|
Pharmaceutical economics & health policy: economic assessment of the community-acquired pneumonia intervention trial employing levofloxacin |
Palmer C S, Zhan C, Elixhauser A, Halpern M T, Rance L, Feagan B G, Marrie T J |
|
|
Record Status This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn. Health technology A critical pathway (CP) designed to manage community-acquired pneumonia (CAP) more efficiently than conventional care was examined. The CP included three elements:
the use of the Pneumonia Severity Scoring Index (PSSI) to avoid unnecessary hospitalisation of CAP patients;
criteria for switching from intravenous to oral therapy to reduce intravenous antibiotic use and length of stay (LOS); and
the use of levofloxacin as the antibiotic, with the aim of having fewer patients treated with more than one antibiotic.
In the conventional therapy arm, decisions on admission to the hospital, choice of antibiotic (excluding levofloxacin), switching from intravenous to oral antibiotics, use of diagnostic tests and procedures, and discharge from the hospital were left to the discretion of the treating physician.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised adult patients whose chest radiography indicated opacity compatible with CAP, and who had two or more symptoms of CAP. The exclusion criteria were shock, immune deficiency, a requirement for intubation or immediate admission to an intensive care unit, pregnancy or nursing an infant, alcoholism, chronic renal failure, or a known or suspected resistance to levofloxacin.
Setting The setting was primary, secondary and tertiary care, as patients were treated in community and teaching hospitals as well as on an outpatient basis. The economic study was conducted in Canada.
Dates to which data relate The effectiveness and resource use evidence was obtained from a trial that enrolled patients from January to July 1998. The price year was 1998.
Source of effectiveness data The effectiveness data were derived from a single study, which was in press at the time of this economic evaluation (see Other Publications of Related Interest).
Link between effectiveness and cost data The effectiveness and cost data were both collected prospectively from the same patient sample as that used in the economic analysis.
Study sample The details of the power calculations were not reported in the economic assessment. The study sample included 716 eligible patients in the 9 CP hospital arms and 1,027 in the 10 conventional therapy hospitals. The total number of patients was 1,743 subjects. The sample appears to have been appropriate for the clinical study question.
Study design This was a cluster-design, multi-centre, randomised, controlled clinical trial. The units of the computer-generated randomisation were the 19 hospitals, stratified by type of institution (teaching or community hospitals) and by historical LOS for CAP. The follow-up was 6 weeks. Loss to follow-up was not reported in the paper.
Analysis of effectiveness The basis of the effectiveness analysis was intention to treat. The health outcomes reported in the published clinical trial included health-related quality of life (measured with the SF-36), complication rates, readmission rates and survival (see Other Publications of Related Interest). The baseline sociodemographic and clinical characteristics of the two groups were similar.
Effectiveness results The effectiveness results were reported in detail in the published clinical trial (see Other Publications of Related Interest). Briefly, there were no significant differences between the two arms in terms of health-related quality of life, complication rates, readmission rates and mortality. The difference in mortality was a non significant 0.1% between the two arms.
Clinical conclusions The results of the study showed that both strategies were associated with similar clinical results.
Measure of benefits used in the economic analysis No summary measure of benefit was used since both strategies produced similar health benefits in the effectiveness analysis. This study was therefore a cost-minimisation analysis.
Direct costs The unit costs were linked to each patient's inpatient and outpatient medical resource use and lost productivity. No discounting was carried out due to the short time horizon of the study. The quantities and the costs were analysed separately. The unit costs were not reported, but are available from the authors on request. The resources were derived using actual data from the sample. They included detailed inpatient and outpatient medical resource use, truncated at 42 days. To minimise protocol-driven resource use and bias, the case report forms differed slightly between the two arms. The authors constructed a composite resource use measure, which they referred to as bed days per patient managed (BDPM). The BDPM was the mathematical product of hospital LOS and admission rate.
All the perspectives evaluated included inpatient physician visits and outpatient direct medical costs. The health care system perspective accounted for units of hospital resources used, for example, laboratory tests, diagnostic imaging, procedures, days on mechanical ventilation and medication. The per diem costs from this perspective were assumed to include all inpatient costs, but excluded inpatient physician visits, which were added to the per diem costs. The per diem costs varied depending on the number of days spent in a less costly nursing ward compared to a more costly intensive care unit (ICU). The government perspective included a mean value for the per diem cost, which was independent of the site of inpatient care.
