Sage Journals: Discover world-class research

Abstract

Introduction:

Meal timing can influence metabolism and health. Correct meal timing data is crucial to accurately assess eating patterns and develop effective interventions. However, commonly used methods like dietary recalls are often inaccurate for caloric and nutrient intake. In this pilot study, we aimed to assess the mealtime agreement between dietary recalls and documentation in real-time.

Methods:

This is an observational 2-week study in older adults with prediabetes and obesity. Participants reported their diet with (1) 2-6 non-consecutive dietary recalls (Automated Self-Administered 24-h (ASA24^®) Dietary Assessment Tool), and (2) time-stamped photographs of all caloric eating and drinking events via the myCircadianClock (mCC) mobile application. Data analysis was performed on days when both methods were used to assess the first eating occasion (FEO) and the last eating occasion (LEO) of the day. Significance set at P < .05.

Results:

Adults (n = 43, 14 men, age 60 ± 7 years, body mass index (BMI) 33.0 ± 6.2 kg/m²) documented 4 ± 1 ASA24 recalls over 2 weeks. Agreement between tools ranged between 62% to 73% for the FEO and 49% to 61% for the LEO. The FEO and LEO agreement were positively correlated (β = .4, P = .01). Differences between tools were higher with fewer recalls (β = −.3, P = .04) for the LEO, but were not associated with age, sex, or body composition for neither the FEO or the LEO. The FEO agreement between tools was higher when the FEO was consumed earlier, when agreement was defined by ±45 min (estimate = −0.94, P = .001) and ±60 min (estimate = −1.11, P < .001).

Conclusions:

Our results suggest that mealtime agreement between dietary recalls and a time-stamped tool is consistently below 75%, and lowers for meals recorded later in the day. The inter-tool agreement is higher with increased overlap tolerance and greater number of dietary recalls.

Clinical trials registration number:

NCT04465721

https://clinicaltrials.gov/study/NCT04465721

Keywords

self-report dietary recall meal timing eating patterns first eating occasion last eating occasion mobile application

Introduction

Emerging evidence suggests that the alignment of meal timing and endogenous circadian rhythms is an important determinant of metabolic disorders.¹ The temporal pattern of food consumption may interact with dietary composition and modulate overall health and metabolic function.^2,3 In addition, habitually consuming meals later in the day can disrupt physiological functions and biological processes that ultimately promote weight gain⁴ and increase metabolic risk, independent of caloric intake or adiposity.⁵ Therefore, meal timing could be a modifiable lifestyle component that healthcare providers can integrate with caloric intake and physical activity recommendations. Contrary to animal models, evidence on meal timing interventions are conflicting, as different study designs rely on varying methodologies to assess meal timing, with numerous studies relying on self-reported methods.^6
-9

To assess energy intake in free-living environments, researchers have commonly relied on dietary recalls, surveys, or food frequency questionnaires.¹⁰ These tools are designed to capture various aspects of dietary intake, including specific food items, portion sizes, preparation methods, and the timing of consumption, ranging from the past 24 h for dietary recalls, to weeks or months for food frequency questionnaires to capture typical usual dietary habits. However, memory lapses are associated with any mentally effortful task¹¹ and the aforementioned methods are notoriously inaccurate when self-reported energy intake is compared against the gold-standard method (doubly-labeled water). The assessment of dietary recalls accuracy has often been limited to energy and nutrient intake.^10,12,13 However, inaccuracies may also be present in other components of dietary recalls, including temporal eating patterns. While there is a promising role of meal timing interventions in lowering body weight, BMI, waist circumference, glycosylated hemoglobin, and fasting glucose,^14
-17 a universally accepted “gold-standard” method for assessing meal timing in the ambulatory setting remains unavailable. Inaccurate measurements of meal timing can significantly hinder research efforts on emerging areas of interest regarding eating patterns, including the duration of the overnight fasting, eating midpoint, and eating window. An accurate and precise documentation of meal timing is critical to correctly estimate these variables, and inaccuracies can lead to erroneous conclusions about the impact of specific eating patterns and health outcomes. For instance, misreporting of the last meal time can lead to over or underestimating of fasting duration, or mask the true impact of late eating on metabolic health.

Passive approaches to assess eating patterns, including wrist motion and automatic ingestion monitors, have the potential to eliminate recall bias, and the effect of self-monitoring. However, these approaches often involve intrusive monitoring and have primarily been evaluated in controlled settings, presenting important limitations for widespread applicability.^18
-20 The myCircadianClock (mCC) phone application (app)^21,22 is a validated tool developed to capture eating patterns, and offers a non-invasive alternative to assess temporal energy intake by timestamping all caloric entries in real-time. In this pilot study, we aimed to assess and characterize the extent of agreement between self-reported mealtime by dietary recalls compared to real-time documented via the phone app. Based on previous findings on inaccuracies of self-reported caloric intake, we hypothesized that the agreement between dietary recalls and real-time tools would be less than 75%. We also hypothesized that agreement would be higher with greater number of dietary recalls and at earlier times of the day.

