
6 shared publications
Latest Publications
- Open Access
- 27 August 2026
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 28, e97501
Addendum: Exploring the Importance of Race and Gender Concordance Between Patients and Physical Therapists in Digital Rehabilitation for Musculoskeletal Conditions: Observational, Longitudinal Study
- 27 August 2026
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 28, e97501
Open AccessAddendum: Exploring the Importance of Race and Gender Concordance Between Patients and Physical Therapists in Digital Rehabilitation for Musculoskeletal Conditions: Observational, Longitudinal Study
- Open Access
- 19 November 2024
- Published by JMIR Publications Inc. in JMIR Medical Informatics
- Vol. 12 (1) , e64806
Predicting Pain Response to a Remote Musculoskeletal Care Program for Low Back Pain Management: Development of a Prediction Tool
Background: Low back pain (LBP) presents with diverse manifestations, necessitating personalized treatment approaches that recognize various phenotypes within the same diagnosis, which could be achieved through precision medicine. Although prediction strategies have been explored, including those employing artificial intelligence (AI), they still lack scalability and real-time capabilities. Digital care programs (DCPs) facilitate seamless data collection through the Internet of Things and cloud storage, creating an ideal environment for developing and implementing an AI predictive tool to assist clinicians in dynamically optimizing treatment. Objective: This study aims to develop an AI tool that continuously assists physical therapists in predicting an individual’s potential for achieving clinically significant pain relief by the end of the program. A secondary aim was to identify predictors of pain nonresponse to guide treatment adjustments. Methods: Data collected actively (eg, demographic and clinical information) and passively in real-time (eg, range of motion, exercise performance, and socioeconomic data from public data sources) from 6125 patients enrolled in a remote digital musculoskeletal intervention program were stored in the cloud. Two machine learning techniques, recurrent neural networks (RNNs) and light gradient boosting machine (LightGBM), continuously analyzed session updates up to session 7 to predict the likelihood of achieving significant pain relief at the program end. Model performance was assessed using the area under the receiver operating characteristic curve (ROC-AUC), precision-recall curves, specificity, and sensitivity. Model explainability was assessed using SHapley Additive exPlanations values. Results: At each session, the model provided a prediction about the potential of being a pain responder, with performance improving over time (P<.001). By session 7, the RNN achieved an ROC-AUC of 0.70 (95% CI 0.65-0.71), and the LightGBM achieved an ROC-AUC of 0.71 (95% CI 0.67-0.72). Both models demonstrated high specificity in scenarios prioritizing high precision. The key predictive features were pain-associated domains, exercise performance, motivation, and compliance, informing continuous treatment adjustments to maximize response rates. Conclusions: This study underscores the potential of an AI predictive tool within a DCP to enhance the management of LBP, supporting physical therapists in redirecting care pathways early and throughout the treatment course. This approach is particularly important for addressing the heterogeneous phenotypes observed in LBP. Trial Registration: ClinicalTrials.gov NCT04092946; https://clinicaltrials.gov/ct2/show/NCT04092946 and NCT05417685; https://clinicaltrials.gov/ct2/show/NCT05417685
- 19 November 2024
- Published by JMIR Publications Inc. in JMIR Medical Informatics
- Vol. 12 (1) , e64806
Open AccessPredicting Pain Response to a Remote Musculoskeletal Care Program for Low Back Pain Management: Development of a Prediction Tool
