TECHNOLOGY OVERVIEW: PHARMACEUTICALS

Transcription

1 TECHNOLOGY OVERVIEW: PHARMACEUTICALS ISSUE 10.0 DECEMBER 1997 CLINICAL AND ECONOMIC CONSIDERATIONS IN THE USE OF FLUOROQUINOLONES based primarily on the Technical Report: An Analysis of the Use of Fluoroquinolones for Uncomplicated Urinary Tract Infections, Prostatitis, and Community-Acquired Pneumonia: Clinical and Economic Considerations B Barrett, M Doyle, P Parfrey, J Fardy, S Crewe, G Kent, J McDonald, K White, V Gadag, J Feehan Overview prepared by Judith L. Glennie, FCSHP Pharmaceutical Consultant to CCOHTA

2 Cite as: Canadian Coordinating Office for Health Technology Assessment. Clinical and Economic Considerations in the Use of Fluoroquinolones. Ottawa: Canadian Coordinating Office for Health Technology Assessment (CCOHTA); Reproduction of this document for non-commercial purposes is permitted provided appropriate credit is given to CCOHTA. Legal Deposit National Library of Canada ISSN

3 The Canadian Coordinating Office for Health Technology Assessment (CCOHTA) is a non-profit organization, funded by the federal, provincial and territorial governments. It was established to encourage the appropriate use of health technology by influencing decision-makers through the scientific evaluation of medical procedures, devices and drugs. The effectiveness and cost of technology and its impact on health are examined. This overview has been prepared by staff at CCOHTA and is based primarily on a technical report: An Analysis of the Use of Fluoroquinolones for Uncomplicated Urinary Tract Infections, Prostatitis and Community-Acquired Pneumonia: Clinical and Economic Considerations. 1 This overview attempts to put the original study into a clinical perspective. This overview does not necessarily reflect the opinions of the original investigators. To obtain copies of publications please contact: CCOHTA Publications Green Valley Crescent Ottawa, Ontario, Canada, K2C 3V4 Telephone: (613) Facsimile: (613) pubs@ccohta.ca or download full-text from Vous pouvez aussi vous procurer la version française Considérations d ordre clinique et économique sur l emploi des fluoroquinolones à l OCCETS.

4 REVIEWERS Reviewers for overview are: Mr. Brenden Barrett Dr. Andreas Laupacis Bayer Inc. Faculty of Medicine and Clinical Epidemiology Unit Memorial University of Newfoundland St. John s, Newfoundland Chair, CCOHTA Scientific Advisory Panel Division of Clinical Epidemiology Ottawa Civic Hospital Ottawa, Ontario Etobicoke, Ontario Reviewers for technical report were: Dr. Andreas Laupacis Dr. David Feeny Dr. Lawrence Joseph Dr. John Conly Bayer Inc. Janssen-Ortho Inc. Merck-Frosst Canada Inc. Chair, CCOHTA Scientific Advisory Panel Division of Clinical Epidemiology Ottawa Civic Hospital Ottawa, Ontario CCOHTA Scientific Advisory Panel Health Sciences Centre McMaster University Hamilton, Ontario CCOHTA Scientific Advisory Panel Division of Clinical Epidemiology Montreal General Hospital Montreal, Quebec Hospital Epidemiologist The Toronto Hospital Toronto, Ontario Etobicoke, Ontario North York, Ontario Kirkland, Quebec This report was reviewed by external reviewers and by members of CCOHTA s Scientific Advisory Panel. These individuals kindly provided comments on drafts of this report. This final document incorporates most of the Reviewers' comments, however, CCOHTA takes sole responsibility for its form and content.

