FOOD PRESCRIPTION PILOTS: Feasibility, acceptability and affordability of improving diet through menu planning and grocery delivery

Abstract

Background:

Lack of adherence is a primary reason people fail to maintain a healthy diet or lose weight. Multiple environmental factors, including aggressive marketing and convenience of nutrient-poor food undermine people’s best intentions.

Objective:

To assess the feasibility, acceptability and impact of food prescriptions in which participants’ exposure to commercial food outlets is reduced, because the groceries are delivered with weekly menu plans and recipes.

Methods:

This is a series of pre-post pilot proof-of-concept studies. We recruited 37 members of Kaiser Permanente interested in improving their diet or losing weight. Weekly meal plans meeting >90% of recommended dietary allowances (RDAs) were designed to be low cost, in line with Supplemental Nutrition Assistance Program (SNAP) allowances. Five separate pilots targeted different populations. Participants were required to provide 24 hr. dietary recalls (ASA24) before and during the interventions. Weight management pilot participants had height, weight and BP measured before and after 4-week pilots and followed sustainability guidelines, limiting meat and dairy.

Results:

Across pilots, the Healthy Eating Index (HEI) improved (+21.1 pts; 95% C, I. 15.9, 26.3). For the weight management pilots, most participants lost weight (average 10.3lbs for men, 5.7 lbs for women; 95% C.I. −10.2, −5.4). The majority of participants liked the program and considered it the easiest weight loss program they ever tried.

Conclusions:

These pilots suggest that meal planning and grocery delivery can be affordable, acceptable and could ultimately have a major impact on diet-related chronic diseases. Longer term studies are needed to confirm how long compliance will endure.

Graphical Abstract:

Introduction

The quality of the American diet is poor, with the average healthy eating index (HEI) estimated at 59 out of a scale from 0–100 that measures how well consumption aligns with the US National Dietary Guidelines.¹ Moreover, the percentage of Americans who are overweight and obese continues to grow, now affecting an estimated 73.6% of the population.² A major underlying reason for such poor diet quality and high obesity rates is the high consumption of highly processed nutrient-poor foods, which are aggressively promoted in the majority of supermarkets, small groceries, and convenience stores, as well as in most restaurants.

The current food environment fosters impulse buying and eating.³ Aggressive advertising and convenience of nutrient-poor foods are ubiquitous. Convenience stores and supermarkets tend to be open very late in the day, and some all night, so, particularly in urban areas people can obtain food at any time, obviating the need for planning. There is also an explosion of restaurants, so people can obtain prepared food outside of home, which also does not require any planning. A 2019 survey found that 29% of consumers don’t ever plan their meals and go one day at a time.⁴ Another market research firm found that only 24% of dinners are made from scratch and 53% are planned within an hour of eating.⁵ Planned meals are associated with healthier meals and healthier outcomes than those unplanned.^6,7

When food is selected on impulse, there is an increased likelihood that less healthy options are chosen, as our automatic mode of thinking favors ingredients high in sugar and fat when decisions are made quickly.^8–10 Today’s world is characterized by information overload, with a multitude of messages, advertisements, and multiple forces competing for people’s attention. Given that context, planning meals and avoiding all the distractions in grocery stores and the built environment overall that promotes nutrient-poor foods constitutes a significant burden that few households can overcome.

Structuring the home food environment with planned meals and appropriate groceries and eliminating nutrient-poor foods may be more effective for promoting a healthy diet than strategies relying on education and self-control.¹¹ Behavioral economics and choice architecture posit that changes in the decision environment affect individual behavior while preserving freedom of choice.^12–15 Modifying situations/environments to prevent impulses may be more effective than cognitive or educational strategies intended to control impulses after they are triggered (e.g., mindfulness).^16,17

Consistent with this premise, interventions that directly provide nutritious foods have yielded greater compliance than interventions that do not provide foods directly.^18,19 For example, trials have examined the effectiveness of modifying participants’ household food environment on weight management (e.g. cabinet cleanout).^20,21 Treatment resulted in significant weight loss, and outcomes seemed particularly promising for women and other adults living in the household.^20,21 Another study in which 15 participants ordered groceries for 8 weeks showed it resulted in fewer impulse buys and fewer high fat foods at home.²²

Cost or affordability has also been invoked as a barrier of healthful eating. However, in recent years the notion that healthy food is too expensive has been challenged. In fact, USDA findings support that nutritious foods – such as grains, vegetables, fruit and dairy – typically cost less per serving than items high in saturated fat and added sugars.²³ The current maximum SNAP benefit for a family of 4 has now increased to $939 per month, or about $2.61 per meal, so cost may be less of a hurdle for low-income households to eat a balanced diet.

