Healthy food prescription incentive programme for adults with type 2 diabetes who are experiencing food insecurity: protocol for a randomised ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Open access Protocol
BMJ Open: first published as 10.1136/bmjopen-2021-050006 on 15 February 2022. Downloaded from http://bmjopen.bmj.com/ on February 25, 2022 by guest. Protected by copyright.
Healthy food prescription incentive
programme for adults with type 2
diabetes who are experiencing food
insecurity: protocol for a randomised
controlled trial, modelling and
implementation studies
Dana Lee Olstad ,1 Reed Beall,1 Eldon Spackman,1 Sharlette Dunn,1
Lorraine L Lipscombe,2 Kienan Williams,3 Richard Oster,4 Sara Scott,1
Gabrielle L Zimmermann,1,5 Kerry A McBrien,1,6 Kieran J D Steer,1
Catherine B Chan,4,7,8 Sheila Tyminski,9 Seth Berkowitz ,10 Alun L Edwards,11
Terry Saunders-Smith,1 Saania Tariq,1 Naomi Popeski,8 Laura White,12
Tyler Williamson,1 Mary L'Abbé,13 Kim D Raine,14 Sara Nejatinamini,1 Aruba Naser,1
Carlota Basualdo-Hammond,9 Colleen Norris,15,16 Petra O’Connell,8 Judy Seidel,1,17
Richard Lewanczuk,18 Jason Cabaj,1 David J T Campbell 1,11,19
To cite: Olstad DL, Beall R, ABSTRACT
Spackman E, et al. Healthy Strengths and limitations of this study
Introduction The high cost of many healthy foods poses
food prescription incentive a challenge to maintaining optimal blood glucose levels for
programme for adults with type ► We will investigate the reach, effectiveness, adop-
adults with type 2 diabetes mellitus who are experiencing food
2 diabetes who are experiencing tion, implementation and maintenance of a healthy
insecurity, leading to diabetes complications and excess acute
food insecurity: protocol for a food prescription incentive programme for adults
randomised controlled trial, care usage and costs. Healthy food prescription programmes who are experiencing food insecurity and persistent
modelling and implementation may reduce food insecurity and support patients to improve hyperglycaemia.
studies. BMJ Open their diet quality, prevent diabetes complications and avoid acute ► A randomised controlled trial and a modelling study
2022;12:e050006. doi:10.1136/ care use. We will use a type 2 hybrid-effectiveness design to will demonstrate the short- and longer-term impacts
bmjopen-2021-050006 examine the reach, effectiveness, adoption, implementation and of the programme on glycosylated haemoglobin,
► Prepublication history and maintenance (RE-AIM) of a healthy food prescription incentive other health-related outcomes, resource use and
additional supplemental material programme for adults experiencing food insecurity and persistent costs.
for this paper are available hyperglycaemia. A randomised controlled trial (RCT) will ► An implementation study will support translation of
online. To view these files, investigate programme effectiveness via impact on glycosylated findings into practice by examining determinants of
please visit the journal online haemoglobin (primary outcome), food insecurity, diet quality effective implementation and reasons behind pro-
(http://dx.doi.org/10.1136/ and other clinical and patient-reported outcomes. A modelling
bmjopen-2021-050006).
gramme successes and failures.
study will estimate longer-term programme effectiveness in ► Patients’ medication/insulin regimes may be inten-
Received 08 February 2021
reducing diabetes-related complications, resource use and sified/de-intensified during the study and thus sen-
Accepted 19 October 2021 costs. An implementation study will examine all RE-AIM domains sitivity analyses will be conducted to examine the
to understand determinants of effective implementation and potential impact of such changes on study findings.
reasons behind programme successes and failures.
Methods and analysis 594 adults who are experiencing food
insecurity and persistent hyperglycaemia will be randomised Ethics and dissemination Ethical approval was obtained
© Author(s) (or their to a healthy food prescription incentive (n=297) or a healthy from the University of Calgary and the University of
employer(s)) 2022. Re-use food prescription comparison group (n=297). Both groups will Alberta. Findings will be disseminated through reports,
permitted under CC BY-NC. No receive a healthy food prescription. The incentive group will lay summaries, policy briefs, academic publications and
commercial re-use. See rights additionally receive a weekly incentive (CDN$10.50/household
and permissions. Published by conference presentations.
member) to purchase healthy foods in supermarkets for 6 Trial registration number NCT04725630.
BMJ.
months. Outcomes will be assessed at baseline and follow- Protocol version Version 1.1; February 2022
For numbered affiliations see
up (6 months) in the RCT and analysed using mixed-effects
end of article.
regression. Longer-term outcomes will be modelled using the
Correspondence to UK Prospective Diabetes Study outcomes simulation model-2. BACKGROUND
Dr Dana Lee Olstad; Implementation processes and outcomes will be continuously Type 2 diabetes mellitus (T2DM) imposes a
dana.olstad@ucalgary.ca measured via quantitative and qualitative data. tremendous burden on healthcare systems
Olstad DL, et al. BMJ Open 2022;12:e050006. doi:10.1136/bmjopen-2021-050006 1Open access
BMJ Open: first published as 10.1136/bmjopen-2021-050006 on 15 February 2022. Downloaded from http://bmjopen.bmj.com/ on February 25, 2022 by guest. Protected by copyright.
