Use of Indicators to Evaluate the Quality of Community-Acquired Pneumonia Management

Transcription

1 MAJOR ARTICLE Use of Indicators to Evaluate the Quality of Community-Acquired Pneumonia Management Dilip Nathwani, 1 Fiona Williams, 3 John Winter, 2 Janet Winter, 2 Simon Ogston, 3 and Peter Davey 1 1 Infection and Immunodeficiency and 2 Respiratory Units, Tayside University Hospitals Trust, and 3 Department of Epidemiology, University of Dundee, Dundee, United Kingdom Quality-assessment indicators for community-acquired pneumonia (CAP) founded on health care structure, process, and outcome have been recommended as a potential audit tool to evaluate the delivery of care. We prospectively audited the treatment of 205 patients admitted with CAP to 2 hospitals in Dundee against some of these key standards. Patients with severe CAP were more likely to die (mortality rate, 42% versus 7%) and to receive antibiotics by the intravenous route (relative risk [RR], 1.81; 95% confidence interval [CI], ) and within 4 hours of admission to the hospital (RR, 1.22; 95% CI, ). There was a lack of uniformity regarding the amount of oxygen prescribed, with evidence of poor case record and drug prescription chart documentation related to oxygen therapy. Adherence to the recommended antibiotic policy was associated with reduced risk of death or readmission to the hospital (RR, 0.58; 95% CI, ). However, in a multivariate analysis, severity of pneumonia was the strongest predictor of death or readmission ( P p.004), and adherence to the antibiotic policy was not statistically significant ( P p.154). Our study has confirmed the value of quality indicators in evaluating our CAP management and has stimulated the development and implementation of a local hospital based integrated care pathway. Community-acquired pneumonia (CAP) remains a significant cause of morbidity, mortality, and economic burden [1, 2]. Many guidelines from around the world have recognized the importance of CAP [3]. In the United Kingdom, written guidelines for antibiotic therapy for CAP exist in most hospitals [4]; most broadly follow the 1993 guidelines of the British Thoracic Society (BTS) [5]. The aim of these best-practice guidelines is to reduce variation in key aspects of care and, by doing so, to improve the efficiency and effectiveness of health care Received 28 February 2001; revised 15 August 2001; electronically published 13 December Reprints or correspondence: Dr. Dilip Nathwani, Consultant Physician, Infection Ward (Ward 42), East Block, Ninewells Hospital, Tayside University Hospitals, Dundee, United Kingdom (nathwani@globalnet.co.uk). Clinical Infectious Diseases 2002; 34: by the Infectious Diseases Society of America. All rights reserved /2002/ $03.00 [6]. We recently reviewed the current evidence for the clinical and cost effectiveness of CAP guidelines [7]. For guidelines to influence outcome, there should be evidence that links process to outcome and evidence that there are potentially important variations in the process of routine care [7, 8]. The key supposition of many guidelines or care pathways is that a number of processes are linked to outcome. Where there is an association between variations in process and variations in outcome, the implication is that improving the process of care will improve outcome [9]. For CAP, a number of such processes have been identified. These include assessment of severity at the time of presentation [10 13]; time to administration of first antibiotic [14 18], the prescription of appropriate antibiotic (choice and route according to the guideline or protocol [19, 20]); the prescription of an antibiotic by the iv route, for patients with severe pneumonia [21, 22]; the measurement and administration of appropriate amounts of oxygen [11, 13, 15, 23]; and the obtaining of blood 318 CID 2002:34 (1 February) Nathwani et al.

