Balancing best practice and reality in behavioral intervention development: A survey of principal investigators funded by the National Institutes of Health

Preliminary studies play a prominent role in the development of large-scale behavioral interventions. Though recommendations exist to guide the execution and interpretation of preliminary studies, these assume optimal scenarios which may clash with realities faced by researchers. The purpose of this study was to explore how principal investigators (PIs) balance expectations when conducting preliminary studies. We surveyed PIs funded by the National Institutes of Health to conduct preliminary behavioral interventions between 2000 and 2020. Four hundred thirty-one PIs (19% response rate) completed the survey (November 2021 to January 2022, 72% female, mean 21 years post-terminal degree). Most PIs were aware of translational models and believed preliminary studies should precede larger trials but also believed a single preliminary study provided sufficient evidence to scale. When asked about the relative importance of preliminary efficacy (i.e. changes in outcomes) and feasibility (i.e. recruitment, acceptance/adherence) responses varied. Preliminary studies were perceived as necessary to successfully compete for research funding, but among PIs who had peer-reviewed federal-level grants applications (n = 343 [80%]), responses varied about what should be presented to secure funding. Confusion surrounding the definition of a successful, informative preliminary study poses a significant challenge when developing behavior interventions. This may be due to a mismatch between expectations surrounding preliminary studies and the realities of the research enterprise in which they are conducted. To improve the quality of preliminary studies and advance the field of behavioral interventions, additional funding opportunities, more transparent criteria in grant reviews, and additional training for grant reviewers are suggested.

Consolidated guidance for behavioral intervention pilot and feasibility studies

BACKGROUND

In the behavioral sciences, conducting pilot and/or feasibility studies (PFS) is a key step that provides essential information used to inform the design, conduct, and implementation of a larger-scale trial. There are more than 160 published guidelines, reporting checklists, frameworks, and recommendations related to PFS. All of these publications offer some form of guidance on PFS, but many focus on one or a few topics. This makes it difficult for researchers wanting to gain a broader understanding of all the relevant and important aspects of PFS and requires them to seek out multiple sources of information, which increases the risk of missing key considerations to incorporate into their PFS. The purpose of this study was to develop a consolidated set of considerations for the design, conduct, implementation, and reporting of PFS for interventions conducted in the behavioral sciences.

METHODS

To develop this consolidation, we undertook a review of the published guidance on PFS in combination with expert consensus (via a Delphi study) from the authors who wrote such guidance to inform the identified considerations. A total of 161 PFS-related guidelines, checklists, frameworks, and recommendations were identified via a review of recently published behavioral intervention PFS and backward/forward citation tracking of a well-known PFS literature (e.g., CONSORT Ext. for PFS). Authors of all 161 PFS publications were invited to complete a three-round Delphi survey, which was used to guide the creation of a consolidated list of considerations to guide the design, conduct, and reporting of PFS conducted by researchers in the behavioral sciences.

RESULTS

A total of 496 authors were invited to take part in the three-round Delphi survey (round 1, N = 46; round 2, N = 24; round 3, N = 22). A set of twenty considerations, broadly categorized into six themes (intervention design, study design, conduct of trial, implementation of intervention, statistical analysis, and reporting) were generated from a review of the 161 PFS-related publications as well as a synthesis of feedback from the three-round Delphi process. These 20 considerations are presented alongside a supporting narrative for each consideration as well as a crosswalk of all 161 publications aligned with each consideration for further reading.

CONCLUSION

We leveraged expert opinion from researchers who have published PFS-related guidelines, checklists, frameworks, and recommendations on a wide range of topics and distilled this knowledge into a valuable and universal resource for researchers conducting PFS. Researchers may use these considerations alongside the previously published literature to guide decisions about all aspects of PFS, with the hope of creating and disseminating interventions with broad public health impact.

The Day-Level Association Between Child Care Attendance and 24-Hour Movement Behaviors in Preschool-Aged Children

BACKGROUND

Twenty-four hour movement behaviors (ie, physical activity [PA], screen time [ST], and sleep) are associated with children's health outcomes. Identifying day-level contextual factors, such as child care, that positively influence children's movement behaviors may help identify potential intervention targets, like improving access to child care programs. This study aimed to examine the between- and within-person effects of child care on preschoolers' 24-hour movement behaviors.

METHODS

Children (N = 74, 4.7 [0.9] y, 48.9% girls, 63.3% White) wore an Axivity AX3 accelerometer on their nondominant wrist 24 hours per day for 14 days to measure PA and sleep. Parents completed surveys each night about their child's ST and child care attendance that day. Linear mixed effects models predicted day-level 24-hour movement behaviors from hours spent in child care.

RESULTS

Children spent an average of 5.0 (2.9) hours per day in child care. For every additional hour of child care above their average, children had 0.3 hours (95% CI, -0.3 to -0.2) less ST that day. Between-person effects showed that compared with children who attended fewer overall hours of child care, children who attended more hours had less overall ST (B = -0.2 h; 95% CI, -0.4 to 0.0). Child care was not significantly associated with PA or sleep.

CONCLUSIONS

Child care attendance was not associated with 24-hour PA or sleep; however, it was associated with less ST. More research utilizing objective measures of ST and more robust measures of daily schedules or structure is necessary to better understand how existing infrastructure may influence preschool-aged children's 24-hour movement behaviors. In addition, future research should consider how access to child care may influence child care attendance.

Understanding Accelerated Summer Body Mass Index Gain by Tracking Changes in Children's Height, Weight, and Body Mass Index Throughout the Year

Background: Drivers of summer body mass index (BMI) gain in children remain unclear. The Circadian and Circannual Rhythm Model (CCRM) posits summer BMI gain is biologically driven, while the Structured Days Hypothesis (SDH) proposes it is driven by reduced structure. Objectives: Identify the mechanisms driving children's seasonal BMI gain through the CCRM and SDH. Methods: Children's (N = 147, mean age = 8.2 years) height and weight were measured monthly during the school year, and once in summer (July-August). BMI z-score (zBMI) was calculated using CDC growth charts. Behaviors were measured once per season. Mixed methods regression estimated monthly percent change in children's height (%HΔ), weight (%WΔ), and monthly zBMI for school year vs. summer vacation, seasonally, and during school months with no breaks vs. school months with a break ≥1 week. Results: School year vs. summer vacation analyses showed accelerations in children's %WΔ (Δ = 0.9, Standard Error (SE) = 0.1 vs. Δ = 1.4, SE = 0.1) and zBMI (Δ = -0.01, SE = 0.01 vs. Δ = 0.04, SE = 0.3) during summer vacation, but %HΔ remained relatively constant during summer vacation compared with school (Δ = 0.3, SE = 0.0 vs. Δ = 0.4, SE = 0.1). Seasonal analyses showed summer had the greatest %WΔ (Δ = 1.8, SE = 0.4) and zBMI change (Δ = 0.05, SE = 0.03) while %HΔ was relatively constant across seasons. Compared with school months without a break, months with a break showed higher %WΔ (Δ = 0.7, SE = 0.1 vs. Δ = 1.6, SE = 0.2) and zBMI change (Δ = -0.03, SE = 0.01 vs. Δ = 0.04, SE = 0.01), but %HΔ was constant (Δ = 0.4, SE = 0.0 vs. Δ = 0.3, SE = 0.1). Fluctuations in sleep timing and screen time may explain these changes. Conclusions: Evidence for both the CCRM and SDH was identified but the SDH may more fully explain BMI gain. Interventions targeting consistent sleep and reduced screen time during breaks from school may be warranted no matter the season.

