Healthcare expenditures associated with cardiovascular, chronic kidney, and diabetic conditions, 2021–2023 Authors: Devki Patel, Holly J Kramer, Abigail Silva, Fares Qeadan, Talar W Markossian Published: 2026 ABSTRACT AbstractBackgroundThe rising burden of Type 2 diabetes mellitus (DM), heart disease, and chronic kidney disease (CKD) is associated with substantial healthcare expenditures among U.S. adults. We described total healthcare, out-of-pocket (OOP) healthcare, and OOP prescription medication expenditure among individuals with DM, heart disease, CKD, or ≥ 2 of these conditions.MethodsWe pooled data from the 2021, 2022, and 2023 Medical Expenditure Panel Survey (MEPS). DM, heart disease, and CKD were identified using ICD-10 codes in the Medical Conditions file. The analytic sample included adults with DM only (n = 4805), heart disease only (n = 3229), CKD only (n = 44), or ≥ 2 of these conditions (n = 1423). Generalized linear and two-part models were used to estimate the predicted mean annual total healthcare expenditure and OOP prescription medication expenditure associated with these conditions, adjusting for participants’ socio-demographic characteristics.ResultsParticipants with DM only, heart disease only, and CKD only represented 50.57%, 33.99%, and 0.46% of the sample, respectively; 14.98% had ≥ 2 conditions. Approximately half of the participants were female (~ 49.61%), about half were ≥ 65 years (60.18%), most were non-Hispanic White (62.49%), and nearly two-fifths lived in the South (39.17%). Adjusted mean OOP prescription medication expenditure was lowest among adults with CKD only ($341.87; 95% CI: $184.12 - $499.61) and highest among those with ≥ 2 conditions ($679.01; 95% CI: $584.21 - $773.95). However, CKD estimates should be interpreted with caution because of uncertainty due to low sample size. Adjusted mean total healthcare expenditure was lowest among adults with DM only ($15,481.31, 95% CI: 14,545.56–16,417.09) and highest among those with ≥ 2 conditions only ($30,618.53, 95% CI: 27,900–33,336.78 3).ConclusionsMultimorbidity (≥ 2 conditions) is a growing problem in the U.S. and was associated with the highest annual total healthcare and OOP prescription medication expenditure among U.S. adults, highlighting the need for targeted financial and clinical interventions for this high-burden group.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12882-026-05086-7.Keywords: Chronic conditions, Expenditure, Out-of-pocket, Multimorbidity Introduction Over the past few years, we have seen a drastic rise in Americans diagnosed with diabetes, chronic kidney disease (CKD), and cardiovascular disease [1, 2]. These conditions share overlapping pathophysiology, including insulin resistance, hypertension, and dyslipidemia. In the present study, we examine healthcare expenditures associated with these three major clinical conditions, namely Type 2 diabetes mellitus (DM), heart disease, and CKD, and consider each condition alone and in combination. Approximately one in four have been diagnosed with at least one of these listed chronic conditions [3]. Due to the aging of the U.S. population within a background of an unabated obesity epidemic, the prevalence of diabetes, CKD, or heart disease-related multimorbidity is predicted to increase. By 2030, almost one-third of U.S. adults are predicted to have at least one of these conditions [3, 4]. The rising prevalence and overlap of these chronic conditions may result in escalating healthcare costs [5–9]. In the U.S., heart disease alone is accountable for more than $400 billion dollars in direct and indirect healthcare costs, with that number projected to reach $1 trillion by 2035 [10]. Additionally, annual healthcare costs associated with diabetes have increased by $80 billion just within the past decade [1]. Although kidney failure is rare, the per-person costs for kidney replacement therapy exceed $90,000 per year [11]. CKD also contributes a disproportionately large amount to healthcare spending. In 2021, 13.5% of Medicare fee-for-service (FFS) beneficiaries were above 66 years and were diagnosed with CKD, yet they accounted for 25% of the total expenditures [11]. The biggest increase was seen in Part D prescription spending, which increased by 138% between 2011 and 2021 [11]. CKD costs rise as the disease progresses. One Veterans Affairs study found healthcare expenditures