Rigor and equity in intervention study design in pediatric psychology: a focus on comparator conditions from diabetes research

Abstract

Objectives

To advance the science of health intervention research, pediatric psychologists must carefully design and conduct intervention research studies, including clinical trials. In contrast to guidance about scientific rigor in the selection of comparator groups in clinical trials, far less has been published on equity considerations in this process. The purpose of this paper is to review considerations for centering both equity and rigor in the study design decision, with a focus on the selection of comparator conditions for clinical trials of pediatric psychology interventions and propose potential solutions.

Methods

We reviewed existing guidance on (1) intervention study design with a focus on selection of comparator conditions from health psychology, medicine, and other similar fields, and (2) integration of both rigor and equity considerations into the design of intervention studies.

Results

We present a range of options for study design choices regarding comparator conditions and discuss potential benefits, limitations, and practical considerations for each type of comparator condition. Examples from behavioral intervention trials in pediatric type 1 diabetes or type 2 diabetes were used to illustrate how each comparator condition functions in practice. We developed a practical guide for researchers to consider both rigor and equity in decisions related to intervention study design and comparator condition selection.

Conclusions

The process of selecting an appropriate comparator condition is one aspect of study design that can advance both equity and scientific rigor in pediatric psychology intervention research.

“Do the best you can until you know better. Then when you know better, do better.”—Maya Angelou

The mission of pediatric psychology is to “actively promote the health and psychological well-being of all children, youth, and families” (Vision, n.d.). A primary contribution of pediatric psychologists is the development, testing, and implementation of empirically supported psychosocial and behavioral interventions for youth with health conditions. The success of this mission is based, in part, on the rigorous and ethical design and evaluation of interventions to ensure that treatments are supported by science. Intervention research occurs on a spectrum from early development studies through small proof-of-concept and pilot studies and ultimately, to fully powered efficacy trials (Czajkowski et al., 2015). Efficacy trials have largely followed the “gold standard” of randomized clinical trials (RCTs; Hariton & Locascio, 2018), comparing an intervention of interest to an alternative comparator condition: treatment as usual, a control condition, or a known efficacious treatment (Freedland et al., 2019).

Trial design shapes the value of the data and knowledge generated. It is critical to consider how we can improve trial design with greater attention to both scientific rigor (i.e., use of the scientific method to conduct well-controlled experiments) and equity (i.e., fair and just distribution of resources and access). Therefore, the current paper addresses one component of intervention research design: choosing a comparator condition in clinical trials and other intervention studies in pediatric psychology. The current paper aims to guide researchers to consider rigor and equity in their selection of (1) trial design (i.e., whether a comparator condition is necessary), and (2) the comparator condition itself (i.e., if a comparator condition is deemed necessary, how to address both rigor and equity in the selection process). Previous work has described processes for selecting comparator conditions with a focus on scientific rigor (Freedland et al., 2019), but equity considerations were not included. We advance the literature by considering both scientific rigor and equity in trial design, with a focus on the selection of comparator conditions.

We, as a field, must reckon with the legacy of inequity and harm built into design choices that maintain the status quo (Buchanan et al., 2021; Venkateswaran et al., 2023). For example, theory development, measure selection, and research participant selection have generally been conducted with homogenous samples that were predominantly non-Hispanic White and of higher socioeconomic status (Hays, 2016; Henrich et al., 2010). Further, patient, community, and public involvement in clinical trial development, including study design, has been limited (Gamble et al., 2014; Price et al., 2018). As a result of the confluence of these factors, our science and practice have been skewed and incomplete (Buchanan et al., 2021). To reverse this pattern, as a field, we must reconsider our methodologies to ensure that empirically supported treatments are designed with input from patients and communities to promote health equity, rather than contribute to disparities. We posit that behavioral intervention trials in pediatric psychology can have both rigor and equity. Intentionality in methodological decisions throughout all phases of research is crucial to ensure that design choices are not exclusionary, and do not lead to widening disparity gaps and maintaining inequities.

