Skip Navigation

Routine Use of the Pneumonia Severity Index for Guiding the Site-of-Treatment Decision of Patients with Pneumonia in the Emergency Department: A Multicenter, Prospective, Observational, Controlled Cohort Study

  1. Bertrand Renaud1,
  2. Eva Coma1,14,
  3. Jose Labarere4,
  4. Jan Hayon5,
  5. Pierre-Marie Roy6,
  6. Hélène Boureaux7,
  7. Fabienne Moritz8,
  8. Jean François Cibien9,
  9. Thomas Guérin10,
  10. Emmanuel Carré11,
  11. Armelle Lafontaine13,
  12. Marie Pierre Bertrand3,
  13. Aline Santin1,
  14. Christian Brun-Buisson2,
  15. Michael J. Fine15,
  16. Eric Roupie1,12, and
  17. for the Pneumocom Study Investigators
  1. 1Department of Emergency Medicine, Créteil
  2. 2Intensive Care Unit, Centre Hospitalier Universtaire Henri Mondor (Assistance Publique-Hopitaux de Paris), Créteil
  3. 3Emergency Department, Centre Hospitalier Intercommunal de Créteil, Créteil
  4. 4Quality of Care Unit, Centre Hospitalier Universitaire de Grenoble, Grenoble
  5. 5Intensive Care Unit, Centre Hospitalier Intercommunal de Poissy Saint-Germain, Saint-Germain-en-Laye
  6. 6Emergency Department, Centre Hospitalier Universtaire d'Angers, Angers
  7. 7Emergency Department, Centre Hospitalier Universtaire de Poitiers, Poitiers
  8. 8Emergency Department, Centre Hospitalier Universtaire de Rouen, Rouen
  9. 9Emergency Department, Centre Hospitalier de Montauban, Montauban
  10. 10Emergency Department, Centre Hospitalier de Roanne, Roanne
  11. 11Emergency Department, Centre Hospitalier de Vannes, Vannes
  12. 12Emergency Department, Centre Hospitalier Universitaire de Caen, Caen
  13. 13Centre Hospitalier Intercommunal de Tarbes, Tarbes, France
  14. 14Department of Emergency Medicine, Hospital de Mataró, Consorci Sanitari del Maresme, Mataró, Spain
  15. 15Center for Health Equity Research and Promotion, Veterans Affairs Pittsburgh Healthcare System, Pittsburgh, Pennsylvania
 
Next Section

Abstract

Background. Although the Pneumonia Severity Index (PSI) has been extensively validated, little is known of the impact of its routine use as an aid to site-of-treatment decisions for patients with pneumonia who present to emergency departments (EDs).

Methods. A prospective, observational, controlled cohort study of patients with pneumonia was conducted in 8 EDs that used the PSI (PSI-user EDs) and 8 EDs that did not use the PSI (PSI-nonuser EDs) in France. The outcomes examined included the proportion of “low-risk” patients (PSI risk classes I–III) treated as outpatients, all-cause 28-day mortality, admission of inpatients to the intensive care unit, and subsequent hospitalization of outpatients.

Results. Of the 925 patients enrolled in the study, 472 (51.0%) were treated at PSI-user EDs, and 453 (49.0%) were treated at PSI-nonuser EDs; 449 (48.5%) of all patients were considered to be at low risk. In PSI-user EDs, 92 (42.8%) of 215 patients at low risk were treated as outpatients, compared with 56 (23.9%) of 234 patients at low risk in PSI-nonuser EDs. The adjusted odds ratios for outpatient treatment were higher for patients in PSI risk classes I and II who were treated in PSI-user EDs, compared with PSI-nonuser EDs (adjusted odds ratio, 7.0 [95% confidence interval, 2.0–25.0] and 4.6 [95% confidence interval, 1.3–16.2], respectively), whereas the adjusted odds ratio did not differ by PSI-user status among patients in risk class III or among patients at high risk. After adjusting for pneumonia severity, mortality was lower in patients who were treated in PSI-user EDs; other safety outcomes did not differ between patients treated in PSI-user and PSI-nonuser EDs.

Conclusions. The routine use of the PSI was associated with a larger proportion of patients in PSI risk classes I and II who had pneumonia and who were treated in the outpatient environment without compromising their safety.

