Abstract
Background:
Obesity is a significant risk factor of several cancers that imposes a substantial economic burden on US healthcare that remains to be quantified. We estimated the excess costs and economic burden of obesity-related cancers in the United States.
Methods:
From the Medical Expenditure Panel Survey (2008–2015) data, we identified 19 405 cancer survivors and 175 498 non-cancer individuals. We estimated annual health expenditures using generalized linear regression with log link and gamma distribution by cancer types (stratified by 11 obesity-related cancers and other cancer types), controlling for sociodemographic and clinical characteristics. All cost estimates were adjusted to 2015 USD value.
Results:
The average annual total health expenditures were $21 503 (95% CI, $20 946-$22 061) for those with obesity-related cancer and $13 120 (95% CI, $12 920-$13 319) for those with other cancer types. There was a positive association between body mass index and health expenditures among cancer survivors: for each additional 5-unit increase in body mass index, the average predicted expenditures increase by $1503 among those with obesity-related cancer and by $722 among those with other cancers. With adjustments for sociodemographic and clinical characteristics, the mean incremental expenditures of treating obesity-related cancer were 2.1 times higher than those of other cancers ($4492 vs $2139) and more considerable among the non-elderly cancer population. Obesity-related cancers accounted for nearly 43.5% of total direct cancer care expenditures, estimated at $35.9 billion in 2015.
Conclusion:
The economic burden of obesity-related cancer in the United States is substantial. Our findings suggest a need for the inclusion of comprehensive obesity prevention and treatment in cancer care.
Keywords: economic burden of cancer, cancer burden in the United States, obesity-related cancer, excess costs of cancer care
Introduction
More than 15.5 million Americans were living with a history of cancer in 2016, projected to increase to 20 million by 2026.1 Given that obesity is a significant risk factor for developing many cancers,2,3 studies suggest that the increase in cancer cases coincide with increasing obesity prevalence.4,5 The International Agency for Research on Cancer3 and the Centers for Disease Control and Prevention4 identified 13 cancers (including esophageal, colorectal, endometrial, gallbladder, stomach, kidney, liver, ovarian, pancreatic, thyroid, and postmenopausal breast cancers) that are linked with obesity. Between 2014 and 2016, there was about a 10% increase in the rates of obesity-related cancer cases, particularly among younger-age persons,1,4–6 and obesity-related cancers accounted for 40% of all cancer diagnoses in the United States.4
Cancer survivorship is associated with substantial medical expenditures for patients, their caregivers, and families.7–9 Although greater than 65% of all patients in studies characterizing medical expenditures for cancer care had overweight or obesity,9 the excess costs associated with obesity (stratified by body mass index [BMI] levels) among patients with cancer has never been collectively quantified. Because the prevalence of obesity is rapidly increasing in the United States,10 diagnosis of cancers that are obesity-related likely impose a significant economic burden on the patients and US healthcare system, making it crucial to estimate the impact of obesity on cost and medical expenditures associated with obesity-related cancers.11
The objectives of the current study were (1) to estimate health expenditures among patients with a history of obesity-related cancer and other types of cancer, (2) to compare the marginal increase in health expenditure associated with BMI among for obesity-related and other cancers, and (3) to estimate overall economic burden of obesity-related cancer in the United States.
Methods
Data and Study Population
Pooled data from the 2008 to 2015 Medical Expenditure Panel Surveys (MEPS) was used. The MEPS is a continuous nationally representative survey of the noninstitutionalized population assessing access to and use of health care, costs associated with medical services, sources of payment, health insurance coverage, and patient experience.12 Data were merged with the Medical Condition Files to supplement the dataset with condition-specific information. From 2008 to 2015, there were 284 296 individuals between ages 18 or older in the pooled MEPS dataset. To secure more accurate medical diagnosis and address data inconsistencies, our exclusion criteria included individuals with missing values on medical condition files (n = 34 740), sociodemographic (education, marital status, and family income; n = 11 568) and health-related information (health insurance, smoking status, BMI, and comorbid conditions; n = 39 400), and women who reported pregnancy at the time of survey owing to unstable BMI during pregnancy (n = 3685). These criteria yielded a sample of 194 903 individuals, representing 239 612 464 Americans (see Appendix Fig 1 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.004).
We identified a study sample of cancer patients using Clinical Classification Category code (see Appendix Table 1 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.004). The MEPS respondents describe their health conditions in a narrative form, and the narratives from the interviews are then interpreted and assigned by trained MEPS coders into condition diagnostic codes according to the International Classification of Diseases, Ninth Revision.12 The Clinical Classification Category code is a diagnosis and procedure categorization that aggregates the International Classification of Diseases, Ninth Revision codes into a smaller number of clinically meaningful categories. Based on the International Agency for Research on Cancer’s report,3 we stratified cancer patients as either with obesity-related cancers (including esophageal, stomach, colorectal, liver, pancreatic, other gastrointestinal cancers, breast, kidney, uterus, ovarian, and thyroid cancers) or other types of cancer.