Are the Risk of Generalizability Biases Generalizable? A Meta-Epidemiological Study

Background Preliminary studies (e.g., pilot/feasibility studies) can result in misleading evidence that an intervention is ready to be evaluated in a large-scale trial when it is not. Risk of Generalizability Biases (RGBs, a set of external validity biases) represent study features that influence estimates of effectiveness, often inflating estimates in preliminary studies which are not replicated in larger-scale trials. While RGBs have been empirically established in interventions targeting obesity, the extent to which RGBs generalize to other health areas is unknown. Understanding the relevance of RGBs across health behavior intervention research can inform organized efforts to reduce their prevalence. Purpose The purpose of our study was to examine whether RGBs generalize outside of obesity-related interventions. Methods A systematic review identified health behavior interventions across four behaviors unrelated to obesity that follow a similar intervention development framework of preliminary studies informing larger-scale trials (i.e., tobacco use disorder, alcohol use disorder, interpersonal violence, and behaviors related to increased sexually transmitted infections). To be included, published interventions had to be tested in a preliminary study followed by testing in a larger trial (the two studies thus comprising a study pair). We extracted health-related outcomes and coded the presence/absence of RGBs. We used meta-regression models to estimate the impact of RGBs on the change in standardized mean difference (ΔSMD) between the preliminary study and larger trial. Results We identified sixty-nine study pairs, of which forty-seven were eligible for inclusion in the analysis (k = 156 effects), with RGBs identified for each behavior. For pairs where the RGB was present in the preliminary study but removed in the larger trial the treatment effect decreased by an average of ΔSMD=-0.38 (range - 0.69 to -0.21). This provides evidence of larger drop in effectiveness for studies containing RGBs relative to study pairs with no RGBs present (treatment effect decreased by an average of ΔSMD =-0.24, range - 0.19 to -0.27). Conclusion RGBs may be associated with higher effect estimates across diverse areas of health intervention research. These findings suggest commonalities shared across health behavior intervention fields may facilitate introduction of RGBs within preliminary studies, rather than RGBs being isolated to a single health behavior field.

A Device Agnostic Approach to Predict Children's Activity from Consumer Wearable Accelerometer Data: A Proof-of-Concept Study

INTRODUCTION

This study examined the potential of a device agnostic approach for predicting physical activity from consumer wearable accelerometry compared with a research-grade accelerometry.

METHODS

Seventy-five 5- to 12-year-olds (58% male, 63% White) participated in a 60-min protocol. Children wore wrist-placed consumer wearables (Apple Watch Series 7 and Garmin Vivoactive 4) and a research-grade device (ActiGraph GT9X) concurrently with an indirect calorimeter (COSMED K5). Activity intensities (i.e., inactive, light, moderate-to-vigorous physical activity) were estimated via indirect calorimetry (criterion), and the Hildebrand thresholds were applied to the raw accelerometer data from the consumer wearables and research-grade device. Epoch-by-epoch (e.g., weighted sensitivity, specificity) and discrepancy (e.g., mean bias, absolute error) analyses evaluated agreement between accelerometry-derived and criterion estimates. Equivalence testing evaluated the equivalence of estimates produced by the consumer wearables and ActiGraph.

RESULTS

Estimates produced by the raw accelerometry data from ActiGraph, Apple, and Garmin produced similar criterion agreement with weighted sensitivity = 68.2% (95% confidence interval (CI), 67.1%-69.3%), 73.0% (95% CI, 71.8%-74.3%), and 66.6% (95% CI, 65.7%-67.5%), respectively, and weighted specificity = 84.4% (95% CI, 83.6%-85.2%), 82.0% (95% CI, 80.6%-83.4%), and 75.3% (95% CI, 74.7%-75.9%), respectively. Apple Watch produced the lowest mean bias (inactive, -4.0 ± 4.5; light activity, 2.1 ± 4.0) and absolute error (inactive, 4.9 ± 3.4; light activity, 3.6 ± 2.7) for inactive and light physical activity minutes. For moderate-to-vigorous physical activity, ActiGraph produced the lowest mean bias (1.0 ± 2.9) and absolute error (2.8 ± 2.4). No ActiGraph and consumer wearable device estimates were statistically significantly equivalent.

CONCLUSIONS

Raw accelerometry estimated inactive and light activity from wrist-placed consumer wearables performed similarly to, if not better than, a research-grade device, when compared with indirect calorimetry. This proof-of-concept study highlights the potential of device-agnostic methods for quantifying physical activity intensity via consumer wearables.

Expert Perspectives on Pilot and Feasibility Studies: A Delphi Study and Consolidation of Considerations for Behavioral Interventions

Background

In the behavioral sciences, conducting pilot and/or feasibility studies (PFS) is a key step that provides essential information used to inform the design, conduct, and implementation of a larger-scale trial. There are more than 160 published guidelines, reporting checklists, frameworks, and recommendations related to PFS. All of these publications offer some form of guidance on PFS, but many focus on one or a few topics. This makes it difficult for researchers wanting to gain a broader understanding of all the relevant and important aspects of PFS and requires them to seek out multiple sources of information, which increases the risk of missing key considerations to incorporate into their PFS. The purpose of this study was to develop a consolidated set of considerations for the design, conduct, implementation, and reporting of PFS for interventions conducted in the behavioral sciences.

Methods

To develop this consolidation, we undertook a review of the published guidance on PFS in combination with expert consensus (via a Delphi study) from the authors who wrote such guidance to inform the identified considerations. A total of 161 PFS-related guidelines, checklists, frameworks, and recommendations were identified via a review of recently published behavioral intervention PFS and backward/forward citation tracking of well-know PFS literature (e.g., CONSORT Ext. for PFS). Authors of all 161 PFS publications were invited to complete a three-round Delphi survey, which was used to guide the creation of a consolidated list of considerations to guide the design, conduct, and reporting of PFS conducted by researchers in the behavioral sciences.