Ethical considerations are central to the intersection of scientific rigor and equity and provide a context for the selection of intervention study designs and comparator conditions. Three key principles are beneficence, justice, and equipoise (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979). Beneficence signifies the importance of doing no harm while maximizing the benefits of research and minimizing possible harm. It is important to recognize possible harm caused by clinical trial design decisions. For example, withholding care in a control group could reduce access to potentially beneficial treatments and limit generalizability in applied settings. Justice, in the context of research ethics, refers to treating people fairly and designing research that shares burdens and benefits equally. It is important to recognize how our research designs may disproportionally burden certain groups or how they may unfairly benefit some groups more than others. For example, some comparator conditions may have been developed from a Eurocentric perspective, and as a result, may be less beneficial or acceptable among people from other backgrounds. Therefore, assessing acceptability of comparator conditions is critical to upholding the principle of justice and ensuring that comparator conditions provide comparable levels of benefit across participants. Equipoise stipulates that there must be true uncertainty within a medical/clinical community about which trial arm is most likely to benefit participants (Ashcroft, 1999; Freedman, 1987), and the results of a study cannot consistently be predicted (Djulbegovic, 2009). That is, participants should only be enrolled in a study if there is reason to believe either arm might be beneficial to them. Random assignment is used to promote fairness and reduce systematic bias in who receives the intervention (Hariton & Locascio, 2018). In trial design and comparator condition selection, investigators must select intervention arms about which they maintain equipoise to conduct randomization without introducing risks to beneficence and justice. These principles apply to the process of selecting a comparator condition and every other stage of clinical trial design, conduct, and dissemination.

Therefore, we discuss how our field might design rigorous studies for pediatric populations that intentionally prioritize equity in addition to rigor. Given that other resources (e.g., Freedland et al., 2019) exist to guide more general comparator selections, we focus on outlining important considerations for comparator condition selection in pediatric psychology intervention research and provide a guide for investigators to “incorporate rigor and equity” in their design choices. We illustrate these points with examples from type 1 diabetes (T1D) and type 2 diabetes (T2D). As investigators whose work has largely been in pediatric diabetes behavioral intervention research, T1D and T2D are the focus of our examples because these are common chronic conditions with differences in prevalence: T1D is more prevalent among White youth and T2D is more prevalent among Black and Hispanic youth (Lawrence et al., 2021). Though stigma is common in both T1D and T2D (Abdoli et al., 2018), young people with T2D experience misplaced blame related to health behaviors and the association of an obesity diagnosis with T2D (Mackey et al., 2022). The much greater focus on T1D than T2D in pediatric psychology research (Hilliard et al., 2016) further illustrates inequities in these conditions. Although our examples are drawn from diabetes research, this paper serves as a guide for researchers to balance both rigor and equity broadly in research across populations and topics.

Considerations for study design

Pediatric psychology intervention research spans intervention development, proof-of-concept and pilot studies, and clinical trials to evaluate efficacy (Czajkowski et al., 2015). This continuum includes both randomized and non-randomized studies. Intervention research at early phases may use a range of study designs, either with or without comparator conditions, to answer questions related to intervention development, feasibility, and acceptability (Czajkowski et al., 2015). Trials to test clinical efficacy or effectiveness often use RCT designs, which have the benefit of being able to attribute differences in outcomes to the study intervention given the reduction in bias through randomization (Hariton & Locascio, 2018). However, there are potential drawbacks of RCT designs, including financial and time costs, threats to generalizability, and inadequate transparency of the different treatment conditions.

