The effects of multiple chronic conditions on hospitalization costs and utilization for ambulatory care sensitive conditions in the United States: a nationally representative cross-sectional study

Multiple chronic conditions are highly prevalent among U.S. patients hospitalized for potentially preventable acute and chronic conditions. In this study, more than 90 % of those hospitalized for ambulatory care sensitive chronic conditions had 2 or more chronic conditions, and over 20 % had 6 or more chronic conditions. For those hospitalized for potentially preventable acute conditions, nearly 80 % had MCCs and more than 10 % had 6 or more chronic conditions. This study’s findings on prevalence are consistent with previously published literature on hospitalization trends for people with MCCs. Other studies have reported higher utilization among patients with more chronic conditions [6, 9, 20, 21]. A study of 2009 HCUP discharge data showed that two-thirds of hospitalized patients had MCCs [6]; by 2012 we found that 80–90 % of patients in our sample had MCCs. Other studies have shown increasing rates of MCCs in the United States [9].

The high prevalence of MCCs among hospital patients indicates that we need better understanding and treatment of these patients. What is the pattern of their use of health care services across the care spectrum? Will improving coordination of care, monitoring patient conditions, and directing patients to more appropriate health care settings result in better health outcomes and greater efficiencies of utilization? HHS has developed a strategic framework with goals of strengthening the U.S. health care and public health systems, empowering individuals to use self-care management, equipping health care providers with tools and interventions, and supporting targeted research on MCCs to create effective interventions [22]. For example, AHRQ has developed tools that physicians can use to enhance ambulatory care for patients by improving communication with patients, by encouraging patients to obtain preventive services and to schedule evaluations and treatments on time, and by monitoring patients for medication effectiveness and adverse events, among other evidence-based approaches [23].

Consistent with our inpatient cost results, other studies have shown higher inpatient health care costs for patients with MCCs [6, 11, 20]. Authors of a systematic literature review of 35 studies found that almost all studies showed a positive monotonic relationship between cost and MCCs [20]. For example, in a study on the impact of MCCs among the Medicare population, annual payment amounts per beneficiary for all settings of care and for those with only 1, 2, and 3 or more conditions were $7172, $14,931, and $32,498, respectively, in 2005 dollars [24]. In our study of inpatient costs for all payers in 2012, costs per stay for the reference category, white males 65 and older, were $6583, $7833, $8689, and $8623 for 0 or 1, 2 or 3, 4 or 5, and 6 or more conditions, respectively. In our study, the higher cost of inpatient treatment was driven by longer lengths of stays rather than by higher costs per day on average. This result suggests that patients admitted and initially treated for one condition may require extra days for monitoring and treatment of one or more of their secondary conditions. For example, a patient admitted with hypertension and diabetes might be managed quickly in terms of blood pressure, but not in terms of blood glucose. Achieving a reasonable glucose level may require extra days in the hospital. Diabetes management and control would not require large outlay costs, as a surgical intervention might, but they would require a longer stay for treatment, for titration of medication, and for development of a suitable post-discharge plan. Alternatively, the data may suggest that individuals that present to the hospital with multiple chronic conditions may have more severe complications with their primary condition upon arrival at the hospital that require greater lengths of stay. For example, an individual with diabetes and depression that is admitted to a hospital due to diabetic complications may, on average, have a more severe complication and require longer stays than, on average, an individual that is admitted to a hospital with diabetic complications but has no other chronic diseases. We note that the subgroup with 6 or more chronic conditions has essentially the same costs as those with 4 to 5 such conditions; the lack of a monotonically increasing cost for the 6+ group suggests that additional diagnostic coding, perhaps motivated by reimbursement, does not add information about risk related to the cost of a stay. Counting up to 5 or more chronic conditions may be sufficient for profiling the relationship of chronic conditions to utilization and outcomes.