worldwide, as individuals with T2DM incur twice the programmes also appear to be cost-effective, with one
healthcare costs as their age-matched and sex-matched recent modelling study indicating that a national healthy
counterparts.1 2 The total economic costs of diabetes were food prescription incentive programme in the USA could
US$327 billion in 2017 in the USA, and CDN$30 billion eliminate US$100.2 billion in healthcare costs if imple-
in Canada in 2019, making it among the most expensive mented over the lifetime of beneficiaries.64
chronic conditions in both nations.1 3 4 The human toll Despite some promising initial findings, major
on individuals and their families is also substantial in knowledge gaps remain pertaining to the impact and
terms of reduced quality of life associated with managing optimal implementation of healthy food prescription
the disease.5 6 Many of these human and economic costs programmes.60 Most prior studies have been small and
are avoidable, as adherence to a healthy diet within an uncontrolled, and have examined a small number of self-
overall diabetes management plan can yield clinically reported outcomes using brief dietary44 45 48 63 65 and/or
meaningful improvements in blood glucose levels, which food insecurity screeners,47 50–53 and/or short descriptive
can reduce diabetes complications over time.7–15 Average surveys,44 61 63 rather than objective clinical outcomes. The
blood glucose levels are most often quantified using the majority of prior programmes have also subsidised the
glycosylated haemoglobin level (A1C), which represents purchase of fruits and vegetables alone, without consid-
the average blood sugar level over the previous 3 ering the relevance of entire dietary patterns to blood
months.16 An absolute reduction of 0.5% in A1C is achiev- glucose levels and health outcomes.7 8 60 Moreover, there
able through improving diet quality7 8 and is considered a is virtually no understanding of the effectiveness and cost-
clinically meaningful difference.17 effectiveness of these programmes over the longer-term,
The high and continually escalating costs of many nor of optimal implementation strategies.66
healthy foods18 19 represents a formidable barrier to We will build on these initial findings through
adhering to a healthy dietary pattern for individuals three concurrent studies, including a randomised
with T2DM, particularly for those who are experiencing controlled trial (RCT), a modelling study and an
food insecurity.20–24 Food insecurity refers to inadequate implementation study. We will use a type 2 hybrid
or insecure access to food due to financial constraints25 effectiveness- i mplementation design, which entails
and is a strong predictor of high-cost healthcare use.26 dual testing of the effectiveness and implementation
Evidence indicates that individuals with T2DM who are of an intervention.67 Collective findings will be inte-
experiencing food insecurity have lower diet quality than grated to provide a comprehensive perspective of the
their food secure counterparts, leading to elevations reach, effectiveness, adoption, implementation and
in blood glucose levels,27–35 and high rates of diabetes maintenance (RE-AIM)68 69 of a healthy food prescrip-
complications and acute care use.29 32 36–39 Indigenous tion incentive programme among adults who are
groups (constitutionally recognised as First Nations, experiencing food insecurity and persistent hypergly-
Inuit and Métis) are a population of particular concern, caemia. First, the RCT will provide a basis for causal
given their disproportionately high rates of both T2DM inference pertaining to programme effectiveness. It
and food insecurity.40–42 The coexistence of food insecu- will entail an incentive to purchase a variety of healthy
rity and T2DM, therefore, has major implications for the foods from all food groups, and will be powered to
sustainability of healthcare systems. detect clinically meaningful changes in A1C, along
Although it is well known that food insecurity is a with a comprehensive range of objective and self-
primary driver of acute care usage and costs, health- reported health- r elated outcomes. A linked model-
care providers often lack effective strategies to address ling study will provide a longer-term perspective of
it. One approach to better address this problem is to programme effectiveness in reducing diabetes-related
assist patients who are experiencing food insecurity to complications, along with healthcare use and costs.
purchase diabetes- appropriate foods through healthy Finally, a complementary implementation study will
food prescription programmes, which provide subsidies encompass quantitative and qualitative measures of
or incentives to improve access to healthy foods. Prelim- all RE-AIM domains to support translation of research
inary evidence from several studies suggests that these findings into practice and policy by helping to under-
programmes may improve diet quality and self-reported stand determinants of effective implementation and
health, while reducing food insecurity, A1C, hyper- reasons behind programme successes and failures.
tension and body mass index (BMI), including within
Indigenous communities.43–59 Moreover, a recent meta-
analysis of 13 studies found that healthy food prescription METHODS
programmes may increase fruit and vegetable intake by Overview
0.8 servings/day, reduce BMI by 0.6 kg/m2 and reduce Ethics, privacy and confidentiality
A1C by 0.8%, although the certainty of the evidence was This research has been approved by the Univer-
rated as very low to moderate.60 Qualitative data similarly sity of Calgary Conjoint Health Research Ethics
suggest patients and care providers perceive financial, Board (REB20-0 543) and the University of Alberta
dietary and health benefits from these programmes, and Health Research Ethics Board Biomedical Panel
support their ongoing delivery.43 50 61–63 Food prescription (Pro00107116). Any protocol deviations will be
2 Olstad DL, et al. BMJ Open 2022;12:e050006. doi:10.1136/bmjopen-2021-050006Open access
BMJ Open: first published as 10.1136/bmjopen-2021-050006 on 15 February 2022. Downloaded from http://bmjopen.bmj.com/ on February 25, 2022 by guest. Protected by copyright.
approved in advance by the board and updated Implementation support team
in the clinical trials registry. All participants will The implementation support team will consist of research
provide informed consent prior to data collection coordinators and assistants who will execute the daily tasks
(online supplemental additional file 1). Participant required to administer, plan, support, monitor and eval-
data will be anonymised and stored on a password- uate the healthy food prescription incentive programme.
protected University server. Only the principal inves-
tigators and research coordinators will have access Patient and public involvement
to identifiable participant information and the final This research has been informed by substantial prior73–77
trial datasets. and ongoing engagement with patients experiencing
financial barriers to chronic disease care. Patient part-
Setting ners (who are not study participants) will help to pilot
This research will take place in Alberta, Canada between test infrastructural supports (eg, healthy food prescrip-
May 2021 and December 2023. Participants will primarily tion pamphlet, usability of the list of eligible foods), care
be recruited through primary care and diabetes specialty pathways and implementation processes, and will be
clinics located in urban and rural communities, including members of the advisory boards and PCC subcommittee.
clinics with an explicit focus on serving people who iden- Patients who are study participants will provide contin-
tify as Indigenous. uous programme feedback via a dedicated study help-
line/email, and by completing implementation fidelity
checklists. At the conclusion of the study, participants will
Study oversight
be invited to describe their programme experiences via
Scientific steering committee
a postintervention questionnaire and during in-depth,
A scientific steering committee will oversee all aspects of
semistructured interviews.