Guide to facilitate selection of comparator conditions

When selecting a comparator condition, considerations based on trial design, population, resources for the study, and study hypotheses and goals must be accounted for; no one choice is appropriate for all trial designs. Freedland and colleagues (2019) suggest starting with the comparator condition that best answers the study question while being aware of potential alternatives and limitations. To assist investigators, we designed a guide to help investigators systematically consider both rigor and equity in comparator condition selection for studies along the continuum of intervention research in pediatric psychology. Given the constraints of research studies and available resources (e.g., budgets, personnel, time), pragmatic considerations are also relevant to decisions about study design and selecting comparator conditions. Figure 1 illustrates five topics to consider in selecting a study design and comparator condition in relation to rigor and equity. The Trial Design section includes considerations related to inclusion/exclusion criteria and how participants are allocated to the different study arms. The Pragmatics and Resources section includes considerations related to cost, time, and burden for the study team and participants. The Intervention Features section includes considerations related to the way the intervention is delivered in terms of content, timing, and relevance. The Intervention Development section refers to those who contribute to the design of the intervention arms and study protocols. The Usual Care Quality section refers to considerations related to characteristics of usual care. For each decision-making subtopic (e.g., strictness of inclusion criteria, sample size, investigation team involvement), the figure indicates the degree of prioritization of equity (higher on the left side, lower on the right side). For example, less strict inclusion criteria prioritize equity more than stricter inclusion criteria by allowing more people the opportunity to participate. Figure 2 highlights how study design and comparator group choices may depend on the stage of trial design (Onken et al., 2014) . These stages of trial design are based, in part, on the ORBIT model of intervention development prior to efficacy testing (Czajkowski et al., 2015), and scientists are directed to this work for a comprehensive description of stages of trial design and development. With regard to Figure 2, for example, less strict inclusion criteria (greater prioritization of equity) may be appropriate in Stages I, III, IV, and V, but stricter criteria (middle/lower prioritization of equity) may be needed in Stage II. Together, both figures can help guide design decisions based on rigor and equity. Table 1 outlines options for comparator condition selection including benefits and burdens related to both rigor and equity, as well as pragmatic considerations. Additionally, Table 1 proposes investigator-driven solutions to improve equity for each type of comparator condition. Researchers can refer to the considerations listed in Table 1 to guide comparator condition selection in designing and executing intervention studies.

Figure 1.

Considerations for prioritizing equity in comparator condition selection.

Figure 2.

Considerations prioritizing equity in comparator condition selection related to stage of research.

Table 1.

Considerations and solutions for comparator condition selection.