5 BACKGROUND SUMMARY REMARKS Fluoroquinolones (also generally referred to as quinolones ) are a class of antimicrobial drugs which, as a group, have a broad spectrum of activity against organisms involved in a wide variety of infections. An additional advantage is their availability as an oral dosage form, which offers the option of outpatient therapy in some cases where inpatient intravenous treatment has traditionally been used. Quinolone antibiotics currently on the market include products such as ciprofloxacin (Cipro ), ofloxacin (Floxin ) and norfloxacin (Noroxin ). Their broad spectrum of activity has also led to the use of fluoroquinolones to treat infections for which they have not received formal regulatory approval, thus raising concerns of extensive and/or non-selective use. The acquisition costs of fluoroquinolone antibiotics typically exceed those of most alternative oral antibiotics used for urinary and respiratory tract infections. Thus, there is debate regarding the appropriate and cost-effective use of these medications in treating some relatively common infections. CCOHTA commissioned a study, entitled An analysis of the use of fluoroquinolones for uncomplicated urinary tract infections, prostatitis and community acquired pneumonia: clinical and economic considerations 1, in an effort to evaluate the use of these agents. Specifically, the study s objectives were to carry out a comparative therapeutic and economic evaluation of ciprofloxacin, ofloxacin and norfloxacin versus other recommended antibiotic therapies in the treatment of: a) prostatitis; b) community-acquired pneumonia; and c) urinary tract infections (not including prostatitis and sexually transmitted diseases). CONCLUSIONS 1) The literature did not establish any difference among the three fluoroquinolones studied (i.e ciprofloxacin, norfloxacin and ofloxacin), in terms of efficacy or toxicity, when administered for uncomplicated lower urinary tract infection in women (i.e. cystitis). 2) The base case analysis of 3-day treatment regimens involving all three quinolones revealed incremental costs per morbid days avoided which were very similar in magnitude. For instance, the comparison of the least expensive quinolone to trimethoprim-sulfamethoxazole (TMP-SMX) revealed an incremental cost per morbid days avoided of $3.93 (incremental cost per day = $2.87; incremental effect = 0.73 days of avoided excess morbidity). Thus, the marginal cost per day of morbidity avoided by choosing a fluoroquinolone over TMP-SMX is quite modest. 3) In 3-day treatment regimens for cystitis, much of the difference in the acquisition cost of fluoroquinolones versus TMP-SMX is offset by the higher cost of treating persistent or recurrent infections or adverse effects which may occur when TMP-SMX is used. Sensitivity analysis of most of the areas of uncertainty or variation had little impact on the conclusions derived from the model. 4) Deviations from the base case 3-day treatment regimen (i.e. higher doses or longer durations of therapy) increased the incremental cost of fluoroquinolones relative to TMP-SMX. Canadian Coordinating Office for Health Technology Assessment 1

6 5) Given the weakness and/or absence of the clinical data, the investigators did not proceed with an economic evaluation of fluoroquinolones in prostatitis or community-acquired pneumonia. 6) Limitations of this analysis include: data availability limitations (i.e. no direct comparisons amongst the three quinolones, no effectiveness data); as well as limitations of the study itself (i.e. compared fluoroquinolones only to TMP-SMX, use of cost data from two jurisdictions, use of a non-standard clinical outcome unit [ morbid days avoided ], and a perspective limited to that of provincial Ministers of Health). 2 Canadian Coordinating Office for Health Technology Assessment

7 INTRODUCTION Urinary Tract Infections Urinary tract infections (UTIs) are bacterial in nature, and can be divided into two broad categories: complicated and uncomplicated. The uncomplicated category can be further subdivided into lower urinary tract infections (i.e. cystitis) and upper urinary tract infections (i.e. pyelonephritis), both of which may be community-acquired or nosocomial a in nature. The majority of uncomplicated UTIs occur in women, especially those between 18 and 40 years of age. Empiric antibiotic therapy (e.g. nitrofurantoin, trimethoprim, TMP-SMX, amoxicillin, fluoroquinolones) is the norm in community-based patients with cystitis, with the initial choice of therapy being influenced by clinical efficacy, cost, adverse effects, patient allergy history and knowledge of local bacterial resistance patterns. While the duration of therapy may vary (i.e. from 1 to 10 days), a 3-day treatment course is the current recommendation in most cases. For patients with uncomplicated pyelonephritis, therapy may take place in hospital (usually by the IV route) or in the community (usually by the oral route), and the average duration of treatment is 14 days. The literature regarding fluoroquinolones in uncomplicated urinary tract infections was sufficient to establish the relative efficacy and toxicity of fluoroquinolones only against TMP-SMX in cystitis. No analysis was carried out regarding fluoroquinolones in pyelonephritis, or fluoroquinolones versus any other antibiotics that might be used for cystitis, as the data in the available literature were insufficient. Prostatitis Prostatitis is an inflammatory condition of the prostate gland, which may present as an acute or chronic condition with or without an identified bacteriological source. Empiric antibiotic therapy is the norm for most types of prostatitis, as cultures to identify the offending organisms are difficult to perform and are not reliable. The literature evaluating the treatment of prostatitis contains few non-randomized prospective studies, and those that do exist often compare therapies that are not recommended for this condition. (Note that the investigators did not limit their search to randomized controlled trials as it was known that there were few such trials in the literature.) Only one randomized trial was identified that compared two recommended therapies; that is, norfloxacin and TMP-SMX. Most trials evaluated the treatment of chronic prostatitis. Only one trial was identified in which the results for patients with acute prostatitis were presented separately. Even in the trials available for chronic prostatitis, the data were so limited that the investigators were unable to judge the relative clinical efficacy and toxicity of long term use of norfloxacin or TMP-SMX. Very few patients have been studied, resulting in wide confidence intervals for most outcome measures. Given the weakness of the data available, the investigators did not proceed with an economic evaluation of fluoroquinolones in prostatitis. a A hospital-acquired infection. Canadian Coordinating Office for Health Technology Assessment 3

