Background: Multimorbidity (the presence of multiple chronic conditions) is associated with high levels of healthcare utilization and associated costs. We investigated the association between number of chronic conditions and costs of care for nonelderly and elderly Veterans Affairs (VA) patients, and estimated mean VA healthcare costs for the most prevalent and most costly combinations of 3 conditions (triads).
Methods: We identified a cohort of 5,233,994 patients who received care within the VA system in fiscal year 2010. We estimated the costs of VA care for each patient using established methods and aggregated costs for inpatient care, outpatient care, prescription drugs, and contract care. Using ICD-9 diagnosis fields from all inpatient and outpatient records, we determined the prevalence of 28 chronic conditions and all condition triads. We then compared the condition-cost gradient, most prevalent triads, and most costly triads among nonelderly (below 65 y) and elderly (65 y and above) patients.
Results: Almost one third of nonelderly and slightly more than a third of elderly VA patients had ≥3 conditions, but these patients accounted for 65% and 67% of total VA healthcare costs, respectively. The most common triad of chronic conditions for both nonelderly and elderly patients was diabetes, hyperlipidemia, and hypertension (24% and 29%, respectively). Conditions that were present in the most costly triads included spinal cord injury, heart failure, renal failure, ischemic heart disease, peripheral vascular disease, stroke, and depression. Although patients with the most costly triads had average costs that were 3 times higher than average costs among patients with ≥3 conditions, the prevalence of these costly triads was extremely low (0.1%–0.4%).
Conclusions: Patients with multiple chronic conditions account for a disproportionate share of VA healthcare expenditures. Interventions that aim to optimize care and contain costs for multimorbid patients need to incorporate strategies specific to the most prevalent and the most costly combinations of conditions.
*Health Economics Resource Center
†Center for Healthcare Evaluation, VA Palo Alto Healthcare System, Menlo Park
‡Division of General Medical Disciplines, Stanford University, Stanford, CA
§Health Services Research and Development, Durham VA Medical Center
∥Department of Medicine, Division of General Internal Medicine, Duke University Medical Center, Durham, NC
The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veteran Affairs, Stanford University or Duke University.
Supported by the Agency for Healthcare Research and Quality (AHRQ) Multiple Chronic Conditions Research Network R21HS019445 to M.L.M., support from the VA Palo Alto Center for Healthcare Evaluation (LIP 62-108) and Health Economics Resource Center core funding (SDR-ECN-99017-1) to J.Y. M.L.M. was also supported by a Research Career Scientist award from the Department of Veterans Affairs (RCS 10-391) and this work was also supported by the Office of Research and Development, Health Services Research and Development Service, Department of Veterans Affairs.
M.L.M. has received consultation funds from Daichi Sankyo, Takeda Pharmaceuticals, the Surgical Review Corporation, ResDAC at the University of Minnesota, and owns stock in Amgen due to his spouse’s employment. The remaining authors declare no conflict of interest.
Reprints: Matthew L. Maciejewski, PhD, Department of Medicine, Division of General Internal Medicine, Duke University Medical Center, Durham, NC 27705. E-mail: firstname.lastname@example.org.
The number of people with multiple chronic conditions has steadily increased over the past decade and is projected to continue rising by >1% per year until 2030.1,2 Patients with multiple chronic conditions account for a disproportionate share of healthcare costs,3,4 a pattern that is well documented among older adults. For example, in 1999, the mean Medicare expenditures were $1154 among those with 1 condition and $13,973 among those with ≥4 conditions.5 These high costs are attributable, in part, to high rates of hospitalization, many of which may be avoidable.5
Although the positive association between number of chronic conditions and healthcare expenditures is well established, there is a need for greater understanding about specific combinations of conditions contributing to high costs. Recent analyses of Medicare data identified high rates of high cholesterol, hypertension, diabetes, and ischemic heart disease in the most prevalent disease combinations and high rates of stroke and chronic kidney disease in the most costliest disease combinations.2 However, there are a number of reasons why these findings may not be generalizable to other health systems such as the Veterans Affairs (VA). Multimorbidity is more common among patients receiving care within the VA system compared with the general population.6 This is especially true among younger patients within the VA, many of whom have comorbid mental health conditions such as depression, anxiety, posttraumatic stress disorder, and substance use disorders.7–9 These comorbidities can complicate the care of other chronic conditions and increase healthcare utilization and overall costs of care.10–12
Recent evidence showed that close to 90% of older adults receiving care within the VA had ≥3 chronic health conditions in 2007–2008, and that the most prevalent triads (combinations of 3 conditions) among this population included hypertension, hyperlipidemia, coronary heart disease, diabetes, gastroesophageal reflux disease, and benign prostatic hypertrophy.13 The costs associated with these various condition triads were not examined.
