Trial Review

The safety and scientific validity of this study is the responsibility of the study sponsor and investigators. Listing a study does not mean it has been endorsed by the ANZCTR. Before participating in a study, talk to your health care provider and refer to this information for consumers
Trial registered on ANZCTR


Registration number
ACTRN12618001409268
Ethics application status
Approved
Date submitted
16/08/2018
Date registered
22/08/2018
Date last updated
13/08/2019
Date data sharing statement initially provided
9/11/2018
Date results information initially provided
13/08/2019
Type of registration
Prospectively registered

Titles & IDs
Public title
Randomised Controlled Trial to reduce unhealthy snacking: evaluation of planning tools and method of allocation in a sample with diverse health literacy and diabetes or high BMI
Scientific title
RCT to reduce unhealthy snacking: evaluation of planning tools and method of allocation in a sample with diverse health literacy and diabetes or high BMI
Secondary ID [1] 295826 0
None
Universal Trial Number (UTN)
U1111-1219-0835
Trial acronym
Linked study record

Health condition
Health condition(s) or problem(s) studied:
Diabetes 309269 0
Overweight/obesity 309270 0
Unhealthy snacking 309271 0
Condition category
Condition code
Metabolic and Endocrine 308142 308142 0 0
Diabetes
Diet and Nutrition 308143 308143 0 0
Obesity

Intervention/exposure
Study type
Interventional
Description of intervention(s) / exposure
This study is testing the effects of two intervention factors, all conducted online through a survey platform:
1) Allocation method
2) Type of action plan
Allocation method:
Arm 1: Random (Arm A)
Arm 2: Screening (Arm B)
Arm 3: Choice (Arm C)

Type of action Plan
Arm 1: literacy sensitive action plan (labelled for participants as Smart Snacking 101 (basic plan))
Arm 2: standard action plan (labelled for participants as "Smart Snacking Pro (advanced plan))


ALLOCATION METHODS
1. RANDOM (Arm A): Participants randomised to the ‘random’ arm will be further randomised to either 1) assess their prior preferences (i.e. their preferred action plan; Arm A1), and then randomised to the standard or literacy-sensitive action plan; or 2) randomised to the standard or literacy-sensitive action plan without assessment of prior preferences (Arm A2). Prior preference will be measured using the same format as in the ‘participant choice’ arm (Arm C) with additional text stating that participants may not receive their preferred tool.
2. SCREENING (Arm B): Allocation is based on health literacy (NVS) scores. Those scoring less than 4 (scores indicative of inadequate health literacy) were allocated to the literacy sensitive action plan, and the remaining participants were allocated to the standard action plan. Participants will be told that, based on their responses, the researchers have selected an action plan tool that is most suitable for them.
3. CHOICE (Arm C): Participants are provided with a brief description of the action plans and select the plan they will use. Participants have the option of selecting ‘Unsure’ to allow for undecided participants. Participants will then be presented with an alternative description of the study and asked again to make a choice. Participants will be informed that if they select ‘Unsure’ again, the researchers will select a plan for them. In doing so, participants will be randomised to an action plan.

TYPE OF ACTION PLAN:
1. LITERACY SENSITIVE ACTION PLAN: This commences with the text: “We want you to plan how you will change your unhealthy snacking behaviour each day because forming plans has been shown to improve snacking habits”. The intervention consists of 4 steps that guide the user through the process of developing an appropriate plan:
a. Step1: Sometimes we snack because we are hungry, but there are lots of other reasons too. Think about your snacks in the last week. Below is a list of ‘snack moments.’ These are times when people tend to choose unhealthy snacks or eat too much. Choose 3 snack moments from the list that happened to you the most often in the last week. [List of snack moments].
b. Step 2: Below are your top 3 snack moments. Some snack moments will be more important than others. Choose the 1 that you would be happiest to change. [User chooses from 3 previously selected snack moments]
c. Step 3: Great! Your most important snack moment was snacking because you are bored. The last step is to come up with a plan! Choose the solution that you think will work best for you. Drag it into the space on the right. [List of solutions]
d. Step 4: Imagine how your plan might feel.[examples of scenarios when this might happen]. The final step is to make sure the plan is realistic. How hard do you think it will be to do this plan for the next month [Slider from very easy to very hard. If the user selects a number greater than or equal to 7 they will be prompted to revise the plan]
2. STANDARD ACTION PLAN (instructions): We want you to plan how you will change your unhealthy snacking behaviour each day because forming plans has been shown to improve snacking habits. You are free to choose how you do this but we want you to formulate your plans in as much detail as possible. Pay attention to the situations in which you will implement (carry out) these plans. Focus on situations when you are not hungry but find yourself snacking.

