A Randomized Trial to Address Food Insecurity and Promote Smoking Cessation Among Low-Income Adults

Abstract

Many low-income adults who smoke also have unmet social needs, such as food insecurity, which can serve as a barrier to smoking cessation. We developed a novel intervention to jointly address smoking cessation and food insecurity and assessed its feasibility, acceptability, and preliminary outcomes. We enrolled participants who screened for food insecurity, reported smoking daily, and were ready to quit. All participants received 3 months of resources navigation from a community health worker through monthly telephone calls for referrals and check-ins for smoking cessation and food access resources. Participants randomized to the intervention group received an economic intervention equivalent to the cost of 1 week of groceries/month for 3 months. We randomized 55 participants who were smoking on average 13 cigarettes/day. The trial was feasible and acceptable based on 3-month retention rates (80%) and end-of-study qualitative feedback (91% would recommend the study to others). At 3 months, participants in the intervention versus control group reported a longer length of abstinence from smoking and had a higher proportion of serious quit attempts. Results from this pilot study suggest the importance of attending to social needs, particularly food insecurity, as a strategy to promote smoking cessation among low-income adults who smoke.

Introduction

The prevalence of cigarette smoking is disproportionately high among U.S. adults with low-income, ¹ which includes population groups living with unmet social needs such as food insecurity. ² Food insecurity is a household-level economic and social condition characterized by limited or uncertain access to adequate food resources. ³ Epidemiological findings from 1999 to 2014 show that smoking prevalence increased sequentially as the severity of food insecurity increased, with smoking prevalence at 49% among U.S. adults with very low food security. ⁴ More recent findings from 2020 to 2021 continue to show a relationship between food insecurity and smoking, with a study of U.S. adults finding that longitudinal patterns of food insecurity was associated with patterns of cigarette use. ⁵ As food insecurity is posited to impede smoking cessation efforts,^2,6 there is a growing need to address food insecurity and related socioeconomic stressors as a part of smoking cessation treatment to improve upon existing approaches to smoking cessation and better meet the needs of priority populations affected by tobacco-related health disparities. ⁷

Food insecurity is a risk factor for hunger due to not having enough money or resources for food, and addressing food insecurity within smoking cessation treatment can be warranted for several reasons. The experience of food insecurity, which disproportionately impacts low-income individuals and households, ⁸ is highly stressful and associated with poor mental health.^9,10 Stress relief is one of the most commonly cited reasons for smoking, and the persistent stress of food insecurity can influence continued smoking. As nicotine suppresses appetite, ¹¹ qualitative findings with low-income smoking adults have delineated that smoking is a strategy to ignore hunger or to eat less when experiencing food hardship. ¹² Food insecurity can also negatively impact one’s cognitive resources. ¹³ The added cognitive burden of experiencing food insecurity is posited to diminish reasoning and impulse control, ¹³ which may extend to difficulty with overcoming smoking cravings during smoking cessation. Based on the available research, food insecurity and smoking are understood as having a reciprocal and mutually reinforcing association. ¹⁴ This warrants the need to develop and test intervention strategies that can address both health risks in tandem, ¹⁵ which has not yet been done to our knowledge.

To fill this gap, we developed and piloted the Food for SucCess intervention (or Food Security for Successful Smoking Cessation). This intervention aimed to reduce the burden of food insecurity for people undergoing a serious smoking cessation attempt, to examine its effects on cessation-related outcomes. As an adjunct to resources navigation for smoking cessation and for food insecurity, Food for SucCess provided short-term, flexible economic resources via cash transfer, by providing funds that can be used for groceries and food. Such funds were intended to reduce food insecurity and related stress during a critical period of undergoing a serious quit attempt. Providing flexible funds was important because the lived experiences of food insecurity can highly vary across individuals and households, but ultimately occurs due to a lack of money or resources. ⁸ Furthermore, food insecurity tends to co-occur with other social needs, ¹⁶ and researchers have recognized the interconnected nature of addressing social needs. For instance, addressing transportation needs could improve food insecurity by allowing a person to allocate more of their budget toward food, or to increase a person’s ability to get to grocery stores. ¹⁷