Results

A total of 496 authors were invited to take part in the Delphi survey, 50 (10.1%) of which completed all three rounds, representing 60 (37.3%) of the 161 identified PFS-related guidelines, checklists, frameworks, and recommendations. A set of twenty considerations, broadly categorized into six themes (Intervention Design, Study Design, Conduct of Trial, Implementation of Intervention, Statistical Analysis and Reporting) were generated from a review of the 161 PFS-related publications as well as a synthesis of feedback from the three-round Delphi process. These 20 considerations are presented alongside a supporting narrative for each consideration as well as a crosswalk of all 161 publications aligned with each consideration for further reading.

Conclusion

We leveraged expert opinion from researchers who have published PFS-related guidelines, checklists, frameworks, and recommendations on a wide range of topics and distilled this knowledge into a valuable and universal resource for researchers conducting PFS. Researchers may use these considerations alongside the previously published literature to guide decisions about all aspects of PFS, with the hope of creating and disseminating interventions with broad public health impact.

Are parent-reported sleep logs essential? A comparison of three approaches to guide open source accelerometry-based nocturnal sleep processing in children

We examined the comparability of children's nocturnal sleep estimates using accelerometry data, processed with and without a sleep log. In a secondary analysis, we evaluated factors associated with disagreement between processing approaches. Children (n = 722, age 5-12 years) wore a wrist-based accelerometer for 14 days during Autumn 2020, Spring 2021, and/or Summer 2021. Outcomes included sleep period, duration, wake after sleep onset (WASO), and timing (onset, midpoint, waketime). Parents completed surveys including children's nightly bed/wake time. Data were processed with parent-reported bed/wake time (sleep log), the Heuristic algorithm looking at Distribution of Change in Z-Angle (HDCZA) algorithm (no log), and an 8 p.m.-8 a.m. window (generic log) using the R-package 'GGIR' (version 2.6-4). Mean/absolute bias and limits of agreement were calculated and visualised with Bland-Altman plots. Associations between child, home, and survey characteristics and disagreement were examined with tobit regression. Just over half of nights demonstrated no difference in sleep period between sleep log and no log approaches. Among all nights, the sleep log approach produced longer sleep periods (9.3 min; absolute mean bias [AMB] = 28.0 min), shorter duration (1.4 min; AMB = 14.0 min), greater WASO (11.0 min; AMB = 15.4 min), and earlier onset (13.4 min; AMB = 17.4 min), midpoint (8.8 min; AMB = 15.3 min), and waketime (3.9 min; AMB = 14.8 min) than no log. Factors associated with discrepancies included smartphone ownership, bedroom screens, nontraditional parent work schedule, and completion on weekend/summer nights (range = 0.4-10.2 min). The generic log resulted in greater AMB among sleep outcomes. Small mean differences were observed between nights with and without a sleep log. Discrepancies existed on weekends, in summer, and for children with smartphones and screens in the bedroom.

What does it mean to use the mean? The impact of different data handling strategies on the proportion of children classified as meeting 24-hr movement guidelines and associations with overweight and obesity

Background

Despite the widespread endorsement of 24-hour movement guidelines (physical activity, sleep, screentime) for youth, no standardized processes for categorizing guideline achievement exists. The purpose of this study was to illustrate the impact of different data handling strategies on the proportion of children meeting 24-hour movement guidelines (24hrG) and associations with overweight and obesity.

Methods

A subset of 524 children (ages 5-12yrs) with complete 24-hour behavior measures on at least 10 days was used to compare the impact of data handling strategies on estimates of meeting 24hrG. Physical activity and sleep were measured via accelerometry. Screentime was measured via parent self-report. Comparison of meeting 24hrG were made using 1) average of behaviors across all days (AVG-24hr), 2) classifying each day and evaluating the percentage meeting 24hrG from 10-100% of their measured days (DAYS-24hr), and 3) the average of a random sample of 4 days across 10 iterations (RAND-24hr). A second subset of children (N=475) with height and weight data was used to explore the influence of each data handling strategy on children meeting guidelines and the odds of overweight/obesity via logistic regression.

Results

Classification for AVG-24hr resulted in 14.7% of participants meeting 24hrG. Classification for DAYS-24hr resulted in 63.5% meeting 24hrG on 10% of measured days with <1% meeting 24hrG on 100% of days. Classification for RAND-24hr resulted in 15.9% of participants meeting 24hrG. Across 10 iterations, 63.6% of participants never met 24hrG regardless of the days sampled, 3.4% always met 24hrG, with the remaining 33.0% classified as meeting 24hrG for at least one of the 10 random iterations of days. Using AVG-24hr as a strategy, meeting all three guidelines associated with lower odds of having overweight obesity (OR=0.38, p<0.05). The RAND-24hr strategy produced a range of odds from 0.27 to 0.56. Using the criteria of needing to meet 24hrG on 100% of days, meeting all three guidelines associated with the lowest odds of having overweight and obesity as well (OR=0.04, p<0.05).

Conclusions

Varying estimates of meeting the 24hrG and the odds of overweight and obesity results from different data handling strategies and days sampled.

Accuracy and Precision of Opportunistic Measures of Body Composition from the Tanita DC-430U

Background: It is essential to quantify the accuracy and precision of bioelectrical impedance (BIA)-estimated percent body fat (%BF) to better interpret community-based research findings that utilize opportunistic measures. Methods: Study 1 measured the accuracy of a new dual-frequency foot-to-foot BIA device (Tanita DC-430U) compared with dual-energy X-ray absorptiometry (DXA) among healthy elementary school-aged children (N = 50). Study 2 examined the precision of BIA %BF estimates within and between days among children and adults (N = 38). Results: Regarding accuracy, Tanita DC-430U underestimated %BF by 8.0 percentage points compared with DXA (20.6% vs. 28.5%), but correctly ranked children in terms of %BF. Differences in %BF between BIA and DXA were driven by lower BIA-estimated fat mass (7.8 kg vs. 9.9 kg, p < 0.05) and higher BIA-estimated fat-free mass (25.3 kg vs. 24.1 kg, p < 0.05). The absolute agreement between BIA and DXA for estimated %BF was moderate (concordance correlation coefficients = 0.53). Regarding precision, measures taken at the same time, but on different days (root mean square standard deviation [RMSD] = 0.42-0.74) were more precise than the measures taken at different times within a single day (RMSD = 1.04-1.10). Conclusion: The Tanita DC-430U substantially underestimated %BF compared with DXA, highlighting the need to assess accuracy of new BIA devices when they are introduced to the market. Opportunistic measures of %BF estimates were most precise when taken at consistent times and in the morning, but may be utilized throughout the day with an understanding of within- and between-day variability.

Use of guidelines, checklists, frameworks, and recommendations in behavioral intervention preliminary studies and associations with reporting comprehensiveness: a scoping bibliometric review

BACKGROUND

Guidelines, checklists, frameworks, and recommendations (GCFRs) related to preliminary studies serve as essential resources to assist behavioral intervention researchers in reporting findings from preliminary studies, but their impact on preliminary study reporting comprehensiveness is unknown. The purpose of this study was to conduct a scoping bibliometric review of recently published preliminary behavioral-focused intervention studies to (1) examine the prevalence of GCFR usage and (2) determine the associations between GCFR usage and reporting feasibility-related characteristics.