the research, receive and review reports from the study’s
Lands and food hold deep cultural, symbolic and spir-
advisory boards and subcommittees, and will have final
itual significance for Indigenous peoples.78 Staff and
decision-making authority. It will be comprised of the
patients from Indigenous PCCs will codesign clinical care
study’s five co-principal investigators (DO, ES, RB, LLL
pathways and other procedures that are context-specific
and DJTC).
and culturally appropriate for Indigenous patients, and
Advisory board that respect and promote Indigenous worldviews, particu-
A multistakeholder advisory board will provide high- larly those surrounding food procurement and consump-
level oversight for the research and will advise the scien- tion. We will ensure that representatives from Indigenous
tific steering committee on study conduct. Members will clinics are involved at all stages of the research, including
include policy- makers, academic experts, representa- study design, pilot testing infrastructural supports, inter-
tives from Alberta Health Services (the provincial health preting results and formulating conclusions, and that
authority), an Indigenous public health expert and a their agreement is obtained prior to communicating
patient. any research findings that pertain to them. As previously
described, the Indigenous advisory board will also oversee
Indigenous advisory board all aspects of the research.
An Indigenous advisory board will ensure that research
activities within Indigenous clinics proceed in a cultur- Evaluation framework and theory of change
ally sensitive, relevant, responsive, equitable and recip- RE-AIM68 69 will provide a structured means of integrating
rocal manner that is guided by Indigenous Ownership, data from the RCT, modelling and implementation
Control, Access and Possession of data principles studies to understand the RE-AIM of the healthy food
(OCAP)70 and complies with Government of Canada prescription incentive programme.
guidance for Indigenous Research.71 72 The board will Our theory of change (figure 1) draws on Barnard et
include Indigenous elders and patients, along with al’s79 conceptual model linking material needs insecuri-
academic experts, policy- makers, managers and front- ties with diabetes outcomes, and posits that reduced food
line practitioners from the public health and healthcare insecurity and improved diet quality will be key mediators
sectors who are themselves Indigenous, or who work of improved blood glucose levels (quantified via A1C),
closely with Indigenous peoples. which will help to reduce diabetes complications, and
healthcare resource use and costs. Each construct will be
Primary care clinic subcommittee examined to affirm or disprove the proposed pathway.
The primary care clinic (PCC) subcommittee will include
PCC managers, staff and patients. As participant recruit- Randomised controlled trial
ment and implementation of the intervention unfolds, The RCT protocol adheres to the Standard Protocol
PCC managers and staff will collect feedback from their Items: Recommendations for Interventional Trials
respective clinics and will share it with the larger group as (SPIRIT) and Template for Intervention Description
a learning tool and to inform ongoing adaptations. and Replication (TIDieR) reporting guidelines (table 1;
online supplemental additional file 2–4).
Olstad DL, et al. BMJ Open 2022;12:e050006. doi:10.1136/bmjopen-2021-050006 3Open access
BMJ Open: first published as 10.1136/bmjopen-2021-050006 on 15 February 2022. Downloaded from http://bmjopen.bmj.com/ on February 25, 2022 by guest. Protected by copyright.
Figure 1 Healthy food prescription programme logic model. A1C an indicator of average blood glucose levels over the
previous 3 months. A1C, glycosylated haemoglobin.
Study design and objectives persistent hyperglycaemia (ie, A1C 7%–12%), including
A 6-month, parallel-group RCT will examine the effective- patients living in rural and urban areas, and those who
ness of a healthy food prescription incentive programme, identify as Indigenous. Potentially eligible patients will
compared with a healthy food prescription alone, in be invited to complete a brief screening questionnaire
improving the following outcomes among 594 adults who to identify risk of food insecurity based on the Hunger
are experiencing food insecurity and persistent hypergly- Vital Sign,80–82 and perceived income adequacy.83–85
caemia (ie, A1C 7%–12%): Eligible patients will be adults (18–85 years) with T2DM
1. Primary outcome: Average blood glucose levels mea- (or diabetes of unknown aetiology) and persistent hyper-
sured by A1C. glycaemia (ie, A1C 7%–12%) who are experiencing food
2. Secondary outcomes: insecurity and/or perceive that it is difficult/very diffi-
a. Blood glucose levels: Proportion of patients with el- cult to make ends meet, do not reside in a facility that
evated A1C (ie, ≥8.5%); blood glucose measured by provides meals (eg, shelter, long-term care, prison), and
fructosamine. can communicate in English or have someone to trans-
b. Dietary intake: Diet quality; skin carotenoids. late. Patients will be excluded if they have an A1C12% (given the recommendation for antihypergly-
pressure; BMI; waist circumference; need for antihy- caemic treatment escalation for those with A1C>12%),
perglycaemic medication/insulin. have signs/symptoms of metabolic decompensation, have
d. Patient-reported outcomes: Psychosocial well-being; an eating disorder, have experienced diabetic ketoacidosis
self-rated health; diabetes self-efficacy; diabetes self- or a hyperglycaemic hyperosmolar emergency in the past
management; diabetes distress; diabetes competing year, or if they experienced a severe hypoglycaemic event
demands; perceived financial barriers to chronic in the past 3 months. Patients will also be excluded if they
disease care; hypoglycaemic episodes; household are pregnant or trying to conceive, breast feeding, partic-
food insecurity. ipating in other clinical trials, if someone in their house-
3. Exploratory outcomes: Subjective social status; per- hold is currently or has previously participated, if they are
ceived income adequacy; work productivity and ac- unwilling/unable to shop in study-affiliated supermarkets
tivity impairment; medication and physical activity for the next 6 months, if they plan to leave Canada for
adherence. more than 2 weeks in the next 6 months, or if they will not
be able to complete data collection at 6 months.