Comparator Condition	Rigor Considerations		Equity Considerations		Pragmatic Considerations	Ways to Address Inequity
	Benefits	Burdens	Benefits	Burdens
No comparator—single-arm design	Maximizes feasibility and acceptability data Particularly useful for early-stage investigations or new interventions	Limited information on impact/efficacy No causal inference	Everyone receives the active intervention	Could increase disparities if not tailored Possibility of unintended harms	Less cost, time, and complexity for research team	Preliminary community engagement to develop meaningful interventions and minimize harm Contextualize findings as preliminary data for future efficacy/randomized trial
Non-randomized 2+ arm design May be called participant choice *Depending on different comparator arms, other considerations may apply	Provides initial data about appeal of different options May enhance feasibility and engagement	Risk for overinterpreting impact/efficacy No causal inference May introduce self-selection bias Potential threats to equipoise	Gives participants choice Potential to reduce power imbalance between researchers and participants	Possibility of unintended harms and widening inequity gap as an artifact of participant choice	High cost, time, and complexity for research team	Take steps to limit selection bias by obtaining data on participant choice Contextualize findings as preliminary data for future efficacy/randomized trial
Randomized to treatment as usual (TAU) *Defined as medical care which can differ based on the availability of allied health services	Permits causal inference Informative for future implementation or pragmatic trials May provide information about “improvement over current practice” (Mohr et al., 2009)	Unable to distinguish the mechanism of intervention/possible effects of attention TAU varies substantially (quality, content/inclusion of behavioral care) and may not be supervised/standardized—threat to internal/external validity and difficult to interpret differences Increased risk of sample bias (due to concerns with recruitment and retention)	Minimal time and burden to participants	TAU may not always reflect and may fall below standard of care for marginalized groups TAU may not benefit all people equitably, may worsen outcomes, and cause disparities People who need specific types of support may not get it (if not part of local TAU) Risk of sample bias may be amplified in marginalized groups	Less cost, time, and complexity for the research team Larger sample size may be needed if TAU is strong May be a barrier to recruitment/ enrollment due to the chance of not receiving intervention Risk of attrition if disappointed to receive TAU/no new intervention. May be more appropriate for implementation study	Explicitly let participants know they are not limited from accessing allied health services as needed outside study, document use of such services Screen for relevant needs (e.g., mental health, SDOH) and provide relevant referrals/resources Consider waitlist control design—TAU during assessment period, then offer intervention after data collection is complete Consider unequal randomization ratio (e.g., 2:1, 3:1) to maximize the amount of people receiving the intervention
Randomized to enhanced TAU *Defined as medical care as described above plus study-provided supplemental resources/information May be called “Standardized Care” or “Optimized Care” (Freedland et al., 2019)	Permits causal inference May permit examination of potential attention effects to strengthen causal inference (depending on intensity/interactions of enhancements)	Unable to distinguish intervention mechanisms TAU varies substantially (quality, content/inclusion of mental health/behavioral care) and may not be standardized—threat to internal/external validity and difficult to interpret differences Potential threats to equipoise (depending on format of enhanced TAU)	If enhancements are designed to address inequities, could reduce inequities by ensuring everyone receives some content beyond TAU Minimal time and burden to participants	TAU may not benefit all people equitably, may cause disparities People who need specific types of support may not get it (if not part of local TAU or enhanced content)	More cost, time, and complexity than TAU for the research team	Provide enhanced TAU to both groups Consider tailoring enhanced TAU materials to address inequities observed in TAU Consider unequal randomization ratio (e.g., 2:1, 3:1) to maximize the amount of people receiving the intervention
Waitlist control condition	Permits causal inference All the same as TAU, since this is essentially TAU followed by receiving the intervention at the end of the study	May overestimate intervention effects Can only be used if the condition being treated is not expected to improve on its own with time Same as TAU	Everyone receives the active intervention eventually		Cannot be utilized if the target of intervention is time-sensitive (e.g., within 6 months of diagnosis) Less cost, time, and complexity for the research team, given there is only one condition for which to monitor fidelity May require a smaller sample size Requires resources (e.g., interventionist time, intervention materials) later in the grant period, after follow-up data collection is complete	Consider unequal randomization ratio (e.g., 2:1, 3:1) to maximize the amount of people receiving the intervention
Randomized to placebo/attention control condition (mechanism-focused) May be called “sham” intervention	Permits causal inference Ability to distinguish mechanism because it is designed to have equivalent attention and allows for examination of potential attention effects	High risk for contamination between conditions Based on the placebo/sham treatment concept from drug trials, unclear relevance to behavioral interventions, which are more obvious to participants Potential threats to equipoise	Attention may benefit groups who have limited access to healthcare professionals If attention control is designed to address inequities, could be effective (for outcomes other than target trial outcomes)	Purposefully giving/designing an intervention without components hypothesized to be helpful for target outcome(s) may: Waste participant effort/time, especially for people who may already be overburdened/have limited resources Promote mistrust of system/research Widen inequity gaps Ethical concerns with using deceitful approaches for “sham” interventions that are dishonest—may damage trust	More cost, time, and complexity than TAU or single-arm for the research team Difficult to design an attention control that eliminates all components of behavioral intervention and provides equivalent amount of attention	Use community engagement strategies to develop attention control that may be valuable/of interest to participants and that has potential to reduce inequities (i.e., is not a waste of time, may have other benefits aside from trial target outcomes) Maintain ongoing, active engagement with community to also decrease mistrust Consider unequal randomization ratio (e.g., 2:1, 3:1) to maximize the amount of people receiving the intervention
Randomized to active comparator (other effective treatment) May be referred to as an alternative intervention or comparative effectiveness May include comparison with efficacious/existing intervention or with the same intervention at different content, modality, dose, or additional components (Freedland et al., 2019)	Permits causal inference May distinguish mechanism if interventions target different mechanisms May inform the effectiveness of different delivery formats	Risk of non-specific treatment factors (e.g., interventionist styles/ characteristics) may have a large impact on outcomes and obscure treatment-related differences May increase risk of type I error (may result in rejecting intervention due to non-specific effects) Potential threats to equipoise	Comparison group has efficacy data, so everyone gets something that is known or hypothesized to be helpful Could build trust by offering something with degree of evidence, if evidence is available for the specific marginalized group	If efficacy trial underlying comparator was based on homogenous/non-generalizable sample, raises questions about the pros and may lessen trust	High cost, time, complexity for the research team May require a larger sample to detect effect, which has implications for timeline, feasibility, study budget/resources	Select comparator condition with equity in mind by considering characteristics/diversity of sample used for original study/evidence and selecting comparator intervention that is appropriate for marginalized groups When not available, consider culturally adapting comparator condition to enhance relevance (balance with potential impact on rigor) Prioritize collecting data about each arm’s efficacy across groups to improve body of literature for future research Use community engagement methods to ensure trial intervention is culturally tailored/appropriate Consider not only content, but also modality, dose, components, graphics, interventionist characteristics, etc. when addressing cultural appropriateness and equity concerns Consider unequal randomization ratio (e.g., 2:1, 3:1) to maximize the amount of people receiving the intervention

Examples from pediatric diabetes research

Pediatric psychology intervention trials in T1D and T2D have used most of the intervention study designs and comparator condition options detailed below.¹ Given the authors’ collective expertise in diabetes research, the following examples were selected based on our knowledge of the field. In most cases, we use our own studies as examples, as we were involved in the decision-making process of selecting comparator conditions and thus, could explain our rationale for design choices.

