Comorbidities as a driver of the excess costs of community-acquired pneumonia in U.S. commercially-insured working age adults

Community-acquired pneumonia (CAP) ranks among the top causes of death and hospitalization and it is a major driver of health care utilization and cost. Annually, there are 2 to 3 million CAP cases in the US resulting in half a million hospitalizations, 10 million physician visits, and 45,000 deaths [1]. A recent study of US non-elderly adults estimated that CAP was associated with annual excess medical costs of $8.5 billion and a productivity loss of $2.1 billion [2]. While it is understood that pneumonia risks are elevated for certain groups including several common comorbidities such as asthma and COPD [3‐7], the cost burden of pneumonia has not been explored within comorbidity subgroups.

We estimate the total excess cost burden of CAP as the sum of medical costs and productivity losses (work absence and short-term disability) in a large cohort of commercially insured working-age adults throughout the United States for the most common comorbidities among individuals with pneumonia: asthma, diabetes, COPD, and CHF [2, 8]. In addition to being common, these comorbidities have been singled out as potential risk factors for complications and/or mortality during or following a pneumonia episode [4, 9‐12]. As a result, costs of pneumonia may be higher when these comorbidities are present. In this paper we explore this possibility from the perspective of employers who insure adult populations and face the burden of medical costs and costs of absenteeism and disability.

We identified 402,831 patients with CAP from MarketScan Commercial Claims (Commercial) and Encounters database. The overall demographic characteristics and pre-index period clinical characteristics are shown in Table 1. The mean age of CAP patients was 47 years and 45.6% were male. The most prevalent comorbid conditions were cardiovascular disease (22.2%), diabetes (8.1%), asthma (6.3%), cancer (6.0%), and COPD (4.1%). Mean length of follow-up was 318 days. Pre-period outpatient costs were nearly $3,000. This high baseline outpatient cost is driven by the high volume of cases with comorbidities. For example, pre-period outpatient costs were over $6,000 for cases with comorbidities of COPD, diabetes, and CHF.

Demographic characteristics were similar between CAP patients and their 1,425,045 matched controls as the standardized differences were typically in the 1-3% range. The highest standardized differences were COPD at 7.29, comprehensive insurance at 7.39, and pre-period outpatient costs with standardized differences above 10; multivariate models were used to control for these remaining imbalances.

Annualized adjusted excess total medical costs of a case of CAP were estimated at $11,395 (SE=$42) across all patients, as determined by the difference between estimated costs for patients with CAP ($17,640 (SD=$26,706)) and the control patients ($6,245 (SD=$16,189)). These estimates come from costs defined in 2008 dollars as total gross payments to a provider, including deductibles, copayments, and coordination of benefits.

Excess medical costs of CAP among patients with COPD, diabetes, and CHF are higher than the average excess medical costs of CAP. Pneumonia is most costly among CHF patients at $31,593 (SE=$531). Excess costs are also substantially higher for patients with COPD ($17,039 (SE=$186)) and diabetes ($16,351 (SE=$165)). Among the comorbidities of interest, CAP with asthma had the lowest total excess medical costs at $10,158 (SE=$109). All of these costs were higher than the excess costs of CAP for patients without CHF, COPD, diabetes, or asthma ($9,866 (SE=$43)).

The detail of medical costs by each cost component sheds light on the drivers of higher excess medical costs among the comorbidity subgroups. As shown in Table 2, excess inpatient costs were highest among CHF patients ($18,339 (SE=$163)) and lowest among asthma patients ($4,908 (SE=38). Excess outpatient costs were highest among CHF patients ($11,953 (SE=387)) and lowest among asthma patients ($4,210 (SE=62)). Excess outpatient pharmacy costs appeared more similar across the comorbidity groups, ranging from $1,040 (SE=19) for asthma patients to $1,565 (SE=29) for COPD patients.

As shown in Table 3, productivity costs are $3,034 (SE=$55) higher for patients with pneumonia than the controls; approximately two-thirds of this difference is driven by short-term disability costs while the remainder comes from absenteeism costs. Among the comorbidity subgroups, productivity costs were highest for COPD ($6,454 (SE=$329)) and diabetes ($5,044 (SE=$263)) and lowest for CHF ($2,842 (SE=$752).

Table 4 contains the adjusted overall excess cost of CAP, which are also depicted in Figure 1. CHF patients had the highest total burden of CAP, with costs nearly three times the average excess costs of CAP. CAP for patients with COPD and diabetes costs 86% and 70% more than for a typical patient. The excess cost of CAP for patients with asthma appeared similar to the average excess cost of CAP. On average, the excess cost of CAP was 79% excess direct costs and 21% excess indirect costs. The relationship between excess direct and indirect costs was similar for patients with asthma, diabetes and COPD where indirect costs accounted for 24-27% of total excess costs. Among patients with CHF, total excess costs were predominantly direct costs (92%) compared to indirect costs (8%).

