Two independent series of 12-week N-of-1 studies will be conducted to separately assess
the effects of two primary interventions, black tea flavoured tea drink (containing
bioactive compounds equivalent to naturally occurring levels in black tea), and green tea
flavoured tea drink (containing bioactive compounds equivalent to naturally occurring
levels in green tea), against two controls each (caffeine only, and placebo, flavoured
according to the primary interventions), on cognition, mood, and sleep.
For each primary intervention, the 12-week study duration will comprise of four,
three-week cycles and each cycle will comprise of three, one-week intervention periods.
The order of the three interventions will be determined separately for each participant
by randomly selecting four of the six possible combinations of three interventions, and
such that no intervention is given in consecutive weeks. Each one-week intervention
period will be split into six test days and one washout day. The washout day will be a
minimum of 24 hours (00:00 to 23:59). On each of the test days, participants will be
asked to drink the tea intervention allocated to that week, as often as they would like.
Participants will be asked to refrain from drinking any other tea products, and
caffeinated, or decaffeinated, drinks. On washout days participants will also be asked to
refrain from drinking the tea products provided as part of the study.
The randomisation process to determine intervention sequences will be conducted by a
third party who will not be involved in running of the study or data analysis. Each tea
intervention will be coded prior to being delivered to the Rowett Institute and the file
linking each tea code to the tea will be sent from Lipton Teas and Infusions to the
Quality Assurance Manager for the Rowett Institute. The principal and sub-investigators
will be informed which three tea codes relate to the three green tea and three black tea
interventions for the purposes of running the study. The specific tea interventions
assigned to each code will only be revealed to the principal and sub-investigators once
the data analysis for all participants has been completed.
At three time points on each study day, participants will be prompted to complete a
series of questionnaires and cognitive tasks which will form one measurement point. These
time points will be decided on a per participant basis and where possible one measurement
point will fall within each of the following time periods: 0800-1000, 1200-1400,
1600-1800. For participants who work night shifts these time periods will not apply,
however, the time intervals between each measurement point will be maintained where
possible. The cognitive tasks will be the digit vigilance task and attention switching
task, and questionnaires include a sleep questionnaire, mood questionnaire, personalised
questionnaire, and tea recall questionnaire. The investigators will use Gorilla
Experiment Builder to deliver the questionnaires and cognitive tasks. The sleep
questionnaire will only be included in the first measurement point of each day. The
investigators will also ask participants to note accidental consumption of any tea or
caffeine products not related to the study, but participants will not excluded from the
study if they have consumed a product unrelated to the study.
Once potential participants have been screened against the inclusion and exclusion
criteria and informed consent has been obtained, participants will complete a series of
questionnaires that will be used to identify variables that are likely to show some
degree of personal variability over time and may influence their cognition/attention,
mood, sleep, and/or compliance to the intervention. Four validated questionnaires will be
used: the UK Short-form 12 health questionnaire, the short-form International Physical
Activity Questionnaire, the Pittsburgh Sleep Quality Index, and the trait anxiety arm of
the State Trait Anxiety Inventory in order to provide an overview of their typical
lifestyle-related behaviours. Participants will also be asked to complete a questionnaire
detailing their tea consumption and consumption of other caffeinated food and drinks and
a questionnaire to establish other lifestyle factors e.g., work routine, dependents,
whether the participant lives with another person. The tea consumption questionnaire will
also inform whether the participant will be included in the black tea or green tea study.
Assessment of these time-dependent variables during the study is known as ecological
momentary assessment and will be implemented using semi-personalised short questionnaires
delivered via Gorilla Experiment Builder at each of the three daily measurement points
over the course of the study.
After receiving their questionnaire responses, an in-person meeting with each participant
will be arranged at a time convenient to them. This meeting will be used to determine the
prompting times and to familiarise participants with the Gorilla Experiment Builder test
platform, provide them with a PRO-Diary, and the tea products for the first cycle of the
study.
Subsequent visits will occur every three weeks - at the end of each cycle. Participants
will receive their tea products, the PRO-Diary device will be swapped with a fully
charged device, and the visits will provide another opportunity for participants to ask
questions or express any concerns they may have. At the end of week 12, participants will
be invited to the Rowett for a final debrief and to return the PRO-Diary device.
Descriptive, univariable, and multivariable analyses will be performed in R. Owing to the
personalised nature of the ecological momentary assessment and the longitudinal nature of
the outcomes, individualised dynamic modelling will be the primary method of analysis for
investigating both trends in variables over time, and the effect of individual-specific
predictors on the attentional aspect of cognition (mean reaction times and accuracy for
the attention switching task and digit vigilance task), sleep, and mood.
Individualised dynamic modelling analyses will be conducted for each participant to
capture variation in the primary and secondary outcomes, and to develop unique regression
models to consider the influence of individualised time-dependent predictors on the daily
outcomes. Dynamic modelling includes two sets of covariates. The first set of covariates
include the potential predictors exogenous to the participant (e.g., time, day of the
week) and endogenous (e.g., mood, length of sleep) on the model, to identify which
factors affect behaviour while controlling for the effect of time. Dynamic covariates are
also included to summarise the effect of the past in present response and predictors by
including lagged variables (values from previous day or two days before) of both the
response and predictors, which enables us to see how much the previous state of the
individuals can affect present outcomes.
