Prevalence of Food Insecurity Among Cancer Survivors in the United States: A Scoping Review

Abstract

Background

Medical financial hardship is an increasingly common consequence of cancer treatment and can lead to food insecurity. However, food security status is not routinely assessed in the health care setting, and the prevalence of food insecurity among cancer survivors is unknown.

Objective

This scoping review aimed to identify the prevalence of food insecurity among cancer survivors in the United States before the COVID-19 pandemic.

Methods

Five databases (PubMed, Scopus, CINAHL [Cumulative Index to Nursing and Allied Health Literature], Web of Science, and ProQuest Dissertations and Theses) were systematically searched for articles that reported on food security status among US patients receiving active cancer treatment or longer-term cancer survivors and were published between January 2015 and December 2020.

Results

Among the 15 articles meeting the inclusion criteria, overall food insecurity prevalence ranged from 4.0% among women presenting to a gynecologic oncology clinic to 83.6% among patients at Federally Qualified Health Centers. Excluding studies focused specifically on Federally Qualified Health Center patients, prevalence of food insecurity ranged from 4.0% to 26.2%, which overlaps the food insecurity prevalence in the general US population during the same time period (range, 10.5% to 14.9%). Women were more likely than men to report being food insecure, and the prevalence of food insecurity was higher among Hispanic and Black patients compared with non-Hispanic White patients.

Conclusions

Given significant heterogeneity in study populations and sample sizes, it was not possible to estimate an overall food insecurity prevalence among cancer survivors in the United States. Routine surveillance of food security status and other social determinants of health is needed to better detect and address these issues.

IN THE UNITED STATES, FINANCIAL HARDSHIP IMPOSED by both direct health care costs and indirect costs, such as loss of productivity,¹ is an increasingly common consequence of cancer treatment,² one of the most expensive health conditions to treat.³ Reduction in spending for food is a common manifestation of medical financial hardship,⁴ due to the flexible nature of food expenditures. Thus, medical financial hardship can result in food insecurity, the economic or social situation in which individuals or households have limited or uncertain access to adequate food.⁵ However, because data on food security status is not collected routinely and reported in the health care setting, the true prevalence of food insecurity among individuals with a history of cancer in the United States is unknown.

Although no studies have evaluated the effects of food insecurity specifically on cancer treatment outcomes, food insecurity could lead to adverse cancer treatment outcomes through several different mechanisms. Food insecurity has been associated with lower intake of more perishable, nutrient-dense foods (eg, fruits, vegetables, and dairy products) and lower intake of key micronutrients (eg, vitamins A and B6, calcium, magnesium, and zinc).⁶ Numerous studies have reported that malnutrition (especially suboptimal intake of calories, protein, and micronutrients) is associated with poorer cancer treatment outcomes, including higher likelihood of treatment-related toxicity,^7,8 reduced treatment intensity or early treatment discontinuation,^9,10 hospitalizations,^7,8,10 and poorer overall survival^7,11,12 compared with patients who are well-nourished. Food insecurity has also been associated with suboptimal health care utilization. In studies of patients who were food insecure and undergoing cancer treatment, 55% reported not taking a prescribed medication,¹³ and 18% report delaying medical care¹⁴ due to financial hardship.

The US Department of Agriculture (USDA) Economic Research Service (ERS) monitors food security in the general US population using an annually administered 18-item survey to a representative sample of more than 34,000 US households, as part of a larger survey conducted by the US Census Bureau. The survey asks about experiences and behaviors related to household food security, such as being able to afford balanced meals, cutting the size of meals, or being hungry because of too little money for food.¹⁵ In 2019, before the COVID-19 pandemic, the USDA ERS found that an estimated 10.5% of the US population experienced food insecurity at some point during the year, which was the lowest level since the 2007 economic recession.¹⁵ Although the USDA stratifies their findings by various household characteristics, no data are available specifically for households dealing with medical issues, such as cancer.

This scoping review aimed to identify the prevalence of food insecurity among US residents with a history of cancer before the COVID-19 pandemic. The findings are intended as a benchmark for evaluating the effect of the pandemic on the prevalence of food insecurity in this population, and potential characteristics associated with variability in the prevalence of food insecurity, including cancer site and key demographic factors, such as sex and race and ethnicity.

METHODS

The protocol was developed using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews¹⁶ by 1 team member (K.R.), and then reviewed and revised by the entire research team. The final review protocol was made publicly available via Open Science Framework (https://osf.io/8vfgy).

Search Strategy

The following 5 databases were used for the initial identification of eligible articles: PubMed, Scopus, CINAHL (Cumulative Index to Nursing and Allied Health Literature), Web of Science, and ProQuest Dissertations and Theses. References lists were reviewed during full-text review to identify articles not identified through the database searches.

