Better patient instruction for sputum sampling can improve microscopic tuberculosis diagnosis.

To examine the effect of patient instruction for provision of sputum samples, we performed a randomised controlled trial involving 174 patients with suspected tuberculosis (TB) in an urban TB clinic in Indonesia. TB was diagnosed in 50.6% of patients who received additional counselling by paramedics compared with 35.5% of patients in the control group following routine diagnostic procedure, corresponding to a 15.1% higher detection rate (95%CI 2.9-27.4). Significant differences were also found for sputum volume, consistency, positivity rate and density of acid-fast bacilli. Simple interventions such as these may increase TB case detection in Indonesia and elsewhere.