Advertisement
The American Journal of Medicine
Advanced searchSave search
Clinical research study| Volume 127, ISSUE 4, P303-310.e3, April 2014

Download started.

Ok

Treat or Eat: Food Insecurity, Cost-related Medication Underuse, and Unmet Needs

Published:January 21, 2014DOI:https://doi.org/10.1016/j.amjmed.2014.01.002
Treat or Eat: Food Insecurity, Cost-related Medication Underuse, and Unmet Needs
Previous ArticleMortality Rate in Type 2 Myocardial Infarction: Observations from an Unselected Hospital Cohort
Next ArticleAlcohol Quantity and Type on Risk of Recurrent Gout Attacks: An Internet-based Case-crossover Study

      Abstract

      Background

      Adults with chronic disease are often unable to meet medication and food needs, but no study has examined the relationship between cost-related medication underuse and food insecurity in a nationally representative sample. We examined which groups most commonly face unmet food and medication needs.

      Methods

      Cross-sectional analysis of data from chronically ill participants (self-report of arthritis, diabetes mellitus, cancer, asthma, chronic obstructive pulmonary disease, stroke, hypertension, coronary heart disease, or presence of a “psychiatric problem”) aged ≥20 years, in the 2011 National Health Interview Survey. We fit logistic regression models to identify factors associated with food insecurity, cost-related medication underuse, or both.

      Results

      There were 9696 adult National Health Interview Survey (NHIS) participants who reported chronic illness; 23.4% reported cost-related medication underuse; 18.8% reported food insecurity; and 11% reported both. Adults who reported food insecurity were significantly more likely to report cost-related medication underuse (adjusted odds ratio [aOR] 4.03). Participants with both cost-related medication underuse and food insecurity were more likely to be Hispanic (aOR 1.58), non-Hispanic black (aOR 1.58), and have more chronic conditions (aOR per additional chronic condition 1.56) than patients reporting neither. They also were less likely to have public, non-Medicare insurance (aOR 0.70) and report participation in the Special Supplemental Nutrition Assistance Program for Woman, Infants, and Children (aOR 0.39).

      Conclusions

      Approximately 1 in 3 chronically ill NHIS participants are unable to afford food, medications, or both. WIC and public health insurance participation are associated with less food insecurity and cost-related medication underuse.

      Keywords

      Clinical Significance
      • •
        Cost-related medication underuse and food insecurity are common among chronically ill adults, with one third of patients reporting one or both.
      • •
        Food insecurity is strongly associated with cost-related medication underuse.
      • •
        Racial/ethnic minorities and adults with more comorbidities are more likely to report food insecurity and cost-related medication underuse. Participation in programs like Medicaid and the Special Supplemental Nutrition Program for Women, Infants, and Children is associated with less cost-related medication underuse and food insecurity.
      The current economic climate has led to substantial hardship for many Americans. In 2012, 1 in 6 Americans reported that they lack health insurance
      • Ward B.
      • Schiller J.
      • Freeman G.
      • Peregoy J.
      Early Release of Selected Estimates Based on Data from the January–September 2012 National Health Interview Survey.
      National Center for Health Statistics, Hyattsville, MDMarch 2013
      and nearly 1 in 5 said they had trouble meeting basic needs.
      • Gallup Healthways
      Gallup-Healthways Monthly U.S. Well-Being Report - April 2012.
      Healthways, Franklin, TN2012
      As a result, approximately 15% of chronically ill Americans reported cost-related medication underuse, which refers to taking less medication than prescribed, or not taking it at all, due to cost.
      • Piette J.D.
      • Heisler M.
      • Wagner T.H.
      Cost-related medication underuse among chronically ill adults: the treatments people forgo, how often, and who is at risk.
      Am J Public Health. 2004; 94: 1782-1787
      , 
      • Madden J.M.
      • Graves A.J.
      • Zhang F.
      • et al.
      Cost-related medication nonadherence and spending on basic needs following implementation of Medicare Part D.
      JAMA. 2008; 299: 1922-1928
      An equal proportion of American households
      • Coleman-Jensen A.
      • Nord M.
      • Andrews M.
      • Carlson S.
      Statistical Supplement to Household Food Security in the United States in 2011.
      U.S. Department of Agriculture Economic Research Service, Washington, DC2012
      experience food insecurity, which is defined as “limited or uncertain availability of nutritionally adequate and safe foods, or limited or uncertain ability to acquire acceptable foods in socially acceptable ways.”
      • Bickel G.
      • Nord M.
      • Price C.
      • et al.
      Guide to Measuring Household Food Security, Revised 2000.
      U.S. Department of Agriculture Food and Nutrition Service, Alexandria, VA2000
      While prescription drug benefit programs and food security assistance programs, such as the Supplemental Nutrition Assistance Program (SNAP, formerly the Food Stamp Program) do exist, it is not clear if they are sufficient to mitigate the effects of food insecurity on medication underuse or the effects of out-of-pocket medication payments on food insecurity.
      Both cost-related medication underuse and food insecurity are associated with adverse health consequences,
      • Pan L.
      • Sherry B.
      • Njai R.
      • Blanck H.M.
      Food insecurity is associated with obesity among US adults in 12 states.
      J Acad Nutr Diet. 2012; 112: 1403-1409
      , 
      • Parker E.D.
      • Widome R.
      • Nettleton J.A.
      • Pereira M.A.
      Food security and metabolic syndrome in U.S. adults and adolescents: findings from the National Health and Nutrition Examination Survey, 1999-2006.
      Ann Epidemiol. 2010; 20: 364-370
      , 
      • Seligman H.K.
      • Laraia B.A.
      • Kushel M.B.
      Food insecurity is associated with chronic disease among low-income NHANES participants.
      J Nutr. 2010; 140: 304-310
      , 
      • Seligman H.K.
      • Schillinger D.
      Hunger and socioeconomic disparities in chronic disease.
      N Engl J Med. 2010; 363: 6-9
      but no prior study has examined the relationship between them in a nationally representative sample. We used the 2011 National Health Interview Survey to understand how common these 2 aspects of economic insecurity are among individuals with chronic disease and, to inform policy-making, whether specific groups of patients, when both needs cannot be met, are more likely to report food insecurity, or vice versa.

      Methods

      Data Source and Study Sample

      We analyzed data from the 2011 National Health Interview Survey (NHIS).

      National Center for Health Statistics, Centers for Disease Control and Prevention. About the National Health Interview Survey. Available at: http://www.cdc.gov/nchs/nhis/about_nhis.htm. Accessed February 10, 2013.