2 samples for culture prior to therapy [15, 24 27]. The link between obtaining blood samples for culture and outcome is one process measure that does not appear to be supported by a number of validity criteria that assess whether an association can be judged as cause and effect [28]. However, the timely delivery of antibiotics does meet these criteria and appears to be biologically plausible, in that there is experimental [29] and clinical evidence [30] to support the assertion that early antibiotic treatment is more effective than delayed treatment. Similarly, timely measurement of oxygen saturation with subsequent appropriate treatment is crucial, because hypoxemia at presentation is independently associated with short-term mortality [11, 13]. In a study we published elsewhere [7], these and other criteria are recommended as potential audit criteria of CAP management. Some of these parameters for audit are outlined in table 1. The aim of the present analysis was to assess the quality of CAP management in the Dundee cohort against these audit criteria. MATERIALS AND METHODS Ascertainment of cases. A research nurse prospectively identified patients in the 2 hospitals in Dundee who had acute medical emergencies and who were admitted to the hospital from October 1999 through March The acute receiving ward was visited on weekdays. The following case definition was used to enter patients into the study: patients who received an antibiotic for a presumed chest infection and had a documented abnormality on a chest radiograph (new infiltrate on chest radiograph) consistent with pneumonia or those who received an antibiotic for whom the diagnosis of pneumonia had been made and noted in the case notes by a consultant or specialist registrar (fellow) in medicine. Names of patients who received an antibiotic for a chest infection or pneumonia were collected prospectively, and onethird of these were randomly selected for detailed audit on discharge, depending on the availability of complete case records. Of the 640 patients admitted with the presumptive diagnosis of lower respiratory tract infection, 205 were selected for audit and had case records available. Assessment of severity. The following information for each patient was recorded: demographic characteristics; initial (within 24 h of admission) assessment of severity (defined by use of the extended BTS criterion [5], in which severe is defined by the presence of 2 of the following characteristics: respiratory rate of 130 beats/min, diastolic blood pressure of!60 mm Hg, urea of 17 mmol/l, and confusion [new onset] at the time of initial assessment); the presence of comorbid illness; and the process of management (investigations, oxygenation, and antibiotic therapy). The patient s clinical outcome (discharge home, discharge to intensive care unit [ICU], death, or readmission within 2 weeks of discharge with the same problem) were also recorded. Adherence to the local antibiotic policy. The local policy recommends a combination of a b-lactam and a macrolide antibiotic for treatment of CAP, which is in line with current British Thoracic Guidelines. The recommended antibiotics are amoxicillin with or without a macrolide, for patients hospitalized with mild to moderate CAP, and iv cefuroxime (or) coamoxiclav with or without a macrolide, for patients with severe CAP. For the purposes of the audit, we allowed ceftriaxone to be included instead of cefuroxime. Although this is not part of our current protocol, ceftriaxone had been included in the previous edition of the antibiotic policy. However, prescription of ceftazidime for CAP was not allowed. Flucloxacillin could be added for patients with severe cases if indicated. Adherence was assessed solely by the choice of antibiotic on the protocol; we did not include route of administration in the assessment of adherence. Statistical analysis. RRs and the corresponding 95% CIs were calculated by use of EpiInfo, version 5.0 (Centers for Disease Control and Prevention and World Health Organization). We used multiple logistic regression analysis (SPSS for Windows, Advanced Statistics Release 6.0 [SPSS]) to determine the effect on outcome of adherence to the local antibiotic policy, adjusted for severity of pneumonia. Ethics approval. Audit of case records supervised by the consultants responsible for their inpatient management does not require approval by the Tayside Research Ethics committee. RESULTS Of the 205 patients in the study, 105 (51%) were male and 100 (49%) were female; 74 (36%) were aged 60 years, 43 (21%) were aged years, and 88 (43%) were aged 175 years. The time of admission was accurately recorded for 170 patients (83%); the time of admission was between 8:00 a.m. and 4:00 p.m. for 82 patients (48%), 4:00 p.m. and midnight for 71 (42%), and after midnight for the remaining 17 (10%). We were able to compare our performance against 11 of 17 quality indicators, which we identified from a recent review of the literature (table 1). Compliance with protocol and process of treatment. We had complete information on antibiotic therapy for 196 of the 205 patients. The median durations of iv and oral therapy, respectively, were 3 days and 9 days. The median duration of antibiotic treatment was 4.6 days. Only 102 (52%) of 196 of patients received antibiotics within 4 h of admission. Of these patients, those with severe CAP were more likely to receive antibiotics promptly and via the iv route during the first 24 h of hospitalization than were the others (table 2). The time of admission did not influence the proportion of patients who CAP Management CID 2002:34 (1 February) 319