Primary Measures
Our primary outcome measure was total health expenditures quantified as sum of direct payments for care provided during the year, including hospital inpatient care, ambulatory care (outpatient, emergency department, and office-based medical provider visits), prescription medications, other medical services (home health, dental, vision, chiropractic care, and other services), and out-of-pocket payments (deductibles, coinsurance, and other cost-sharing). The MEPS collects expenditure data from household and medical provider components. If no data were available from either source, expenditure data were imputed through a robust imputation procedure.12 All expenditures were adjusted to 2015 USD using the consumer price index for medical care.13 Using self-reported height and weight, we categorized respondents into weight categories based on the World Health Organization BMI classification, including underweight (BMI <18.5 kg/m2), normal weight (BMI 18.5–24.9 kg/m2), overweight (BMI 25–29.9 kg/m2), class I obese (BMI 30–34.9 kg/m2), class II obese (BMI 35–39.9 kg/m2), and class III severely obese (BMI ≥ 40 kg/m2).
Covariates
Other covariates included survey year, age, sex, race or ethnicity (Hispanic, non-Hispanic white, non-Hispanic black, non-Hispanic Asian, and other), marital status, education (less than high school, high school, some college, bachelor’s degree, and graduate or professional degree), family income (based on federal poverty level [FPL]: low income [FPL <200%], middle income [FPL 200%–399%], and high income [FPL ≥400%]), census region (Northeast, Midwest, South, West), smoking status, cancer status, comorbid conditions (hypertension, hyperlipidemia, diabetes, heart disease, and chronic obstructive pulmonary disease), and full-year insurance coverage (any private, Medicaid, Medicare, Medicaid plus private [eg, Medicare Advantage or with private plans of other family members], and uninsured).
Statistical Analyses
Cancer type (obesity-related cancer, other cancer type, and non-cancer control) was the primary independent variable for all our analyses. First, weighted descriptive statistics were used to compare sample characteristics by cancer type. All medical expenditures were estimated using generalized linear models (GLMs) with log link and gamma distribution. The models adjusted for sociodemographic and clinical characteristics that were associated with health service utilization and expenditure14,15: survey year, age, sex, race or ethnicity, marital status, education, family income, census region, insurance, and smoking status. We further adjusted for the following comorbid conditions—hypertension, hyperlipidemia, diabetes, heart disease, and chronic obstructive pulmonary disease—given their higher prevalence and association with obesity.16–18 To examine whether the effect of BMI on health expenditures varied by cancer type, we tested an interaction term between BMI and each cancer type (obesity-related cancer and other cancer) in the same model. Incremental expenditures associated with cancer type and other covariates were estimated from respective regression coefficients (independent differentials) in GLM including all covariates. Incremental expenditures for 5-unit increase in BMI were also estimated for each cancer type. We report the adjusted expenditures of medical service types: hospital inpatient, ambulatory care, prescription medications, and out-of-pocket spending. Finally, we extrapolated the total national excess expenditures associated with cancer care by multiplying the estimated cancer type-specific incremental expenditures by the US cancer population estimates using the MEPS sample weights (prevalence-based estimates).
To test the robustness of our results, we performed several sensitivity and subgroup analyses. First, we compared our estimates with those from a 2-part model (known as mixed discrete-continuous variable regression)19 and the specifications of BMI as a continuous variable. Given that costs are likely to be higher closer to cancer diagnosis, as part of our sensitivity analysis, we examined whether difference in estimates varied across time since cancer diagnosis between persons diagnosed with obesity-related cancer and those with other cancer type: <1 year, 2 to 5 years, 5 to 10 years, and ≥11 years (median time of 5 years). We determined time since diagnosis by subtracting self-reported age of cancer diagnosis from the age during the MEPS interview (analysis restricted to only those with information; n = 7494). Lastly, to investigate the within-group variation, we estimated expenditures for the 11 cancer types in the obesity-related cancer group. All analyses were conducted using survey procedures to generate national estimates in SPSS Complex Samples 24 (IBM Corp, Armonk, NY) and SAS 9.4 statistical software (SAS Institute, Cary, NC). The study was granted exempt review by the institutional review board of the University of Florida.
Results
Table 1 presents the summary characteristics of cancer patients aged 18 or older by cancer type during the 2008 to 2015 period (see Appendix Table 2 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.004 for the baseline characteristics of respondents with and without cancer). Of 19 405 patients with cancer, 26.8% (95% CI, 25.5%–28.1%; approximately 8.0 million cancer survivors nationwide) were diagnosed with a history of obesity-related cancer. Compared with those with other cancer type, patients with obesity-related cancer were more likely to be younger, female, and of racial or ethnic minority, and had lower education and family income. Although severe obesity status (class II and III obesity) was more prevalent among those with obesity-related cancer, those with other cancer types had more obesity-associated comorbidities.
Table 1.
