OBJECTIVE To evaluate the recognition of postoperative pulmonary complications (PPC) following thoracotomy and lung resection using three PPC scoring tools. METHODS Prospective observational study. METHODS Regional thoracic centre. METHODS One hundred and twenty-nine consecutive thoracotomy and lung resection patients (October 2007 and April 2008). METHODS PPC assessment was performed on a daily basis using three sets of criteria described by Brooks-Brunn, Gosselink et al. and Reeve et al.: the Brooks-Brunn Score (BBS), Gosselink Score (GS) and Melbourne Group Scale (MGS), respectively. The results were compared with treatment for PPC and clinical outcomes including mortality, postoperative length of stay and high dependency unit length of stay. RESULTS PPC frequency was 13% (17/129) with the MGS, 6% (8/129) with the GS and 40% (51/129) with the BBS. The clinically observed incidence of treated (requiring antibiotic therapy or bronchoscopy) PPC was 12% (16/129). CONCLUSIONS PPC treatment following thoracotomy is common. Of the three scoring tools, the MGS outperforms the BBS and the GS in terms of PPC recognition following thoracotomy and lung resection. Patients with a PPC-positive MGS score have a worse outcome as defined by mortality, high dependency unit length of stay and postoperative length of stay. The MGS is an easy-to-use multidisciplinary scoring tool, but further work is required into its use in minimally invasive surgery and in targeting high-risk groups for therapy.