Objective: There are a number of factors that go into a recommendation to treat non-tuberculous mycobacterial lung infection (NTM-LI). The purpose of the study is to validate consensus recommendations on clinical considerations before treating a NTM-LI. Methods: We engaged respiratory physicians of the "Mycobactérioses" working group, part of the research and teaching group in pulmonology and infectious diseases (GREPI), on behalf of the SPLF (Société de Pneumologie de Langue Française) in a Delphi survey concerning three areas of interest for physicians before they had to decide to treat (Diagnosis, Definition of the level of severity and Clinical assessment before deciding to treat) in order to create a decision algorithm. Each assertion was subjected to rating and free comments by these experts via an online questionnaire. The scores were weighted by the number of patients seen each year by the experts. The assertions having obtained a weighted average greater than 9/10 were validated. Otherwise they were subject to re-evaluation during round 2 until reaching 9/10. Results: A total of 37 experts responded. At first round, 9/38 assertions obtained a strong agreement (median ≥7 and 80% of responses >7). Other assertions from all domains, in particular the definition of the level of severity, were revised and new assertions added for submission to scoring in round 2. Fifty-one assertions were finally assessed. Agreement was strong in 78% of the assertions (40/51); moderate in 22% (11/51). There was no disagreement. Conclusion: We obtained a consensus from physicians about clinical considerations before deciding to treat NTM-LI and we propose a decision algorithm to guide a treatment decision.

We prospectively assessed the safety and cost saving of a small-bore drain based procedure for outpatient management of first episodes of primary spontaneous pneumothorax. Patients were managed by observation alone or insertion of an 8.5-F "pig-tail" drain connected to a one-way valve, according to size and clinical tolerance of the pneumothorax. All patients were reassessed after 4 h, on the first working day after discharge and on day 7. Patients still exhibiting air leak on day 4 underwent thoracoscopy. The primary end-point was complete lung re-expansion at day 7. 60 consecutive patients entered the study. 48 (80%) met the definition of large pneumothorax. The success rate was 83%. The 1-year recurrence rate was 17%. 36 (60%) patients were discharged after 4 h and 50% had full outpatient management. No severe complication was observed. The mean ± SD length of hospitalisation was 2.3 ± 3.1 days. This policy resulted in about a 40% reduction in hospital stay-related costs. The present study supports the use of a single system combined with a well-defined management algorithm including safe discharge criteria, as an alternative to manual aspiration or chest tube drainage. This approach participates in healthcare cost-savings.