Trial Review

Did you know?

The ANZCTR now automatically displays published trial results and simplifies the addition of trial documents such as unpublished protocols and statistical analysis plans.

These enhancements will offer a more comprehensive view of trials, regardless of whether their results are positive, negative, or inconclusive.

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
ACTRN12616000122459
Ethics application status
Approved
Date submitted
25/01/2016
Date registered
3/02/2016
Date last updated
3/02/2016
Type of registration
Prospectively registered

Titles & IDs
Public title
Examining the effects of health-related food taxes and subsidies on health outcomes for New Zealand adults: the Price Experiment and Modelling Study (Price ExaM)
Scientific title
Price ExaM: a Price Experiment and Modelling Study to examine the effects of health-related food taxes and subsidies on health outcomes for New Zealand adults
Secondary ID [1] 288060 0
Nil
Universal Trial Number (UTN)
Trial acronym
Price ExaM
Linked study record

Health condition
Health condition(s) or problem(s) studied:
Obesity (prevention) 296929 0
Non-communicable disease (prevention) 297409 0
Condition category
Condition code
Public Health 297172 297172 0 0
Other public health
Diet and Nutrition 297173 297173 0 0
Other diet and nutrition disorders

Intervention/exposure
Study type
Interventional
Description of intervention(s) / exposure
Price ExaM combines experimental, economic and simulation modelling methods to derive more accurate and precise food price elasticity values (e.g., change in demand in response to change in price). These values are needed to design better policy advise on health-related food taxes and subsidies. The study contains four steps, including: econometrics (step 1); experimental methods (step 2); economic modelling (step 3); and simulation modelling (step 4). This trial registration covers step 2.

The experimental phase is not a traditional RCT testing intervention versus control. Instead, the experimental phase is designed to obtain results that can feed into the subsequent modelling phases.

During the experimental phase, we will expose study participants to randomly changing food prices (different price sets). We will use our Virtual Supermarket and ask 1,000 participants to complete five household food shops over five consecutive weeks. The Virtual Supermarket is a three-dimensional computer simulation of a real New Zealand supermarket. The Virtual Supermarket has a product selection that is representative for an average NZ supermarket and includes 1450 unique food and beverage products. Participants can navigate a shopping trolley in this virtual tool and select products from the shelves with a mouse click. Prices are clearly marked on the shelve.

Participants complete the entire study online, using our Price ExaM website containing all forms (registration, consent, questionnaires) the Virtual Supermarket software and instruction material (including manuals and videos). Following registration and consent, participants are able to download the Virtual Supermarket on their computer. Next, they login with their registration number. At their first login, participants are asked to complete a tutorial where they have to find and purchase six products in the Virtual Supermarket. This tutorial helps participants to become familiar with the software. After completing the tutorial, participants can start their first shop where they will be randomized to a different price set each time they shop (see below).

In the econometric phase (prior to the experimental phase), we design price sets with varying prices for all 1450 products in the Virtual Supermarket. Participants will be exposed to these prices within the Virtual Supermarket. These price sets are designed using econometric methods in a way that allows us to maximise the precision of price elasticities across individuals and within individuals estimated from the subsequent steps in the study. We develop different price sets for each participant and for each virtual shop, meaning that we will have 5,100 price sets in total (1000 participants * 5 shops + 100 back-up price sets). The price sets will be divided into five broad categories in line with five policy options (sugar tax, saturated fat tax, salt tax, sugar-sweetened beverage (SSB) tax and fruit and vegetable subsidy). Within each price set, the price of all products will vary (baseline variation), and we will add the taxes/subsidies on top of these price sets leading to larger price variations for particular products depending on the particular policy option (for example, we will have more variations in SSB prices in SSB tax scenario). 15% of the price sets will not receive a tax / subsidy (e.g., will only have baseline variation). Using this approach, Price ExaM will provide precise price elasticity data for any food group which can be used to test any food tax, but will maximise precision about PE values for the five most likely policy options.

Participants will be randomly assigned to one of the 5,100 price sets each time they shop (e.g., participants can receive any of the 5,100 price sets). The price sets are allocated from our price set database using a random generator, with the only restriction being that a price set cannot be used more than once. Participants do not necessarily receive price sets within the same policy option for each shop and randomization is not stratified (i.e., it is completely random).

The Virtual Supermarket collects all purchased products and their prices. Participants’ virtual purchase responses will then feed into the next economic modelling phase. Participants are informed that this study aims to measure shopping behaviour. Participants are also informed that prices in the Virtual Supermarket vary and they are able to see the prices when they shop. However participants are not informed: (i) how the prices vary; (ii) how these price changes relate to fiscal policy options; (iii) and how the study relates to health outcomes.