8 Community-Acquired Pneumonia Pneumonia involves infection of the lung tissue, which is due to a variety of organisms (e.g. bacteria, viruses, fungi, parasites, etc.). The term pneumonia describes a group of specific types of lung infections, each of which is defined through its individual epidemiologic, pathologic, and clinical characteristics. The disease is most common among elderly patients and those patients with concurrent illnesses, such as diabetes, chronic obstructive lung disease, heart failure, liver or kidney disease and AIDS. Pneumonia is a serious infection and, due to the variety of organisms which may be involved and the unreliability of culture results, patients are usually treated with empiric antibiotic therapy. The antibiotics typically used to treat community-acquired pneumonia (in hospital and/or in the community) include: beta-lactams, fluoroquinolones, macrolides, aminoglycosides, lincosamides, tetracyclines and TMP-SMX. Specific empiric regimens have been developed for a variety of patient subgroups, based on previous research into the organisms most likely to cause the infection in specific patient populations. The literature evaluating the treatment of community-acquired pneumonia contained no studies which satisfied the criteria for inclusion developed by the investigators (e.g. randomized controlled trial involving a quinolone, clear case definition, well-defined and distinct patient populations and subgroups in study, etc.). Most studies which included quinolone antibiotics did not compare these drugs to a standard pneumonia treatment regimen, thus limiting the ability to establish the place of fluoroquinolones in pneunomia therapy compared to standard treatments. Given the absence of sufficient data to allow reliable analysis, the investigators did not proceed with an economic evaluation of fluoroquinolones in pneumonia. (Readers should note that the results of several controlled trials comparing quinolones to standard treatment regimens for community-acquired pneumonia should be available in the peer reviewed literature within the next year.) PARAMETERS OF THE EVALUATION Therapy Evaluated This report summarizes the comparison of norfloxacin, ciprofloxacin and ofloxacin versus trimethoprim-sulphmethoxazole (TMP-SMX; see Efficacy section below) in the treatment of uncomplicated lower urinary tract infection (i.e cystitis). Perspective and Target Audience The analysis was performed from the perspective of a provincial Ministry of Health. This viewpoint was adopted because information for that perspective was available, whereas information was inadequate to construct analyses from the societal viewpoint. The target audience for this analysis is provincial drug plan managers, third party payers and clinicians dealing with this patient population. 4 Canadian Coordinating Office for Health Technology Assessment

9 Type of Analysis i) Analysis The study used cost-effectiveness analysis to determine the relative role of norfloxacin, ciprofloxacin and ofloxacin versus TMP-SMX in the treatment of uncomplicated lower urinary tract infections. ii) Decision Model A model was developed to determine the cost per day of morbidity avoided if a quinolone were prescribed rather than TMP-SMX as a 3-day regimen for uncomplicated lower urinary tract infection in women. The decision analytic model was constructed using SMLTREE software. The basic scheme for each arm of the tree was as follows: a female patient presents with signs, symptoms and preliminary medical evaluation (i.e. pyuria on microscopy) which indicates an uncomplicated lower urinary tract infection; no urine culture is sent at this point (consistent with clinical general practice), but she is treated on an empiric basis with a 3-day course of antibiotic (i.e. choice node); the patient may return in 3 days if symptoms persist, necessitating a change in drug therapy; or the patient may develop side effects (e.g. rash, nausea, vaginitis), also obliging a change in antibiotic; and, finally, the patient may have a recurrence up to 6 weeks after initial therapy, necessitating additional drug therapy to cure the infection. Standard meta-analytic methods were used to determine point estimates and 95% confidence intervals (CI) for the pertinent clinical outcomes (i.e. relative rates of efficacy and toxicity) described in the decision model and defined below. The meta-analysis used information based on randomized trials for uncomplicated lower urinary tract infection. iii) Time Horizon Cost and outcomes were evaluated out to 6 weeks to capture differences in the risk of recurrent infections between therapies. Due to the short time frame of this particular disease state, future costs and outcomes were not discounted. iv) Assumptions For purposes of analysis of the use of fluoroquinolones versus TMP-SMX in cystitis, the following major assumptions were made: a) It is estimated that of those women presenting with signs and symptoms of cystitis, 86% will have a bacterial urinary tract infection, 6% will have non-bacterial urethritis, and 8% will have other diagnoses. b) Clinical success rates (i.e. alleviation of symptoms) from clinical trials are a more relevant indicator of response to treatment as opposed to bacteriological success rates, as it is the symptoms which prompt patients to seek medical care. Canadian Coordinating Office for Health Technology Assessment 5