Hospitalization rates for community-acquired pneumonia vary from 22% to 42% in European countries [1], with inpatient treatment costing ∼8 times as much as outpatient care [2]. The Pneumonia Severity Index (PSI) is a prediction rule that classifies patients who have pneumonia into 5 strata of increased risk for short-term mortality on the basis of 20 variables that are routinely available at presentation [3]. Patients assigned to risk classes I–III have a cumulative mortality of <1% and are defined as being at “low risk,” whereas patients in risk classes IV and V have mortality rates of 9.0%–30.0% and are defined as being at “high-risk” [3]. This disease-specific prediction rule has been extensively validated in prospective and retrospective studies using a variable duration of follow-up for mortality [3,4,5,6–7].

Although it was initially developed as a prediction rule to identify patients who were at low risk for mortality, studies conducted in North America have shown that implementation of the PSI in the emergency department (ED) increased the outpatient treatment rates of patients at low risk who have community-acquired pneumonia, without compromising their safety [8, 9]. In addition, a recently published, randomized, controlled trial of 224 patients in PSI risk classes II and III revealed that outpatient treatment was as safe as inpatient treatment and provided greater patient satisfaction [10]. As a result, several guidelines recommend the use of the PSI to aid in identification of patients with community-acquired pneumonia at low risk who are appropriate candidates for outpatient treatment [11,12–13].

Although some EDs have systematically adopted the PSI to guide the hospitalization decision for patients who have community-acquired pneumonia, others have not, and little is known about the effectiveness and safety of routine use of the PSI for guiding the site-of-treatment decision in daily practice, particularly outside of North America.

Within this context, we designed a multicenter, prospective, observational, controlled study to assess the effectiveness and safety of routine use of the PSI to guide the initial site-of-treatment decision for patients who have community-acquired pneumonia who present to the ED.

Previous SectionNext Section

Methods

Study sites and design. We approached a total of 25 EDs, each with an annual total volume of >10,000 visits, and 16 EDs eventually volunteered to participate in the study. Eight of these 16 participating EDs (hereafter referred to as PSI-user EDs) had, independently, already implemented the use of the PSI prior to initiation of our study as a part of their pneumonia management policy; the remaining 8 EDs (hereafter referred to as PSI-nonuser EDs) did not use the PSI to manage cases of community-acquired pneumonia. Taken together, the 16 participating EDs had ∼643,000 visits per year (339,000 visits per year in PSI-user and 304,000 visits per year in PSI-nonuser EDs). Four and 3 of the PSI-user and PSI-nonuser EDs were university-affiliated, respectively; 4 PSI-user EDs had an annual volume of <40,000 visits per year, whereas 3 of the 8 PSI-nonuser EDs reached this volume.

Study procedures. Patients were enrolled at each center during a 3–5-month period between February 2002 and July 2003. Before beginning patient enrollment, physicians at each ED attended a single information session about the study' design and data collection methods for patients suspected of having community-acquired pneumonia. The only additional intervention performed at PSI-user EDs was to provide physicians with poster and pocket card reminders that explained how to assign PSI risk class and encouraged outpatient therapy for patients at low risk.

During the study monitoring, baseline chest radiographs and completed data collection forms were reviewed and checked for accuracy by 2 study investigators (B.R. and E.C.). The study protocol and patient informed consent procedures were approved by the French Critical Care Society Ethical review board (Comité d'Éthique de la Société de Réanimation de Langue Française).

Patients. Patient inclusion criteria were: (1) age ⩾ 18 years; (2) acute onset of at least 2 of the following symptoms or signs suggestive of pneumonia: fever, cough, sputum production, dyspnea, chest pain, confusion, and abnormal pulmonary physical findings; (3) acute radiographic evidence of pneumonia at presentation; and (4) provision of informed consent to study participation by the patient or by a relative.

Patients exclusion criteria on the basis of a perceived need for hospitalization were: (1) having HIV infection or being otherwise immunocompromised, (2) having a prior diagnosis of cystic fibrosis, or (3) breathing via a tracheotomy or receiving chronic mechanical ventilation. We also excluded patients who had been discharged from the hospital within the 7 days preceding presentation to the ED (i.e., patients with health care—associated pneumonia).

Standardized screening forms were used in each center to help identify eligible patients; these forms were placed, according to local logistics and staffing patterns, either in the medical records of each patient, on chest radiograph request forms, or in computerized patient medical records. These forms reminded medical staff of a patient' potential for enrollment in the study if at least 1 diagnostic code suggested a lower respiratory tract infection. Patients who met inclusion criteria, consented to participate, and had none of the exclusion criteria were enrolled in the study. All participating EDs enrolled patients 24 h a day, 7 days a week, with the exception of 1 ED that enrolled patients only on weekdays.