Participants can complete the study using any computer or laptop. Each shop takes around 30 minutes to complete and there is not time limit to complete the shop.

Intervention code [1] 293597 0
Behaviour
Comparator / control treatment
This is not a traditional randomized controlled trial comparing the effects of an intervention among participants in control and intervention groups. Every participant gets a different price set each time they shop. Within each price set, the prices of all 1450 products in the Virtual Supermarket varies (baseline variation), with extra variation for the price sets aligning with the five policy options. 15% Of the price sets will have no allocated tax or subsidy and won't have this extra variation (but will still have the baseline variation). These price sets can be used as control in the experimental analysis. However, not every participant will necessarily receive a control price set as part of their five shops.



Control group
Active

Outcomes
Primary outcome [1] 297026 0
The main outcome of the Virtual Supermarket experiment is the multiple observations of price configurations and purchased amounts which can then be used to calculate three main outcomes:

Difference in food purchases between broad food pricing policy options (no tax/subsidy; sugar tax, saturated fat tax, salt tax, sugar sweetened beverage tax and fruit and vegetable subsidy). Note that this does not mean that we are constrained to these policy options only. Rather, they are the most likely policy options we will assess and they are the ones for which we need to maximise precision about price elasticities. Differences in food purchasing will be quantified as:
a. Mean quantity (g) (adjusted for household size) of key nutrients, including saturated fat, total sugar, sodium and energy of the total shopping basket.
b. Nutrient profiling score of the total shopping basket.
c. Mean quantity (g/ml) (adjusted for household size) of food groups most impacted by the five fiscal food policies (e.g., soft drinks, snacks, fruit and vegetables, meat, etc)




Timepoint [1] 297026 0
Data from all five weekly shops will feed into the models.
Primary outcome [2] 297196 0
Food price elasticities calculated using the Virtual Supermarket output and traditional and/or Bayesian economic modelling methods.
Timepoint [2] 297196 0
Data from all five weekly shops will feed into the models.
Primary outcome [3] 297197 0
Health gains/losses for each tax/subsidy policy calculated through the BODE3 DIET model, by two methods:
a. Directly inputting the average amount of each food product purchased in the VS for each policy option
b. Using the price elasticities.
Timepoint [3] 297197 0
Data from all five weekly shops will feed into the models.
Secondary outcome [1] 319878 0
We aim to derive specific price elasticity estimates for Maori and non-Maori using the Virtual Supermarket output and traditional and/or Bayesian economic modelling methods.

Timepoint [1] 319878 0
Data from all five weekly shops will feed into the models.
Secondary outcome [2] 320182 0
We aim to derive specific price elasticity estimates for men and women using the Virtual Supermarket output and traditional and/or Bayesian economic modelling methods
Timepoint [2] 320182 0
Data from all five weekly shops will feed into the models.
Secondary outcome [3] 320313 0
We aim to derive specific price elasticity estimates for different age groups using the Virtual Supermarket output and traditional and/or Bayesian economic modelling methods
Timepoint [3] 320313 0
Data from all five weekly shops will feed into the models.
Secondary outcome [4] 320314 0
Price sensitivity (e.g., how sensitive are people to pricing) measured by a selection of questions from a validated questionnaire (Lichtenstein et al.., 1993; "PRICE PERCEPTION CONSTRUCT SCALE ITEMS"). We aim to include this measure as a confounder / effect modifier
Timepoint [4] 320314 0
After the first virtual shop and after the last virtual shop
Secondary outcome [5] 320402 0
This is a primary outcome:
Net health system costs for each tax/subsidy policy calculated through the BODE3 DIET model, by two methods:
a. Directly inputting the average amount of each food product purchased in the VS for each policy option
b. Using the price elasticities.
Timepoint [5] 320402 0
Data from all five weekly shops will feed into the models.

Eligibility
Key inclusion criteria
-Adults (18 years or over)
-Contributes to household grocery shopping
-Speaks and reads English
-Has access to a computer and internet and able to download the software onto that computer (tablets and smartphones are not compatible with the Virtual Supermarket software).
-Has an email address (provided at screening stage)
-Able to use basic functions of a computer such as opening the internet browser, using email and saving files.
-Available to conduct five consecutive weekly shops in the Virtual Supermarket during the required study period.
Minimum age
18 Years
Maximum age
No limit
Sex
Both males and females
Can healthy volunteers participate?
Yes
Key exclusion criteria
Another person in the same household already participating in the study.

