1-4 PROGNOSIS AND DECISIONS IN PNEUMONIA
It is possible that clinical prediction rules will help reduce the variation in hospitalizations which occur in different areas. But, most of the variation in admission policies undoubtedly relates to older patients, those with coexisting illnesses, and those with abnormalities on physical examination—cases in which clinical judgment must always supersede the rules. NEJM January 23, 1997; 336: 288-89

1-3 A PREDICTION RULE TO IDENTIFY LOW-RISK PATIENTS WITH COMMUNITY-ACQUIRED PNEUMONIA

Hospital admission rates for pneumonia vary markedly from one geographic region to the next, suggesting that the criteria for hospitalization are inconsistent. Physicians often rely on their subjective impressions of a patient’s clinical appearance in making the initial decision about the site of care. Physicians tend to overestimate the risk of death in patients with pneumonia, and these overestimates are associated with the decision to hospitalize patients at low risk.
The purpose of this study was to develop a prediction rule for prognosis that would accurately identify patients with community-acquired pneumonia who are at low risk of dying within 30 days of presentation.

STUDY

1. Analyzed (in 1989) over 14 000 adult inpatients with community-acquired pneumonia to derive a prediction rule stratifying patients into 5 classes of risk of death within 30 days.
2. Validated the rules with 1991 data on over 30,000 in patients.
3. The rule assigned points based on age; the presence of coexisting disease; respiratory rate > 30; temperature > 400 C; pH < 7.35; blood urea nitrogen > 30; or a sodium concentration < 130 mmol/L at presentation.
4. The prediction rule was developed in 2 steps to parallel more closely physicians’ decision-making process: 1) Step 1 identified patients at low risk solely on the basis of history and physical-examination, and 2) Step 2 added selected laboratory and radiographic data.

RESULTS

1. Step 1. The following were independently associated with mortality:

• 1. Age > 50
• 2. Five coexisting illnesses
• A. Neoplastic disease
• B. Congestive heart failure
• C. Cerebrovascular disease
• D. Renal disease
• E. Liver disease

3. Five physical examination findings

• A. Altered mental status
• B. Pulse > 125
• C. Respiratory rate > 30
• D. Systolic BP < 90
• E. Temperature < 350 C or > 400 C

2. Patients with none of the above were assigned a low risk class (class 1) — mortality ranging from 1/1000 to 4/1000

• 3. Patients with any one of the above were assigned to step 2, adding the following.
• A. Male sex
• B. Nursing home residence
• C. Laboratory and x-ray:
• 1) BUN > 30
• 2) Glucose > 250 mg/dL
• 3) Hematocrit < 30
• 4) Sodium < 130 mmol/L
• 5) Partial pressure O2 < 60 mmol Hg
• 6) pH < 7.35
• 7) Pleural effusion
• 4. According to presence of each of the 20 points, patients were classified into 5 risk categories
• (table 2 p. 247).
• 5. Class I patients were all young (median age 36) and had none of the coexisting illnesses or abnormalities on physical examination. Class II were typically middle-aged (58 years) most were assigned this class by virtue of age alone. Class III were typically older (72 years) and most had a least one pertinent coexisting illness, a new physical examination abnormality, or one laboratory or radiographic abnormality. Class IV & V were somewhat older, the majority having 2 or 3 pertinent risk factors.
• 6. Morality was low for risk categories I, II, & III. Of over 1,500 patients in these categories, 7 died (1 in class I; 3 in class II; and 3 in class III). Four of the deaths were pneumonia related. None of these deaths was judged to have been preventable.
• 7. The rate of hospitalization within 30 days ranged from 5% in class I to 20% for class IV. None of the class I, II, or III outpatients who were subsequently hospitalized died. Of 8 outpatients in class IV & V who were subsequently hospitalized, 3 died.
• 8. Death rates rose dramatically for class IV & V patients (9% to 30%).

DISCUSSION

1. "Our prediction rule was designed to reduce uncertainty and to foster more appropriate use of hospitals in the management of this illness."
2. These predictor variables are all explicitly defined and can readily be assessed at the time of patient presentation.
3. Patients can be assigned the lowest risk class (class I) on the basis of information from the initial history and physical examination alone. This avoids ordering costly laboratory tests that are difficult to perform in an outpatient setting.
4. The prediction rule identifies 3 distinct risk classes (I, II, & III) of patients who at sufficiently low risk of death and other adverse medical outcomes that physicians can consider outpatient treatment or an abbreviated course of inpatient care. All patients 50 years of age or less who have none of the coexisting illnesses or physical examination abnormalities (class I) should be candidates for outpatient treatment. Many patients in risk classes II & III are also potential candidates for outpatient treatment — those who are under age 50 and have only a single pertinent coexisting illness or only one abnormal finding on physical examination or laboratory testing.
5. Application of these prediction rules could reduce the proportion of patients receiving traditional inpatient care by about 1/3.
6. An additional margin of safety could be provided by hospitalization of class 1, 2, and 3 patients who are hypoxemia (pO2 < 60 mm Hg on room air).
7. Individualization is important. Patients classified as low risk may have important psychosocial contraindications to outpatient care.

CONCLUSION

The prediction rule accurately identified patients with community-acquired pneumonia who were at low risk for death and other adverse outcomes. This may help physicians make more rational decisions about hospitalization for patients with pneumonia.
NEJM January 23, 1997; 336: 247-50 Original study, first author from Univ. Of Pittsburgh, PA

Comment:
I consider this a reference paper. Figure 1 page 246 gives a simple algorithm. Clinicians have always used these points in judging severity, prognosis, and need for hospitalization. The important point is to judge when a patient may be safely treated as an outpatient. RTJ

1-4 PROGNOSIS AND DECISIONS IN PNEUMONIA
(This editorial comments and expands on the preceding.)
The truth of an epidemiological association can be supported by such factors as the strength of the association and the consistency with which the association is demonstrated in different studies by different investigators. The remarkable consistency of four studies (including the preceding) suggests that each study offered a portion of the truth about prognosis in pneumonia. The full truth in science has been likened to a mosaic constructed by piecing together the partial truths of many studies.
Will physicians use the prediction rules? The authors emphasize that the rule should not supersede clinical judgment. Physicians know that patients with pneumonia who have highly abnormal vital signs, an altered mental status, a serious underlying illness, or who are extremely old should be admitted to the hospital, and most patients with pneumonia are already being cared for at home. In some settings, however, it is possible that larger proportions of patients with pneumonia may be admitted to the hospital unnecessarily.
It is possible that clinical prediction rules will help reduce the variation in hospitalizations which occur in different areas. But, most of the variation in admission policies undoubtedly relates to older patients, those with coexisting illnesses, and those with abnormalities on physical examination—cases in which clinical judgment must always supersede the rules.
NEJM January 23, 1997; 336: 288-89 Editorial from Univ. of Virginia Health Sciences Center, Charlottesville

[index]