Chest ACCP Member Benefits
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (10)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bauer, T. T.
Right arrow Articles by Schultze-Werninghaus, G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bauer, T. T.
Right arrow Articles by Schultze-Werninghaus, G.
(Chest. 2005;128:2238-2246.)
© 2005 American College of Chest Physicians

Cost Analyses of Community-Acquired Pneumonia From the Hospital Perspective*

Torsten Thomas Bauer, MD; Tobias Welte, MD; Carolin Ernen, PhD; Barbara Monika Schlosser, MD; Inga Thate-Waschke, MD; Justus de Zeeuw, MD and Gerhard Schultze-Werninghaus, MD

* From the Department of Internal Medicine III (Drs. Bauer, Schlosser, de Zeeuw, and Schultze-Werninghaus), University Hospital "Bergmannsheil," Bochum; Department of Pneumology (Dr. Welte), University of Hannover, Hannover; Institute of Empirical Health Economics (Dr. Ernen), Burscheid; and Bayer Vital GmbH (Dr. Thate-Waschke), Leverkusen, Germany.

Correspondence to: Torsten T. Bauer, MD, Medical Clinic III, Bergmannsheil, Clinic of the Ruhr-University Bochum, Bürkle-de-la-Camp-Platz 1, D-44789 Bochum, Germany; e-mail: torsten.bauer{at}rub.de


   Abstract
TOP
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Study objectives: Community-acquired pneumonia (CAP) is a widespread disease with important implications for health-care systems worldwide. This study investigated direct costs, treatment patterns, and outcomes associated with two patient cohorts hospitalized with CAP.

Design: The study design was naturalistic, prospective, and open.

Patients: The study enrolled 580 patients. Two hundred sixty-one patients were treated initially with IV moxifloxacin (45%, cohort M); the remaining 319 patients received nonstandardized treatment (cohort S).

Setting: Twenty-two hospitals in Germany.

Results: Clinical success rates were similar between treatment groups (cohort M, 242 of 256 patients, 95%; cohort S, 286 of 312 patients, 92%; p = 0.208). Mean ± SD length of hospital stay was 10.8 ± 5.2 days, with cohort M having a significantly shorter hospital stay (10.0 ± 4 days) compared to cohort S (11.5 ± 6 days; p < 0.001). Median of all direct costs was $1,333 (minimum, $127; maximum, $9,488), with direct costs of $1,250 in cohort M (minimum, $372; maximum, $9,488) and $1,409 in cohort S (minimum, $127; maximum, $9,366) per treated episode of CAP (p = 0.066).

Conclusions: Major determinants of costs were length of hospital stay and ICU admission, whereas costs for staff and hotel were major contributors to direct costs. Initial antibiotic therapy with moxifloxacin resulted in similar clinical efficacy and direct costs compared to nonstandardized therapy; however, patients treated with moxifloxacin benefited with an earlier hospital discharge.

Key Words: community-acquired pneumonia • cost • length of hospital stay • moxifloxacin


   Introduction
TOP
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Community-acquired pneumonia (CAP) is a widespread disease with important implications for health-care systems worldwide. For industrialized countries, CAP represents the leading cause of death due to infectious diseases.12 The incidence of CAP in Europe ranges from 1.6 to 10.8 per 1,000 adults per year, but to our knowledge, exact figures are available only for Finland, Spain, and the United Kingdom.3 Based on the German government health report, the annual number of CAP infections in Germany is estimated at 1.4 million, with approximately 230,000 patients requiring hospitalization. As such, the number of patients hospitalized with CAP is higher than with other widespread diseases, such as myocardial infarction (132,000 hospital admissions) and stroke (162,000 hospital admissions). In 2001, the mortality due to CAP amounted to 16,900 patients and represented 2.0% of all deaths in Germany, with CAP taking the ninth position for death due to all causes.4

CAP is associated with a significant use of health-care resources, primarily due to its high incidence. In 1997, the total direct and indirect costs of pneumonia in Germany were estimated to be $1.64 billion, of which $983 million were direct costs and $656 million were indirect costs (eg, losses claimed by employers). Outpatient treatment costs, consisting of drug acquisitions of $32.8 million and treatment costs of $45.9 million, were comparably low. Costs due to inpatient treatment (estimated at $896 million) account for 90% of the direct costs and therefore represent a considerable burden to the health-care system. In 2004, Germany introduced the diagnosis-related group for the reimbursement of hospitals in the statutory health insurance system. To this date, reimbursement was paid on a lump sum-per-day basis. From the perspective of the hospital, any means of reducing economic costs per case is of great interest.