Baseline characteristics of the study population by cancer type: MEPS 2008–2015.
| Characteristics | Total cancer survivor |
Obesity-related cancer* |
Other cancer |
P value |
|---|---|---|---|---|
| sample no. (Weighted %) | Weighted %, (95% CI) | |||
| Sample n | 19 405 | 5692 | 13 713 | |
| Population estimates | 29 828 236 | 7 995 359 | 21 832 878 | |
| 26.8 (25.5–28.1) | 73.2 (72.1–74.3) | |||
| Survey year | .037 | |||
| 2008 | 2080 (10.8) | 11.1 (9.9–12.2) | 10.7 (10.1–11.3) | |
| 2009 | 2515 (12.2) | 13.2 (12.1–14.3) | 11.8 (11.2–12.5) | |
| 2010 | 2175 (11.2) | 12.4 (11.3–13.5) | 10.8 (10.1–11.4) | |
| 2011 | 2402 (12.1) | 12.5 (11.3–13.7) | 12.0 (11.2–12.7) | |
| 2012 | 2612 (12.8) | 12.1 (10.7–13.5) | 13.0 (12.1–13.9) | |
| 2013 | 2470 (13.1) | 11.9 (10.6–13.2) | 13.6 (12.7–14.5) | |
| 2014 | 2416 (13.2) | 13.3 (11.9–14.6) | 13.2 (12.4–14.0) | |
| 2015 | 2735 (14.5) | 13.7 (12.1–15.3) | 14.9 (13.9–15.9) | |
| Age, years | <.001 | |||
| 18–34 | 1125 (4.7) | 6.0 (5.0–7.0) | 4.3 (3.7–4.8) | |
| 35–44 | 1461 (6.4) | 7.6 (6.3–8.9) | 5.9 (5.3–6.6) | |
| 45–54 | 2662 (13.2) | 13.9 (12.2–15.6) | 12.9 (12.0–13.9) | |
| 55–64 | 4419 (23.1) | 23.0 (21.0–25.0) | 23.2 (21.6–24.7) | |
| 65–74 | 4737 (25.0) | 22.7 (20.5–24.9) | 25.9 (24.5–27.2) | |
| 75+ | 5001 (27.5) | 26.7 (24.0–29.5) | 27.8 (25.8–29.8) | |
| Sex | <.001 | |||
| Female | 11 290 (55.9) | 75.9 (73.9–77.9) | 48.6 (47.0–50.1) | |
| Male | 8115 (44.1) | 24.1 (22.1–26.1) | 51.4 (49.9–53.0) | |
| Race/ethnicity | <.001 | |||
| NH white | 13 658 (85.5) | 78.8 (76.7–81) | 88.0 (86.9–89.1) | |
| NH black | 2588 (6.0) | 9.0 (7.8–10.2) | 4.9 (4.4–5.5) | |
| Hispanic | 2251 (5.5) | 7.7 (6.5–9.0) | 4.6 (4.0–5.3) | |
| NH Asian | 584 (1.6) | 2.7 (1.9–3.6) | 1.2 (0.8–1.5) | |
| Other | 324 (1.4) | 1.7 (1.1–2.3) | 1.3 (0.9–1.8) | |
| Education | .004 | |||
| Less than high school | 3035 (10.9) | 12.0 (10.4–13.6) | 10.5 (9.5–11.5) | |
| High school or GED | 6743 (34.3) | 37.3 (34.7–39.8) | 33.2 (31.6–34.8) | |
| Some college | 2513 (13.2) | 12.1 (11.0–13.1) | 13.6 (12.7–14.4) | |
| Bachelor’s | 4525 (25.7) | 24.0 (21.8–26.1) | 26.4 (25.0–27.8) | |
| Graduate or professional | 2589 (15.9) | 14.7 (12.6–16.7) | 16.3 (15.1–17.6) | |
| Marital status | <.001 | |||
| Single | 2003 (8.8) | 10.1 (8.6–11.6) | 8.3 (7.4–9.1) | |
| Married | 10 734 (59.7) | 54.6 (51.6–57.6) | 61.6 (59.7–63.6) | |
| Separate/divorced/widowed | 6668 (31.5) | 35.3 (32.8–37.8) | 30.1 (28.4–31.8) | |
| Family income† | ||||
| Low income | 7042 (28.1) | 31.2 (28.9–33.5) | 27.0 (25.6–28.4) | <.001 |
| Middle income | 5462 (27.0) | 28.5 (26.5–30.5) | 26.5 (25.2–27.8) | |
| High income | 6901 (44.9) | 40.3 (37.5–43.1) | 46.5 (44.6–48.5) | |
| Census region | ||||
| Northeast | 3091 (17.8) | 19.5 (16.3–22.7) | 17.1 (15.2–19.0) | .292 |
| Midwest | 4248 (22.6) | 23.0 (20.5–25.4) | 22.5 (20.3–24.6) | |
| South | 7472 (37.8) | 37.2 (34.0–40.4) | 38.0 (35.2–40.8) | |
| West | 4594 (21.8) | 20.3 (17.9–22.7) | 22.4 (20.3–24.5) | |
| Health insurance | ||||
| Private | 11143 (43.0) | 36.3 (34.1–38.6) | 44.6 (42.9–46.2) | <.001 |
| Private + Medicare | 6367 (26.6) | 28.3 (25.9–30.7) | 26.3 (24.8–27.7) | |
| Medicaid | 3162 (7.1) | 9.9 (8.6–11.2) | 6.5 (5.9–7.1) | |
| Medicare | 5535 (18.2) | 21.1 (18.9–23.4) | 17.6 (16.4–18.7) | |
| Uninsured | 1941 (4.9) | 4.3 (3.5–5.2) | 5.1 (4.6–5.6) | |
| Smoking | .869 | |||
| No | 15 224 (85.9) | 85.8 (84.0–87.5) | 86.0 (84.8–87.1) | |
| Yes | 2826 (14.1) | 14.2 (12.5–16.0) | 14.0 (12.9–15.2) | |
| BMI category | <.001 | |||
| Underweight | 1066 (4.9) | 6.5 (5.4–7.6) | 4.4 (3.9–4.9) | |