The objective of this study was to advance our understanding of multimorbidity within the VA healthcare system by: (1) estimating mean costs per patient and across the population by number of chronic conditions, (2) identifying the most prevalent triads and associated costs, and (3) identifying the most costly comorbidity triads among nonelderly (below 65 y) and elderly (65 y and above) patients. Examining mean VA healthcare costs for the most prevalent triads can identify a specific, large group of veterans who may benefit from efforts to develop clinical guidelines that address multimorbidity because they account for a disproportionate share of VA healthcare costs due to their sheer numbers. Examining VA healthcare costs for the most costly triads can identify veterans who would directly benefit from intensive primary care interventions to coordinate the care that best addresses their needs.
Data Sources and Measures
We identified a cohort of 5,233,994 patients who received VA inpatient or outpatient care in fiscal year 2010 (October 1, 2009–September 30, 2010) from the VA Medical SAS files that record all VA utilization. We obtained patients’ age, sex, diagnoses, all inpatient and outpatient utilization, and costs of VA-provided and VA-sponsored services during fiscal year 2010.
Using ICD-9 diagnosis fields from all inpatient and outpatient records, we coded 28 chronic conditions that have been the focus of quality improvement efforts and research within the VA because of their prevalence, management challenges, and costs.14–16 Each inpatient stay included fields for up to 13 conditions, whereas each outpatient encounter included fields for up to 10 conditions. The study conditions included: acid-related diseases [gastric, duodenal, peptic, gastrojejunal, and esophagus ulcers, gastritis, esophageal reflux (GERD), and dyspepsia], alcohol/drug use disorders, Alzheimer disease, arthritis, asthma, cancer, chronic renal failure, chronic obstructive pulmonary disease, depression, diabetes, headache, heart failure, hepatitis C, HIV/AIDS, hyperlipidemia, hypertension, ischemic heart disease (including coronary heart disease), low back pain, multiple sclerosis, other psychiatric conditions (including schizophrenia and manic depression), osteoporosis, Parkinson disease, posttraumatic stress disorder, prostatic hyperplasia, peripheral vascular disease, spinal cord injury, stroke, and thyroid disorder. Conditions were indicated for each patient if they had at least 2 separate instances of a diagnosis in inpatient or outpatient (including primary care and specialty care) encounter records to avoid counting conditions recorded to rule out diagnoses.
The overall cost of VA care for each patient was estimated by aggregating costs for inpatient care, outpatient care, prescription drugs, and contract care from several sources as there are no billing records in the VA. The costs of VA outpatient visits and inpatient stays were obtained from the 2010 Average Costs files, which generate VA healthcare costs based on Medicare payments for similar care using information such as DRG codes in inpatient records and procedure codes in outpatient records.17–19 Prescription drug costs were obtained from the 2010 Decision Support System Pharmacy files that report the supply and dispensing costs per each prescription filled in the VA. The costs of contract care provided by non-VA providers was obtained from 2010 Fee Basis files, which report the actual payments per service made to providers.
We stratified the sample into elderly (age, 65 y and above) and nonelderly (18–64 y) subgroups to examine whether the cost-condition gradient, most prevalent triads, and most costly triads were similar for elderly and nonelderly.20 For both age cohorts, we report the number and proportion of patients by number of study conditions, unadjusted mean costs by number of conditions, and proportion of total VA healthcare costs by number of conditions.