For both types of action plan interventions, immediately before submitting the baseline survey, participants will be presented with their plan a final time, instructed to write down, take a screenshot or make a copy of it, and will be asked to indicate that they have a copy of the plan.

All participants will be emailed a reminder of their personal plan at baseline (within the first week), and before then end of the 2nd and 3rd weeks. Participants will complete a follow-up survey after 4 weeks.
Both interventions will be administered using an online survey hosted by Qualtrics. At no point do participants have any direct interaction from researchers. Email reminders will be automated using a mail merge function.
The baseline intervention will be delivered initially over a short period (approximately 1 week) in order to ascertain that the survey is collecting data correctly with a smaller number of participants (N=~200). As the sample size estimation assumes an equal proportion of participants choosing the literacy sensitive and standard action plans in the choice arm, and an equal proportion of participants identified as having low and high health literacy in the screening arm, at this point these proportions will be calculated. If it is found that these proportions are substantially different from 50%, the sample size will be adjusted to ensure adequate power is maintained. After these participants are recruited, further recruitment will be paused until follow up data is collected from this initial sample, again to ensure that the survey is correctly collecting data. We expect that recruitment of the remaining participants could be completed within a period of 3 weeks.
Adherence to the intervention in both action planning groups will be ensured by forcing a response for items that help the participant to create their plan. Coding (see analysis section) of standard action plans will assess adherence to the intervention.
Intervention code [1] 312159 0
Behaviour
Intervention code [2] 312181 0
Lifestyle
Comparator / control treatment
The control for method of allocation (screening tool, free choice) is the random allocation arm.
The control for type of action plan (health-literate action plan) is the standard action plan arm.
Control group
Active

Outcomes
Primary outcome [1] 307117 0
Updateself-reported unhealthy snacking (average number of serves of unhealthy snacks per week). The measure consists of 7-items measure of snacking was based on a diet score developed and validated by a well-established and highly regarded Australian scientific organization (CSIRO) (Hendrie et al., 2016). Items were drawn from the ‘discretionary foods’ category which the Australian Guidelines to Healthy Eating define as 'not considered necessary for a healthy diet’. Alcohol and takeaway foods were excluded in this study as the focus was on ‘snacks. ’ Participants answered how many serves of each category of unhealthy snacks they had eaten in the past month. Participants could answer according to the number of serves per day, week or month. Average weekly servings of unhealthy snacks were calculated from these scores. Timepoint: 4-week follow-up (also collected at baseline. Analysis will control for baseline unhealthy snacking). (Hendrie, G. A., Baird, D., Golley, R. K., & Noakes, M. (2017). The CSIRO Healthy diet score: an online survey to estimate compliance with the australian dietary guidelines. Nutrients, 9(1), 47.)
Timepoint [1] 307117 0
follow-up (4 weeks after baseline, administered through online survey)
Secondary outcome [1] 350749 0
Difficulty using the planning tool. A single item asked participants to rate how hard it was to use the planning tool (1=not at all hard, 5=extremely hard). This was measured immediately following the intervention.
Timepoint [1] 350749 0
baseline (measured immediately after the intervention)
Secondary outcome [2] 350750 0
intention, based on the 3-item measure of intention used by (Verhoeven, Adriaanse, de Ridder, de Vet, & Fennis, 2013; Verhoeven, Adriaanse, de Vet, Fennis, & de Ridder, 2014), adapted to refer specifically to unhealthy snacking. Responses will be recorded on a 7-point Likert scales (strongly disagree to strongly agree).