for veterans diagnosed with diabetic kidney disease between January 2016 and March 2022 ranged from $1,597 per month for stage 1 CKD to $7,000 by stage 5 [5]. Multiple new medication classes are now available to treat diabetes, CKD, and heart disease conditions, but these medications may not be accessible to all patients due to costs. For some patients with these conditions, multiple medications may be needed but may be cost-prohibitive. The risk of CKD among those with diabetes and heart disease thus increases the potential financial burden on many U.S. adults [6, 12]. While healthcare expenditures for each of the individual conditions have been described, previous studies have not examined out-of-pocket (OOP) costs and OOP burden specifically for coinsurance, deductibles, and prescription medications across combinations of conditions [13–15]. In previous work, CKD alone was associated with high OOP healthcare costs in 2014 [15]. Therefore, our objective was to quantify and compare annual total healthcare expenditure, OOP expenditures, OOP burden, and OOP prescription medication costs among U.S. adults with DM, heart disease, CKD only, and two or more conditions. We hypothesize that OOP cost burden would be highest among adults with two or more conditions (DM, heart disease, and/or CKD). Methods Study design This cross-sectional study pooled 2021, 2022, and 2023 full-year Household Component data files (HC) and the Medical Conditions files (MC) from the Medical Expenditure Panel Survey (MEPS) published by the Agency for Healthcare Research and Quality. MEPS is a national data source that collects data about healthcare use and expenditure among the noninstitutionalized American population. MEPS data are weighted to generate estimates of expenditures representative of the noninstitutionalized American population. In our study, MEPS was used to identify chronic conditions and their association with healthcare expenditures among U.S. adults diagnosed with type 2 diabetes, CKD, heart disease, or a combination of at least two of these conditions. Chronic conditions Presence of heart disease included a clinician diagnosis of coronary heart disease, angina, or other heart disease based on specific ICD-10 codes from the MC file (Table 1). Presence of Type 2 DM (excluding gestational diabetes) and CKD diagnosis was categorized using any observations reporting ICD-10 codes “E11” and “N18” respectively from the Medical Conditions file (Table 1). Table 1Sources of population data using MEPSConditionaSourceICD-10-CM CodesDefinitionDiabetesMedical Conditions FileE11Type 2 diabetes mellitusHeart disease Medical Conditions FileI20, I21, I25, I26, I34, I38, I48, I49, I50, I51Angina, Chronic heart disease, and Other (Acute myocardial infarction, Pulmonary embolism, Nonrheumatic mitral valve disorders, Endocarditis, Atrial fibrillation and flutter, Other cardiac arrhythmias, Heart failure, Complications and ill-defined heart disease)Chronic Kidney Disease (CKD)Medical Conditions FileN18Chronic Kidney Disease (CKD)a = All participants were 18 years or older Sources of population data using MEPS a = All participants were 18 years or older For this analysis, we operationalized conditions as the presence of clinician-diagnosed DM, heart disease, or CKD (excluding N19 “Unspecified kidney failure conditions”). Participants were categorized into four mutually exclusive groups including DM only, heart disease only, CKD only, or ≥ 2 of these conditions (Fig. 1). Although the composition of ≥ 2 conditions subgroups (DM only + heart disease, heart disease + CKD only, DM only + CKD only, and all three) were reported, there was no separate modelling performed for these subgroups because of small CKD-containing sample sizes and unstable estimates. Participants were excluded if they were younger than 18 years or had missing or unknown responses for diagnosis, insurance type, poverty status, race, ethnicity, marital status, region of residence, perceived health or sex. A total of 69,686 people participated in the pooled MEPS survey, and our final analytical sample included 9,501 participants after applying these exclusion criteria (Fig. 2). Overlap of Diabetes, Heart Disease, and CKD in the Study Population, 2021–2023. This figure shows the composition of the chronic conditions in the sample. The overlapping areas represent subgroups of those in the ≥ 2 conditions. The sample