All database searches were conducted by 1 team member (E.S.) on November 16, 2020, and each was searched using every combination of the following terms. For the concept of food security: (Food insecur* OR Food secur* OR Food insufficien* OR Hunger OR Food access OR Food supply OR Food assistance). For the concept of cancer: Cancer OR Oncology OR Neoplasm were used. For the PubMed searches, Medical Subject Headings were also used, specifically Food supply OR Hunger Or Food assistance for the food security concept, and Neoplasms OR Neoplasms by site (explode) OR Medical oncology OR Radiation oncology OR Surgical oncology for the cancer concept. The searches were repeated in January 2022 to verify that no articles had been missed in the original search. No new articles meeting the original eligibility criteria were identified.

Eligibility Criteria

Included studies had been published (online or in print) between January 1, 2015 and the original search date of November 16, 2020; collected data before March 1, 2020 (to avoid confounding due to the COVID-19 pandemic); and collected data on food security status among US patients receiving active treatment for cancer or longer-term cancer survivors. Observational studies were the primary study design of interest, given the purpose of identifying prevalence of food insecurity among individuals with a history of cancer. Intervention studies were also included if food insecurity was measured as part of the study population’s baseline characteristics. Studies that reported food security status of patients treated outside of the United States (due to differences in health care costs, payment systems, and social determinants of food security in other countries) and/or studies of food insecurity in patients with illnesses other than cancer were excluded.

Review and Data Extraction

Two team members (K.R., M.C.) conducted the review of research articles and performed the data abstraction using the Covidence systematic review management system.¹⁷ Discrepancies were reconciled via discussion between the 2 reviewers/abstractors.

Data abstracted from each article included authors, publication year, dates of data collection, study design, study population (including age, sex, race and ethnicity, insurance type, and cancer site), sampling and recruitment strategies, sample size, food security measurement tool, food security reference period, and food insecurity prevalence.

Quality Assessment

The quality of each article was assessed using a scale developed previously by Nikolaus and colleagues^18,19 for a scoping review on the prevalence of food security among US college students. The quality rating system (Figure 1) evaluates 5 criteria, with each criterion weighted equally and evaluated on a scale of 0 to 2. Total possible scores ranged from 0 to 10, with a higher score indicating better quality.

Figure 1.

Quality assessment criteria and scoring for manuscripts and reports included in a scoping review of food insecurity among cancer survivors in the United States.

Data Synthesis

For each study, the prevalence of food insecurity is reported for the overall study population and, if possible, stratified by cancer site and key demographic factors, such as age, sex, and race and ethnicity. Considerable differences between studies related to patient populations, sample size, age groups, cancer sites, treatment phase (eg, active treatment vs post treatment), type of health care organization, and socio-demographic factors (eg, family income, history of food insecurity before cancer diagnosis, and immigration status) and reporting quality precluded the ability to calculate an overall weighted estimate of the prevalence of food insecurity.

Quality assessment scores were summarized as medians and ranges for all studies combined and then, when possible, stratified by treatment phase (active treatment vs post treatment), age group (pediatric, adult, and 65 years or older), and race and ethnicity.

RESULTS

Ultimately, 15 studies met the inclusion criteria (Figure 2). Individual study characteristics and findings are summarized in the Table. Two studies^21,22 were reported as conference abstracts and 1 study²³ was published in an oncology trade publication; the remaining studies (n = 12)^{24–30,31–34} were published in peer-reviewed scientific journals. Five studies that met the inclusion criteria used data from large national datasets^{27,29,30,32,33} maintained by the Centers for Disease Control and Prevention (National Health Information Survey [NHIS], Behavioral Risk Factor Surveillance Survey [BRFSS], and the National Health and Nutrition Examination Survey [NHANES]). Study sizes ranged from 32²¹ to 463,974³² participants.

Figure 2.

Flow diagram of the literature search and filtering results for a scoping review of food insecurity among cancer survivors in the United States. From Page and colleagues,³⁵ adapted with permission.

Table.

Basic characteristics and findings of studies examining the prevalence of food insecurity among US cancer survivors