Background: Low back pain (LBP) presents with diverse manifestations, necessitating personalized treatment approaches that recognize various phenotypes within the same diagnosis, which could be achieved through precision medicine. Although prediction strategies have been explored, including those employing artificial intelligence (AI), they still lack scalability and real-time capabilities. Digital care programs (DCPs) facilitate seamless data collection through the Internet of Things and cloud storage, creating an ideal environment for developing and implementing an AI predictive tool to assist clinicians in dynamically optimizing treatment. Objective: This study aims to develop an AI tool that continuously assists physical therapists in predicting an individual’s potential for achieving clinically significant pain relief by the end of the program. A secondary aim was to identify predictors of pain nonresponse to guide treatment adjustments. Methods: Data collected actively (eg, demographic and clinical information) and passively in real-time (eg, range of motion, exercise performance, and socioeconomic data from public data sources) from 6125 patients enrolled in a remote digital musculoskeletal intervention program were stored in the cloud. Two machine learning techniques, recurrent neural networks (RNNs) and light gradient boosting machine (LightGBM), continuously analyzed session updates up to session 7 to predict the likelihood of achieving significant pain relief at the program end. Model performance was assessed using the area under the receiver operating characteristic curve (ROC-AUC), precision-recall curves, specificity, and sensitivity. Model explainability was assessed using SHapley Additive exPlanations values. Results: At each session, the model provided a prediction about the potential of being a pain responder, with performance improving over time (P<.001). By session 7, the RNN achieved an ROC-AUC of 0.70 (95% CI 0.65-0.71), and the LightGBM achieved an ROC-AUC of 0.71 (95% CI 0.67-0.72). Both models demonstrated high specificity in scenarios prioritizing high precision. The key predictive features were pain-associated domains, exercise performance, motivation, and compliance, informing continuous treatment adjustments to maximize response rates. Conclusions: This study underscores the potential of an AI predictive tool within a DCP to enhance the management of LBP, supporting physical therapists in redirecting care pathways early and throughout the treatment course. This approach is particularly important for addressing the heterogeneous phenotypes observed in LBP. Trial Registration: ClinicalTrials.gov NCT04092946; https://clinicaltrials.gov/ct2/show/NCT04092946 and NCT05417685; https://clinicaltrials.gov/ct2/show/NCT05417685
- Open Access
- 29 October 2024
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 26 (1) , e65354
Exploring the Importance of Race and Gender Concordance Between Patients and Physical Therapists in Digital Rehabilitation for Musculoskeletal Conditions: Observational, Longitudinal Study
Background: Race/ethnicity and gender concordance between patients and providers is a potential strategy to improve health care interventions. In digital health, where human interactions occur both synchronously and asynchronously, the effect of concordance between patients and providers is unknown. Objective: This study aimed to evaluate the impact of race/ethnicity or gender concordance between patients and physical therapists (PTs) in engagement and the clinical outcomes following a digital care program (DCP) in patients with musculoskeletal (MSK) conditions. Methods: This secondary analysis of 2 prospective longitudinal studies (originally focused on assessing the acceptance, engagement, and clinical outcomes after a remote DCP) examined the impact of both race/ethnicity concordance and gender concordance between patients and PTs on outcomes for a digital intervention for MSK conditions. Outcomes included engagement (measured by the completion rate and communication, assessed by text interactions), satisfaction, and clinical outcomes (response rate, ie, percentage of patients achieving at least a minimal clinically important change in pain, measured by the Numerical Pain Rating Scale [NPRS]; anxiety, measured by the Generalized Anxiety Disorder 7-item scale [GAD-7]; depression, measured by the Patient Health Questionnaire 9-item [PHQ-9]; and daily activity impairment, measured by the Work Productivity and Activity Impairment [WPAI] questionnaire). Results: Of 71,201 patients, 63.9% (n=45,507) were matched with their PT in terms of race/ethnicity, while 61.2% (n=43,560) were matched for gender. Concordant dyads showed a higher completion rate among White (adjusted odds ratio [aOR] 1.11, 95% CI 1.05-1.19, P<.001) and Hispanic (aOR 1.27, 95% CI 1.08-1.54, P=.009) groups, as well as women (aOR 1.10, 95% CI 1.06-1.18, P<.001), when compared to discordant dyads. High and similar levels of interaction between patients and PTs were observed across race/ethnicity and gender dyads, except for Asian concordant dyads (adjusted