Non-randomized, early-stage designs

Single-arm designs are typically used in early-stage research or development of novel interventions (ORBIT Phase 1). In these designs, all participants receive active treatment, prioritizing data on feasibility and acceptability over data on efficacy or causality (Holtz et al., 2021). Non-randomized 2+ arm designs, otherwise known as participant choice designs, are also appropriate for early-stage trials for new approaches, particularly to assess feasibility and acceptability (see Figure 2). In these designs, participants self-select a treatment arm rather than being randomly assigned to a condition. These types of non-randomized designs focus less on scientific rigor (e.g., isolating the effects of each condition) but can promote equity by giving participants agency and mitigating the power imbalance between researchers and participants (Table 1). Unintended consequences, however, may be possible if participant selection unintentionally leads to widening existing inequities (e.g., a particular group is more or less likely to self-select into a treatment group). See Table 1 for additional rigor and equity benefits and burdens, and pragmatic considerations for non-randomized designs. While non-randomized and early-stage trials represent an important aspect of the ORBIT model, they are beyond the scope of this paper and will not be discussed further.

Comparator conditions for randomized trials

TAU comparator condition

Also known as usual care, treatment as usual (TAU) is defined as routine treatments or medical care for the healthcare setting in which a trial takes place. In most intervention trials in pediatric psychology, participants in both arms continue receiving their routine medical care, regardless of the comparator condition. For RCT designs with TAU as the comparator condition, the intervention group receives routine medical care plus the intervention being studied, and the TAU group receives routine medical care with no additional intervention. Pragmatic and equity-related benefits of TAU include low burden for participants and research teams, as well as continuity of medical care (see Table 1). Rigor and equity-related concerns include a low potential for added benefit for participants in the TAU comparator. Further, variability in TAU quality across settings may contribute to inequities and difficulty ascertaining the true value of the intervention.

First STEPS was a behavioral intervention for parents of young children at new onset of T1D and used TAU (i.e., routine medical care) as the comparator (Hilliard et al., 2017). Due to the time-sensitive nature of the intervention (i.e., delivered in the first 1.5 years post-diagnosis) a waitlist comparator condition was not feasible. As this was the first RCT of the First STEPS intervention, the investigators had equipoise about the intervention’s impact and felt comparison to an alternative intervention was premature. First STEPS used a stepped-care intervention design, in which participants received up to three intervention components if parent depressive symptoms or child A1c exceeded clinically relevant thresholds at pre-specified times; to ensure sufficient power to evaluate efficacy of each step, participants were randomized 3:1 intervention to TAU. In addition to enhancing rigor, this approach also addressed equity as outlined in Table 1, as more participants received intervention contacts and resources above and beyond what they received in routine care.

Enhanced TAU comparator condition

Enhanced TAU (also known as enhanced usual care, standardized care, optimized care) is a version of the TAU comparator condition defined as TAU plus additional study-provided resources or information (Freedland et al., 2019). Equity-related benefits include minimal additional time and burden for participants and possible added value for participants, including the potential to address inequities of TAU alone (see Table 1). A potential downside is that without intentional tailoring, the inequities of TAU may persist (see Figure 1).

For example, in a study that evaluated a monetary reinforcement intervention for self-monitoring blood glucose (SMBG), participants were randomized to either the reinforcer intervention condition or enhanced TAU (Wagner et al., 2019). The enhanced TAU condition included an educational session with a diabetes care team, focused on establishing an SMBG regimen, assistance with setting alarm reminders, and weekly requests to upload SMBG data. The intervention group received the educational session plus earned monetary reinforcements for specific SMBG goals (e.g., meeting a minimum number of SMBG measurements within a window).