Study, first author, year, location	Study population	Year of data collection	Food security assessment	Prevalence of food insecurity	Quality ratinga
Gany, 201524 Cancer Portal Project (10 hospital-based cancer clinics with large numbers of immigrant and low-income patients) New York, NY	1,390 participants (45% Hispanic, 41% non-Hispanic Black, 5% Asian, and 9% other race and ethnicity) Adults with immigrant or ethnic minority status All cancer sites Mean (SD) age: 55 (13.7) y	2008–2013	USDAb Household Food Security Survey Module15 - 18 items Reporting period: past year	809 (58.2%) reported being food insecure (n = 571 [41.1%] low food security, n = 238 [17.1%] very low food security) Food insecurity by sex: Female: n = 542/908 (59.7%) Male: n = 265/479 (55.3%) Food insecurity by cancer site: Breast cancer: n = 353/561 (62.9%) Prostate cancer: n = 101/186 (54.3%) Gastrointestinal cancers: n = 126/238 (52.9%) Head and neck cancers: n = 32/53 (60.4%) Food insecurity by race and ethnicity: Hispanic: n = 385/626 (61.5%) Black: n = 294/513 (57.3%) Asian or Pacific Islander: n = 39/74 (52.7%)	8/10 Unmet criteria: rigorous sampling and recruitment
Gany, 201525 Cancer Portal Project (10 hospital-based cancer clinics with large numbers of immigrant and low-income patients) New York, NY	201 participants Adults who qualified to participate in the food pantry program between 10/3/2011 and 1/1/2013 and completed the food security screening All cancer sites Mean age: 57 y	2011–2013	USDA Household Food Security Survey Module15 - 18 items Reporting period: past year	168 (83.6%) reported being food insecure (90 [44.8%] reported low food security, 78 [38.8%] reported very low food security).	5/10 Unmet criteria: rigorous sampling/recruitment Partially met criteria: specific research question; specific study population; sufficient sample reporting
Bona, 201626 Dana-Farber/Boston Children’s Cancer and Blood Disorders Center Boston, MA	99 families Cancer diagnosis within 30 d of study enrollment Childhood cancers Mean age of patient: 8.9 y (interquartile range, 2.9 to 15.0 y)	2011–2013	USDA Household Food Security Survey Module15 - 6 item Reporting period: past year	11 families (11%) were food insecure at baseline and 19 families (20%) were food secure at 6-mo follow-up	7/10 Unmet criteria: rigorous sampling/recruitment Partially met criteria: sufficient sample reporting
Garcia, 201827 National	14,879 participants (not all with a history of cancer) National Health Interview Survey respondents Any history of a cancer diagnosis (self-reported) Older adults: aged 50–85 y	2011–2014	USDA Household Food Security Survey Module15 - 10 item Reporting period: past 30 d	10.8% of the participants with a history of cancer reported being food insecure	7/10 Partially met criteria: specific research question; specific study population; sufficient sample reporting
Paden, 201928 Ohio State University (OSU) James Cancer Hospital, OSU Hospital East, Stefanie Spielman Comprehensive Breast Center Columbus, OH	288 participants (87.1% non-Hispanic White, 9.4% non-Hispanic Black, 3.5% other race and ethnicity) Any stage (I– IV) or variant of cancer receiving any treatment 18 y and older	2016–2017	USDA Household Food Security Survey Module15 - 10 item Reporting period: past year	27 (9.4%) reported marginal, low, or very low food security	5/10 Unmet criteria: rigorous sampling/recruitment, sufficient sample reporting Partially met criteria: specific research question
Poghosyan, 201929 National	6,481 participants (85.4% non-Hispanic White, 10.5% non-Hispanic Black, 1.4% Hispanic, 2.7% other race and ethnicity) Behavioral Risk Factor Surveillance System survey participants Adults with a history of cancer residing in 12 states that administered the Social Context Module in 2015 18 y and older	2015	Single question: “How often in the past 12 months would you say you were worried or stressed about having enough money to buy nutritious meals?” Reporting period: past year	26.2% (95% CI 24.2% to 28.2%, weighted) reported food insecurity Food insecurity by race and ethnicity: Non-Hispanic White: 24.0% (95% CI 21.9% to 26.1%) Non-Hispanic Black: 33.1% (95% CI 26.4% to 40.4%) Non-Hispanic other: 44.0% (95% CI 32.1% to 56.8%) Hispanic: 45.6% (95% CI 31.7% to 60.1%) Food insecurity by age group: 18–64 y: 36.8% (95% CI 33.4% to 40.3%) 65+ y: 15.7% (95% CI 13.8% to 17.7%)	9/10 Partially met criteria: sufficient sample reporting
Trego, 201930 National	1,022 participants (86.2% non-Hispanic White, 5.3% non-Hispanic Black, 5.3% Hispanic, 3.2% other race/ethnicity) National Health and Nutrition Examination Survey participants All cancer sites (including 30 adult survivors of childhood cancers) 20 y and older	2011–2014	USDA Household Food Security Survey Module15 - 10 item Reporting period: past year	132 (8.4%) cancer survivors reported being food insecure Food insecurity by sex: Female: n = 89 (10.7%) Male: n = 43 (5.3%) Food insecurity by race and ethnicity: Non-Hispanic White: 57 (6.25%) Non-Hispanic Black: 31 (20.6%) Mexican American: 18 (30.5%) Other Hispanic: 18 (25.5%) Other or multiracial: 8 (12.8%) Food insecurity by time from diagnosis: <1 y: 11 (8.6%) 1–3 y: 30 (8.6%) 4–8 y: 25 (7.2%) 9–15 y: 28 (8.2%) >15 y: 38 (9.2%)	7/10 Partially met criteria: specific research question; specific study population; sufficient sample reporting
Brown, 202021 MD Anderson Children’s Cancer Hospital Houston, TX	32 participants Caregivers for childhood cancer patients Age of patients/caregivers not specified	Not specified in this abstract	“validated measures of food insecurity” (no additional details provided) Reporting period: not specified	40.6% of households reported food insecurity	3/10 Unmet criteria: rigorous sampling/recruitment, reliable/valid measure of food security Partially met criteria: specific research question; specific study population; sufficient sample reporting
Rocha, 202022 Rady Children’s Hospital San Diego, CA	37 participants Caregivers for children with newly diagnosed cancer (past 6 mo) Age of patients/caregivers not specified	Not specified in this abstract	Food security assessment tool not specified in the abstract Reporting period: not specified	7 (18.9%) reported food insecurity	1/10 Unmet criteria: specific study population; rigorous sampling/recruitment; sufficient sample reporting; reliable/valid measure of food security Partially met criteria: specific research question
