Timing Patterns of Coffee, Dairy, and Protein Intake and Their Measurement Reliability Using Repeated 24-Hour Dietary Recalls in the iDATA Study

Principal Investigator

Name
Qibin Qi

Degrees
Ph.D.

Institution
Albert Einstein College of Medicine

Position Title
Professor of Epidemiology

Email
qibin.qi@einsteinmed.org

About this CDAS Project

Study
IDATA (Learn more about this study)

Project ID
IDATA-98

Initial CDAS Request Approval
Feb 23, 2026

Title
Timing Patterns of Coffee, Dairy, and Protein Intake and Their Measurement Reliability Using Repeated 24-Hour Dietary Recalls in the iDATA Study

Summary
The timing of food and beverage intake has emerged as an important dimension of dietary behavior with potential relevance to circadian alignment and metabolic regulation. Beyond what and how much people eat, when intake occurs across the 24-hour day may influence physiological processes through circadian mechanisms. However, the characterization of dietary timing patterns and the reliability of their measurement remain limited, largely due to the lack of datasets that combine detailed time-stamped dietary intake with repeated assessments.
Particularly, coffee, dairy, and protein are dietary components that influence sleep health, exhibit substantial within-person variability across the day, and are relevant to ongoing research in chrononutrition. The Interactive Diet and Activity Tracking in AARP (iDATA) study provides a unique opportunity to address this gap. iDATA collected repeated 24-hour dietary recalls with detailed timing information, along with sleep data and demographic characteristics, enabling the systematic characterization of intake timing for dairy, protein, and coffee.
The overarching objective of this project is to apply a multidimensional dietary timing framework to iDATA 24-hour recall data to: (1) generate empirical timing patterns of dairy and protein intake, (2) characterize coffee intake timing patterns, and (3) evaluate the reliability of timing metrics derived from different numbers of dietary recalls.
First, we will derive timing features of coffee intake across the 24-hour day using a multidimensional timing framework that captures intake timepoints, temporal distribution, and eating window duration. These dimensions will be used to generate coffee intake timing clusters. This approach allows a comprehensive description of how coffee consumption is distributed across daily time periods rather than relying on single summary measures.
Second, we will derive timing features of dairy and protein intake across the 24-hour day using a multidimensional timing framework that captures intake timepoints, temporal distribution, eating window duration, and day-level patterns. This approach allows a comprehensive description of how dairy and protein consumption are distributed across daily time periods rather than relying on single summary measures.
Third, we will evaluate the reliability of timing metrics for coffee, dairy, and macronutrient intake by comparing estimates derived from varying numbers of 24-hour recalls. By quantifying within- and between-person variability and estimating reliability coefficients, this aim will inform the minimum number of recalls required to reasonably capture habitual food- and nutrient-specific intake timing patterns.
Collectively, this project will provide a rigorous characterization of dietary timing behaviors in iDATA and generate empirical evidence on the measurement reliability of timing-based dietary metrics. With external validation using the data from the National Health and Nutrition Examination Survey and the Hispanic Community Health Study/Study of Latinos, these findings will directly inform future chrononutrition studies and support the appropriate use of repeated 24-hour dietary recalls for evaluating intake timing in epidemiologic research.

Aims

The overall goal of this project is to characterize dietary intake timing patterns for coffee, dairy, and protein, and to assess the reliability of timing-based metrics using repeated 24-hour dietary recalls in the iDATA study. Leveraging detailed intake timing information and sleep data, the specific aims are as follows:
Aim 1. Generate coffee intake timing patterns using repeated 24-hour dietary recalls. Coffee intake is highly time-dependent and typically occurs in discrete episodes, making it an ideal exposure for timing analyses. We will characterize coffee intake timing by deriving metrics such as timepoints (daily coffee intake midpoint; whether consume coffee at 6:00-9:59 am (morning), 10:00 am-1:59 pm (noon), 2:00-5:59 pm (afternoon), 6:00-9:59 pm (evening), and 10:00 pm-5:59 am (overnight); whether consume coffee 1, 2, and 3 hours before sleep), duration (intervals from first to last coffee intake, from wake to first coffee intake, and from last coffee intake to sleep), and distribution (proportions of daily coffee intake across five clock-time blocks [morning, noon, afternoon, evening, and overnight] and four wake-period quartiles). Using these features, we will identify common coffee intake timing patterns through data-driven approaches (e.g., two-step clustering). We will also examine demographic correlates of identified coffee timing patterns to describe population heterogeneity.
Aim 2. Generate timing patterns of dairy and protein intake using a multidimensional dietary timing framework. Dairy foods and protein are consumed at varying times across the day and may follow distinct temporal patterns that are not captured by total intake alone. Using repeated 24-hour dietary recalls, we will derive timing metrics for dairy and protein intake, including intake timepoints, duration, distribution, and day-level intake patterns. These features will be integrated within a multidimensional timing framework to comprehensively characterize habitual dairy and protein intake timing at the individual level.
Aim 3. Evaluate the reliability of dairy, coffee, and macronutrient intake timing metrics using different numbers of dietary recalls. Reliable measurement of intake timing is essential for future etiologic and interventional studies. Using repeated recalls, we will quantify within- and between-person variability for timing metrics of dairy, coffee, and macronutrient intake (carbohydrate, protein, and fat). We will estimate reliability coefficients for timing features derived from varying numbers of recalls (e.g., 1, 2, 3, or more days) to assess how reliability improves with additional dietary assessments. This aim will provide empirical guidance on the number of recalls needed to capture habitual intake timing patterns with acceptable reliability.
Impact: This project will establish a structured approach to characterizing dietary intake timing and evaluating its measurement reliability using iDATA. The analyses will be replicated in the National Health and Nutrition Examination Survey and the Hispanic Community Health Study/Study of Latinos, and the findings will inform best practices for dietary timing assessment and support future chrononutrition research using large-scale observational data.

Collaborators

Qibin Qi Albert Einstein College of Medicine
Qibin Qi Albert Einstein College of Medicine
Yanbo Zhang Albert Einstein College of Medicine