| Normal | 5786 (31.4) | 32.1 (29.5–34.6) | 31.1 (29.7–32.6) | |
| Overweight | 6617 (35.0) | 32.1 (29.6–34.5) | 36.1 (34.9–37.2) | |
| Class I obese | 3502 (17.4) | 15.9 (14.1–17.7) | 17.9 (16.8–18.9) | |
| Class II obese | 1523 (7.2) | 8.3 (7.2–9.5) | 6.8 (6.0–7.5) | |
| Class III obese | 911 (4.2) | 5.2 (4.2–6.2) | 3.8 (3.2–4.4) | |
| Comorbid conditions | ||||
| Hypertension | .826 | |||
| No | 8323 (43.8) | 43.6 (41.2–46.0) | 43.9 (42.5–45.4) | |
| Yes | 11 074 (56.2) | 56.4 (54.0–58.8) | 56.1 (54.6–57.5) | |
| Hyperlipidemia | ||||
| No | 9040 (45.1) | 47.1 (45.0–49.3) | 44.3 (42.6–46.0) | .034 |
| Yes | 10 341 (54.9) | 52.9 (50.7–55.0) | 55.7 (54.0–57.4) | |
| Diabetes | ||||
| No | 15 756 (83.2) | 80.6 (78.7–82.5) | 84.2 (83.0–85.4) | <.001 |
| Yes | 3647 (16.8) | 19.4 (17.5–21.3) | 15.8 (14.6–17.0) | |
| Heart disease | ||||
| No | 13 373 (67.9) | 69.9 (67.7–72.1) | 67.1 (65.7–68.6) | .037 |
| Yes | 6032 (32.1) | 30.1 (27.9–32.3) | 32.9 (31.4–34.3) | |
| COPD | ||||
| No | 17 562 (91.0) | 91.0 (89.7–92.3) | 91.0 (90.2–91.9) | .956 |
| Yes | 1843 (9.0) | 9.0 (7.7–10.3) | 9.0 (8.1–9.8) |
BMI indicates body mass index; COPD, chronic obstructive pulmonary disease; GED, general education development; MEPS, Medical Panel Expenditure Survey; NH, non-Hispanic.
Includes esophageal, stomach, colorectal, liver, pancreatic, other gastrointestinal cancers, breast, kidney, uterus, ovarian, and thyroid cancers.
Based on household income relative to poverty threshold; low income (less than 200% of poverty level), middle income (200%–399%), and high income (≥400%).
Total Health Expenditures for Patients With Cancer
Overall, the mean predicted total expenditure from the GLM was $7665 (95% CI, $7553–$7777), which was relatively close to the actual average of $7466 (95% CI, $7293-$7639) for the general population in 2015 (see Appendix Table 3 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.004). Specifically by cancer type, the mean predicted total direct expenditures were $21 503 (95% CI, $20 946-$22 061) for those with obesity-related cancer and $13 120 (95% CI, $12 920-$13 319) for those with other cancer types (Fig 1A). Although the mean predicted total expenditures followed a U-shape (increasing with low and high BMI), regardless of cancer type, it appeared that the increase was greater with higher BMI among those with obesity-related cancer (Fig 1B). For instance, among those with normal weight, the mean predicted total expenditures were $18 409 (95% CI, $17 620-$19 182) for those with obesity-related cancer and $11 286 (95% CI, $11 042-$11 524) for those with other cancer. Obesity-related cancer patients with class III obesity had the mean predicted total expenditures of $29 745 (95% CI, $26 354-$33 068), compared with $18 465 (95% CI, $16 919–$19 980) among those with other cancer. Our regression model estimated that, for an additional 5-unit increase in BMI, the average predicted expenditures increased by $1502 (95% CI, $820-$2185) among those with obesity-related cancer and $722 (95% CI, $518-$926) among those with other cancer types.
Figure 1.
Mean predicted total health expenditures among cancer population by cancer type (A) and BMI category (B). Predicted from the model including cancer type, BMI, interactions between cancer type and BMI, age, sex, race or ethnicity, marital status, educational attainment, family income level, census region, type of, smoking status, comorbid conditions and survey year. Error bars represent 95% CIs.
BMI indicates body mass index.