For patients with ≥3 of the study conditions, we coded all possible combinations of 3 coexisting conditions (triads) as in previous research13 and determined which triads had the greatest prevalence and costs by age group. When determining the highest cost triads, there were triad combinations that had very low rates (<0.1% of patients), hence we report only the highest cost triad combinations where the prevalence was at least 0.1%. We report the prevalence of triads and the mean costs of patients within triads for the top 5 most prevalent and top 5 most expensive triads. As some patients had >3 conditions, the triads do not represent mutually exclusive groups of patients.
The Cost-Condition Gradient for Elderly and Nonelderly Patients
Among the 2,888,989 nonelderly VA patients, 35% had none of the 28 chronic conditions examined, 18% had only 1 condition, 16% had 2 conditions, and almost a third had ≥3 conditions (Fig. 1). The unadjusted mean cost for nonelderly patients with 3 conditions ($10,809) was twice the mean cost for patients with only 1 condition and >5 times the cost for patients with no documented conditions. Nearly all (92%) VA healthcare costs were spent on patients with ≥1 chronic conditions. Nonelderly patients with ≥3 conditions accounted for 65% of total VA healthcare costs although they made up only 32% of the nonelderly VA population.
Among the 2,345,005 elderly VA patients, 36% had none of the study conditions, and 35% had ≥3 conditions (Fig. 2). The same pattern of costs by number of chronic conditions was observed in patients 65 years and older although the mean cost per patient was lower than for patients under 65 (as would be expected because dual eligibility for Medicare tends to result in decreased use of VA care). Patients with ≥3 conditions incurred 67% of total VA healthcare costs among the older cohort.
The Most Prevalent Condition Triads Among Nonelderly and Elderly Veterans
Among the 897,614 nonelderly patients with ≥3 conditions, the most common triad of chronic conditions was diabetes, hyperlipidemia, and hypertension, which affected 24% of nonelderly patients with an unadjusted mean cost of $15,856. These 3 conditions along with ischemic heart disease, depression, PTSD, and low back pain in various combinations accounted for the 5 most prevalent triads (Table 1). Total costs of patients in any of the top 5 triads represented 37% of costs of all patients with ≥3 conditions.
In the cohort of 832,300 elderly VA patients, the most prevalent triad was also diabetes, hyperlipidemia, and hypertension, which affected 29% of elderly patients and had an unadjusted mean cost of $13,174. These 3 conditions along with ischemic heart disease and acid-related diseases composed the 5 most prevalent triads. Contrary to our expectations, mental health conditions were not implicated in the most prevalent conditions for elderly patients. For the elderly, the total costs of patients in any of the top 5 triads were 50% of total costs of all patients with ≥3 conditions.
The Most Costly Condition Triads Among Nonelderly and Elderly Veterans
Chronic heart failure, chronic renal failure, and chronic obstructive pulmonary disorder was the most costly triad for nonelderly patients, which affected only 0.2% of patients but was associated with an unadjusted mean cost of $81,685 (Table 2). The total costs of patients in the most expensive triads represented 3% of total costs among patients with ≥3 conditions. Diabetes, hypertension, and spinal cord injury was the most costly triad for elderly patients, affecting 0.1% of patients but was associated with an unadjusted mean cost of $65,905, and the total costs of elderly patients in the most expensive triads represented 4% of costs of all patients with ≥3 conditions. The combination of chronic heart failure and chronic renal failure was included in 3 of the 5 most costly triads for both age cohorts. The other conditions involved in the most costly triads in both age cohorts included ischemic heart disease, peripheral vascular disease, stroke, depression, and other psychiatric disorders. None of the prevalence rates for these triads was >0.4% of patients with ≥3 chronic conditions.
Multimorbidity is highly prevalent among elderly and nonelderly veterans using the VA healthcare system, and patients with multiple chronic conditions incurred a disproportionate share of resources. Hypertension and hyperlipidemia were the most frequent conditions occurring in the 5 most prevalent triads for elderly and nonelderly veterans, highlighting the need for preventive services to encourage dietary changes and physical activity. Hypertension and hyperlipidemia are considered concordant conditions in terms of treatment, hence effective treatment of one condition can positively affect the outcomes of the other.21 As a result, clinical guidelines responsive to multimorbidity may reasonably place greater emphasis on the joint management of these conditions.