The current study aimed to determine the feasibility and acceptability of the Food for SucCess intervention that was delivered as an adjunct to resources navigation (ie, referrals for smoking cessation and food insecurity). We conducted a pilot randomized controlled trial (RCT) targeting low-income adults with food insecurity who smoke cigarettes daily and are ready to quit, comparing resources navigation versus Food for SucCess (ie, resources navigation with an economic intervention). We examined preliminary estimates of smoking cessation-related indicators at 3 months (making a serious quit attempt; length of smoking abstinence). We also conducted qualitative interviews with a subset of participants to understand perceptions of the Food for SucCess intervention and its helpfulness on food insecurity and smoking cessation, to inform considerations for future research.

Methods

Study Setting and Overview

This study was conducted through The MetroHealth System, a safety-net hospital system serving the needs of patients throughout Greater Cleveland, Ohio, and its Institute for H.O.P.E.™ which leads the health system’s work to identify and act on the social drivers of health. As described elsewhere,^{18 -20} starting in May 2019, MetroHealth began implementing social needs screening for their patient population, initially beginning with patients needing care coordination and social worker involvement, then expanding to primary care patient populations and beyond in 2020. Using an electronic referral platform, patients who screen for various social needs are referred to relevant community organizations to address those needs. ¹⁸

Cleveland is an important setting for this research, as combustible tobacco use remains particularly high in the city compared to state and national levels.^21,22 Furthermore, Cleveland has one of the highest rates of poverty across U.S. cities at 31%, ²³ putting its population at high risk for food insecurity. It is estimated that food insecurity affects 1 in 7 people in Greater Cleveland Area, ²⁴ reflecting a 50% increase since 2019. ²⁵ A prior study of over 45 000 adult primary care patients at MetroHealth (2019-2021) found that smoking prevalence was 34% among patients who screened positive for food insecurity, compared to 17% among patients who did not screen positive for food insecurity. ¹⁹

Study recruitment entailed a prescreening procedure using data from the electronic health record (EHR) to generate an initial list of potential participants, followed by outreach to provide study information and elicit interest. This outreach included screening for study eligibility, and those who were eligible and interested in participating provided informed consent. A pre-intervention baseline survey was administered prior to randomization. Regardless of treatment arm, all eligible and consented participants received resources navigation (described below). Participants randomized to the Food for SucCess arm additionally received an economic intervention for the 3-month duration of the study, and pilot study outcomes were assessed by a follow-up survey also at 3 months. Three months is generally considered “near future” for acting on one’s quitting intentions, ²⁶ and this duration is similar to other smoking cessation studies for low-income smokers.^27,28 Participants in the intervention group who completed the study were invited to participate in a semi-structured interview to share their experiences of the economic intervention. All study activities were conducted remotely, taking place between February 2022 and April 2023. The study was approved by the Institutional Review Boards at the MetroHealth System and at Case Western Reserve University and was registered at ClinicalTrials.gov (identifier NCT05259852). Further detail on each aspect of the study is provided in the following sections.

Recruitment and Eligibility Screening

We used available EHR data to generate a list of potential participants, who made a non-urgent primary care visit within the past year, drawn from selected MetroHealth primary care sites throughout Greater Cleveland. Prescreening was based on age (21 years or older), address (within Cuyahoga County), smoking status (currently smoking per EHR), and food insecurity status per the 2-item Hunger Vital Sign ²⁹ in reference to the past 12 months, assessed at most recent primary care visit. The Hunger Vital Sign is a validated tool for screening adolescents and adults for food insecurity, used widely in medical and community-based settings to screen for the presence of any food insecurity. ²⁹ All individuals in the prescreening list were initially sent study information by mail to their address on file, and interested individuals contacted the research team by telephone. The research team also proactively called individuals on the prescreening list to inform them about the study and conduct formal eligibility screening for those who expressed interest.

