Recent evidence suggests that more than half of Americans suffer from 1 or more chronic diseases, with the estimated cost of treating these conditions exceeding $1 trillion annually.1 Unless preventative measures are taken, the rate of chronic diseases is expected to increase annually and with it the cost of treatment. By 2050, it is estimated that treatment costs will exceed $6 trillion.1 California will be significantly impacted by the projected growth of chronic diseases in the United States, as it has the largest population of all US states. Currently, in California, there are an estimated 14 million people living with at least one chronic condition and more than half of this group have multiple chronic conditions.2
The advent of the Affordable Care Act provides an opportunity to address these challenges. As more people are covered by some form of health insurance, there is an opportunity to expand prevention services, especially to vulnerable and at-risk populations. Reducing the rates of chronic conditions and their associated costs will require a concerted and coordinated effort that integrates all levels of government, private providers, and the public to address issues such as poor nutrition, tobacco use, excessive alcohol use, and lack of physical activity.
Information on the cost of chronic conditions to counties can help community members understand the scope of the problems facing the people in their region and identify high priority areas to target interventions and programs. As the primary organization responsible for monitoring the health of the people in the region, local health departments (LHDs) are expected to play an increasingly important role in combating the spread of chronic diseases in the future.3 Accurate estimates of the rates and costs of chronic conditions can help LHDs identify the amount of health care expenditure that could be saved if chronic conditions were prevented or controlled, information that is vital to identifying and planning cost-effective prevention and control efforts.
The goal of a costing study is to report only the costs associated with a condition (the “attributable” or “marginal” costs).4 Estimating the attributable cost of care for a particular condition requires differentiating between those costs that are incurred because of a specific condition and those that are incurred for other reasons. From the standpoint of identifying and planning prevention activities, the attributable costs are more relevant than the total costs since they provide an estimate of the amount of health care expenses that could be saved by avoiding a particular condition. This study combines estimates from the Centers for Disease Control and Prevention's (CDC's) Chronic Disease Cost Calculator,5 with county-level data on rates of 6 chronic conditions and attributable cost of each condition to estimate the burden of chronic conditions in each county.
The methodology involved the following steps:
- Use the CDC's Chronic Disease Cost Calculator to develop estimates of the additional medical expenditure (cost) associated with chronic conditions for the state of California by condition, age group, and gender.
- Develop a cost per case of each of the chronic conditions in each county of California by adjusting for price differences in health care services between counties.
- Estimate the prevalence of chronic disease within each county of California by age, gender, and race/ethnicity. We used 5 age strata (0-17, 18-44, 45-64, 64-79, and ≥80 years), 2 gender categories (male and female), and 5 race/ethnicity categories (Hispanic, non-Hispanic white, non-Hispanic black, non-Hispanic Asian, and non-Hispanic other).
- Combine estimates for the rates of chronic conditions with county-level census population data (by age strata, gender, and ethnicity) to form estimates for the number of cases of chronic conditions in that county.
- Estimate the total additional cost of each chronic condition in each county by multiplying and then adding the cost per case of each chronic condition in that county by the number of cases for each of the 6 chronic conditions in that county.
The research was provided an exemption from the institutional review board at the University of California, Merced.
Cost estimates from the CDC Chronic Disease Cost Calculator
Estimates of the medical payments attributable to each chronic condition were estimated using the CDC Chronic Disease Cost Calculator, Version 2 (referred to as the CDC Cost Calculator).5 The CDC Cost Calculator provides state-level estimates of medical payments for arthritis, asthma, cancer, cardiovascular diseases (CVDs), depression, and diabetes by age group and gender. Because CVD is such a broad category, when estimating CVD, the CDC calculated costs of congestive heart failure (CHF), coronary heart disease (CHD), hypertension, stroke, and other heart diseases separately. The estimates are provided by gender for 5 age bands or strata: (1) 0 to 17, (2) 18 to 44, (3) 45 to 64, (4) 65 to 79, and (5) 80 years or older. All information used in this report is publically available from the CDC, including tables of the cost of each condition by age and gender for California.