METHODS

A systematic search was conducted for preliminary studies of behavioral-focused interventions published between 2018 and 2020. Studies were limited to the top 25 journals publishing behavioral-focused interventions, text mined to identify usage of GCFRs, and categorized as either not citing GCFRs or citing ≥ 2 GCFRs (Citers). A random sample of non-Citers was text mined to identify studies which cited other preliminary studies that cited GCFRs (Indirect Citers) and those that did not (Never Citers). The presence/absence of feasibility-related characteristics was compared between Citers, Indirect Citers, and Never Citers via univariate logistic regression.

RESULTS

Studies (n = 4143) were identified, and 1316 were text mined to identify GCFR usage (n = 167 Citers). A random sample of 200 studies not citing a GCFR were selected and categorized into Indirect Citers (n = 71) and Never Citers (n = 129). Compared to Never Citers, Citers had higher odds of reporting retention, acceptability, adverse events, compliance, cost, data collection feasibility, and treatment fidelity (ORrange = 2.62-14.15, p < 0.005). Citers also had higher odds of mentioning feasibility in purpose statements, providing progression criteria, framing feasibility as the primary outcome, and mentioning feasibility in conclusions (ORrange = 6.31-17.04, p < 0.005) and lower odds of mentioning efficacy in purpose statements, testing for efficacy, mentioning efficacy in conclusions, and suggesting future testing (ORrange = 0.13-0.54, p < 0.05). Indirect Citers had higher odds of reporting acceptability and treatment fidelity (ORrange = 2.12-2.39, p < 0.05) but lower odds of testing for efficacy (OR = 0.36, p < 0.05) compared to Never Citers.

CONCLUSION

The citation of GCFRs is associated with greater reporting of feasibility-related characteristics in preliminary studies of behavioral-focused interventions. Researchers are encouraged to use and cite literature that provides guidance on design, implementation, analysis, and reporting to improve the comprehensiveness of reporting for preliminary studies.

Evaluation of a device-agnostic approach to predict sleep from raw accelerometry data collected by Apple Watch Series 7, Garmin Vivoactive 4, and ActiGraph GT9X Link in children with sleep disruptions

GOAL AND AIMS

Evaluate the performance of a sleep scoring algorithm applied to raw accelerometry data collected from research-grade and consumer wearable actigraphy devices against polysomnography.

FOCUS METHOD/TECHNOLOGY

Automatic sleep/wake classification using the Sadeh algorithm applied to raw accelerometry data from ActiGraph GT9X Link, Apple Watch Series 7, and Garmin Vivoactive 4.

REFERENCE METHOD/TECHNOLOGY

Standard manual PSG sleep scoring.

SAMPLE

Fifty children with disrupted sleep (M = 8.5 years, range = 5-12 years, 42% Black, 64% male).

DESIGN

Participants underwent to single night lab polysomnography while wearing ActiGraph, Apple, and Garmin devices.

CORE ANALYTICS

Discrepancy and epoch-by-epoch analyses for sleep/wake classification (devices vs. polysomnography).

ADDITIONAL ANALYTICS AND EXPLORATORY ANALYSES

Equivalence testing for sleep/wake classification (research-grade actigraphy vs. commercial devices).

CORE OUTCOMES

Compared to polysomnography, accuracy, sensitivity, and specificity were 85.5, 87.4, and 76.8, respectively, for Actigraph; 83.7, 85.2, and 75.8, respectively, for Garmin; and 84.6, 86.2, and 77.2, respectively, for Apple. The magnitude and trend of bias for total sleep time, sleep efficiency, sleep onset latency, and wake after sleep were similar between the research and consumer wearable devices.

IMPORTANT ADDITIONAL OUTCOMES

Equivalence testing indicated that total sleep time and sleep efficiency estimates from the research and consumer wearable devices were statistically significantly equivalent.

CORE CONCLUSION

This study demonstrates that raw acceleration data from consumer wearable devices has the potential to be harnessed to predict sleep in children. While further work is needed, this strategy could overcome current limitations related to proprietary algorithms for predicting sleep in consumer wearable devices.

The mysterious case of the disappearing pilot study: a review of publication bias in preliminary behavioral interventions presented at health behavior conferences

BACKGROUND

The number of preliminary studies conducted and published has increased in recent years. However, there are likely many preliminary studies that go unpublished because preliminary studies are typically small and may not be perceived as methodologically rigorous. The extent of publication bias within preliminary studies is unknown but can prove useful to determine whether preliminary studies appearing in peer-reviewed journals are fundamentally different than those that are unpublished. The purpose of this study was to identify characteristics associated with publication in a sample of abstracts of preliminary studies of behavioral interventions presented at conferences.

METHODS

Abstract supplements from two primary outlets for behavioral intervention research (Society of Behavioral Medicine and International Society of Behavioral Nutrition and Physical Activity) were searched to identify all abstracts reporting findings of behavioral interventions from preliminary studies. Study characteristics were extracted from the abstracts including year presented, sample size, design, and statistical significance. To determine if abstracts had a matching peer-reviewed publication, a search of authors' curriculum vitae and research databases was conducted. Iterative logistic regression models were used to estimate odds of abstract publication. Authors with unpublished preliminary studies were surveyed to identify reasons for nonpublication.

RESULTS

Across conferences, a total of 18,961 abstracts were presented. Of these, 791 were preliminary behavioral interventions, of which 49% (388) were published in a peer-reviewed journal. For models with main effects only, preliminary studies with sample sizes greater than n = 24 were more likely to be published (range of odds ratios, 1.82 to 2.01). For models including interactions among study characteristics, no significant associations were found. Authors of unpublished preliminary studies indicated small sample sizes and being underpowered to detect effects as barriers to attempting publication.

CONCLUSIONS

Half of preliminary studies presented at conferences go unpublished, but published preliminary studies appearing in peer-reviewed literature are not systematically different from those that remain unpublished. Without publication, it is difficult to assess the quality of information regarding the early-stage development of interventions. This inaccessibility inhibits our ability to learn from the progression of preliminary studies.

Scientists' perception of pilot study quality was influenced by statistical significance and study design

OBJECTIVES

Preliminary studies play a key role in developing large-scale interventions but may be held to higher or lower scientific standards during the peer review process because of their preliminary study status.

STUDY DESIGN AND SETTING

Abstracts from 5 published obesity prevention preliminary studies were systematically modified to generate 16 variations of each abstract. Variations differed by 4 factors: sample size (n = 20 vs. n = 150), statistical significance (P < 0.05 vs. P > 0.05), study design (single group vs. randomized 2 groups), and preliminary study status (presence/absence of pilot language). Using an online survey, behavioral scientists were provided with a randomly selected variation of each of the 5 abstracts and blinded to the existence of other variations. Respondents rated each abstract on aspects of study quality.