Study design
Purpose of the study
Prevention
Allocation to intervention
Randomised controlled trial
Procedure for enrolling a subject and allocating the treatment (allocation concealment procedures)
Allocation will be concealed. Participants complete the entire study online. Following registration and consent, eligible participants can download the Virtual Supermarket software on their computer. Each time a participant logs in, a computerized sequence allocation will select one of the 5,100 price sets and employ this into the Virtual Supermarket (so the participants will see the prices of that price set in the Virtual Supermarket).
Methods used to generate the sequence in which subjects will be randomised (sequence generation)
Computerized sequence allocation
Masking / blinding
Blinded (masking used)
Who is / are masked / blinded?
The people receiving the treatment/s
The people administering the treatment/s
The people assessing the outcomes
The people analysing the results/data
Intervention assignment
Other
Other design features
Randomization:
For each of the five shops, participants are exposed to a different price set (1 of 5,100) and the price sets will differ for all participants (e.g. a different price set for each participant AND for each shop). The price sets are allocated from our price set database using a random generator, with the only restriction being that a price set cannot be used more than once. Participants don't receive the price sets in any particular order and they can receive price sets from any of the policy options. Randomization is not stratified (i.e., it is completely random).
Phase
Not Applicable
Type of endpoint/s
Efficacy
Statistical methods / analysis
For the experimental phase, statistical analyses are performed by the study biostatistician, who is blinded to the price sets participants are allocated to. All randomised participants are included in the analysis (intention to treat), using their shopping data collected during each of the five virtual shops.

For each shop, the policy option applied to the price set is flagged using a binary indicator (yes/no). Random effect mixed models are used to test the effect of individual policy options on the purchased products and nutrient outcomes, adjusting for baseline demographics and shopping basket number. Correlation between repeated shopping data collected from the same participant will be taken into account using a random subject effect. Potential interaction effects between the five policy options will be explored in the model. The consistency of effects across important subgroups will be tested. Where sub-group numbers are sufficient, pre-specified subgroup analyses may be undertaken when a significant overall effect is observed with a policy option.

For the subsequent modelling phases, food price elasticities will be calculated using the Virtual Supermarket output and traditional and/or Bayesian economic modelling methods. Health gains/losses and net health system costs for each tax/subsidy policy calculated through the BODE3 DIET model, by two methods
a. Directly inputting the average amount of each food product purchased in the VS for each policy option
b. Using the price elasticities.

The sample size for this study was determined by combining previous price elasticity estimations and previous Virtual Supermarket data. Using the soft drink tax as an example, a sample size of 188 will provide 85% power at 5% overall level of significance, to detect a group difference of 0.8 litres in soft drink purchase per household/week between two treatment strategies (based on our previous trial, Waterlander et al., 2014). This sample size assumes a standard deviation of 2.5 litres. N=1000 participants will therefore be sufficient to measure the impact of all five fiscal food policies.

Recruitment
Recruitment status
Not yet recruiting
Date of first participant enrolment
Anticipated
10/02/2016
Actual
Date of last participant enrolment
Anticipated
30/12/2016
Actual
Date of last data collection
Anticipated
Actual
Sample size
Target
1000
Accrual to date
Final
Recruitment outside Australia
Country [1] 7512 0
New Zealand
State/province [1] 7512 0

Funding & Sponsors
Funding source category [1] 292676 0
Government body
Name [1] 292676 0
Health Research Council of New Zealand
Country [1] 292676 0
New Zealand
Primary sponsor type
Individual
Name
Dr Wilma Waterlander
Address
National Institute for Health Innovation
School of Population Health
The University of Auckland – Tamaki Campus
Private Bag 92019, Auckland Mail Centre
Auckland 1142 – New Zealand
Country
New Zealand
Secondary sponsor category [1] 291396 0
Individual
Name [1] 291396 0
Jo Michie
Address [1] 291396 0
National Institute for Health Innovation
School of Population Health
The University of Auckland – Tamaki Campus
Private Bag 92019, Auckland Mail Centre
Auckland 1142 – New Zealand
Country [1] 291396 0
New Zealand

Ethics approval
Ethics application status
Approved
Ethics committee name [1] 294145 0
University of Auckland Human Participants Ethics Committee (UAHPEC)
Ethics committee address [1] 294145 0
The University of Auckland
Level 10, 49 Symonds Street
Grafton
Auckland 1010
Ethics committee country [1] 294145 0
New Zealand
Date submitted for ethics approval [1] 294145 0
28/09/2015
Approval date [1] 294145 0
10/11/2015
Ethics approval number [1] 294145 0
016151

Summary
Brief summary
The aim of this research is: 1) to derive accurate and more precise food price elasticity (PE) values; 2) to quantify the impact of price changes on quantity and quality of discrete food groups and; 3) to model the potential health and disease impacts of these food pricing policy options.