Waitlist comparator condition

A waitlist comparator condition (also known as delayed intervention or crossover design) involves participants receiving TAU initially, then receiving the intervention being evaluated in the trial. This approach is common in pediatric diabetes trials (O’Donnell et al., 2023; Straton et al., 2024). Waitlist conditions may be useful when investigators wish to examine pre-post change in outcomes or gain intervention feedback from a larger number of participants. They may also be useful when a ‘no-treatment’ comparison group would not be acceptable or ethical, or when investigators feel it would be more equitable to offer the intervention to all study participants at some point. In some cases, it involves waitlist participants receiving the full intervention (Straton et al., 2024) or an abbreviated version of the intervention after completion of the primary outcome time point. Benefits of waitlist conditions include that all participants are offered the intervention content, with timing differing based on randomization, and data are collected from more participants receiving the intervention than in a no-treatment comparison group (see Table 1). The group with the ‘delayed’ intervention serves as a no-treatment comparison group to allow for examination of intervention efficacy, before receiving the intervention. Downsides include feasibility challenges, such as for interventions targeting a specific event (e.g., intervention delivery during the transition between pediatric and adult care; Carreon et al., 2024), and logistical challenges, such as constraints on funds, resources, and time within the project period to deliver a second round of intervention. For example, ROUTE-T1D was a pilot RCT assessing the feasibility and acceptability of a behavioral telemedicine intervention to optimize use of a continuous glucose monitor (CGM) that used a waitlist control design (Straton et al., 2024). Participants were randomized 1:1 after enrollment to either the immediate or delayed intervention group. The intervention consisted of monthly individualized telemedicine sessions on behavioral topics related to CGM use. The immediate intervention group sessions occurred following baseline data collection, while the delayed intervention group received TAU for 6 months and then received the intervention. Waitlist controls also include providing intervention materials (e.g., a digital app) after the conclusion of the study.

Attention control comparator condition

Behavioral interventions often involve interactions with the study team, such as group sessions, individual phone calls, or time spent on an app/web module, which raises the question of whether any effects are due to the increased attention or can be attributed to the theoretically informed content of the intervention (Aycock et al., 2018). Thus, researchers may use attention control conditions to evaluate the effects of the “active” intervention by controlling for equal attention (i.e., contact with the research team). Attention control conditions may involve a similar format (e.g., group sessions) and delivery schedule as the intervention with different content. In diabetes research, a common attention control comparator is diabetes education without behavioral intervention (Basch et al., 2024; Jaser et al., 2020).

A risk of attention control conditions may be the creation of “busy work” for participants, which can create frustration, lack of interest, or higher rates of drop-out in those assigned to the attention control condition. Additionally, attention control conditions require more time for participants and research teams. On the other hand, attention control comparator conditions have potential for providing benefit to participants if the content is relevant or designed to address inequities (see Table 1). With intentional design and community engagement, this approach has the potential to reduce participant mistrust in research and eliminate the risk of “busy work” and minimize risk of drop-out or loss to follow-up.

For example, BREATHE-T1D was a pilot RCT that assessed the feasibility and acceptability of a group mindfulness-based intervention called Learning to BREATHE (L2B-T1D) tailored for adolescents with T1D with elevated negative affect (Basch et al., 2024). The comparator condition was a health education curriculum focused on T1D management, developed with input from interest holders, as suggested in Table 1. This attention control condition paralleled the intervention delivery format and included the social aspect of the group intervention, but without mindfulness training. Because the target population reported elevated negative affect, the research team prioritized providing social support and information about diabetes management that might still be beneficial to participants in some way. This example did not contain any of the “active ingredients” of the mindfulness-based intervention and illustrates both rigor and equity: this potentially beneficial comparator did not target key outcomes and mechanisms of the active intervention, addressed needs and preferences of the population, and therefore may offset risks of attention control conditions.

Active comparator condition

Active comparator conditions compare an intervention to either an existing efficacious intervention or to the same intervention with different content, modality, dose, or additional components (Freedland et al., 2019). These comparator conditions may be useful for comparative effectiveness studies or implementation-related studies that aim to determine the most important intervention components or the best intervention delivery method. Active comparator conditions may also be selected to address equity concerns related to ensuring all participants receive some type of intervention (see Table 1). For instance, BREATHE-T2D was a pilot RCT that assessed fidelity, feasibility, and acceptability of a group mindfulness-based intervention called Learning to BREATHE (L2B-T2D) in adolescents at-risk for T2D with elevated negative affect (Sanchez et al., 2024). In this trial, L2B-T2D was compared to an active comparator condition as well as an attention control condition. The active comparator condition was a cognitive behavioral therapy group, called the Blues Program, which has demonstrated empirical evidence of effectiveness in reducing depressive symptoms in adolescents (Marchand et al., 2010). Other examples of active comparators in behavioral diabetes research include comparing an in-person behavioral intervention to delivery of the same intervention via a digital format (Harris et al., 2015).