To be eligible for the study, participants were required to: report smoking cigarettes daily (defined as having smoked daily over the past 7 days), be willing to make a serious quit attempt in the next 30 days (response of “yes” to the question “In the next 30 days, are you willing to try quitting smoking for at least 24 h?”), be currently experiencing food insecurity (per Hunger Vital Sign in reference to the past 30 days, or by virtue of having visited a food pantry for food in the past 30 days, in efforts to capture individuals who may be missed by formal screening measures), considered to be the primary food shopper of their household (defined as doing half or more of the household food shopping), and have a reliable telephone number and local mailing address where they may be reached throughout the duration of the study. Informed consent was obtained from eligible and interested participants.

Treatment Arms

Upon completing the baseline survey, participants were randomly assigned to resources navigation (control) or resources navigation with Food for SucCess (intervention). Resources navigation addressed needs related to food and smoking cessation, and involved 3 monthly telephone contacts with a community health worker (CHW) who is embedded in the health system and skilled in using the EHR and referral platform to refer patients and follow up for various social needs. Food access resources included referrals to the Greater Cleveland Food Bank (for assistance with applications for the Supplemental Nutrition Assistance Program and other state-level nutrition incentive programs), as well as information on nearby food pantries. Smoking cessation resources included electronic referrals to the Ohio Tobacco Quit Line (which offers free nicotine replacement therapy), linkage with group counseling classes (American Lung Association’s Freedom from Smoking™), a primary care appointment to discuss cessation pharmacotherapy options, and self-help materials. This is consistent with prior navigation-based interventions for smoking cessation. ³⁰

For this study, the study team including 2 CHWs from the health system co-developed the initial semi-structured resources navigation protocol, in line with usual practices for food access referrals and capturing study-related data. We expanded the usual scope of CHW duties (in this particular health system) to include providing referrals for smoking cessation resources, which is aligned with overall CHW functions described in the literature ³¹ and CHW-involved smoking cessation studies. ³² We then made as-needed modifications throughout the study through regular team meetings (eg, adding comment boxes for CHWs to make relevant notes; adding prompts to monitor referral status). In the first resource call, CHWs explained the resources available and referred participants to the ones of interest. In the 2 subsequent calls, CHWs followed up on the referrals made and provided any additional referrals based on participants’ needs.

In addition to resources navigation, the Food for SucCess intervention group also received a $250 economic intervention, or approximately $83/month, intended to cover approximately 1 week of groceries, per the USDA Monthly Cost of Food - Moderate Plan. ³³ In recognition that several neighborhoods in and around Greater Cleveland lack a full-service grocer and that transportation remains a key barrier to food access, funds were disbursed as a Visa card to allow participants the agency and flexibility to shop for their preferred groceries at their preferred locations (vs a gift card to a specific grocer). Given this, while we informed participants that the purpose of the Visa card funds was to help pay for groceries and food costs while undergoing a quit attempt, they were not restricted to using the funds exclusively for groceries and food. Participants provided consent for the research team to access Visa card transaction data (ie, dates, names and locations of retailers, and amounts spent), which also allowed the research team to provide assistance when cards were lost or misplaced.

Data Collection and Measures

Participants completed a baseline survey prior to randomization, with feasibility, acceptability, and primary outcomes assessed at the 3-month follow-up survey. Surveys were either verbally administered by telephone by research staff or self-completed through a REDCap survey, and participants were compensated $20 and $30 for the baseline and follow-up surveys, respectively. Surveys assessed sociodemographics, smoking history and current smoking behaviors, level of food security (USDA 6-item Food Security Survey), and questions related to food access and general health status. Data from the CHW resource calls, such as length of telephone calls, which referrals were provided and accepted, and any CHW remarks about participant circumstances were collected through semi-structured call log forms.