      The NHIS is conducted annually by trained in-person interviewers in English, Spanish, or with an interpreter. Full details of NHIS methodology have been described previously.
      • Division of Health Interview Statistics
      • National Center for Health Statistics
      NHIS Survey Description.
      Centers for Disease Control, Atlanta, GA2012
      To focus on participants with chronic illness, who may be forced to choose month after month between medications and food and for whom such tradeoffs may have important clinical consequences, we included adult participants (age ≥20 years) in the 2011 NHIS who report at least 1 of 8 common chronic conditions surveilled by the NHIS: arthritis, diabetes mellitus, cancer, asthma, chronic obstructive pulmonary disease (COPD), stroke, hypertension, or coronary heart disease; or presence of a “psychiatric problem.” For subgroup analyses, we created the following groups: cardiovascular disease or hypertension, defined as participants reporting hypertension, coronary heart disease (including myocardial infarction), or stroke; diabetes mellitus; noncardiovascular disease (asthma, COPD, arthritis, or cancer); and presence of a “psychiatric problem.” We also conducted sensitivity analyses using all adult participants, regardless of whether they reported a chronic illness.
      The Partners HealthCare Human Research Committee exempted this study from review.

      Food Insecurity Assessment

      Food insecurity was assessed using the 10 adult-referenced items of the US Department of Agriculture's household-level 30-day food security scale. These items are used to ascertain national rates of food insecurity.
      • Bickel G.
      • Nord M.
      • Price C.
      • et al.
      Guide to Measuring Household Food Security, Revised 2000.
      U.S. Department of Agriculture Food and Nutrition Service, Alexandria, VA2000
      We dichotomized responses into Food Secure (affirmative response to 0-2 items) and Food Insecure (affirmative response to 3-10 items) according to standard protocol.
      • Bickel G.
      • Nord M.
      • Price C.
      • et al.
      Guide to Measuring Household Food Security, Revised 2000.
      U.S. Department of Agriculture Food and Nutrition Service, Alexandria, VA2000

      Cost-related Medication Underuse

      Cost-related medication underuse was determined by response to 4 validated items adapted from the Medication Expenditure Panel Survey.

      Agency for Healthcare Research and Quality. Medical Expenditures Panel Survey (MEPS); 2004. Available at: http://meps.ahrq.gov/mepsweb/survey_comp/survey.jsp. Accessed February 11, 2013.

      Similar to prior studies,
      • Soumerai S.B.
      • Pierre-Jacques M.
      • Zhang F.
      • et al.
      Cost-related medication nonadherence among elderly and disabled medicare beneficiaries: a national survey 1 year before the medicare drug benefit.
      Arch Intern Med. 2006; 166: 1829-1835
      a respondent was considered to have cost-related medication underuse if he or she responded “Yes” to any of 4 items that asked if, in the last 12 months: 1) there was a prescription a participant did not fill due to cost; 2) the participant skipped medication doses to save money; 3) the participant took less medication than prescribed to save money; or 4) the participant delayed filling a prescription to save money.

      Demographic and Socioeconomic Variables

      We adjusted for several demographic and socioeconomic variables that may be associated with food insecurity or cost-related medication underuse.
      • Madden J.M.
      • Graves A.J.
      • Zhang F.
      • et al.
      Cost-related medication nonadherence and spending on basic needs following implementation of Medicare Part D.
      JAMA. 2008; 299: 1922-1928
      , 
      • Seligman H.K.
      • Laraia B.A.
      • Kushel M.B.
      Food insecurity is associated with chronic disease among low-income NHANES participants.
      J Nutr. 2010; 140: 304-310
      Demographic factors included age (in years), sex, self-report race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, or other/mixed race/ethnicity), and nativity status (born in the US or not). We categorized health insurance status with regard to both source and presence of a prescription drug benefit. Because almost 90% of Medicare beneficiaries have creditable prescription drug coverage,

      Kaiser Family Foundation. Number of Medicare Beneficiaries with Creditable Prescription Drug Coverage, by Type. Available at: http://kff.org/medicare/state-indicator/medicare-rx-drug-coverage/. Accessed May 8, 2013.

      and prescription drug benefits are offered by Medicaid in all 50 states,

      Kaiser Family Foundation. Medicaid Benefits: Prescription Drugs. Available from: http://kff.org/medicaid/state-indicator/prescription-drugs/#. Accessed May 8, 2013.

      we categorized patients into 5 mutually exclusive insurance groups: none, private with drug benefit, private without drug benefit, Medicare, and other non-Medicare public insurance (including Medicaid, Medicare/Medicaid dual eligible respondents, and Veterans Administration coverage). For socioeconomic factors, we categorized educational attainment as less than high school diploma, high school diploma, or some college or higher. We classified income as a percentage of the federal poverty level (FPL), which accounts for household size, in categories of ≤100%, 101-200% FPL, and ≥200% FPL. We also ascertained self-reported household participation in the 2 main federal nutrition assistance programs, SNAP (formerly the Food Stamp Program) and the Special Supplemental Nutrition Assistance Program for Woman, Infants, and Children (WIC).

      Clinical and Other Variables

      Because the number of chronic medical conditions that a patient has is associated with a higher likelihood of cost-related medication underuse,
      • Madden J.M.
      • Graves A.J.
      • Zhang F.
      • et al.
      Cost-related medication nonadherence and spending on basic needs following implementation of Medicare Part D.
      JAMA. 2008; 299: 1922-1928
      we constructed an index of the number of chronic medical conditions, similar to the method of Madden et al.
      • Madden J.M.
      • Graves A.J.
      • Zhang F.
      • et al.
      Cost-related medication nonadherence and spending on basic needs following implementation of Medicare Part D.
      JAMA. 2008; 299: 1922-1928
      As above, these conditions included arthritis, diabetes mellitus, cancer, asthma, COPD, stroke, hypertension, coronary heart disease, or presence of a “psychiatric problem.” We also adjusted for tobacco use because the cost associated with this may lead to food insecurity or cost-related medication underuse, categorized as current, former, and never users based on self-report.

      Statistical Analysis

      Unadjusted comparisons were made using chi-squared tests for categorical variables and t-tests for continuous measures. We then used multivariable logistic regression to test our hypothesis that food insecurity was independently associated with cost-related medication underuse, adjusting for the demographic, social, and clinical variables described above. In order to see if these results, obtained in a chronically ill population, were generalizable to the overall American population, we then fit a model with the same variables, but in all adult NHIS participants who were asked food security and cost-related medication underuse questions.
      We also hypothesized that participation in insurance programs with lower prescription medication co-pays, such as Medicaid, or programs that reduced out-of-pocket spending for food, such as WIC, might be associated with lower food insecurity and lower cost-related medication underuse. In order to test these hypotheses, we fit 3 multivariable logistic regression models comparing those with food insecurity but not cost-related medication underuse, those reporting cost-related medication underuse but no food insecurity, and those reporting both, with those who report neither food insecurity nor cost-related medication underuse.
      Finally, in order to understand what factors were associated with unmet food and medication needs in situations where both cannot be fulfilled, we fit multivariable logistic regression models comparing those who were food insecure but did not report cost-related medication underuse to those who reported cost-related medication underuse but were not food insecure. This represents a group that may be subject to a “treat or eat” trade-off between prescription medication and household food. We also conducted this analysis among participants with incomes <200% of the FPL in order to capture a more homogenous population with few resources—those for whom a “treat or eat” trade-off would be most likely.
      Analyses were conducted using SAS version 9.3 (SAS Institute Inc., Cary, NC) survey procedures with domain analyses to account for the complex multistage sampling design and provide appropriate weights, as recommended in the 2011 NHIS data release documentation.
      National Center for Health Statistics
      2011 National Health Interview Survey (NHIS) Public Use Data Release.
      U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics, Hyattsville, MD2012

