Australian Regenerative Medicine Institute

Identification of Low-Risk Hospitalized Patients With Pneumonia: Validity of the Practice Guideline

Severity of Illness
The average severity of illness score for patients in the study using the British Thoracic Society criteria was 0.9 ±0.9, using the University of Pittsburgh criteria was 2.53 ±2.91, and using the University of Rochester criteria was 3.95 ± 2.53.
Instability Upon Discharge
The mean instability score for low-risk patients was 1.17 ±0.91 on the fourth hospital day (when patients were suitable for discharge according to the guideline). This instability score compares with a mean score of 1.37 among Medicare patients with pneumonia at the time of hospital discharge.
The validity of the practice guideline was determined by measuring the ability of the guideline to detect patients who suffered life-threatening complications or died during their hospitalization. One hundred seventy-one patients were excluded because they suffered life-threatening complications within the first 3 days of hospitalization or did not have a length of stay of more than 3 days (inclusion of these patients would have overestimated the discriminatory ability of the guideline). Canadian health care mall The remaining 332 patients were analyzed. The sensitivity and specificity of the guideline for detecting patients who suffered life-threatening complications were 75.9 and 52.5 percent, respectively (Table 1). Moreover, the negative-predictive value of the guideline for detecting patients at low risk for life-threatening complications was 95.8 percent. Life-threatening complications were 3.5 times more likely (95 percent confidence interval [Cl] 1.4, 9.2 p=0.007) if the patient was designated as high risk rather than low risk.
The practice guideline had a sensitivity of 93.3 percent and specificity of 52.6 percent for detecting patients who died during the hospitalization (Table 2). The negative-predictive value of the guidelines for detecting death was 99.4 percent. Death was 15.5 times more likely (95 percent Cl 2.1, 73.1, p=0.001) if the patient was designated as high risk rather than if the patient was designated as low risk. The mortality rates for low- and high-risk patients were 0.6 and 8.6 percent, respectively.

Table 1—Practice Guideline Recommendations and Life-threatening Complications

Practice Guideline Yes No
High risk 22 144
Low risk 7 159
Sensitivity, % 75.9
Specificity, % 52.5
Negative-predictive value, % 95.8
Positive-predictive value, % 13.3
Odds ratio 3.5 (95% Cl 1.4, 9.2, p=0.007)

Table 2—Practice Guideline Recommendations and Inhospital Deaths

Practice Guideline Determination Inhospital Death
Yes No
High risk 14 148
Low risk 1 164
Sensitivity, % 93.3
Specificity, % 52.6
Negative-predictive value, % 99.4
Positive-predictive value, % 8.6
Odds Ratio 15.5 (95% Cl 2.1, 73.1, p=0.001)

Category: Respiratory Symptoms

Tags: antimicrobial therapy, congestive heart failure, healthcare costs, hospitalization, pneumonia