Cost estimates from the CDC Cost Calculator include all additional or attributable medical expenditures for the entire state population (all payers and the uninsured) and include estimates of absenteeism. The estimates were derived from the 2004-2008 Medical Expenditure Panel Survey (MEPS) Consolidated Data Files,6 a nationally representative survey of the civilian noninstitutionalized population that provides data on annual medical expenditures, sources of payment, insurance coverage, and days missed from work due to illness or injury for each participant. Diseases were defined using ICD-9 (International Classification of Diseases, Ninth Revision) codes based on self-reported diseases that were transcribed by professional coders and reported in the MEPS Medical Conditions files for years 2004-2008 (see Table 1). The combined 5-year MEPS sample included 153 012 persons of all ages living in the United States. All expenditure data were inflated to 2010 dollars using the gross domestic product general price index.
To account for multiple chronic conditions and avoid double-counting (ie, overlap of disease costs) of the associated medical costs, the CDC Cost Calculator uses estimates from a 2-stage statistical analysis that generates estimates of cost of multiple conditions. The resulting estimates of the costs of arthritis, asthma, cancer, CHF, CHD, stroke, other heart diseases, and depression are thus independent of the other conditions, although the cost estimates for hypertension and diabetes do include the costs of complications such as CHD, CHF, and stroke. The costs of hypertension and diabetes are therefore not mutually exclusive of the costs of other reported diseases. We report the estimated cost of CHF, CHD, stroke, and hypertension separately but report only the combined cost of CVD that includes these 4 conditions (we do not estimate costs for other heart diseases).
Adjustment of cost of chronic conditions by county
The CDC Cost Calculator provides estimates of the average attributable cost for each chronic condition at the state level by age strata and gender. To account for differences between counties in the cost per case due to variations in the price of health care services across the state, prices were adjusted using the Geographic Adjustment Factor reported by the Institute of Medicine and based on the Centers for Medicare & Medicaid Studies Medicare geographic practice cost index for California.7 The Geographic Adjustment Factor takes account of geographic differences due to 3 factors: cost of physician services, practice expenses due to location (eg, rent and cost of operating a facility), and geographic differences in malpractice or professional indemnity. The Geographic Adjustment Factor, which divides California into 9 distinct regions for which geographic practice cost indexes are calculated, was applied to the cost estimate for each condition, age, and gender for each region of California. The cost adjusters, ranging from 1.0323 to 1.1817, were applied to the cost estimates from the CDC Cost Calculator.
Prevalence rates for California
Estimates of the prevalence of the 6 chronic conditions of interest for California were derived from a variety of data sources, including AskCHIS online query system for the California Health Interview Survey8 and SEER (Surveillance Epidemiology and End Results)-Medicare data.9 The rates summarized in Table 1 were derived from 3 sources: AskCHIS,8 SEER-Medicare data,9 and the American Diabetes Association.10 The number of cases was derived by multiplying the prevalence rates by the population per county as obtained from the 2010 Census.11
State-level prevalence rates for (a) arthritis, (b) asthma, (c) CVD, and (d) diabetes were obtained from the 2011 to 2012 California Health Interview Survey (CHIS). The CHIS is a representative population-based, random-dial, health survey of noninstitutionalized individuals in California12 that is used to estimate prevalence rates for various health conditions at the state level and for large- and medium-sized counties at the county level (smaller counties are grouped together).13 State-, regional-, and county-level estimation of various diseases and health-related behaviors surveyed in the CHIS can be obtained from the online Web tool AskCHIS.14 Because the 2011-2012 CHIS sample size was too small to obtain county-level prevalence rates for each age strata by race/ethnicity and gender, the analysis was conducted with state-level prevalence rates (rather than county-level rates).