β coefficient 5.32, 95% CI 3.28-7.36, P<.001). Concordance did not affect satisfaction, with high values (>8.52, 95% CI 8.27-8.77) reported across all dyads. Response rates for pain, anxiety, and daily activity impairment were unaffected by race/ethnicity concordance. An exception was observed for depression, with White patients reporting a higher response rate when matched with PTs from other races/ethnicities (aOR 1.20, 95% CI 1.02-1.39, P=.02). In terms of gender, men had a slightly higher pain response rate in discordant dyads (aOR 1.08, 95% CI 1.01-1.15, P=.03) and a higher depression response rate in concordant dyads (aOR 1.23, 95% CI 1.05-1.47, P=.01). Conclusions: Race/ethnicity and gender concordance between patients and PTs does not translate into higher satisfaction or improvement for most clinical outcomes, aside from a positive effect on treatment completion. These results highlight the importance of other PT characteristics, in addition to race/ethnicity or gender concordance, suggesting the potential benefit of experience, languages spoken, and cultural safety training as ways to optimize care. Trial Registration: ClinicalTrials.gov NCT04092946, NCT05417685; https://clinicaltrials.gov/study/NCT05417685, https://clinicaltrials.gov/study/NCT04092946
- 29 October 2024
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 26 (1) , e65354
Open AccessExploring the Importance of Race and Gender Concordance Between Patients and Physical Therapists in Digital Rehabilitation for Musculoskeletal Conditions: Observational, Longitudinal Study
Background: Race/ethnicity and gender concordance between patients and providers is a potential strategy to improve health care interventions. In digital health, where human interactions occur both synchronously and asynchronously, the effect of concordance between patients and providers is unknown. Objective: This study aimed to evaluate the impact of race/ethnicity or gender concordance between patients and physical therapists (PTs) in engagement and the clinical outcomes following a digital care program (DCP) in patients with musculoskeletal (MSK) conditions. Methods: This secondary analysis of 2 prospective longitudinal studies (originally focused on assessing the acceptance, engagement, and clinical outcomes after a remote DCP) examined the impact of both race/ethnicity concordance and gender concordance between patients and PTs on outcomes for a digital intervention for MSK conditions. Outcomes included engagement (measured by the completion rate and communication, assessed by text interactions), satisfaction, and clinical outcomes (response rate, ie, percentage of patients achieving at least a minimal clinically important change in pain, measured by the Numerical Pain Rating Scale [NPRS]; anxiety, measured by the Generalized Anxiety Disorder 7-item scale [GAD-7]; depression, measured by the Patient Health Questionnaire 9-item [PHQ-9]; and daily activity impairment, measured by the Work Productivity and Activity Impairment [WPAI] questionnaire). Results: Of 71,201 patients, 63.9% (n=45,507) were matched with their PT in terms of race/ethnicity, while 61.2% (n=43,560) were matched for gender. Concordant dyads showed a higher completion rate among White (adjusted odds ratio [aOR] 1.11, 95% CI 1.05-1.19, P<.001) and Hispanic (aOR 1.27, 95% CI 1.08-1.54, P=.009) groups, as well as women (aOR 1.10, 95% CI 1.06-1.18, P<.001), when compared to discordant dyads. High and similar levels of interaction between patients and PTs were observed across race/ethnicity and gender dyads, except for Asian concordant dyads (adjusted β coefficient 5.32, 95% CI 3.28-7.36, P<.001). Concordance did not affect satisfaction, with high values (>8.52, 95% CI 8.27-8.77) reported across all dyads. Response rates for pain, anxiety, and daily activity impairment were unaffected by race/ethnicity concordance. An exception was observed for depression, with White patients reporting a higher response rate when matched with PTs from other races/ethnicities (aOR 1.20, 95% CI 1.02-1.39, P=.02). In terms of gender, men had a slightly higher pain response rate in discordant dyads (aOR 1.08, 95% CI 1.01-1.15, P=.03) and a higher depression response rate in concordant dyads (aOR 1.23, 95% CI 1.05-1.47, P=.01). Conclusions: Race/ethnicity and gender concordance between patients and PTs does not translate into higher satisfaction or improvement for most clinical outcomes, aside from a positive effect on treatment completion. These results highlight the importance of other PT characteristics, in addition to race/ethnicity or gender concordance, suggesting the potential benefit of experience, languages spoken, and cultural safety training as ways to optimize care. Trial Registration: ClinicalTrials.gov NCT04092946, NCT05417685; https://clinicaltrials.gov/study/NCT05417685, https://clinicaltrials.gov/study/NCT04092946
