Group‐Based Medical Mistrust and Logistical Factors Influencing Rheumatology Clinical Trial Enrollment: A Single‐Center Cross‐Sectional Survey

Abstract

Objective

To examine whether group‐based medical mistrust predicts willingness to enroll in a rheumatology clinical trial and, among those unwilling, to identify logistical factors that most influence their decision.

Methods

We conducted a single‐center, cross‐sectional survey at an academic rheumatology clinic between November 2024 and February 2025. Adults (≥18 years) with physician‐confirmed rheumatic disease completed the 12‐item Group‐Based Medical Mistrust Scale (GBMMS; range 12–60, higher = more mistrust), rated 11 logistical factors on a 7‐point scale (1 = extremely unlikely to influence, 7 = extremely likely), and answered a yes or no item on trial willingness. We assessed GBMMS internal consistency, compared mean mistrust by willingness using Welch's t‐test, and fit a univariable logistic regression with GBMMS predicting willingness. Among those unwilling, we ranked logistical factors by mean importance.

Results

Forty‐seven patients participated (79% women; mean age 49.2 ± 13.6 years); 66% reported willingness to enroll. Mean mistrust did not differ by willingness (willing 24.7 ± 11.6 vs unwilling 25.3 ± 13.4; t ₂₆ = 0.13, P = 0.893), and GBMMS did not predict willingness (odds ratio 0.996, 95% confidence interval 0.95–1.05; P = 0.89). Among the 16 patients unwilling to participate, the highest‐rated influences were the prospect of stopping disease progression (mean 5.38), physician encouragement (mean 4.56), travel or transportation (mean 4.13), and shared language with the doctor (mean 4.13).

Conclusion

Group‐based medical mistrust did not predict self‐reported willingness to join a rheumatology clinical trial. Instead, expectation of concrete disease improvement and personal endorsement by one's treating rheumatologist emerged as the strongest drivers of participation. Recruitment strategies that emphasize realistic potential benefits, incorporate physician‐delivered invitations, and address logistics may improve enrollment.

INTRODUCTION

Clinical trials are the cornerstone of evidence‐based rheumatology, yet recruiting and retaining representative patient cohorts remains a challenge. ¹ Underrepresentation of racial and ethnic groups in clinical research limits the generalizability of findings and exacerbates disparities among populations disproportionately affected by rheumatic diseases. ² , ³ , ⁴ Group‐based medical mistrust, the belief that health care systems and providers may act against the interests of one's own ethnic group, has been associated with lower engagement in preventive and therapeutic care. ⁵ , ⁶

Despite extensive research on medical mistrust in preventive settings, its impact on participation in rheumatology clinical trials remains unexplored. The 12‐item Group‐Based Medical Mistrust Scale (GBMMS) reliably measures perceptions of systemic bias and institutional distrust based on group identity. ⁵ In addition to psychosocial factors, practical concerns such as transportation, time commitment, and physician recommendation play important roles in patients’ decisions to participate in research. ⁴ , ⁷ , ⁸ , ⁹ , ¹⁰ Specifically, distance to trial sites and insufficient transportation support pose significant barriers, ¹¹ whereas language concordance between patients and health care providers has been shown to enhance communication and satisfaction. ¹² Patients also routinely weigh the potential benefits of stopping disease progression against the burdens of trial participation. ¹³ However, no study in rheumatology has examined both trust‐related attitudes and logistical considerations simultaneously, leaving a critical gap in our understanding of enrollment drivers.

To address these gaps, we conducted a single‐center cross‐sectional survey in an academic rheumatology clinic to test whether higher GBMMS scores are associated with lower self‐reported willingness to participate in a rheumatology clinical trial. Secondarily, among patients who reported unwillingness, we sought to identify which logistical concerns most strongly influenced their decision. By clarifying the relative importance of psychosocial and practical barriers, this study aims to inform more effective, patient‐centered recruitment strategies for rheumatology research.

PATIENTS AND METHODS

Study design and setting

We conducted a single‐center, cross‐sectional survey at the Jersey City Medical Center rheumatology clinic between November 1, 2024, and February 28, 2025. The study protocol was approved by the Rutgers University Institutional Review Board (Approval ID Pro2025000039), and all participants provided written informed consent before enrollment.