Data collection. Baseline demographic characteristics and clinical data at presentation were collected by ED physicians through patient interviews and standardized review of medical records. Variables that were recorded included coexisting illnesses, symptoms, and clinical findings (confusion, respiratory rate, heart rate, blood pressure, temperature, and pulse O2 saturation), as well as the results of laboratory tests, if performed (e.g., WBC count and hematocrit, blood urea nitrogen, glucose, sodium, and arterial blood gas levels). Chest radiographs and reports were reviewed by the local coinvestigators to confirm the presence of an infiltrate and to grade their extension (as 1 vs. ⩾ 2 lobes involved), and to determine the presence of pleural effusion. When decided, the reason for hospital admission (including the PSI class at PSI-user EDs) was documented by ED physicians.

After completion of the prospective data collection, we calculated the PSI risk score from all 20 prognostic variables and determined the PSI risk class for each patient, regardless of the PSI-user status of the ED [3]. For all dichotomous variables included in the PSI, missing values were assumed to be normal (as was done in the original derivation and validation cohorts of the PSI) [3, 14]. Patients aged <51 years who had none of the 5 comorbid illnesses and none of the 5 physical examination prognostic factors were assigned to risk class I. All remaining patients were assigned to risk classes II–V based on their calculated PSI score. We defined patients in risk classes I–III as being at “low risk” and those in risk classes IV or V as being at “high risk.”

Study outcomes. The primary outcome measure was the proportion of patients at low risk who were treated as outpatients. We also assessed this proportion among patients at high risk, a subgroup for which inpatient treatment was recommended by practice guidelines [11,12–13]. We defined outpatient treatment as discharge from the ED to any outpatient environment after the index ED visit. All other patients were defined as inpatients, regardless of their length of stay in the hospital, and we examined the actual reason—including the PSI score at PSI-user EDs—for hospital admission.

Secondary safety outcomes, assessed at 28 days following presentation, were all-cause mortality, admission to an intensive care unit (ICU) for inpatients, and subsequent hospitalization for outpatients. Local investigators conducted follow-up telephone interviews with the patient, a relative, or the family practitioner at least 28 days after presentation. Follow-up hospital data were collected using a structured review of medical records covering the first 28 days after admission.

Previous SectionNext Section

Analysis

We compared baseline characteristics for patients enrolled at PSI-user and PSI-nonuser EDs using the χ 2 test or Fisher' exact test, where appropriate. In bivariate and multivariable analyses, we used multilevel logistic regression to analyze the binary primary and secondary outcomes, with the 2 levels defined by patient and ED site [15]. In multivariable analysis, we calculated adjusted ORs and associated 95% CIs for outpatient treatment of patients at low risk, using a parsimonious 2-level logistic regression model with a cutoff point of P = .05. Secondary safety outcomes were compared between PSI-user and PSI-nonuser EDs, after adjusting for PSI risk class. Two-level logistic regression models were estimated using the Metropolis-Hastings Markov Chain Monte Carlo algorithm in MLwiN, version 2 (University of London) [16]. Two-sided P values < .05 were considered to be statistically significant.

Previous SectionNext Section

Results

Of the 1295 patients presenting with a clinical suspicion of community-acquired pneumonia, 970 fulfilled all inclusion criteria and were enrolled. Subsequently, 45 were excluded because of post hoc discovery of an exclusion criterion (n = 7) or assignment of a final diagnosis other than community-acquired pneumonia (n = 38). The final study population included 925 patients, 472 (51.0%) of whom were from PSI-user EDs and 453 (49.0%) of whom were from PSI-nonuser EDs. The median number of patients enrolled per site was 56 (interquartile range [IQR], 39–74) and the overall incidence rate of community-acquired pneumonia in participating EDs was 0.7%. At PSI-user EDs, the proportion of case report forms that included evidence of PSI calculation exceeded 98%.

Patient characteristics. The median age of all patients was 71 years (IQR, 52–82 years); 590 patients (63.8%) were male, and 449 (48.5%) were patients at low risk (i.e., assigned to PSI risk classes I–III) ( table 1). With the exception of recent influenza immunization, clinical characteristics not included in the PSI did not differ among patients treated in PSI-user EDs and patients treated in PSI-nonuser EDs. Patients who were treated in PSI-nonuser EDs had a higher prevalence of cerebrovascular disease within the high-risk patient group (54 [21.0%] vs. 28 [12.8%]; P < .01), and of chronic renal failure within PSI risk class IV (13 [8.1%] vs. 5 [3.0%]; P = .05).