contains 1328 (13.98%) with DM only + heart disease, 33 (0.35%) with heart disease + CKD only, 31 (0.33%) with DM only + CKD only, and 31 (0.33%) with all three. These subgroups were aggregated into ≥ 2 conditions category for analysis Study flow chart Demographic characteristics Demographic variables were obtained from three MEPS panels with multiple interview rounds each year of 2021, 2022, and 2023. Age was self-reported as the age of the participant at the end of the year. Race and ethnicity were self-reported and categorized as Hispanic, non-Hispanic Black, non-Hispanic other, and non-Hispanic White. Sex was self-reported as male and female. Region of residence was grouped as Northeast, Midwest, West, and South. Perceived health was self-rated from a survey question asking respondents to rate general physical health as excellent, very good, good, fair, and poor. Perceived health was included in descriptive analysis but not in regression analysis. Marital status was classified as married and not married. Poverty status was determined by calculating each participant’s family income over the federal poverty level (FPL) for that household size and expressed as a percentage. The FPL is calculated based on the information given on family size and composition using data from the Current Population Survey (CPS). Individuals were then classified into one of five categories: poor (less than 100% FPL), low income (100% to less than 200% FPL), middle income (200% to less than 400% FPL), and high income (400% FPL or greater). Those without a CPS family identifier were treated as one-person families for classification. Insurance coverage was categorized by MEPS into three groups: private, public, and uninsured. Participants reporting private insurance, including TRICARE/CHAMPVA, at any point were classified as “Any Private”. Participants with public insurance coverage (excluding TRICARE/CHAMPVA) throughout the year were placed in “Public only”. Healthcare expenditures OOP healthcare expenditure was self-reported payments for coinsurance and deductibles, and cash outlays for services, supplies, and other items not covered by health insurance. Total OOP costs for prescription medications captured the sum of OOP costs for all prescription fills and refills obtained in the calendar year. Total healthcare expenditure in MEPS was the total sum of payments during the calendar year, which included expenditures made by insurance carriers and OOP payments. The expenditure data were collected for the following health care services: non-physician visits, dental and vision services, other medical equipment and services, and home health care not provided by an agency, office-based visits to physicians, hospital-based events (inpatient stays, outpatient, and emergency room visits), and prescribed medicines. Expenditures were inflated to 2023 U.S. $ (latest data year) using the Consumer Price Index for medical care (CPI-M) available from the Bureau of Labor Statistics and the MEPS website [16, 17]. Out-of-pocket burden OOP healthcare burden was computed using the ratio of OOP spending over personal income and reported as a percentage ranging from 0 to 100% [18]. Observations of individuals who reported zero for personal income were not included in these calculations (n = 902 observations) and all percentage burdens over 100 were truncated to 100. Sensitivity analysis of OOP burden To test the robustness of the OOP burden estimates after truncation and excluding those with zero for personal income, a sensitivity analysis was performed using an alternative method to calculate OOP burden. In a separate sensitivity analysis, observations of individuals who reported zero personal income but had a non-zero OOP healthcare expenditure (n = 418) were included and assigned an OOP burden of 100% to bypass the division-by-zero calculation error and still have financially vulnerable populations in the analysis. Unadjusted mean and median estimates of OOP burden derived from this method were also reported in our results. Statistical analysis Descriptive and bivariate analyses were performed using survey-weighted procedures to account for the complex MEPS design. We used PROC SURVEYFREQ to estimate weighted frequencies and percentages and to conduct Rao-Scott χ² tests for differences in categorical variables across the four condition groups (DM only, heart disease only, CKD