A systematic review of ecological momentary assessment studies of appetite and affect in the experience of temptations and lapses during weight loss dieting

Summary Dietary temptations and lapses challenge control over eating and act as barriers toward successful weight loss. These are difficult to assess in laboratory settings or with retrospective measures as they occur momentarily and driven by the current environment. A better understanding of how these experiences unfold within real‐world dieting attempts could help inform strategies to increase the capacity to cope with the changes in appetitive and affective factors that surround these experiences. We performed a narrative synthesis on the empirical evidence of appetitive and affective outcomes measured using ecological momentary assessment (EMA) during dieting in individuals with obesity and their association with dietary temptations and lapses. A search of three databases (Scopus, Medline, and PsycInfo) identified 10 studies. Within‐person changes in appetite and affect accompany temptations and lapses and are observable in the moments precipitating a lapse. Lapsing in response to these may be mediated through the strength of a temptation. Negative abstinence‐violation effects occur following a lapse, which negatively impact self‐attitudes. Engagement in coping strategies during temptations is effective for preventing lapses. These findings indicate that monitoring changes in sensations during dieting could help identify the crucial moments when coping strategies are most effective for aiding with dietary adherence.


S.1.1 Descriptions of quality assessments
A modified Newcastle-Ottawa cohort scale adapted for cross sectional studies (NOS; Modesti et al., 2016) We assessed quality of included studies using the Newcastle-Ottawa cohort scale adapted for cross sectional studies which rates quality of selection, comparability, and outcome (NOS for cross-sectional scale; Modesti et al., 2016).This was modified to assess the qualities of EMA observational investigations.
Selection is comprised of four items with a maximum score of five.These items assess sample representativeness, sample size, non-respondents and ascertainment of the exposure, all with a score of one except the latter item which has a maximum score of two.Items on sample representativeness were modified to assess representativeness of within-person assessments as these are the target of inference in EMA.A point was awarded for representativeness if RAs were utilised in the study design, and the time scheduling used for these assessments were either completely random times throughout the day (e.g.notified to perform an assessment at four random points throughout the day at any given time) or random timeframes (e.g.notified to complete between the hours of 8-10am, 10-12pm, and 12-4pm).A point was not awarded if RAs were not utilised or the time scheduling used for assessments were at fixed times (e.g.complete an assessment at 8am, 4pm, and 6pm).
Comparability is measured using two items with a maximum score of two.These items assess the most important factor for comparability (i.e.statistical analyses to account for clustering such as mixed models) as well as other important factors for comparability (i.e.controlling for differences in compliance and response rate during analyses, providing appropriate instructions to participants regarding assessment procedure and definitions of temptations and lapses).
Outcome has a maximum score of three and is measured by two items which assessed the type of assessment for the outcome used (maximum score of two) and appropriateness/description of statistical tests used (maximum score of one).

Checklist for reporting EMA studies (CREMAS; Liao et al., 2016)
An adapted STROBE checklist for EMA studies was used to assess the quality of reporting of the included studies.The STROBE is a commonly used checklist of items for observational studies.It contains 22 items that relate to the title, abstract, introduction, methods, results, and discussion sections of papers with the goal to improve the quality of reporting.Building on the STROBE checklist and the EMA design guidelines by Stone and Shiffman 27 , a comprehensive checklist of specific items to be reported for EMA studies was also developed.In addition to STROBE checklist, additional methodological features, responses and compliance information is assessed with 16 items attaining to the following 5 main areas: Sampling and measures: sample characteristics and tools used in the EMA protocol Schedule: monitoring periods (number of waves from which data were collected), duration (number of days that each monitoring period lasted), prompt frequency (frequency of EMA prompts per day), and prompt interval (the time between each EMA prompt) Technology and administration: use or lack of technology and method of administration of EMAs Prompting strategy: methods used to cue participants-interval contingent (EMA prompts were set for certain intervals that were not random), random interval contingent (EMA prompts were set to be randomized throughout the day), event based (EMAs were recorded when eating occasions or physical activity occurred), or evening report (EMAs administered in the evenings to summarize the events of the day) Response and compliance: participation rate, gathered data, missing data (i.e., unanswered and/or unprompted EMA surveys), latency (i.e., the time period between when participants receive an EMA prompt and when the EMA is answered), and attrition (i.e., the number of participants who dropped out of the study for any reason)