Feasibility was assessed primarily by examining the enrollment rate and retention rate. To assess acceptability, the follow-up survey included yes/no and open-ended questions (eg, “Would you recommend this study to others? Why or why not?”). The primary quantitative outcomes were the number of 24-h quit attempts (categorized as 1 or more vs none) and the longest length of abstinence in days over the past 3 months. The number of cigarettes per day on a typical day was used as a proxy for smoking cessation at the 3-month follow-up.

As exploratory measures, we examined Visa card transaction data for all participants in the intervention group, mainly to categorize the types of retailers where cards were used (granular itemized data in the form of payment receipts were not requested from participants). Retailers were categorized as grocery store/supermarket, restaurant, gas station/convenience store, dollar store, pharmacy, or other non-food business. A subset of participants in the intervention group were invited for a more in-depth, post-study qualitative interview to understand their experiences in the study related to the economic intervention, such as how they personally used the Visa card, and its perceived impacts on food security and smoking cessation. These interviews (lasting 30-45 min) used a semi-structured interview guide, were conducted by telephone or Zoom with audio-recording, and participants were compensated $50 for their time.

Data Analysis

We compared participant sociodemographic and smoking-related characteristics by treatment assignment, using non-parametric t-tests, chi-square, and Fisher’s Exact tests to assess whether there were baseline differences in groups. We also compared baseline characteristics by group among those who completed the study. The length of the CHW resource calls were compared by treatment group. Three-month outcomes were examined using both complete case and intent-to-treat analysis, in which missing data on smoking variables were assumed as a negative outcome (ie, no quit attempt since baseline; 0 days abstinent from smoking since baseline; no change in cigarettes per day since baseline). Quantitative data were analyzed using SPSS and R.

We used thematic analysis ²¹ to inductively and deductively analyze the qualitative interview transcripts that were derived from the audio-recorded interviews. Research team members, including those who were not directly involved in any aspect of the interviews, read selected transcripts and provided memos, with several group discussions to develop an initial set of codes. Two authors (JKM and PN) refined the codes and applied them to transcripts to identify and resolve discrepancies, and overall themes were identified and finalized by group discussion (JKM, PN, and MCC).

Results

Figure 1 depicts the flow of participants in a CONSORT diagram, from screening to randomization. The overall recruitment pool consisted of 568 unique patients included in the study pre-screening list based on EHR data (not shown in Figure 1), all of whom were sent study information by mail. Of them, 226 individuals were subsequently reached by telephone to assess for eligibility, and 29% (n = 65) were eligible, interested, and provided consent. The remainder were either not eligible (31%) or not interested/available to participate (40%). Additionally, 10 individuals who consented did not complete the baseline survey, and therefore, a total of 55 participants were enrolled and randomized. Over the course of the 3-month study, 20% (n = 11) were lost to follow-up, with a greater attrition rate in the control group compared to the intervention group (30% vs 11%, respectively). A total of 44 participants completed the 3-month follow-up survey, with most but not all completing the 3 resource calls with the CHW, as displayed in Figure 1.

Figure 1.

CONSORT diagram showing flow of participants through the food for SucCess pilot study.

Participant Characteristics and Outcomes

Table 1 displays participant characteristics at baseline for the overall sample (n = 55). The mean age was 50 years, with 73% identifying as female, 44% identifying as Black or African American, 62% reporting an annual household income of less than $20 000/year, and 58% reporting being single/never married. At baseline, participants reported smoking on average 13 cigarettes per day, having smoked regularly for an average of 29 years. Over half (55%) reported very low food security in reference to the past 3 months. No differences were observed between participants randomized in the control arm versus intervention arm for the full study sample (n = 55) and for those who completed the 3-month follow-up survey (n = 44).

Table 1.

Participant Characteristics at Baseline, by Treatment Assignment, and Study Completion Status (N = 55).