Primary care clinics
Eligible patients will be asked to provide consent
PCCs will be recruited, including urban, rural and
Indigenous clinics. To be eligible, clinics must serve to their healthcare provider to be contacted by the
lower-income patients, agree to allow their physicians, research team. A research assistant will contact patients
registered dietitians and/or nurses to dispense healthy to confirm all eligibility criteria have been met, obtain
food prescriptions, appoint a staff member to liaise with informed consent and provide instructions for collection
the implementation support team, and be willing to of baseline data. Participants may elect to report baseline
receive training. The final list of study sites, currently patient-reported data immediately over the telephone, or
projected at 30 clinics, will be available in the clinical- independently via the study’s online data collection plat-
trials.gov registry. forms. Any patients identified as at risk of food insecu-
rity at screening, but who do not respond affirmatively
Participants to any of the items on the full 18-item Household Food
Healthcare providers will use information from electronic Security Survey Module or who do not indicate that it is
medical records to identify patients with T2DM and difficult/very difficult to make ends meet during baseline
4 Olstad DL, et al. BMJ Open 2022;12:e050006. doi:10.1136/bmjopen-2021-050006Table 1 SPIRIT flow diagram
Study period
Enrolment Allocation Postallocation Close-out Follow-up
T1 T2 T3 T4
Time point −T1 0 Baseline +6 mos +12 mos +10 years
Enrolment
Eligibility screen X
Informed consent X
Randomisation X
Allocation X
Interventions
Healthy food prescription incentive group
Healthy food prescription comparison group
Assessments
Sociodemographic and health-related characteristics X X
Primary outcome: haemoglobin A1C X X
Olstad DL, et al. BMJ Open 2022;12:e050006. doi:10.1136/bmjopen-2021-050006
Patient-reported outcomes: psychosocial well-being, X X
self-rated health, diabetes self-efficacy, diabetes self-
management, diabetes distress, diabetes competing
demands, perceived financial barriers to chronic disease
care, hypoglycaemic episodes, work productivity and
activity impairment, household food insecurity, subjective
social status, perceived income adequacy, medication and
physical activity adherence
Biochemical measures: fructosamine, skin carotenoids, X X
blood lipids, serum creatinine, albumin-to-creatinine ratio,
haemoglobin
Physical measures: X X
weight, height, waist circumference, blood pressure, heart
rate
Urinalysis: X X
albuminuria
Diet quality Twice Twice
Administrative health data: Medication/insulin type and X X
dose, comorbidities, diabetes complications, haemoglobin
A1C
Patient, care provider and clinic characteristics, reasons
for non-participation and drop-out
Open access
Continued
5
BMJ Open: first published as 10.1136/bmjopen-2021-050006 on 15 February 2022. Downloaded from http://bmjopen.bmj.com/ on February 25, 2022 by guest. Protected by copyright.Open access
BMJ Open: first published as 10.1136/bmjopen-2021-050006 on 15 February 2022. Downloaded from http://bmjopen.bmj.com/ on February 25, 2022 by guest. Protected by copyright.
data collection will be excluded. Participants will have
biochemical and physical measurements performed at a
community laboratory or at their PCC. Participants with
an A1C outside the 7%–12% range will be excluded at
+10 years
that point. All participants will subsequently receive a
healthy food prescription pamphlet from a healthcare
T4
provider (ie, physician, nurse, registered dietitian) and
a brief, high-
level overview of its contents using stan-
dardised teaching guidelines, either virtually or in-person
during a clinic visit.
Follow-up
Sample size calculation
+12 mos
Based on local administrative data, we expect a mean
baseline A1C of 8.5% (SD=1.4%) in our population.86
T3
Assuming 5% type I error, 30% attrition87 and potential
design effects based on sampling in different clinics (25%
inflation), 594 participants are required for 90% power to
detect a difference in A1C of 0.5%, which is often consid-
ered a minimally important clinical difference.17
Close-out
+6 mos
Randomisation and blinding
Following baseline data collection and delivery of the
A1C, glycosylated haemoglobin; SPIRIT, Standard Protocol Items: Recommendations for Interventional Trials.
T2
X
healthy food prescription pamphlet, participants will
be randomised to a healthy food prescription incentive
Enrolment Allocation Postallocation
(n=297) or a healthy food prescription comparison group
(n=297) with a 1:1 allocation ratio using a computer-
Baseline
generated, concealed, blocked randomisation sequence
created by an independent statistician. Blocking vari-
T1
X
X
ables will include gender, clinic type/location (urban,
rural, Indigenous) and baseline A1C (7%–8.5%, 8.6%–
12%). Allocation concealment will be ensured via secure
storage of the randomisation sequence separately from
the participant database, which will only be accessible by
0
Study period
the statistician. To ensure researcher blinding, allocation
assignment will be operationalised via REDCap (Research
Electronic Data Capture) following completion of base-
Qualitative interviews, observations, meeting notes, notes
−T1
line data collection. Intervention assignment will be
communicated by research assistants via a telephone call.
Participants cannot be blinded to treatment allocation,
Healthy food incentives offered, earned and redeemed
however details of the healthy food incentive, including
its monetary value and the types of foods that are eligible,
will not be divulged to participants in the comparison
group. Care providers, individuals who collect biochem-
ical and physical measurements and data analysts will be
blinded to group allocation.