The costs were determined from Ontario source documents. They were for inpatient and outpatient physician visits, laboratory and diagnostic tests, procedures, prescription and over-the-counter drugs, and dispensing fees. To estimate the room and board costs (nursing wards and ICU per diem for the health care perspective and average per diem for the government one), the Ontario Case Cost Project performed an ad hoc analysis. These costs included all medical resources named for inpatients, but excluded inpatient physician visits. The data came from a 1994-5 database of 11 Ontario hospitals. The analysis was performed for typical cases (excluding transfers and LOS outliers). The home care costs included data on home visiting, hours of homemaker services, and administrative costs. In the economic analysis, as well as in the clinical trial, the initial hospital LOS and readmissions were truncated at 42 days (6 weeks).
The costs were reflated as necessary with the medical component of the Consumer Price Index. The price year was 1998.
Statistical analysis of costs The analysis took into account the clustered structure of the trial. Data were aggregated by hospital and then by study arm. The analysis was performed using SAS software (version 6.2). Non-parametric tests were used to compare differences in resource use and lost productivity between the arms. Missing outpatient and lost productivity cost data were imputed through a mixed model prediction and bootstrapping, and 95% confidence intervals (CIs) were calculated. The variables included in the model were baseline sociodemographics, clinical severity, admission to hospital and hospital indicator. A significant proportion of the data on outpatient data and lost productivity was missing, although data on the main cost drivers for outpatients (readmissions and medications at discharge) were available for all patients.
Indirect Costs No discounting was carried out in view of the 6-week time horizon. The societal perspective added the indirect costs of lost days of normal activity due to CAP, as reported by the patients, to the government perspective. Data on CAP-related lost productivity were collected from the study participants by mail and telephone, 2 and 6 weeks after completion of the 10-day antibiotic course. The caregivers' lost productivity was not included. No indirect costs were included for patients who died, as there was no difference in the death rates. The costs were based on the mean hourly age and gender-specific Canadian wages for full-time workers. The costs were reflated as necessary with the medical component of the Consumer Price Index. The price year was 1998.
Sensitivity analysis Variability of the cost data was evaluated and estimated through the 95% CI of the parameters, using bootstrapping.
Estimated benefits used in the economic analysis This was a cost-minimisation analysis since the benefits were similar in both arms. See the 'Effectiveness Results' section.
Cost results From all three perspectives, CP produced cost-savings that ranged from $457 to $994 per patient and implied cost decreases of 16 to 24% in comparison with conventional treatment.
As hypothesised, the proportion of patients discharged from the emergency department was larger at the CP hospitals (53.5%) than in the conventional hospitals (37.8%), (p=0.001). Also, patients in the CP arm were admitted less often with a low-risk CAP based on the PSSI (33.2% versus 46,8%; p<0.001).
Variations in resource use and costs were large between compared groups, and varied widely from hospital to hospital. Hospitals in the CP had 1.6 fewer BDPM than the comparators, (p=0.05). A larger proportion of CP patients received antibiotic monotherapy (35.5% versus 11.3%; p<0.001), and for a shorter period overall (3.8 versus 5.3 days; p=0.001). More outpatients discharged from the CP hospitals were prescribed monotherapy than the comparators (64.9% versus 47.1%; p<0.05). There were no other significant differences in medical resource use between the treatment arms.
>From the health care perspective, the mean total costs were $2,382 (standard deviation, SD=697) for CP versus $2,839 (SD=929) for conventional treatment. The corresponding costs from the government perspective were $3,114 (SD=865) for CP versus $4,108 (SD=1,354) for conventional treatment, and from the societal perspective $3,687 (SD=1,140) versus $4,560 (SD=1,418). A sensitivity analysis using the bootstrap method resulted in 95% CIs for savings per patient of $22 to $975 from the health care system perspective, $372 to $1,486 from the government perspective, and $192 to $1,331 from the societal perspective.
The direct medical costs were very similar between study arms from all three perspectives. The initial hospitalisation costs represented between 85 and 87% of the total costs from both the government and the health care perspectives. Rehospitalisation costs accounted for 9% of the total costs from the health care perspective to 9 to 11% from the government perspective.
The small number of patients treated in the ICU meant that the lower per diem cost of nursing ward used from the health care perspective resulted in overall lower hospitalisation/rehospitalisation and total costs, compared with the government perspective.
Missing data were more common in the conventional arm (74%) than the CP arm (47%) and were imputed for lost productivity, outpatient physician visits and home care. Nevertheless, outpatient care constituted 3 to 4% of the total costs from the government perspective and 5 to 6% of the total costs from the health care system perspective.
Synthesis of costs and benefits Authors' conclusions The critical pathway (CP), which comprised the Pneumonia Severity Scoring Index (PSSI), criteria for switching from intravenous to oral antibiotics, hospital discharge criteria and the use of levofloxacin, was associated with cost-savings in comparison with conventional therapy for pneumonia. Use of the CP did not compromise the health outcomes.