      Results

      Our sample comprised 9696 adults who reported chronic illness. Overall, 23.4% of respondents, representing approximately 7.2 million Americans, reported cost-related medication underuse in the prior 12 months. In addition, 18.8% of respondents, representing 5.8 million American households, reported food insecurity. More than 11% of respondents reported both food insecurity and cost-related medication underuse. The baseline characteristics of our sample, stratified by the presence of cost-related medication underuse, food insecurity, both, or neither, are presented in Table 1. Those with no cost-related medication underuse or food insecurity were older, more likely to be white, and had higher incomes than those in the other groups. Poverty was more common among those participants reporting food insecurity. Lack of insurance was more prevalent in groups with cost-related medication underuse.
      Table 1Demographics
      Overall

      n = 9696

      % or Mean (SE)      		Neither Food Insecure nor CRMU

      n = 6441

      % or Mean (SE)      		Food Secure

      but CRMU

      % or mean (SE)

      n = 1258      		Food Insecure

      but no CRMU

      % or mean (SE)

      n = 908      		Both Food

      Insecure and CRMU

      n = 1082

      % or Mean (SE)      		P-Value
      Age (y)60.9 (0.2)64.8 (0.2)53.7 (0.4)54.2 (0.6)49.7 (0.4)<.0001
      Male38.840.036.338.134.6.003
      Race/ethnicity<.0001
       Hispanic8.06.29.814.611.9
       Non-Hispanic white76.480.973.759.564.0
       Non-Hispanic black12.59.813.922.220.8
       Asian2.02.31.51.81.4
       Other1.00.81.11.91.8
      No usual place of care5.83.39.97.715.8<.0001
      Insurance<.0001
       None10.34.323.410.832.9
       Private, w/Rx coverage27.228.533.116.619.5
       Private, no Rx coverage1.31.32.30.60.5
       Medicare38.946.126.121.921.1
       Public, non-Medicare22.319.715.050.126.0
      Income<.0001
       <100% FPL19.513.317.346.240.1
       100-200% FPL22.718.926.130.836.1
       >200% FPL57.767.856.623.023.8
      Education<.0001
       <HS diploma20.418.217.934.826.4
       HS diploma29.830.128.527.431.8
       >HS diploma49.851.753.637.841.9
      Born outside US9.69.18.913.310.9.0004
      SNAP beneficiary17.710.119.446.341.8<.0001
      WIC beneficiary2.61.54.55.15.7<.0001
      Smoking status
       Current21.515.029.634.543.7<.0001
       Former32.6536.428.524.120.3
       Never45.948.741.941.436.0
      Number of chronic conditions2.2 (0.01)2.1 (0.02)2.2 (0.04)2.3 (0.05)2.4 (0.05)<.0001
      CRMU = cost-related medication underuse; FPL= federal poverty level; SNAP = Supplemental Nutrition Assistance Program; WIC = Special Supplemental Nutrition Program for Women, Infants, and Children.

      Relationship between Cost-related Underuse and Food Insecurity

      Cost-related medication underuse was more common in respondents with food insecurity compared with those without (55.5% vs 16.0%, P <.001). After adjustment for demographic factors, insurance, and clinical comorbidities, food insecurity remained strongly associated with cost-related medication underuse (odds ratio [OR] 4.03; 95% confidence interval [CI], 3.44-4.73) (Table 2). The odds ratio was similar in a model including all adult NHIS participants, not only those with chronic disease (eAppendix Table 1). The observed association also was similar when looking at the individual subtypes of cost-related medication underuse (adjusted for the same covariates as above): unable to afford a prescription (OR 4.15; 95% CI, 3.52-4.89), delaying a prescription due to cost (OR 4.01; 95% CI, 3.38-4.76), skipping doses due to cost (OR 3.43; 95% CI, 2.90-4.06), and taking less medication than prescribed due to cost (OR 3.64; 95% CI, 3.06-4.33). Finally, the results were similar when looking at 4 subgroups of chronically ill patients: those with cardiovascular disease or hypertension, those with diabetes, those with noncardiovascular disease (asthma, COPD, arthritis, or cancer), and those who reported history of a “psychiatric problem” (eAppendix Table 2).
      Table 2Factors Associated with Cost-related Medication Underuse in All Chronically Ill Participants (n = 9696)
      aOR
      Adjusted odds ratio (aOR) represents multivariable logistic regression results with adjustment for all variables in table.
      (95% CI)
      Food insecure4.03 (3.44-4.73)
      P < .05.
      Age (y)0.96 (0.95-0.97)
      P < .05.
      Male0.76 (0.67-0.86)
      P < .05.
      Race/Ethnicity
       Hispanic1.22 (0.96-1.54)
       Non-Hispanic white (ref)1.0 (–)
       Non-Hispanic black1.15 (0.98-1.35)
       Asian0.90 (0.58-1.39)
       Other0.92 (0.51-1.67)
      No usual place of care1.20 (0.90-1.61)
      Insurance
       None3.38 (2.66-4.29)
      P < .05.
       Private, w/ Rx coverage (ref)1.0 (–)
       Private, no Rx coverage1.49 (0.86-2.59)
       Medicare0.97 (0.80-1.19)
       Public, non-Medicare0.47 (0.37-0.59)
      P < .05.
      Income
       <100% FPL1.08 (0.87-1.35)
       100-200% FPL1.67 (1.40-1.99)
       >200% FPL (ref)1.0 (–)
      Education
       <HS Diploma0.86 (0.72-1.03)
       HS Diploma0.94 (0.81-1.09)
       >HS Diploma (ref)1.0 (–)
      Born outside US0.89 (0.70-1.14)
      SNAP beneficiary1.12 (0.90-1.39)
      WIC beneficiary0.91 (0.65-1.29)
      Smoking status
       Current1.45 (1.25-1.70)
      P < .05.
       Former1.05 (0.91-1.22)
       Never (ref)1.0 (–)
      Number of chronic conditions1.41 (1.33-1.49)
      P < .05.
      CI = confidence interval; FPL = federal poverty level; HS = high school; SNAP = Supplemental Nutrition Assistance Program; WIC = Special Supplemental Nutrition Program for Women, Infants, and Children.
      ∗ Adjusted odds ratio (aOR) represents multivariable logistic regression results with adjustment for all variables in table.
      † P < .05.