Rigor and equity in sample representation and measurement

While a full account is beyond the scope of this paper, it is important to acknowledge that ensuring representative inclusion in intervention research is essential to both scientific rigor and equity. Using homogenous samples can exacerbate health disparities and compromise the validity and generalizability of trial findings (National Academies of Sciences, Engineering, and Medicine; Policy and Global Affairs; Committee on Women in Science, Engineering, and Medicine; Committee on Improving the Representation of Women and Underrepresented Minorities in Clinical Trials and Research, 2022). Despite growing awareness and earlier policy efforts (e.g., NIH Revitalization Act of 1993; Piantadosi, 1995; The All of Us Research Program Investigators, 2019), many trials still overlook these considerations. Measurement tools often assume fairness across all participants, yet few studies assess or consider measurement bias or differential validity across participant characteristics (Helms, 2006; Modi et al., 2024; Straton et al., 2025). This can distort outcomes independent of true effects, whether in questionnaires or biomarkers like HbA1c (Hamdan et al., 2016).

Even when samples include people from multiple groups, subgroup analyses are frequently underpowered. To rigorously assess treatment effects across groups, researchers should consider targeted inclusion and analytic strategies—such as oversampling, modeling subgroup interactions, or adjusting for group status—to ensure valid inferences (Sies et al., 2019; Wang et al., 2021). Person-centered analytic approaches offer promise for understanding heterogeneity of intervention outcomes and informing tailored interventions (Gates et al., 2023; Pyatak et al., 2024).

Discussion

In this paper, we highlight both rigor and equity considerations in decision-making about intervention study design and make suggestions for selecting comparator conditions for clinical trials in pediatric psychology, adding to the literature on guidelines for rigorous conduct of clinical trials. At the same time, it is important to be explicit that bias occurs at many stages of research (e.g., investigator grant submissions, reviews, awards; participant recruitment, retention; study analysis, and dissemination) and great care is needed to enhance rigor and move toward equity (Buchanan et al., 2021; Nguyen et al., 2023; Niranjan et al., 2020; Venkateswaran et al., 2023). Further, to combat health inequities, it is crucial to ensure equitable opportunity for all pediatric populations to participate in intervention studies that may benefit them. As such, the scientific community must remain committed to minimizing harmful bias in clinical trials to combat health disparities, pursue equity, and advance rigor in clinical research. The considerations and suggestions for comparator condition selection in intervention research in pediatric psychology in this paper with a specific focus on rigor and equity represent one step toward this aim.

The selection of a comparator condition in intervention research design influences what research questions can be answered, how findings are interpreted, and the potential impact on clinical care, which can widen or reduce health inequities. Therefore, this selection should be made with careful consideration of the scientific approach, stage of trial design, health equity, and input from interest holders. The purpose of the current paper is not to criticize existing science or minimize the importance of rigorous conduct of clinical trials. Rather, we aim to encourage the field to consider trial design and conduct decisions with an intentional focus on equity as well. The guide for considering equity as well as scientific rigor in the selection of comparator conditions, and the considerations for measurement and statistical approaches, can help investigators in pediatric psychology advance science. Notably, the use of tools such as the ORBIT model (Czajkowski et al., 2015) allows behavioral scientists to be thoughtful about issues of rigor and equity in early stages of intervention development and refinement, and we encourage scientists to consider the current guide in the context of tools such as the ORBIT model as well.