Methods

Study Design

Analyses were conducted in a sub-sample of the New York Time Restricted EATing to improve cardiometabolic health (NY-TREAT),²³ a single-center, randomized, open-label trial. After enrollment, participants completed a 2-week baseline, which occurred in ambulatory conditions, except for outpatient clinic visits on days 1, 13, and 14. Only data prior to randomization were used for this analysis.

Study Population

Older adults from the New York City area were recruited between June 2021 and December 2023. Inclusion criteria included age 50 to 75 years, overweight or obesity, prediabetes or early-onset type 2 diabetes, English speaking, in possession of a smartphone, habitual breakfast eater, weight stable for at least 3 months, successfully enrolled in the NY-TREAT trial,²³ that is, having demonstrated a ≥70% of days with logging adherence with the app (see Eating Patterns section) during the 2-week remote screening and with a habitual eating window of ≥14 h duration. Participants with sleep disorders and active shift work, with a history of bariatric surgery or on weight loss medications were excluded. Additional information about inclusion and exclusion criteria have been previously described.²³ Informed consent was obtained from all participants, and all procedures of this study were approved by the Columbia University Institutional Review Board (IRB AAAS7791). All measurements in this study were performed in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Anthropometric Measurements

On the first day of a 2-week period, body weight was measured in triplicate to the nearest 0.1 kg (Ohaus Champ General Purpose Bench Scale, Ohaus Corp., Pine Brook, NJ, USA) and height was measured in triplicate to the nearest 1 mm using a stadiometer (Holtain Ltd., Crymych, UK). Body mass index (BMI) was calculated as weight (kg) divided by height squared (m²). Waist circumference was measured in triplicate to the nearest 0.1 cm between the lowest point of the costal margins on the mid-axillary lines and the highest point of the iliac crests using a Gulick II measurement tape (Country Technology, Inc., Gays Mills, WI). Prior to these measurements, participants were instructed to empty their bladders, remove garments and jewelry, and changing into hospital gown and slippers provided by the study staff.

Dietary Recalls

Participants completed up to six 24-h dietary recalls on non-consecutive weekdays and at least 1 weekend day using the web-based 24-h dietary recalls via the Automated Self-Administered 24-h^® (ASA24^®) Dietary Assessment Tool¹² and received a practice session under staff supervision on the first study day. Participants were instructed to report all their meals from the preceding day, and maintain their habitual diet throughout the baseline assessment study period. All recalls were completed under free-living conditions, except for the last entry, which was completed during a study visit at the end of the 2-week period under staff supervision. The ASA24^® is a web-based tool provided by the National Cancer Institute, and registration to use the full version of the researcher website is available with open-access.

Eating Patterns

Participants used the mCC mobile application to record their daily meal timing during the 2-week baseline assessment. After consent, participants were coached to use the mCC, and underwent a 2-week remote screening period to ensure adherence to logging and to measure the duration of their daily eating window. Adherence was defined as logging at least 2 meals per day, separated by a minimum of 5 h on at least 70% of logging days.²⁴ The mCC was developed by the Salk institute for Biological studies, and is accessible on iOS and Android platforms, allows logging of all food, snack, and beverage intake by time-stamped photos or brief text entry, without data on calories, nutrients or portion sizes. Participants were instructed to continue their typical dietary intake and to log all entries using a time-stamped photograph or text entry in real-time, immediately before starting ingestion, and to limit same-day retroactive logging to exceptional circumstances. All entries were completed under free-living conditions, except for the last entry, which was completed during a study visit at the end of the 2-week period under staff supervision. All entries were transferred to a Health Insurance Portability and Accountability Act (HIPAA)-compliant and double-encrypted remote server immediately after data submission. Participants did not have access to a summary of their eating patterns within the app after dietary entries. Details about the smartphone application software, data annotation and analyses have been previously published.²¹ The times of the first eating occasion (FEO) and last eating occasion (LEO) were assessed for days when both dietary recall data and real-time data were available. Formal data use agreement to access and analyze the collected data was secured with the Salk Institute for Biological Studies prior to the commencement of the study.