Site of treatment. Overall, 373 patients (79%) who presented at PSI-user EDs and 388 (85.6%) of those who presented at PSI-nonuser EDs were hospitalized (P < .01). Among patients at low risk, 148 (33%) of 449—including 92 (42.8%) of 215 patients at PSI-user EDs and 56 (23.9%) of 234 patients at PSI-nonuser EDs—were treated as outpatients (P < .01). Conversely, 16 (3.4%) of 476 patients at high risk were treated as outpatients; this rate did not differ between PSI-user and PSI nonuser EDs (OR, 0.6; 95% CI, 0.2–2.5) ( figure 1).

Figure 1
View larger version:
Figure 1

The percentage of patients who were treated as outpatients within each pneumonia severity index (PSI) class at emergency departments that used the PSI (PSI-user EDs) and emergency departments that did not use the PSI (PSI-nonuser EDs).

After excluding the 90 patients at low risk who had evidence of oxygen desaturation (O2 pressure <60 mmHg or O2 saturation <90%) and 41 patients at low risk for whom assessment of arterial oxygenation was not performed, 87 (50.9%) of 171 patients who were treated at PSI-user EDs and 46 (29.3%) of 157 patients who were treated at PSI-nonuser EDs were treated as outpatients (P < .01). Only 3 patients at low risk who had evidence of oxygen desaturation were treated as outpatients—2 (5.1%) at PSI-user EDs and 1 (2.0%) at a PSI-nonuser ED, respectively (P = .57).

A patient' PSI risk class often appeared to be incorporated into the decision process for hospital admission at PSI-user centers ( table 2). In bivariate analysis, patients at low risk presenting at PSI-user EDs were more frequently treated as outpatients ( table 3). The proportion of patients at low risk treated as outpatients ranged from 0.0% to 59.1% at PSI-nonuser EDs and from 22.2% to 75.0% at PSI-user EDs; however, this higher rate for PSI-user EDs was restricted to PSI risk classes I and II ( figure 1).

Table 1
View larger version:
Table 1

Baseline characteristics of patients treated in emergency departments (EDs) that used the pneumonia severity index (PSI) to aid in site-of-treatment decisions for patients with community-acquired pneumonia (PSI-user EDs) and EDs that did not use the PSI (PSI-nonuser EDs).

Table 2
View larger version:
Table 2

Reasons for hospitalization for patients at low risk and patients at high risk documented by physicians at emergency departments (EDs) that used the pneumonia severity index (PSI) to aid in site-of-treatment decisions for patients with community-acquired pneumonia (PSI-user EDs) and EDs that did not use the PSI (PSI-nonuser EDs).

Table 3
View larger version:
Table 3

Comparisons of baseline characteristics between 449 outpatients and inpatients within the low-risk category.

In multivariable analysis, the odds of receiving outpatient treatment were higher at PSI-user EDs, compared with PSI-nonuser EDs, for patients in PSI risk classes I or II, whereas the odds of receiving outpatient treatment did not differ by ED status for patients in PSI risk class III ( table 4). The presence of ⩾ 1 comorbid condition not included in the PSI risk score assessment, an abnormal WBC count, and multilobar radiographic infiltrates were all independently associated with decreased odds of receiving outpatient treatment.

Table 4
View larger version:
Table 4

Independent associations of patients and emergency department (ED) characteristics with outpatient treatment for 430 patients at low risk for mortality.

Microbiological characteristics and antibiotic management. Microbiological documentation of an etiology was obtained for 19% of patients overall, with no difference observed between PSI-user EDs and PSI-nonuser EDs. Overall, bacteremia was recorded for 47 patients (5.1%), including 13 patients at low risk (2.9%) and 34 patients at high risk (7.1%) ( table 5). There were some differences in initial antibiotic management between patients treated at PSI-user and PSI-nonuser EDs, including a higher rate of single therapy with amoxicillin and a lower rate of combined therapy prescribed to inpatients at PSI-user EDs.

Table 5
View larger version:
Table 5

Microbiological etiology of bacteremia among patients with community-acquired pneumonia, by risk of mortality.

