A CRITERION BASED AUDIT OF COMMUNITY-ACQUIRED PNEUMONIA

We undertook a prospective audit of patients admitted to a specialist respiratory unit in order to assess: (a) the completeness of the objec tive assessment of severity of community-acquired pneumonia in patient s on admission to hospital, and whether such indices were useful in pr edicting outcome; (b) the completeness of microbiological testing and whether maximizing the number of specimens sent to the laboratory woul d increase the frequency of positive microbiological diagnoses and; (c ) the influence of the above on clinical management. Forty-eight patie nts with consolidation on chest X-ray were studied in two audit period s: 1 February 1991 to 1 May 1991 and 2 May 1991 to 16 March 1992. Afte r the first audit period, a rubber stamp with a check list of microbio logical investigations was used in the patients' notes. Seven 'markers of severity' of pneumonia, were recorded consistently throughout the study and these were useful in predicting outcome (P<0.01). In the fir st audit period sputum culture, blood culture and acute serology were recorded in approximately 50% of patients. Following the introduction of the rubber stamp, the requesting of sputum, blood and urine antigen rose from none in the first audit period to 40-60%. The increase in t he number of specimens obtained was not associated with an increase in the proportion of positive microbiological diagnoses. Clinical manage ment was altered in 10 patients because of the microbiological reports and in four patients with negative microbiology (chi(2)=8.19; P<0.01) .Conclusion: The standard of initial assessment was high in this speci alist unit, and the presence of two or more 'markers of severity' did help predict outcome. A simple change in work practice resulted in a s ignificant improvement in the thoroughness of microbiological investig ations. This did not result in an increase in the frequency of positiv e microbiological diagnoses, although in the 19 patients with a microb iological diagnosis antibiotic treatment was modified significantly mo re often than in patients with no pathogen isolated.