Adjusted Incremental Effect of Cancer Type and BMI on Health Expenditures
Table 2 shows the adjusted incremental effects of health expenditures across cancer type, BMI category, and other covariates. When standardizing expenditures to the covariates of the overall general population, the patients with obesity-related cancer incurred approximately 2.1 times higher excess expenditures than those with other cancer types. In absolute-dollar terms, the net excess annual expenditures of treating obesity-related cancer and other cancer types were estimated to be $4492 (95% CI, $3422-$5741) and $2139 (95% CI, $1868-$2425), respectively. As an example, a male patient in his early 50s with obesity-related cancer and class II obesity is estimated to spend an excess of $8524 on total annual health expenditure compared with counterparts (same characteristics but without cancer) (see Appendix Table 4 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.004 for the regression results). The interactions for cancer type and BMI were statistically significant, suggesting that there was a significant difference in the association of BMI and total expenditures by cancer type (see Appendix Table 5 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.004). Additional 5-unit increase in BMI incurred incremental expenditures of $135 (95% CI, $99-$182) for treating obesity-related cancers and $21 (95% CI, $17-$26) for other cancer (see Appendix Table 6 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.004). Table 3 shows the adjusted incremental expenditures of medical service types. Except for out-of-pocket spending (P = .134), the mean incremental expenditures of treating obesity-related cancers were significantly higher than those of other cancer types across the all types of health services (P<.001).
Table 2.
Adjusted regression results for incremental effects on total medical expenditures (2015 USD).
| Variables | Estimates | 95% CI | P value |
|---|---|---|---|
| Cancer type | |||
| Non-cancer control | Ref. | - | - |
| Obesity-associated cancer | 4492 | 3422–5741 | <.001 |
| Other cancer | 2139 | 1868–2425 | <.001 |
| BMI category | |||
| Normal weight | Ref. | - | - |
| Underweight | 756 | 406–1143 | <.001 |
| Overweight | 4 | −150 to 166 | .962 |
| Class I obese | 55 | −120 to 241 | .547 |
| Class II obese | 466 | 245–702 | <.001 |
| Class III obese | 515 | 224–834 | <.001 |
| Hypertension | 1243 | 1031–1467 | <.001 |
| Hyperlipidemia | 521 | 360–690 | <.001 |
| Diabetes | 1858 | 1634–2094 | <.001 |
| Heart disease | 2082 | 1840–2336 | <.001 |
| COPD | 1535 | 1230–1862 | <.001 |
| Smoking | −77 | −246 to 101 | .388 |
| Age | |||
| 18–34 | Ref. | - | - |
| 35–44 | 391 | 183–612 | <.001 |
| 45–54 | 1008 | 737–1298 | <.001 |
| 55–64 | 1925 | 1574–2304 | <.001 |
| 65–74 | 1103 | 768–1470 | <.001 |
| 75+ | 1397 | 1001–1833 | <.001 |
| Sex | |||
| Female | Ref. | - | - |
| Male | −787 | −891 to −678 | <.001 |
| Race/ethnicity | |||
| NH white | Ref. | - | - |
| NH black | −522 | −656 to −380 | <.001 |
| Hispanic | −820 | −956 to −675 | <.001 |
| NH Asian | −790 | −1159 to −346 | .001 |
| Other | −112 | −409 to 219 | .491 |
| Education | |||
| Less than high school | Ref. | - | - |
| High school or GED | 278 | 36–540 | .024 |
| Some college | 195 | −73 to 487 | .159 |
| Bachelor’s | 577 | 304–875 | <.001 |
| Graduate or professional | 755 | 474–1059 | <.001 |
| Marital status | |||
| Single | Ref. | - | - |
| Married | 141 | −39 to 333 | .128 |
| Separate/divorced/widowed | 196 | −13 to 420 | .066 |
| Family income | |||
| Low income | Ref. | - | - |
| Middle income | −309 | −436 to −176 | <.001 |
| High income | −128 | −285 to 38 | .127 |
| Census region | |||
| Northeast | Ref. | - | - |
| Midwest | −8 | −199 to 197 | .939 |
| South | −358 | −509 to −198 | <.001 |
| West | −105 | −301 to 107 | .322 |
| Health insurance | |||
| Private | Ref. | - | - |
| Private + Medicare | 1704 | 191–3950 | .023 |
| Medicaid | 1479 | 982–2039 | <.001 |
| Medicare | 1651 | 157–3865 | .027 |
| Uninsured | −1904 | −1979 to −1822 | <.001 |
BMI indicates body mass index; COPD, chronic obstructive pulmonary disease; Ref., reference; NH, non-Hispanic.
Table 3.
Adjusted incremental expenditures of cancer care by medical service types by age group (2015 USD).*
| Obesity-associated cancer |
Other cancer |
|||
|---|---|---|---|---|
| Estimates | 95% CI | Estimates | 95% CI | |
| Type of medical services | ||||
| Total medical expenditure | 4492 | 3422–5741 | 2139 | 1868–2425 |
| Hospital inpatient care | 1343 | 1169–1533 | 711 | 642–784 |
| Ambulatory care | 2022 | 1641–2455 | 1035 | 932–1144 |
| Prescription medications | 350 | 327–373 | 231 | 218–243 |
| Total out-of-pocket | 130 | 100–160 | 102 | 88–115 |
Calculated by exponentiating the respective effects on expenditures in the model including cancer type, BMI category, age, sex, race/ethnicity, marital status, educational attainment, family income level, census region, type of, smoking status, comorbid conditions (hypertension, hyperlipidemia, diabetes, heart disease, chronic obstructive pulmonary disease), and survey year.