RESULTS

Behavioral scientists (n = 271, 79.7% female, median age 34 years) completed 1,355 abstract ratings. Preliminary study status was not associated with perceived study quality. Statistically significant effects were rated as more scientifically significant, rigorous, innovative, clearly written, warranted further testing, and had more meaningful results. Randomized designs were rated as more rigorous, innovative, and meaningful.

CONCLUSION

Findings suggest reviewers place a greater value on statistically significant findings and randomized control design and may overlook other important study characteristics.

Feasibility indicators in obesity-related behavioral intervention preliminary studies: a historical scoping review

BACKGROUND

Behavioral interventions are often complex, operate at multiple levels, across settings, and employ a range of behavior change techniques. Collecting and reporting key indicators of initial trial and intervention feasibility is essential to decisions for progressing to larger-scale trials. The extent of reporting on feasibility indicators and how this may have changed over time is unknown. The aims of this study were to (1) conduct a historical scoping review of the reporting of feasibility indicators in behavioral pilot/feasibility studies related to obesity published through 2020, and (2) describe trends in the amount and type of feasibility indicators reported in studies published across three time periods: 1982-2006, 2011-2013, and 2018-2020.

METHODS

A search of online databases (PubMed, Embase, EBSCOhost, Web of Science) for health behavior pilot/feasibility studies related to obesity published up to 12/31/2020 was conducted and a random sample of 600 studies, 200 from each of the three timepoints (1982-2006, 2011-2013, and 2018-2020), was included in this review. The presence/absence of feasibility indicators, including recruitment, retention, participant acceptability, attendance, compliance, and fidelity, were identified/coded for each study. Univariate logistic regression models were employed to assess changes in the reporting of feasibility indicators across time.

RESULTS

A total of 16,365 unique articles were identified of which 6873 of these were reviewed to arrive at the final sample of 600 studies. For the total sample, 428 (71.3%) studies provided recruitment information, 595 (99.2%) provided retention information, 219 (36.5%) reported quantitative acceptability outcomes, 157 (26.2%) reported qualitative acceptability outcomes, 199 (33.2%) reported attendance, 187 (31.2%) reported participant compliance, 23 (3.8%) reported cost information, and 85 (14.2%) reported treatment fidelity outcomes. When compared to the Early Group (1982-2006), studies in the Late Group (2018-2020) were more likely to report recruitment information (OR=1.60, 95%CI 1.03-2.49), acceptability-related quantitative (OR=2.68, 95%CI 1.76-4.08) and qualitative (OR=2.32, 95%CI 1.48-3.65) outcomes, compliance outcomes (OR=2.29, 95%CI 1.49-3.52), and fidelity outcomes (OR=2.13, 95%CI 1.21, 3.77).

CONCLUSION

The reporting of feasibility indicators within behavioral pilot/feasibility studies has improved across time, but key aspects of feasibility, such as fidelity, are still not reported in the majority of studies. Given the importance of behavioral intervention pilot/feasibility studies in the translational science spectrum, there is a need for improving the reporting of feasibility indicators.

Influence of pilot and small trials in meta-analyses of behavioral interventions: a meta-epidemiological study

BACKGROUND

Pilot/feasibility or studies with small sample sizes may be associated with inflated effects. This study explores the vibration of effect sizes (VoE) in meta-analyses when considering different inclusion criteria based upon sample size or pilot/feasibility status.

METHODS

Searches were to identify systematic reviews that conducted meta-analyses of behavioral interventions on topics related to the prevention/treatment of childhood obesity from January 2016 to October 2019. The computed summary effect sizes (ES) were extracted from each meta-analysis. Individual studies included in the meta-analyses were classified into one of the following four categories: self-identified pilot/feasibility studies or based upon sample size but not a pilot/feasibility study (N ≤ 100, N > 100, and N > 370 the upper 75th of sample size). The VoE was defined as the absolute difference (ABS) between the re-estimations of summary ES restricted to study classifications compared to the originally reported summary ES. Concordance (kappa) of statistical significance of summary ES between the four categories of studies was assessed. Fixed and random effects models and meta-regressions were estimated. Three case studies are presented to illustrate the impact of including pilot/feasibility and N ≤ 100 studies on the estimated summary ES.

RESULTS

A total of 1602 effect sizes, representing 145 reported summary ES, were extracted from 48 meta-analyses containing 603 unique studies (avg. 22 studies per meta-analysis, range 2-108) and included 227,217 participants. Pilot/feasibility and N ≤ 100 studies comprised 22% (0-58%) and 21% (0-83%) of studies included in the meta-analyses. Meta-regression indicated the ABS between the re-estimated and original summary ES where summary ES ranged from 0.20 to 0.46 depending on the proportion of studies comprising the original ES were either mostly small (e.g., N ≤ 100) or mostly large (N > 370). Concordance was low when removing both pilot/feasibility and N ≤ 100 studies (kappa = 0.53) and restricting analyses only to the largest studies (N > 370, kappa = 0.35), with 20% and 26% of the originally reported statistically significant ES rendered non-significant. Reanalysis of the three case study meta-analyses resulted in the re-estimated ES rendered either non-significant or half of the originally reported ES.

CONCLUSIONS

When meta-analyses of behavioral interventions include a substantial proportion of both pilot/feasibility and N ≤ 100 studies, summary ES can be affected markedly and should be interpreted with caution.

Classroom teachers' "off-the-shelf" use of movement integration products and its impact on children's sedentary behavior and physical activity

Movement integration (MI) products are one of many MI strategies that aim to reduce students' sedentary behavior (SB) and increase physical activity (PA) during classroom time. This study examined elementary classroom teachers' off-the-shelf (i.e., no researcher support) use of MI products (GoNoodle Plus [GN], ABC for Fitness [ABC], Take10) and their impact on students' SB and PA. Teachers (N = 57) at five schools received one MI product and reported MI strategy uses/day while student (n = 1,098, 52% female, 66% Black) accelerometer-determined SB and PA was assessed. Mixed regression models estimated changes in MI uses/day and SB and PA during the school day prior to and after teachers received the MI product. GoNoodle was the only MI product where overall MI strategy uses/day increased (∆ = 0.8, 95% CI = 0.1, 1.4). Across products, students' SB increased (∆ = 2.2, 95% CI = 1.2, 3.1) while light (∆ = -1.7, 95% CI = 1.2, 3.1) and MVPA (∆ = -0.5, 95% CI = -0.8, -0.2) decreased. For GN SB (∆ = -3.3, 95% CI = -7.8, 1.3), light (∆ = 2.5, 95% CI = -0.7, 5.7), and MVPA (∆ = 0.8, 95% CI = -0.9, 2.5), did not show statistically significant change. For Take10 SB (∆ = 1.0, 95% CI = -0.2, 2.2) and MVPA (∆ = 0.1, 95% CI = -0.3, 0.6) did not change while light PA decreased (∆ = -1.1, 95% CI = -2.0, -0.3). For ABC SB increased (∆ = 11.1, 95% CI = 8.4, 13.9) while light (∆ = -7.0, 95% CI = -8.9, -5.0) and MVPA (∆ = -4.2, 95% CI = -5.2, -3.1) decreased. GN shows promise for classroom teacher use. However, given limited uptake of the other products and the lack of change in children's SB and PA, this study suggests that off-the-shelf MI products cannot be integrated into classroom routines without additional support.