Healthy food prescriptions prescribed
INTERVENTION
on emails/calls to help-line
Development of the healthy food prescription incentive
programme was informed by the social prescribing liter-
Implementation fidelity
ature, Research to Equip Primary Healthcare for Equity
Continued
principles of equity-oriented healthcare,88 initiatives
Questionnaires
elsewhere (eg, Wholesome Wave89), and stakeholder
consultation. The comparison group will receive a one-
Time point
time healthy food prescription pamphlet. The incentive
Table 1
group will receive a one-time healthy food prescription
pamphlet and a weekly incentive valued at CDN$10.50/
household member (ie, CDN$1.50/household member
6 Olstad DL, et al. BMJ Open 2022;12:e050006. doi:10.1136/bmjopen-2021-050006Open access
BMJ Open: first published as 10.1136/bmjopen-2021-050006 on 15 February 2022. Downloaded from http://bmjopen.bmj.com/ on February 25, 2022 by guest. Protected by copyright.
per day) to purchase healthy foods in study- affiliated
Table 2 Foods that qualify for the healthy food incentive
supermarkets. Thus, the study is designed to test the
impact of a healthy food incentive, which is an inter- Food group Eligible items
vention that targets economic rather than knowledge- Vegetables and fruits Fresh vegetables and fruit
related barriers to healthy eating. Aside from labelling Frozen vegetables and fruit
the nutritional advice delivered as a ‘prescription’ (which Canned vegetables
may have some independent impact on participants’
Meat, poultry and fish Fresh meat, poultry and fish
behaviour), the healthy food prescription closely mimics
current care (ie, nutrition counselling) and is unlikely Canned fish
to substantially change dietary intake in the context of Meat alternatives Dried or canned lentils, chickpeas
significant economic constraints.90–95 The value of the or beans
incentive exceeds the benefit provided by many similar Whole eggs
US programmes56 89 in order to more closely bridge the Whole almonds
gap in food spending between food secure and insecure
Dairy products White cow’s milk
households in Canada.96 The value of the incentive that
each household will receive will be calculated based on Unsweetened fortified soy
beverage
the number of household members at baseline and will
remain consistent throughout the intervention regard- Plain yoghurt
less of changes in household size. A household member Hard cheddar cheese
is defined as a partner or a dependent child or adult who Whole grain foods Whole grain pasta
resides at the same location at least 50% of the time. The Brown rice
intervention will be delivered over 6 months to allow suffi-
Large flake rolled oats
cient time for dietary changes to be reflected in approxi-
mately two A1C cycles.97 100% whole wheat bread
The healthy food prescription pamphlet was designed Bran flakes cereal
by registered dietitians and modelled after a previous
food prescription programme to be a visually appealing,
low literacy resource98 (online supplemental additional receive a loyalty card points payback with a redeemable
file 5). The cover page contains the following preprinted value of CDN$21. The value of the points incentive is
prescription ‘I prescribe a healthy eating pattern of capped at CDN$10.50/household member, meaning that
minimally processed foods that have little to no added households that exceed this spending threshold will not
fat, sugar or salt,’ with space for the care provider to add receive additional points, while those that do not meet
their signature, date and patient information. The inner this threshold will not receive any points that week. The
pages outline an evidence-based healthy dietary pattern, offers will be renewed weekly. While progress towards the
with key messages to consume a variety of whole, mini- minimum spend for triggering the points payback will be
mally processed foods from all food groups with little to reset weekly, loyalty card points never expire and will carry
no added fat, sugar or salt, to spread carbohydrate foods over between weeks if left unspent. Loyalty card points
over the day, to satisfy thirst with water, and to avoid can be redeemed in CDN$10 increments to purchase
sugary drinks, refined grains, sweets, confectionary and anything in store, with no restrictions. Importantly, while
desserts.7 99 A diabetes- appropriate recipe is provided there is no requirement to do so, participants may use
along with links to connect patients with sources of free/ loyalty card points as payment for purchases that will earn
lower-cost food, additional recipes, nutrition information, them even more points in return (ie, by using their points
other helpful community services and sources of emer- to purchase incentive-eligible foods).
gency food assistance. Feedback from PCC staff, patients At baseline, participants’ loyalty cards will be preloaded
and the advisory boards was incorporated into the final with the dollar amount of points that matches their house-
version of the pamphlet. hold size so that they can earn their first points payback
The healthy food incentive consists of a weekly incen- by purchasing incentive-eligible foods without paying out-
tive valued at CDN$10.50/household member to of-pocket. Participants will then be encouraged to repeat
purchase healthy foods in study-affiliated supermarkets. this pattern of redeeming loyalty card points weekly to
The list of incentive-eligible foods includes whole, mini- earn more loyalty card points for shopping the following
mally processed foods with little to no added fat, sugar week. Participants who run out of loyalty card points
or salt from all food groups7 99 (table 2). Once a house- to meet their offer’s spending threshold can request a
hold reaches their spending threshold they will receive second allocation of loyalty card points for resuming the
an immediate payback in loyalty card points of the same cycle of redeeming points to earn more points without
value (ie, a redeemable value of CDN$10.50/household spending out-of-pocket.
member). For instance, if a two-person household spends A booklet was created with pictures of incentive-eligible
CDN$21 over a 1-week period on incentive-eligible foods foods to assist participants to locate them. The process
(in a single shop or across multiple shops), they will of collecting and redeeming loyalty card points using the
Olstad DL, et al. BMJ Open 2022;12:e050006. doi:10.1136/bmjopen-2021-050006 7Open access
BMJ Open: first published as 10.1136/bmjopen-2021-050006 on 15 February 2022. Downloaded from http://bmjopen.bmj.com/ on February 25, 2022 by guest. Protected by copyright.
