The definitive diagnosis of idiopathic pulmonary fibrosis (IPF) requires a pathologic pattern of usual interstitial pneumonia (UIP). However, it is not clear that whether there is a difference in outcome between patients with IPF diagnosed by clinical findings, and those diagnosed by pathological findings. We investigated differences in survival between patients diagnosed with IPF based on clinical findings, and patients who underwent surgical biopsy for a pathological diagnosis of IPF/UIP based on findings showing a UIP pattern. Sixty-one patients who received a clinical diagnosis of IPF were selected by random sampling. Fifty-nine patients underwent open lung biopsy or video assisted thoracoscopic biopsy and receive a pathological diagnosis of IPF/UIP. At diagnosis, 23 clinical features were identified and the hazard ratio of pathological diagnosis was estimated using the Cox proportional hazard model adjusted for those features that differed between the two groups. The adjusted hazard ratio of pathological diagnosis, taking these factors into account, was 1.09 (95% confidence interval, 0.42 to 2.80, p = 0.861) to about 1.0. Therefore, it is reasonable to diagnose IPF based on clinical findings from the perspective of outcome.