Participants

Eligible participants were adults (≥18 years) under rheumatologic care with a physician‐confirmed diagnosis of a rheumatic disease. Diagnoses included rheumatoid arthritis (classified per 2010 American College of Rheumatology/EULAR criteria), systemic lupus erythematosus (per 2012 Systemic Lupus International Collaborating Clinics criteria), and other rheumatic conditions (eg, psoriatic arthritis, mixed connective tissue disease, gout, and Sjögren disease) at the treating clinician's discretion. Patients unable to provide informed consent or complete the questionnaire were not enrolled. A convenience sample of eligible patients presenting for routine care during the study period was approached in the clinic, and all participants provided informed consent before survey completion. No financial incentive was provided.

Measures

Medical mistrust

Group‐based mistrust was assessed using the 12‐item GBMMS. Items are rated on a 5‐point Likert scale (1 = strongly disagree to 5 = strongly agree) and summed to a total score ranging from 12 to 60, with higher scores indicating greater mistrust. Negatively worded items were reverse scored per instrument guidance.

Trial willingness

Willingness to participate in a rheumatology clinical trial was assessed with a single yes or no item (“Would you enroll in a clinical trial if offered?”).

Logistical factors

Patients rated the likelihood that specific practical or motivational factors would influence their decision about trial participation on a 7‐point scale (1 = extremely unlikely to influence me to 7 = extremely likely). Factors included travel or transportation, scheduling or time commitment, prospect of receiving a new treatment, blood draws, sharing or storing personal health information, level of compensation, prospect of stopping disease progression, encouragement from one's doctor to join a trial, prospect of patient education by the doctor, shared ethnic group with the doctor, and shared language with the doctor.

Demographics and clinical characteristics

Age, sex, and primary rheumatic diagnosis were recorded. Race and ethnicity was obtained from the electronic health record according to the patient's self‐identification.

Outcomes

The primary objective was to test the association between GBMMS total score and willingness to participate in a clinical trial (yes or no). Secondary and exploratory objectives were (1) among participants unwilling to enroll, to rank logistical factors by mean importance and (2) in the full sample, to examine correlations between GBMMS score and each logistical factor rating.

Data collection and management

Surveys were administered on electronic tablet devices in the clinic waiting area, examination rooms or by telephone, with bilingual research staff available to provide assistance in English and Spanish. Responses were captured using a Qualtrics survey. Demographic and clinical data (age, sex, race and ethnicity, rheumatic diagnosis, disease duration, and prior clinical trial experience) were abstracted from the electronic medical record.

Statistical analysis

Statistical analyses were performed using Python (version 3.13.3) with the Pandas (version 2.2.3), SciPy (version 1.15.3), and statsmodels (version 0.14.4) packages. Participant characteristics are presented as means (SD) for continuous variables and counts (percentage) for categorical variables. Internal consistency of the GBMMS was evaluated using Cronbach's alpha.

For the primary analysis, mean GBMMS scores were compared between willing versus unwilling groups using Welch's t‐test. We also fit a univariable logistic regression with willingness (yes = 1 and no = 0) as the dependent variable and GBMMS total score (per one‐point increase) as the predictor, reporting odds ratios (ORs) with 95% confidence intervals (CIs).

For secondary analyses, among participants unwilling to enroll, logistical factors were summarized by mean (SD) and ranked by magnitude. In the full sample, Pearson correlation coefficients (r) quantified associations between GBMMS total score and each logistical factor rating. Two‐sided tests with α = 0.05 were used throughout; no adjustments were made for multiple comparisons given the exploratory nature of the study and small sample size.

Sample size and study classification

This was an exploratory, hypothesis‐generating study; no formal a priori sample size calculation was performed. The sample size was determined by clinic flow and the fixed study period.

Ethics

The study protocol was approved by the Rutgers University Institutional Review Board, and all participants provided informed consent.

RESULTS

Participant characteristics

We enrolled 47 rheumatology patients (79% women; mean ± SD age 49.2 ± 13.6 years). The cohort was racially and ethnically diverse (White American 28%, African American 23%, Latino/a 21%, Asian/Pacific Islander 15%, other 13%). Diagnoses included rheumatoid arthritis (47%), systemic lupus erythematosus (15%), and other rheumatic conditions. Overall, 31 of 47 patients (66.0%) indicated they would be willing to enroll in a clinical trial (Table 1). The GBMMS (range 12–60) showed excellent internal consistency (Cronbach's α = 0.90). GBMMS scores were, on average, in the lower‐moderate range (mean 24.9 ± 12.1).

Table 1.

