Random forests: The new generation of machine learning algorithms to predict survival in breast cancer
Asmaa Al‐Allak, Gianfilippo Bertelli, Paul D. Lewis · International Journal of Surgery · 2013
Methods: All patients (n¼101) undergoing breast surgery in November 2012 were prospectively audited.The dose, timing and appropriateness of LMWH administration were reviewed and compared to the UHL protocol.Results: 100/101 patients received LMWH with 95% of patients receiving 2500 units and 94% receiving LMWH pre-operatively.Furthermore 7% of patients received LMWH despite it not being indicated.Conclusions: The current practice is not standardised and a proportion of patients (7%) were over-treated.There are no published RCTs concerning the appropriate timing or dose of LMWH for general surgery and none concerning the use of VTE prophylaxis for breast surgery specifically.The incidence of VTEs or haematomas/haemorrhage is extremely low so large scale trials would be required to provide statisticially significant evidence.