Feasibility of Measuring Screen Time, Activity, and Context Among Families With Preschoolers: Intensive Longitudinal Pilot Study

Background

Digital media has made screen time more available across multiple contexts, but our understanding of the ways children and families use digital media has lagged behind the rapid adoption of this technology.

Objective

This study evaluated the feasibility of an intensive longitudinal data collection protocol to objectively measure digital media use, physical activity, sleep, sedentary behavior, and socioemotional context among caregiver-child dyads. This paper also describes preliminary convergent validity of ecological momentary assessment (EMA) measures and preliminary agreement between caregiver self-reported phone use and phone use collected from passive mobile sensing.

Methods

Caregivers and their preschool-aged child (3-5 years) were recruited to complete a 30-day assessment protocol. Within 30-days, caregivers completed 7 days of EMA to measure child behavior problems and caregiver stress. Caregivers and children wore an Axivity AX3 (Newcastle Upon Tyne) accelerometer to assess physical activity, sedentary behavior, and sleep. Phone use was assessed via passive mobile sensing; we used Chronicle for Android users and screenshots of iOS screen time metrics for iOS users. Participants were invited to complete a second 14-day protocol approximately 3-12 months after their first assessment. We used Pearson correlations to examine preliminary convergent validity between validated questionnaire measures of caregiver psychological functioning, child behavior, and EMA items. Root mean square errors were computed to examine the preliminary agreement between caregiver self-reported phone use and objective phone use.

Results

Of 110 consenting participants, 105 completed all protocols (105/110, 95.5% retention rate). Compliance was defined a priori as completing ≥70%-75% of each protocol task. There were high compliance rates for passive mobile sensing for both Android (38/40, 95%) and iOS (64/65, 98%). EMA compliance was high (105/105, 100%), but fewer caregivers and children were compliant with accelerometry (62/99, 63% and 40/100, 40%, respectively). Average daily phone use was 383.4 (SD 157.0) minutes for Android users and 354.7 (SD 137.6) minutes for iOS users. There was poor agreement between objective and caregiver self-reported phone use; root mean square errors were 157.1 and 81.4 for Android and iOS users, respectively. Among families who completed the first assessment, 91 re-enrolled to complete the protocol a second time, approximately 7 months later (91/105, 86.7% retention rate).

Conclusions

It is feasible to collect intensive longitudinal data on objective digital media use simultaneously with accelerometry and EMA from an economically and racially diverse sample of families with preschool-aged children. The high compliance and retention of the study sample are encouraging signs that these methods of intensive longitudinal data collection can be completed in a longitudinal cohort study. The lack of agreement between self-reported and objectively measured mobile phone use highlights the need for additional research using objective methods to measure digital media use.

International Registered Report Identifier (IRRID)

RR2-36240

Fund behavioral science like the frameworks we endorse: the case for increased funding of preliminary studies by the National Institutes of Health

Innovative, groundbreaking science relies upon preliminary studies (aka pilot, feasibility, proof-of-concept). In the behavioral sciences, almost every large-scale intervention is supported by a series of one or more rigorously conducted preliminary studies. The importance of preliminary studies was established by the National Institutes of Health (NIH) in 2014/2015 in two translational science frameworks (NIH Stage and ORBIT models). These frameworks outline the essential role preliminary studies play in developing the next generation of evidence-based behavioral prevention and treatment interventions. Data produced from preliminary studies are essential to secure funding from the NIH’s most widely used grant mechanism for large-scale clinical trials, namely the R01. Yet, despite their unquestionable importance, the resources available for behavioral scientists to conduct rigorous preliminary studies are limited. In this commentary, we discuss ways the existing funding structure at the NIH, despite its clear reliance upon high-quality preliminary studies, inadvertently discourages and disincentivizes their pursuit by systematically underfunding them. We outline how multiple complementary and pragmatic steps via a small reinvestment of funds from larger trials could result in a large increase in funding for smaller preliminary studies. We make the case such a reinvestment has the potential to increase innovative science, increase the number of investigators currently funded, and would yield lasting benefits for behavioral science and scientists alike.

Healthy Summer Learners: An explanatory mixed methods study and process evaluation

Healthy Summer Learners (HSL), a novel, 6-week summer program for 2-4th grade children from low-income families in the Southeastern United States, aimed to prevent accelerated summer BMI gain and academic learning loss by providing healthy meals and snacks, 15 min of nutrition education, 3 h of physical activity opportunities and 3.5 h of reading instruction daily. This three-armed pilot quasi-experimental study used a repeated measure within- and between-participant design to compare HSL, to an active comparator-21st Century Summer Learning Program (21 C), and no-treatment control. A mixed-methods process evaluation was employed to evaluate program implementation and provide insight for future program development. Though the program was well received, student attendance was lower than anticipated and full program fidelity was not achieved. During interviews, both parents and teachers noted that the bussing schedule was inconsistent, making attendance difficult for some families. These process evaluation findings may help explain why no statistically significant group-by-time interactions at 3- or 12-month follow up were found for the primary outcomes of zBMI or MAP reading score. Future iterations of HSL should seek to extend program hours, lengthen program duration, and explore ways to lower projected cost of attendance.

Impact of risk of generalizability biases in adult obesity interventions: A meta-epidemiological review and meta-analysis

Biases introduced in early-stage studies can lead to inflated early discoveries. The risk of generalizability biases (RGBs) identifies key features of feasibility studies that, when present, lead to reduced impact in a larger trial. This meta-study examined the influence of RGBs in adult obesity interventions. Behavioral interventions with a published feasibility study and a larger scale trial of the same intervention (e.g., pairs) were identified. Each pair was coded for the presence of RGBs. Quantitative outcomes were extracted. Multilevel meta-regression models were used to examine the impact of RGBs on the difference in the effect size (ES, standardized mean difference) from pilot to larger scale trial. A total of 114 pairs, representing 230 studies, were identified. Overall, 75% of the pairs had at least one RGB present. The four most prevalent RGBs were duration (33%), delivery agent (30%), implementation support (23%), and target audience (22%) bias. The largest reductions in the ES were observed in pairs where an RGB was present in the pilot and removed in the larger scale trial (average reduction ES -0.41, range -1.06 to 0.01), compared with pairs without an RGB (average reduction ES -0.15, range -0.18 to -0.14). Eliminating RGBs during early-stage testing may result in improved evidence.