Demographics of participants

Characteristics	Participants (n = 47), n (%)
Age, y
18–35	10 (21.28)
36–45	9 (19.15)
46–65	24 (51.06)
66+	4 (8.51)
Sex
Female	37 (78.72)
Male	10 (21.28)
Race and ethnicity
African American	11 (23.40)
Asian/Pacific Islander	7 (14.89)
Latino/a	10 (21.28)
Other	6 (12.76)
White American	13 (27.66)
Education level
Incomplete high school	8 (17.02)
High school graduate	11 (23.4)
Technical certification	6 (12.77)
Some college (no degree)	2 (4.26)
4‐y college degree	13 (27.66)
Postgraduate degree	7 (14.89)
Rheumatic diseases
Rheumatoid arthritis	22 (46.80)
Systemic lupus erythematosus	7 (14.89)
Psoriatic arthritis	3 (6.38)
Mixed connective tissue disease	3 (6.38)
Gout	2 (4.26)
Sacroiliitis	2 (4.26)
Sjögren disease	2 (4.26)
Other	6 (12.77)
Perceived control of disease
Yes	23 (48.94)
No	24 (51.06)
Trial willingness
Yes	31 (65.96)
No	16 (34.04)

Primary analysis: Mistrust and trial willingness

Mean ± SD GBMMS scores did not differ between participants who were willing versus unwilling to enroll (24.74 ± 11.62 vs 25.25 ± 13.36; Welch's t ₂₆ = 0.13, P = 0.893) (Table 2). In a univariable logistic regression with willingness as the outcome, mistrust was not associated with willingness (OR per one‐point GBMMS increase = 0.996; 95% CI 0.95–1.05; P = 0.89). Given the small sample and subgroup sizes, multivariable modeling was not pursued. Exploratory, descriptive checks did not reveal obvious differences in age band, sex, or race and ethnicity distributions between willing and unwilling participants.

Table 2.

Mean (SD) scores on individual GBMMS items for all participants and by willingness to participate in a clinical trial^*

Items	Mean (SD)			P
	All	Willing	Unwilling
Mean GBMMS score	24.91 (12.10)	24.74 (11.62)	25.25 (13.36)	0.893
1. Doctors sometimes hide information from patients who belong to my ethnic group.	2.55 (1.61)	2.74 (1.71)	2.19 (1.38)
2. Doctors have the best interests of people in my ethnic group. a	2.15 (1.43)	2.16 (1.39)	2.13 (1.54)
3. People of my ethnic group should not confide in doctors because it will be used against them.	1.74 (1.24)	1.58 (1.23)	2.06 (1.24)
4. People of my ethnic group should be suspicious of information from doctors.	1.60 (1.10)	1.52 (1.06)	1.75 (1.18)
5. People of my ethnic group cannot trust doctors.	1.55 (1.05)	1.48 (1.00)	1.69 (1.20)
6. People of my ethnic group should be suspicious of modern medicine.	1.74 (1.19)	1.71 (1.13)	1.81 (1.33)
7. Doctors treat people of my ethnic group like “guinea pigs.”	2.21 (1.52)	2.10 (1.58)	2.44 (1.41)
8. People of my ethnic group receive the same medical care from doctors as people from other groups. a	2.23 (1.62)	2.23 (1.61)	2.25 (1.69)
9. Doctors do not take the medical complaints of my ethnic group seriously.	2.79 (1.69)	2.81 (1.74)	2.75 (1.65)
10. People of my ethnic group are treated the same as people of other groups by doctors. a	2.34 (1.67)	2.29 (1.68)	2.44 (1.71)
11. In most hospitals, people of different ethnic groups receive the same kind of care. a	2.21 (1.60)	2.13 (1.57)	2.38 (1.71)
12. I have personally been treated poorly or unfairly by doctors because of my ethnicity.	1.89 (1.58)	2.03 (1.66)	1.63 (1.41)

^*GBMMS, Group‐Based Medical Mistrust Scale. Scale: 1 = strongly disagree to 5 = strongly agree.

^a Reverse scored.

Secondary or exploratory analyses: Logistical and motivational factors

Among the 16 participants who were unwilling to enroll, the highest‐rated influences (1 = extremely unlikely to influence; 7 = extremely likely) were the prospect of stopping disease progression (mean 5.38), physician encouragement (mean 4.56), travel or transportation (mean 4.13), and shared language with the doctor (mean 4.13) (Table 3).