3 Table 1. Suggested quality indicators for hospitalized patients with community-acquired pneumonia (CAP; adapted from [7]). Quality or outcome indicator Data collected in Dundee during the study period Processes that are likely to improve the outcome of CAP Posterior-anterior chest radiograph performed within 24 h of admission 80% Proven indicators of severity of CAP documented in the medical case records 82% of patients with severe pneumonia; 48% of patients with nonsevere pneumonia with all 4 severity indicators documented Brief interval between admission to hospital and initiation of appropriate 52% of patients received antibiotics within 4 h of admission antibiotic therapy Receipt of an antibiotic regimen active against all of the likely causative pathogens 72% Receipt of iv antibiotics for severe pneumonia 71% (within 24 h of admission) Receipt of adequate oxygenation and respiratory support for severe pneumonia 82% of all patients received oxygen therapy prescribed; 98% of those in the severe group received oxygen therapy Receipt of adequate fluid replacement for severe pneumonia Brief median interval between diagnosis of respiratory failure and transfer to the intensive care unit Processes that are likely to increase the costs of management of CAP Admission to the hospital for patients at low risk Receipt of unnecessarily intensive or expensive antibiotic treatment 17% Failure to switch from iv to oral therapy, according to existing criteria and clinical stability Discharge at 124 h after switching to oral therapy Outcome indicators a Duration of hospital stay, median days 7 Requirement of intensive care unit admission, % 2 30-day mortality rate, % 14.6 Readmission within 30 days of hospital discharge with an associated illness 7.2% admitted within 2 weeks of discharge Cost of care a One would need to identify the presence of comorbid illness and severity indicators in all patients. received antibiotics within 4 h of admission. During the 3 admission time frames of 8:00 a.m. to 4:00 p.m., 4:00 p.m. to midnight, and midnight to 8:00 a.m., similar percentages of patients received antibiotics within 4 h (44 [54%] of 82, 44 [62%] of 71, and 10 [61%] of 17, respectively; x 2, 1.02 with 2 degrees of freedom; P p.594). The antibiotic choice was correct according to the local protocol in 138 (70%) of the 196 patients in whom this could be assessed. Failure to use a macrolide or combination of antibiotics without activity against atypical organisms accounted for 40 cases (69%) of noncompliance with the protocol. The proportion of patients who received antibiotics that are active against atypical bacteria was not related to severity of pneumonia (38 [79%] of 48 patients with severe cases vs. 118 [80%] of 148 patients with nonsevere cases). Only in a very small number of cases (10 of 58) was noncompliance the result of treatment with such agents as ceftazidime, doxycycline, or quinolones. Compliance with the recommended antibiotic policy was associated with reduced risk of death or readmission to hospital (RR, 0.58; 95% CI, ; table 2). However, in the multivariate analysis (table 3), severity of pneumonia was the strongest predictor of death or readmission ( P p.004), and adherence to the antibiotic policy was not statistically significant ( P p.154). Data about oxygenation were available for all 205 patients audited. According to either medical or nursing notes, 82% of the patients received oxygen therapy. Only 70% of the patients in this group had oxygen prescribed in the drug prescription chart. The concentration of oxygen administered varied considerably, from 24% to 100%, but this did not appear to be guided by initial oximetry or blood gas levels. Overall, 170 (83%) of 205 of patients had oxygen saturations performed within 24 h of admission, but patients with severe pneumonia were no more likely than other patients to have oxygen saturations performed or repeated within the first 24 h of admission (table 2). DISCUSSION Our data confirm that patients with cases stratified as severe, according to the BTS criteria [5], are significantly more likely to die (irrespective of the presence of comorbid illness) or to remain in the hospital for a longer period of time. These patients should be treated in a high-dependency unit or ICU. The 320 CID 2002:34 (1 February) Nathwani et al.