booklet was pilot tested with two participants, with good will also assess diabetes competing demands114 and
results. Research assistants will review the booklet and perceived financial barriers to chronic disease care, the
rules pertaining to how loyalty card points may be earned latter of which has undergone testing via focus groups
and redeemed with participants prior to the intervention. and cognitive interviews.115 Subjective social status will be
They will also assist participants to download the super- assessed using the MacArthur Scale of Subjective Social
market’s app where they can review details of the healthy Status national and community ladders.116 117 Participants
food incentive, monitor their loyalty card points balance, will report perceived income adequacy by answering the
and their progress towards meeting their weekly spending question: ‘Thinking about your total monthly income,
threshold. Participants without mobile phones can also how difficult or easy is it for you to make ends meet?’.83–85
login to their loyalty card account via computer or consult Quality of dietary intake will be assessed using the
the bottom of their store receipt to view the number of online Automated Self- Administered 24-hour Dietary
loyalty card points they have accumulated. Research assis- Recall for Canada (ASA24- Canada-2018) whereby all
tants will email/text participants at the beginning of the participants will report all foods and beverages consumed
intervention to identify and resolve any difficulties they from midnight to midnight the previous day, including
may have had in collecting and/or redeeming loyalty location of consumption and dietary supplements.118–120
card points. Participants will also have continuous access The ASA24 has demonstrated good correspondence with
to a study email and telephone help-line where they can standardised interviewer administered dietary recalls
ask study-related questions and inquire about their loyalty and with true intakes.119 121 Participants will receive an
card points balance or spending progress. unannounced email/text 2–4 days later prompting them
to complete a second dietary recall to provide a more
precise estimate of usual intake.122 Dietary intake data
DATA COLLECTION will be used to calculate subscores and an overall Healthy
To support retention, all participants (regardless of treat- Eating Index-2015 score from 0 to 100 for each partic-
ment allocation) will receive CDN$100 following comple- ipant, which provides a valid assessment of overall diet
tion of data collection at baseline (0 months) and again quality consistent with recommendations in the healthy
at follow-up (6 months). food prescription pamphlet.123 124
To reduce missing data, REDCap will be configured to
Questionnaires require a response prior to proceeding to the next ques-
Electronic questionnaires will encompass sociodemo- tion, although ‘don’t know’ and ‘refuse to answer’ will be
graphic and health-related items, dietary intake in the response options. Research assistants will also review all
previous 24 hours, and a variety of patient- reported
completed questionnaires and will telephone participants
outcomes. The final questionnaires will be reviewed by
within 24 hours to request responses to any unanswered
the advisory boards and scientific steering committee to
questions.
establish face and content validity, and will be pretested
with patients.
Sociodemographic and health- related variables will Clinical measurements
be recorded in REDCap using existing items from the Biochemical measurements will include quantifica-
Canadian Community Health Survey where available,100 tion of blood glucose levels via A1C (standardised to
including date of birth, sex at birth, gender identity, the Diabetes Complications and Control Trial)125–128
race/ethnicity, years lived in Canada, household size and fructosamine, as A1C can be unreliable for some
and composition, number of household members with patients and fructosamine is more sensitive to acute
T2DM, educational attainment, employment status, changes.129 Blood lipids (total, HDL and LDL choles-
marital status, annual household income, main income terol, triglycerides, apolipoprotein B), serum creati-
source, access to extended health benefits, participation nine (to calculate estimated glomerular filtration rate),
in income support programmes, smoking status, housing albumin-to-creatinine ratio and haemoglobin concen-
status, medication/insulin type and dose, duration of tration will also be quantified. Participants will provide
diabetes and physical activity level.101 a urine sample to detect albuminuria. All samples will
Patient-reported outcomes will be assessed using the be analysed by Alberta Precision Laboratories and
following validated scales: WHO Well- Being Scale,102 DynaLIFE Medical Labs.
Stanford Diabetes Self-Efficacy Scale, 103 104
Diabetes Self- Physical measurements will adhere to standardised
Management Questionnaire,105 106 Problem Areas in measurement protocols and will be performed a minimum
Diabetes Scale-5 to assess diabetes distress,107 EQ-5D-5L of two times by trained researchers/clinicians, including
to assess self-rated health,108 hypoglycaemic episodes,30 weight and height to calculate BMI, waist circumference,
Work Productivity and Activity Impairment Ques- systolic and diastolic blood pressure (using oscillometric
tionnaire,109 Health Canada’s Household Food Secu- devices approved by Hypertension Canada) and heart
rity Survey Module to assess experiences of marginal, rate. Skin carotenoids will be assessed using Pharmanex
moderate and severe household food insecurity in the Biophotonic Scanners as biomarkers of fruit and vege-
past 6 months110–112 and medication adherence.113 We table intake.130
8 Olstad DL, et al. BMJ Open 2022;12:e050006. doi:10.1136/bmjopen-2021-050006Open access
BMJ Open: first published as 10.1136/bmjopen-2021-050006 on 15 February 2022. Downloaded from http://bmjopen.bmj.com/ on February 25, 2022 by guest. Protected by copyright.
Administrative health data Sensitivity analyses
A1C levels and information on comorbidities and Sensitivity analyses will examine outcomes among patients
diabetes complications will be obtained from Alberta whose antihyperglycaemic medication/insulin regimen
Health Services’ Analytics, Data Integration, Measure- was unchanged during the 3 months prior to the study,
ment and Reporting database. The secondary outcome throughout the study period, and when patients taking
of need for antihyperglycaemic medication/insulin will insulin are excluded. We will also examine the impact of
be quantified by monitoring changes in medication/ excluding patients who were started on lipid-lowering or
insulin use (ie, initiation or discontinuation), type (ie, anti-hypertensive therapy from models assessing impact
Metformin, Sulfonylureas, Repaglinide, DPP- 4 inhibi- on blood lipids and blood pressure, respectively. Addi-
tors, GLP1 receptor agonists, SGLT2- inhibitors, Acar- tional sensitivity analyses will consider the impact on find-
bose, Thiazolidinediones, Statins or other lipid-lowering ings when food insecurity is modelled as a continuous,
agents, Renin- angiotensin aldosterone antagonists and rather than as a categorical outcome,132 when diet quality
other anti-hypertensive agents) and dosage recorded in is assessed via the new Healthy Eating Food Index-2019,133
the Pharmaceutical Information Network Database.131 and when an indicator of energy intake misreporting (ie,
We will also collect administrative data on health events the ratio of reported energy intake to estimated energy
and healthcare use on an ongoing basis postintervention expenditure) is included in models assessing impact on
to support understanding of longer-term outcomes. diet quality.