Meal Timing Agreement Assessment

Meal intake was defined as any caloric intake taking place throughout a 24-h day, except water. The 24-h period was adjusted to start at 4:00 a.m. and end at 3:59 a.m. the following day to account for eating events occurring after 12:00 a.m. The timing of the beginning of the FEO (first intake after 4:00 am) and LEO (last intake prior to 3:59 am) were determined for all paired entries in both the recall and the real-time tools. For each individual entry, the time between both tools was compared to determine the agreement between both methods, with tolerance periods of ±30, ±45, and ±60 min. These tolerance periods were selected based on prior literature showing “slight” and “fair” meal timing reliability for the LEO with dietary recalls use,²⁵ and a range of 52% to 80% overlap of dietary recalls against wrist motion activity and continuous glucose monitoring with ±30, ±60, and ±120 min.²⁰ Agreement between both methods for the self-reported entry was dichotomized as “concordant.” Non-agreement between both methods was dichotomized as “non-concordant.” To determine the percentage agreement over 2 weeks, the number of concordant entries was summed and divided by the total number of dietary entries.

Statistical Analyses

No formal power analysis was conducted to determine the sample size for this pilot study. All baseline data from participants enrolled in the parent grant were included in these analyses. Categorical variables, including sex, age, race, and ethnicity were compared with a chi-squared test. For continuous data, Pearson’s and Spearman’s correlations were performed to assess relationships between parametric and non-parametric variables, respectively, which were determined by the Shapiro-Wilk test. The Mann-Whitney U test was used to compare non-parametric unpaired data. The Sign test was used to compare non-parametric paired data. The absolute difference between dietary tools was calculated to determine the overall magnitude of the difference between methods. Linear mixed models were used to evaluate the effect of call sequence on agreement. Systematic bias for method comparison was assessed with Bland-Altman analysis.²⁶ To further quantify the degree of agreement, and specifically, to test if the meal timing agreement between methods was higher with earlier meals compared to later meals, the difference between methods was assessed for the first meal and last meals of the day recorded by both methods by 95% confidence interval (CI) using a linear regression, and dichotomized as binary variables (1 = yes, 0 = no) at ±30, ±45, and ±60 min, where “1” denoted that the constructed 95% confidence interval fell into the ±30, ±45, and ±60 min time windows, respectively, and the derived binary variables were correlated with real meal timings under a logistic regression model. Data reported as count (percentage) for categorical variables and mean ± standard deviation (SD) for continuous variables unless otherwise noted. Statistical analyses were conducted using SPSS (version 29.0.0.0, SPSS Statistics, IBM, Armonk, NY, USA) and R software (version 4.2.0, R Foundation for Statistical Computing, Vienna, Austria). The level of significance was set at .05.

Results

Overview

A total 43 healthy adults completed 2 to 6 dietary recalls for a total of 174 paired recalls and real-time entries (4 ± 1 per participant over 2 weeks). The participants flow diagram is presented in Supplemental Figure 1, and descriptive characteristics are shown in Table 1. There were no patterns of lower or higher agreement between the 2 tools across repeated entries in linear mixed model analyses (Supplemental Table 1), therefore, all analyses were conducted with the average agreement among all entries.

Table 1.

Descriptive Characteristics.

Variable		N (%), Mean ± SD
Age (years)		60.0 ± 7.0
Sex	Male	14 (32.6)
Sex	Female	29 (67.4)
Race	White	14 (32.6)
	Black	14 (32.6)
	Asian	9 (20.9)
	One or more race	3 (7)
	Unknown	3 (7)
Ethnicity	Hispanic	14 (32.5)
Ethnicity	Not Hispanic	29 (67.4)
Anthropometric measurements	Height (cm)	166.1 ± 9.4
	Weight (kg)	91.4 ± 20.5
	BMI (kg/m²)	33.0 ± 6.2
	Waist circumference (cm)	97.2 ± 46.0
Dietary recalls entries (ASA24^®)	Total number of ASA24 recalls over 2 weeks	4 ± 1
	Average FEO time (hh:mm)	8:43 ± 2:02
	Average LEO time (hh:mm)	19:24 ± 2:09
Real-time entries (mCC app)	Average FEO time (hh:mm)	8:58 ± 2:48
Real-time entries (mCC app)	Average LEO time (hh:mm)	19:04 ± 1:58

Note. Descriptive characteristics.

BMI = body mass index; cm = centimeters; FEO = first eating occasion; hh:mm = hours and minutes; kg = kilograms; LEO = last eating occasion; m² = meter squared; N = number; SD = standard deviation.

The mean ± SD FEO intake time was 8:43 ± 2:02 h with recalls versus 8:58 ± 2:48 h with real-time tool, respectively, resulting in an average meal timing difference of 0:25 ± 1:16 h with no significant difference between tools (P = .08). The mean ± SD LEO intake time was 19:24 ± 2:09 h with recalls and 19:04 ± 1:58 h with real-time tool, respectively, resulting in an average meal timing difference of 0:20 ± 2:13 h with no significant difference between tools (P = .25; Figure 1A). The percentage of meal timing agreement between 2 dietary tools with a ±30 min tolerance periods was 62% and 49% for the FEO and LEO, respectively. At ±45 min tolerance, the meal timing agreement increased to 67% for the FEO and 54% for the LEO. At ±60 min tolerance, the meal timing agreement was 73% for the FEO and 61% for the LEO (Figure 1B). We also assessed whether lower meal timing agreement showed any relationships with known predictors of caloric intake misreporting. Age, sex, nor anthropometric measurements (weight, waist circumference, BMI) were not correlated with the FEO or the LEO agreement.