McDougall, 202031 State-wide survey New Mexico	394 participants (56% non-Hispanic White, 42% Hispanic, 8% other race and ethnicity) Stage I-III breast, colorectal, or prostate cancer identified through the New Mexico Cancer Registry Individuals who (at the time of their cancer diagnosis): spoke English or Spanish; had private insurance, Medicaid, or were uninsured; and their income was <$24,000 (approximately 200% of the Federal Poverty Level for an individual) Aged 21–64 y	2018–2019 (patients diagnosed with cancer between 2008 and 2016)	Hunger Vital Sign screener20 Reporting period: past year	229 (58%) were food secure in both the year before and the year after cancer diagnosis (persistently food secure) 38 (10%) were food secure in the year before and food insecure in the year after diagnosis (newly food insecure) 101 (26%) were food insecure at both times (persistently food insecure)	10/10
Berger, 202032 National	463,974 participants National Health Interview Survey respondents Adult patients with a history of throat/pharynx head and neck cancer, thyroid cancer, and colon cancer 18 y and older	2014–2018	USDA Household Food Security Survey Module15 - 10 item Reporting period: 30 d	Marginal, low, or very low food security by cancer site: Throat/pharynx: 17.7%, (95% CI 10.5% to 28.1%) Thyroid: 10.7% (95% CI 7.7% to 14.7%) Colon: 10.1% (95% CI 7.8% to 13.2%) Marginal, low, or very low food security by sex: Female: Throat/pharynx: 25.3% (95% CI 18.0% to 34.5%) Thyroid: 11.5% (95% CI 8.2% to 15.9%) Colon: 12.1% (95% CI 8.7% to 16.5%) Male Throat/pharynx: 14.9% (95% CI 7.1% to 28.8%) Thyroid: 8.3% (95% CI 5.2% to 13.0%) Colon: 7.8% (95% CI 5.4% to 11.1%) Black participants were significantly more likely than White participants to report being marginally food secure or food insecure (P = .008). Throat/pharynx: Other: 48.2% (95% CI 10.4% to 88.1%) Black: 23.4% (95% CI 7.3% to 54.2%) White: 15.2% (95% CI 9.7% to 23.3%) Thyroid: Other 19.6% (95% CI 7.4% to 42.6%) Black 23.2% (95% CI 13.4% to 37.1%) White 8.9% (95% CI 6.5% to 12.1%) Colon: Other: 11.2% (95% CI 4.1% to 27.0%) Black: 20.9% (95% CI 13.4% to 31.3%) White: 8.7% (95% CI 6.7% to 11.2%) No consistent patterns were observed related to Hispanic ethnicity (P = .522)	8/10 Unmet criteria: sufficient sample reporting
Zheng, 202033 National	12,141 cancer survivors, and 143,664 individuals without a cancer history National Health Interview Survey respondents All cancer sites, except nonmelanoma skin cancer or skin cancer of unknown type 18 y and older	2013–2017	USDA Household Food Security Survey Module15 — 10 item Reporting period: 30 d	Among individuals aged 18–39 y: cancer survivors were more likely to report “often true” (7.9% vs 4.6%; P = .004) about food running out, “often true” (7.6% vs 3.3%; P = .003) or “sometimes true” (14.2% vs 10.5%; P = .010) about food not lasting, and “often true” (6.3% vs 3.4%; P = .007) about inability to afford balanced meals compared with individuals without a cancer history Among individuals aged 40–64 y: compared with those without a cancer history, cancer survivors were more likely to report “often true” (4.3% vs 3.6%; P = .032) about inability to afford balanced meals compared with individuals without a cancer history Among individuals 65 y or older: Compared with individuals without a cancer history, cancer survivors reported similar patterns of food insecurity as individuals without a cancer history	9/10 Partially met criteria: sufficient sample reporting
Beavis, 202034 Gynecologic oncology clinic at The Johns Hopkins Hospital Baltimore, MD	752 women presenting to the gynecologic oncology clinic (60.2% White, 26.7% Black, 13.0% other race): 291 (38.7%) did not have cancer, 185 (24.6%) had uterine cancer, 173 (23.0%) had ovarian cancer, 61 (8.1%) had cervical cancer, and 42 (5.6%) had vulvar cancer	2017–2018	Health Leads Social Needs Screening Toolkit36 (single yes/no question) Reporting period: past year	Note: the prevalence of food security was not stratified by cancer diagnosis (yes/no) in this study. The prevalence of food insecurity reported below includes all study participants, both women with (61.3%) and without (38.7%) a cancer diagnosis. 30 women (4%) reported food insecurity, of whom 14 (7%) were Black women and 9 (2%) were White women Difference in food insecurity prevalence between Black and White women was statistically significant (P < .05)	5/10 Unmet criteria: rigorous sampling/recruitment; reliable/valid measure of food security Partially met criteria: specific research question; specific study population; sufficient sample reporting
Weisberg, 202023 New England Cancer Specialists Maine	All cancer sites Inclusion criteria, number of study participants and ages of participants not described	2017	Single question - patients considered food insecure if they agreed with one of the statements below: I was worried about running out of food and not being able to buy more I ran out of food and was not able to buy more Reporting period: past year	61% of male participants and 39% of female participants reported being food insecure Patients between the ages of 50 and 59 y had the highest percentage of food insecurity	3/10 Unmet criteria: rigorous sampling/recruitment Partially met criteria: specific research question; specific study population; sufficient sample reporting
Gany, 202137 Integrated Cancer Care Access Network (hospitals with large low-income immigrant and minority populations) New York, NY	All cancer sites Inclusion/exclusion criteria not specified, although it is implied that all patients enrolled in the Integrated Cancer Care Access Network for a cancer diagnosis (and up to 1 y post-treatment completion) were eligible to participate (n = 1,607) 681 individuals completed the assessment (43% non-Hispanic Black, 32% Hispanic, 5% non-Hispanic White, and 17% other race/ethnicity) Mean (SD) age: 56 (12.16) y	2012–2017	USDA Household Food Security Survey Module - 18 item15 Reporting period: past year	470 participants (69%) reported being food insecure Food insecurity by sex: Female: 318 (68%) Male: 149 (74%) Food insecurity by race and ethnicity: Non-Hispanic White: 22/30 (73%) Non-Hispanic Black: 201/307 (66%) Hispanic: 162/217 (75%) Other race: 83/118 (70%) By cancer site: Breast cancer: 206/297 (70%) Prostate cancer: 38/65 (59%) Lung cancer: 30/44 (68%) Colon cancer: 34/44 (77%) Lymphoma: 15/25 (60%) Other cancers: 145/197 (74%)	4/10 Unmet criteria: rigorous sampling/recruitment; sufficient sample reporting Partially met criteria: specific research question; specific study population