      Risk Factors for Cost-related Medication Underuse, Food Insecurity, or Both

      Compared with participants with neither cost-related medication underuse nor food insecurity, participants reporting both were more likely to be Hispanic, non-Hispanic black, have no usual place of care, have no insurance, and have more chronic conditions (Table 3). These participants also were less likely to have public, non-Medicare insurance and to be WIC beneficiaries.
      Table 3Multivariable Logistic Regression Results Comparing Those Reporting Food Security and No Cost-related Medication Underuse (CRMU)
      Both Food Insecure and CRMU
      Represents results from multivariable logistic regression with “both food insecurity and cost-related medication underuse” status as dependent variable in a sample of chronically ill patients with neither (n = 6441) or both (n = 1082).


      aOR (95% CI)      		Food Insecure but no CRMU
      Represents results from multivariable logistic regression with “food insecurity but no cost-related medication underuse” status as dependent variable in a sample of chronically ill patients with neither (n = 6441) or food insecurity but no CRMU (n = 908)


      aOR (95% CI)      		Food Secure but CRMU
      Represents results from multivariable logistic regression with “cost-related medication underuse but no food insecurity” status as dependent variable in a sample of chronically ill patients with neither (n = 6441) or food secure patients with CRMU (n = 1258).


      aOR (95% CI)
      Age (y)0.94 (0.93-0.94)
      Represents P <.05.
      		0.96 (0.95-0.97)
      Represents P <.05.
      		0.95 (0.95-0.96)
      Represents P <.05.
      Male0.70 (0.58-0.84)
      Represents P <.05.
      		0.98 (0.81-1.17)0.74 (0.63-0.87)
      Represents P <.05.
      Race/ethnicity
       Hispanic1.58 (1.12-2.22)
      Represents P <.05.
      		1.80 (1.35-2.39)
      Represents P <.05.
      		1.39 (1.05-1.85)
      Represents P <.05.
       Non-Hispanic white (ref)1.0 (–)1.0 (–)1.0 (–)
       Non-Hispanic black1.58 (1.24-2.00)
      Represents P <.05.
      		1.64 (1.30-2.08)
      Represents P <.05.
      		1.15 (0.94-1.39)
       Asian1.04 (0.54-1.97)1.02 (0.59-1.78)0.85 (0.49-1.47)
       Other1.44 (0.68-3.08)1.55 (0.77-3.15)0.90 (0.36-2.30)
      No usual place of care1.73 (1.17-2.56)
      Represents P <.05.
      		1.40 (0.92-2.14)1.30 (0.90-1.87)
      Insurance
       None3.73 (2.70-5.17)
      Represents P <.05.
      		1.10 (0.73-1.65)3.62 (2.74-4.78)
      Represents P <.05.
       Private, w/Rx coverage1.0 (–)1.0 (–)1.0 (–)
       Private, no Rx coverage0.62 (0.23-1.71)0.65 (0.25-1.71)1.68 (0.95-2.97)
       Medicare1.23 (0.88-1.72)0.96 (0.68-1.35)0.95 (0.75-1.20)
       Public, non-Medicare0.70 (0.50-0.96)
      Represents P <.05.
      		1.48 (1.11-1.99)
      Represents P <.05.
      		0.55 (0.41-0.72)
      Represents P <.05.
      Income
       <100% FPL3.87 (2.82-5.33)
      Represents P <.05.
      		3.86 (2.92-5.11)
      Represents P <.05.
      		0.89 (0.67-1.18)
       100-200% FPL4.81 (3.72-6.20)
      Represents P <.05.
      		3.77 (2.94-4.84)
      Represents P <.05.
      		1.71 (1.38-2.12)
      Represents P <.05.
       >200% FPL (ref)1.0 (–)1.0 (–)1.0 (–)
      Education
       <HS Diploma0.97 (0.7-1.23)1.39 (1.11-1.74)
      Represents P <.05.
      		0.96 (0.76-1.22)
       HS Diploma1.05 (0.85-1.31)0.95 (0.75-1.21)0.90 (0.75-1.08)
       >HS Diploma (ref)1.0 (–)1.0 (–)1.0 (–)
      Born outside US0.98 (0.71-1.35)1.08 (0.80-1.46)0.89 (0.63-1.24)
      On SNAP1.61 (1.25-2.08)
      Represents P <.05.
      		1.82 (1.41-2.35)
      Represents P <.05.
      		1.36 (1.02-1.81)
      Represents P <.05.
      On WIC0.39 (0.25-0.61)
      Represents P <.05.
      		0.50 (0.30-0.84)
      Represents P <.05.
      		0.89 (0.58-1.36)
      Smoking status
       Current2.01 (1.61-2.50)
      Represents P <.05.
      		1.43 (1.15-1.79)
      Represents P <.05.
      		1.51 (1.25-1.83)
      Represents P <.05.
       Former0.98 (0.79-1.21)0.92 (0.72-1.16)1.09 (0.91-1.31)
       Never (ref)1.0 (–)1.0 (–)1.0 (–)
      Number of chronic conditions, per condition1.56 (1.44-1.70)
      Represents P <.05.
      		1.17 (1.08-1.27)
      Represents P <.05.
      		1.45 (1.36-1.55)
      Represents P <.05.
      aOR = adjusted odds ratio; CI = confidence interval; FPL = federal poverty level; HS = high school; SNAP = Supplemental Nutrition Assistance Program; WIC = Special Supplemental Nutrition Program for Women, Infants, and Children.
      ∗ Represents results from multivariable logistic regression with “both food insecurity and cost-related medication underuse” status as dependent variable in a sample of chronically ill patients with neither (n = 6441) or both (n = 1082).
      † Represents results from multivariable logistic regression with “food insecurity but no cost-related medication underuse” status as dependent variable in a sample of chronically ill patients with neither (n = 6441) or food insecurity but no CRMU (n = 908)
      ‡ Represents results from multivariable logistic regression with “cost-related medication underuse but no food insecurity” status as dependent variable in a sample of chronically ill patients with neither (n = 6441) or food secure patients with CRMU (n = 1258).
      § Represents P <.05.
      Individuals reporting only food insecurity were more likely to be Hispanic, non-Hispanic black, and have more chronic conditions but were less likely to be WIC beneficiaries than adults reporting neither food insecurity nor cost-related medication underuse. Adults reporting only cost-related medication underuse were more likely to have no insurance and more chronic conditions.
      Finally, we conducted an analysis among those reporting only cost-related medication underuse but not food insecurity and those reporting only food insecurity but not cost-related medication underuse (Table 4). Participants who were Hispanic and non-Hispanic black were less likely to report cost-related medication underuse but more likely to report food insecurity. Adults who met food but not medication needs were more likely to have dependent children in the household, no insurance or private insurance without drug coverage, and more chronic conditions. Sensitivity analyses restricted to adults with incomes <200% FPL revealed similar results in a smaller sample (eAppendix Table 3).
      Table 4Associations with Report of Cost-related Medication Underuse (CRMU) But No Food Insecurity Compared with Report of Food Insecurity But No CRMU
      aOR
      Adjusted odds ratio (aOR) >1 represents more likely to report CRMU but no food insecurity; aOR <1 represents more likely to report food insecurity but no CRMU. Adjusted OR represents adjustment for all variables in table in a multivariable logistic regression analysis among participants with CRMU but no food insecurity (n = 1258) and food insecurity but not CRMU (n = 908). “CRMU but no food insecurity” status was the dependent variable.
      (95% CI)
      Age (y)1.00 (0.99-1.01)
      Male0.84 (0.66-1.06)
      Race/ethnicity
       Hispanic0.69 (0.47-0.99)
      P <.05.
       Non-Hispanic white (ref)1.0 (–)
       Non-Hispanic black0.63 (0.48-0.83)
      P <.05.
       Asian0.78 (0.40-1.53)
       Other0.69 (0.29-1.66)
      No usual place of care1.09 (0.71-1.69)
      Insurance
       None2.79 (1.85-4.22)
      P <.05.
       Private, w/Rx coverage (ref)1.0 (–)
       Private, no Rx coverage2.79 (0.94-8.29)
       Medicare0.98 (0.67-1.43)
       Public, non-Medicare0.38 (0.26-0.55)
      P <.05.
      Income
       <100% FPL0.26 (0.19-0.37)
      P <.05.
       100-200% FPL0.42 (0.31-0.57)
      P <.05.
       >200% FPL (ref)1.0 (–)
      Education
       <HS diploma0.63 (0.47-0.84)
      P <.05.
       HS diploma0.97 (0.74-1.28)
       >HS diploma (ref)1.0 (–)
      Born outside US0.88 (0.59-1.30)
      SNAP beneficiary0.75 (0.54-1.02)
      WIC beneficiary1.49 (0.81-2.73)
      Dependent children1.45 (1.08-1.96)
      P <.05.
      Smoking status
       Current0.98 (0.76-1.27)
       Former1.16 (0.87-1.54)
       Never (ref)1.0 (–)
      Number of chronic conditions1.19 (1.09-1.30)
      P <.05.
      CI = confidence interval; FPL = federal poverty level; HS = high school; SNAP = Supplemental Nutrition Assistance Program; WIC = Special Supplemental Nutrition Program for Women, Infants, and Children.
      ∗ Adjusted odds ratio (aOR) >1 represents more likely to report CRMU but no food insecurity; aOR <1 represents more likely to report food insecurity but no CRMU. Adjusted OR represents adjustment for all variables in table in a multivariable logistic regression analysis among participants with CRMU but no food insecurity (n = 1258) and food insecurity but not CRMU (n = 908). “CRMU but no food insecurity” status was the dependent variable.
      † P <.05.