We will also consider the impact on findings when
Implementation fidelity models are adjusted for changes in medication/insulin
Implementation fidelity will be continuously assessed type and dosage that occurred between baseline and
using a combination of quantitative and qualitative follow-up. We propose to use a novel scoring system that
measures as part of the implementation study (described attempts to match the changes made with the expected
below). clinical impact on A1C.134 The following changes will be
assigned one point (expected change in A1C of ~0.5%):
less than full dose of Metformin (Open access
BMJ Open: first published as 10.1136/bmjopen-2021-050006 on 15 February 2022. Downloaded from http://bmjopen.bmj.com/ on February 25, 2022 by guest. Protected by copyright.
from Alberta Health Services’ Analytics, Data Integration, Implementation study
Measurement and Reporting database, including emer- Study design and objectives
gency department, inpatient, specialist, general practice A mixed- methods implementation study will eval-
and urgent care costs. The costs of the intervention will uate the RE- AIM68 69 of the healthy food prescription
include administrative costs associated with implementing incentive programme in order to understand determi-
the programme and the costs of incentives at a house- nants of effective implementation and reasons behind
hold level. The time to administer the programme will be programme successes and failures (table 3). The Consol-
based on the non-research hours of the study personnel idated Framework for Implementation Research (CFIR)
and unit costs will be obtained from the Alberta Wage and consolidates determinants of effective implementation
Salary Survey.140 into five domains (intervention characteristics, inner
setting, outer setting, characteristics of individuals, imple-
Data analysis mentation process), and will accordingly structure our
The potential longer-term health and economic impacts investigation of determinants of effective implementa-
of the healthy food prescription incentive programme tion, including barriers and facilitators, within RE-AIM’s
will be modelled using the validated UK Prospective implementation domain.143
Diabetes Study outcomes simulation model-2.141 Health-
related model outputs will include differences between Implementation process
the incentive and comparison groups in cardiovascular The implementation process will unfold according to
events, amputation, blindness, renal failure, diabetic the four phases and action-oriented steps in the Quality
foot ulcers, and mortality over 1 year, 5 years and lifetime Implementation Framework (QIF).144 145
scenarios. Model outputs related to the economic impacts QIF phase 1: Initial considerations regarding the host
of the intervention will include the incremental differ- setting
ence in costs and quality-adjusted life-years142 between 1. Stakeholder buy-in: Partnership agreements will be fi-
the incentive and comparison groups, the incremental nalised with all stakeholders.
cost-effectiveness ratio and the net benefit. A budget 2. Implementation support team: An implementation
impact analysis will explore the difference in costs (eg, support team will be formed to administer, plan, sup-
prescriptions, physician visits, hospitalisations) consid- port, monitor and evaluate implementation of the in-
ering the RE-AIM domains, and return on investment tervention.
from the public payer perspective over 1 year, 5 years and 3. Training: Study personnel will be trained in principles
lifetime scenarios. of equity-oriented care88 146 147 and study procedures.
Table 3 Logic model for the implementation of a healthy food prescription incentive programme
Goal: to support adults who are experiencing food insecurity and persistent hyperglycaemia to manage their diabetes with a healthy diet.
Situation: In Alberta, more than 54 000 adults are experiencing food insecurity and type 2 diabetes, including 13 600 Indigenous individuals who bear
a disproportionately high burden171–174 This group of Albertans incurs CDN$842 million/year in healthcare costs, with a small subset of 9600 individuals
with persistent hyperglycaemia incurring nearly one-quarter of these costs.20 24 170 171 175 Nevertheless, although it is well known that food insecurity is a
primary driver of acute care usage and costs, primary care providers often lack effective strategies to address it.
Inputs Activities Outputs Short-term outcomes Longer-term Impact
outcomes
► Patient-oriented research, ► Development ► Healthy food ► Successful integration ► Improved quality ► Decreased
with patients as partners of partnership prescriptions of care pathways of care acute care
► A type 2 hybrid agreements prescribed within PCC workflows ► Improved patient usage
effectiveness- ► Readiness, capacity, ► Healthy food ► Increased awareness satisfaction ► Decreased
implementation study barriers/facilitators incentives offered, of effective strategies ► Improved acute care
design and implementation earned and redeemed to reduce food glycaemia costs
► Scientific committee, assessments ► Patient, care provider insecurity ► Reduced
advisory boards and PCC ► Cocustomisation and PCC participation ► Increased chronic diabetes
subcommittee of care pathways ► Staff training empowerment for complications
► PCC support and and implementation ► Patient and provider patients and care ► Commitments
infrastructure strategies experiences and providers from Alberta
► Organisational champions ► Education and perceived outcomes ► Increased care Health Services,
► Funding and in-kind training, including ► Determinants provider motivation to PCCs, Alberta
support from Alberta booster training of effective sustain care pathways Blue Cross and
Innovates, Alberta Blue ► Ongoing monitoring implementation ► Improved diet quality supermarkets
Cross, Alberta Health and evaluation ► Reasons for ► Reduced food to collaborate
Services and Nu-Skin ► Regular programme insecurity for longer-term
► Implementation support communication, successes/failures ► Improved diabetes sustainability
team including continuous ► Cost-effectiveness management
► Technical support implementation analysis
feedback
PCC, primary care clinic.
10 Olstad DL, et al. BMJ Open 2022;12:e050006. doi:10.1136/bmjopen-2021-050006Open access
BMJ Open: first published as 10.1136/bmjopen-2021-050006 on 15 February 2022. Downloaded from http://bmjopen.bmj.com/ on February 25, 2022 by guest. Protected by copyright.