only, ≥ 2 conditions). For continuous variables (e.g., expenditures and income), we used PROC SURVEYMEANS to obtain survey-weighted means and standard errors and PROC SURVEYREG to test unadjusted mean differences across condition groups. We also report medians and interquartile ranges to reflect the skewed distribution of cost data. The generalized linear model (GLM) was used to estimate mean total expenditures, and the two-part model [17] was used to estimate out-of-pocket medications (Rx) expenditures for each diagnosis group. These models were built in Stata, and adjusted for participant characteristics including sex, age, race, marital status, poverty level, insurance, region of residence, as well as complex survey design. GLM was used to model total expenditures since only a few (n = 11) observations had 0 total expenditures. The two-part model was used for out-of-pocket Rx expenditures because there were some observations with zero expenditures in the data, although these were uncommon, they may still exist. These models are included in the supplementary material. All models incorporated the MEPS design variables (PERWT as the person-level weight, VARSTR as the variance estimation stratum, and VARPSU as the primary sampling unit). Unadjusted descriptive analyses (Tables 2 and 3) were conducted in SAS OnDemand for Academics (version 9.4; SAS Institute Inc., Cary, NC) using PROC SURVEYFREQ and PROC SURVEYREG. Adjusted analyses (Table 4) were conducted in Stata (version 18; StataCorp LLC., College Station, TX). All tests were two-sided with a significance level of 0.05. Table 2Demographics and health characteristics overall and by the cardiovascular, kidney, and diabetic conditions, 2021–2023Totaln(%)**DM only n(%)**Heart onlyn(%)**CKD only n(%)**≥ 2 conditions n(%)**P-value Total 9,501 (100)4,805 (50.57)3,229 (33.99)44 (0.46)1,423(14.98) Age** 18–2455 (0.58)35 (1.17)20 (1.33)0 (0.00)0 (0.00)< 0.0001*25–643,633 (38.24)2,381 (54.85)871 (31.33)8 (22.72)373 (28.17)65+5,813 (61.18)2,389 (43.98)2,338 (67.34)36 (77.28)1,050 (71.83) Marital status Married4,382 (46.13)2,245 (54.09)1,530 (52.91)16 (44.38)591 (48.14)0.0385Not Married5,119 (53.87)2,560 (45.91)1,699 (47.09)28 (55.62)832 (51.86) Race Hispanic1,337 (14.07)959 (16.76)220 (6.73)0 (0.00)158 (9.74)< 0.0001*Non-Hispanic White5,937 (62.49)2,484 (56.74)2,527 (80.06)34 (71.72)892 (67.61)Non-Hispanic Black1,553 (16.35)942 (15.99)328 (7.72)9 (18.14)274 (14.10)Non-Hispanic Other or Multi674 (7.09)420 (10.51)154 (5.49)1 (10.14)99 (8.55) % Male Male4,698 (50.39)2,220 (50.39)1,689 (53.18)14 (27.73)775 (59.28)0.0001* Region Northeast1,596 (16.81)709 (14.97)655 (21.12)2 (4.70)230 (18.13)Midwest2,099 (22.09)992 (20.77)747 (22.77)20 (44.89)340 (22.99)< 0.0001*South3,721 (39.17)2,016 (43.19)1,120 (37.08)13 (34.05)572 (40.69)West2,085 (21.93)1,088 (21.07)707 (19.03)9 (16.36)281 (18.19) Family income status (poverty) a High income1,472 (10.26)793 (10.26)403 (8.38)0 (0.00)276 (11.54)Middle income2,116 (19.11)1,094 (19.11)654 (18.61)10 (18.75)358 (22.79)< 0.0001*Low income2,667 (30.90)1,392 (30.90)875 (28.66)17 (46.19)383 (29.26)Poor3,246 (39.73)1,526 (39.73)1,297 (44.36)17 (35.06)406 (36.41) Insurance Any private4,628 (58.38)2,431 (58.38)1,622 (54.45)27 (64.29)548 (44.34)Public only4,701 (38.86)2,234 (38.86)1,581 (44.53)17 (35.71)869 (55.17)0.0002*Uninsured172 (2.75)140 (2.75)26 (1.02)0 (0.00)6 (0.50) Perceived Health Excellent646 (5.64)288 (5.64)310 (10.02)3 (4.04)45 (2.89)Very Good2,340 (24.60)1,147 (24.60)935 (29.74)14 (28.82)244 (18.88)< 0.0001*Good3,595 (41.32)1,932 (41.32)1,147 (36.07)12 (25.37)504 (34.90)Fair2,184 (21.66)1,108 (21.66)633 (18.18)9 (27.92)434 (28.71)Poor736 (6.78)330 (6.78)204 (5.99)6 (13.86)196 (14.63)* p < 0.05 (statistically significant)** Demographics are listed as unweighted N and weighted column percentagesa Poor (less than 100%), low income (100% to less than 200%), middle income (200% to less than 400%), and high income (400% or greater) Demographics and health characteristics overall and by the cardiovascular, kidney, and diabetic conditions, 2021–2023 * p < 0.05 (statistically significant) ** Demographics are listed as unweighted N and weighted column percentages a Poor (less than 100%), low income (100% to less than 200%), middle income (200% to less than 400%), and high income (400% or greater) Table 3Unadjusted healthcare spending by condition group (Mean and Median, 