      Discussion

      In this nationally representative sample of adults with chronic disease, food insecurity and cost-related medication underuse were common, with 22% reporting food insecurity or cost-related medication underuse. Another 11% reported both food insecurity and cost-related medication underuse. The high overall prevalence of food insecurity and cost-related medication underuse highlights how difficult successful chronic disease management in the current social environment is. These findings suggest residual unmet needs for food-insecure participants and thus have clear implications for health policy. Additionally, our finding that respondents with incomes from 100%-200% FPL, an income range associated with the phase out of many means-tested benefits, report high rates of food insecurity and cost-related medication underuse may be an important consideration when setting program eligibility rules.
      We report an association between public insurance programs, such as Medicaid, and lower cost-related medication underuse compared with private insurance prescription benefits. This likely reflects the lower levels of cost sharing, such as lower copayments, in Medicaid plans. Low- or no-cost sharing prescription drug benefits have been associated with improved health outcomes in a general population,
      • Choudhry N.K.
      • Avorn J.
      • Glynn R.J.
      • et al.
      Full coverage for preventive medications after myocardial infarction.
      N Engl J Med. 2011; 365: 2088-2097
      as well as the reduction of socioeconomic disparities in health outcomes.
      • Booth G.L.
      • Bishara P.
      • Lipscombe L.L.
      • et al.
      Universal drug coverage and socioeconomic disparities in major diabetes outcomes.
      Diabetes Care. 2012; 35: 2257-2264
      Similarly, we found that patients enrolled in WIC had a lower prevalence of food insecurity and cost-related medication underuse. While the WIC program serves a population different from many chronically ill Americans, it may be useful to investigate whether aspects of the program itself are particularly effective in reducing economic insecurity. The observation that Medicaid and WIC participation is associated with lower odds of both food insecurity and cost-related medication underuse suggest that there may be important “spill-over” effects from programs that target food insecurity or cost-related medication underuse, by freeing up available resources. Interestingly, SNAP participation was associated with higher rates of food insecurity. Prior analyses also have noted greater food insecurity in SNAP participants compared with income-matched nonparticipants, which is often attributed to self-selection into the program.
      • Wilde P.E.
      Measuring the effect of food stamps on food insecurity and hunger: research and policy considerations.
      J Nutr. 2007; 137: 307-310
      Our analysis identified particular population subgroups that, when faced with insufficient resources to meet both food and medication needs, were more likely to report food insecurity rather than cost-related medication underuse and vice versa. These patient groups may particularly benefit from targeted interventions. Specifically, Hispanic and non-Hispanic black participants with insufficient resources were more likely to report food insecurity while following their prescribed treatment plan, which may represent an unforeseen harm of medical interventions. Alternatively, those with more chronic conditions and dependent children are more likely to forgo needed medications—with potentially important consequences both for their own health and for their children if their health worsens. This is consistent with observations that caregivers are less likely to adhere to their medications.
      • Shrank W.H.
      • Liberman J.N.
      • Fischer M.A.
      • et al.
      Are caregivers adherent to their own medications?.
      J Am Pharm Assoc (2003). 2011; 51: 492-498
      Because of these findings, clinicians should be attuned to the difficult trade-offs between food and medication-related expenses that patients face. It also is worthwhile to keep in mind the larger social context of the patient's world—children, relatives, and others to whom the patient devotes scarce resources. Assessing for both household food insecurity and cost-related medication underuse, especially in patients from groups at high risk, may be warranted. If identified, engagement of the patient in state, local, or private (such as pharmaceutical company-based) benefit programs may help reduce the burden of competing choices. However, the ability of individual clinicians to rectify unequally distributed social determinants of health is limited. We need organized, broad-based support for multidisciplinary programs to improve social determinants in vulnerable populations.
      Our analysis is consistent with and expands upon prior work. In an analysis from a representative population-based sample of Californians with diabetes, Billimek and Sorkin
      • Billimek J.
      • Sorkin D.H.
      Food insecurity, processes of care, and self-reported medication underuse in patients with type 2 diabetes: results from the California Health Interview Survey.
      Health Serv Res. 2012; 47: 2159-2168
      found that food insecurity was associated with one component of cost-related medication underuse, delays in filling prescriptions. However, whether this was generalizable to other patient populations or other domains of cost-related medication underuse was previously unknown. Work by Kushel et al
      • Kushel M.B.
      • Gupta R.
      • Gee L.
      • Haas J.S.
      Housing instability and food insecurity as barriers to health care among low-income Americans.
      J Gen Intern Med. 2006; 21: 71-77
      demonstrated that food insecurity was associated with increased acute care utilization such as emergency department visits, using data on low-income families drawn from a nationally representative population-based sample, which suggests one possible consequence of cost-related medication underuse. An analysis of Medicare participants,
      • Madden J.M.