- Open Access
- 13 August 2024
- Published by JMIR Publications Inc. in JMIR Preprints
Exploring the Importance of Race and Gender Concordance Between Patients and Physical Therapists in Digital Rehabilitation for Musculoskeletal Conditions: Observational, Longitudinal Study (Preprint)
BACKGROUND Race/ethnicity and gender concordance between patients and providers is a potential strategy to improve health care interventions. In digital health, where human interactions occur both synchronously and asynchronously, the effect of concordance between patients and providers is unknown. OBJECTIVE This study aimed to evaluate the impact of race/ethnicity or gender concordance between patients and physical therapists (PTs) in engagement and the clinical outcomes following a digital care program (DCP) in patients with musculoskeletal (MSK) conditions. METHODS This secondary analysis of 2 prospective longitudinal studies (originally focused on assessing the acceptance, engagement, and clinical outcomes after a remote DCP) examined the impact of both race/ethnicity concordance and gender concordance between patients and PTs on outcomes for a digital intervention for MSK conditions. Outcomes included engagement (measured by the completion rate and communication, assessed by text interactions), satisfaction, and clinical outcomes (response rate, ie, percentage of patients achieving at least a minimal clinically important change in pain, measured by the Numerical Pain Rating Scale [NPRS]; anxiety, measured by the Generalized Anxiety Disorder 7-item scale [GAD-7]; depression, measured by the Patient Health Questionnaire 9-item [PHQ-9]; and daily activity impairment, measured by the Work Productivity and Activity Impairment [WPAI] questionnaire). RESULTS Of 71,201 patients, 63.9% (n=45,507) were matched with their PT in terms of race/ethnicity, while 61.2% (n=43,560) were matched for gender. Concordant dyads showed a higher completion rate among White (adjusted odds ratio [aOR] 1.11, 95% CI 1.05-1.19, P<.001) and Hispanic (aOR 1.27, 95% CI 1.08-1.54, P=.009) groups, as well as women (aOR 1.10, 95% CI 1.06-1.18, P<.001), when compared to discordant dyads. High and similar levels of interaction between patients and PTs were observed across race/ethnicity and gender dyads, except for Asian concordant dyads (adjusted β coefficient 5.32, 95% CI 3.28-7.36, P<.001). Concordance did not affect satisfaction, with high values (>8.52, 95% CI 8.27-8.77) reported across all dyads. Response rates for pain, anxiety, and daily activity impairment were unaffected by race/ethnicity concordance. An exception was observed for depression, with White patients reporting a higher response rate when matched with PTs from other races/ethnicities (aOR 1.20, 95% CI 1.02-1.39, P=.02). In terms of gender, men had a slightly higher pain response rate in discordant dyads (aOR 1.08, 95% CI 1.01-1.15, P=.03) and a higher depression response rate in concordant dyads (aOR 1.23, 95% CI 1.05-1.47, P=.01). CONCLUSIONS Race/ethnicity and gender concordance between patients and PTs does not translate into higher satisfaction or improvement for most clinical outcomes, aside from a positive effect on treatment completion. These results highlight the importance of other PT characteristics, in addition to race/ethnicity or gender concordance, suggesting the potential benefit of experience, languages spoken, and cultural safety training as ways to optimize care. CLINICALTRIAL ClinicalTrials.gov NCT04092946, NCT05417685; https://clinicaltrials.gov/study/NCT05417685, https://clinicaltrials.gov/study/NCT04092946
- 13 August 2024
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessExploring the Importance of Race and Gender Concordance Between Patients and Physical Therapists in Digital Rehabilitation for Musculoskeletal Conditions: Observational, Longitudinal Study (Preprint)