2021–2023) in 2023 U.S. $Diabetes only (n = 4,805)Heart Disease only (n = 3,229)CKD onlya (n = 44)≥ 2 conditions (n = 1,423)Total (n = 9,501)P-valueOut-of-Pocket Healthcare ExpenditureMean (SE)1,638.18 (102.35)2,003.97 (89.15)1,459.01 (110.63)2,300.28 (181.98)1,852.17 (70.17)0.0116*Median (IQR)580.94 (151.24, 1595.09)871.38 (277.68, 2118.83)817.63 (371.69, 1862.05)1036.09 (287.83, 2550.43)723.52 (201.93, 1935.10)Out-of-Pocket Prescription ExpenditureMean (SE)435.30 (28.70)463.45 (32.92)363.68 (15.53)683.23 (47.19)478.38 (20.85)0.0240*Median (IQR)158.99 (39.08, 469.22)155.31 (48.20, 436.60)222.38 (60.46, 443.17)305.20 (83.16, 801.28)168.90 (47.68, 502.02)Total Health Expenditure a, (all sources of payment)Mean (SE)14,902 (447.50)19,868 (757.72)16,593 (2,069.92)30,953 (1,306.70)18,789 (395.02)< 0.0001*Median (IQR)8,095.75 (2,609.93, 18,170.00)9,172.29 (3,985.75, 20,997.00)8,455.89 (3,551.19, 17,795.00)17,752 (8,479.70, 39,137.00)9,529.88 (3,495.31, 21,256.00)Personal IncomeMean (SE)49,427 (1,201.27)54,198 (1,566.42)44,574 (1,196.42)43,087 (1,679.69)50,164 (938.92)0.0337*Median (IQR)35,798 (17,719, 65,136)36,377 (18,952, 69,869)32,240 (21,800, 49,902)28,296 (16066, 53,237)34,994 (17,991, 65,091)Out-of-Pocket BurdenbcMean (SE)5.48% (0.25)6.54% (0.31)5.35% (1.58)8.14% (0.44)6.21% (0.20)< 0.0001*Median (IQR)0.72% (0, 3.58)1.38% (0, 5.18%)1.83% (0, 5.31%)1.99% (0 7.47%)1.11% (0, 4.62%)Out-of-Pocket Burden (sensitivity analysis) dMean (SE)16.08% (5.71)28.64% (11.14)6.15% (1.21)22.02% (8.48)25.95% (6.70)0.5005Median (IQR)0.72% (0, 3.58)1.38% (0, 5.18%)1.83% (0, 5.31%)1.99% (0 7.47%)1.20% (0, 4.92%)* p < 0.05 (statistically significant)a Total expenditure includes prescription expenditure and includes payments from all sources: out-of-pocket, Medicare, Medicaid, private insurance, VA/CHAMPVA, TRICARE, workers compensation, other federal, state, local sources, and other miscellaneous sourcesb Out-of-Pocket Burden was calculated as a ratio of out-of-pocket spending over personal income and reported as a percentage ranging from 0-100%c All individuals reporting personal income of zero were excluded from the analysisd Individuals reporting personal income of zero but a non-zero OOP healthcare expenditure above zero was assigned 100% OOP burden and included in the analyses Unadjusted healthcare spending by condition group (Mean and Median, 2021–2023) in 2023 U.S. $ * p < 0.05 (statistically significant) a Total expenditure includes prescription expenditure and includes payments from all sources: out-of-pocket, Medicare, Medicaid, private insurance, VA/CHAMPVA, TRICARE, workers compensation, other federal, state, local sources, and other miscellaneous sources b Out-of-Pocket Burden was calculated as a ratio of out-of-pocket spending over personal income and reported as a percentage ranging from 0-100% c All individuals reporting personal income of zero were excluded from the analysis d Individuals reporting personal income of zero but a non-zero OOP healthcare expenditure above zero was assigned 100% OOP burden and included in the analyses Table 4Adjusted out-of-pocket prescription and total health expenditure by chronic condition in 2023 U.S. $VariablesAdjusteda out-of-pocket prescription expenditureAdjusteda total expenditures* (yearly)Mean (95% CI) **Mean (95% CI) ** Diagnoses DM only458.27 (415.68–500.86)15,481.32 (14,545.56–16,417.09)Heart Disease only432.06 (368.96–495.16)18,953.72 (17,512.22–20,395.22)CKD only341.87 (184.11–499.62)16,632.64 (9,403.88–2,3861.40)≥ 2 conditions679.08 (584.21–773.95)30,618.53 (27,900.28–33,336.78)a Model adjusted for participant sex, age, race, marital status, poverty level, payor, and region of residence*Total expenditure includes prescription expenditure and includes payments from all sources: out-of-pocket, Medicare, Medicaid, private insurance, VA/CHAMPVA, TRICARE, workers compensation, other federal, state, local sources, and other miscellaneous sources**Estimates show predicted means adjusted Adjusted out-of-pocket prescription and total health expenditure by chronic condition in 2023 U.S. $ a Model adjusted for participant sex, age, race, marital status, poverty level, payor, and region of residence *Total expenditure includes prescription expenditure and includes payments from all sources: out-of-pocket, Medicare, Medicaid, private insurance, VA/CHAMPVA, TRICARE, workers compensation, other federal, state, local