Not only is overconsumption of nutrient-poor food the primary driver of morbidity and mortality in the US,²⁴ but climate change is also affected by overconsumption.^25–27 Evidence suggests that consumption of less meat would not only improve heath, but it would also reduce the forces that drive climate change and make diets more sustainable.^25–28

Because diet is the number one underlying cause of mortality in the US among all groups, responsible for nearly 530,000 deaths in 2016²⁹ and food now accounts for one-third of all carbon emissions,³⁰ we developed an intervention which would reduce the burden of planning and shopping for participants, limit their exposure to the marketing of nutrient-poor foods, help participants consume healthy meals, and limit their carbon footprint. An additional goal was to make eating healthy meals affordable for everyone, including those who rely on the Supplemental Nutrition Assistance Program (SNAP).

Below we describe 5 pilots built on the following premises 1) Avoiding retail food outlets will reduce impulse buys and help control the quality of the food at home; 2) Pre-planning meals and ensuring appropriate food availability will make it easier to prepare healthy meals and will decrease the need to rely on fast-food and restaurants; 3) Careful menu planning will make the cost of healthy foods affordable. These pilots tested the feasibility, acceptability and impact of food prescriptions, consisting of providing meal plans, recipes and the basic groceries needed to prepare the recommended meals.

Methods

We conducted 5 different pilot studies among members of Kaiser Permanente (KP), Southern California to investigate the feasibility and acceptability of combining menu planning with grocery delivery. All pilots were approved by the institutional IRB. Recruitment was a convenience sample of eligible KP members who lived near the grocery store where food was purchased or the clinic when they had to appear for measurement. Participants were invited via email or via telephone calls. (Flow charts are in supplemental materials.)

Pilot 1: One week of grocery delivery to post-partum Latino households to improve diet quality

We invited 5 women who had given birth in the previous 12 months to participate. Eligibility included having at least 4–5 household members, including the newborn. The women had to provide two dietary recalls prior to the week of grocery delivery and 2 during the week. They were also requested to give us feedback about the experience after the week. During the week of grocery delivery, they received a suggested meal plan that covered breakfast, lunch, dinner and a snack for meals that, when combined, provided 2,000 calories per day. The entire week of meals were designed to provide all the recommended dietary allowances (RDAs). Participants received recipes for the suggested meals and snacks. All the groceries needed to prepare the suggested meals were delivered to their homes, with the exception of pantry staples, like spices, condiments and oils.

Results:

Added sugar decreased by 67%, fruit consumption more than doubled and vegetable consumption increased by 79%. Solid fat consumption was reduced by 22%. Over this one week the HEI increased from 60.5 to 66. Participants said they liked participating for 2 main reasons, first, having a meal plan made it easier for them to figure out what to eat and second it saved them a lot of time by not having to go to the grocery store. During this week none of the participants dined out, when previously they would dine out several times per week. Most of the participants said they received way too much food and did not use it all. Some said they did not drink milk and others did not like all of the types and variety of vegetables that were provided. Actual grocery costs were $120.

Pilot 2. Weight loss pilot with 4 weeks of grocery delivery:

In addition to creating a meal plan that would limit calories to 1500/day, we created a menu that followed the EAT-Lancet sustainability guidelines that limited quantities of meat and dairy (2 servings each of fish and poultry per week, no more than one serving of red meat per week, and dairy limited to 1 cup-equivalent of milk/day).³¹ The quantity of food to be prepared for dinner was usually doubled, so participants could eat the leftovers for lunch the next day. The menu plans met at least 90% of the USDA RDAs for 22 nutrients (protein, fiber, vitamins A, B1, B2, B3, B6, B12, C, E, K, Folate, Copper, Iron, Mg, Mn, Phosphorus, Potassium, Selenium, Zinc, Omegas 3 and 6). The Calcium sustainability guidelines were set at 700mg/day, lower than the USDA RDAs.³¹ The quantities of fat, carbohydrates and sodium were considerably lower than maximal amount recommended for weight maintenance.

Eligibility included individuals who had been referred to Kaiser Permanente’s Center for Healthy Living for weight control counseling and a household with two adult members with a BMI ≥ 27 who both wanted to lose weight. Both members were required to come to the clinic for checks of height, weight, and blood pressure, provide two dietary recalls prior to the week of grocery delivery and once during each week that the meals were delivered. An exit interview was also requested.