Results for NOS for EMA (Table S.1.2a)
Most studies were of reasonable quality as assessed by the NOS.However, they uniformly performed poorly on sample size justification.No studies mentioned how sample sizes were determined or if power analyses were performed.If appropriate multilevel forms of analyses are employed to account for repeated measurements, large sample sizes are less of a problem than for other statistical approaches, as units of analyses are the within-person assessments and these are usually sufficiently powered.However, in multilevel analyses the major restricting factor is usually the group level sample size as these are usually lower than the number of within-person assessments, contain a greater amount of variation than withinperson assessments, and have been shown to produce biased group-level estimates in smaller sample sizes McNeish & Stapleton (2016).Future investigations in this area should cite guidelines that describe appropriate group-level sample sizes for multilevel modelling (Maas & Hox, 2015) to avoid scepticism surrounding sample sizes.
Most investigations used appropriate methods of analyses to account for multileveled datasets.Accounting for nesting of datapoints is important in repeated measures designs as within-person assessments are likely to be highly correlated which violates the assumption of independence of errors as datapoints.Most studies also reported appropriate statistics though a few failed to report confidence intervals for their associated p-values.
Reporting of differences between respondents and non-respondents or controlling for these in analyses was generally good in included studies, and most reported average compliance rates of EMA assessment protocol.Compliance with assessment protocol is a limitation of EMA as rates can have an impact on the statistical power of the study, particularly if data are missing not at random and are systematic (e.g.missing prompts due to working hours) (Graham, 2009).There is currently no 'gold-standard' rate of compliance, though a rule of thumb is that compliance rates of at least 80% are considered acceptable (Jones et al., 2019).Providing descriptive information on compliance rates is essential as it may indicate whether a particular EMA assessment procedure may be too burdensome and allows for reviews to be conducted that examine overall compliance rates across studies as well as predictors that may influence compliance that could be used to facilitate higher rates (Jones et al., 2019).
All studies used subjective measures and self-report which are associated with information bias such as socially desirability and demand characteristics.Furthermore, one problem of EMA investigations relate to reactivity to experimental procedure which could also introduce bias into measures (Rowan et al., 2007).However, given the subjective nature of appetite ratings these experiences would be difficult to measure otherwise.Future investigations could combine both free-living and laboratory-based approaches to validate changes in average levels of real-world subjective appetite ratings such as hunger throughout weight loss with physiological markers.

Results for CREMAS (Table S.1.2b)
Most (7/10) studies identified the paper as an EMA study in the title, however all performed better providing a rationale for EMA in the introduction (10/10).
Regarding methodology, only three studies (7/10) did not report (or refer to a primary study which detailed) how participants were trained in the EMA protocol.All (10/10) papers referred to the technology which was used to administer EMA, and 9 out of 10 papers referred to the wave duration (i.e. 1 period of EMA) and all studies reported the number of days each wave lasted.All studies reported the type of prompts utilised, and 9 out of 10 studies reported how many random prompts were sent per day (if the study utilised this type of prompt).4 out of 10 studies utilised a design feature to control for EMA methodological limitations such as reactivity or participant burden.4 out of 10 studies reported on the participant attrition.All but one study reported on the results of the prompt delivery (e.g.how many random prompts were received/events were reported).No studies reported on the latency between receiving a prompt and reporting of the prompt.7/10 reported on the compliance rate of EMA prompts, and 4 out of 10 studies reported on whether missing data was related to time or demographic-related variables.
Regarding discussion points, 6 out of 10 studies reported the limitations of the study in light of using an EMA methodology, and all studies discussed the benefits of using EMA and how it helped achieved the desired aims of the paper.Eligibility criteria 6 Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale.

5-6
Information sources 7 Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched.

5
Search 8 Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated.

5
Study selection 9 State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis).

5-6
Data collection process 10 Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators.

6
Data items 11 List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made.

None
Additional analyses 16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which were pre-specified.

Study selection
17 Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram.

Figure 1
Study characteristics 18 For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations.

6
Risk of bias in individual studies12 Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at the study or outcome level), and how this information is to be used in any data synthesis.7Summarymeasures 13 State the principal summary measures (e.g., risk ratio, difference in means).NASynthesis of results 14 Describe the methods of handling data and combining results of studies, if done, including measures of consistency (e.g., I 2 ) for each meta-analysis.Risk of bias across studies 15 Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies).

Table S .1.2a
-Quality assessment of studies using modified Newcastle-Ottawa scales for assessing included studies of appetite measures with EMA during ER

Table 1
Risk of bias within studies19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12).