Figure 1.

Overview of meal timing using two dietary tools: (A) comparison of the time difference between dietary tools for the FEO and LEO. The median value shown by the horizontal line, the mean value marked with an “x” within each box, whiskers represent the minimum and maximum values excluding outliers, outliers are shown as individual points outside of the boxes, (B) percentage of the average meal timing agreement between two dietary tools at different time intervals, and (C) correlations between age, sex, anthropometric measurements, number of dietary recalls, and the absolute difference for the FEO and LEO. Correlation coefficients and P-values are displayed in Supplemental Table 2.

There was a significantly positive correlation between the absolute difference between tools for the FEO and the LEO (Figure 1C). No significant relationships were seen between the number of dietary recalls and the absolute difference between the 2 tools for the FEO. However, there was a significantly negative correlation between the number of dietary recalls and absolute difference in time of LEO between the 2 tools. In further analyses, we found that participants that completed at least 4 dietary recalls had a lower absolute difference in meal timing between tools for the LEO compared to those who completed 3 dietary recalls or less (Supplemental Figure 2).

Extent of Agreement Between Two Dietary Tools

The average meal timing between dietary recalls and real-time tools were positively correlated for the FEO (r = .745, P < .001) and the LEO (r = .542, P < .001; Supplemental Figure 3). The Bland-Altman plots show a range of under-estimation and over-estimation of meal timing using the recall tool for both the FEO and LEO (up to 4 h and 8 min for FEO, and up to 5 h and 46 min for the LEO; Supplemental Figure 4). For the FEO, there was no relationship between both dietary tools agreement and real-time data with a ±30 min tolerance period (P = .238). However, there was an inverse relationship between dietary tools agreement and real-time data with ±45 min tolerance (P = .001) and ±60 min tolerance (P < .001; Table 2). These relationships were absent for the LEO. In summary, the FEO inter-tool agreement is higher at earlier times of the day when an agreement is defined by ±45 and ±60 min. We did not find any significant relationships between dietary tools agreement and real-time data for the LEO (Table 2).

Table 2.

Agreement Between 2 Dietary Tools with a Tolerance of 30, 45, and 60 min.

Tolerance period (min)	FEO			LEO
Tolerance period (min)	Estimate	SE	P-value	Estimate	SE	P-value
±30	−0.4877	0.41	.24	−0.00000024	27 770	.99
±45	−0.9400	0.29	<.01	0.1806	0.24	.45
±60	−1.105	0.31	<.01	0.1919	0.18	.28

Note. The extent of agreement between dietary recalls and real-time tools, using a logistic regression model. Significance is shown in bold.

FEO = first eating occasion; LEO = last eating occasion; SE = standard error.

Discussion

The present study aimed to evaluate the concordance in reporting meal timing between dietary recalls and documentation in real-time with a smartphone application. Our findings indicate that there are discrepancies and high variability between the 2 methods. These discrepancies are higher for meals consumed at later times of the day. These findings add to the existing literature on the inferiority of recall methods to assess dietary intake, as this method is subject to biases. Prior studies have focused on the validity of dietary recalls for caloric and nutrient intake,^{10,12,13,27
-31} and our findings indicate that errors are not limited to caloric and nutrient inaccuracies, but may also apply to meal timing. Understanding these limitations is crucial to inform future researchers about the use of dietary recall data in studies investigating meal patterns and health outcomes. Furthermore, data from publications that employ dietary recall methods to assess eating patterns should be interpreted with nuances and the understanding of the limitations of the recall methods.

Our data show a low level of agreement, 62% and 49%, between dietary recall and real-time phone application for the FEO and LEO meal timing, respectively. While increasing the tolerance period from 30 to 45 and 60 min did improve agreement for both, as expected, the highest agreement reached was only 73% and 61% for the FEO and LEO, respectively, with a 60-min tolerance period. These findings indicate that discrepancies in meal timing between dietary recalls and a time-stamped measure persist regardless of relaxation in agreement criteria to account for potential variations in the accuracy and precision of both tools. These results align with previous research, which found that even with generous tolerance periods, there is no perfect agreement between self-reported recalls and objective measures of meal timing.²⁰ This suggests that relying solely on recalls may not accurately capture meal timing, particularly when precise measurements of all components of eating patterns are crucial in research settings.