      • Graves A.J.
      • Zhang F.
      • et al.
      Cost-related medication nonadherence and spending on basic needs following implementation of Medicare Part D.
      JAMA. 2008; 299: 1922-1928
      before and after Medicare Part D implementation, found that prescription drug coverage increased disposable income available for “basics” such as food. Also consistent with prior work,
      • Zivin K.
      • Ratliff S.
      • Heisler M.M.
      • et al.
      Factors influencing cost-related nonadherence to medication in older adults: a conceptually based approach.
      Value Health. 2010; 13: 338-345
      we observed that educational attainment is not associated with cost-related medication underuse alone, but is associated with food insecurity (and food insecurity plus cost-related medication underuse). The reasons for this apparent discrepancy are not entirely clear. It may be that those with chronic illness and lower educational attainment receive assistance in enrolling for prescription drug benefits when they contact the health care system, while there is less assistance in accessing the safety net for food insecurity. Alternatively, the positive association between lower education and food insecurity may represent residual confounding.
      NHIS data are cross-sectional and observational, which prevent us from evaluating time ordering or establishing cause. For example, the association between SNAP participation, food insecurity, and cost-related medication underuse most likely represents confounding by indication—those participating in SNAP are the more severely under-resourced—rather than a causal association between SNAP and cost-related medication underuse. Despite this inability to generate causal inference, our results do demonstrate that food-insecure adults are a group with high rates of cost-related medication underuse and may particularly benefit from targeted interventions for both food insecurity and cost-related medication underuse. Next, because participants were not directly asked whether they sacrificed food for medications, or medications for food, we cannot be sure of the reason a food or medication need went unmet. Nevertheless, identifying groups that are more likely to have food insecurity when their medication needs are met, or vice versa, may be important for targeting policies to these vulnerable groups. Another limitation is that clinical outcomes, such as laboratory measurements or health care records, are not available in the NHIS dataset. Finally, reporting cost-related medication underuse may be subject to social desirability bias. However, there is no reason to suspect a differential effect on reporting.
      These limitations however, are balanced by several strengths. Our analysis used data collected from a nationally representative, population-based sample with well-validated items. The large effect size after adjustment makes residual confounding a less likely explanation for the observed associations. Finally, food insecurity represents an easily identifiable risk factor for cost-related medication underuse. Because several programs are known to improve both cost-related medication underuse and health outcomes,
      • Choudhry N.K.
      • Avorn J.
      • Glynn R.J.
      • et al.
      Full coverage for preventive medications after myocardial infarction.
      N Engl J Med. 2011; 365: 2088-2097
      , 
      • Booth G.L.
      • Bishara P.
      • Lipscombe L.L.
      • et al.
      Universal drug coverage and socioeconomic disparities in major diabetes outcomes.
      Diabetes Care. 2012; 35: 2257-2264
      it may be worth evaluating whether utilizing food insecurity as an eligibility criterion for similar programs could help improve health in chronically ill patients with food insecurity.
      In conclusion, food insecurity is strongly associated with cost-related medication underuse, and approximately 1 in 3 chronically ill NHIS participants are unable to afford food, medications, or both, despite participation in assistance programs. Interventions targeted to under-resourced groups who may face “treat or eat” choices could produce substantial health gains for these vulnerable patients.

      Appendix

      eAppendix Table 1Factors Associated with Cost-related Medication Underuse in All Adult Participants (n = 32,092)
      aOR
      Adjusted odds ratio (aOR) represents multivariable logistic regression results with adjustment for all variables in the table.
      (95% CI)
      Food insecure4.56 (4.13-5.03)
      P <.05.
      Age (y)0.99 (0.98-0.99)
      P <.05.
      Male0.63 (0.58-0.69)
      P <.05.
      Race/ethnicity
       Hispanic1.09 (0.95-1.24)
       Non-Hispanic white (ref)1.0 (–)
       Non-Hispanic black1.10 (0.99-1.23)
       Asian0.84 (0.66-1.05)
       Other1.09 (0.73-1.61)
      No usual place of care0.76 (0.67-0.87)
      P <.05.
      Insurance
       None3.11 (2.77-3.49)
      P <.05.
       Private, w/Rx coverage1.0 (–)
       Private, no Rx coverage1.16 (0.87-1.53)
       Medicare0.72 (0.62-0.84)
       Public, non-Medicare0.53 (0.46-0.63)
      Income
       <100% FPL1.11 (0.97-1.26)
       100-200% FPL1.37 (1.23-1.52)
      P <.05.
       >200% FPL (ref)1.0 (–)
      Education
       <HS Diploma0.77 (0.67-0.88)
      P <.05.
       HS Diploma0.88 (0.80-0.97)
      P <.05.
       >HS Diploma (ref)1.0 (–)
      Born outside US0.98 (0.86-1.12)
      On SNAP1.37 (1.21-1.56)
      P <.05.
      On WIC0.93 (0.79-1.09)
      Smoking status
       Current1.73 (1.57-1.91)
      P <.05.
       Former1.34 (1.20-1.49)
      P <.05.
       Never (ref)1.0 (–)
      Number of chronic conditions1.56 (1.51-1.62)
      P <.05.
      CI = confidence interval; FPL = federal poverty level; HS = high school; SNAP = Supplemental Nutrition Assistance Program; WIC = Special Supplemental Nutrition Program for Women, Infants, and Children.
      ∗ Adjusted odds ratio (aOR) represents multivariable logistic regression results with adjustment for all variables in the table.
      † P <.05.
      eAppendix Table 2Factors Associated with Cost-related Medication Underuse in Chronically Ill Subgroups
      CVD