10 Outcomes of Interest Table 1 outlines the estimates for and pooled risk differences in safety and efficacy between fluoroquinolones and TMP-SMX as used in the model. These were derived from the meta-analysis of the available literature. In addition, the number of days of morbidity was used as the unit of clinical effect to incorporate the disutility b of both the symptoms of disease and potential drug-related adverse effects. Avoidance of these days of morbidity represents the measure of the effect of antibiotic therapy. Estimates of morbid days attributable to the various clinical outcomes which might be associated with uncomplicated lower urinary tract infections in women are outlined in Table 2. Table 1: Meta analysis-derived estimates of and pooled risk differences in efficacy and safety between fluoroquinolones and TMP-SMX Outcome measure Fluoroquinolones* TMP-SMX (95% CI) Pooled risk difference #, fixed effects (95% CI) Proportion free of bacteriuria soon after therapy Proportion still free of bacteriuria up to 6 weeks after therapy Proportion free of symptoms soon after therapy Proportion free of symptoms up to 6 weeks after therapy Proportion with adverse effects Proportion stopping drug due to adverse effects 95.4% 93.6% (90.9% to 96.3%) 86.5% 84.9% (81% to 88.7%) 90.5% 82.6% (76.8% to 88.4%) 86.8% 80.3% (73.9% to 86.7%) 12.2% 19.2% (15.2% to 22.8%) 0.7% 3.1% (1.2% to 5%) 1.8% (-0.9% to 4.5%) 1.6% (-2.2% to 5.5%) 7.9% (2.1% to 13.7%) 6.6% (0.2% to 13.0%) -7.0% (-10.6% to -3.3%) -2.4% (-4.3% to -0.5%) * No 95% CIs reported in Technical Document; see pooled risk difference column (and respective 95% CIs) in reference to importance of differences between quinolones and TMP- SMX # Risk difference reported is quinolone minus TMP-SMX b Disutility refers to an outcome which has a negative impact on preference rankings. 6 Canadian Coordinating Office for Health Technology Assessment

11 Table 2: Estimates of morbid days attributable to clinical outcomes in cystitis Variable Number of days with symptoms Rationale or source of information Initial urinary tract infection Persistent urinary infection Recurrent urinary infection Non-bacterial urethritis responsive to initial therapy; or other diagnosis if self-limited Non-bacterial urethritis not responsive to initial therapy Other diagnosis, not self-limited 2 meta-analysis, reference days while on initial antibiotic treatment, then 2 days waiting for urine culture, then 3 days until symptom response to follow-up therapy 4 2 days of initial antibiotic therapy before representation, then 2 days until symptom response to follow-up therapy 3 has to be 3 days, as otherwise would be seen again 8 3 days while on initial antibiotic therapy, then seen again, then 5 days until symptom response to follow-up therapy 11 assumed Side effects 2 assumed, based on nature of usual side effects EFFICACY The clinical literature comparing fluoroquinolones to alternatives in uncomplicated lower urinary tract infections was only sufficient to establish their relative efficacy and toxicity against TMP-SMX in cystitis. The literature did not establish any difference amongst the fluoroquinolones in terms of efficacy or toxicity for this particular indication. Meta-analysis of the trials comparing fluoroquinolones to TMP-SMX indicated that fluoroquinolones are associated with higher initial rates of symptom relief and fewer recurrent infections and side effects. Specific rates are outlined in Table 1 above. Canadian Coordinating Office for Health Technology Assessment 7