Our analyses did not reveal any significant association between meal timing agreement and either demographic factors (age and sex) or anthropometric measurements (weight, waist circumference, and BMI). These results contradict our hypotheses, developed from previous literature on dietary recall underreporting of caloric intake, which have shown that women, older individuals, and individuals with higher BMI are more likely to underreport their caloric intake^29,31,32 and an epidemiological study that found sex differences in meal timing reliability.²⁵ Our results could indicate that inaccuracies in meal timing reporting may not be susceptible to the negative societal biases commonly associated with caloric intake reporting^33
-36 that could potentially impact the behavior of specific demographic groups, (eg, women).³⁷ Nevertheless, other potential moderators could influence these correlations (eg, body satisfaction), and our study may have not captured these correlations as it is a small sample size in a cohort that only included older participants with overweight or obesity.

To investigate the influence of the number of recalls on the overall accuracy of meal timing, we assessed meal timing agreement relationships with the total number of recalls completed by participants. Our findings show that individuals who completed at least 4 dietary recalls had a lower absolute time difference between tools for the LEO. In addition, in correlation analyses, the absolute difference between the 2 dietary tools was lower with an increasing number of dietary recalls for the LEO, that is, the greater the number of recalls, the better agreement in time reporting between recalls and time-stamped app. These results agree with previous epidemiological findings that have shown increased meal timing reliability with 3 or more ASA24 over a 1-to-3-year period,²⁵ and reinforce that in the absence of real-time tools for meal timing measurement, 3 or more dietary recalls are necessary to increase the likelihood of accurate estimates of average meal timing.

We also observed a positive relationship between the difference among both dietary tools for the FEO and LEO, suggesting that individuals with a high agreement for the FEO also have a greater likelihood to accurately recall their meal timing at other times of the day. These results indicate that high agreement of meal timing is not a random or isolated occurrence. We speculate that lower LEO agreement may indicate retroactive interference associated with mentally effortful tasks¹¹ or dietary recall completion failure.³⁸ Therefore, identifying the factors that potentially contribute to higher agreement (or discrepancies) in reporting of meal timing are warranted, as meal timing can be influenced by other mediators that have not been previously addressed as predictors of caloric misreporting, including chronotype, societal norms, work schedules, seasonal variations, or daylight savings.^25,39

The average meal timing between dietary recalls and real-time tools was positively correlated for the FEO and LEO, and the mean difference between the 2 dietary tools was of 25 and 20 min for the FEO and LEO, respectively. However, the standard deviation of the mean was 1 h and 16 min for the FEO, and 2 h and 13 min for the LEO, indicating high variability, therefore, the difference between the mean was not statistically significant. The Bland-Altman plots demonstrated a wide range of individual differences, indicating substantial disagreement between the 2 methods. The logistic regressions showed that higher agreement between the dietary recalls and real-time tool, as defined by overlaps of 45 and 60 min, was significantly associated with earlier FEO. That is, the earlier the FEO was consumed in the day, the better the agreement between both dietary tools. These relationships were not present for the FEO agreements by 30 min, nor any definition of agreement for the LEO. To our knowledge, there is no data on the degree of agreement or reliability of ASA24 or other recall methods for the FEO, but our findings agree with previous data that has shown that the meal timing reliability (calculated by Intraclass correlation coefficients (ICC), Light’s Kappa estimates, and 95% CI’s) of 24-h dietary recalls was “slight” (CI = 0.01-0.20) and “fair” (CI = 0.21-0.40) for the LEO when this variable was evaluated on 3 separate datasets from 3 epidemiologic studies.²⁵ These findings underscore the limitations of recall methods to capturing the precise timing of eating events, particularly those occurring later in the day.

The strengths of this study include the serial meal timing assessments over a 2-week period with simultaneous meal pattern measurements via a recall tool and a real-time tool under free-living conditions in older adults of racially and ethnically diverse backgrounds. Nevertheless, although the real-time smartphone application mCC is of great value to assess meal timing, and has been validated, it still relies on user-reported data. Furthermore, although all participants were instructed to log meals in real-time with time-stamped photographs, same-day retroactive logging was permitted, however, coaching on this function emphasized the importance to limit retroactive text entries to unavoidable circumstances (eg, smartphone device malfunction or battery depletion). Similarly, good adherence to app use prior to enrollment was required, therefore, these findings may have reduced applicability in groups that are not comfortable with the use of a mobile application to track their meals. Nonetheless, passive tools like chewing detectors, wrist motion devices, and portable camera images are still in development, and although the subjectivity of manual-logging and self-motivated entries is still involved with a mobile app, the inaccuracies of a time-stamped photograph are limited to incomplete entries. Other limitations include the relatively small sample size with specific inclusion criteria for older individuals with overweight or obesity. Additionally, as we proceeded with secondary analyses of data collected for the parent trial for this pilot study, a power analysis was not performed, therefore, the results may not be generalizable to other populations.