Characteristics	Total	All randomized, by group			All completed, by group
		Control	Intervention	P value	Control	Intervention	P value
	N = 55	n = 27	n = 28		n = 19	n = 25
Age, M (SD)	50.4 (13.9)	49.9 (12.6)	50.8 (15.3)	.822	53.4 (11.6)	51.0 (15.7)	.578
Gender, %				.826			.419
Female	72.7	74.1	71.4		78.9	68.0
Male	27.3	25.9	28.6		21.1	32.0
Race, %				.923			.490
Black/African American	43.6	40.7	46.4		52.6	48.0
White	43.6	44.4	42.9		36.8	48.0
Another race	3.6	3.7	3.6		0.0	4.0
2 or more races	5.5	7.4	3.6		5.3	0.0
Educational attainment, %				.123			.147
HS diploma or less	45.5	48.1	42.9		47.4	44.0
Some college	43.6	33.3	53.6		31.6	52.0
College degree or greater	10.9	18.5	3.6		21.1	4.0
Household income, %				.825			.850
Less than $10 000	27.3	22.2	32.1		26.3	32.0
$10 000-19 999	34.5	37.0	32.1		36.8	36.0
$20 000-39 999	29.1	33.3	25.0		26.3	24.0
$40 000 or greater	7.3	7.4	7.1		10.5	4.0
Marital status, %				.932			.334
Single, never married	58.2	55.6	60.7		61.2	60.0
Married/partnered	10.9	11.1	10.7		5.3	12.0
Divorced/separated	25.5	25.9	25.0		21.1	28.0
Widowed	5.5	7.4	3.6		10.5	0.0
Cigarette use, M (SD)
# cigarettes per day	13.4 (9.3)	14.3 (10.4)	12.5 (8.1)	.487	14.1 (11.2)	13.5 (8.0)	.562
# years smoked regularly	28.9 (14.1)	30.2 (14.7)	27.6 (13.5)	.498	34.2 (13.3)	28.5 (13.8)	.186
Quit attempt in past 3 months, % yes	43.6	37.0	50.0	.333	31.6	52.0	.176
Longest length of quit attempt in days, M (SD)	10.7 (11.0)	8.6 (9.4)	12.1 (12.1)	.469	5.6 (5.0)	12.5 (12.5)	.116
Food pantry use, % yes	38.2	37.0	39.3	.864	31.6	36.0	.759
Food security, %				.534			.435
High/marginal food security	10.9	7.4	14.3		10.5	16.0
Low food security	34.5	40.7	28.6		42.1	24.0
Very low food security	54.5	51.9	57.1		47.4	60.0

Data from the resource navigation telephone calls with the CHW revealed no group differences in engagement among those who engaged with the CHW, as measured by the length of calls. On average, the initial resource call lasted 32 min, and the 2 follow-up calls lasted 29 and 19 min, respectively. For referrals made by the CHW, 92% (45 of 49) agreed to at least 1 smoking cessation resource referral (ie, quitline, group counseling, and/or pharmacotherapy appointment with doctor), with no group differences. For referrals used, 68% (27 of 40) used 1 or more referrals (ie, connected with quitline, group counseling, and/or doctor), with no group differences.

Table 2 displays smoking- and food insecurity-related outcomes at the 3-month follow-up, using both intent-to-treat and complete-case analysis. Participants in the intervention arm versus control arm had greater quit attempts, reductions in cigarettes per day, a longer length of abstinence, and higher 3-month smoking cessation rates. Statistically significant differences were detected when using intent-to-treat analysis for the number of cigarettes per day (P = .039) and whether made a quit attempt in the past 3 months (P = .013), whereas quit attempts trended toward significance when using complete-case analysis. While there were no statistically significant group differences in 3-month food insecurity levels by group, rates of very low food security were lower among the intervention group compared to the control group. Lastly, the follow-up survey indicated that 91% of study participants who completed the study perceived all aspects of the research as being acceptable, with 89% of control group participants (17 of 19) and 92% of intervention group participants (23 of 25) indicating that they would recommend this study to others. In terms of open-ended responses, a control group participant remarked “I would definitely recommend this study to others. There’s a lot of people that don’t think there is help out there. Just receiving that check-in call means a lot.” An intervention group participant remarked “Yes, I would [recommend the study to others] because quitting is important and taking the stress of food off of someone. Stress is the major reason a lot of people I know smoke and the majority of stress in their lives right now with prices rising.”