      n = 6271

      aOR (95% CI)      		Diabetes

      n = 2256

      aOR (95% CI)      		Non-CVD

      n = 7476

      aOR (95% CI)      		Psychiatric

      n = 8625

      aOR (95% CI)
      Food insecure4.25 (3.44-5.24)5.86 (4.18-8.22)4.12 (3.40-5.00)4.01 (3.34-4.81)
      Age (y)0.95 (0.94-0.96)0.95 (0.94-0.97)0.96 (0.95-0.96)0.96 (0.96-0.97)
      Male0.79 (0.68-0.93)0.94 (0.71-1.23)0.74 (0.64-0.86)0.78 (0.68-0.91)
      Race/ethnicity
       Hispanic1.02 (0.76-1.38)1.06 (0.70-1.59)1.36 (1.02-1.81)1.32 (1.03-1.70)
       Non-Hispanic white (ref)
       Non-Hispanic black1.15 (0.94-1.41)1.29 (0.92-1.80)1.14 (0.95-1.36)1.19 (0.99-1.43)
       Asian0.84 (0.49-1.44)0.52 (0.21-1.28)1.10 (0.66-1.83)1.04 (0.68-1.59)
       Other0.50 (0.22-1.15)0.48 (0.13-1.83)1.05 (0.54-2.05)1.18 (0.63-2.19)
      No usual place of care1.64 (1.06-2.54)2.30 (1.08-4.92)1.36 (0.95-1.93)1.29 (0.92-1.80)
      Insurance
       None3.32 (2.44-4.54)3.16 (1.90-5.25)3.72 (2.83-4.91)3.32 (2.56-4.32)
       Private, w/Rx coverage (ref)
       Private, no Rx coverage2.11 (1.03-4.30)1.32 (0.39-4.47)1.68 (0.88-3.21)1.72 (0.98-3.04)
       Medicare1.03 (0.81-1.32)0.76 (0.50-1.15)1.06 (0.83-1.34)0.88 (0.71-1.10)
       Public, non-Medicare0.44 (0.33-0.59)0.34 (0.21-0.55)0.47 (0.36-0.61)0.46 (0.36-0.59)
      Income
       <100% FPL1.14 (0.86-1.51)1.09 (0.69-1.72)0.98 (0.76-1.24)1.21 (0.95-1.52)
       100-200% FPL1.80 (1.46-2.22)1.69 (1.21-2.38)1.64 (1.34-2.00)1.79 (1.49-2.14)
       >200% FPL (ref)
      Education
       <HS diploma0.89 (0.71-1.12)0.67 (0.48-0.93)0.89 (0.73-1.09)0.85 (0.71-1.03)
       HS diploma0.86 (0.70-1.04)0.89 (0.65-1.23)1.00 (0.84-1.18)0.96 (0.81-1.12)
       >HS diploma (ref)
      Born outside US1.02 (0.76-1.38)1.11 (0.72-1.74)0.82 (0.60-1.11)0.90 (0.69-1.17)
      On SNAP0.98 (0.76-1.27)0.71 (0.47-1.06)1.22 (0.95-1.56)1.05 (0.83-1.32)
      On WIC0.90 (0.53-1.55)1.07 (0.42-2.72)0.85 (0.55-1.30)1.02 (0.69-1.51)
      Smoking status
       Current1.45 (1.16-1.80)1.22 (0.83-1.80)1.42 (1.20-1.69)1.39 (1.16-1.65)
       Former1.03 (0.86-1.23)1.34 (0.99-1.79)1.04 (0.88-1.23)1.05 (0.90-1.23)
       Never (ref)
      Number of chronic conditions1.44 (1.35-1.54)1.39 (1.23-1.56)1.41 (1.32-1.50)1.39 (1.31-1.49)
      Cardiovascular disease (CVD) analysis includes participants who report a history of stroke, hypertension, or coronary heart disease (including myocardial infarction). Diabetes analysis includes patients who report a history of diabetes. Non-CVD analysis includes participants who report a history of asthma, chronic obstructive pulmonary disease, arthritis, or malignancy. Psychiatric analysis includes participants who report a history of a “psychiatric problem.”
      aOR = adjusted odds ratio; CI = confidence interval; FPL = federal poverty level; HS = high school; SNAP = Supplemental Nutrition Assistance Program; WIC = Special Supplemental Nutrition Program for Women, Infants, and Children.
      eAppendix Table 3Associations with Report of Cost-related Medication Underuse (CRMU) But No Food Insecurity Compared with Report of Food Insecurity But No CRMU, Restricted to Participants with Income <200% FPL (n = 3972)
      aOR
      Adjusted odds ratio (aOR) >1 represents more likely to report CRMU but no food insecurity; aOR < 1 represents more likely to report food insecurity but no CRMU. Adjusted OR represents adjustment for all variables in table in a multivariable logistic regression analysis among participants with CRMU but no food insecurity and food insecurity but not CRMU. ‘CRMU but no food insecurity’ status was the dependent variable.
      (95% CI)
      Age (y)1.01 (0.99-1.02)
      Male0.99 (0.73-1.33)
      Race/ethnicity
       Hispanic0.74 (0.47-1.16)
       Non-Hispanic white (ref)1.0 (–)
       Non-Hispanic black0.61 (0.43-0.87)
      P <.05.
       Asian0.61 (0.27-1.40)
       Other0.65 (0.23-1.91)
      No usual place of care0.84 (0.49-1.44)
      Insurance
       None1.63 (0.91-2.94)
       Private, w/ Rx coverage (ref)1.0 (–)
       Private, no Rx coverage0.95 (0.19-4.62)
       Medicare0.51 (0.29-0.90)
      P <.05.
       Public, non-Medicare0.20 (0.11-0.36)
      P <.05.
      Income
       <100% FPL0.61 (0.45-0.82)
      P <.05.
       100-200% FPL (ref)1.0 (–)
      Education
       <HS Diploma0.83 (0.59-1.18)
       HS Diploma1.24 (0.88-1.75)
       >HS Diploma (ref)1.0 (–)
      Born outside US0.85 (0.52-1.41)
      On SNAP0.88 (0.63-1.22)
      On WIC1.69 (0.90-3.20)
      Dependent children1.55 (1.07-2.24)
      P <.05.
      Smoking status
       Current1.09 (0.77-1.54)
       Former1.08 (0.75-1.56)
       Never (ref)1.0 (–)
      Number of chronic conditions1.10 (0.99-1.23)
      CI = confidence interval; FPL = federal poverty level; HS = high school; ref = reference; SNAP = Supplemental Nutrition Assistance Program; WIC = Special Supplemental Nutrition Program for Women, Infants, and Children.
      ∗ Adjusted odds ratio (aOR) >1 represents more likely to report CRMU but no food insecurity; aOR < 1 represents more likely to report food insecurity but no CRMU. Adjusted OR represents adjustment for all variables in table in a multivariable logistic regression analysis among participants with CRMU but no food insecurity and food insecurity but not CRMU. ‘CRMU but no food insecurity’ status was the dependent variable.
      † P <.05.

