“Data safety and monitoring board in non-industry trials: learning it the hard way”

In the majority of studies, no Data and Safety Monitoring Board (DSMB) is either needed or instituted. We report an investigator initiated study where we should have done this earlier than we did and discuss the lessons we learned. The EOLUS study was a single center, randomized controlled trial of the initiation of chronic home mechanical ventilation (HMV) at home. Typical indications for HMV are neuromuscular diseases such as amyotrophic lateral sclerosis (ALS) and Duchenne's disease next to thoracic cage deformities. The study was set up to investigate whether the initiation of HMV at home with the help of telemonitoring was not inferior to our usual in hospital start. The primary outcome measure was change in arterial carbon dioxide pressure from baseline to 6 months, for which we calculated a necessary sample size of 52 evaluable patients. The study was performed by an experienced nurse practitioner, as part of a PhD project. Weekly informal supervision was done by a senior pulmonologist and monthly progress meetings were held in a larger group. At these meetings, inclusion, lists of those excluded, and dropouts with their reasons and all deaths were discussed and minutes were always made and distributed. At four occasions over the first 14 months, a death of a patient, all with ALS, was discussed. The number of deaths occurring in this severely ill patients group, nor the circumstances were deemed remarkable. After the inclusion of 33 patients, all events were totaled and presented as such. We were shocked to realize that all 4 patients had died in the home group and none in the hospital group. The study was immediately put on hold and we reported this to the medical ethics committee. The ethics committee