Table 3.

Mean (SD) importance ratings of logistical factors for all participants and by willingness to participate in a clinical trial^*

Factors	Mean (SD)
	All	Willing	Unwilling
1. Travel or transportation	5.14 (1.47)	5.67 (1.45)	4.13 (0.89)
2. Scheduling or time commitment	3.13 (1.39)	3.42 (1.31)	2.56 (1.41)
3. Prospect of getting a new treatment	3.45 (2.14)	3.90 (2.18)	2.56 (1.82)
4. Blood draws	3.38 (1.26)	3.81 (1.01)	2.56 (1.31)
5. Having personal health information stored or shared	3.96 (0.81)	4.03 (0.91)	3.81 (0.54)
6. Level of compensation or payment	4.72 (2.01)	5.38 (1.73)	3.44 (1.97)
7. Prospect of stopping worsening disease	6.08 (1.59)	6.45 (1.56)	5.38 (1.40)
8. Trial encouragement from doctor	5.42 (1.54)	5.87 (1.31)	4.56 (1.63)
9. Prospect of patient education by doctor	4.49 (1.12)	4.70 (1.32)	4.06 (0.25)
10. Shared ethnic group with doctor	4.40 (1.15)	4.58 (1.29)	4.06 (0.77)
11. Shared language with doctor	4.38 (1.05)	4.52 (1.26)	4.13 (0.34)

^*Scale: 1 = extremely unlikely to 7 = extremely likely.

Descriptively, participants willing to enroll tended to rate most logistical and motivational factors as more influential than those unwilling. Examples include travel or transportation (mean 5.67 vs 4.13), compensation or payment (mean 5.38 vs 3.44), blood draws (mean 3.81 vs 2.56), scheduling or time commitment (mean 3.42 vs 2.56), and stopping disease progression (mean 6.45 vs 5.38). These between‐group differences were not formally tested and should be interpreted cautiously.

Across the full sample, higher mistrust showed small, nonsignificant inverse correlations with the perceived influence of certain facilitators, notably the prospect of receiving a new treatment (r = −0.28; P = 0.056) and physician encouragement (r = −0.23; P = 0.11). Correlations with other logistical items were smaller in magnitude and nonsignificant.

Age, sex, race and ethnicity, education level, rheumatic disease perceived control, and willingness to participate in a clinical trial

Across demographic and clinical characteristics, stated willingness to participate in a clinical trial did not differ by age (P = 0.750), sex (P = 0.071), race and ethnicity (P = 0.989), education level (P = 0.407), or perceived disease control (P = 0.609) (Table 4). Distributions across categories were similar in the willing and unwilling groups, albeit with limited power due to small sample sizes in some subgroups.

Table 4.

Age, sex, race and ethnicity, education level, rheumatic diseases and perceived control of rheumatic disease in relation to willingness to participate in a clinical trial

	Willing, n (%)	Unwilling, n (%)	All, n (%)	P
Age, y				0.750
18–35	7 (14.89)	3 (6.38)	10 (21.28)
36–45	5 (10.64)	4 (8.51)	9 (19.15)
46–65	17 (36.17)	7 (14.89)	24 (51.06)
66+	2 (4.26)	2 (4.26)	4 (8.51)
Sex				0.071
Female	22 (46.81)	15 (31.91)	37 (78.72)
Male	9 (19.15)	1 (2.13)	10 (21.28)
Race and ethnicity				0.989
African American	7 (14.89)	4 (8.51)	11 (23.40)
Asian/Pacific Islander	5 (10.64)	2 (4.26)	7 (14.89)
Latino/a	7 (14.89)	3 (6.39)	10 (21.28)
Other	4 (8.52)	2 (4.26)	6 (12.76)
White American	8 (17.02)	5 (10.64)	13 (27.66)
Education level				0.407
Incomplete high school	7 (14.89)	1 (2.13)	8 (17.02)
High school graduate	7 (14.89)	4 (8.51)	11 (23.40)
Technical certification	4 (8.51)	2 (4.26)	6 (12.77)
Some college (no degree)	2 (4.26)	0 (0)	2 (4.26)
4 y degree	6 (12.77)	7 (14.89)	13 (27.66)
Postgraduate	5 (10.64)	2 (4.26)	7 (14.89)
Rheumatic diseases				0.055
Rheumatoid arthritis	17 (36.17)	5 (10.64)	22 (46.81)
Systemic lupus erythematosus	5 (10.64)	2 (4.26)	7 (14.90)
Psoriatic arthritis	2 (4.26)	1 (2.13)	3 (6.38)
Mixed connective tissue disease	0 (0)	3 (6.38)	3 (6.38)
Gout	2 (4.26)	0 (0)	2 (4.26)
Sacroiliitis	2 (4.26)	0 (0)	2 (4.26)
Sjögren disease	0 (0)	2 (4.26)	2 (4.26)
Other	3 (6.38)	3 (6.38)	6 (12.76)
Perceived control of disease				0.609
No	15 (31.91)	9 (19.15)	24 (51.06)
Yes	16 (34.04)	7 (14.89)	23 (48.94)