Safety outcomes. Overall, mortality was 10.6% (98 of 925 patients died) for the total study cohort, including 7 (1.6%) of 449 patients at low risk, with all but 1 death occurring in patients who were assigned PSI risk class III ( table 6). No death occurred in the 148 patients at low risk who were initially treated as outpatients, and 7 deaths (2.3 %) occurred among the 301 patients at low risk who were treated as inpatients. Among patients at high risk, death occurred in 4 (25%) of 16 persons treated as outpatients and in 87 (18.9%) of 460 inpatients. After adjusting for PSI risk class, mortality was significantly lower among patients treated in PSI-user EDs. No differences in subsequent hospitalization for outpatients or ICU admission for inpatients were observed between patients treated in PSI-user EDs and PSI-nonuser EDs.

Table 6
View larger version:
Table 6

Comparisons of secondary outcomes between patients treated at emergency departments (EDs) that used the pneumonia severity index (PSI) to aid in site-of-treatment decisions for patients with community-acquired pneumonia (PSI-user EDs) and EDs that did not use the PSI (PSI-nonuser EDs).

Specifically, outcomes of patients at low risk who were treated as inpatients at PSI-user EDs and PSI-nonuser EDs were similar in regard to mortality (1 patient [0.8%] and 6 patients [3.4%] died, respectively; P = .148), ICU admission rate (5 patients [4.1%] and 14 patients [7.9%], respectively; P = .179), and length of stay (8 days [IQR, 5–14 days and 7.5 days [range, 5–12 days], respectively; P = .769). Likewise, no differences in secondary outcomes were observed for the 90 patients at low risk who had evidence of oxygen desaturation. Only 3 of these patients were treated as outpatients, of whom 1 patient who was treated at a PSI-nonuser ED was subsequently hospitalized (P = .33). Among the 87 inpatients, 14 were admitted to an ICU (4 [10.3%] at PSI-user EDs and 10 [19.6%] at PSI-nonuser EDs; P = .26), and all 4 patients (4.4%) who died within 28 days of the index visit were enrolled at PSI-nonuser EDs (P = .13).

Overall, 8 bacteremic patients (17%) died. No death occurred among the 13 patients at low risk who had bacteremia, of whom 3 were treated as outpatients (including 1 who was subsequently hospitalized). All 34 patients at high risk with bacteremia were hospitalized, and 8 (23.5%) died within 28 days of the index visit (3 of these cases of bacteremia were due to Escherichia coli, 2 cases were due to Streptococcus pneumoniae, 2 cases were due to Staphylococcus aureus, and 1 case was due to Streptococcus species).

Previous SectionNext Section

Discussion

In this prospective, observational, controlled study of patients with community-acquired pneumonia who were treated in the ED, we found that routine use of the PSI was associated with a proportion of patients at low risk who were treated in the outpatient environment that was almost twice as high as that for nonuse of the PSI (42.8% vs. 23.9%), without a compromise of the patients' safety. The actual rate of outpatient care was 18.9% higher for patients who were treated in EDs that routinely used the PSI. After adjusting for pneumonia severity, short-term mortality was lower in patients treated in PSI-user EDs, whereas subsequent hospitalization for outpatients and ICU admission for inpatients did not differ between patients who were treated at PSI-user EDs and patients who were treated at PSI-nonuser EDs. These findings support the existing evidence that the PSI is a safe and effective tool to augment clinical judgment for the decision of the initial site-of-treatment of patients who have community-acquired pneumonia [8, 9, 17].

Few patients in PSI risk class III were treated in the outpatient environment, in either PSI-user EDs or PSI-nonuser EDs. In contrast, the Low Risk Community Acquired Pneumonia (LOCAP) study reported a major impact of the PSI in the decreasing rate of hospitalization of patients in PSI risk classes II and III [8]. The low rate of outpatient treatment of patients in PSI risk class III in our study may be related to the fact that French pneumonia practice guidelines that were available at the time of this study did not support outpatient treatment of PSI risk class III patients [18].

The proportion of patients at low risk who were treated in the outpatient environment in the PSI-nonuser EDs (23.9%) was also substantially lower than those proportions observed in the control arms of the LOCAP (42%) and CAPITAL (51%) studies performed in the United States [8, 9]. However, by including in our study several subgroups of patients who were excluded from these previous studies (e.g., homeless persons, alcohol abusers, and nursing home residents), we increased the proportion of patients who had an implied psychosocial indication for hospitalization. Furthermore, unlike the LOCAP study, our study included 90 patients (20%) at low risk who had evidence of oxygen desaturation at presentation. Alternatively, differences in medical practice patterns from one country to another, patient and proxy preferences [19], organization of outpatient health care services, [20], and secular trends in hospitalization rates [21, 22] could explain the lower outpatient treatment rates that were observed in our study.

