P38 Is it possible to predict ambulatory oxygen (AO) requirements

Aim To determine whether AO flow rate requirements can be predicted in patients with chronic lung diseas. Background The process for AO assessment as per current UK guidance can be extremely time consuming, necessitating multiple walking tests with significant rest periods between walks. We set out to explore whether a model could accurately predict the AO flow rate required to abolish/minimise desaturation, thus simplifying the process. Method Retrospectivedatawere analysed for all patients who attended AO clinic from April 2009 to January 2013. AO needs were assessed as per current UK guidance. For those who met the criteria for AO (n = 100), resting Sp02, desaturation on the endurance shuttle walk test (ESWT) and AO flow rate required to minimise this desaturation were recorded, along with age and diagnosis. Results Multiple regression analysis was undertaken to test how well the variables included in the data could predict the likely level of AO required. A range of models were constructed and tested against the actual AO flow rate that had been required. The model which accounted for the greatest proportion of the variance (r = 0.671, F = 79.40, p Conclusion Using this model, 91% of patients were predicted correctly with an error of +/- 1L of the actual flow rate. It would therefore, 9 times out of 10, give a flow rate that was either correct ( +/- 1L). Use of this model will reduce the number of walking tests required when performing AO assessment, saving both time and valuable healthcare resources.