Tuberculosis (TB) suspects from Rhodes Chest Clinic, Nairobi, Kenya, were subjected to three sputum smear microscopy (Ziehl-Neelsen) examinations and a chest X-ray (CXR). Results were compared with Löwenstein-Jensen culture as the gold standard to establish the efficiency of the routine diagnostic process. All laboratory tests and the CXR were available for 993 (71%) of the 1,398 enrolled suspects. Of these, 554 (56%) were culture-positive. The routine diagnostic process was very sensitive, able to detect 92% of culture-positive cases but missing 8%. The specificity was low (66%), and 23% of the patients started on treatment were culture-negative, mainly due to the low specificity of the CXR. It may be possible to increase the efficiency of the diagnostic process by specifying better criteria for CXR examination, improving the quality of CXR reading and counselling patients to return when complaints persist.