sources, and other miscellaneous sources **Estimates show predicted means adjusted Results Demographics Table 2 shows the characteristics of MEPS participants who reported a diagnosis of DM only, heart disease only, CKD only, or at least two of these three chronic conditions. Overall, 50.57% had DM only, 33.99% had heart disease, 0.46% had CKD only, and 14.98% had ≥ 2 conditions. Among those with ≥ 2 conditions, 1,328 (13.98%) had DM only + heart disease, 33 (0.35%) had heart disease + CKD only, 31 (0.33%) had DM only + CKD only, and 31 (0.33%) had all three. About half of the sample was ≥ 65 years of age (61.18%), male (50.39%), married (46.13%), non-Hispanic White (62.49%), and had private insurance (58.38%). Over one-third of the sample lived in the South (39.17%). Perceived health status was most commonly reported as “good” (41.32%). Age, race/ethnicity, poverty status, insurance type, region of residence, and perceived health were significantly associated with the condition category (p < 0.05). Mean and median yearly healthcare spending across chronic conditions The lowest unadjusted mean OOP healthcare expenditure was among individuals with CKD only and DM only at $1,459.01 and $1,638.18, and the highest mean OOP healthcare expenditure was among individuals with two or more conditions at $2,300.38. Unadjusted mean OOP annual prescription expenditure was lowest among CKD only and DM only at $363.68 and $435.30, followed by heart disease at $381.66, and highest among individuals with at least ≥ 2 conditions at $478.38. DM only showed the lowest unadjusted mean total health expenditure with $14,902, while the presence of ≥ 2 conditions had roughly two-fold higher total health expenditures at $30,953, respectively (Table 3). Importantly, personal income was lowest among those with ≥ 2 conditions ($43,087) and highest among those with heart disease alone ($54,198) (Fig. 3). Distribution of unadjusted mean personal income and total healthcare expenditure in 2023 U.S. $ DM only and CKD only showed the lowest OOP mean burden at 5.48% and 5.35%, while ≥ 2 conditions showed the highest burden with a substantially higher percentage at 8.14%. Table 3 also shows the medians of the expenditure variables and expresses similar trends outlined above for the mean. Sensitivity analysis of OOP burden In the sensitivity analysis, mean OOP burden substantially increased across all conditions. In this group of individuals, heart disease alone showed the highest mean OOP burden with 28.64%, while DM only and CKD only remained comparatively the lowest (16.08% vs. 6.15%). Median OOP burdens remained unchanged for every condition. Despite the changes in the magnitude of OOP burden, the results were not statistically significant in the sensitivity analysis (p = 0.5005). The observed differences in OOP burden across conditions may be sensitive to the treatment of individuals with zero personal income but non-zero OOP healthcare expenditure in the analysis. Predicted means for yearly out-of-pocket and total healthcare expenditure by chronic conditions The CKD only group reported the lowest mean yearly out-of-pocket prescription expenditure at 341.87 (95% CI: 184.11–499.62) and those with ≥ 2 conditions showed the highest at 679.08 (584.21–773.95) (Table 4). Individuals who reported ≥ 2 conditions had the highest predicted mean costs for total expenditure ($30,618.53, 95% CI: 27,900.28–33,336.78) and DM only group showed the lowest ($15,481.32 95% CI: 14,545.56–16,417.09) (Table 4). Discussion In this nationally representative study of U.S. adults with DM, heart disease, CKD, or combinations of these conditions, having ≥ 2 conditions were associated with roughly two-fold higher total healthcare expenditure than having each condition alone. Comorbidity has previously shown higher expenditure, due to increased hospitalizations, prescription medications, and multiple emergency visits [7, 19]. In a large number of individuals in our sample, these conditions occurred simultaneously, particularly among patients with CKD. Most patients with CKD in our study had a comorbid condition (95 individuals versus 44 individuals with CKD alone). Nichols et al. showed that cardiovascular disease (CVD) was the most common cause for hospitalization among individuals with CKD, emphasizing the inevitable increase in healthcare costs among those with multiple conditions. One study