12 EFFECTIVENESS There were no data available regarding the community effectiveness of these drugs with respect to the indication studied. COSTS Given the perspective chosen for analysis, only direct costs were included. Standard costs are outlined in Table 3. To ensure consistency, most costs were estimated from a single source. Specifically, costs were derived from: Alberta Health (October 1996; cost estimates for all antibiotics studied, as well as other relevant drugs); Schedule of Medical Benefits, Alberta Health Care Insurance Plan, 1992 (cost of medical services); and the General Hospital, St. John s, Newfoundland (estimates of laboratory test costs). Table 3: Costs used in the analysis of fluoroquinolones versus TMP-SMX in cystitis Variable Value Initial visit to MD $40.40 Revisit with side effects or persistent symptoms $22.30 Revisit for recurrent infection $22.30 Daily cost, ciprofloxacin 250 mg bid $4.44 Daily cost, ofloxacin 200 mg bid $4.14 Daily cost, norfloxacin 400 mg bid $4.36 Daily cost, TMP-SMX 960 mg bid $0.24 Daily cost, doxycycline 100 mg bid* $1.17 Daily cost, drug used to treat persistent or recurrent urinary tract infection $4.31 (i.e. average of fluororquinolone costs) Dispensing fee $8.90 Course of clotrimazole for vaginitis $10.41 Urine culture $24.45 * For treatment of suspected/probable urethritis 8 Canadian Coordinating Office for Health Technology Assessment

13 COST EFFECTIVENESS ANALYSIS Cost Effectiveness Analysis Based on the clinical outcome and cost information outlined above, expected costs and morbidity days avoided were calculated from the decision model. Outputs of the analysis included the average cost and number of morbid days associated with each drug, along with the incremental cost per day of morbidity avoided. Results of the base case analysis are outlined in Table 4. The results reveal that the incremental costs per morbid days avoided for all of the fluoroquinolones are very similar in magnitude, with most of the difference among them being due to different drug acquisition costs. The analysis showed that the least expensive quinolone cost $2.87 more than TMP-SMX per patient treated, but provided 0.73 fewer days of excess morbidity (i.e. incremental cost per day of $3.93). Thus, the marginal cost per day of morbidity avoided by choosing any one of the fluoroquinolones over TMP-SMX is quite modest. Table 4: Base case analysis* Strategy Cost, $ Morbid days Incremental cost, $ Incremental days Incremental cost per day ($/d) TMP-SMX Ofloxacin (0.73) 3.93 Norfloxacin (0.70) 5.60 Ciprofloxacin (0.71) 5.58 * Note: Confidence intervals for parameters in Table 4 not provided in original Technical Document. Sensitivity Analysis (SA) A number of one-way sensitivity analyses were carried out to determine the effect of altering the estimates of the variables for which reliable figures do not exist, or which are known to have a high degree of variability. The values for the probability of cure, recurrent UTI and side effects with TMP- SMX; number of morbid days under various clinical scenarios; dispensing fees; drug costs for persistent or recurrent UTI; and doctors fees were varied over the limits of plausibility. The model was most sensitive to the probability of cure of the initial infection and the risk of recurrent infection with TMP-SMX. (The authors only provided the results of one particular comparison involving TMP-SMX and ofloxacin in their report. This example was representative of the results which would be obtained with all of the quinolones, and not meant to support one particular fluoroquinolone over another.) At the lower limit of the probability of cure of initial infection with TMP- SMX, ofloxacin was dominant (i.e. both cheaper and more effective); while at the upper limit of TMP- SMX cure, the incremental cost per day of morbidity avoided by using ofloxacin increased to $ When the probability of recurrent infection with TMP-SMX went from highest to lowest, the incremental cost per day of morbidity avoided by using ofloxacin went from $0.95 to $8.95. Canadian Coordinating Office for Health Technology Assessment 9