Table 2.

3-Month Outcomes Related to Smoking Characteristics and Food Insecurity Status.

Characteristics	Intent-to-treat a			P value	Complete case			P value
	Total	Control	Intervention		Total	Control	Intervention
	N = 55	n = 27	n = 28		N = 44	n = 19	n = 25
	M (SD) or %				M (SD) or %
Number of cigarettes per day	9.5 (9.4)	12.2 (10.0)	7.0 (8.2)	.039	9.0 (9.5)	11.2 (10.5)	7.3 (8.6)	.184
% smoking zero cigarettes/day	14.5	7.4	21.4	.252	18.2	10.5	24.0	.433
Quit attempt in past 3 months	61.8	44.4	78.6	.013	77.3	63.2	88.0	.090
Longest length of quit attempt in days b	15.2 (19.5)	9.3 (15.4)	18.7 (21.0)	.186	15.2 (19.5)	9.3 (15.5)	18.7 (21.0)	.186
Evidence-based cessation methods in past 3 months	80.0	74.1	85.7	.329	84.1	73.7	92.0	.100
Advice or counseling	30.9	37.0	25.0	.391	34.1	47.4	24.0	.123
Group counseling class	32.7	33.3	32.1	1.00	38.6	47.4	32.0	.359
Quitline	38.2	22.2	53.6	.124	45.5	31.6	56.0	.125
Nicotine replacement therapy	65.5	51.9	78.6	.050	68.2	47.4	84.0	.020
Other medication	32.7	37.0	28.6	.573	36.4	42.1	32.0	.540
Food security				.594				.753
High/marginal food security	14.5	14.8	14.3		18.2	21.1	16.0
Low food security	45.5	37.0	50.0		43.2	36.8	48.0
Very low food security	41.8	48.1	35.7		38.6	42.1	36.0

a

Missing data on smoking variables were assumed as a negative outcome (ie, no quit attempt since baseline; 0 days abstinent from smoking since baseline; no change in cigarettes per day since baseline).

b

Among those who made a quit attempt.

Exploratory Findings From Intervention Participants

In examining Visa card transaction data, there were a total of 306 unique transactions made across participants in the intervention group. Across all the transactions, 49% of the funds were used at grocery stores or supermarkets, 26% at gas stations or convenience stores, 9% at restaurants, 5% at dollar stores, and <1% at pharmacies, and 11% at businesses that do not sell food (eg, paying a bill and medical supply store). Qualitative interviews were conducted with 8 participants specifically from the intervention group (6 women and 2 men), which provided further context to the transaction information. We identified 3 themes regarding the perceived effects of the economic intervention.

First, the economic intervention allowed participants to meet immediate or shorter-term needs that were not covered by other sources. For instance, participants purchased fruits and vegetables that they usually cannot afford, or they bought other household necessities that are not allowed to be purchased on the Supplemental Nutrition Assistance Program (SNAP).

“I bought fresh vegetables for the first time in I can’t tell you how long. . . The best part of it was the vegetables. That was great. . . I could buy things that normally I couldn’t afford. You know, as I said, I can’t tell you how much time I spent in the produce department. You know, I’m going to get a nectarine. I hadn’t had a nectarine in years.” (p6)

“[The money] filled in on places where I wouldn’t have been able to, you know, [get] some of those over-the-counter medicines. I wouldn’t have felt real great without them.” (p26)

“[The money] was noticeable because I would usually try to buy in bulk, like the big packs of paper towels, and I had to go to the store and spend $2 or $3 for one roll of paper towels. . . It was like a pot of gold at the end of the rainbow. I could run and go get this, because I was running low on salsa, or I needed to put gas in my car. Or let me go purchase some toilet paper or some cleaning supplies so I could clean this weekend.” (p50)

Secondly, the timing of this study coincided with a period of high economic inflation in the U.S., particularly for rising food costs. Participants were already very vigilant about their usual grocery budget, as many were living on fixed incomes. Therefore, the effects of the economic intervention on food insecurity was dampened by rising food costs.