Verhoeven, A. A. C., Adriaanse, M. A., de Ridder, D. T. D., de Vet, E., & Fennis, B. M. (2013). Less is more: The effect of multiple implementation intentions targeting unhealthy snacking habits. European Journal of Social Psychology, 43(5), 344-354. doi:10.1002/ejsp.1963
Verhoeven, A. A. C., Adriaanse, M. A., de Vet, E., Fennis, B. M., & de Ridder, D. T. D. (2014). Identifying the ‘if’ for ‘if-then’ plans: Combining implementation intentions with cue-monitoring targeting unhealthy snacking behaviour. Psychology & Health, 29(12), 1476-1492. doi:10.1080/08870446.2014.950658
Timepoint [2] 350750 0
4-week follow-up (also collected at baseline before and immediately after the intervention
Secondary outcome [3] 350751 0
habit strength. This measure is based on Verplanken & Orbell's (2003) 12-item measure of habit strength, adapted to refer specifically to unhealthy snacking. Responses will be recorded on a 7-point Likert scales (strongly disagree to strongly agree).

Verplanken, B., & Orbell, S. (2003). Reflections on Past Behavior: A Self-Report Index of Habit Strength. Journal of Applied Social Psychology, 33(6), 1313-1330. doi:10.1111/j.1559-1816.2003.tb01951.x
Timepoint [3] 350751 0
4-week follow-up (also collected at baseline)
Secondary outcome [4] 350754 0
Perceived extent of unhealthy snacking in the previous week (2-item, 7-point likert scale)
Timepoint [4] 350754 0
4-week follow-up (also collected at baseline. )
Secondary outcome [5] 350756 0
action control. 6-item measure based on those used by Sniehotta, Scholz, & Schwarzer, 2005, by referring to the plan to reduce unhealthy snacking. Responses will be recorded on a 7-point Likert scales (strongly disagree to strongly agree).

Sniehotta, F. F., Scholz, U., & Schwarzer, R. (2005). Bridging the intention–behaviour gap: Planning, self-efficacy, and action control in the adoption and maintenance of physical exercise. Psychology & Health, 20(2), 143-160. doi:10.1080/08870440512331317670
Timepoint [5] 350756 0
4 week follow up
Secondary outcome [6] 350757 0
difficulty following the plan. This was a single item measure ("How hard was it to follow your plan?" assessed on an 11-point likert scale (0=very easy, 10=very hard).
Timepoint [6] 350757 0
4-week followup
Secondary outcome [7] 350758 0
Action plan preference. Participants will be reminded of the name and logo used for their plan (literacy-sensitive plan = Smart Snacking 101 (basic plan); standard action plan =Smart Snacking Pro (advance plan)). Participants will then be shown screenshots on an image slider from the plan that they had not used at baseline. Participants are then asked: "If you were given the choice, which would you prefer to use next time?" Participants select ONE of the two action plan interventions. Participants are also able to select "unsure" and then will be provided with alternative descriptions. Participants are then provided with alternative descriptions of the action plans and asked again to select their preference. There is also an option to select 'unsure' again.
Timepoint [7] 350758 0
4 week follow-up
Secondary outcome [8] 350834 0
Perceived extent of healthy snacking in the previous week (2-item, 7-point likert scale)
Timepoint [8] 350834 0
4-week follow-up (also collected at baseline. )

Eligibility
Key inclusion criteria
Participants self-reported either:
1) a diagnosis of type 2 diabetes; or
2) self-reported weight and height that placed them in the overweight/obese BMI category.
Minimum age
18 Years
Maximum age
No limit
Sex
Both males and females
Can healthy volunteers participate?
No
Key exclusion criteria
Inability to speak English.

Study design
Purpose of the study
Treatment
Allocation to intervention
Randomised controlled trial
Procedure for enrolling a subject and allocating the treatment (allocation concealment procedures)
Allocation method was concealed as participants are not aware of which allocation method was used, and is carried out on the Qualtrics (survey host) server.. Allocation to standard or health-literate action plan was not concealed from participant.
Methods used to generate the sequence in which subjects will be randomised (sequence generation)
Randomisation was achieved using Qualtrics ‘Randomizer’ tool. This software randomly allocates the intervention to participants, such that allocations are presented evenly across the sample. Randomisation will take place separately for people with diabetes and people who do not have diabetes, such that roughly half the people in each arm of the allocation method will have diabetes.
Masking / blinding
Blinded (masking used)
Who is / are masked / blinded?