14 Monte Carlo analysis was carried out to examine the effect on costs, morbid days and cost per morbid day avoided of simultaneously varying the value of some of the variables assessed via one-way sensitivity analysis (Table 5). The analysis demonstrated little variation around the number of morbid days associated with any of the drugs versus the base case analysis (Table 6). The only substantial variation was around the cost of treatment with TMP-SMX, whose upper 95% CI was very similar to the expected cost of the fluoroquinolones. Table 5: Range of values for variables used in the Monte Carlo analysis Variable Base value (range) Probability of cure of initial UTI with TMP-SMX (0.768 to 0.884) Probability of recurrent UTI with TMP-SMX (0.133 to 0.261) Probability of side effects with TMP-SMX (0.152 to 0.228) Number of morbid days due to persistent UTI 8 (4 to 12) Number of morbid days due to recurrent UTI 4 (2 to 6) Table 6: Expected costs and morbid days from Monte Carlo analysis TMP-SMX Ofloxacin Norfloxacin Ciprofloxacin Cost, $ (SD)* (2.14) (0) # (0) # (0) # Morbid days (SD)* 4.73 (0.28) 4.00 (0.13) 4.03 (0.13) 4.02 (0.13) * Standard deviation # Fluoroquinolone costs were not one of the variables run in the Monte Carlo analysis, hence there was no standard deviation calculated for them in the analysis. The only factor likely to have a substantial impact on the relative costs of the drugs compared (but not efficacy, according to clinical trials) is the dose and duration of therapy chosen. Given that the previous differences are based on 3-day treatment regimens for both initial and recurrent disease, if therapy is given for 7 days and all else remains the same the fluoroquinolones would cost $17.81 to $20.11 more per case than TMP-SMX, translating into an incremental cost per day of morbidity avoided of $24.40 to $28.32 (Table 7). 10 Canadian Coordinating Office for Health Technology Assessment

15 Table 7: Effect of changing duration of therapy from 3 to 7 days Strategy Cost, $ Morbid days Incremental, cost $ Incremental days Incremental cost per day ($/d) TMP-SMX Ofloxacin (0.73) Norfloxacin (0.70) Ciprofloxacin (0.71) Interprovincial Analysis The investigators were unable to estimate the Provincial budgetary impact of switching from TMP- SMX to fluoroquinolones for cystitis due to the lack of reliable epidemiologic and prescribing data for this community-based disease state. STUDY LIMITATIONS There were limitations to this evaluation, in terms of the data used to develop the model as well as the methods of the analyses. In terms of the limitations of the available data, the most pressing problem was the fact that there were no controlled studies comparing the three quinolones head on. In addition, the investigators were unable to determine the extent of use of any of these agents, which precluded the development of an accurate estimate of the overall financial impact of any policies which might limit or promote the use of fluoroquinolones for uncomplicated cystitis. Finally, data regarding the effectiveness of these medications in clinical practice were not available. The clinical trials used in the evaluation excluded patients from the analysis if they were lost to follow-up. Without the information regarding cure, recurrence and adverse effects from these patients, it is difficult to determine the overall effectiveness of each drug therapy. There were no standardized definitions used to describe adverse events among the clinical trials reviewed. However, the impact of any differences among these trials would be minimized when studies were pooled in the meta-analysis. Finally, most studies evaluated only those patients with bacteriologically proven UTI, which differs significantly from the most common approach used in daily clinical practice as well as the empiric model developed for the analysis. In terms of the analysis itself, it was limited to a comparison of fluoroquinolones with TMP-SMX. This was due to the limited availability of studies providing comparative efficacy or toxicity of fluoroquinolones versus other antibiotics also commonly used to treat these infections. The costing analysis included data from two jurisdictions; Alberta and Newfoundland. A more comprehensive sensitivity analysis (i.e. two or three way SA) around the cost parameters using data from these Canadian Coordinating Office for Health Technology Assessment 11

16 provinces would have been useful to better demonstrate the impact of using data from two different health care systems. The use of the morbid day as the unit of clinical outcome was a unique means of approaching but not actually achieving a cost-utility analysis. This is not a standard clinical outcome unit, and as such could be challenged for its clinical usefulness. In developing this outcome unit, the investigators assumed that the disutility of adverse effects and the symptoms of urinary tract infections were equal as well as additive. This assumption may or may not be valid. Finally, the perspective of the analysis was limited to that of provincial Ministries of Health. However, the quality of the underlying evidence available for taking the societal perspective was inadequate. 12 Canadian Coordinating Office for Health Technology Assessment

17 REFERENCES 1. Barrett B, Doyle M, Parfrey P, Fardy J, Crewe S, Kent G, et al. An analysis of the use of fluoroquinolones for uncomplicated urinary tract infections, prostatitis and community acquired pneumonia: clinical and economic considerations. Ottawa: Canadian Coordinating Office for Health Technology Assessment (CCOHTA); Carlson KJ, Mulley AG. Management of acute dysuria; A decision analysis model of alternative strategies. Annals of Internal Medicine 1985;102:(2) Canadian Coordinating Office for Health Technology Assessment 13

18 Canadian Coordinating Office for Health Technology Assessment Green Valley Crescent Ottawa, Ontario, Canada K2C 3V4