“Now that food has gone up so damn much, excuse my expression. It is very difficult to make what I get in food stamps – which is $213 – stretch. I mean, when I went to [name of grocery store chain] the other day, I almost had a heart attack when I saw the price, just like everybody’s been talking about the eggs, $4.44 a dozen. More expensive than milk. It’s outrageous. It’s very, very, very tight, and it’s very stressful.” (p22)

“Milk is so doggone high. Eggs are so freaking expensive. I went to go get some eggs and it was $6 and some change. The smaller egg was $3.99, so I had to go with my better choice, which was the smaller eggs. . . I just get a little less [food] not to spend more money. So I mean, before you could spend a little less and get more for your money. But now, it’s not as much as I used to get.” (p50)

Lastly, for some participants, the economic intervention increased awareness and motivated behavior change related to smoking cessation, in terms of cutting down, stopping, or seriously considering change. Participants varied in their perceived mechanisms of how the money motivated behavior change related to smoking cessation, such as increasing their awareness of the cost of smoking and perceiving the money as someone caring about their smoking cessation.

“I think what it did was just make me much more aware of what I was doing, like that I was smoking, and almost made me feel guilty if I wasn’t making some attempt to change that habit. Even if that wasn’t what it was intended for. . . It did make me much more aware of my smoking. . . I mean that’s how I interpreted this. . . I used to smoke a pack a day. . . I have been very aware of how much I smoke and how much I spend on cigarettes. So that’s a big change for me.” (p31)

“Actually, [receiving $80 a month] gave me a reason to quit. Where before, it was like who really cares—[my] cat? I mean, really, who cares? And yeah, there was a reason to quit.” (p6)

Discussion

This study demonstrated the overall feasibility and acceptability of a novel intervention that aimed to co-address food insecurity and smoking cessation in a diverse sample of low-income adults residing in a U.S. city with high levels of poverty and poverty-related health disparities. Notably, more than half of participants were considered to have very low food security at baseline, speaking to the level of socioeconomic disadvantage and related social needs. While prior studies have identified that food insecurity is a barrier to smoking cessation, this study is among the first to include a food insecurity reduction intervention as a key component of a smoking cessation study, in efforts to address this critical social need that intersects with smoking. The quantitative findings suggested that the 3-month Food for SucCess intervention that provided money for food and necessities in addition to resources navigation, compared to resources navigation only, was beneficial for reducing the number of cigarettes smoked and promoting a serious quit attempt, whereas the qualitative findings point to a diverse set of potential mechanisms.

Several aspects of this study related to feasibility are worth noting as strengths and related opportunities for improvement. First, our recruitment method successfully leveraged available data from the EHR, particularly recent food insecurity screening information and documented tobacco use status, to initially reach potential participants by mail. Our prior EHR-based study found that primary care patients who were screened for social needs by web versus telephone were significantly less likely to smoke, which may indicate disparities in access to and use of technology. Coupled with population-based studies showing that most people who smoke, regardless of their socioeconomic circumstances, are interested in quitting, ³⁴ the current enrollment rate (29%) provides further evidence that many lower-income primary care patients who smoke are not only interested, but also ready to quit. Even as all participants were considered food insecure, the fully remote nature of the study along with minimal technological requirements for participation enabled us to recruit a relatively diverse sample in terms of racial composition and socioeconomic status.

