Introduction Extracorporeal membrane oxygenation (ECMO) is a valuable but limited resource for the management of sick neonates in respiratory failure. Appropriate and timely referral to this service is thus paramount. We have noted a substantially increased referral rate from our unit to the local ECMO centre relative to equivalent level 2 units within the same region. We have thus evaluated the referrals made from our unit between 2011 and 2014 and considered factors which may have influenced the management pathway in each case. We consider specifically the site of delivery, initial ventilatory support, early evidence of respiratory compromise, and additional pertinent factors including surfactant, nitric oxide, and inotrope usage. Analysis of this data and comparison to other units within the region enables us to develop hypotheses regarding influential factors for ECMO requirement. This in turn enables the formulation of strategies to both reduce referral rates and allow the earlier identification of potential ECMO recipients. This offers a host of future service benefits in a streamlined NHS including aiding transport planning, appropriate resource management, and a potential reduction in the duration of intensive respiratory support through earlier intervention.

Method This was a retrospective analysis of the patients born at Birmingham City Hospital between 2009–2014 (n = 15). Inclusion criterion was transfer to an ECMO centre. Those patient transferred to an ECMO centre but who subsequently did not receive ECMO were also included. Those patients referred but not accepted were not included in the results. Patients were identified through internal digital search of BadgerNet and those identified correlated with physical medical records in all cases. There were no other exclusion criteria.

Results The medical records of the 15 identified infants were reviewed. Parameters evaluated include birth and referral times, identified pathology (including meconium and persistent pulmonary hypertension), methods of ventilation and intubation times, PaO2 and PaCO2 ranges, oxygen indices, mean arterial blood pressures, use of pharmacological adjuvants (including inotropes and nitric oxide), and referral times for ECMO. These were correlated with need for ECMO, the time of ECMO initiation and subsequent duration, and the total duration of hospital stay.

Analysis Using the data we present here we are able to develop upon the hypothesised antenatal risk factors and neonatal management strategies which affect the likelihood of deteriorating ventilation and the subsequent need for ECMO intervention. Our hope is that this will lead to a reduced ECMO referral rate from our centre and that the strategies will be applicable more widely. Learning from these experiences becomes ever more important in the modern NHS where intensive care resources are limited and the need to tailor specialist therapies ever more pertinent.