National Total Excess Expenditures of Cancer Care
The total excess expenditures associated with cancer care were estimated by multiplying the mean incremental expenditures of cancer care ($4492 for obesity-related cancer and $2139 for other cancer type) by the weighted population with a history of cancer (7 995 359 individuals with obesity-related cancer and 21 832 878 with other cancer). The aggregate net excess health expenditures of cancer care were estimated at $82.6 billion in 2015. Although only 26.8% of cancer patients were diagnosed with obesity-related cancers, those with obesity-related cancers accounted for nearly 43.5% of the total excess expenditures, estimated at $35.9 billion annually (Fig 2A). Specifically, by the type of medical services, ambulatory care services contributed 46.9% ($38.8 billion) of total excess expenditures of cancer care, followed by inpatient care (31.8%; $26.2 billion) and prescription medications (9.5%; $7.8 billion) (Fig 2B). The regression results of expenditures for each service type are available in Appendix Tables 7, 8, 9, and 10 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.004.
Figure 2.
Estimated national prevalence of cancer population (A) and aggregate economic burden of cancer care (B). Estimated by multiplying next excess cancer-related expenditures by cancer survivor population estimates from the pooled MEPS sample weights.
MEPS indicates Medical Expenditure Panel Surveys.
Sensitivity and Subgroup Analysis
The results of the sensitivity analyses using the 2-part model (see Appendix Table 11 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.004) and BMI as a continuous variable were relatively consistent with the main results (see Appendix Table 5 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.004). In subgroup analyses, the differences in estimates between obesity-related cancer and other cancers were relatively consistent; obesity-related cancer incurred more expenditure than other cancer types in both short-term and longer-term cancer survivorship (see Appendix Table 12 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.004). Among obesity-related cancers, patients with ovarian cancer had the highest average predicted total annual health expenditures ($58 581), followed by those with liver cancer ($55 090), pancreatic cancer ($36 433), and colorectal cancer ($31 434) (see Appendix Table 13 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.004).
Discussion
Patients with obesity-related cancer had approximately 2.1 times higher direct excess health expenditures than those with other cancer types. In 2008 to 2015, the mean annual excess expenditures of cancer care were $4492 per patient with obesity-related cancer and $2139 per patient with other cancer. Aggregate national excess health expenditures for obesity-related cancer was estimated at $36 billion per annum, accounting for 43.5% of the total direct excess expenditures associated with cancer care ($82.6 billion). Our estimates of economic burden of cancer care are comparable with the previous estimates reported by the Agency for Healthcare Research and Quality and American Cancer Society ($80.2 billion in 2015).20 Although there were no significant changes in the trends for health expenditures among cancer patients during the study period, we observed consistently higher costs attributable to obesity-related cancer across all health services type, 51.5% higher in prescription medications to 95.4% in ambulatory care, compared with those for other cancer.
Our findings suggest that obesity-related cancers impose a significantly higher burden on healthcare, and this burden becomes greater with higher BMI among those with obesity-related cancer than those with other cancer types, even after controlling for other comorbid conditions. These findings may suggest an independent effect of BMI on health expenditure among all cancer patients. Although it remains unclear how weight change or preexisting condition trajectories affect their health services use, the differences in expenditure between obesity-related and other cancers could also be explained, partly, by the poorer prognosis cited with obesity.21–23 Studies have demonstrated that having obesity and excess weight gain are significant predictors of mortality.24–26 Cancer cachexia may play a role in these differences, given that cancer cachexia is more common in obesity-related cancer than other cancer types (based on weight loss; about 80% of gastric/pancreatic cancer patients versus about 50% of those with lung, prostate cancer, or leukemia).27,28 For instance, those with obesity-related cancer are less likely to experience cachexia (ie, loss of adipose tissue) or a significant decline in physical functioning resulting from cancer cachexia; this may be linked to improved survival outcome,25,29 but increased health expenditures for extended care. On the contrary, it is possible that a worsening condition often manifested by weight loss among cancer patients may be obscured by obesity, leading to a more severe condition and excess costs.23 This may result in more costly care along with more likelihood of being hospitalized for a more extended period or in facilities for palliative or end-of-life care.21,22
Our findings have important implications for patients, payers, health systems, and policymakers. To date, less efforts have been made to evaluate the economic impact of obesity on cancer survivorship, including potential short-term and long-term cost savings resulting from obesity prevention or treatment interventions among cancer patients. Although it may be argued that cancer care itself is already associated with a significant medical cost burden,7 it is crucial to identify preventable and modifiable factors (eg, behavioral and psychological factors) or identify and target high-risk populations to potentially maximize cost savings and minimize the disease burden on patients. Our analyses reveal differential association patterns of having overall cancer and obesity-related cancer across individual characteristics. For instance, although most cancers are common in older adults,6 the prevalence of obesity-related cancer was higher in working-age adults (18–64 years) in this study. The observed variations across age groups in this study are likely to reflect true differences in the prevalence of obesity-related cancer. In the past decade, the prevalence of obesity has increased markedly among youth and adolescents,10 which may be contributing to the higher prevalence of obesity-related cancer among younger adults. A study noted that incidence rates of colorectal cancer among young adults increased, while there was a decreasing trend in overall incidence of colorectal cancer.30,31 Considering both direct and indirect morbidity costs (eg, lost productivity) owing to cancer, those working-age cancer patients are likely incurring higher expenditures, and their financial burden of care is expected to grow at a faster rate with increasing prevalence of obesity and obesity-related cancer.