Because the CHIS does not ask children and teens certain health questions, the prevalence for arthritis, CVDs, and diabetes for individuals aged 0 to 17 years was not available from the AskCHIS query system. We obtained diabetes prevalence rates for individuals younger than 20 years from the American Diabetes Association 2011 National Diabetes Fact Sheet.10 We could not obtain prevalence rates for arthritis or CVD for children or teens, so we set the prevalence at zero in our estimations.
Cancer prevalence for each age strata by race/ethnicity and gender was calculated for each of the 58 California counties using the program SEER Stat and 2009 and 2010 SEER (written communication, January 31, 2014).9 The county-level prevalence rates were estimated for the following strata to match the age groups used by SEER: 0-19, 20-44, 45-64, 65-79, and 80 years or older. Because of small sample sizes, no estimates were available for small counties or small demographic subgroups within counties. Small counties were therefore combined and average rates applied to each small county.
Population per county was calculated from the 2010 Census Summary File 1, Table PCT12, provided by the California State Data Center.11 This Table PCt12 presents the population of California and each Californian county race, gender, age, and Hispanic ethnicity. We used this table to calculate the population for each of the age, gender, and race/ethnicity condition for each of the 58 counties in California.
Compiling the cost estimates
The estimated cost of each chronic condition in each county of California was derived by multiplying:
- Estimate of the attributable cost for the age and gender for each condition (Table 2) adjusted for prices in the county;
- Prevalence of the condition by age group and gender for each county; and
- The number of people in each county by age, gender, and ethnicity.
This procedure resulted in estimates of each of the 6 conditions for each of the 58 counties by age, gender, and ethnicity. These estimates were then combined to calculate a total attributable cost per county. To provide counties with information on the relative importance of chronic diseases to their county, the results also report the percentage of their total health care expenditure that is due to chronic diseases. The total health care expenditure for each county was calculated by multiplying the population in the county by the average estimated health care expenditures for each person in California in 2010: $6238.15
Table 3 summarizes the results for California. Overall, there are more than 28 million cases of chronic conditions in California, with the most common being CVDs (36.4%), followed by arthritis (19.4%), asthma (14.2%), and depression (11.7%). Cancer is the least common condition (3.3%), with diabetes also being relatively infrequent. These 28 million cases do not suggest that 28 million people have a chronic condition, since people can have more than 1 chronic condition. On the basis of previous studies suggesting that 14 million (39% of the population) Californians have at least one chronic condition,16 the estimates reported here suggest that these 14 million people have an average of nearly 2 chronic conditions (28 million conditions divided by 14 million people with a chronic condition).
The total cost for each condition and the percentage of the cost to overall health care expenditures in California are also shown in Table 3. Befitting its prevalence in the state, CVD is associated with the greatest expense, with more than $37 billion spent annually, or 16% of all health care costs in California. Despite its prevalence (14.2%), asthma has the lowest overall cost, contributing approximately 4% to the total health care expenditure. And cancer, despite its relatively low prevalence, is similar to the total health care expenditures of other conditions ($13 billion, or 6% of total costs). Overall, the additional cost of treating chronic conditions is $98 million or approximately 42% of all health care expenditures in California.
The cost differences between counties (Table 4) reflect their relative populations, ethnicities, price adjustments, and rates of chronic conditions, with total costs varying from a high of $23 billion for Los Angeles County, $8 billion for Orange and San Diego counties, and $5 billion for Riverside County to $3.5 million for Alpine County and $12 million for Sierra County. Relative to the estimated total cost of health care in the county ($6238 per person in 2010),11 counties with the estimated lowest spending on chronic conditions tend to be those with the youngest populations, including Kings and Kern counties (32% and 34% of total, respectively) whereas the counties with the highest spending have an older age distribution, including Amador (62%), Marin (61%, and Tuolumne (63%) counties.
The purpose of this report was to estimate the cost of chronic conditions for each county in California. We combined estimates from the CDC Cost Calculator of the cost per person for each of 6 chronic conditions—arthritis, asthma, cancer, CVD, diabetes, and depression—with prevalence rates and census data from counties in California to develop estimates of the number of cases in each county and the cost associated with these conditions. The results suggest that approximately $98 billion is currently spent on treating chronic conditions in California. This represents approximately 42% of all health care expenditure in the state.