Ideally, clinical trials would enroll participants representing identities that reflect the population of interest and use rigorous methodology to design studies optimal to support efficacious interventions and reduce health inequities. However, many barriers to these ideal situations exist, including recruitment of underrepresented populations who report a lack of trust in clinical research and lack of access to clinical trial opportunities (Heller et al., 2014), as well as the cost and time burden involved in evaluating and implementing efficacy and effectiveness trials (Jaramillo et al., 2023). Therefore, decisions about trial design and reporting often prioritize rigor over equity rather than balancing both, resulting in the widening of health inequities. We hope researchers, reviewers, and funders benefit from the resources in this paper on how to balance both considerations and allow for trial design decisions that prioritize equity as well as rigor. Our intention is not to suggest a dichotomy between rigor and equity; rather, we encourage the field of pediatric psychology to push toward science that values both.

It is important to recognize the limitations of the present work. First, the current paper focused solely on the selection of comparator conditions, just one aspect of intervention research design. There are many other important considerations in research methodology, including pragmatism, participant compensation, community collaborations, and measurement and use of demographic data (Call et al., 2023). Although these considerations were beyond the scope of the present work, researchers are encouraged to take them into consideration when making decisions about study design and conduct. Additionally, given the expertise of the authors, we provided examples of comparator conditions from pediatric T1D and T2D behavioral intervention trials. Focusing on examples from one domain of pediatric psychology may limit the generalizability of some of these recommendations. For example, it is possible that TAU comparator conditions may not be applicable in research conducted in rarer chronic conditions or those with less structured standards of care. However, our focus on one health area also serves as a strength, as we demonstrate how different comparator conditions operate within similar research populations.

Future directions

As the scientific community has continued to promote RCTs as ideal and essential for determining empirical support, researchers may use the guide presented in the current paper to consider available options for comparator conditions to address both rigor and equity when designing intervention studies. Future research can and should be both rigorous and equitable. While the present paper focuses on just one aspect of trial design, namely selecting comparator conditions, there is a need within pediatric psychology to continue exploring avenues and strategies to enhance rigor and equity throughout the research process.

Conclusions

We encourage behavioral scientists to prioritize equity along with scientific rigor in clinical trial design and all other research decisions to maximize benefits for participants. We urge researchers, reviewers, and funders to be open to the incorporation of decisions that promote equity in trial design. While selection of a comparator condition is just one piece of trial design, concrete steps toward equity in research are needed throughout the research process to minimize harm and push our field toward improved outcomes for all children, adolescents, and families.

Author contributions

Francesca Lupini (Conceptualization [equal], Writing—original draft [equal], Writing—review & editing [equal]), Marisa E. Hilliard (Conceptualization [equal], Visualization [equal], Writing—original draft [equal], Writing—review & editing [equal]), Idia B. Thurston (Conceptualization [equal], Visualization [equal], Writing—original draft [equal], Writing—review & editing [equal]), Sarah Jaser (Conceptualization [equal], Visualization [equal], Writing—original draft [equal], Writing—review & editing [equal]), Samantha Carreon (Conceptualization [equal], Writing—original draft [equal], Writing—review & editing [equal]), Ana Gutierrez-Colina (Conceptualization [equal], Writing—original draft [equal], Writing—review & editing [equal]), Randi Streisand (Conceptualization [equal], Writing—original draft [equal], Writing—review & editing [equal]), Kristoffer S. Berlin (Conceptualization [equal], Writing—original draft [equal], Writing—review & editing [equal]), and Eleanor Mackey (Conceptualization [equal], Writing—original draft [equal], Writing—review & editing [equal])

Funding

This work was supported by grants 1R01DK102561 (National Institute of Diabetes and Digestive and Kidney Diseases) to M.E.H.; R01DK119246 (National Institute of Diabetes and Digestive and Kidney Diseases) and R01DK119246-03S1 (National Institute of Diabetes and Digestive and Kidney Diseases) to M.E.H. and S.A.C.; R01DK121316 (National Institute of Diabetes and Digestive and Kidney Diseases) to S.S.J.; R01DK132557-01S2 (National Institute of Diabetes and Digestive and Kidney Diseases) to A.M.G.-C.; R01DK131026 (National Institute of Diabetes and Digestive and Kidney Diseases) and 1R01DK102561 (National Institute of Diabetes and Digestive and Kidney Diseases) to R.S.; and 5R34AT011035 (National Center for Complementary and Integrative Health) and 1U01AT011008/5R01AT011008 (National Center for Complementary and Integrative Health) to E.R.M.

Conflicts of interest: The authors have no conflicts of interest to disclose.

Data availability

No new data were generated or analyzed in support of this research.