Our analysis also shows that the prevalence of obesity-related cancer was higher among racial and ethnic minorities (32%–46%) than non-Hispanic whites (25%), which is contrary to what we observed among those with any cancer. Considering higher prevalence rates of colorectal cancer and breast cancer, major obesity-related cancers among racial/ethnic minorities, these differences may be attributed to continued lower screening uptake rates for colorectal cancer or breast cancer among racial/ethnic minorities.32,33 Multifaceted prevention strategies including education and sustained intervention programs to tackle obesity among the general population and cancer survivors may help reduce the economic burden of obesity on cancer.11,34 Future studies using claims data would be useful to improve our knowledge of the association between obesity and cancer care and to evaluate how and to what extent specific health-related factors or services provided contribute to the burden of the disease.
Despite its strengths, this study has limitations. We may be underestimating the health expenditures among cancer patients because our estimated incremental costs and total health expenditures were based on noninstitutionalized US populations. Furthermore, we did not capture indirect or nonmedical expenses such as travel time to health providers, loss of productivity owing to diseases, or burden of caregivers, which also contributes substantially to the overall cost burden of cancer patients. Estimated expenditures in our study pertain to short-term health consequences, which may differ significantly by the time of diagnosis and length of cancer survivorship owing to different treatment patterns in early diagnosis and later periods.35 Specifically, the most substantial cost associated with cancer care is incurred during the first or last year of life among cancer patients.36,37 The cross-sectional nature of the MEPS and the lack of information regarding cancer diagnosis time precluded us from estimating the phase-based cost of cancer care. Furthermore, survivors of high-mortality cancers were likely to be less represented in the data. Despite these limitations, using MEPS data allowed us to estimate most up-to-date and comprehensive population-level national annual expenditures associated with obesity-associated cancers, given its nationally representative sample reflecting characteristics of cancer population in aggregate.
Our findings demonstrate that the economic burden of obesity-related cancer is substantial in the United States. This burden is expected to be greater as the prevalence of obesity increases. Policy interventions for implementing obesity prevention or treatment programs among the general population and cancer patients could have an impact on overall cancer care expenditures. Continued monitoring of obesity-related cancer outcomes is needed, and further research investigating the burden of specific obesity-related cancers is especially warranted to inform the development of effective cancer prevention and control programs.
Supplementary Material
Acknowledgments
The authors have no other financial relationships to disclose.
Footnotes
Conflict of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article.
Supplemental Material
Supplementary data associated with this article can be found in the online version at https: https://doi.org/10.1016/j.jval.2019.07.004.
REFERENCES
- 1.American Cancer Society. Cancer facts & figures. 2017. https://www.cancer.org/research/cancer-facts-statistics/all-cancer-facts-figures/cancer-facts-figures-2017.html. Accessed November 1, 2018.
- 2.Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen M. Body-mass index and incidence of cancer: a systematic review and meta-analysis of prospective observational studies. Lancet 2008;371(9612):569–578. [DOI] [PubMed] [Google Scholar]
- 3.Lauby-Secretan B, Scoccianti C, Loomis D, Grosse Y, Bianchini F, Straif K. Body fatness and cancer — viewpoint of the IARC Working Group. N Engl J Med 2016;375(8):794–798. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Steele CB, Thomas CC, Henley SJ, et al. Vital signs: trends in incidence of cancers associated with overweight and obesity — United States, 2005–2014. MMWR Morb Mortal Wkly Rep 2017;66(39):1052–1058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sung H, Siegel RL, Rosenberg PS, Jemal A. Emerging cancer trends among young adults in the USA: analysis of a population-based cancer registry. Lancet Public Health. 2019;4(3):e137–e147. [DOI] [PubMed] [Google Scholar]
- 6.Siegel RL, Miller KD, Jemal A. Cancer statistics, 2018. CA Cancer J Clin 2018;68(1):7–30. [DOI] [PubMed] [Google Scholar]
- 7.American Cancer Society Action Network. The costs of cancer. https://www.acscan.org/sites/default/files/CostsofCancer-FinalWeb.pdf. Accessed April 13, 2018.
- 8.Guy GP, Ekwueme DU, Yabroff KR, et al. Economic burden of cancer survivorship among adults in the United States. J Clin Oncol 2013;31(30):3749–3757. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lorgelly PK, Neri M. Survivorship burden for individuals, households and society: estimates and methodology. J Cancer Policy. 2018;15:113–117. [Google Scholar]
- 10.Hales CM, Fryar CD, Carroll MD, Freedman DS, Ogden CL. Trends in obesity and severe obesity prevalence in US youth and adults by sex and age, 2007–2008 to 2015–2016. JAMA 2018;319(16):1723. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Basen-Engquist K, Alfano CM, Maitin-Shepard M, et al. Agenda for translating physical activity, nutrition, and weight management interventions for cancer survivors into clinical and community practice. Obesity. 2017;25:S9–S22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Cohen JW, Cohen SB, Banthin JS. The Medical Expenditure Panel Survey: a national information resource to support healthcare cost research and inform policy and practice. Med Care. 2009;47(7 suppl 1):S44–S50. [DOI] [PubMed] [Google Scholar]
- 13.Bureau of Labor Statistics. Medical care in Consumer Price Index. http://data.bls.gov/timeseries/CUUR0000SAM?output_view=pct_12mths. Accessed October 18, 2018.
