Transbronchial needle aspiration (TBNA) via flexible bronchoscopy is a well-established sampling modality for lung masses. The procedure is useful in the diagnosis of neoplastic and non-neoplastic lesions as well as for staging of bronchogenic carcinoma. Rapid on-site evaluation (ROSE) adds value as it has the advantage of triaging material during the procedure so avoiding a battery of investigations. Frequently used rapid stains are the modified Wright-Giemsa water-based stain (WG-ROSE) and the alcohol-based modified Papanicolaou stain (Pap-ROSE). Final review of laboratory-based Giemsa and Pap stains supplemented by ancillary investigations is essential for quality assurance. To investigate whether and how ROSE influenced the quantity and quality of the material submitted to the laboratory we randomized 126 patients to WG-ROSE, requiring only one pathologist on-site, or combined WG- and Pap-ROSE, requiring an additional person on-site to assist with staining. In those patients with positive TBNA we graded the laboratory-based slides of the first pass containing diagnostic material into insufficient, suspicious, adequate and excellent. The first diagnostic pass was found after 3.06 ± 1.94 (SD) passes and 3.13 ± 2.16 passes with WG-ROSE and combined ROSE (P = 0.87), respectively. Following WG-ROSE and combined ROSE 69% and 71.1% (P = 0.509) of slides were diagnostic (adequate or excellent) on laboratory-based Giemsa stains, and 93.3% and 100% (P = 0.134) were scored adequate or excellent on laboratory-based Pap stains. We concluded that the less costly and labour intensive WG-ROSE procedure is adequate for TBNA. This has cost implications especially in resource poor settings. Diagn. Cytopathol. 2012. © 2011 Wiley Periodicals, Inc.