Conclusion

Our results suggest that dietary recalls agreement with an app documenting meals in real-time is consistently below 75% in older adults. Inter-tool agreement was lower for meals consumed later in the day, and was higher with a higher number of recalls. This highlights the limitation of dietary recalls for the assessment of temporal eating patterns, particularly in populations that typically skip breakfast or are late eaters. While our data support the need to use tools allowing assessment of free-living dietary behavior in real-time, they also suggest that there is a great need to develop methods for passive monitoring of dietary behavior. Future studies should focus on cohorts with different age, adiposity levels and chronotypes, address whether agreement differs by meal classification (eg, snack, dinner, etc.), and once apps are more performant,⁴⁰ leverage objective measures of caloric and nutrient intake and integrate it with real-time meal timing assessments.

Supplemental Material

sj-docx-2-inq-10.1177_00469580261419162 – Supplemental material for Comparison of Self-Reported Dietary Recalls and Real-Time Tools to Track Mealtimes in Older Adults: A Pilot Study

Supplemental material, sj-docx-2-inq-10.1177_00469580261419162 for Comparison of Self-Reported Dietary Recalls and Real-Time Tools to Track Mealtimes in Older Adults: A Pilot Study by Leinys S. Santos-Báez, Omer Kazmi, Diana Díaz-Rizzolo, Collin J. Popp, Emily N.C. Manoogian, Satchidananda Panda, Bin Cheng and Blandine Laferrère in INQUIRY: The Journal of Health Care Organization, Provision, and Financing

Supplemental Material

sj-pdf-1-inq-10.1177_00469580261419162 – Supplemental material for Comparison of Self-Reported Dietary Recalls and Real-Time Tools to Track Mealtimes in Older Adults: A Pilot Study

Supplemental material, sj-pdf-1-inq-10.1177_00469580261419162 for Comparison of Self-Reported Dietary Recalls and Real-Time Tools to Track Mealtimes in Older Adults: A Pilot Study by Leinys S. Santos-Báez, Omer Kazmi, Diana Díaz-Rizzolo, Collin J. Popp, Emily N.C. Manoogian, Satchidananda Panda, Bin Cheng and Blandine Laferrère in INQUIRY: The Journal of Health Care Organization, Provision, and Financing

Footnotes

Acknowledgements

The authors express their gratitude to the study participants for their participation, and the NY-TREAT trial staff, Ana P. Sordi-Guth, Rabiah B. Borhan, and Danny G. DeBonis, for the recruitment, administrative and technical support.

ORCID iD

Leinys S. Santos-Báez

Ethical considerations

This study was approved by the Columbia University Institutional Review Board (IRB AAAS7791) on June 30, 2020. This research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki.

Consent to participate

All participants provided written informed consent prior to enrollment in the trial.

Consent for publication

Not applicable.

Author contributions

L-S.S-B. was involved in the conception of the study, design of the study, conduct of the study, results interpretation, and wrote the first draft of the manuscript. L-S.S-B. and O.K. performed data curation. L-S.S-B. and B.C. performed statistical analyses. B.L. provided project administration, resources, and supervision. All authors edited, reviewed and approved the final version of the manuscript.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the National Institute on Aging (NIA) R01AG065569-05, the National Center for Advancing Translational Sciences (NCATS) UL1TR001873. Santos-Báez was supported by R01AG065569-03S1 and the Naomi Russ Berrie Fellowship. Díaz-Rizzolo was supported by Fundación Alfonso Martin Escudero and the Naomi Russ Berrie Fellowship. Popp was supported by R01NR018916 and K99HL163474.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: S.P. receives author royalties from PenguinRandomHouse for the books “The Circadian Code” and “The Circadian Diabetes Code,” is a consultant and scientific advisory board member at Hooke London and WndrHlth, and is a co-founder of Circadian Biosystems. B.L. is editor UpToDate obesity section and Co-Editor ENDOTEXT obesity chapters.

Data availability statement

The datasets generated during and/or analyzed during the current study are not publicly available as the parent study is ongoing, but are available (as allowable according to institutional IRB standards) from the corresponding author on reasonable request.

Supplemental material

Supplemental material for this article is available online.