      References

        • Ward B.
        • Schiller J.
        • Freeman G.
        • Peregoy J.
        Early Release of Selected Estimates Based on Data from the January–September 2012 National Health Interview Survey.
        National Center for Health Statistics, Hyattsville, MDMarch 2013
        View in Article
        • Gallup Healthways
        Gallup-Healthways Monthly U.S. Well-Being Report - April 2012.
        Healthways, Franklin, TN2012
        View in Article
        • Piette J.D.
        • Heisler M.
        • Wagner T.H.
        Cost-related medication underuse among chronically ill adults: the treatments people forgo, how often, and who is at risk.
        Am J Public Health. 2004; 94: 1782-1787
        View in Article
        • Madden J.M.
        • Graves A.J.
        • Zhang F.
        • et al.
        Cost-related medication nonadherence and spending on basic needs following implementation of Medicare Part D.
        JAMA. 2008; 299: 1922-1928
        View in Article
        • Coleman-Jensen A.
        • Nord M.
        • Andrews M.
        • Carlson S.
        Statistical Supplement to Household Food Security in the United States in 2011.
        U.S. Department of Agriculture Economic Research Service, Washington, DC2012
        View in Article
        • Bickel G.
        • Nord M.
        • Price C.
        • et al.
        Guide to Measuring Household Food Security, Revised 2000.
        U.S. Department of Agriculture Food and Nutrition Service, Alexandria, VA2000
        View in Article
        • Pan L.
        • Sherry B.
        • Njai R.
        • Blanck H.M.
        Food insecurity is associated with obesity among US adults in 12 states.
        J Acad Nutr Diet. 2012; 112: 1403-1409
        View in Article
        • Parker E.D.
        • Widome R.
        • Nettleton J.A.
        • Pereira M.A.
        Food security and metabolic syndrome in U.S. adults and adolescents: findings from the National Health and Nutrition Examination Survey, 1999-2006.
        Ann Epidemiol. 2010; 20: 364-370
        View in Article
        • Seligman H.K.
        • Laraia B.A.
        • Kushel M.B.
        Food insecurity is associated with chronic disease among low-income NHANES participants.
        J Nutr. 2010; 140: 304-310
        View in Article
        • Seligman H.K.
        • Schillinger D.
        Hunger and socioeconomic disparities in chronic disease.
        N Engl J Med. 2010; 363: 6-9
        View in Article

      11. National Center for Health Statistics, Centers for Disease Control and Prevention. About the National Health Interview Survey. Available at: http://www.cdc.gov/nchs/nhis/about_nhis.htm. Accessed February 10, 2013.

        View in Article
        • Division of Health Interview Statistics
        • National Center for Health Statistics
        NHIS Survey Description.
        Centers for Disease Control, Atlanta, GA2012
        View in Article

      13. Agency for Healthcare Research and Quality. Medical Expenditures Panel Survey (MEPS); 2004. Available at: http://meps.ahrq.gov/mepsweb/survey_comp/survey.jsp. Accessed February 11, 2013.

        View in Article
        • Soumerai S.B.
        • Pierre-Jacques M.
        • Zhang F.
        • et al.
        Cost-related medication nonadherence among elderly and disabled medicare beneficiaries: a national survey 1 year before the medicare drug benefit.
        Arch Intern Med. 2006; 166: 1829-1835
        View in Article

      15. Kaiser Family Foundation. Number of Medicare Beneficiaries with Creditable Prescription Drug Coverage, by Type. Available at: http://kff.org/medicare/state-indicator/medicare-rx-drug-coverage/. Accessed May 8, 2013.

        View in Article

      16. Kaiser Family Foundation. Medicaid Benefits: Prescription Drugs. Available from: http://kff.org/medicaid/state-indicator/prescription-drugs/#. Accessed May 8, 2013.

        View in Article
        • National Center for Health Statistics
        2011 National Health Interview Survey (NHIS) Public Use Data Release.
        U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics, Hyattsville, MD2012
        View in Article
        • Choudhry N.K.
        • Avorn J.
        • Glynn R.J.
        • et al.
        Full coverage for preventive medications after myocardial infarction.
        N Engl J Med. 2011; 365: 2088-2097
        View in Article
        • Booth G.L.
        • Bishara P.
        • Lipscombe L.L.
        • et al.
        Universal drug coverage and socioeconomic disparities in major diabetes outcomes.
        Diabetes Care. 2012; 35: 2257-2264
        View in Article
        • Wilde P.E.
        Measuring the effect of food stamps on food insecurity and hunger: research and policy considerations.
        J Nutr. 2007; 137: 307-310
        View in Article
        • Shrank W.H.
        • Liberman J.N.
        • Fischer M.A.
        • et al.
        Are caregivers adherent to their own medications?.
        J Am Pharm Assoc (2003). 2011; 51: 492-498
        View in Article
        • Billimek J.
        • Sorkin D.H.
        Food insecurity, processes of care, and self-reported medication underuse in patients with type 2 diabetes: results from the California Health Interview Survey.
        Health Serv Res. 2012; 47: 2159-2168
        View in Article
        • Kushel M.B.
        • Gupta R.
        • Gee L.
        • Haas J.S.
        Housing instability and food insecurity as barriers to health care among low-income Americans.
        J Gen Intern Med. 2006; 21: 71-77
        View in Article
        • Zivin K.
        • Ratliff S.
        • Heisler M.M.
        • et al.
        Factors influencing cost-related nonadherence to medication in older adults: a conceptually based approach.
        Value Health. 2010; 13: 338-345
        View in Article

      Article info

      Publication history

      Published online: January 21, 2014

      Footnotes

      Funding: SAB was supported by an Institutional National Research Service Award #T32HP10251, the Ryoichi Sasakawa Fellowship Fund, and by the General Medicine Division at Massachusetts General Hospital. The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of data; or preparation, review, decision to submit for publication, or approval of the manuscript.

      Conflict of Interest: NKC has received unrestricted research grants from CVS Caremark, Aetna, the Commonwealth Fund, and the Robert Wood Johnson Foundation to study medication adherence. SAB and HKS report no conflicts of interest.

      Authorship: SAB had full access to all data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. He also drafted the manuscript. HKS and NKC had access to the data and were critical in writing the manuscript.

      Identification

      DOI: https://doi.org/10.1016/j.amjmed.2014.01.002

      Copyright

      © 2014 Elsevier Inc. Published by Elsevier Inc. All rights reserved.

      ScienceDirect

      Access this article on ScienceDirect

      The content on this site is intended for healthcare professionals.


      We use cookies to help provide and enhance our service and tailor content. To update your cookie settings, please visit the Cookie Preference Center for this site.
      Copyright © 2023 Elsevier Inc. except certain content provided by third parties.


      RELX