reported an increase in adjusted annualized healthcare costs by $6,789 for diabetes, $21,573 for CVD, and $10,316 for CKD [7]. These results are consistent with our findings and may indicate that cardiovascular diseases substantially contribute to the total healthcare expenditures between these related conditions. CKD severity is also associated with high healthcare costs with the costs increasing sevenfold by stage 5, which may reflect greater resource utilization [11, 20, 21]. CKD patients also face many challenges with timely diagnosis as many clinicians fail to recognize early stages, which may result in preventable disease progression and contribute to higher costs, especially due to increased hospitalizations among later stage CKD patients [22]. Additionally, among the individual condition groups, DM only was associated with the highest adjusted OOP prescription medication expenditure. A possible explanation for lower prescription costs observed in CKD-only and heart disease groups could be due to the availability of generic CKD and heart disease medications compared to the new influx of brand-name diabetes medications. Patients with CVD experienced an increase in prescription-related medical expenditure after developing diabetes, suggesting prescription cost burden stemming from diabetes in comorbid conditions [9]. The American Heart Association also reported Medicare Part D beneficiaries spent $11 billion on generic and brand-name medications for cardiovascular disease. However, patent exclusivity from pharmaceutical companies has made it more difficult for generic diabetes medications to flood the market [23]. In general, the medications prescribed for any of these conditions have high OOP costs. Young et al. reported that OOP costs for cardiovascular-kidney-metabolic medications among those with Medicare coverage still exceeds $1500 despite the availability of various therapies for these conditions [24]. This helps explain how the nature of these conditions is associated with having high OOP Rx costs, even with coverage, with costs increasing further as comorbidities arise. It was also noteworthy that individuals with ≥ 2 conditions showed the lowest mean personal income yet the highest total healthcare expenditure. This pattern is consistent with prior literature on how individuals with a lower socioeconomic status (SES) are disproportionately affected by certain chronic conditions. Numerous studies note that many chronic conditions are more prevalent among low socioeconomic groups [25–29]. For instance, one study found that individuals with a family income to poverty ratio less than 300% had 125% higher risk of having multiple cardio-kidney-diabetic conditions compared to those with higher income ratio [30]. Limited access to healthcare services due to financial barriers can also contribute to the mismanagement of chronic disease diagnoses, potentially risking the accumulation of multiple related conditions and higher healthcare utilization in later stages of life [31]. This can suggest that financially disadvantaged populations may represent a particularly vulnerable population with higher prevalence of comorbidities and elevated healthcare expenditure [32]. This study has limitations that may have affected the results of the study. First, the confidence intervals around the CKD values were large because of a small sample of individuals with CKD only in the study which may have reduced statistical power and may not have been representative of the true CKD population [29]. This study acknowledges the CKD only estimates were exploratory since CKD is often undiagnosed, especially within the early stages [33]. The use of ICD-10-based CKD identification in MEPS may not ascertain undiagnosed or early stages of CKD. Thus, the true population living with CKD was likely underestimated, especially among those with diabetes. Additionally, due to the use of publicly available data, MEPS has restricted access to dialysis information due to confidentiality provisions. Therefore, differentiation between CKD stages was not clearly defined. The lack of CKD staging or dialysis status limits the full interpretation of CKD-related expenditure. There were a few extreme values of healthcare expenditure that may have slightly skewed the data; however, the use of the two-part models and reporting medians with interquartile ranges helps mitigate the