Unlike a recent paper by Steinman et al13 that found hypertension, hyperlipidemia, and coronary heart disease to be the most prevalent triad among elderly veterans in 2007–2008, we found that hypertension, hyperlipidemia, and diabetes was the most prevalent triad in 2010. The consequences of comorbid diabetes and hypertension and/or hyperlipidemia in older age is demonstrated by the high rates of ischemic heart disease co-occurring with these conditions in 3 of the 5 most common condition triads for the elderly cohort. Preventing diabetes onset and managing diabetes symptoms when they develop is a major issue in reducing morbidity and adverse outcomes as patients age. We also found that mental health conditions (depression and PTSD) occurred frequently in triad combinations in nonelderly patients. Finding innovative ways to comanage mental health and medical problems for these patients will continue to be an area of priority for VA providers given the high rates of mental health conditions in recent veterans and substantial evidence that adults with mental health comorbidities have greater difficulty managing chronic medical conditions and have higher rates of hospitalizations and higher costs.12,22,23 Mental health conditions did not arise in the most prevalent triads for elderly patients, possibly due to undercoding of mental health diagnoses.24
Unlike the most prevalent condition triads, the most costly condition triads included many discordant conditions affecting multiple organ systems. The co-occurrence of chronic heart failure and chronic renal failure in many of the top 5 most costly triads for both age cohorts highlights the high costs associated with managing complex and discordant conditions. This finding builds on literature describing challenges to self-management and quality of care as well as adverse health outcomes that arise in the presence of discordant conditions.25–29 For example, patients with discordant conditions often face challenges adhering to multiple drug prescriptions, experience adverse drug interactions, and have difficulty coordinating care with multiple specialists.30–33 In addition, patients in some of the most costly triads had 7–8 total chronic conditions, on average, compared with patients in less costly triads who had 5–6 total chronic conditions, hence their costs also reflect a greater total comorbidity burden (consistent with literature describing the association between number of chronic conditions and overall costs of care).5
Although relatively few patients were in the most costly triads, their mean costs were almost 3 times the costs of patients in other triads. These findings highlight the need for interventions that target the sickest patients who have high resource use to provide more cost-effective care. In recent years a number of innovative healthcare delivery models have emerged with the goal of improving outcomes and containing utilization in these complex and high-risk patients. This strategy has been tested in a number of settings,34–36 including clinics serving older adults.37,38 Some models, such as CareMore (a for-profit corporation that provides intensive management for frail and chronically ill members) and the Ambulatory-Intensive Care Unit (developed for Boeing employees and later disseminated to other settings such as Atlantic City and rural Humboldt County in California), report savings as high as 15%–20% in nonpeer-reviewed observational studies.36,39 These savings are generally attributed to lower spending for emergency care and hospitalizations.
This analysis is subject to several limitations. We presented the mean costs of patients in different triads, but we did not determine how each condition separately attributed to the total costs of triads. As patients in some of the most costly triads had a greater number of total conditions than patients in less costly triads, the high costs may be due to specific conditions in the triad, or the total burden of conditions in addition to the triad, or another, more costly but rarer condition.
We identified chronic health conditions from diagnosis fields in inpatient and outpatient utilization records, hence there may have been underreporting of diagnosed conditions. We focused on the 28 conditions that are common or a focus of VA quality improvement efforts, and results may have changed if we considered a broader array of conditions. Steinman et al13 considered 23 conditions among elderly veterans in 2007–2008 including atrial fibrillation, anemia, epilepsy, gout, and chronic renal insufficiency, which were not included in our study and obtained a significantly higher estimate of elderly patients with ≥3 conditions. Steinman also used only 1 instance of a diagnosis to identify a condition, hence our estimates using 2 instances are more conservative. We did not have access to Medicare data to supplement VA diagnoses for Medicare-eligible veterans, hence we may have undercounted conditions in older patients, especially those who were less reliant on VA care. If we had used Medicare data, then the prevalence of conditions would have increased, and the proportion of elderly patients with ≥3 conditions would likely have been higher. Steinman and colleagues also included patients with at least 2 VA visits or hospitalizations, whereas we did not, hence we may have more casual VA users than the cohort of Steinman and colleagues.