^aThe quality of each article was assessed using a scale developed for a previous scoping review on the prevalence of food security among US college students by Nikolaus and colleagues.^18,19 The quality rating system (Figure 1) evaluates 5 criteria, with each criterion weighted equally and evaluated on a scale of 0 to 2. Possible scores ranged from 0 to 10.

^bUSDA = US Department of Agriculture.

Measures of Food Insecurity

Most of the studies (9 of 15 [60.0%])^{24–28,30,32,33,37} used a version of the USDA Core Food Security modules (6-, 10-, or 18-question formats) to assess participants’ food security status.¹⁵ Three of these studies^27,32,33 used data from the NHIS, which has included the 10-item adult 30-day USDA food security module since 2011.³⁸ Of the 9 studies that used the USDA food security screening modules, all but 3 studies^28,32,33 defined food insecurity as any participant reporting “low” or “very low” food security as per the USDA ERS definition. Two reports^28,32 defined food insecurity among their study population as anyone reporting “marginal,” “low,” or “very low” food security. One report³³ described food insecurity by response to individual USDA Core Food Security module questions, and also created a unique summary measure that synthesized responses to various measures of financial worry along with food insecurity.

Of the 6 studies that did not use the USDA Core Food Security modules to assess food security, 1 study³¹ used the validated 2-question Hunger Vital Signs screener,²⁰ and another study³⁴ used the 1-question food security screener from the Health Leads Social Needs Screening Toolkit.³⁶ Poghosyan and Scarpino²⁹ used BRFSS data in the 12 states that administered the 2015 Social Context Module, which included a single question that asked “How often in the past 12 months would you say you were worried or stressed about having enough money to buy nutritious meals?” In 1 study,²³ the investigators developed a single screening question in which patients were considered food insecure if they agreed to either of the following statements: “I was worried about running out of food and not being able to buy more” or “I ran out of food and was not able to buy more.” Two studies,^21,22 specifically focused on pediatric patients and reported as conference abstracts, did not specify the tool used to assess food security status.