4 Table 2. Comparison of process and outcome measures for management of community-acquired pneumonia (CAP) in patients with severe or nonsevere CAP. Relation to antibiotic protocol Characteristic Patients with severe CAP (n p 48) Patients with nonsevere CAP (n p 148) Outside protocol (n p 58) Within protocol (n p 146) RR (95% CI) Process or outcome of CAP management Died 20 (42) 10 (7) 6.17 ( ) Received antibiotics within 4 h of hospital admission 29 (61) 73 (49) 1.22 ( ) Received iv antibiotics within 24 h of hospital admission 34 (71) 58 (39) 1.81 ( ) Blood gas levels determined within 24 h of hospital admission 44 (92) 126 (85) 1.08 ( ) Blood gases or O 2 saturation repeated within 24 h of hospital admission, n/n (%) 29/44 (66) 80/126 (63) 1.12 ( ) Antibiotic therapy in relation to protocol Patient died or was readmitted to the hospital 17 (29) 25 (17) 0.58 ( ) NOTE. Data are no. (%) of patients, unless otherwise indicated. low mortality rate in the patients with nonsevere pneumonia confirms increasing evidence [17] that many patients currently admitted to the hospital with pneumonia can be managed in the ambulatory or outpatient setting by use of well-validated risk stratification criteria [5, 18 21]. Compliance with the choice of antibiotics recommended by the BTS was associated with reduced risk of death and of readmission within 2 weeks of hospital discharge. However, the association was not statistically significant after adjustment for severity of pneumonia. Nonetheless, our results are consistent with those of a number of studies that have supported the potential positive effect of guidelines or protocols on the outcome of patients who are hospitalized with CAP [7]. Simple cohort studies are always subject to confounding and bias. However, randomized trials of organizational interventions in hospitals present considerable methodological challenges, because, by necessity, they require the participation of several hospitals, given that the unit of randomization is the hospital, not the patient. [31] The Cochrane Effective Practice and Organisation of Care (EPOC) group recommends the following 2 quasi experimental study designs as alternatives to randomized trials: controlled before-and-after studies and interrupted Table 3. Multiple logistic regression of the relationship between adherence to antibiotic protocol or severity of pneumonia and outcome (death or readmission). Variable b SE Wald Exp (b) P Protocol Severity Constant NOTE. Exp (b), exponential function (logistic regression coefficient). time-series analyses. Unfortunately, the published literature on quality improvement interventions in CAP consists of uncontrolled before-and-after studies [7], which would not meet the quality criteria for inclusion in an EPOC systematic review (see the Data Collection Checklist at health/hsru/epoc/). It is important that future studies are designed to meet the EPOC quality criteria. Some recent evidence [32] has suggested that antibiotic choice may be particularly important for elderly patients. This may have influenced the guidelines developed by the Infectious Disease Society for CAP [33], in which a new 4-fluoroquinolone has been introduced as an alternative regimen to a macrolide combined with either cefuroxime or ceftriaxone for hospitalized patients. The use of a 4-fluoroquinolone is thought to be particularly of importance in the elderly and nursing home populations [34]. Good compliance with our protocol is encouraging, because it provides some measurable evidence of our success in implementation of such protocols through posters and a continuing education program. Although compliance may not result directly in improved health outcome, it shows that clinicians are supporting good practice by reducing the use of unnecessary agents [35] or combinations. There is growing concern about inappropriate use of these agents in particular, extended-spectrum cephalosporins, which are important drivers of antibiotic resistance [36] and the occurrence of Clostridium difficile infections [37]. Although it was encouraging that patients with severe pneumonia were more likely to receive antibiotics by the iv route and within 4 h of admission, there is still plenty of room for improvement, because 39% of the patients with severe pneu- CAP Management CID 2002:34 (1 February) 321

5 monia did not receive prompt appropriate treatment (table 2). This may indicate a lack of appreciation of the factors that should stimulate iv therapy, despite the fact that these are clearly outlined in our empiric sepsis-management protocol, which is prominently displayed in all the medical units. The use of oral therapy in the patients with severe pneumonia may also reflect confidence in oral therapy for patients with severe cases [38], which many may argue is appropriate [21] in patients with uncomplicated cases. In fact, decisions regarding the use of the iv route are largely made on the basis of the theoretical advantage that high serum levels of antibiotics have for therapeutic cure, a fact that may be relevant in intravascular infections or cases in which there are barriers to penetration, such as necrotizing pneumonia; however, this would appear to be less relevant in patients with pneumonia, given that the lung is a vascular structure and that penetration is excellent [39]. Nonetheless, the existing evidence in favor of oral therapy comes from studies with a low proportion of severe cases [38]. The median duration of iv therapy in our study (3 