Results:

We recruited 5 households, but one dropped out because they had 4 children, that made the preparation of the meals for two adults too difficult in addition to feeding the children. Of the remaining participants, everyone lost weight, with men losing an average of 13 lbs. in 4 weeks (range 5.5–17.9) and women losing 6.8 lbs. (range 4.9–9.6). Similar improvements were seen in increased fruit consumption (97.7% higher) and a significant decrease in sold fat (−70.0%) Added sugar was reduced by 51.6%. The HEI index increased from an average of 49 to 78. Participants were pleased with the results, but many found the food preparation requirements challenging. Grocery costs in February/March 2021 were an average of $77.35 per week for two people.

Pilot 3. Weight loss pilot with 4 weeks of grocery delivery targeting Latino households:

This was implemented the same way as Pilot 2, only the meal plans focused on Mexican/Latino recipes and only Latino households were eligible for participation.

Results:

We started with 5 Latino households, but one family dropped out in the last week due to a car crash and hospitalization. Of the remaining 8 participants, 5 lost weight. For men the average was 10 lbs. (range 0–18) and 5 lbs. for women (range 0–10) Dietary quality also improved with the average HEI moving from 51.3 to 63.0, but, overall, this change was not significant. The consensus from all participants was that the menu plans involved too much work. They found many of the traditional Mexican recipes too complicated and demanding. The couple that did not lose weight was totally overwhelmed and had little cooking experience and did not follow the plan. However, those who lost weight liked the convenience of grocery delivery and meal plan. For many, understanding appropriate portion sizes was a revelation. Grocery costs during June 2021 were an average of $88.00/week.

Pilot 4. Weight loss pilot targeting Type 2 non-insulin dependent diabetics with 6 weeks of grocery delivery:

This pilot focused on individuals with Type 2 diabetes with poor glucose control. Eligibility included one household participant having a HgA1C >8.0 and this person as well as one other household member was required to have a BMI≥27. In addition to meeting >90% of all critical RDAs and sustainability guidelines, the menu plans limited carbohydrate intake to < 50 grams per meal and daily calories did not exceed 1500. The menu was simplified to limit preparation time to less than 30 minutes per meal. To reduce the frequency of cooking, we planned for batch cooking, the preparation of larger quantities that would provide leftovers for additional meals beyond next day’s lunch. In addition to weight, height and blood pressure, we also measured HbA1c prior to the pilot and after 6 weeks of grocery delivery. While HbA1c improvement is usually measured after 3 months, 6 weeks should capture about 60% of the expected improvement.³²

Results:

Four households (8 individuals) were recruited. Only half lost weight, between 10–18 pounds). Among those with T2DM, HbA1C improved, with 2 persons showing dramatic improvement, a drop of 2–3 points after only 6 weeks. (See Table 2). The healthy eating index also improved from an average of 50.7 to 70.7 during the weeks of grocery delivery. One couple did not follow the meal plan and continued to eat candy and snacks frequently. The others liked the variety and meal plans One person photographed all the meals prepared. (See supplemental material for samples.) Grocery costs were an average of $95.62/week during September 2021.

Table 2.

HbA1c changes for Diabetic pilot

Diabetic participants	Baseline HbA1c	6-week follow-up HbA1c
1	8.9	8.2
2	9.8	9.6
3	11.8	9.9
4	13	10.2

Pilot 5. Weight loss pilot for 4 weeks, with participants paying $80/week for groceries.

In this pilot we simplified the menu that was used in the second pilot by repeating more of the meals to reduce the frequency of preparation to 4–5 days of cooking per week. To test the sustainability and scalability, participants had to pay $80.00 per week for the groceries that were delivered each week.

Results:

We recruited 4 households for this pilot and all lost weight (range 5–17 lbs.). The weight loss was very similar to pilot #2. Men lost an average of 13.4 pounds and women an average of 6.4 pounds over the 4 weeks. The HEI index increased from 48.2 to 77.0. During February-March 2022, average actual grocery costs were $98.27/week for a household of 2 adults.

Table 1 shows the combined results and averages across the weight loss pilots. Although in the diabetes pilot some non-adherent participants gained weight, overall the weight loss for men exceed 10 lbs. and 5 lbs. for women. Blood pressure did not change significantly as a whole.

Table 1.