The objective of this study was to estimate, for selected comorbidities, the excess health care costs and employer productivity costs of working age adults with CAP. The excess costs per case of CAP in the working-age adult population was $14,429 overall, but could be nearly 3 times as high among patients with CHF and around 75% higher among patients with COPD and diabetes. When considering the reduced cost burden from prevention of CAP, it is not sufficient to consider the risks of acquiring pneumonia, the fact that the cost per case can vary among different risk groups must be taken into account. As a result, we have provided key estimates of the burden of disease among key comorbidity risk groups.

The cost per CAP patient is highest for patients with CHF ($34,436) and is substantially higher than the cost of CAP for patients with other conditions such as COPD ($23,439), diabetes ($21,395) and asthma ($13,307), and particularly those without any of these selected comorbidities ($12,601). The higher costs, particularly for CHF and even for COPD may be explained by the clinical evidence which suggests that the pre-existing condition has a detrimental effect on pneumonia prognosis [10, 21], but there is also evidence that pneumonia contributes to an exacerbation of the original condition [9, 22]. These estimates can be used to better understand the cost of preventing a case of CAP, which would clearly be higher among patients with CHF than if prevention was driven only by the risks of being diagnosed with pneumonia.

With no other comorbidity-specific estimates of pneumonia costs, these results cannot be directly compared to other studies in the literature, but the overall estimates are in line with the literature. For example, our work suggests that productivity costs are about 20% of total burden. The one previous analysis that included productivity costs suggested that they represent 29% of total cost [22]. Colice et al. performed a retrospective claims-based analysis of CAP based on employer-payments for medical care made by a single Fortune 100 company in the time period from 1996 to 1998 [23]. This analysis found that the mean cost for an inpatient episode of pneumonia was $8,942, while an ambulatory CAP episode cost $466. The discrepancy in CAP costs between the two studies is likely due, in large part, to the escalation of health care expenditures over the previous decade and the cost time horizon.

A recent European study of patients with CAP reported no difference in median hospitalization costs between patients with and without diabetes, CHF, or COPD [24]. The analysis did not include outpatient costs or a non-CAP comparison cohort. The study reported a significant lack of literature describing the impact of comorbidities on CAP-related costs. Bartolome [25] and Niederman [26] evaluated age as a risk factor for higher CAP-related costs. Reyes [24], Bauer [27], and Kaplan [28] all reported higher pneumonia-related costs associated with either more severe CAP cases or in the presence of pneumonia-related complications, such as pleural effusion or respiratory failure.

The analyses described herein have limitations. While the MarketScan Research Databases represent a wide variety of nationally-representative employed patients and their dependents, it is not a random sample. As a result, the study results may not be generalizable to groups beyond the commercially insured, such as Medicaid beneficiaries or unemployed patients. Misclassification of pneumonia due to inaccurate or absent diagnosis coding is possible when analyzing claims data and basing the pneumonia diagnosis on ICD-9 codes alone and not on review of medical records or X-ray results. For example, it is possible that some mild cases were coded as “respiratory infection” and a few severe cases were coded as “sepsis”. Likewise, the intention of this analysis was to describe the burden of CAP separate from HAP; our ability to differentiate CAP from HAP may be limited by information available in an administrative claims database. Including HAP cases in our cohort of CAP patients could lead to an overestimation of the burden of CAP. However, to reduce the possibility of including HAP, we excluded from the analysis any diagnosis of pneumonia in an inpatient claim that was based on a secondary diagnosis or even in a principal diagnosis if there was evidence of an inpatient or long-term care stay in the previous 30 days. Productivity costs are estimated from the employer perspective and include only missed work time, not work time completed at a reduced productivity level. If illness affects labor force participation or has an effect on long-term disability, these productivity costs will not be captured. Finally, our method of estimating “excess costs” depends on the degree to which our matched control group is an accurate representation of the counterfactual of what the pneumonia patients would have experienced had they not contracted pneumonia. This assumption is not testable so our results must be interpreted in light of this uncertainty.

In summary, we found that CAP is associated with substantial excess medical and productivity costs that can vary substantially by comorbidity group. Excess costs per case are highest for patients with a comorbidity of CHF and also elevated for patients with diabetes and COPD. Returns on prevention of pneumonia would therefore be higher in adults with these comorbidities than in adults without any of the selected comorbidities.