The people assessing the outcomes
The people analysing the results/data
Intervention assignment
Other
Other design features
Assignment to allocation arm is 'parallel' (groups of participants receive different interventions during the same time span of the study.)
Assignment to action plan is also parallel, but is only randomised for participants in the 'randomisation' arm. Participants in the 'screening tool' arm are allocated to an action plan based on their health literacy scores, and participants in the 'free choice' arm are allocated to an action plan based on the action plan they select.
Phase
Not Applicable
Type of endpoint/s
Efficacy
Statistical methods / analysis
Sample size justification
We aim to recruit, using SSI, 2,000 participants with high BMI and/or type 2 diabetes, who will be randomised (stratified by diabetes status) at a ratio of 1:1:1 to each study arm, for a total of 500 per study arm (random, screening, or choice of intervention). With a two sided alpha of 0.05 and power of 80%, a sample of this size will allow us to detect a small main effect of f = 0.08 in a univariate ANOVA comparing the three study arms; this corresponds to a minimum pairwise difference between the two most extreme mean values of approximately 0.18 standard deviations .
Based on our previous studies recruiting through this provider, we anticipate an attrition rate no greater than 15% by one month follow-up. Therefore we estimate a total of 2,352 should be recruited at baseline to ensure sufficient sample size for analysis.
A sample of this size will also ensure that there is at least 80% power for secondary analysis conducted to estimate treatment and preference effects(1, 2) with a treatment effect between the two interventions (literacy-sensitive and standard planning tools) as small as 0.25 standard deviations, and a preference effect, comparing those who received their preferred tool to those who did not, as small as 0.35 standard deviations. This assumes an equal proportion of participants choose the literacy-sensitive and standard planning tools in the choice arm, and equal proportion of participants allocated to each of these interventions in the screening arm.
The sample size estimation assumes an equal proportion of participants choosing the literacy-sensitive version versus the standard in the choice arm and an equal proportion identified as having low health literacy versus non-low health literacy in the screened arm. After enrolment of the first 200 participants these proportions will be calculated; if found to be substantially different from 50% the sample size will be adjusted to ensure adequate power is maintained.
Primary analysis
The primary analysis will use regression (equivalent to ANOVA) to test for a difference in self-reported unhealthy snacking across the three randomised arms (random, the screening and choice) whilst adjusting for any effect of diabetes status. An adjusted model will also be constructed to allow for any baseline imbalances in the potential confounders including age, English as a second language, level of education. The analysis will be repeated on the secondary outcomes of perceived unhealthy snacking in the previous week, snacks consumed the previous day, difficulty using the planning tool, action control, and habit strength with similar adjustment for diabetes status and any baseline imbalances.
Treatment, preference and selection effects will be estimated for the primary outcome(1) The treatment effect compares the efficacy of the health literate tool with the standard tool in the random arm. The preference effect and selection effects are estimated using both the random and choice arms. The preference effect measures the difference in self-reported unhealthy snacking for those who received their preferred treatment compared to those who did not receive their preferred treatment. The selection effect measures the difference between those who would select the literacy-sensitive intervention with those who would select the standard intervention regardless of which intervention they received. These treatment, preference and selection effects will also be estimated for the secondary outcomes described above. An analysis that is analogous to that described above for treatment, preference and selection effects will be conducted on the participants in the random and screening arms.
Analysis of assessment of prior preference:
Participants whose action plan preference was assessed prior to randomisation to an intervention will be compared to those whose preference was not assessed. Multiple linear regression (controlling for health literacy, age, level of education, language spoken at home and baseline snacking) will evaluate the effect of preference assessment on unhealthy snacking behaviour. For participants who provided a preference, an additional multiple linear regression will evaluate the effect of participants receiving their preferred intervention compared to those not receiving their preferred intervention on unhealthy snacking behaviour.