One implication of our research is the potential value from incorporating MCCs or numbers of MCCs into risk adjustment, quality and performance measurement, and other analytics. The statistically significant findings that more MCCs result in longer hospital stays and higher costs per stay suggest that the number of MCCs, if not specific MCC combinations, should be analyzed whenever severity of a patient’s condition is a potential contributing factor to an outcome or a potential distinguishing feature of a population. The number of chronic coexisting conditions typically is not used as a measure of severity in health services research or performance measurement, although particular comorbidities may be specified in various risk adjustment schemes. Findings from this study suggest that the number of MCCs could be a simple and useful predictor of utilization and costs of health care services that account for the complexity of patients’ conditions. Furthermore, health policymakers might test the effect of number of MCCs on reimbursement formulas and provider performance measures to create equitable comparisons and adjustments of payments among providers.

Our study benefited from the use of the HCUP SID, a large discharge database of experience for all payers. Use of the PQI conditions facilitated examining MCCs for patients who share potentially preventable conditions. Despite these strengths, our results should be interpreted in light of their limitations. First, only a portion of the total costs of care for chronic disease can be studied using hospital data. Although hospital stays account for a large portion of the cost of some chronic conditions such as heart failure, they do not include office visits, clinical tests, drug therapy, and other ancillaries. Second, only 38 of the 50 states provided data that could be used for this study, so the study is not nationally representative, despite over 3.5 million discharges for the 12 conditions reflected in the composite PQIs.. Third, variation in the presence of MCCs reflects only a portion of the variability of hospital costs among patients in our analysis. In particular, we did not address other patient, hospital, and community factors that influence costs. Our intent for this analysis was to explore the value of future work on MCCs, which we confirmed. Fourth, enumerating the number of chronic conditions present in hospital records depends on the extent of coding of secondary diagnoses in the data. To the extent that recording may be truncated, the number of MCCs may have been underestimated. In addition, our use of an enumeration of chronic diseases, rather than incorporating measures of the severity of an individual’s illness, such as the Charlson co-morbidity index [25], or a consideration of each individuals specific disease combinations based on their specific disease profile, as done by Kadam and colleagues [26], prohibits us from identifying which combinations of diseases are most strongly associated with the outcomes of interest. However, the enumeration and categorization of conditions aligns our results with operating definitions of MCC outlined by the United States Department of Health and Human Services initiatives that underlie research and policy considerations on this topic and does allow for a general understanding of MCCs that would not be possible considering disease combinations in isolation.

In this large sample of records of U.S. hospitalizations, we observed a positive association between the presence of MCCs and hospitalization costs for potentially preventable conditions. The higher costs associated with MCCs was driven by longer hospital stays rather than by higher costs per day. To our knowledge, this is the first study to examine MCC in relation to costs and length of stay for hospitalizations that are potentially preventable through high-quality ambulatory care. e. When considering trends in costs and utilization related to preventable hospitalizations, studies have indicated that the occurrence of preventable hospitalizations is decreasing in the United States [8]. However, it is important to consider that at the same time, the prevalence of patients presenting with MCCs is increasing [6, 8]. The higher costs and utilization associated with these more complicated patients bring a greater burden with each preventable hospitalization. Thus, although the number of preventable hospitalizations is declining, our analyses indicate that some of the benefit anticipated with that reduction may be offset by an increase in costs and utilization driven by a rise in the prevalence of MCCs. Patients with MCCs are a challenge to treat because of the number, complexity, and interaction of their illnesses. Patients with a high number of morbidities may be experiencing end-of-life crises that lead them to the hospital repeatedly; recent insights on how the medical profession deals with patients at the end of life suggest that the health care system could benefit from educating clinicians on how to deal with these situations realistically and humanely [27].

In this study we specifically examined the hospital experience, but treatment spans all settings of care. A greater understanding of how MCCs influence care costs and utilization from ambulatory care through the inpatient setting may help providers across this spectrum identify ways to work together to achieve optimum health outcomes for their patients and to reduce hospitalization costs [12].