Study Quality Ratings

Quality ratings for the studies included in this scoping review varied widely, ranging from 1 to 10 with a median of 7. For studies that included adult participants only,^{24,25,27–34,37} the median quality score was also 7.0 (range, 4 to 10). For studies that included pediatric patients only,^21,22,26 (2 of which were conference abstracts), the mean quality score was 3 (range, 1 to 7). Among studies that included patients in active treatment only,^{21–26,28,31,34,37} the median quality rating was 5 (range, 1 to 10). Studies that included both patients in active treatment and those post treatment^{27,29,30,32,33} had a median quality score of 8 (range, 7 to 9). The most common reasons for not receiving higher-quality scores included use of convenience sampling as the participant recruitment strategy,^{21–26,29,34} no report of response rate or nonresponse bias evaluation,^{21–23,25,27,28,30,32,37} and use of food security measures that were not specified or had not been psychometrically tested.^21–23

Prevalence Estimates

Overall prevalence of food insecurity reported across studies ranged from 4.0% among all women presenting to a gynecologic oncology clinic in Baltimore, MD,³⁴ to 83.6% among patients at Federally Qualified Health Centers (FQHCs) in New York, NY, who qualified to participate in a hospital-based food pantry program.²⁵ When restricted to studies that were not focused specifically on patients being treated at FQHCs, prevalence of food insecurity ranged from 4.0%³⁴ to 26.2%.²⁹ Figure 3 depicts the prevalence of food insecurity across the included studies with USDA data for the general US population by time of data collection for comparison.

Figure 3.

Timing of data collection and prevalence of food insecurity among cancer survivors in the United States in comparison with US Department of Agriculture (USDA) data on prevalence of food insecurity.¹⁵. ^aIndicates that prevalence estimates include both marginal food security and low/very low food security.

Nine^{21–26,28,34,37} of the 15 studies reported a prevalence of food insecurity among patients actively undergoing cancer treatment. Three of these studies^21,22,26 focused specifically on the prevalence of household food insecurity among pediatric cancer patient populations, which ranged from 11.0%²⁶ to 40.6%.²¹ Because most of the included studies and reports focused on adult cancer patient populations, the overall prevalence of food insecurity across all articles is the same range as the range when restricted solely to studies among adults during active treatment (4.0%³⁴ to 83.6%²⁵).

McDougall and colleagues³¹ surveyed 394 individuals with a history of early-stage (stages I to III) breast, colorectal, or prostate cancer identified through the New Mexico Cancer Registry, and asked them to report on food security status for both the year before and the year after their cancer diagnosis. The survey revealed that 229 (58%) were food secure both before and after their cancer diagnosis (persistently food secure), 38 (10%) were food secure the year before the cancer diagnosis but became food insecure in the year after the cancer diagnosis (newly food insecure), and 101 (26%) were food insecure before and after the cancer diagnosis (persistently food insecure). Bona and colleagues²⁶ also found that, among families of pediatric cancer patients, the prevalence of food insecurity increased from 11% of study participants at the time of the cancer diagnosis to 20% at the 6-month post-diagnosis follow-up.

The NHIS, BRFSS, and NHANES ask participants to self-report whether they have ever been diagnosed with a chronic condition such as cancer, and thus likely include both individuals currently receiving cancer treatment and long-term cancer survivors. The reported prevalence of food insecurity among surveillance study participants with a history of cancer ranged from 8.4% among NHANES 2011–2014 participants³⁰ to 17.7% among NHIS respondents with a history of esophageal or pharyngeal cancers.³²

Food Insecurity Estimates by Age, Sex, and Race

Zheng and colleagues³³ compared food insecurity experiences among 3 age groups: 18 to 39 years, 40 to 64 years, and 65 years and older. They found that, among cancer survivors, individuals in the 2 younger age groups were more likely to experience food insecurity than individuals 65 years and older. Similarly, among 2015 BRFSS participants, Poghosyan and Scarpino²⁹ found that the prevalence of food insecurity was significantly higher among 18- to 64 year-old cancer survivors (36.8%; 95% CI 33.4% to 40.3%) compared with cancer survivors 65 years and older (15.7%; 95% CI 13.8% to 17.7%). The study by McDougall and colleagues³¹ focused specifically on individuals diagnosed between 21 and 64 years of age because prior studies had identified inverse associations between age and financial hardship. Although the study from McDougall and colleagues did not include a comparison group of individuals diagnosed at or after age 65 years, 35% of the study population reported either new or persistent food insecurity after their cancer diagnosis, which is consistent with the findings reported by Poghosyan and Scarpino for this age group.

Of the 3 studies evaluating the prevalence of food insecurity among pediatric patient populations, 2 were conference abstracts, and were only able to provide limited details.^21,22 In contrast, Bona and colleagues²⁶ conducted a longitudinal, single-center study of 99 families with children being treated for a cancer diagnosis, and found that 11 families (11%) reported being food insecure at the time of diagnosis, which increased to 19 families (20%) reporting being food insecure at the 6-month follow-up.

Several studies evaluated the prevalence of food insecurity by sex^{23,24,30,32,37} with inconsistent findings. Of these, 3 studies with higher-quality ratings and larger sample sizes^24,30,32 reported that female cancer survivors were more likely to report being food insecure than male cancer survivors.

