The study protocol is initiated after enrollment with a Food Frequency Questionnaire and
a Three-factor eating questionnaire conducted during the screening visit after the
participant's qualifications have been verified. After the questionnaires have been
completed, the study visits and food pick ups are scheduled according to the availability
of the subject and the study calendar. At the end of the screening appointment, the
subject selects the frozen dinners to be consumed the evening prior to the study visit.
The study is a crossover design in which investigators will assess the immediate effects
of a single meal of two commercially available snack bars on subjective feelings of
satiety over 2 hours and on energy and nutrient intake from a subsequent meal 2:15 to
2:30 following consumption of the bars. Study visits will be scheduled between 1 and 14
days apart.
Subjects will consume a standard meal between 7 and 8 PM the evening prior to each study
visit and complete a satiety questionnaire after finishing the meal. The subjects record
all food consumed the day prior to the study visit including the dinner meals provided by
the study. Participants will pick up food at the Indiana University-Bloomington School of
Public Health the day prior to the scheduled study visits. This visit consists of
dispensing the evening meals in lunch coolers packed with icepacks and the frozen meals
the participant selected in the screening appointment. The study visit protocol for the
following day is reviewed briefly. Appointment reminders are set up as emails and
calendar invitations instructing the participant to consume the meals between 7-8 PM that
night, to complete a satiety questionnaire, the link to which is embedded in the emails,
and to remain fasted except for water and non-caloric beverages until the study visit the
next day. Participants are offered a paper version of the satiety questionnaire if they
so desire. The satiety questionnaire asks the participants to evaluate their feelings of
hunger, fullness, desire to eat, and satisfaction with the previous meal. Desire to eat
is differentiated to the participant as a desire to consume food regardless of their
state of hunger while satisfaction is differentiated from fullness to the participants as
a psychological sensation of contentedness and not a physical sensation of fullness. The
same questionnaire is used for all time points during each study visit. Participants are
asked to refrain from consuming caffeine in the morning. A specific arrival time is set
for arrival at the Innovation Center the following day between 8 and 8:40 AM.
After the subject arrives at the IU-Bloomington Innovation Center, coolers and icepacks
are collected, compliance with the study protocol is confirmed. Next the diet record is
reviewed by study personnel. Participants may be asked to clarify entries made on the
diet record.
Participants are invited to a room where as many as 3 participants at a time will undergo
the study protocol. Subjects are allowed to interact with each other but are instructed
to refrain from discussing the snack bars. A baseline satiety questionnaire is completed
before the snack bars are offered with a bottle of water. Participants are instructed to
enter "1" with regard to satisfaction with the meal at this fasted time point. Ad libitum
water intake is allowed throughout the study visit. The participants are instructed to
consume two snack bars comprising 300 kcal as quickly and comfortably as possible. After
the subjects consume the bars, study personnel provide the participant with a sensory
evaluation, either printed or electronic, which asks the subject to evaluate the bars in
terms of overall taste, sweetness, saltiness, savory, desire to eat more, with the meal,
and fullness/satiation. Participants complete the sensory evaluation after having
consumed both bars. If participants are unable to consume both bars, the sensory
evaluation is not completed, and the experimental protocol is terminated for that
subject.
The subjects remain in the room are are allowed bathroom breaks, to entertain themselves
with their own audio devices, to read, and to work. Participants are prohibited from
bringing and consuming outside food and from exercising during the study protocol.
Satiety questionnaires are completed every 30 minutes from fasting. An electronic timer
is provided with a flow sheet denoting the times for the satiety questionnaires to be
completed. Participants write the times that they completed each iteration of the satiety
questionnaire. Participants are monitored periodically for compliance and electronic
recording of the questionnaires is monitored.
After 2 hours the penultimate satiety questionnaire for the study visit is completed.
Subjects are informed that they will be escorted to a lunch buffet consisting of grilled
hamburger patties, grilled chicken, brioche buns, sliced tomatoes, sliced medium cheddar
cheese, mayonnaise, mustard and ketchup, a fruit tray with individual portions of
watermelon, cantaloupe, strawberries and blueberries, a bowl of potato chips, a bowl of
mixed, salted nuts, chocolate chip cookies, and whipped cream to add to any of the food
they choose. Participants are instructed to "serve themselves as much of the food as they
care to consume and to consume it until comfortably full as quickly and comfortably as
possible." Participants log the time that they finished the meals and complete a final
satiety questionnaire after lunch. Each food item is weighed by study personnel before
and after the participants dine.