BACKGROUND Race/ethnicity and gender concordance between patients and providers is a potential strategy to improve health care interventions. In digital health, where human interactions occur both synchronously and asynchronously, the effect of concordance between patients and providers is unknown. OBJECTIVE This study aimed to evaluate the impact of race/ethnicity or gender concordance between patients and physical therapists (PTs) in engagement and the clinical outcomes following a digital care program (DCP) in patients with musculoskeletal (MSK) conditions. METHODS This secondary analysis of 2 prospective longitudinal studies (originally focused on assessing the acceptance, engagement, and clinical outcomes after a remote DCP) examined the impact of both race/ethnicity concordance and gender concordance between patients and PTs on outcomes for a digital intervention for MSK conditions. Outcomes included engagement (measured by the completion rate and communication, assessed by text interactions), satisfaction, and clinical outcomes (response rate, ie, percentage of patients achieving at least a minimal clinically important change in pain, measured by the Numerical Pain Rating Scale [NPRS]; anxiety, measured by the Generalized Anxiety Disorder 7-item scale [GAD-7]; depression, measured by the Patient Health Questionnaire 9-item [PHQ-9]; and daily activity impairment, measured by the Work Productivity and Activity Impairment [WPAI] questionnaire). RESULTS Of 71,201 patients, 63.9% (n=45,507) were matched with their PT in terms of race/ethnicity, while 61.2% (n=43,560) were matched for gender. Concordant dyads showed a higher completion rate among White (adjusted odds ratio [aOR] 1.11, 95% CI 1.05-1.19, P<.001) and Hispanic (aOR 1.27, 95% CI 1.08-1.54, P=.009) groups, as well as women (aOR 1.10, 95% CI 1.06-1.18, P<.001), when compared to discordant dyads. High and similar levels of interaction between patients and PTs were observed across race/ethnicity and gender dyads, except for Asian concordant dyads (adjusted β coefficient 5.32, 95% CI 3.28-7.36, P<.001). Concordance did not affect satisfaction, with high values (>8.52, 95% CI 8.27-8.77) reported across all dyads. Response rates for pain, anxiety, and daily activity impairment were unaffected by race/ethnicity concordance. An exception was observed for depression, with White patients reporting a higher response rate when matched with PTs from other races/ethnicities (aOR 1.20, 95% CI 1.02-1.39, P=.02). In terms of gender, men had a slightly higher pain response rate in discordant dyads (aOR 1.08, 95% CI 1.01-1.15, P=.03) and a higher depression response rate in concordant dyads (aOR 1.23, 95% CI 1.05-1.47, P=.01). CONCLUSIONS Race/ethnicity and gender concordance between patients and PTs does not translate into higher satisfaction or improvement for most clinical outcomes, aside from a positive effect on treatment completion. These results highlight the importance of other PT characteristics, in addition to race/ethnicity or gender concordance, suggesting the potential benefit of experience, languages spoken, and cultural safety training as ways to optimize care. CLINICALTRIAL ClinicalTrials.gov NCT04092946, NCT05417685; https://clinicaltrials.gov/study/NCT05417685, https://clinicaltrials.gov/study/NCT04092946
- Open Access
- 26 July 2024
- Published by JMIR Publications Inc. in JMIR Preprints
Predicting Pain Response to a Remote Musculoskeletal Care Program for Low Back Pain Management: Development of a Prediction Tool (Preprint)
BACKGROUND Low back pain (LBP) presents with diverse manifestations, necessitating personalized treatment approaches that recognize various phenotypes within the same diagnosis, which could be achieved through precision medicine. Although prediction strategies have been explored, including those employing artificial intelligence (AI), they still lack scalability and real-time capabilities. Digital care programs (DCPs) facilitate seamless data collection through the Internet of Things and cloud storage, creating an ideal environment for developing and implementing an AI predictive tool to assist clinicians in dynamically optimizing treatment. OBJECTIVE This study aims to develop an AI tool that continuously assists physical therapists in predicting an individual’s potential for achieving clinically significant pain relief by the end of the program. A secondary aim was to identify predictors of pain nonresponse to guide treatment adjustments. METHODS Data collected actively (eg, demographic and clinical information) and passively in real-time (eg, range of motion, exercise performance, and socioeconomic data from public data sources) from 6125 patients enrolled in a remote digital musculoskeletal intervention program were stored