Non-VA healthcare costs were excluded from these comparisons, and results might have changed if Medicare and Medicaid costs had been included. These results also may not generalize to non-VA samples because veterans are predominantly male. In addition, the most prevalent and most costly triads may differ in nonveteran samples, which would be useful to compare in future research.
The rise in reported chronic conditions in US adults suggests multimorbidity will pose a growing challenge for healthcare systems including the VA system. Currently, >90% of VA healthcare resources are devoted to patients with chronic conditions. We found a significant cost-condition gradient among elderly and nonelderly veterans, which is consistent with findings from similar studies in Medicare beneficiaries. As patients with multiple chronic conditions account for a disproportionate share of VA healthcare expenditures, targeted treatment guidelines and interventions focused on this population are critical to managing the growth of healthcare costs. Greater efforts around prevention, guidelines for joint management of comorbid conditions, and intensive management of high utilizing and high cost patients may benefit multimorbid patients and healthcare systems overall.
1. Freid VM, Bernstein AB, Bush MA.Multiple Chronic Conditions Among Adults Aged 45 and Over: Trends Over the Past 10 years. NCHS Data Brief, No. 100.2012.Hyattsville, MD:National Center for Health Statistics.
2. .Chronic Conditions Among Medicare Beneficiaries, Chartbook.2012:2012 ed.Baltimore, MD:Centers for Medicare and Medicaid Services.
4. Lehnert T, Heider D, Leicht H, et al..Review: health care utilization and costs of elderly persons with multiple chronic conditions.Med Care Res Rev.2011;68:387–420.
5. Wolff J, Starfield B, Anderson G.Prevalence, expenditures, and complications of multiple chronic conditions in the elderly.Arch Intern Med.2002;162:2269–2276.
6. Agha Z, Lofgren RP, VanRuiswyk JV, et al..Are patients at veterans affairs medical centers sicker?: a comparative analysis of health status and medical resource use.Arch Intern Med.2000;160:3252–3257.
7. Cohen BE, Gima K, Bertenthal D, et al..Mental health diagnoses and utilization of VA non-mental health medical services among returning Iraq and Afghanistan veterans.J Gen Intern Med.2010;25:18–24.
8. Hankin CS, Spiro A III, Miller DR, et al..Mental disorders and mental health treatment among US Department of Veterans Affairs outpatients: the Veterans Health Study.Am J Psychiatry.1999;156:1924–1930.
9. Hoge CW, Auchterlonie JL, Milliken CS.Mental health problems, use of mental health services, and attrition from military service after returning from deployment to Iraq or Afghanistan.JAMA.2006;295:1023–1032.
10. Dickey B, Normand SL, Weiss RD, et al..Medical morbidity, mental illness, and substance use disorders.Psychiatr Serv.2002;53:861–867.
11. Druss BG, Rosenheck RA, Desai MM, et al..Quality of preventive medical care for patients with mental disorders.Med Care.2002;40:129–136.
12. Himelhoch S, Weller WE, Wu AW, et al..Chronic medical illness, depression, and use of acute medical services among Medicare beneficiaries.Med Care.2004;42:512–521.
13. Steinman MA, Lee SJ, John Boscardin W, et al..Patterns of multimorbidity in elderly veterans.J Am Geriatr Soc.2012;60:1872–1880.
14. Yoon J, Scott JY, Phibbs CS, et al..Recent trends in veterans affairs chronic condition spending.Popul Health Manag.2011;14:293–298.
15. Yu W, Ravelo A, Wagner TH, et al..Prevalence and costs of chronic conditions in the VA health care system.Med Care Res Rev.2003;60suppl146S–167S.
16. Demakis JG, McQueen L, Kizer KW, et al..Quality Enhancement Research Initiative (QUERI): a collaboration between research and clinical practice.Med Care.2000;38suppl 1I17–I25.