The rate of outpatient treatment in the PSI-user EDs (42.8%) was also lower than the rates observed in the intervention arms of the above-mentioned earlier studies (57% for the LOCAP study and 69% for the Community Acquired Pneumonia Intervention Trial Assessing Levofloxacin (CAPITAL) study [8, 9]). In addition to all of the reasons stated above, this lower rate of outpatient treatment may be related to the observational design of our study. The use of PSI at user sites had been determined prior to the beginning of the study. Because our intervention only consisted of a single reminding session at initiation of the study, we did not expect to observe outpatient treatment rates as high as those resulting from strong interventions to implement the PSI [8, 9, 17]. Finally, our findings also support the concept that physicians routinely consider other variables not included in the PSI (e.g., other comorbid conditions, abnormal WBC counts, and extensions of radiographic involvement) in their hospital admission decisions [23].

Several potential limitations to our study must be acknowledged. First, its observational design is prone to selection bias and confounding of the study outcomes; however, our 2-level multivariable regression modeling allowed us to adjust for potential confounders related to differences in the patient populations and between PSI-user EDs and PSI-nonuser EDs. Second, we cannot exclude that the initial information session influenced physicians' behaviors. However, it seems unlikely that a single information session could notably modify physicians' practices, as was indirectly observed in the Emergency Department Community-Acquired Pneumonia (EDCAP) study [17, 24]. Third, we did not assess other parameters that reflected the quality of care and that might have confounded the interpretation of the differences in mortality between PSI-user EDs and PSI-nonuser EDs. Nonetheless, the academic affiliation characteristics and annual volume were similar between PSI-user EDs and PSI-nonuser EDs. Additionally, a higher proportion of patients at high risk who were treated at PSI-nonuser EDs had severe comorbidities, especially cerebrovascular disease, which may account, in part, for the higher mortality that was recorded within PSI risk classes IV and V. Fourth, our study was conducted in France and, as such, our findings may not be generalizable to other countries; however, the consistency of our findings with those from previous studies performed in less heterogeneous patient populations [8, 9] supports the generalizability of the use of the PSI in various environments.

To summarize, our study demonstrates that the routine use of the PSI to help guide the initial site-of-treatment decision was associated with a larger proportion of PSI risk class I and II patients who were treated in the outpatient environment without compromising their safety. These findings provide further support for the incorporation of the PSI into guidelines for treatment of patients with community-acquired pneumonia and additional evidence supporting its more widespread clinical use in EDs.

Previous SectionNext Section

Pneumocom Study Investigators

Dr. Laurent Delaire and Dr. Sylvie Betoulle (Centre Hospitalier Général d'Angoulême, Angoulême, France), Dr. Philippe Grippon (Centre Hospitalier Général de Fontainebleau, Fontainebleau, France), Dr. Pierre Mardegan, Dr. Jean-François Cibien, and Dr. Cécile Noyez (Centre Hospitalier Général de Montauban, Montauban, France), Dr. Alain Cannamela, Dr. Thomas Guérin, and Dr. Emmanuelle Fritsch (Centre Hospitalier Général de Roanne, Roanne, France), Dr. Jean-Pierre Bal and Dr. Marie-Pierre Bertrand (Centre Hospitalier Inter-Communal de Créteil, Créteil, France), Dr. Nicolas Simon and Dr. Luce Guérin (Centre Hospitalier Intercommunal de Poissy-Saint-Germain-en-Laye, Poissy, France), Dr. Jérome Khazakha and Dr. Armelle Lafontaine (Centre Hospitalier Intercommunal de Tarbes, Tarbes, France), Dr. Jean-Baptiste Driencourt (Centre Hospitalier Régional d'Annecy, Annecy, France), Dr. Didier Jan and Dr. Emmanuel Carre (Centre Hospitalier Régional de Vannes, Vannes, France), Dr. Isabelle Claude, Dr. V. Moulin, and Dr. Gilles Mehu (Centre Hospitalier de Quimper, Quimper, France), Dr. Alain Delhumeau, Dr. Betty Mazet, and Dr. Pierre Marie Roy (Centre Hospitalier Universitaire d'Angers, Angers, France), Dr. Dominique Pateron and Dr. Joelle Benkel (Centre Hospitalier Universitaire de Bondy, Bondy, France), Dr. Françoise Carpentier, Dr. Marc Blancher, and Dr. Caroline Douchant (Centre Hospitalier Universitaire de Grenoble, Grenoble, France), Dr. Gilles Potel, Dr. Philippe Leconte, and Dr. Celine Longo (Centre Hospitalier Universitaire de Nantes, Nantes, France), Dr. Jean Rouffineau and Dr. Hélène Boureaux (Centre Hospitalier Universitaire de Poitiers, Poitiers, France), Dr. Jacques Bouget, Dr. Isabelle Jouannic, and Dr. Marie-Hélène Marquez (Centre Hospitalier Universitaire de Rennes, Rennes, France), and Dr. Jean-Michel Muller, Dr. Fabienne Moritz, Dr. Joel Jenvrin, and Dr. Iliasse Idrissi (Centre Hospitalier Universitaire de Rouen, Rouen, France).