Of 6 studies that evaluated the prevalence of food insecurity by race and ethnicity,^{24,29,30,32,34,37} a more consistent picture emerged. Most studies reported that food insecurity was highest among Hispanic patients, followed by non-Hispanic Black patients, with several studies noting a statistically significant higher prevalence among non-Hispanic Black patients than White patients.^32,34 Three^29,30,32 of the 6 studies presented the prevalence of food insecurity by race and ethnicity using data from the large, nationally representative data sets (NHIS, BRFSS, and NHANES). These studies reported that the prevalence of food insecurity among Hispanic patients ranged from 25.5%³⁰ to 45.6%,²⁹ non-Hispanic Black patients ranged from 20.6%³⁰ to 33.1%,²⁹ compared with non-Hispanic White patients ranged from 6.2%³⁰ to 24.0%.²⁹

Only three studies^24,32,37 included evaluation of food insecurity prevalence by cancer site. An analysis focusing on cancers of the throat or pharynx, thyroid, and colon³² revealed that individuals with throat or pharynx cancers had significantly higher prevalence of food insecurity (17.7%; 95% CI 10.5% to 28.1%) than those with thyroid (10.7%; 95% CI 7.7% to 14.7%) or colon (10.1%; 95% CI 7.8% to 13.2%) cancers. Two studies^24,37 indicated a high prevalence (>50% of participants) of food insecurity among low-income patients receiving treatment at FQHCs in New York City, regardless of cancer site.

DISCUSSION

Given the significant heterogeneity in research study populations and sample sizes, it was not possible to estimate an overall prevalence of food insecurity among cancer survivors in the United States. Not surprisingly, the reported prevalence of food insecurity was highest among low-income individuals receiving care at FQHCs, and reports using data from the large, population-based health surveillance surveys indicated that the prevalence of food insecurity among cancer survivors is more consistent with that reported by the USDA ERS for the general US population (which declined from a high of 14.9% in 2011 to 10.5% in 2019¹⁵). Although strengths of the population-based health survey data include sampling strategies that are representative of the US population and the availability of comparison data from a non-cancer control group, data on time since diagnosis are not available. Thus, the “cancer survivors” in these large datasets include both individuals currently receiving cancer treatment and longer-term cancer survivors—2 distinct periods of cancer survivorship with potentially different financial issues and risk of food insecurity. Studies show that cancer-related costs are generally lower among long-term survivors compared with survivors diagnosed more recently.¹ Thus, if a greater proportion of the cancer survivor participants in these population-based health surveys were long-term survivors, the prevalence of food insecurity would be expected to be closer to that of the non-cancer population mean, as these studies report. However, longer-term cancer survivors, especially younger cancer survivors,^39–41 may still experience other factors that put them at higher risk of medical financial hardship and food insecurity, such as the inability to work, expense of ongoing medications, and treatment costs for late effects of cancer treatment and treatment-related comorbidities.^42,43

This scoping review identified several limitations of the study designs that have been used to try to estimate the prevalence of food insecurity among cancer survivors. Single-center studies typically included select patient groups only (eg, low-income, pediatric, or certain cancer sites) and many lacked non-cancer control groups. Studies using national surveillance data are cross-sectional, so it is unclear whether participants were food insecure before the cancer diagnosis, or whether they became food insecure as a result of out-of-pocket costs of cancer treatment. There were also few data available related to the prevalence of food insecurity by specific cancer site, which is important to understand because site-specific differences in treatment regimens, duration, costs, and toxicity can all contribute to differences in risk of financial hardship and food insecurity. For example, certain cancers, such as oral, esophageal, and gastrointestinal cancers, can lead to the inability to consume regular foods by mouth, and use of alternative feeding routes, such as enteral or parenteral nutrition, can add to overall costs of cancer treatment. Although some of the studies and reports included in this scoping review stratified the prevalence of food insecurity by cancer site, none of the studies considered the treatment type received, such as surgery, radiation, or pharmacologic therapy (including chemotherapy, targeted therapy, hormone therapy, or immunotherapy).

The limitations of the currently available data on the prevalence of food insecurity among cancer survivors underscores the need for routine, standardized, and comprehensive surveillance and data reporting on food insecurity among cancer survivors. In 2017, the American Association for Cancer Research, American Cancer Society, American Society of Clinical Oncology, and the National Cancer Institute issued a position statement calling for routine collection of data on the social determinants of health (SDOH), including food security status, followed by “meaningful action and continued monitoring to ensure that the action was successful.”⁴⁴

Yet, despite these calls for action, SDOH data are not yet collected routinely in most health care systems. Recent surveys of oncology health care organizations (including National Cancer Institute–designated Comprehensive Cancer Centers,⁴⁵ National Comprehensive Cancer Network member institutions,⁴⁶ and the National Cancer Institute Community Oncology Research Program⁴⁷) indicate that although many centers are beginning to screen patients for financial hardship, the screening tools used vary widely, and often do not include questions about food security status specifically. Studies evaluating the barriers to implementing SDOH and food security status screening indicate that the time required to carry out screening and the absence of provider training on how to respond to identified social needs are key concerns.⁴⁸ Health care organizations that do screen for food insecurity often use the validated, 2-question Hunger Vital Sign screener,²⁰ which has been shown to have good sensitivity and specificity for identifying food insecurity compared with the USDA 18-item Household Food Security Survey.¹⁵ In addition, although not specific to oncology care, several of the widely used SDOH screeners, such as the Protocol for Responding to and Assessing Patient Assets, Risks and Experiences screener⁴⁹ and the Health Leads Social Need Screening Toolkit,³⁶ include a single question to assess food security status.