Previous studies have estimated the cost of chronic conditions at the national level,17–20 with some estimates also available for individual states. However, while individual counties have attempted to estimate the cost of chronic conditions in their region (eg, San Diego County21), this is the first attempt to estimate the cost of chronic conditions for multiple counties in a state. Because the CDC Cost Calculator provides estimates of the attributable or additional cost of the chronic conditions, the estimates represent the amount that could be saved if chronic diseases were eliminated or reduced.
Understanding the savings that could be made through enhanced prevention activities is important for LHD planning. Under the Affordable Care Act, many previously uninsured people now have access to affordable prevention services. This provides a unique opportunity for LHDs to reduce the future burden of chronic conditions by investing in or helping people gain access to effective prevention activities. But because budgets are likely to remain limited, LHDs must identify prevention activities that are “worth the money,” meaning that they need to identify whether the health outcomes from a prevention activity justify the investment they will have to make. For instance, since 2007, the CDC has made a program called PRISM (Prevention Impacts Simulation Model) available to help local health officials understand the health and cost outcomes from a number of interventions, including medical care (14 separate interventions), smoking (n = 5), nutrition and weight loss (n = 8), physical activity (n = 4), emotional distress (n = 2), and particulate air pollution (n = 1).22 The diseases and conditions modeled in detail include heart disease, stroke, diabetes, hypertension, high cholesterol level, and obesity, and the model also accounts for cancers and respiratory diseases related to smoking, obesity, poor nutrition, and physical inactivity. Combining this type of information with the cost of the condition can provide LHDs with an understanding of their return on investment from possible prevention activities.
The estimates derived in this study differ somewhat from the estimates from the CDC regarding the cost of chronic conditions. As summarized in Table 5, the overall cost estimate of $98 billion is approximately 14% higher than the CDC's estimate of $87 billion for California. While the estimated cost of each chronic condition was the same (Table 2), the prices were adjusted to account for differences between counties (Table 4). Most significantly, the prevalence rates used in the current study differed significantly: the CDC estimates relying on prevalence rates from the MEPS compared with the use of AskCHIS in this study. The California-specific AskCHIS rates were determined by an advisory group of regional experts as being more accurate than the MEPS rates. However, because the CHIS relied on self-reports of chronic conditions whereas the MEPS used ICD-9 codes from clinical diagnosis, the rates do differ, sometimes significantly (Table 1). The difference in prevalence rates translates into differences in costs, with the largest difference being the estimated cost associated with asthma ($6 billion higher than the CDC estimate). Neither approach is without its limitations (the MEPS having a relatively small California sample size, the CHIS relying on self-reports) and thus the results might best be viewed as providing a different perspective on the costs of chronic conditions in California. Finally, the calculation of the percentage of cost attributable to chronic conditions (Table 3) should be interpreted with caution since the total cost was derived from a different estimate (Centers for Medicare & Medicaid Studies National Health Expenditure Accounts) than the cost estimates (derived from the MEPS).
As this was the first attempt to adapt the CDC Cost Calculator to estimate costs at the county level, the study encountered a number of methodological issues. These include the following:
- Accuracy of county prevalence rates: The majority of prevalence rates were obtained from the CHIS. While the CHIS is widely respected and commonly used, the sample size for 1 year is not large enough to produce reliable estimates for some groups (especially in some underserved regions) and reliance on self-reported diagnoses (as opposed to medical records as does the MEPS) raise questions about the accuracy of the estimates. Additional study is needed to develop adjusters for county-level rates. In addition, because the CHIS does not include people in institutional care, the costs reported are underestimates of the total cost. Finally, while it was not possible to develop standard deviations or confidence intervals given the components of these estimates (eg, CDC cost estimates and census data), additional analysis might explore using probabilistic sensitivity analysis to explore the robustness of the results to changes in key parameters.