influence of outliers and skewed distributions. In this study, because of small subgroup sample sizes, we collapsed different combinations of conditions into one overall ≥ 2 condition group. Nonetheless, there were many strengths in the study to highlight the growing financial burden of cardiovascular, kidney, and diabetic comorbidity. For one, MEPS is a nationally representative data source that shows weighted healthcare expenditure estimates that can be applied to the entirety of the U.S. population with the mentioned condition. The use of ICD-10 codes versus self-reported data to determine diagnoses also helped reduce information bias and accurately include conditions that are within the scope of the study. It is also important to note that our findings are specific to the U.S. healthcare system and are not generalizable to other countries due to differing healthcare systems. Multimorbidity (≥ 2 conditions) was linked to the highest yearly healthcare expenditures among U.S. adults with these conditions. As a result, investments in prevention and treatment strategies that reduce the development and clustering of these conditions may lessen both the individual health burden and the overall healthcare expenditures associated with Type 2 DM, heart disease, and CKD. Further studies should research the differences between each combination of conditions and prescription classes separately to provide a more detailed understanding of clinical profiles and expenditures of comorbidity. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements This work was completed as part of Ms. Patel’s Master of Public Health capstone project and would acknowledge the Public Health Sciences Faculty for their support in the program. Abbreviations Diabetes mellitus Chronic kidney disease Medical Expenditure Panel Survey Fee for service Out-of-pocket Household Component data file Medical Conditions file International Classification of Diseases, 10th Edition, Clinical Modification Cardiovascular disease Federal poverty level Medications Current Population Survey Socioeconomic status Author contributions All the authors made substantive intellectual contributions in this study. DP, HK, and TM made substantial contributions to conception and design, acquisition of data, analysis and interpretation of data, drafting the manuscript and revising it critically for important intellectual content. AS and FQ, has made substantial contributions to interpretation of data and revising it critically for important intellectual content. All the authors have given final approval of the version to be published, and they have participated sufficiently in the work to take public responsibility for appropriate portions of the content. Funding No funding was received to conduct this study. Data availability The data used in this study is publicly available for scientists from the Agency for Healthcare Research and Quality’s (AHRQ) Medical Expenditure Panel Survey (MEPS) website upon MEPS data use agreement. We used the household component full-year and medical conditions files for 2022 for these analyses. These datasets are downloadable from the following website: https://meps.ahrq.gov/mepsweb/data_stats/download_data_files.jsp. Declarations Ethics approval and consent to participate This study did not require IRB approval since we used publicly available data from the Agency for Healthcare Research and Quality’s (AHRQ) Medical Expenditure Panel Survey (MEPS). All information for MEPS is collected with assurances of confidentiality, and individual identifiers have been removed from the micro data contained in these files. Nevertheless, by using this data, the authors signified their agreement to comply with the MEPS data use agreement. Therefore, consent to participate was not collected by the authors in this study. The ‘Data Use Agreement’ for MEPS can be found on the following website, section A. MEPS HC 243. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The data used in this study is publicly available for scientists from the Agency for Healthcare Research and Quality’s (AHRQ) Medical Expenditure Panel Survey (MEPS) website upon MEPS data use agreement. We used the household component full-year and medical conditions files for 2022 for these analyses. These datasets are downloadable from the following website: https://meps.ahrq.gov/mepsweb/data_stats/download_data_files.jsp.