17. Phibbs CS, Bhandari A, Yu W, et al..Estimating the costs of VA ambulatory care.Med Care Res Rev.2003;60suppl54S–73S.
18. Wagner TH, Chen S, Barnett PG.Using average cost methods to estimate encounter-level costs for medical-surgical stays in the VA.Med Care Res Rev.2003;60suppl15S–36S.
19. Yu W, Wagner TH, Chen S, et al..Average cost of VA rehabilitation, mental health, and long-term hospital stays.Med Care Res Rev.2003;60suppl40S–53S.
20. Petersen LA, Byrne MM, Daw CN, et al..Relationship between clinical conditions and use of veterans affairs health care among Medicare-enrolled veterans.Health Serv Res.2010;43:762–791.
21. Lagu T, Weiner M, Hollenbeak C, et al..The impact of concordant and discordant conditions on the quality of care for hyperlipidemia.J Gen Intern Med.2008;23:1208–1213.
22. Sullivan G, Han X, Moore S, et al..Disparities in hospitalization for diabetes among persons with and without co-occurring mental disorders.Psychiatr Serv.2006;57:1126–1131.
23. Katon WJ, Lin E, Russo J, et al..Increased medical costs of a population-based sample of depressed elderly patients.Arch Gen Psychiatry.2003;60:897–903.
24. Kim HM, Smith EG, Stano CM, et al..Validation of key behaviourally based mental health diagnoses in administrative data: suicide attempt, alcohol abuse, illicit drug abuse and tobacco use.BMC Health Serv Res.2012;12:18.
25. Frayne SM, Halanych JH, Miller DR, et al..Disparities in diabetes care: impact of mental illness.Arch Intern Med.2005;165:2631–2638.
26. Harman JS, Edlund MJ, Fortney JC, et al..The influence of comorbid chronic medical conditions on the adequacy of depression care for older Americans.J Am Geriatr Soc.2005;53:2178–2183.
27. Redelmeier DA, Tan SH, Booth GL.The treatment of unrelated disorders in patients with chronic medical diseases.N Engl J Med.1998;338:1516–1520.
28. Turner BJ, Hollenbeak CS, Weiner M, et al..Effect of unrelated comorbid conditions on hypertension management.Ann Intern Med.2008;148:578–586.
29. Wang PS, Avorn J, Brookhart MA, et al..Effects of noncardiovascular comorbidities on antihypertensive use in elderly hypertensives.Hypertension.2005;46:273–279.
30. Dickson VV, Buck H, Riegel B.A qualitative meta-analysis of heart failure self-care practices among individuals with multiple comorbid conditions.J Card Fail.2011;17:413–419.
31. Noel PH, Parchman ML, Williams JW, et al..The challenges of multimorbidity from the patient perspective.J Gen Intern Med.2007;22suppl 3419–424.
32. Schoenberg NE, Leach C, Edwards W.“It’s a toss up between my hearing, my heart, and my hip”: prioritizing and accommodating multiple morbidities by vulnerable older adults.J Health Care Poor Underserved.2009;20:134–151.
33. Sevick MA, Trauth JM, Ling BS, et al..Patients with complex chronic diseases: perspectives on supporting self-management.J Gen Intern Med.2007;22supp 3438–444.
34. Brown KE, Levine JM, Fiellin DA, et al..Primary intensive care: pilot study of a primary care-based intervention for high-utilizing patients.Dis Manag.2005;8:169–177.
35. Raven MC, Doran KM, Kostrowski S, et al..An intervention to improve care and reduce costs for high-risk patients with frequent hospital admissions: a pilot study.BMC Health Serv Res.2011;11:270.
36. Reuben DB.Physicians in supporting roles in chronic disease care: The CareMore model.J Am Geriatr Soc.2011;59:158–160.
37. Boult C, Reider L, Leff B, et al..The effect of guided care teams on the use of health services: results from a cluster-randomized controlled trial.Arch Intern Med.2011;171:460–466.
38. Counsell SR, Callahan CM, Clark DO, et al..Geriatric care management for low-income seniors: a randomized controlled trial.JAMA.2007;298:2623–2633.