CHIME News
A A A

Publication: Primary total knee replacement as an example of predicting length of stay from electronic patient record system data

23 April 2014

A paper, published on April 4th in BMC Medical Informatics & Decision Making, shows how hospital length of stay can be modelled using data available in an electronic patient record system, using primary total knee replacements as an example (Carter EM, Potts HWW (2014). “Predicting length of stay from an electronic patient record system: A primary total knee replacement example.” BMC Medical Informatics & Decision Making, 14, 26. doi:10.1186/1472-6947-14-26).


The paper by Evelene Carter (Oxfordshire University Hospitals NHS Trust) was based on the dissertation that she completed as part of her MSc in Health Informatics at UCL CHIME. The paper is co-authored by her dissertation supervisor, Dr Henry Potts (UCL). CHIME encourages all of its MSc students to undertake real-world projects and, whenever possible, to turn them into publications.

Length of stay data is generally highly skewed and previous work in this field has often struggled to deal with this. The paper by Carter & Potts shows how a negative binomial regression fits such data well and can be readily applied. It also shows how the type of data generally available in hospital electronic patient systems is predictive of length of stay, allowing useful modelling to be done and insights to be gained.