Pilot results summary

	Dura-tion	#	Avg Age	Avg Weight change for men in lbs (range)	Avg. Weight change for women (range)	95% C.I. All participants p-value	Baseline HEI	Intervention HEI	Post Intervention HEI	Avg. HEI Change (95% CI)	Base-line BP	Post BP	Average cost of groceries per week	Dates
Latina Postpartum* (2000 cal/day)	1 wk	5*	32.4	n/a	n/a	n/a	60.5	66.0	55.1	+10.68 (−3.620, +24.972); p=0.10	n/a	n/a	$120*	Sept, 2020
1st weight loss Menu	4 wks	8	39.6	−13.0 (−17.9 to −5.5)	−6.8 (−9.6 to −4.9)	(−13.90, −5.68) p< 0.0008	49.0	78.0	61.1	+30.36 (+23.97, +36.76); p<0.0001	124/74	118/73	$77.30	Feb 2021
Mexican Menu	4 wks	8	35.7	−10.0 (−18.0 to −1.4)	−5.0 (−10.0 to 0)	(−12.42, −1.68) p =0.02	51.4	63.0	65.1	+11.97 (−4.52, +28.47); p=0.13	109/71	109/68	$88.00	June, 2021
Diabetic Menu	6 wks	8	47.8	−4.9 (−18 to +7.6)	−4.5 (−15.2 to +4.0)	(−12.50, 3.30) p= 0.21	50.7	70.7	68.8	+20.01 (+7.47, +32.55); p=0.007	132/78	127/75	$95.62	Sept, 2021
Weight loss-$80/week	4 wks	8	50.1	−13.4 (−15.5 to −11.3)	−6.4 (−7.6 to −5.1)	(−13.32, −6.32) p< 0.0003	48.2	77.0	63.8	+28.42 (+18.74, +38.10); p=0.0002	121/72	121/72	$98.27	Feb/March 2022
Total average weight change participants (Average)	--	32	(41.1)	−10.3	−5.7	(−10.20, −5.43) p <0.0001	(49.8)	(72.2)	(63.2)	+21.09 (+15.92, +26.27); p<0.0001	(122/74)	(119/71)	$89.80	--

^*Targeted family of 4–5, but only one person completed dietary recalls, not for weight loss; HEI= Healthy Eating Index

Participant Feedback:

The majority of participants had positive views of the intervention and liked the fact that their planning and shopping burden was reduced. Many were surprised by the portion sizes of meat, which were significantly smaller than previous consumption, but found they did not need the larger quantities. Others were surprised by the large quantity of fruit and vegetables supplied and acknowledged that they ate more because they were now readily available. Everyone thought they received too much food and ended up with a lot of leftovers. They liked most of the recipes and considered it somewhat of an adventure to discover new meals and tastes. Some liked to have more variety, while others were happy to have some recipes where they could prepare in bulk and heat up leftovers for a few different meals. However, a small number of participants that did not like to cook or have a schedule that was conducive to regular cooking did not do well with the prescribed routine. In nearly all cases, men lost twice as much weight as the women, and several women were disappointed that their weight loss was not as dramatic as their partners. Some participants thought they might be able to shop for the groceries on their own once the pilot ended. For some, having the recipes and ingredients customized to their personal tastes would be necessary to for them to buy into this kind of service over the long term.

Discussion

This series of pilots demonstrates that over a short period of time food prescriptions can be a very effective method for improving diet quality and for supporting weight loss for most participants. For diabetics it can also support glucose control. It also demonstrates that a healthy diet can be affordable for anyone. However, obtaining foods at reasonable prices did require shopping at multiple locations and capitalizing on sales and specials at each business.

The best results for weight loss were among the group that paid for the groceries, possibly because they were more motivated and committed. The group that didn’t do as well with weight loss was the pilot targeting out of control diabetics, possibly because they already were struggling to reduce their intake of sugar and carbohydrates. We provided a reduced average carbohydrate intake meal plan that may not fit each person’s carbohydrate needs with varying sizes and activity levels. It was a conservative guide to meet a wider range of people, but we did see benefits in HbA1c even at six weeks.

Other studies have previously shown that when foods are delivered with the appropriate caloric and nutrient content, people do have a higher quality diet.^33–35 However, delivering prepared foods is relatively expensive and likely unsustainable for the majority of Americans. Being able to help families plan and prepare meals on their own is something that would not only support families in obtaining a healthy diet, but also potentially lower their food costs, by reducing nutrient-poor foods and capitalizing on the lower costs of seasonal fruits and vegetables. We essentially provided an immersive experience of meal planning and cooking, with the opportunity to try balanced nutrient rich eating, previously not known or experienced.