in the cloud. Two machine learning techniques, recurrent neural networks (RNNs) and light gradient boosting machine (LightGBM), continuously analyzed session updates up to session 7 to predict the likelihood of achieving significant pain relief at the program end. Model performance was assessed using the area under the receiver operating characteristic curve (ROC-AUC), precision-recall curves, specificity, and sensitivity. Model explainability was assessed using SHapley Additive exPlanations values. RESULTS At each session, the model provided a prediction about the potential of being a pain responder, with performance improving over time (P<.001). By session 7, the RNN achieved an ROC-AUC of 0.70 (95% CI 0.65-0.71), and the LightGBM achieved an ROC-AUC of 0.71 (95% CI 0.67-0.72). Both models demonstrated high specificity in scenarios prioritizing high precision. The key predictive features were pain-associated domains, exercise performance, motivation, and compliance, informing continuous treatment adjustments to maximize response rates. CONCLUSIONS This study underscores the potential of an AI predictive tool within a DCP to enhance the management of LBP, supporting physical therapists in redirecting care pathways early and throughout the treatment course. This approach is particularly important for addressing the heterogeneous phenotypes observed in LBP. CLINICALTRIAL ClinicalTrials.gov NCT04092946; https://clinicaltrials.gov/ct2/show/NCT04092946 and NCT05417685; https://clinicaltrials.gov/ct2/show/NCT05417685
- 26 July 2024
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessPredicting Pain Response to a Remote Musculoskeletal Care Program for Low Back Pain Management: Development of a Prediction Tool (Preprint)
BACKGROUND Low back pain (LBP) presents with diverse manifestations, necessitating personalized treatment approaches that recognize various phenotypes within the same diagnosis, which could be achieved through precision medicine. Although prediction strategies have been explored, including those employing artificial intelligence (AI), they still lack scalability and real-time capabilities. Digital care programs (DCPs) facilitate seamless data collection through the Internet of Things and cloud storage, creating an ideal environment for developing and implementing an AI predictive tool to assist clinicians in dynamically optimizing treatment. OBJECTIVE This study aims to develop an AI tool that continuously assists physical therapists in predicting an individual’s potential for achieving clinically significant pain relief by the end of the program. A secondary aim was to identify predictors of pain nonresponse to guide treatment adjustments. METHODS Data collected actively (eg, demographic and clinical information) and passively in real-time (eg, range of motion, exercise performance, and socioeconomic data from public data sources) from 6125 patients enrolled in a remote digital musculoskeletal intervention program were stored in the cloud. Two machine learning techniques, recurrent neural networks (RNNs) and light gradient boosting machine (LightGBM), continuously analyzed session updates up to session 7 to predict the likelihood of achieving significant pain relief at the program end. Model performance was assessed using the area under the receiver operating characteristic curve (ROC-AUC), precision-recall curves, specificity, and sensitivity. Model explainability was assessed using SHapley Additive exPlanations values. RESULTS At each session, the model provided a prediction about the potential of being a pain responder, with performance improving over time (P<.001). By session 7, the RNN achieved an ROC-AUC of 0.70 (95% CI 0.65-0.71), and the LightGBM achieved an ROC-AUC of 0.71 (95% CI 0.67-0.72). Both models demonstrated high specificity in scenarios prioritizing high precision. The key predictive features were pain-associated domains, exercise performance, motivation, and compliance, informing continuous treatment adjustments to maximize response rates. CONCLUSIONS This study underscores the potential of an AI predictive tool within a DCP to enhance the management of LBP, supporting physical therapists in redirecting care pathways early and throughout the treatment course. This approach is particularly important for addressing the heterogeneous phenotypes observed in LBP. CLINICALTRIAL ClinicalTrials.gov NCT04092946; https://clinicaltrials.gov/ct2/show/NCT04092946 and NCT05417685; https://clinicaltrials.gov/ct2/show/NCT05417685