A second study visit is conducted that mirrors the first with a different type of snack
bar. Start times for the protocol and the buffet lunch are matched as closely to the
first study visit as possible.
Statistical Plan (This plan was articulated in the proposal and agreed upon when the
proposal was accepted, but was uploaded here on March 24, 2026)
This study will be conducted as a single-blind randomized cross-over design in which
consented participants will be randomly assigned to two treatment orders (Tx 1 then Tx 2
or Tx 2 then Tx 1) according to a block randomization scheme created by the
IU-Bloomington SPH Biostatistics Consulting Center. Enrolling 150 subjects, randomized to
which diet is received first, and assuming 20% attrition, a sample size of 120 completers
would provide 80% power for detecting significant differences (two-sided alpha=0.05)
between the two diets in the crossover study for "medium" effect sizes, Cohen's dz=0.268.
Rather than the classic d used for between-group comparisons, dz is the standardized
difference from first calculating a difference score between each person's responses in
each condition and computing dz=Mean(diff)/ SD(diff). This effect size is consistent with
previous literature where dz was observed to be 0.49. From multiple other crossover
studies using preloads and test meals, we conservatively estimate that standard
deviations of within-subject differences in intake between conditions is around 150 kcal.
This would imply that our study is powered to true mean differences even if the actual
causal effect is as small as 39 kcal.
The statistical plan includes the crossover data as described by Senn utilizing the
intent-to-treat (ITT) principle methodology previously described [7] which will include
all randomized subjects in the final analysis.[8] Linear mixed models will be used to
analyze the 1-100 scale satiety outcome collected across 5 SMS measurements in each
phase. Energy intake and macronutrient intake will be compared using a paired t test.
Subject will be considered a random effect.
The tertiary variables will be the standardized score from the 51-question "3 factor
eating" questionnaire and the 155 item FFQ. The model will include fixed effects for the
questionnaire score, visit (time points), and their interaction. Random effects will
account for repeated measures within subjects.
The 51-question survey to be collected includes both true/false and Likert scale (1-5)
questions. To standardize the scores we will do the following:
True/False Questions: We will convert responses to a numerical scale (True = 1,
False = 0).
Likert Scale Questions: These questions are already on a numerical scale (1 to 5)
and will be used as-is.
IRB#: 24245 Version Date: 10/25/2024 Page 15 of 27
Standardization: To standardize the scales we will calculate z-scores for each
question type to. This standardization method involves subtracting the mean and
dividing by the standard deviation for each question.
Composite Score: We will then combine the standardized scores across all 51
questions by averaging them to create a single composite score for each participant
for this survey.
Interaction Effects If the interaction between the questionnaire scores and visit, or
visit alone is significant, we will then generate estimated marginal means and pairwise
comparisons for the satiety scale outcome at each visit.
Model Variations Two different models will be considered for the outcomes: i. Unadjusted
Model: This model will include only the primary explanatory variables (standardized
questionnaire score, visit, and their interaction). ii. Adjusted Model: This model will
adjust for additional covariates such as age, sex, and baseline health status.
Area Under the Curve (AUC) Finally, we will also apply a SAS macro for obtaining Areas
Under the Curve (AUC) for the 1-100 scale across the 5 measurements during the test day.
The SMS the evening prior will be used as a covariate if necessary. These AUC values will
be statistically compared by the macros to assess the overall relationship between the
standardized 3 factor eating questionnaire score and the FFQ on the AUC of the outcome
across both visits following
https://www.lexjansen.com/wuss/2004/posters/c_post_the_sas_calculations_.pdf).
Gadbury GL, Coffey CS, Allison DB. Modern statistical methods for handling missing
repeated measurements in obesity trial data: beyond LOCF. Obes Rev 2003;4:175-84.
https://doi.org/10.1046/j.1467-789X.2003.00109.x.
Senn SS. Cross-over Trials in Clinical Research. 2nd edition. Chichester, Eng. ; New
York: Wiley; 2002.