Measuring Micro Temporal Processes Underlying Preschoolers Screen Use and Behavioral Health: Protocol for the Tots & Tech Study (Preprint)

Background

Excessive screen time is associated with poor health and behavioral outcomes in children. However, research on screen time use has been hindered by methodological limitations, including retrospective reports of usual screen time and lack of momentary etiologic processes occurring within each day.

Objective

This study is designed to assess the feasibility and utility of a comprehensive multibehavior protocol to measure the digital media use and screen time context among a racially and economically diverse sample of preschoolers and their families. This paper describes the recruitment, data collection, and analytical protocols for the Tots and Tech study.

Methods

The Tots and Tech study is a longitudinal, observational study of 100 dyads: caregivers and their preschool-age children (aged 3-5 years). Both caregivers and children will wear an Axivity AX3 accelerometer (Axivity Ltd) for 30 days to assess their physical activity, sedentary behavior, and sleep. Caregivers will complete ecological momentary assessments (EMAs) for 1 week to measure child behavioral problems, caregiver stress, and child screen time.

Results

The Tots and Tech study was funded in March 2020. This study maintains rolling recruitment, with each dyad on their own assessment schedule, depending on the time of enrollment. Enrollment was scheduled to take place between September 2020 and May 2022. We aim to enroll 100 caregiver-child dyads. The Tots and Tech outcome paper is expected to be published in 2022.

Conclusions

The Tots and Tech study attempts to overcome previous methodological limitations by using objective measures of screen time, physical activity, sedentary behavior, and sleep behaviors with contextual factors measured by EMA. The results will be used to evaluate the feasibility and utility of a comprehensive multibehavior protocol using objective measures of mobile screen time and accelerometry in conjunction with EMA among caregiver-child dyads. Future observational and intervention studies will be able to use this study protocol to better measure screen time and its context.

International Registered Report Identifier (IRRID)

DERR1-10.2196/36240

Differences in ​the proportion of children meeting behavior guidelines ​between summer ​and school by socioeconomic status and race

OBJECTIVE

Children who fail to meet activity, sleep, and screen-time guidelines are at increased risk for obesity. Further, children who are Black are more likely to have obesity when compared to children who are White, and children from low-income households are at increased risk for obesity when compared to children from higher-income households. The objective of this study was to evaluate the proportion of days meeting obesogenic behavior guidelines during the school year compared to summer vacation by race and free/reduced priced lunch (FRPL) eligibility.

METHODS

Mixed-effects linear and logistic regressions estimated the proportion of days participants met activity, sleep, and screen-time guidelines during summer and school by race and FRPL eligibility within an observational cohort sample.

RESULTS

Children (n = 268, grades = K - 4, 44.1%FRPL, 59.0% Black) attending three schools participated. Children's activity, sleep, and screen-time were collected during an average of 23 school days and 16 days during summer vacation. During school, both children who were White and eligible for FRPL met activity, sleep, and screen-time guidelines on a greater proportion of days when compared to their Black and non-eligible counterparts. Significant differences in changes from school to summer in the proportion of days children met activity (-6.2%, 95CI = -10.1%, -2.3%; OR = 0.7, 95CI = 0.6, 0.9) and sleep (7.6%, 95CI = 2.9%, 12.4%; OR = 2.1, 95CI = 1.4, 3.0) guidelines between children who were Black and White were observed. Differences in changes in activity (-8.5%, 95CI = -4.9%, -12.1%; OR = 1.5, 95CI = 1.3, 1.8) were observed between children eligible versus uneligible for FRPL.

CONCLUSIONS

Summer vacation may be an important time for targeting activity and screen-time of children who are Black and/or eligible for FRPL.

Impact of a year-round school calendar on children's BMI and fitness: Final outcomes from a natural experiment

BACKGROUND

Structure may mitigate children's accelerated summer BMI gain and cardiorespiratory-fitness (CRF) loss.

OBJECTIVES

Examine BMI and CRF change during school and summer for year-round and traditional calendar school children.

METHODS

Three schools (N = 2279, 1 year-round) participated in this natural experiment. Children's BMI z-score (zBMI) and CRF (PACER laps) were measured from 2017 to 2019 each May/August. Mixed effects regression estimated monthly zBMI and CRF change during school/summer. Secondary analyses examined differences by weight status and race. Spline regression models estimated zBMI and CRF growth from kindergarten-sixth grade.

RESULTS

Compared to traditional school, children attending a year-round school gained more zBMI (difference = 0.015; 95CI = 0.002, 0.028) during school, and less zBMI (difference = -0.029; 95CI = -0.041, -0.018), and more CRF (difference = 0.834; 95CI = 0.575, 1.093) monthly during summer. Differences by weight status and race were observed during summer and school. Growth models demonstrated that the magnitude of overall zBMI and CRF change from kindergarten-sixth grade was similar for year-round or traditional school children.

CONCLUSIONS

Contrary to traditional school children zBMI increased during the traditional 9-month school calendar and zBMI decreased during the traditional summer vacation for year-round school children. Structured summer programming may mitigate accelerated summer BMI gain and CRF loss especially for overweight or obese, and/or Black children.

COVID-19 Leads to Accelerated Increases in Children's BMI z-Score Gain: An Interrupted Time-Series Study

INTRODUCTION

The COVID-19 pandemic may have negatively impacted children's weight status owing to the closure of schools, increased food insecurity and reliance on ultraprocessed foods, and reduced opportunities for outdoor activity.

METHODS

In this interrupted time-series study, height and weight were collected from children (n=1,770 children, mean age=8.7 years, 55.3% male, 64.6% Black) and were transformed into BMI z-score in each August/September from 2017 to 2020. Mixed-effects linear regression estimated yearly BMI z-score change before the COVID-19 pandemic year (i.e., 2017-2019) and during the COVID-19 pandemic year (i.e., 2019-2020). Subgroup analyses by sex, race (i.e., Black, White, other race), weight status (overweight or obese and normal weight), and grade (i.e., lower=kindergarten-2nd grade and upper=3rd-6th grade) were conducted.

RESULTS

Before the COVID-19 pandemic, children's yearly BMI z-score change was +0.03 (95% CI= -0.10, 0.15). Change during the COVID-19 pandemic was +0.34 (95% CI=0.21, 0.47), an acceleration in BMI z-score change of +0.31 (95% CI=0.19, 0.44). For girls and boys, BMI z-score change accelerated by +0.33 (95% CI=0.16, 0.50) and +0.29 (95% CI=0.12, 0.46), respectively, during the pandemic year. Acceleration in BMI z-score change during the pandemic year was observed for children who were Black (+0.41, 95% CI=0.21, 0.61) and White (+0.22, 95% CI=0.06, 0.39). For children classified as normal weight, BMI z-score change accelerated by +0.58 (95% CI=0.40, 0.76). Yearly BMI z-score change accelerated for lower elementary/primary (+0.23, 95% CI=0.08, 0.37) and upper elementary/primary (+0.42, 95% CI=0.42, 0.63) children.