The treatment of CAP with a fluoroquinolone leads to faster resolution of the infection and its associated signs and symptoms when compared to nonstandardized or combination (ß-lactams and macrolides) therapies. In two studies, patients receiving fluoroquinolone therapy had their hospital stay shortened by approximately 1 day.56 Since the length of hospital stay is an important variable affecting treatment costs, shortening the length of stay may influence the overall cost-effectiveness of the therapy.

A recently published study7 provided initial insight into the actual costs of CAP therapy. However, this study evaluated treatment costs on the basis of officially registered fees of the health-care service and was restricted to one area in Spain. In addition, this study used data from 1993 to 1995 and severity criteria that were published 2 years later, and costs could only be established retrospectively.8 Total costs associated with the treatment of patients hospitalized with CAP have not been evaluated systematically, and the relative contribution of single cost factors such as treatment and hotel costs are currently unknown. This prospective cohort study therefore assessed all costs associated with the treatment of patients hospitalized with CAP. The study also evaluated the influence of fluoroquinolone therapy on the overall cost-effectiveness of CAP treatment.


   Materials and Methods
TOP
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
The study was conducted as a cohort study with a naturalistic design in 22 community hospitals in Germany. The study design was open, active controlled, and prospective. Five hundred eighty patients were enrolled from January to July 2003. Pneumonia was defined as a new pulmonary infiltrate together with symptoms and signs of a lower respiratory tract infection. Patients were included in the study if they received at least one dose of IV antimicrobial drug and in the absence of any of the following exclusion criteria: (1) age < 18 years, or pregnant or lactating/breast feeding; (2) hospital admission within the last 2 weeks; (3) participation in a clinical trial during the period of observation; (4) pneumonia as an expected terminal event of a severe chronic disabling comorbidity; and (5) discharge from the hospital at the patient’s request.

The study design was approved by an ethical committee (University of Freiburg, Germany). Informed consent was judged not to be necessary since there was no intervention to the physicians’ treatment decision. Patients were observed in two cohorts: the first cohort was comprised of patients who received initial IV therapy with moxifloxacin (cohort M); the second cohort consisted of patients who had received any other initial IV antibiotic (cohort S). To reduce possible bias by selective recruitment and therapy in and between hospitals, we only documented patients of one cohort per hospital and included only the first 30 consecutive patients meeting the inclusion criteria.

Data Collection and Evaluation
All patients were assessed on hospital admission and during follow-up, according to a standardized data sheet. In this study, the following parameters, recorded on hospital admission, were evaluated: date of presentation (month, year); age; gender; height; weight; smoking habits (current smokers or ex-smokers ≤ 1 year); comorbidity; mode of hospital admission; residence in nursing home; duration of symptoms; clinical symptoms (body temperature, respiratory rate, heart rate, arterial systolic and diastolic BP, pneumonia-associated confusion, ie, disorientation with regard to person, place, or time that is not known to be chronic, stupor, or coma); blood gas analysis (pH, PaO2, PaCO2); chest radiography (number of lobes affected, pleural effusion); laboratory parameters (hematocrit, BUN, sodium, blood glucose); and ICU admission and/or mechanical ventilation. The following three end points were defined: hospital discharge, conclusion of the inpatient treatment due to CAP, and death. Efficacy of the therapeutic regimen was graded by the treating physician. Clinical success was defined as the disappearance of signs and symptoms related to the infection. Clinical failure was defined as the requirement for additional systemic antibacterial therapy for pneumonia, and/or the requirement for intubation and mechanical ventilation because of clinical worsening after at least 48 h of study treatment. Clinical response was rated as not assessable if the treating physician was unable to assess the signs and symptoms because of lack of information or interference of the assessment by concomitant medical or surgical conditions.

For comparison of pneumonia severity within cohorts, the pneumonia severity index (PSI) was calculated for each patient on admission. The PSI, as defined by Fine and coworkers,8 is a validated index compiled with the variables of age, gender, comorbidity, vital sign abnormalities, as well as several laboratory, blood gas, and radiographic parameters. The PSI results in a five-class point scoring system reflecting an increasing risk of mortality.

Microbiology
In order to not influence overall costs, requests for microbiological evaluation were left to the discretion of the treating physician. Therefore no standardized microbiological workup was recommended or performed.

Cost Calculation
All cost calculations were performed by economists on the basis of the case report forms and a standardized hospital controller interview (C.E.). Direct costs were defined according to Drummond and associates9 as category I (organizing and operating costs within the health sector). Total costs consisted of quantity aspects and the charge per measure. For each patient, the quantity of the type and number of diagnostic and therapeutic measures, the administered drugs, the medical devices, and the duration of medical care by physicians, nurses, medical-technical staff, and physiotherapists were documented by the responsible physician. In addition, the empiric average of a standard treatment was documented to calculate the costs of the diagnostic and therapeutic measures separately for the intensive and the regular care unit. These average durations were multiplied by the staff costs per hour of the respective responsible staff.