DISCUSSION

In this single‐center survey of 47 rheumatology patients, group‐based medical mistrust, as measured by the 12‐item GBMMS, was not associated with stated willingness to participate in a clinical trial. Instead, expectations of tangible clinical benefit and logistical considerations, including physician encouragement, transportation, and shared language with the clinician, were more prominent. Taken together, these findings suggest that among patients already connected to rheumatology services, practical and interpersonal determinants may overshadow broader, historically rooted mistrust when contemplating research participation. This pattern is consistent with work showing that structural and clinical barriers often dominate the trial enrollment process once a trial opportunity exists. ⁷ , ⁸

Medical mistrust is a well‐documented barrier to research participation among racially underrepresented populations. ¹⁴ Thompson et al first validated the GBMMS and linked higher scores to lower adherence to breast‐cancer screening among African American and Latina women, ⁵ a relationship subsequently replicated in African American men. ⁶ The GBMMS measures perceptions of the health care system toward one's ethnic group, not trust in a personal physician or specific research team, which may underlie our null association. Previous research shows that physician‐specific trust and direct clinician recommendations more strongly drive trial enrollment; in oncology, for example, explicit recommendations measurably increase participation, ¹⁰ , ¹³ and similar patterns likely occur in other specialties. Consistent with this distinction, our cohort's willingness rate (66%) contrasts with a southeastern US lupus cohort, in which only 28% of patients were willing to join drug trials, with no racial differences. In that study, willingness was associated with demographic, insurance, and disease‐related factors, and African Americans respondents had notable knowledge gaps about trial definitions. ¹⁵ Taken together, these data indicate that group‐based mistrust carries less weight when patients receive sustained specialty care and develop therapeutic clinician–patient relationships, ⁹ , ¹⁶ as in our cohort, where interpersonal trust appeared to support higher willingness.

Several factors may explain this finding. Interpersonal trust in the patient's clinician may supersede broader group‐based mistrust; patients often differentiate “the system” from “my doctor,” and strong therapeutic alliances can buffer skepticism, muting GBMMS effects in settings where clinician continuity is high. Restricted score variability in our sample may also have limited our ability to detect an association. Mean GBMMS values clustered in the lower‐moderate range (≈25 of 60), reducing statistical power to identify meaningful gradients. In addition, sample engagement bias may be at play: all participants were receiving ongoing rheumatology care, and individuals with the highest mistrust may never attend such clinics, creating a floor effect in which most respondents scored at the low end of mistrust, leaving little variation to detect meaningful differences.

More broadly, systematic reviews across conditions emphasize that structural and clinical barriers (eg, trial availability, eligibility criteria, and site access) account for the majority of nonenrollment, with attitudes playing a smaller role once an offer is made. This matches our own observations that patients tend to focus on practical barriers and clear benefits. ⁷ , ⁸ People who believe the trial will help them personally, and who understand how it works, are more likely to participate. This highlights the importance of explaining the potential benefits clearly and realistically. ¹³ , ¹⁷ , ¹⁸

In our cohort, neither education level nor perceived disease control was associated with willingness to participate (Table 4); however, previous work links lower educational attainment and lower income to reduced clinical trial participation, often through health literacy barriers, transportation limitations, and competing time or financial demands. Health literacy, in particular, shapes patients’ understanding of trial purpose, procedures, and risks or benefit, elements that are central to informed consent and decision‐making. ¹⁹ We did not collect income data, nor did we measure objective disease severity (only perceived control), so both socioeconomic status and clinical burden may operate as unmeasured confounders in our analysis. In addition, disease‐specific factors such as severity, duration, progression, and treatment history may shape patients’ perceived benefits from trial participation; those with advanced disease, multiple treatment failures, or substantial functional impact may view investigational options more favorably than those with early, stable, or well‐controlled illness. ¹⁵ Future studies should therefore capture granular socioeconomic variables (including income and health literacy) alongside objective clinical markers (disease activity indices, time since diagnosis, previous therapy failures, and functional or imaging measures) and examine how these interact with trust, logistical barriers, and physician endorsement to predict actual enrollment.