CONCLUSIONS

If similar BMI z-score accelerations occurred for children across the world, public health interventions to address this rapid unhealthy BMI gain will be urgently needed.

Identifying effective intervention strategies to reduce children's screen time: a systematic review and meta-analysis

BACKGROUND

Excessive screen time ([Formula: see text] 2 h per day) is associated with childhood overweight and obesity, physical inactivity, increased sedentary time, unfavorable dietary behaviors, and disrupted sleep. Previous reviews suggest intervening on screen time is associated with reductions in screen time and improvements in other obesogenic behaviors. However, it is unclear what study characteristics and behavior change techniques are potential mechanisms underlying the effectiveness of behavioral interventions. The purpose of this meta-analysis was to identify the behavior change techniques and study characteristics associated with effectiveness in behavioral interventions to reduce children's (0-18 years) screen time.

METHODS

A literature search of four databases (Ebscohost, Web of Science, EMBASE, and PubMed) was executed between January and February 2020 and updated during July 2021. Behavioral interventions targeting reductions in children's (0-18 years) screen time were included. Information on study characteristics (e.g., sample size, duration) and behavior change techniques (e.g., information, goal-setting) were extracted. Data on randomization, allocation concealment, and blinding was extracted and used to assess risk of bias. Meta-regressions were used to explore whether intervention effectiveness was associated with the presence of behavior change techniques and study characteristics.

RESULTS

The search identified 15,529 articles, of which 10,714 were screened for relevancy and 680 were retained for full-text screening. Of these, 204 studies provided quantitative data in the meta-analysis. The overall summary of random effects showed a small, beneficial impact of screen time interventions compared to controls (SDM = 0.116, 95CI 0.08 to 0.15). Inclusion of the Goals, Feedback, and Planning behavioral techniques were associated with a positive impact on intervention effectiveness (SDM = 0.145, 95CI 0.11 to 0.18). Interventions with smaller sample sizes (n < 95) delivered over short durations (< 52 weeks) were associated with larger effects compared to studies with larger sample sizes delivered over longer durations. In the presence of the Goals, Feedback, and Planning behavioral techniques, intervention effectiveness diminished as sample size increased.

CONCLUSIONS

Both intervention content and context are important to consider when designing interventions to reduce children's screen time. As interventions are scaled, determining the active ingredients to optimize interventions along the translational continuum will be crucial to maximize reductions in children's screen time.

Small studies, big decisions: the role of pilot/feasibility studies in incremental science and premature scale-up of behavioral interventions

Background

Careful consideration and planning are required to establish “sufficient” evidence to ensure an investment in a larger, more well-powered behavioral intervention trial is worthwhile. In the behavioral sciences, this process typically occurs where smaller-scale studies inform larger-scale trials. Believing that one can do the same things and expect the same outcomes in a larger-scale trial that were done in a smaller-scale preliminary study (i.e., pilot/feasibility) is wishful thinking, yet common practice. Starting small makes sense, but small studies come with big decisions that can influence the usefulness of the evidence designed to inform decisions about moving forward with a larger-scale trial. The purpose of this commentary is to discuss what may constitute sufficient evidence for moving forward to a definitive trial. The discussion focuses on challenges often encountered when conducting pilot/feasibility studies, referred to as common (mis)steps, that can lead to inflated estimates of both feasibility and efficacy, and how the intentional design and execution of one or more, often small, pilot/feasibility studies can play a central role in developing an intervention that scales beyond a highly localized context.

Main body

Establishing sufficient evidence to support larger-scale, definitive trials, from smaller studies, is complicated. For any given behavioral intervention, the type and amount of evidence necessary to be deemed sufficient is inherently variable and can range anywhere from qualitative interviews of individuals representative of the target population to a small-scale randomized trial that mimics the anticipated larger-scale trial. Major challenges and common (mis)steps in the execution of pilot/feasibility studies discussed are those focused on selecting the right sample size, issues with scaling, adaptations and their influence on the preliminary feasibility and efficacy estimates observed, as well as the growing pains of progressing from small to large samples. Finally, funding and resource constraints for conducting informative pilot/feasibility study(ies) are discussed.

Conclusion

Sufficient evidence to scale will always remain in the eye of the beholder. An understanding of how to design informative small pilot/feasibility studies can assist in speeding up incremental science (where everything needs to be piloted) while slowing down premature scale-up (where any evidence is sufficient for scaling).

Comparison of multichannel and single-channel wrist-based devices with polysomnography to measure sleep in children and adolescents

STUDY OBJECTIVES

To compare sleep parameters produced by the Fitbit Charge 3 (Fitbit) and Actigraph GT9X accelerometer (Actigraph) to polysomnography in children and adolescents.

METHODS

Participants (n = 56, ages 9.2 ± 3.3 years) wore a Fitbit and an Actigraph on their nondominant wrist concurrently with polysomnography during an overnight observation at a children's sleep laboratory. Total sleep time, sleep efficiency, wake after sleep onset, sleep onset, and sleep offset were extracted from the Fitabase and Actilife software packages, respectively, with the Sadeh algorithm. Bland-Altman plots were used to assess the agreement between wearable devices and polysomnography.

RESULTS

Seventy-nine percent of participants were diagnosed with OSA. Compared with polysomnography, the Fitbit and the Actigraph underestimated total sleep time by 6.1 minutes (absolute mean bias [AMB] = 27.7 minutes) and 31.5 minutes (AMB = 38.2 minutes), respectively. The Fitbit overestimated sleep efficiency by 3.0% (AMB = 6.3%), and the Actigraph underestimated sleep efficiency by 12.9% (AMB = 13.2%). The Fitbit overestimated wake after sleep onset by 18.8 minutes (AMB = 23.9 minutes), and the Actigraph overestimated wake after sleep onset by 56.1 minutes (AMB = 54.7 minutes). In addition, the Fitbit and the Actigraph underestimated sleep onset by 1.2 minutes (AMB = 13.9 minutes) and 10.2 minutes (AMB = 18.1 minutes), respectively. Finally, the Fitbit and the Actigraph overestimated sleep offset by 6.0 minutes (AMB = 12.0 minutes) and 10.5 minutes (AMB = 12.6 minutes). Linear regression indicated significant trends, with the Fitbit underestimating wake after sleep onset and sleep efficiency at higher values.

CONCLUSIONS

The Fitbit provided comparable and in some instances better sleep estimates with polysomnography compared to the Actigraph. Findings support the use of multichannel devices to measure sleep in children and adolescents. Additional studies are needed in healthy children over several nights and in free-living settings.