days) was consistent with that in studies published elsewhere [19, 39]. It was reassuring that 92% of the patients with severe pneumonia had oxygen saturations performed with 24 h of admission. However, only 66% of the patients in the severe group had an additional assessment performed after another 24 h, as recommended in our protocol. Furthermore, only 18% of the patients who received oxygen therapy had any documentation of this treatment in the medical or nursing case notes, despite the legal requirement that oxygen prescriptions be written in the notes. The fact that only 62% of the patients had oxygen prescribed in the drug prescription chart is also a cause of concern. Even when oxygen treatment was documented, the amount of oxygen therapy (e.g., concentrations of 35%, 60%, or 100%) was not adequately noted or justified in the case notes. We have identified similar evidence of poor case record documentation in the management of respiratory tract infection and other infections elsewhere [40]. An improvement in the process of administering oxygen therapy by implementing a multidisciplinary multimethod pathway can lead to changes in oxygen prescription behavior, but this has been found to be time consuming and costly, and it has not led to changes in patient outcome. [41]. Despite this, we need to encourage more-careful documentation of oxygen therapy in our practice and to ensure that patients with severe pneumonia have additional assessments of their oxygenation when appropriate. Regular evaluation of clinical performance through audit and feedback of information to stakeholders as a means of promoting change is now encouraged as part of the quality improving agenda of Clinical Governance within the United Kingdom [42]. The emphasis is on implementation and evaluation of clinical care. Identification of a champion for managing CAP effectively is a first important step toward realizing this goal. This may be a physician or specialist nurse who may potentially also be involved in the care of several patients with other illnesses that require antibiotics. The benefit, for example, of use of a nurse practitioner for providing ambulatory-care parenteral antibiotic therapy in our infection service has already been shown [43]. We aim to complete the audit loop by introducing a care pathway for CAP [44 46] in 1 of our hospitals, reauditing our practice in months. The effect of this intervention on outcome will be assessed by use of a controlled before-andafter study design. References 1. Woodhead MA, McFarlane JT, McCraken JS, Rose DH, Finch RG. Prospective study of the aetiology and outcome of pneumonia in the community. Lancet 1987; 1: Guest JF, Morris A. Community acquired pneumonia: the annual cost to the National Health Service in the United Kingdom. Eur Respir J 1997; 10: Woodhead M. Community-acquired pneumonia guidelines an international comparison: a view from Europe. Chest 1998; 113:183S 7S. 4. Woodhead M, McFarlane J, for the BTS CAP Guideline Committee. Local antibiotic guidelines for adult community acquired pneumonia (CAP): a survey of UK hospital practice in J Antimicrob Chemother 2000; 46: British Thoracic Society. Guidelines for the management of community acquired pneumonia in adults admitted to hospital. Br J Hosp Med 1993; 49: Miller J, Petrie J. Development of practice guidelines. Lancet 2000; 355: Nathwani D, Rubenstein E, Barlow G, Davey P. Do guidelines for community acquired pneumonia improve the cost-effectiveness of hospital care? Clin Infect Dis 2001; 32: Thomson R, Lavender M, Madhok R. How to ensure that guidelines are effective. BMJ 1995; 311: Kahn KL, Rogers WH, Rubenstein TV, et al. Measuring quality of care with explicit process criteria before and after implementation of the DRG-based prospective payment system. JAMA 1990; 264: Atlas SJ, Benzer TI, Borowsky LH, et al. Safely increasing the proportion of patients with community-acquired pneumonia treated as outpatients: an interventional trial. Arch Intern Med 1998; 158: Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify lowrisk patients with community acquired pneumonia. N Engl J Med 1997; 336: Fine MJ, Hanusa BH, Lave JT, et al. Comparison of disease specific and a generic severity illness measure of patients with community acquired pneumonia. J Gen Intern Med 1995; 10: Neill AM, Martin IR, Weir R, et al. Community acquired pneumonia: aetiology and usefulness of severity criteria on admission. Thorax 1996; 51: Kahn KL, Rogers WH, Rubenstein LV, et al. Measuring quality of care with explicit process criteria before and after implementation of the DRG-based prospective payment system. JAMA 1990; 264: Meehan TP, Fine MJ, Krumholz HM, et al. Quality of care, process, and outcomes in elderly patients with pneumonia. JAMA 1997; 278: Gleason PP, Meehan TP, Fine MJ, Galusha DH, Fine MJ. Associations between initial antimicrobial therapy and medical outcomes for hospitalized elderly patients with pneumonia. Arch Intern Med 1999; 159: Stahl JE, Barza M, DesJardin J, Martin R, Eckman MH. Effect of macrolides as part of initial empiric therapy on length of stay in patients 322 CID 2002:34 (1 February) Nathwani et al.