Consistent with national surveys, lack of reliable or affordable transport continues to limit clinical research access. ²⁰ Trials demanding frequent visits or specialized imaging impose particular burdens on rheumatology patients who may already face mobility challenges from pain or other factors. Decentralized and hybrid trial designs, such as home phlebotomy, tele‐visits, and ride‐share vouchers, are increasingly advocated to mitigate these barriers. ²¹

Participants who declined trials rated shared language with their physician highly, echoing evidence that language discordance suppresses enrollment of underrepresented ethnic groups. ³ Employing bilingual coordinators, translating consent documents, and using culturally tailored multimedia explanations therefore remain essential inclusivity tactics in rheumatology research.

These insights have clear implications for recruitment strategies. First, rheumatology trial teams should engage trusted clinicians in recruitment by having treating rheumatologists personally introduce studies and share concise, disease‐specific benefit messages during routine visits. Our results, together with an evidence that physicians’ recommendations influence patients’ decisions to participate and that professional rapport is a top motivator, ¹ , ¹⁰ suggest that encouragement from a trusted physician can substantially increase enrollment. Second, all recruitment materials should clearly communicate the potential for symptom relief and disease‐modifying effects, while also providing balanced and transparent information about associated risks. Emphasizing the possibility of slowing or halting disease progression, rather than only framing research as altruism, is likely to resonate, because disease control is a primary motivator. ¹ Third, investigators should provide logistical support for patients, such as arranging transportation or conducting home visits, to help address travel barriers, particularly for older adults. ¹¹ Finally, the language congruence of staff and materials must be ensured (eg, by providing bilingual coordinators or interpreters), as difficulty understanding trial information could deter non‐English speakers. By combining physician‐led invitations, transparent benefit messaging, and practical support for travel and communication, investigators can develop a more patient‐centered recruitment strategy that improves enrollment rates and promotes diversity and equity in rheumatology clinical research.

This exploratory study has several limitations. The small sample size limited statistical power to detect small‐to‐moderate effects, increasing the risk of a Type II error. Accordingly, the null association with GBMMS should be interpreted cautiously, as larger studies may reveal subtler influences. The single‐center design and inclusion of patients already in specialty care constrain generalizability to community or underserved settings with weaker clinician ties. We measured stated willingness rather than actual enrollment; intention does not always translate into behavior in research participation contexts. Finally, limited demographic diversity (eg, race and age) may have restricted our ability to evaluate subgroup effects.

Future research should include multicenter, adequately powered studies to examine how trust in one's physician, trial knowledge, and logistical support jointly influence actual enrollment, not just willingness, using designs that identify sequential barriers across the recruitment and enrollment process. Pragmatic trials of physician‐endorsed invitations, transportation assistance, and language‐concordant patient navigation, with enrollment and equity as key outcomes, are warranted in rheumatology. Embedding these strategies aligns with evidence that structural and clinical barriers are the primary limiting factors in trial participation and may yield more inclusive, representative enrollment.

In our single‐center cohort, group‐based medical mistrust did not predict patients’ self‐reported willingness to participate in a rheumatology clinical trial. Instead, the expectation of tangible disease improvement and direct encouragement by the treating rheumatologist emerged as the influential drivers of participation. These findings suggest that recruitment strategies focused on highlighting clear clinical benefits and leveraging trusted physician–patient relationships may be more effective than approaches focused solely on reducing mistrust. Confirmatory multisite studies and pragmatic intervention trials are needed to validate and operationalize these insights at scale.

AUTHOR CONTRIBUTIONS

All authors contributed to at least one of the following manuscript preparation roles: conceptualization AND/OR methodology, software, investigation, formal analysis, data curation, visualization, and validation AND drafting or reviewing/editing the final draft. As corresponding author, Dr Martinez Paulino confirms that all authors have provided the final approval of the version to be published and takes responsibility for the affirmations regarding article submission (eg, not under consideration by another journal), the integrity of the data presented, and the statements regarding compliance with institutional review board/Declaration of Helsinki requirements.

Supporting information

Disclosure form.