- Accuracy of estimates for small counties/younger ages: The cost calculations likely underestimate the cost of chronic disease for individuals aged 0 to 17 years and for small counties or small demographic subgroups. Because the CHIS does not ask children and teens certain health questions, it was not possible to determine the prevalence for arthritis and CVDs, as well as for diabetes for individuals aged 0 to 17. As cancer prevalence rates were obtained by strata, race/ethnicity, and gender for each of the 58 California counties using SEER data, small populations in some counties or demographic subgroups within counties made it necessary to pool the small counties when determining cancer rates.
- Adjusting for differences in health services usage between counties: The estimates shown earlier account for differences in the cost to counties based on age, gender, prices in the region, population in the region, and (to some degree) rates of chronic conditions per county. Left unaccounted for are differences in the intensity of health care usage for a given condition. This can arise because of differences in the availability of medical facilities, differences in practice between regions, and differences in ability to pay.
A reliable source of data on health services usage in California is the state hospital (OSHPD) data. This data set contains all hospitalizations for the state in a given year. Estimating differences in rates of hospitalizations will give an indication of the differences in hospital use between counties. A thorough estimation of the differences between counties in health services usage was not feasible for this study, thus suggesting that differences between counties may be understated, with counties with relatively few health services (ie, medically underserved regions) likely to exhibit lower overall costs of care whereas those in more affluent areas likely to exhibit higher overall costs.
This study represents an initial attempt to estimate the cost of chronic conditions in counties in California. Additional information is needed to help counties identify the return on investment from prevention activities, including the health gains that would result, the parties that would incur additional costs and those who would financially benefit, and the health care providers and other organizations that would need to be involved to make the prevention activities successful. Having accurate estimates of the cost of chronic diseases provides LHDs with an understanding of the challenges facing the people in their region. Information on the burden chronic conditions place on each provider (eg, private insurers, Medicare, and Medicaid) will provide more accurate estimates of the amount of expenditure that each funder could avoid by engaging in prevention activities. As more people are covered by some form of health insurance in California, there is an opportunity to expand prevention services, especially to vulnerable and at-risk populations. This expansion will require increased planning and organization, particularly from local public health departments, as reducing the number of individuals with chronic disease will take not only individual-level prevention services but also institutional- and population-level interventions. Information on the economic cost of chronic conditions is important for planning prevention and control efforts, as it provides decision makers with an estimate of the amount that could be saved if chronic conditions were prevented or controlled.
1. DeVol R, Bedroussian A. An Unhealthy America: The Economic Burden of Chronic Disease. Santa Monica, CA: Milken Institute; 2007. http://www.milkeninstitute.org
. Accessed October 12, 2013.
2. California Department of Public Health. The burden of chronic disease and injury, California. http://www.cdph.ca.gov/programs/Documents/BurdenReportOnline%2004-04-13.pdf
. Published 2013. Accessed March 17, 2014.
3. Maylahn C, Fleming D, Birkhead G. Health departments in a brave new world. Prev Chronic Dis. 2013;10:130003. doi:http://dx.doi.org/10.5888/pcd10.130003
4. Drummond M, Sculpher M, Torrance G, O'Brien B. Methods for the Economic Evaluation of Health Care Programmes. 3rd ed. Oxford, England: Oxford University Press; 2005.
5. Centers for Disease Control and Prevention
. Chronic disease and health promotion. http://www.cdc.gov/chronicdisease/overview/index.htm
. Accessed March 17, 2014.
6. Agency for Healthcare Research and Quality. Mean Expenses per Person With Care for Selected Conditions by Type of Service: United States, 2011. Medical Expenditure Panel Survey Household Component Data. Rockville, MD: Agency for Healthcare Research and Quality. http://meps.ahrq.gov/data_stats/tables_compendia_hh_interactive.jsp?_SERVICE=MEPSSocket0&_PROGRAM=MEPSPGM.TC.SAS&File=HC2Y2011&Table=HC2Y2011%5FCNDXP%5FCA&
. Published 2011. Accessed on January 10, 2014.
