Clinical trials of new therapeutics for community-acquired pneumonia (CAP) have typically used a subjective endpoint of clinical response. However, as this endpoint is not quantitative, it is subject to observer bias and renders the conduct of multicenter trials difficult. For the purposes of conducting a clinical trial of filgrastim, as an adjunct to antibiotics for the treatment of CAP, a set of clinical criteria were developed prospectively to determine the time when a clinical cure was achieved, based on respiratory rate, temperature, oxygenation and roentgenographic findings, which was termed the time to resolution of morbidity (TRM). The TRM was evaluated on the first 100 patients entered in this clinical trial. As no clear reference standard exists, the predictive value for the duration of parenteral antibiotics (AB) and the length of hospital stay (LOS) was compared with that provided by a widely used classification system for severity of disease, APACHE II. The TRM was found to correlate significantly better with AB or LOS than APACHE II (P < 0.001). Furthermore, TRM offers the benefit over the endpoints of LOS and AB of being specifically designed to measure the patient's response to therapy, and, in fact, may aid physicians in determining the duration of parenteral antibiotic therapy. Hence, TRM is relevant to the clinician and is a useful tool to ensure uniformity in the assessment of the response to a new therapeutic in a multicenter clinical trial.