CRD COMMENTARY - Selection of comparators The intervention studied was complex, with a three-part strategy. It combined the use of a severity score to avoid hospitalisation in low-risk patients with guidelines for switching from intravenous to oral antibiotics and the use of levofloxacin. It seems adequate to have chosen hospitals as the unit of randomisation, and to compare the CP with conventional therapy in comparator hospitals. The comparator strategy was not described in sufficient detail, and as it was carried out in different hospital settings, it is probably subject to a lot of within-arm variability. This lack of information on the conventional treatment arm may make it more difficult to translate the results to other settings. However, as randomisation was stratified by type of hospital (teaching and community) and by historic LOS in CAP patients, this gives more validity to the results obtained.
Validity of estimate of measure of effectiveness The analysis used a cluster-design, randomised controlled trial that involved 19 hospitals. It was appropriate for the study question. The study sample was representative of the study population and the patient groups were comparable at baseline. The total number of patients in each group differed significantly, but this may have been due to fewer (one less) hospitals in the CP group. The analysis was conducted by intention to treat and respected the special design of the trial. For example, adjustments were made for confounding factors as well as by hospital. These factors suggest that the validity of the effectiveness should be high. However, many of the details of the clinical effects were reported elsewhere, as indicated throughout the abstract.
Validity of estimate of measure of benefit The analysis of benefits was not reported in depth in this paper (see Other Publications of Related Interest). Since the benefits were shown to be similar, the economic analysis concentrated only on the cost differences.
Validity of estimate of costs The reporting and analysis of resource and costs were very ambitious and extensive. All the categories of costs relevant to each of the three evaluated perspectives were included, thus enhancing the generalisability of results. The authors reported that they excluded the caregivers' time valuation in the societal perspective, but this does not seem to have affected the conclusions. The societal perspective may have overestimated the total savings of the CP, as the patients were mostly old (mean age 65 years) and unemployed or retired (more than half), and full-time age- and gender-adjusted wages were used. The costs and the quantities were not reported completely separately, but the authors offered to provide this information upon request.
The resource use data were obtained from the trial and an appropriate statistical analysis of the quantities was performed. The amount of missing data was high for outpatients and indirect costs. The method used to impute missing data, especially in the conventional therapy arm, was adequate. This did not have a large impact on the results, accounting for only a small proportion of the total costs.
Although the hospitals were from different Canadian provinces, prices were all taken from Ontario institutional sources. Different prices were used for the health care and government or societal perspectives. The cost analysis did not include the costs of implementing the CP.
Other issues The authors made extensive and appropriate comparisons with findings from other studies. As they pointed out, the effects of the individual components of the CP strategy and their relative importance cannot be disentangled by the present study and should be studied further. The generalisability to other settings was not specifically addressed, but the study was conducted in an important sample of teaching and community hospitals and thus fosters the transferability of their results to similar settings. The authors stated that there was an important variability in resource use and costs between the hospitals, but did not expand on this issue, although they stated that data were available upon request. This information could be important in applying the results to other specific settings. The authors also stated that even with CP implementation, attention must be paid to barriers to timely hospital discharge of clinically stable patients, which depend to a large extent on the behaviour of health professionals and hospital priorities. The authors did not present their results selectively. Their results can be applied to patients who present to an emergency department and have CAP.
Implications of the study The CP was associated with cost-savings in comparison with conventional therapy for pneumonia. Use of the intervention did not compromise health outcomes. Future studies, as indicated by the authors, should examine the relative importance of each element of the CP, should make efforts to capture every element of the economic analysis, and should consider the costs of implementing the intervention.
Source of funding Supported by Janssen-Ortho, Toronto, Canada.
Bibliographic details Palmer C S, Zhan C, Elixhauser A, Halpern M T, Rance L, Feagan B G, Marrie T J. Pharmaceutical economics & health policy: economic assessment of the community-acquired pneumonia intervention trial employing levofloxacin. Clinical Therapeutics 2000; 22(2): 250-64 Other publications of related interest Marrie TJ, Lau CY, Wheeler SL, Wong CJ, Vandervoort MK, Feagan BG. 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 9;283:749-55.
Indexing Status Subject indexing assigned by NLM MeSH Aged; Anti-Infective Agents /economics /therapeutic use; Canada; Cluster Analysis; Community-Acquired Infections /classification /drug therapy /economics; Critical Pathways /economics; Delivery of Health Care /economics; Female; Humans; Length of Stay; Levofloxacin; Male; Middle Aged; Ofloxacin /economics /therapeutic use; Pneumonia /classification /drug therapy /economics; Severity of Illness Index AccessionNumber 22003007536 Date bibliographic record published 31/03/2004 Date abstract record published 31/03/2004 |
|
|
|