7. O'Brien-Strain M, Addison W, Theobald N. Final Report on the Sixth Update of the Geographic Practice Cost Index for the Medicare Physician Fee Schedule. Burlingame, CA: Acumen LLC; 2010. http://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/PhysicianFeeSched/downloads/GPCI_Report.pdf
. Accessed March 17, 2010.
8. California Health Interview Survey. CHIS [computer file]. Los Angeles, CA: UCLA Center for Health Policy Research; 2011-2012. http://ask.chis.ucla.edu
. Accessed January-March 2014.
9. Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov
) SEER*Stat Database: Incidence—SEER 9 Regs Research Data, Nov 2011 Sub (1973-2010) <Katrina/Rita Population Adjustment>—Linked To County Attributes—Total U.S., 1969-2010 Counties, National Cancer Institute, DCCPS, Surveillance Research Program, Surveillance Systems Branch, released April 2013, based on the November 2012 submission.
10. American Diabetes Association. Data from the 2011 National Diabetes Fact Sheet. http://www.diabetes.org/diabetes-basics/statistics
. Released January 26, 2011. Accessed March 17, 2014.
11. California Department of Finance. 2010 Census Summary File 1, Table PCT12 by race/ethnicity California and counties. http://www.dof.ca.gov/research/demographic/state_census_data_center/census_2010/#SF1
. Extracted July 22, 2010. Accessed January 28, 2014.
12. UCLA Center for Health Policy Research. Who does the CHIS interview. http://healthpolicy.ucla.edu/chis/about/Pages/who.aspx
. Accessed March 17, 2014.
13. California Health Interview Survey. CHIS 2011-2012 Methodology Series: Report 2—Data Collection Methods. Los Angeles, CA: UCLA Center for Health Policy Research; 2014. http://healthpolicy.ucla.edu/chis/design/Documents/chis2011-2012-method-2_2014-02-21.pdf
. Accessed March 17, 2014.
14. California Health Interview Survey. CHIS 2005-2012 [computer file]. Los Angeles, CA: UCLA Center for Health Policy Research. http://ask.chis.ucla.edu
. Accessed January-March 2014.
16. California Healthcare Foundation. Chronic conditions
in California: facts and figures 2006. http://www.chcf.org/∼/media/MEDIA%20LIBRARY%20Files/PDF/C/PDF%20ChronicDiseaseFactsFigures06.pdf
. Accessed January 31, 2014.
17. DeVol R, Bedroussian A. An Unhealthy America: The Economic Burden of Chronic Disease Charting a New Course to Save Lives and Increase Productivity and Economic Growth. Santa Monica, CA: Milken Institute. http://www.milkeninstitute.org/pdf/chronic_disease_report.pdf
. Accessed October 24, 2013.
18. Thorpe K, Florence C, Joski P. Which medical conditions account for the rise in health care spending? The fifteen most costly medical conditions accounted for half of the overall growth in health care spending between 1987 and 2000. Health Aff. 2004;23:w437–w445.
19. Druss B, Marcus S, Olfson M, Tanielian T, Elinson L, Pincus H. Comparing the national economic burden of five chronic conditions
. Health Aff. 2001;20(6):233–241.
20. Garrett N, Martini EM. The boomers are coming: a total cost of care model of the impact of population aging on the cost of chronic conditions
in the United States. Dis Manag. 2007;10(2):51–60. doi:10.1089/dis.2006.630.
21. County of San Diego Health & Human Services Agency. The economic burden of chronic disease in San Diego County. www.sdhealthstatistics.com
. Published October 2010. Accessed March 17, 2014.
22. Hirsch G, Homer J, Evans E, Zielinski A. A system dynamics model for planning cardiovascular disease interventions. Am J Public Health. 2010;100(4):616–622.