This study combines experimental and simulation modelling methods using four steps, including: econometrics (step 1); experimental methods (step 2); economic modelling (step 3); and simulation modelling (step 4).

This trial registry focuses on the experimental phase. This experiment is not a traditional RCT testing intervention control, but is designed to measure multiple responses to changing food prices that can feed into simulation models to estimate more precise price elasticity values.

For the experimental phase, we aim to randomize N=1,000 adults from the general NZ population and ask them to complete five weekly shops for their household in a New Zealand Virtual Supermarket for five consecutive weeks. The Virtual Supermarket includes around 1400 unique food and beverage items. Participants can complete this study from home. This study is experimental and people will pay with virtual money. Real groceries will not be provided.

During the virtual shopping tasks, participants will be exposed to random varying sets of food prices in the Virtual Supermarket. The price sets will be constructed in the simulation modelling phase that occurs prior to this experimental phase. In total, we will construct 5,100 different price sets (1,000 participants * 5 shops + 100 back up) and each study participant will be exposed to a different price set during each of their five virtual shops. The price sets will be divided into five categories in line with five pre-determined fiscal policy options (sugar tax, saturated fat tax, salt tax, sugar sweetened beverage tax and fruit and vegetable subsidy). This means that we will have around 1,000 different price sets per policy option (e.g., 1,000 price sets for a soft drink tax, 1,000 price sets for a fruit and vegetable subsidy, etc.). Within each price set, the price of all products will vary, but we will simulate larger price variations for particular products depending on the policy option of interest (for example, we will have more variation in soft drink prices for the soft drink tax scenario).

Data from this experiment will then be put into econometric models to derive more accurate food price elasticity values which will subsequently feed into macro-simulation models to quantify the impact of price changes on quantity and quality of food choices and to model the potential health and disease impacts of these policy options.
Trial website
https://diet.auckland.ac.nz/content/price-exam
Trial related presentations / publications
- Blakely T, Waterlander W, Nghiem N, Genc M, Eyles H, Cobiac L, Ni Mhurchu C. A novel approach to improving price elasticity estimation: an experimental study using a virtual supermarket and Bayesian methods. 2015 International Health Economics Association Congress, Milan, Italy, 12 Jul 2015 - 15 Jul 2015. 12 Jul 2015
Public notes
n/a

Contacts
Principal investigator
Name 61634 0
Dr Wilma Waterlander
Address 61634 0
National Institute for Health Innovation
School of Population Health
The University of Auckland – Tamaki Campus
Private Bag 92019, Auckland Mail Centre
Auckland 1142 – New Zealand
Country 61634 0
New Zealand
Phone 61634 0
+64 (09) 9234612
Fax 61634 0
Email 61634 0
[email protected]
Contact person for public queries
Name 61635 0
Jo Michie
Address 61635 0
National Institute for Health Innovation
School of Population Health
The University of Auckland – Tamaki Campus
Private Bag 92019, Auckland Mail Centre
Auckland 1142 – New Zealand
Country 61635 0
New Zealand
Phone 61635 0
+64 (0) 9234721
Fax 61635 0
Email 61635 0
[email protected]
Contact person for scientific queries
Name 61636 0
Cliona Ni Mhurchu
Address 61636 0
National Institute for Health Innovation
School of Population Health
The University of Auckland – Tamaki Campus
Private Bag 92019, Auckland Mail Centre
Auckland 1142 – New Zealand
Country 61636 0
New Zealand
Phone 61636 0
+64 (09) 9234612
Fax 61636 0
Email 61636 0
[email protected]

No information has been provided regarding IPD availability


What supporting documents are/will be available?

No Supporting Document Provided



Results publications and other study-related documents

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

Documents added automatically
SourceTitleYear of PublicationDOI
EmbaseImpact of taxes on purchases of close substitute foods: analysis of cross-price elasticities using data from a randomized experiment.2021https://dx.doi.org/10.1186/s12937-021-00736-y
N.B. These documents automatically identified may not have been verified by the study sponsor.