In an effort to improve documentation of SDOH-related activities in clinical settings, The Gravity Project (https://thegravityproject.net/), a national collaborative that develops consensus-based data standards for SDOH data, identified codes within commonly used medical coding systems (SNOMED CT [Systemized Nomenclature of Medicine Clinical Terms], LOINC [Logical Observation Identifiers Names and Codes], and ICD-10-CM [International Classification of Diseases, 10^th Revision, Clinical Modification) that can be used to document SDOH (including food security status) assessment, diagnosis, treatment, and intervention.⁵⁰ These codes were then submitted to the Office of the National Coordinator for Health Information Technology, which included the codes in the July 2021 release of the United States Core Data for Interoperability, version 2⁵¹ data standards. Ultimately, requirements by accrediting bodies, such as the Commission on Cancer, might be needed to ensure that standardized and comprehensive surveillance of financial hardship and SDOH (including food security status) becomes routine in oncology settings.

Routine screening for food insecurity in the clinical setting would allow practitioners to identify and refer patients with food insecurity to supportive services,⁵² potentially improving cancer treatment outcomes. Health care organizations across the country are exploring various interventions to address food insecurity, including passive referrals or active navigation to community organizations or government agencies that provide food and nutrition assistance, food vouchers for local grocery stores or farmer’s markets,^53,54 medically tailored meal programs,^55,56 and food provision through hospital- and clinic-based food pantries.^25,57–59 In addition, some cancer centers offer financial navigation and counseling,⁴⁷ including guidance in applying to nutrition-assistance programs or identifying other sources of free or low-cost foods. At the community level, aggregated data on food security status would support the Community Health Needs Assessment process required of tax-exempt hospitals under the Affordable Care Act,⁶⁰ allowing available resources to be better targeted toward at-risk populations.⁶¹ On a broader level, routine data collection on food security status and other SDOH would allow researchers to better evaluate the impact of medical financial toxicity on cancer treatment outcomes and to identify populations at higher risk in order to develop strategies to mitigate adverse economic and health outcomes. The data will also support health disparities research and policy making aimed at addressing SDOH issues, including food insecurity

Previous studies have shown that a patient’s financial and food security status can change over the course of cancer treatment and survivorship.^1,2 Therefore, screening for financial hardship and food insecurity should occur throughout the disease trajectory, not just at diagnosis or among high-risk patient populations. Although low-income, minority, and underinsured/underserved patients may be at higher risk, patients with private health insurance may also experience financial hardship, especially those with high-deductible health plans.⁶² Changes in employment, insurance coverage, caregiver support, and disease status, as well as the intensity and duration of treatment, may contribute to changes in financial and food security statuses. Future research is also needed to better understand the impact of a cancer diagnosis on household food insecurity, including factors influencing decisions around food allocation among household members and the use of nutrition-assistance programs or other sources of free or low-cost food.

Strengths and Limitations of This Scoping Review

Strengths of this scoping review include use of an a priori review protocol prepared by a research team with expertise in knowledge synthesis through systematic and scoping review. The search strategy included 5 large electronic bibliographic databases, as well as the review of reference lists for eligible articles and reports. Each article or report was reviewed by 2 independent reviewers who met to review and resolve conflicts. Use of systematic review software (Covidence)¹⁷ ensured that all articles and reports were tracked properly and accounted for through the entire review process. Limitations of the review include the possible failure to identify all potentially eligible articles and reports, despite attempts to be as comprehensive as possible.

CONCLUSIONS

This scoping review found that the ability to estimate the prevalence of food insecurity among US cancer survivors is extremely limited at this time due to significant heterogeneity in study populations and sample sizes. Routine, standardized, and comprehensive data reporting and surveillance of food security status among cancer survivors is needed as part of the broader effort to support future research and address SDOH.

RESEARCH SNAPSHOT.

Research Question:

What was the prevalence of food insecurity among cancer survivors in the United States before the COVID-19 pandemic?

Key Findings:

Across the 15 studies included in this scoping review, prevalence of food insecurity ranged from 4% among women presenting to a gynecologic oncology clinic in Baltimore, MD, to 83.6% among patients at Federally Qualified Health Centers in New York, NY. Given the significant heterogeneity in study populations and sample sizes, it was not possible to estimate an overall prevalence of food insecurity. Routine surveillance of food security status among all patients being treated for cancer is needed.