4. Assess needs, fit, capacity, readiness and adaptations: 7. Study planning: The implementation support team
Implementation strategies will be tailored by clinic us- will develop study protocols and procure materials for
ing a theory-informed modified conjoint analysis148 in all three studies.
which PCC staff will complete a questionnaire to iden- QIF phase 2: Creating a structure for implementation
tify potential implementation barriers and facilitators 1. Implementation planning and adaptations: The im-
within the five domains of CFIR. Researchers will use plementation support team will use findings from the
the CFIR-Expert Recommendations for Implementing modified conjoint analysis to develop a detailed im-
Change compilation matching tool to identify strate- plementation plan and timeline, cocustomise it with
gies to mitigate the barriers and leverage the facilita- PCCs, and assign specific roles and responsibilities.
tors identified by each clinic.143 148–152 Incentive-related procedures will be finalised with our
5. Preimplementation planning and adaptations: The supermarket partner.
implementation support team will develop a preim- QIF phase 3: Implementation and ongoing implemen-
plementation plan and timeline and will execute it, tation structure
including codeveloping infrastructural supports, train- 1. Programme implementation and data collection: The
ing modules and care pathways with PCC staff and healthy food prescription incentive programme will
patients. The Indigenous advisory board will progress be implemented and data collection for the RCT and
relationship building with Indigenous PCCs and will modelling study will proceed (figure 2).
work with them to adapt infrastructural supports and 2. Technical support and communication: The implemen-
care pathways as required. tation support team will provide ongoing support to
6. Capacity building and supportive organisational cli- PCC staff, including via weekly meetings with staff desig-
mate: PCC staff will be trained in principles of equity- nates. Booster training sessions will be held when new/
oriented care88 146 147 and study procedures. Training modified processes are introduced and for new staff.
sessions and codesign processes will enhance buy-in 3. Implementation study and feedback mechanisms: The
and readiness to change. One staff designate per or- implementation support team will collect data contin-
ganisation will liaise with the implementation support uously for the implementation study. Ongoing moni-
team weekly and will serve as an organisational cham- toring and provision of feedback to PCCs will support
pion. continuous quality improvement.
Figure 2 Healthy food prescription programme care pathway.
Olstad DL, et al. BMJ Open 2022;12:e050006. doi:10.1136/bmjopen-2021-050006 11Open access
BMJ Open: first published as 10.1136/bmjopen-2021-050006 on 15 February 2022. Downloaded from http://bmjopen.bmj.com/ on February 25, 2022 by guest. Protected by copyright.
QIF phase 4: Improving future applications reported by patients via quantitative checklists and semi-
1. Data analysis: The research team will analyse and inte- structured interviews.
grate data from all three studies.
2. Knowledge translation and learning from experience: Data analysis
The research team and advisory boards will jointly in- Quantitative data
terpret and disseminate findings. Outcomes from a de- Quantitative findings pertaining to all five RE- AIM
liberative dialogue153 154 knowledge translation event domains will be summarised using descriptive statistics
will inform sustainability planning. and will inform areas for subsequent in-depth qualitative
exploration. We will stratify our analyses by clinic type
Data collection (urban, rural, Indigenous) to examine any meaningful
Implementation processes and outcomes related to differences between them.
all five RE-AIM domains, and determinants of effective
implementation within CFIR domains, will be repeatedly Qualitative data
measured via quantitative and qualitative data collected Qualitative data will be coded by two trained researchers
by trained research assistants. Participants will receive using directed content analysis,155 whereby development
CDN$30 for participating in interviews. of an initial coding scheme for each set of interviews will
be informed by RE-AIM, CFIR and other frameworks as
Quantitative data appropriate. Concurrent data collection and analysis and
Quantitative data will be collected via the following: (1) regular meetings between researchers will permit itera-
Administrative records of patients, care providers and tive adjustments to the interview questions and coding
PCCs that did and did not participate (reach and adop- schemes, and continuous evaluation of the adequacy of
tion); care providers trained, healthy food prescrip- the samples.156 Sampling will end when new concepts are
tions prescribed, and healthy food incentives offered, no longer being identified in the data.
earned and redeemed, including redemption location
Data integration
(implementation); (2) Implementation fidelity check-
Quantitative and qualitative data will be integrated during
lists (implementation) and (3) Quantitative question-
the analysis stage for the purposes of expansion (eg, qual-
naire items completed by PCC staff and patients to
itative data will help to elaborate and explain quantitative
report perceived programme outcomes (effectiveness);
findings) and convergence (eg, to examine whether quan-
perceived programme experiences, facilitators, barriers,
titative and qualitative fidelity ratings correspond).157
mechanisms of impact, quality of infrastructural supports
and determinants of effective implementation (imple- Data integration and dissemination
mentation); and longer- term programme feasibility, Data from each of the three studies will be published
acceptability and willingness to participate in or deliver separately, with an additional final publication that will
it, success in integrating the programme within existing integrate and synthesise their collective findings across
workflows and how aspects of the programme were all RE-AIM domains (table 4). These fully integrated data
sustained over time (maintenance). will be disseminated via technical reports, lay summaries,
infographics, policy briefs, academic publications and
Qualitative data oral/poster presentations.
Qualitative data will be collected via the following: (1)
Reported reasons why patients, care providers and PCCs
decline to participate in or drop out from the study DISCUSSION
(adoption and maintenance); (2) Qualitative ques- Adults who are experiencing food insecurity cannot
tionnaire items completed by PCC staff and patients to consume the healthy foods they require to manage their
provide suggestions for programme improvement (main- diabetes if they lack sufficient funds to purchase them.
tenance); (3) Notes from patient emails/calls to the However, primary care providers often lack access to
study help-line (all domains); (4) Notes from meetings resources that could assist them to alleviate their patients’
with PCC and supermarket staff liaisons (all domains); experiences of food insecurity. By addressing income-
(5) Semistructured interviews with patients and members related causes of unhealthy dietary patterns and persistent
of the Indigenous advisory board (all domains) and (6) hyperglycaemia, healthy food prescription programmes
Qualitative observations of Indigenous advisory board can equip clinicians with resources that assist their
meetings (all domains). patients to maintain a healthier dietary pattern. Over the
longer-term, maintenance of a healthier dietary pattern
Implementation fidelity can improve health and reduce diabetes-related health-
From the measures summarised above, objective measures care expenditures.7–15 64
of implementation fidelity will include administrative We will investigate the RE-AIM of a healthy food prescrip-
records of healthy food prescriptions prescribed, and of tion incentive programme for adults who are experi-
healthy food incentives offered, earned and redeemed. encing food insecurity and persistent hyperglycaemia.
Perceived measures of implementation fidelity will be Through an RCT, modelling and implementation studies,
12 Olstad DL, et al. BMJ Open 2022;12:e050006. doi:10.1136/bmjopen-2021-050006You can also read