- 14.Agency for Healthcare Research and Quality. MEPS topics: medical expenditures. https://meps.ahrq.gov/mepsweb/data_stats/MEPS_topics.jsp?topicid=33Z-1. Accessed November 4, 2016.
- 15.Deb P, Norton EC. Modeling health care expenditures and use. Annu Rev Public Health. 2018;39(1):489–505. [DOI] [PubMed] [Google Scholar]
- 16.Agency for Healthcare Research and Quality. MEPS topics: priority conditions. https://meps.ahrq.gov/data_stats/MEPS_topics.jsp. Accessed October 15, 2018.
- 17.The GBD 2015 Obesity Collaborators. Health effects of overweight and obesity in 195 countries over 25 years. N Engl J Med 2017;377(1):13–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.The Emerging Risk Factors Collaboration. Separate and combined associations of body-mass index and abdominal adiposity with cardiovascular disease: collaborative analysis of 58 prospective studies. Lancet 2011;377(9771):1085–1095. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Belotti F, Deb P, Manning WG, Norton EC. twopm: Two-part models. Stata J 2015;15(1):3–20. [Google Scholar]
- 20.American Cancer Society. Cancer facts & figures 2018. https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-statistics/annual-cancer-facts-and-figures/2018/cancer-facts-and-figures-2018.pdf. Accessed February 2, 2019.
- 21.Chastek B, Harley C, Kallich J, Newcomer L, Paoli CJ, Teitelbaum AH. Health care costs for patients with cancer at the end of life. J Oncol Pract 2012;8(6S):75s–80s. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Harris JA, Byhoff E, Perumalswami CR, Langa KM, Wright AA, Griggs JJ. The relationship of obesity to hospice use and expenditures. Ann Intern Med 2017;166(6):381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Fearon K, Arends J, Baracos V. Understanding the mechanisms and treatment options in cancer cachexia. Nat Rev Clin Oncol 2013;10(2):90–99. [DOI] [PubMed] [Google Scholar]
- 24.Flegal KM, Graubard BI, Williamson DF, Gail MH. Cause-specific excess deaths associated with underweight, overweight, and obesity. JAMA 2007;298(17):2028. [DOI] [PubMed] [Google Scholar]
- 25.Tsang NM, Pai PC, Chuang CC, et al. Overweight and obesity predict better overall survival rates in cancer patients with distant metastases. Cancer Med 2016;5(4):665–675. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Grabowski DC, Campbell CM, Ellis JE. Obesity and mortality in elderly nursing home residents. J Gerontol A Biol Sci Med Sci 2005;60(9):1184–1189. [DOI] [PubMed] [Google Scholar]
- 27.von Haehling S, Anker SD. Cachexia as a major underestimated and unmet medical need: facts and numbers. J Cachexia Sarcopenia Muscle. 2010;1(1):1–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Del Fabbro E, Inui A, Strasser F. Cancer Cachexia. Tarporley, UK: Springer Healthcare Ltd; 2012. [Google Scholar]
- 29.Demark-Wahnefried W, Platz EA, Ligibel JA, et al. The role of obesity in cancer survival and recurrence. Cancer Epidemiol Biomarkers Prev 2012;21(8):1244–1259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Siegel RL, Jemal A, Ward EM. Increase in incidence of colorectal cancer among young men and women in the United States. Cancer Epidemiol Biomarkers Prev 2009;18(6):1695–1698. [DOI] [PubMed] [Google Scholar]
- 31.Bailey CE, Hu CY, You YN, et al. Increasing disparities in the age-related incidences of colon and rectal cancers in the United States, 1975–2010. JAMA Surg 2015;150(1):17–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Fedewa SA, Goodman M, Flanders WD, et al. Elimination of cost-sharing and receipt of screening for colorectal and breast cancer. Cancer. 2015;121(18):3272–3280. [DOI] [PubMed] [Google Scholar]
- 33.American Cancer Society. Breast Cancer Facts & Figures 2015–2016. Atlanta, GA; 2015. [Google Scholar]
- 34.Jain R, Denlinger CS. Incorporating weight management into clinical care for cancer survivors: challenges, opportunities, and future directions. Obesity. 2017;25:S27–S29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Yabroff KR, Davis WW, Lamont EB, et al. Patient time costs associated with cancer care. J Natl Cancer Inst 2007;99(1):14–23. [DOI] [PubMed] [Google Scholar]
- 36.Mariotto AB, Robin Yabroff K, Shao Y, Feuer EJ, Brown ML. Projections of the cost of cancer care in the United States: 2010–2020. J Natl Cancer Inst 2011;103(2):117–128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Deshmukh AA, Zhao H, Franzini L, et al. Total lifetime and cancer-related costs for elderly patients diagnosed with anal cancer in the United States. Am J Clin Oncol Cancer Clin Trials. 2018;41(2):121–127. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.