Previous SectionNext Section

Acknowledgements

We thank Mrs. Mary B. Walsh (Center for Health Equity Research and Promotion, Veterans Affairs Pittsburgh Healthcare System, PA).

Potential conflicts of interest. All authors: no conflicts.

Financial support. Direction de la Recherche Clinique d'Ile de France as a part of the Programme Hospitalier de Recherche Clinique (AOM 89-145). Département de la Formation Continue des Médecins de l'Assistance Publique des Hopitaux de Paris, l'ARMUR (Association de Recherche en Médecine d'Urgence, Henri Mondor, Créteil) France, AQUARE (Association pour la QUAlité, la Recherche et l'Enseignement á l'Hopital Saint-Joseph (Paris)), and GlaxoSmithKline France (to B.R.), and Egide Foundation, Paris (Programme Lavoisier to J.L.).

  • Received June 27, 2006.
  • Accepted August 7, 2006.
Previous Section
 

References

  1. ↵
    1. Fine MJ,
    2. Auble TE,
    3. Yealy DM,
    4. et al
    . A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 1997;336:243-50.
    CrossRefMedlineWeb of Science
  2. ↵
    1. Bartolome M,
    2. Almirall J,
    3. Morera J,
    4. et al
    . A population-based study of the costs of care for community-acquired pneumonia. Eur Respir J 2004;23:610-6.
    Abstract/FREE Full Text
  3. ↵
    1. Fine MJ,
    2. Auble TE,
    3. Yealy DM,
    4. et al
    . A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 1997;336:243-50.
    CrossRefMedlineWeb of Science
  4. ↵
    1. Roson B,
    2. Carratala J,
    3. Dorca J,
    4. et al
    . Etiology, reasons for hospitalization, risk classes, and outcomes of community-acquired pneumonia in patients hospitalized on the basis of conventional admission criteria. Clin Infect Dis 2001;33:158-65.
    Abstract/FREE Full Text
  5. ↵
    1. Espana PP,
    2. Capelastegui A,
    3. Quintana JM,
    4. et al
    . A prediction rule to identify allocation of inpatient care in community-acquired pneumonia. Eur Respir J 2003;21:695-701.
    Abstract/FREE Full Text
  6. ↵
    1. Marras TK,
    2. Gutierrez C,
    3. Chan CK
    . Applying a prediction rule to identify low-risk patients with community-acquired pneumonia. Chest 2000;118:1339-43.
    Abstract/FREE Full Text
  7. ↵
    1. Arnold FW,
    2. Ramirez JA,
    3. McDonald LC,
    4. et al
    . Hospitalization for community-acquired pneumonia: the pneumonia severity index vs. clinical judgment. Chest 2003;124:121-4.
    Abstract/FREE Full Text
  8. ↵
    1. Atlas SJ,
    2. Benzer TI,
    3. Borowsky LH,
    4. et al
    . Safely increasing the proportion of patients with community-acquired pneumonia treated as outpatients: an interventional trial. Arch Intern Med 1998;158:1350-6.
    Abstract/FREE Full Text
  9. ↵
    1. Marrie TJ,
    2. Lau CY,
    3. Wheeler SL,
    4. et al
    . A controlled trial of a critical pathway for treatment of community-acquired pneumonia: community-acquired pneumonia intervention trial assessing levofloxacin. CAPITAL Study Investigators. JAMA 2000;283:749-55.
    Abstract/FREE Full Text
  10. ↵
    1. Carratala J,
    2. Fernandez-Sabe N,
    3. Ortega L,
    4. et al
    . Outpatient care compared with hospitalization for community-acquired pneumonia: a randomized trial in low-risk patients. Ann Intern Med 2005;142:165-72.
    Abstract/FREE Full Text
  11. ↵
    1. Mandell LA,
    2. Bartlett JG,
    3. Dowell SF,
    4. et al
    . Update of practice guidelines for the management of community-acquired pneumonia in immunocompetent adults. Clin Infect Dis 2003;37:1405-33.