Based on the costs of the medical and the nursing staff per hour, the specific costs of care per patient were determined by multiplication of the average daily duration of care per patient, by the number of days spent in the respective unit. Specific hotel costs per patient were calculated by the Institute of Empirical Health Economics, Burscheid, Germany. Analyses of hospital costs were blinded; therefore, no interhospital differences can be disclosed here. Total costs were calculated from the following cost factors comprising the listed single positions.

Diagnostic Measures:
Diagnostic measures included auscultation; percussion; chest radiography; microbiology, including resistance, serology, and antigen testing; bronchoscopy, including costs for biopsy and lavage; pulmonary function testing; blood gas analysis; and ECG.

Therapeutic Measures (Excluding Drug Therapy):
Therapeutic measures included inhalation therapy, supplemental oxygen, and physiotherapy, including mobilization.

Drug Acquisition:
Drug acquisition included all drugs indicated by the treating physician to be necessary for treatment of CAP.

Hotel:
Hotel costs included estimated costs per day and patient, administration, housekeeping (eg, gas, electricity), and food for regular wards and ICUs.

Staff:
Staff included physicians, technical staff, physiotherapists, and nurses for regular wards and ICUs.

All costs have been converted from Euros ({euro}) to US dollars on the basis of the exchange rate on April 1, 2003 (1 {euro} = $1.0925). Calculated values have been rounded.

Statistics and Sample Size Calculation
The primary objective of the study was the evaluation of the overall economic costs associated with the inpatient treatment of CAP patients. The secondary objectives of this study were the evaluation of the efficacy of CAP therapy, length of hospital stay, and mortality. Primary and secondary outcome variables were compared between cohort M and cohort S.

All data were analyzed and processed using statistical software (Statistical Package for Social Sciences, Version 10.01; SPSS, Chicago, IL) on a Windows XP operating system (Microsoft; Redmond, WA). Statistics not provided by this software package were calculated using MedCalc software (version 7.0.0.2; MedCalc Software; Mariakerke, Belgium). Results are expressed as frequencies or as mean ± SD unless indicated otherwise. The {chi}2 test was used to compare proportions, and Student t test was used to compare means. All continuous variables were tested for normal distribution using the Kolmogorov-Smirnov test. For variables in which this test indicated nonnormal distribution, the Mann-Whitney U test was employed for comparison, and median, minimum, and maximum values were reported. Logistic regression was employed for multivariable testing of factors associated with mortality. Linear regression for multiple variable testing of costs included known contributors such as length of hospital stay and ICU admission together with PSI class and the type of initial antibiotic therapy (cohort M vs cohort S). All multivariable analyses used a forward model (p[in] = 0.05 and p[out] = 0.10 as inclusion and exclusion criteria, respectively). The significance level of all analyses was set to 5%, and exact p values are reported.

Because no published costs associated with the inpatient treatment of CAP in Germany were available, sample size calculations were based on the following assumptions: supposed costs under supposition of a normal distribution of $2,185 and a SD of 50% ($1,095). The study was then designed to establish a difference of 15% of total costs ($328) between cohorts with an {alpha} error of 5% and a ß error of 10% (minimal required sample size of n = 235 in each cohort).


   Results
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Patients and Demographic Data
A total of 580 patients were included in the analysis (cohort M, 261 of 580 patients, 45%; cohort S, 319 of 580 patients, 55%). Demographic characteristics and findings on hospital admission are summarized in Table 1 . Patients in cohort M tended to be older, but no significant differences were found between cohorts for age, sex, and smoking status. There was a trend toward more frequent comorbidity in cohort M, with fewer patients free of an accompanying disease. This was, in part, due to the high incidence of diabetes mellitus in cohort M but was also true if only insulin-dependent diabetes mellitus was included as a definition (cohort M, 89 of 261 patients, 34%; cohort S, 33 of 319 patients, 10%; p < 0.001). However, severity of CAP was more often categorized into PSI classes IV-V in cohort S; however, mean PSI scores differed only by eight points on average (Table 1). This is in contrast to the less frequent presentation of pneumonia-associated symptoms in cohort S compared to cohort M, such as cough, fever, general weakness, headache, and myalgia. A comparable proportion of both cohorts had to be treated in the ICU, and days spent there were also not significantly different.