Confirmatory analysis
A confirmatory analysis will replicate the analysis previously reported (3) to examine if the treatment effect is modified by health literacy (as measured by the Newest Vital Sign (NVS), measured at baseline). Multiple linear regression models including an intervention group × health literacy (NVS score) interaction term will be used to predict follow-up snacking scores and perceived difficulty using the plan. Important correlates of health literacy (age, level of education, language spoken at home) (4) and baseline snacking will be controlled for in the model.
Additional analysis
Two researchers will also independently code standard action plans to indicate the extent that participants followed standard action plan instructions and the extent that plans differed from the pre-determined options presented in the literacy-sensitive action plans. Coders will be blind to the health literacy level of participants. Any disagreements will be resolved through discussion. Results from this content analysis will also inform a secondary analysis of the impact of allocation method and action plan on snacking scores.




References

1. Rücker G. A two-stage trial design for testing treatment, self-selection and treatment preference effects. Statistics in Medicine. 1989;8(4):477-85.
2. Turner RM, Walter SD, Macaskill P, McCaffery KJ, Irwig L. Sample Size and Power When Designing a Randomized Trial for the Estimation of Treatment, Selection, and Preference Effects. Medical Decision Making. 2014;34(6):711-9.
3. Ayre J, Bonner C, Cvejic E, McCaffery K, editors. Effect of a health-literate design for an online planning tool on unhealthy snacking behaviour: An experimental study. 15th International Congress of Behvioral Medicine in Santiago; 2018 November 14th-17th; Santiago, Chile.
4. Berkman ND, Sheridan SL, Donahue KE, Halpern DJ, Viera AJ, Crotty K, et al. Health literacy interventions and outcomes: An updated systematic review. Evidence report/technology assessment. 2011(199):1-941.

Recruitment
Recruitment status
Completed
Date of first participant enrolment
Anticipated
15/02/2019
Actual
14/02/2019
Date of last participant enrolment
Anticipated
4/04/2019
Actual
7/06/2019
Date of last data collection
Anticipated
8/05/2019
Actual
18/07/2019
Sample size
Target
2352
Accrual to date
Final
2370
Recruitment in Australia
Recruitment state(s)
ACT,NSW,NT,QLD,SA,TAS,WA,VIC

Funding & Sponsors
Funding source category [1] 300229 0
University
Name [1] 300229 0
The University of Sydney
Country [1] 300229 0
Australia
Funding source category [2] 301141 0
Charities/Societies/Foundations
Name [2] 301141 0
Diabetes Australia
Country [2] 301141 0
Australia
Primary sponsor type
University
Name
The University of Sydney
Address
School of Public Health
Edward Ford Building (A27)
Fisher Road
The University of Sydney, NSW, 2006
Country
Australia
Secondary sponsor category [1] 299640 0
None
Name [1] 299640 0
Address [1] 299640 0
Country [1] 299640 0

Ethics approval
Ethics application status
Approved
Ethics committee name [1] 301052 0
Human Research Ethics Committee at University of Sydney
Ethics committee address [1] 301052 0
Human Ethics Office
Level 3, Administration Building (F23)
University of Sydney
NSW 2006
Ethics committee country [1] 301052 0
Australia
Date submitted for ethics approval [1] 301052 0
17/08/2018
Approval date [1] 301052 0
17/09/2018
Ethics approval number [1] 301052 0
2018/793

Summary
Brief summary
Health literacy interventions typically address health literacy by presenting information in a simpler format. Whilst there is strong evidence that this supports comprehension of health information, the evidence is less clear for self-management behaviours. This is consistent with reviews reporting that key strategies to promote action (e.g. action plans) are often underutilised in health literacy interventions. Existing strategies to promote action are often unsuitable for people with lower health literacy because they are cognitively demanding.