References

Sletten

Cappuccio

Davidson

Van Cauter

Rajaratnam

SMW

Scheer

FAJL

. Health consequences of circadian disruption. Sleep. 2020;43(1):1-3. doi:10.1093/sleep/zsz194

Xiao

Garaulet

Scheer

FAJL

. Meal timing and obesity: interactions with macronutrient intake and chronotype. Int J Obes. 2019;43(9):1701-1711. doi:10.1038/s41366-018-0284-x

Santos-Báez

Díaz-Rizzolo

Popp

, et al. Diet and meal pattern determinants of glucose levels and variability in adults with and without prediabetes or early-onset type 2 diabetes: a pilot study. Nutrition. 2024;16(9). doi:10.3390/nu16091295

Vujović

Piron

Qian

, et al. Late isocaloric eating increases hunger, decreases energy expenditure, and modifies metabolic pathways in adults with overweight and obesity. Cell Metab. 2022;34(10):1486-1498.e7. doi:10.1016/j.cmet.2022.09.007

Díaz-Rizzolo

Santos Baez

Popp

, et al. Late eating is associated with poor glucose tolerance, independent of body weight, fat mass, energy intake and diet composition in prediabetes or early onset type 2 diabetes. Nutr Diabetes. 2024;14(1):90. doi:10.1038/s41387-024-00347-6

Antoni

Robertson

Johnston

. A pilot feasibility study exploring the effects of a moderate time-restricted feeding intervention on energy intake, adiposity and metabolic physiology in free-living human subjects. J Nutr Sci. 2018;7:e22.

Anton

Lee

Donahoo

, et al. The effects of time restricted feeding on overweight, older adults: a pilot study. Nutrition. 2019;11(7). doi:10.3390/nu11071500

Lin

Cienfuegos

Ezpeleta

, et al. Time-restricted eating without calorie counting for weight loss in a racially diverse population: a randomized controlled trial. Ann Intern Med. 2023;176(7):885-895. doi:10.7326/m23-0052

Pavlou

Cienfuegos

Lin

, et al. Effect of time-restricted eating on weight loss in adults with type 2 diabetes: a randomized clinical trial. JAMA Netw Open. 2023;6(10):e2339337. doi:10.1001/jamanetworkopen.2023.39337

10.

Bingham

Gill

Welch

, et al. Comparison of dietary assessment methods in nutritional epidemiology: weighed records v. 24 h recalls, food-frequency questionnaires and estimated-diet records. Br J Nutr. 1994;72(4):619-643. doi:10.1079/bjn19940064

11.

Dewar

Cowan

Sala

. Forgetting due to retroactive interference: a fusion of Müller and Pilzecker’s (1900) early insights into everyday forgetting and recent research on anterograde amnesia. Cortex. 2007;43(5):616-634. doi:10.1016/s0010-9452(08)70492-1

12.

Subar

Kirkpatrick

Mittl

, et al. The automated self-administered 24-hour Dietary Recall (ASA24): a resource for researchers, clinicians and educators from the National Cancer Institute. J Acad Nutr Diet. 2012;112(8):1134-1137.

13.

Whitton

Collins

Mullan

, et al. Accuracy of energy and nutrient intake estimation versus observed intake using 4 technology-assisted dietary assessment methods: a randomized crossover feeding study. Am J Clin Nutr. 2024;120(1):196-210. doi:10.1016/j.ajcnut.2024.04.030

14.

Liu

Eso

Cook

O'Neill

Albarqouni

. Meal timing and anthropometric and metabolic outcomes: a systematic review and meta-analysis. JAMA Netw Open. 2024;7(11):e2442163. doi:10.1001/jamanetworkopen.2024.42163

15.

Panagiotou

Stefanou

Kourlaba

Athanasopoulos

Kassari

Charmandari

. The effect of time-restricted eating on cardiometabolic risk factors: a systematic review and meta-analysis. Nutrition. 2024;16(21). doi:10.3390/nu16213700

16.

Wilkinson

Manoogian

ENC

Zadourian

, et al. Ten-hour time-restricted eating reduces weight, blood pressure, and atherogenic lipids in patients with metabolic syndrome. Cell Metab. 2020;31(1):92-104.e5. doi:10.1016/j.cmet.2019.11.004

17.

Manoogian

ENC

Wilkinson

O'Neal

, et al. Time-restricted eating in adults with metabolic syndrome: a randomized controlled trial. Ann Intern Med. 2024;177(11):1462-1470. doi:10.7326/M24-0859

18.

Ghosh

Han

Raju

, et al. Integrated image and sensor-based food intake detection in free-living. Sci Rep. 2024;14(1):1665. doi:10.1038/s41598-024-51687-3

19.

Stankoski

Kiprijanovska

Gjoreski

, et al. Controlled and real-life investigation of optical tracking sensors in smart glasses for monitoring eating behavior using Deep Learning: cross-sectional study. JMIR Mhealth Uhealth. 2024;12:e59469. doi:10.2196/59469

20.

Popp

Wang

Hoover

, et al. Objective determination of eating occasion timing: combining self-report, wrist motion, and continuous glucose monitoring to detect eating occasions in adults with prediabetes and obesity. J Diabetes Sci Technol. 2024;18(2):266-272. doi:10.1177/19322968231197205

21.