View this table:
[in this window]
[in a new window]

  		Table 1. Baseline Characteristics and Clinical Presentation at Hospital Admission of Patients in Cohort M (Initial IV Therapy With Moxifloxacin) and Cohort S (Initial IV Therapy With any Other Antibiotic Drug)*

 
Antibiotic Therapy
The type of antibiotic regimens used in the study are summarized in Table 2 . Monotherapy and sequential therapy were prescribed more often in cohort M. Aminopenicillin/ß-lactamase inhibitors, in combination with a macrolide, were the most common dual-antibiotic regimen in cohort S, but monotherapy was used in 65% of the cases (Table 2). A ß-lactam monotherapy was used in 172 of 319 patients (54%), and a combination of a ß-lactam plus macrolide was administered to 75 of 319 patients in cohort S (24%). In cohort S, more patients had received an antibiotic before hospital admission, but the percentage of previously untreated patients was high in this study (473 of 580 patients, 82%).


View this table:
[in this window]
[in a new window]

  		Table 2. Antibiotic Treatment During Hospital Stay of Patients in Cohort M (Initial IV Therapy With Moxifloxacin) and Cohort S (Initial IV Therapy With any Other Antibiotic Drug)*

 
Microbiology
A microbiological workup was not requested or not reported in 233 of 580 patients (40%), with a trend toward less frequent requests in cohort S (cohort M, 94 of 261 patients, 36%; vs cohort S, 139 of 319 patients, 44%; p = 0.065). Among the positive findings in any microbiological sample, Streptococcus pneumoniae (n = 45) was the most frequently identified pathogen, followed by Staphylococcus aureus (n = 28) and Klebsiella pneumoniae (n = 13). No further data were available to group these findings in terms of a definite or presumptive etiology.

Clinical Efficacy and Adverse Events
Overall clinical efficacy of antibiotic therapy was reported by the treating physicians as a success (528 of 580 patients; 91%), a failure (40 of 580 patients; 7%), or not assessable (12 of 580 patients; 2%). When both cohorts were compared, no significant differences were found between cohort M (success, 242 of 256 patients, 95%) and cohort S (286 of 312 patients, 92%; p = 0.208).

A total of 65 adverse events were documented during CAP therapy (26 events in cohort M and 39 events in cohort S), involving 25 cohort M patients and 39 cohort S patients. Of these adverse events, 13 of 26 events in cohort M and 38 of 39 events in cohort S were classified as serious adverse events (SAEs) and not study drug related. The rates of adverse events were 10% in cohort M and 12% in cohort S. All non-SAEs in cohort M were cases of diarrhea and were assessed as related to study drug administration. In cohort S, only one non-SAE, colon bleeding, was documented and not deemed related to the administered drug. Tolerability of the antibiotic treatments was generally good.

Length of Hospital Stay and Mortality
The mean length of hospital stay was 10.8 ± 5.2 days, with a significantly shorter duration in cohort M (10.0 ± 4 days) compared to cohort S (11.5 ± 6 days, p < 0.001). The proportion of patients admitted to the ICU and the length of stay within the ICU were not significantly different between the two cohorts (Table 1).

Overall mortality was 48 of 580 patients (8.3%), with mortality being significantly higher in cohort S (36 of 319 patients; 11.3%) compared to cohort M (12 of 261 patients; 5%). Mortality was not significantly different when patients receiving ß-lactam monotherapy (18 of 172 patients, 10.5%) were compared to those receiving a combination of a ß-lactam antibiotic plus a macrolide (8 of 75 patients, 10.7%; p = 1.00). On average, nonsurvivors died after 8.8 ± 9 days, resulting in a significantly shorter hospital stay compared to survivors (11.0 ± 5 days, p = 0.005). A total of 16 of 48 patients died in the ICU (33%). When mortality was stratified according to the PSI, the following values were calculated: PSI classes I-II (0 of 143 patients, 0%), PSI classes III (6 of 116 patients, 5%), and PSI classes IV-V (42 of 321 patients, 13%). Among patients with PSI classes IV-V, there was a nonsignificant trend toward a lower mortality in cohort M compared to cohort S (cohort M, 9 of 119 patients, 8%; vs cohort S, 33 of 193 patients, 16%; p = 0.061). In the multivariate logistic regression model, with mortality as a dependent variable, the five risk classes according to the PSI were the strongest predictors of death, with a risk ratio of 2.38 (95% confidence interval, 1.6 to 3.5; p < 0.001) per increase in class. Therapy with moxifloxacin was also associated in this multivariable statistical model with a decreased risk ratio (0.43; 95% confidence interval, 0.21 to 0.85; p = 0.015).