    FREE Full Text
  12. ↵
    1. Mandell LA,
    2. Marrie TJ,
    3. Grossman RF,
    4. et al
    . Canadian guidelines for the initial management of community-acquired pneumonia: an evidence-based update by the Canadian Infectious Diseases Society and the Canadian Thoracic Society. The Canadian Community-Acquired Pneumonia Working Group. Clin Infect Dis 2000;31:383-421.
    FREE Full Text
  13. ↵
    1. Bartlett JG,
    2. Dowell SF,
    3. Mandell LA,
    4. et al
    . Practice guidelines for the management of community-acquired pneumonia in adults. Infectious Diseases Society of America. Clin Infect Dis 2000;31:347-82.
    FREE Full Text
  14. ↵
    1. Aujesky D,
    2. Auble TE,
    3. Yealy DM,
    4. et al
    . Prospective comparison of three validated prediction rules for prognosis in community-acquired pneumonia. Am J Med 2005;118:384-92.
    CrossRefMedlineWeb of Science
  15. ↵
    1. Fine MJ,
    2. Auble TE,
    3. Yealy DM,
    4. et al
    . A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 1997;336:243-50.
    CrossRefMedlineWeb of Science
  16. ↵
    1. Fine MJ,
    2. Auble TE,
    3. Yealy DM,
    4. et al
    . A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 1997;336:243-50.
    CrossRefMedlineWeb of Science
  17. ↵
    1. Yealy DM,
    2. Auble TE,
    3. Stone RA,
    4. et al
    . Effect of increasing the intensity of implementing pneumonia guidelines: a randomized, controlled trial. Ann Intern Med 2005;143:881-94.
    Abstract/FREE Full Text
  18. ↵
    1. Labarere J,
    2. Fourny M,
    3. Pavese P,
    4. et al
    . Agreement between clinical practice guidelines for management of community-acquired pneumonia. A retrospective study of 101 hospitalized patients [in French]. Rev Mal Respir 2003;20:858-70.
    MedlineWeb of Science
  19. ↵
    1. Fine MJ,
    2. Hough LJ,
    3. Medsger AR,
    4. et al
    . The hospital admission decision for patients with community-acquired pneumonia: results from the pneumonia Patient Outcomes Research Team cohort study. Arch Intern Med 1997;157:36-44.
    Abstract/FREE Full Text
  20. ↵
    1. Weinberger M,
    2. Oddone EZ,
    3. Henderson WG
    . Does increased access to primary care reduce hospital readmissions?. Veterans Affairs Cooperative Study Group on Primary Care and Hospital Readmission. N Engl J Med 1996;334:1441-7.
    CrossRefMedlineWeb of Science
  21. ↵
    1. Anderson GF,
    2. Poullier JP
    . Health spending, access, and outcomes: trends in industrialized countries. Health Aff (Millwood) 1999;18:178-92.
    Abstract/FREE Full Text
  22. ↵
    1. Fielding JE,
    2. Lancry PJ
    . Lessons from France—`vive la difference': the French health care system and US health system reform. JAMA 1993;270:748-56.
    Abstract/FREE Full Text
  23. ↵
    1. Halm EA,
    2. Atlas SJ,
    3. Borowsky LH,
    4. et al
    . Understanding physician adherence with a pneumonia practice guideline: effects of patient, system, and physician factors. Arch Intern Med 2000;160:98-104.
    Abstract/FREE Full Text
  24. ↵
    1. Yealy DM,
    2. Auble TE,
    3. Stone RA,
    4. et al
    . The emergency department community-acquired pneumonia trial: methodology of a quality improvement intervention. Ann Emerg Med 2004;43:770-82.
    CrossRefMedlineWeb of Science
| Table of Contents

This Article

  1. Clin Infect Dis. (2007) 44 (1): 41-49. doi: 10.1086/509331
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 54 (5)
  1. Clinical Infectious Diseases
  1. Alert me to new issues

Most