There are other unique strengths of this study. By partnering with a large safety-net hospital system that serves many low-income patients throughout the region, we leveraged the vital role of CHWs who were embedded in the system and adept in providing social needs navigation. As frontline public health personnel and trusted members of the communities they serve, ³⁵ CHWs provide critical connection to community resources via a wide range of roles and responsibilities, such as providing social support, access to information, and continuity in follow up. ³¹ Studies show that active involvement of CHWs in the care of low-income patients has measurable improvements in health outcomes, such as less urgent care use and more symptom-free days among children with asthma, ³⁶ and reduction of mental health symptoms among people with poor mental health. ³⁷ Given high rates of tobacco use in certain patient population groups, researchers have also highlighted the value of providing CHWs with tailored tobacco cessation training to maximize CHW-involved patient care. ³⁸ As state Medicaid programs vary on coverage or reimbursement models for CHW services, the involvement of CHWs in smoking cessation studies may also have important health policy implications. However, we note that healthcare systems and clinics can highly vary in CHW availability and involvement.

There are some notable study limitations. The overall 3-month retention rate (80%) varied by treatment assignment, with a higher retention rate in the treatment versus control group (90% vs 70%, respectively; see Figure 1). Due to the nature of the economic intervention and funds disbursement, blinding was not possible, and appears to have differentially impacted attrition through loss to follow-up. Although all participants understood and consented to random assignment, loss to follow-up in the control group tended to occur even prior to the first contact with the CHW. This suggests the need to revise aspects of the study design to mitigate the impact of group assignment expectations on dropout, such as conducting a pre-consent randomization as done in other EHR-based recruitment studies, ³⁹ or varying the amount of the economic intervention by group as done in the Baby’s First Years study, ⁴⁰ or offering a waitlist control condition. Despite differential dropout rates, we note that there were no significant baseline differences based on study completion status (Table 1), and the majority of participants in the control group found the study to be acceptable and would recommend the study to others. Nevertheless, the pilot findings suggest that it will be important to achieve a balance between the ethical and scientific considerations in randomized group assignment.

This study was also limited to participants who were willing to make a quit attempt, who had a reliable working phone and local mailing address which could suggest a form of selection bias. Those experiencing significant economic distress without reliable phone access as well as those who are unhoused, who may be at the greatest risk for food insecurity, were not included in this study thus impacting the generalizability of our findings. Additionally, as a short-term pilot study, the modest amount of the economic intervention was likely insufficient to show measurable reductions in food insecurity. However, the flexibility of the funds was an important feature of this study given that different areas of social needs are correlated with one another.^16,17 Future directions from this work might involve comparing the flexibility of funds with more “traditional” methods to address food insecurity (eg, providing groceries and food items, or funds restricted to food items only). The qualitative findings suggested that participants benefited from the economic intervention in other ways that, although not directly related to food, alleviated some financial burden. We also note that the timing of this study coincided with economic inflation, and emergency increases to SNAP benefits due to the pandemic also came to an end. ⁴¹ These are additional factors that may impact the interpretation of the current study findings, in full recognition that addressing food insecurity remains a complex and structural challenge.

Most people who smoke, regardless of their socioeconomic circumstances, are interested in quitting and a substantial proportion attempt to quit each year. Yet, disparities in successful quitting has persisted, which suggests that novel approaches, such as the one examined in this feasibility study, are needed to reduce socioeconomic disparities in tobacco use and promote health equity. This will very likely involve concerted attention and further research to understand the broader social conditions and individual social needs that act as a barrier to smoking cessation to inform more robust structural interventions that more equitably address smoking-related health outcomes. In 2017 to 2019, U.S. health systems reported investing nearly $300 million in programs to address food insecurity, ⁴² recognizing that such investments are likely to have wide-ranging positive impacts in the long run. Therefore, this type of research and subsequent findings may also inform future cost-effectiveness research, in understanding how strategic investments may offer direct and indirect health benefits that can reduce health disparities.