Costs Calculations
The standardized controller interviews revealed the following average staff and hotel costs: ward physician ($37.69 ± 9.83/h); other medical staff ($11.33 ± 3.28/h); technical laboratory staff ($20.24 ± 5.46/h); physiotherapist ($19.92 ± 5.46/h); nursing ($21.50 ± 5.46 /h); nursing intensive care ($22.55 ± 5.46/h); hotel ($58.27 ± 21.8 /d); and hotel intensive care ($98.16 ± 69.92/d). The total costs of all episodes of hospitalized CAP were $381,906 in cohort M (n = 261) and $532,683 in cohort S (n = 319). The overall median costs were $1,333 (minimum, $127; maximum, $9,488), with $1,250 in cohort M (minimum, $372; maximum, $9,488) and $1,409 in cohort S (minimum, $127; maximum, $9,366) per treated episode of CAP. The difference of $159 was not statistically significant (p = 0.066). Figure 1 shows the total costs according to the individual cost factors investigated in this study. Total median costs for diagnostics and nondrug therapy were $61 and $43, respectively. As expected from the study design, ranges were large and some patients did not receive any nondrug therapy. Median drug acquisition costs associated with the therapy of CAP were $124, with a significantly higher value in cohort M. The largest factors driving total treatment costs were expenses for staff and hotel; both were significantly lower in cohort M (Fig 1).



View larger version (25K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Median costs (in US dollars) of in-hospital treatment of CAP in a prospective cohort analysis. Costs were calculated for initial IV moxifloxacin therapy (cohort M) and compared to a nonstandardized therapy (cohort S). All comparisons were made with the Mann-Whitney U test, and exact levels of significance are given. Since bars indicate median values, statistically significant differences may occur in comparisons with identical median; therefore, values and range are also reported.

 
Figure 2 summarizes the total median treatment costs associated with hospitalized CAP according to PSI score on hospital admission. The median increased from risk classes I-III and leveled off for risk classes IV-V. This is, in part, due to the median values. Mean values were $1,253, $1,372, $1,638, $1,672, and $1,676 for the PSI classes I through V, respectively. However, because patients who died were only present in risk classes III or higher, and hospital stay for those patients was shorter compared to survivors (see above), total median costs were reduced. Total median costs were significantly higher in patients with ICU admission ($2,300; minimum, $372; maximum, $9,488) compared to non-ICU patients ($1,242; minimum, $127; maximum, $6,325; p < 0.001). The multivariable analysis revealed the length of hospital stay (ß = 0.797, p < 0.001) and ICU admission (ß = 0.334, p < 0.001) as the two most important single determinants of total treatment costs. Risk class according to Fine score did not influence total treatment costs in this multivariable analysis.



View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Median treatment costs associated with hospitalized CAP according to PSI on hospital admission (in US dollars). The median increases from PSI risk classes I-III and levels off for PSI risk classes IV-V. This is in part due to the nonnormally distributed costs and the consequent display of median values, but it is also due to the fact that nonsurvivors had a shorter treatment duration and were present only in PSI classes III-V.

 

   Discussion
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
The major findings of this study were as follows: (1) median direct costs of treating patients hospitalized with CAP in Germany were $1,333 per case; (2) major contributors to the total direct costs were expenses for staff and hotel; (3) major determinants of total costs were length of hospital stay and ICU admission; (4) initial therapy with moxifloxacin was highly effective in the treatment of patients hospitalized with CAP; and (5) higher drug acquisition costs in cohort M were compensated for by a shorter duration of hospitalization. To our knowledge, this is the first study to prospectively investigate the costs of CAP treatment from the hospital perspective. So far, only data calculated from the claims of the hospitals or insurance companies were available. These claims do not reflect the direct costs of the hospital and cannot be used to give complete insight into the effectiveness of therapy. Colice and coworkers10 estimated costs of inpatient CAP treatment from a database of employer administrative claims in the United States and reported substantially higher figures ($10,227 ± 15,342). The major differences between their data and the values calculated in our study are of a methodologic nature, as this investigation calculated the costs from the hospital perspective. Therefore, the results of Colice and associates10 reflect only the higher treatment costs. Interestingly, the SD in the study by Colice et al10 indicates the large range of costs that may be associated with the treatment of patients hospitalized with CAP. To avoid bias by cost outliers and extremes, median values were used in all comparisons. However, this study10 was able to approximate the previously published data by Bartolomé and associates, who investigated the costs of CAP in a population-based European study. In this study, a total of 224 episodes of CAP were included, and of those 134 patients (60%) were treated in the hospital. The average cost estimate given in the study by Bartolomé et al7 for hospitalized CAP was $1,697 ± 592 and apparently relates closely to the mean figure calculated in the current multicenter study (mean ± SD, $1,581 ± 1,013). Nevertheless, as in the study by Colice and coworkers,10 these figures do not allow calculation of direct costs from the hospital perspective and therefore do not allow an evaluation of the profitability of drug therapy. They do, however, reflect the narrowing standard of care in the European Community because outcome variables, such as length of hospital stay, morbidity, and mortality, were comparable between these two trials.7 In contrast to the study by Bartolomé et al,7 this study was able to identify length of hospital stay and ICU admission as the two most important single predictors of resource utilization in the treatment of patients hospitalized with CAP. This is an important finding and may help to guide inpatient treatment of CAP in the future. The study by Bartolomé et al7 could not identify length of hospital stay as an independent predictor of total treatment costs. This is most likely due to the smaller sample size and the fact that the Spanish study used official figures, rather than exact calculations based on individual controller interviews as in the current study. Even so, Bartolomé et al7 calculated 15% of the total treatment costs—including inpatients and outpatients—to be associated with intensive care treatment, supporting ICU admission as the second important determinant of treatment costs, as was confirmed in the present study.7