Limitations:

The main limitations are the small sample sizes and the limited duration of the intervention. Between one and six weeks, compliance was good for most participants, as we saw dramatic results in diet quality improvement and significant weight loss among most participants. In addition, household sizes were limited to two individuals, so it remains to be seen as to how this can be scaled to meet the needs of households of different sizes and caloric needs.

Although we demonstrated that a healthy diet is affordable, the costs that we covered did not include pantry staples, like condiments, spices and oils. Those items would increase the cost, but divided over an entire year, may increase the overall costs only slightly. In addition, we did shop in California which does have year-round fresh produce that is grown relatively locally, which may underlie the relatively low cost of fresh fruits and vegetables.

For a low-calorie diet, without a careful selection of recipes, it is difficult to meet all of the most critical RDAs, especially to achieve sufficient levels Vitamin E, omega 6, calcium and B12. Limiting consumption of dairy and meat makes it particularly difficult to achieve all the daily recommended dietary allowances (RDAs), particularly for vitamins like B12, but could be met with careful planning.³⁶ When people do try to plan meals and reduce their food-based carbon footprint, it is not intuitive which foods need to be included to meet all the RDAs. We could not achieve the RDAs without using nutrient analysis software. When we established the menu we used a standard person’s height and weight, which may have been lower than some of our participants; therefore, for participants who were considerably taller, we may have not reached their real RDAs.

Our pilots were limited to a short period of time, so it is still unknown whether participants would stick to this, once the grocery delivery and recipes were no longer novel. The families did appreciate new recipes so future work to extend this service will need to constantly provide new and tasty recipes to keep participants engaged.

While the first 4 pilots provided food for free, to see whether the concept would work and be acceptable, the final pilot in which people paid for their groceries demonstrates that this kind of program can eventually be sustainable in the future. Most participants wished they could continue the service and the majority ended up saving a lot of money, because having the food at home, reduced the frequency of dining out which is considerably more expensive than eating at home.

We did not include the costs for shopping and delivery done by the study team. We did test the feasibility of using a commercial shopping and delivery service with a $5.99 delivery charge plus 10%. However, there were some logistical problems such as shoppers not following directions, not providing the specified quantities, or selecting the wrong products. It is possible that these kinds of services may be acceptable in the future with appropriate shopper training.

Our future work will focus on adjusting meal plans and recipes to meet the needs of an entire household, not only to support weight loss, but to also support healthy growth for children, and weight management for other family members who need to maintain their current weight.

Conclusion:

It appears completely feasible, acceptable and affordable to provide meal planning and grocery delivery to meet the US dietary guidelines as well as to reduce the food system’s carbon footprint and at the same time improve dietary quality and support weight management.

Key Points.

• Healthy groceries meeting RDAs and sustainability guidelines are affordable and fall within SNAP maximum allotments for < 2000 calories per person per day

• Menu planning and grocery delivery resulted in significant improvements in the healthy eating index (HEI)

• For overweight/obese individuals, 1500 calorie meals plans/groceries resulted in significant weight loss among most participants.

Biographies

Deborah A. Cohen, MD, MPH is a Research Scientist at Kaiser Permanente, Southern California. Her research focuses on interventions that address the built and social environments to influence health behaviors.

Erika L Estrada: Research Associate Kaiser Permanente Department of Research and Evaluation, Health Services Research Division

Monica Montes, is a Registered Dietitian Nutritionist and an advocate for a healthy lifestyle, blending nutrition, exercise and wellness principles to the communities she serves. She founded NEW Health Consultants.

Allison C. Voorhees, MPH: Research Associate with the Kaiser Permanente Department of Research & Evaluation supporting a portfolio of nutrition and physical activity-related studies in the Behavioral Research Division.

Galina Inzhakova, MPH is the Lead Sr. Research Project Manager of the Division of Behavioral Research for the Department of Research & Evaluation, Kaiser Permanente Southern California.

Claudia Rios, MS, RD: Bilingual, bicultural registered dietitian with the Goran Lab in the Division of Endocrinology at Children’s Hospital Los Angeles

Dr. Timothy Hsieh: Chief of Endocrinology at the Kaiser West Los Angeles Medical Center. Regional Co-lead for Diabetes Complete Care for the Southern California Permanente Medical Group.

Justin Tayag, MPH: Research Associate II at Kaiser Permanente Department of Research and Evaluation. Justin received his Master of Public Health with a specialization in Healthcare Administration.

Amorette Castillo: Research Associate II in Behavioral Research at Kaiser Permanente Research and Evaluation

Dr. Sean Hashmi: Regional Physician Director of Clinical Nutrition and Obesity Medicine at Southern California Permanente Medical Group.