Our recent study addressed this issue by evaluating a ‘health-literate’ action plan. This included simple language and images, and guided users to generate effective plans. After four weeks, participants with lower health literacy who used the health-literate action plan significantly reduced their reported intake of unhealthy snacks compared to those using a ‘standard’ action plan. The reverse was true for participants with higher health literacy, indicating that the health-literate action plan must target the appropriate audience
to be optimally effective.
This tool was effective in a general population for a behaviour that is key to diabetes self-management. The current study will test a similar intervention to improve self-management for people with diabetes and/or overweight/obese BMI. The current study will also explore two practical options for targeting people with low health literacy: a validated screening question for low health literacy; and allowing people
to freely choose between the two action plans. The latter has the added advantage of increasing intervention satisfaction, which is likely to increase engagement.
AIMS AND HYPOTHESES
AIM 1. To replicate the findings of the previous study in a clinical sample of Australians with type 2 diabetes, and identify the most effective action plan formats for this population.
AIM 2. To investigate the optimal method to allocate people to the action plan that is most appropriate for their health literacy level, to inform implementation via two diabetes apps.
AIM 3: to evaluate whether assessment of participant preference for an intervention prior to
random allocation influences the effectiveness of the intervention.
Hypotheses:
1. The health-literate action plan will be more effective at reducing unhealthy snacking for people with lower health literacy, whereas the standard action plan will be more effective for higher health literacy.
2. The intervention will be more effective at reducing unhealthy snacking for people who are allocated an action plan using the health literacy screening tool compared to those who are asked to select their preferred plan. Both of these allocation methods will be more effective than random allocation to an action plan.
3. Assessing preference will negatively impact plan effectiveness, an effect which will be greater for those who are randomised to the plan which is discordant with their preference.
Trial website
Trial related presentations / publications
Public notes

Contacts
Principal investigator
Name 85718 0
Prof Kirsten McCaffery
Address 85718 0
Rm128B, Edward Ford Building (A27),
The University of Sydney, NSW, 2006
Country 85718 0
Australia
Phone 85718 0
+61293517220
Fax 85718 0
Email 85718 0
[email protected]
Contact person for public queries
Name 85719 0
Prof Kirsten McCaffery
Address 85719 0
Rm128B, Edward Ford Building (A27),
The University of Sydney, NSW, 2006
Country 85719 0
Australia
Phone 85719 0
+61293517220
Fax 85719 0
Email 85719 0
[email protected]
Contact person for scientific queries
Name 85720 0
Prof Kirsten McCaffery
Address 85720 0
Rm128B, Edward Ford Building (A27),
The University of Sydney, NSW, 2006
Country 85720 0
Australia
Phone 85720 0
+61293517220
Fax 85720 0
Email 85720 0
[email protected]

Data sharing statement
Will individual participant data (IPD) for this trial be available (including data dictionaries)?
Yes
What data in particular will be shared?
cleaned data as used in the analyses
When will data be available (start and end dates)?
5 after publication of results
Available to whom?
Data will be made available upon request to anyone wishing to access it who provides a methodologically sound proposal to the principal investigator.
Available for what types of analyses?
Replication and meta-analysis.
How or where can data be obtained?
Data will be made available upon direct contact with the principal investigator. Contact details of the principal investigator are [email protected]


What supporting documents are/will be available?

Doc. No.TypeCitationLinkEmailOther DetailsAttachment
4021Study protocolAyre, J., et al. (2019). "Accounting for health literacy and intervention preferences when reducing unhealthy snacking: protocol for an online randomised controlled trial." BMJ Open 9(5): e028544.http://dx.doi.org/10.1136/bmjopen-2018-028544 
4022Statistical analysis planAyre, J., et al. (2019). "Accounting for health literacy and intervention preferences when reducing unhealthy snacking: protocol for an online randomised controlled trial." BMJ Open 9(5): e028544.http://dx.doi.org/10.1136/bmjopen-2018-028544 
4023Ethical approvalAyre, J., et al. (2019). "Accounting for health literacy and intervention preferences when reducing unhealthy snacking: protocol for an online randomised controlled trial." BMJ Open 9(5): e028544.http://dx.doi.org/10.1136/bmjopen-2018-028544 
4024Informed consent formSupplementary file in Ayre, J., et al. (2019). "Accounting for health literacy and intervention preferences when reducing unhealthy snacking: protocol for an online randomised controlled trial." BMJ Open 9(5): e028544.http://dx.doi.org/10.1136/bmjopen-2018-028544 



Results publications and other study-related documents

Documents added manually
No documents have been uploaded by study researchers.

Documents added automatically
SourceTitleYear of PublicationDOI
EmbaseAccounting for health literacy and intervention preferences when reducing unhealthy snacking: Protocol for an online randomised controlled trial.2019https://dx.doi.org/10.1136/bmjopen-2018-028544
N.B. These documents automatically identified may not have been verified by the study sponsor.