This investigation found hotel and personnel costs to be the two major contributors to health-care resource utilization in hospitalized CAP patients. German hospitals have a high standard of care, with most wards equipped with two or three bedrooms, which increases hotel and staff costs. Nursing accounted for two thirds of the total personnel costs because wards outside the ICU are staffed 24 h/d with nurses, but doctors are available on average 8 h/d with emergency care outside office hours. There is also a trend in Germany to cut costs by reducing nursing staff during night shifts, but this will eventually affect the quality of patient care.11 Since hospitals are in close competition and individual accommodation and quality of care are important factors for the selection of a hospital, it is unlikely that average costs for personnel and hotel can be reduced. Therefore, length of hospital stay is the most important single variable tackling hotel and personnel costs. Fine and coworkers12 investigated the importance of a reduction of length of stay for patients hospitalized for CAP. The mean length of stay was 8.2 ± 5.4 days in US hospitals, and revealed considerable savings of up to $680 (from the third-party payer perspective) if the length of stay could be reduced by a single day. The length of stay in the current study was determined under reimbursement conditions on a lump sum-per-day basis. The German mean length of stay shows a remarkable difference of approximately 2.5 days compared to the United States. Since diagnoses-related groups will be the basis of reimbursement for all hospitals outside the university, it is expected that the length of hospital stay will decrease in the near future. However, it has to be kept in mind that all studies calculating costs associated with hospitalized CAP in the United States found much higher expenses for the individual episode.1013 Early patient discharge requires an efficient infrastructure usually associated with increases in administrative overhead costs. Woolhandler and associates14 calculated that administration accounted for up to 31% of health-care expenditures in the United States. Optimization of treatment and/or discharge pathways should, therefore, be performed in an integrative manner among physicians, nursing, and administrative staff to ensure that savings in one field do not result in higher costs for others.15

The newer fluoroquinolones, and particularly moxifloxacin, have been shown to induce faster symptom relief and shorten hospital stay. In a randomized controlled trial,16 CAP patients receiving moxifloxacin had a shorter time to resolution of fever (median time, 2 days vs 3 days) compared to amoxiclav with or without clarithromycin. The mean duration of the hospital stay was 9.49 ± 7.29 days for the moxifloxacin group (n = 301) and 10.41 ± 7.49 days for the comparator group (n = 321). The duration of the hospital stay, however, was not the subject of formal statistical testing in the study by Finch et al16 and was planned for descriptive analysis only. In contrast, length of hospital stay was an outcome variable in the present study and was shortened by almost 1.5 days for patients receiving initial moxifloxacin therapy. Early discharge was not standardized due to the naturalistic design, but the shorter hospital stay associated with moxifloxacin is most likely due to the faster symptom relief associated with moxifloxacin therapy.16 In a Canadian study5 comparing two treatment pathways involving levofloxacin, length of hospital stay and duration of IV treatment could also be shortened. However, this effect was not attributed to the use of fluoroquinolones alone because the major intervention was introduction of a treatment pathway and not antibiotic therapy.

Although drug acquisition costs were significantly higher in the moxifloxacin cohort, they did not influence the comparison of the total costs. While this is likely to be the result of the shorter hospital stay in this cohort, it might also be attributed to the higher nondrug costs in cohort S associated with the high proportion of IV therapies.17 Previous studies61819 have shown that an early switch from IV to oral therapy is safe and cost-effective, even in severe CAP. Early switch from IV to oral moxifloxacin has been shown to be efficient and economic.618 Our study confirms that this is well received in clinical practice, because the switch rate in cohort M was significantly higher compared to the nonstandardized cohort.