6 hospitalized with community-acquired pneumonia. Arch Intern Med 1999; 159: McGarvey RN, Harper JJ. Pneumonia mortality reduction and quality improvement in a community hospital. QRB Qual Rev Bull 1993; 19: Halm EA, Fine MJ, Marrie TJ, et al. Time to clinical stability in patients hospitalized with community-acquired pneumonia: implications for practice guidelines. JAMA 1998; 279: Davis D, O Brien MA, Freemantle N, Wolf FM, Mazmanian P, Taylor- Vaisey A. Impact of formal continuing medical education: do conferences, workshops, rounds, and other traditional continuing education activities change physician behaviour or health care outcomes? JAMA 1999; 282: Chan R, Hemeryck L, O Regan M, Clancy L, Feely J. Oral versus intravenous antibiotics for community acquired lower respiratory tract infection in a general hospital: open, randomised controlled trial. BMJ 1995; 310: MacGregor RR, Grazianni AL. Oral administration of antibiotics: a rational alternative to the parenteral route. Clin Infect Dis 1997; 24: Fulmer JD, Snider GL. American College of Chest Physicians/National Heart, Lung and Blood Institute National Conference on Oxygen Therapy. Heart Lung 1984; 13: Arbo MD, Snydman DR. Influence of blood culture results on antibiotic choice in the treatment of bacteremia. Arch Intern Med 1994; 154: Bryan CS. Clinical implications of positive blood cultures. Clin Microbiol Rev 1989; 2: Rintala E, Kairisto V, Eerola E, Nikoskelainen J, Lehtonen O-P. Antimicrobial therapy of septicaemic patients in intensive care units before and after blood culture reporting. Scand J Infect Dis 1991; 23: Schonheyder HC, Hojbjerg T. The impact of the first notification of positive blood cultures on antibiotic therapy: a one-year survey. APMIS 1995; 103: Hennekens CH, Buring JE. Statistical association and cause-effect relationships. In: Mayrent SL, ed. Epidemiology in medicine, first ed. Toronto: Little, Brown and Company, 1987: Davey PG, Barza M, Stuart M. Dose response of experimental Pseudomonas endophthalmitis to ciprofloxacin, gentamicin and imipenem: evidence for resistance to late treatment of infections. J Infect Dis 1987; 155: Fine MJ, Medsger AR, Stone RA, et al. The hospital discharge decision for patients with community-acquired pneumonia: results from the Pneumonia Patient Outcomes Research Team cohort study. Arch Intern Med 1997; 157: Ukoumunne C, Gulliford MC, Chinn S, Sterne JA, Burney PG, Donner A. Methods in health service research: evaluation of health interventions at area and organisation level. BMJ 1999; 319: Gleason PP, Meehan TP, Fine JM, et al. Associations between initial antimicrobial therapy and medical outcomes for hospitalised elderly patients with pneumonia. Arch Intern Med 1999; 159: Bartlett JG, Breiman RF, Mandell LA, File TM. Community-acquired pneumonia in adults: guidelines for management. Clin Infect Dis 1998; 26: Marrie TJ. Community-acquired pneumonia in the elderly. Clin Infect Dis 2000; 31: Department of Health Advisory Committee Sub-Group on Antimicrobial Resistance. The path of least resistance. London, United Kingdom: Department of Health, John JF, Fishman NO. Programmatic role of the infectious diseases physician in controlling antimicrobial costs in the hospital. Clin Infect Dis 1997; 24: Zadik PM, Moore AP. Antimicrobial associations of an outbreak of diarrhoea due to Clostridium difficile. J Hosp Infect 1998; 40: Bartlett JG. Impact of new oral antibiotics on the treatment of infectious diseases. Infect Dis Clin Pract 1994; 2: Siegel RE. How long a stay in hospital is needed for patients with community acquired pneumonia. Am J Med 2000; 109: Seaton RA, Nathwani D, Philips G, Millar R, Davey P. Clinical record keeping in patients receiving antibiotics in hospital. Health Bull 1999; 57: Wong C, Visram F, Cook D, et al. Development, dissemination, implementation and evaluation of a clinical pathway for oxygen therapy. CMAJ 2000; 162: Scally J, Donaldson LJ. Clinical governance and drive for quality improvement in the new NHS in England. BMJ 1998; 317: Nathwani D, Morrison J, Seaton R, et al. Outpatient and home parenteral antibiotic therapy (OHPAT): evaluation and impact of one years experience in Tayside. Health Bull 1999; 57: Marrie TJ, Lau CY, Wheeler SL, Wong CJ, Vandervoort MK, Feagan BG. A controlled trial of a critical pathway for treatment of communityacquired pneumonia. CAPITAL Study Investigators. Community-Acquired Pneumonia Intervention Trial Assessing Levofloxacin. JAMA 2000; 283: Al-Eidan FA, McElnay JC, Scott MG, Kearney MP, Corrigan J, Mc- Connell JB. Use of a treatment protocol in the management of community-acquired lower respiratory infection. J Antimicrob Chemother 2000; 45: Fine MJ, Pratt HM, Oborsky S, et al. Relation between length of hospital stay and cost of care for patients with community acquired pneumonia. Am J Med 2000; 109: CAP Management CID 2002:34 (1 February) 323