Gill

Panda

. A smartphone app reveals erratic diurnal eating patterns in humans that can Be modulated for health benefits. Cell Metab. 2015;22(5):789-798. doi:10.1016/j.cmet.2015.09.005

22.

Manoogian

ENC

Wei-Shatzel

Panda

. Assessing temporal eating pattern in free living humans through the myCircadianClock app. Int J Obes. 2022;46(4):696-706. doi:10.1038/s41366-021-01038-3

23.

Santos-Báez

Garbarini

Shaw

, et al. Time-restricted eating to improve cardiometabolic health: the New York time-restricted EATing randomized clinical trial - protocol overview. Contemp Clin Trials. 2022;120:106872. doi:10.1016/j.cct.2022.106872

24.

Chow

Manoogian

ENC

Alvear

, et al. Time-restricted eating effects on body composition and metabolic measures in humans who are overweight: a feasibility study. Obesity. 2020;28(5):860-869. doi:10.1002/oby.22756

25.

Peterson

Lee

Huybrechts

, et al. Reliability estimates for assessing meal timing derived from longitudinal repeated 24-hour dietary recalls. Am J Clin Nutr. 2023;117(5):964-975. doi:10.1016/j.ajcnut.2023.02.026

26.

Bland

Altman

. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1(8476):307-310.

27.

Lichtman

Pisarska

Berman

, et al. Discrepancy between self-reported and actual caloric intake and exercise in obese subjects. N Engl J Med. 1992;327(27):1893-1898. doi:10.1056/nejm199212313272701

28.

Lissner

Troiano

Midthune

, et al. OPEN about obesity: recovery biomarkers, dietary reporting errors and BMI. Int J Obes. 2007;31(6):956-961. doi:10.1038/sj.ijo.0803527

29.

Price

Paul

Cole

Wadsworth

MEJ

. Characteristics of the low-energy reporters in a longitudinal national dietary survey. Br J Nutr. 1997;77(6):833-851. doi:10.1079/bjn19970083

30.

Freedman

Commins

Moler

, et al. Pooled results from 5 validation studies of dietary self-report instruments using recovery biomarkers for energy and protein intake. Am J Epidemiol. 2014;180(2):172-188. doi:10.1093/aje/kwu116

31.

Johansson

Wikman

Ahrén

Hallmans

Johansson

. Underreporting of energy intake in repeated 24-hour recalls related to gender, age, weight status, day of interview, educational level, reported food intake, smoking habits and area of living. Public Health Nutr. 2001;4(4):919-927. doi:10.1079/phn2001124

32.

Klesges

Eck

Ray

. Who underreports dietary intake in a dietary recall? Evidence from the Second National Health and Nutrition Examination Survey. J Consult Clin Psychol. 1995;63(3):438-444. doi:10.1037//0022-006x.63.3.438

33.

Goris

Westerterp-Plantenga

Westerterp

. Undereating and underrecording of habitual food intake in obese men: selective underreporting of fat intake. Am J Clin Nutr. 2000;71(1):130-134. doi:10.1093/ajcn/71.1.130

34.

Lafay

Mennen

Basdevant

, et al. Does energy intake underreporting involve all kinds of food or only specific food items? Results from the Fleurbaix Laventie Ville Santé (FLVS) study. Int J Obes Relat Metab Disord. 2000;24(11):1500-1506. doi:10.1038/sj.ijo.0801392

35.

Gemming

Ni Mhurchu

. Dietary under-reporting: what foods and which meals are typically under-reported? Eur J Clin Nutr. 2016;70(5):640-641. doi:10.1038/ejcn.2015.204

36.

Macdiarmid

Blundell

. Assessing dietary intake: who, what and why of under-reporting. Nutr Res Rev. 1998;11(2):231-253. doi:10.1079/NRR19980017

37.

Bärebring

Palmqvist

Winkvist

Augustin

. Gender differences in perceived food healthiness and food avoidance in a Swedish population-based survey: a cross sectional study. Nutr J. 2020;19(1):140. doi:10.1186/s12937-020-00659-0

38.

Kupis

Johnson

Hallihan

Olstad

. Assessing the usability of the automated self-administered Dietary Assessment Tool (ASA24) among low-income adults. Nutrition. 2019;11(1). doi:10.3390/nu11010132

39.

Dashti

Scheer

FAJL

Saxena

Garaulet

. Timing of food intake: identifying contributing factors to design effective interventions. Adv Nutr. 2019;10(4):606-620. doi:10.1093/advances/nmy131

40.

Laferrère

. Can we measure food intake in humans? Int J Obes. 2023;47(5):391-392. doi:10.1038/s41366-023-01282-9

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.22 MB

0.03 MB