Microbiological workup was not standardized in our study, as it would have influenced cost calculations. The costs were not significantly different between the two cohorts, and further reduction of diagnostic costs can probably not be achieved.20

While costs may become more important for the economic survival of hospitals, effective treatment remains the primary focus. We therefore included standard measurements of treatment efficacy and safety in the study design. Clinical outcomes were generally favorable, and there were no significant differences between the two patient cohorts. Overall mortality was 8%; however, we found higher mortality rates in PSI class III compared to PSI classes IV and V. It is well known that the PSI tends to underestimate the severity in young individuals, probably because it was intended to identify low-risk patients. Our sample size is, however, too small to draw any definitive conclusions from this observation. Mortality was significantly higher in cohort S (11%) compared to cohort M (5%). This observation was made in an earlier randomized comparison but did not reach statistical significance.16 Adding a macrolide to ß-lactam monotherapy has been associated with a lower mortality, for example in bacteremic pneumococcal disease.21 However, in the current study, the proportion of patients who died was not different when patients receiving a ß-lactam monotherapy were compared to patients who received a combination of a ß-lactam and a macrolide. It is, therefore, unlikely that monotherapy with a ß-lactam in cohort S biased the comparison between the cohorts. Nevertheless, it is very important to interpret these findings with caution, as more severe CAP cases were enrolled in cohort S, and other studies failed to report a significant difference between moxifloxacin and nonstandardized treatment.22 Nevertheless, when data in the current study were stratified according to PSI risk classes in a multivariable analysis, therapy with moxifloxacin was still associated with better survival.

A possible limitation of this study is the observational design and the lack of randomization. Since cost measurements were the primary objective of this study, a controlled clinical study with a restricted patient population and increased social and medical care would not have yielded realistic data. The transfer of results of clinical efficacy to daily practice (clinical effectiveness) is problematic, and therefore the naturalistic design was chosen for the study design. Benson and Hartz23 found a good correspondence between clinical results from both randomized clinical trials and observational studies after conducting a meta-analysis involving 19 different indications trials. Since mean PSI was significantly different between the two cohorts, one might argue that the comparison between the two cohorts was biased. However, this difference was small and our major findings were confirmed in multivariable analyses including PSI. Patient discharge was not standardized due to the design of the study. The shorter length of stay within the moxifloxacin cohort may have, therefore, been influenced by other confounding factors including the availability of sequential IV/oral therapy.

In conclusion, this study provides cost estimates for the treatment of patients hospitalized with CAP. The major determinants of costs were length of hospital stay and ICU admission, whereas costs for personnel and hotel were major contributors to the direct costs. Initial antibiotic therapy with moxifloxacin resulted in high clinical efficacy with a trend to lower direct costs and the patient benefit of earlier discharge from the hospital.


   Acknowledgements
 
We thank all contributors to the study, whose identities cannot be disclosed herein due to the confidentiality agreement reached prior to the study. We are especially indebted to Dietmar Daniel and Thomas Pfeil for the preparation of the data file and the valuable help they provided in data analysis and management of the manuscript. Professor R. Rychlik provided valuable information regarding health-related economic issues, and Brian G. Shearer edited the manuscript.


   Footnotes
 
Abbreviations: CAP = community-acquired pneumonia; PSI = pneumonia severity index; SAE = serious adverse event

Dr. Bauer and Dr. Schlosser were supported by official grants of the Ruhr-University Bochum (FoRUM F296–01, F298–01, and F 397–03); Dr. Ernen was supported by an educational grant of Bayer Vital Leverkusen, Germany.

Received for publication September 23, 2004. Accepted for publication March 1, 2005.


   References
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Niederman, MS, McCombs, JS, Unger, AN, et al (1998) The cost of treating community-acquired pneumonia. Clin Ther 20,820-837[CrossRef][ISI][Medline]
  2. Garibaldi, RA Epidemiology of community-acquired respiratory tract infections in adults: incidence, etiology, and impact. Am J Med 1985;78,32-37[CrossRef][ISI][Medline]
  3. Major respiratory diseases: pneumonia. Loddenkemper, R Gibson, GJ Sibille, Y eds. European lung white book: the first comprehensive survey on respiratory health in Europe. 2003,55-64 European Respiratory Society and European Lung Foundation. Sheffield, UK:
  4. Konietzko, N, Fabel, H Weißbuch lunge. 2000 Thieme. Stuttgart, Germany:
  5. Marrie, TJ, Lau, CY, Wheeler, SL, et al A controlled trial of a critical pathway for treatment of community-acquired pneumonia. CAPITAL Study Investigators. Community-Acquired Pneumonia Intervention Trial Assessing Levofloxacin. JAMA 2000;283,749-755[Abstract/Free Full Text]
  6. Drummond, MF, Becker, DL, Hux, M, et al An economic evaluation of sequential IV/po moxifloxacin therapy compared to IV/po co-amoxiclav with or without clarithromycin in the treatment of community-acquired pneumonia. Chest 2003;124,526-535[Abstract/Free Full Text]
  7. Bartolomé, M, Almirall, J, Morera, J, et al A population-based study of the costs of care for community-acquired pneumonia. Eur Respir J 2004;23,610-616[Abstract/Free Full Text]
  8. Fine, MJ, Auble, TE, Yealy, DM, et al A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 1997;336,243-250[Abstract/Free Full Text]
  9. Drummond, MF, Stoddart, GL, Torrance, GW Critical assessment of economic evaluation: methods for the economic evaluation of health care programs. 1987,18-32 Oxford University Press. Oxford, UK:
  10. Colice, GL, Morley, MA, Asche, C, et al Treatment costs of community-acquired pneumonia in an employed population. Chest 2004;125,2140-2145[Abstract/Free Full Text]
  11. Milo, LA, Smucker, W, Logue, E, et al Shoot, ready, aim: pneumonia care quality and costs in a community hospital. Am J Med Qual 2003;18,214-219[Abstract]
  12. Fine, MJ, Pratt, HM, Obrosky, DS, et al Relation between length of hospital stay and costs of care for patients with community-acquired pneumonia. Am J Med 2000;109,378-385[CrossRef][ISI][Medline]
  13. Goss, CH, Rubenfeld, GD, Park, DR, et al Cost and incidence of social comorbidities in low-risk patients with community-acquired pneumonia admitted to a public hospital. Chest 2003;124,2148-2155[Abstract/Free Full Text]
  14. Woolhandler, S, Campbell, T, Himmelstein, DU Costs of health care administration in the United States and Canada. N Engl J Med 2003;349,768-775[Abstract/Free Full Text]
  15. Brown, PD Adherence to guidelines for community-acquired pneumonia: does it decrease cost of care? Pharmacoeconomics 2004;22,413-420[Medline]
  16. Finch, R, Schurmann, D, Collins, O, et al Randomized controlled trial of sequential intravenous (i.v.) and oral moxifloxacin compared with sequential i.v. and oral co-amoxiclav with or without clarithromycin in patients with community-acquired pneumonia requiring initial parenteral treatment. Antimicrob Agents Chemother 2002;46,1746-1754[Abstract/Free Full Text]
  17. Van Zanten, AR, Engelfriet, PM, van Dillen, K, et al Importance of nondrug costs of intravenous antibiotic therapy. Crit Care 2003;7,R184-R190[CrossRef][Medline]
  18. Lode, H, Grossman, C, Choudhri, S, et al Sequential IV/PO moxifloxacin treatment of patients with severe community-acquired pneumonia. Respir Med 2003;97,1134-1142[Medline]
  19. Castro-Guardiola, A, Viejo-Rodriguez, AL, Soler-Simon, S, et al Efficacy and safety of oral and early-switch therapy for community-acquired pneumonia: a randomized controlled trial. Am J Med 2001;111,367-374[CrossRef][ISI][Medline]
  20. Oosterheert, JJ, Bonten, MJ, Buskens, E, et al Algorithm to determine cost savings of targeting antimicrobial therapy based on results of rapid diagnostic testing. J Clin Microbiol 2003;41,4708-4713[Abstract/Free Full Text]
  21. Martinez, JA, Horcajada, JP, Almela, M, et al Addition of a macrolide to a ß-lactam-based empirical antibiotic regimen is associated with lower in-hospital mortality for patients with bacteremic pneumococcal pneumonia. Clin Infect Dis 2003;36,389-395[CrossRef][ISI][Medline]
  22. Torres, A, Muir, JF, Corris, P, et al Effectiveness of oral moxifloxacin in standard first-line therapy in community-acquired pneumonia. Eur Respir J 2003;21,135-143[Abstract/Free Full Text]
  23. Benson, K, Hartz, AJ A comparison of observational studies and randomized, controlled trials. N Engl J Med 2000;342,1878-1886[Abstract/Free Full Text]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (10)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bauer, T. T.
Right arrow Articles by Schultze-Werninghaus, G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bauer, T. T.
Right arrow Articles by Schultze-Werninghaus, G.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS