Prediction of surgery times and scheduling of operation theaters in ophthalmology department.

This paper presents the framework for forecasting the surgery time by taking into account the surgical environment in an ophthalmology department (experience of surgeon in years, experience of anesthetist in years, staff experience in years, type of anesthesia etc.). The estimation of surgery times is done using three techniques, such as the Adaptive Neuro Fuzzy Inference Systems (ANFIS), Artificial Neural Networks (ANN) and Multiple Linear Regression Analysis (MLRA) and the results of estimation accuracy were compared. Though the developed framework is general, it is illustrated for three ophthalmologic surgeries such as the cataract surgery, corneal transplant surgery and Oculoplastic surgery. The framework is validated by using data obtained from a local hospital. It is hypothesized that by accurately knowing the surgery times, one can schedule the operations optimally resulting in the efficient utilization of the operating rooms. This increase in the efficiency is demonstrated through computer simulations of the operating theater.

Evaluation of kidney transplantation programmes using system simulation.

In the case of kidney transplantations, there is always a serious imbalance between the number of kidneys donated for transplantation and the number of persons wishing to receive a transplant. This not only affects the quality of life of those unable to obtain a transplant, but it also has important repercussions on the treatment of End Stage Renal Disease (ESRD) by transplantation and dialysis. Also there are a number of ways in which the kidney transplantation can be achieved, such as the cadaveric kidney transplantation, live donor kidney transplantation, kidney paired donation and list exchange. A simulation study of all the referred programmes is performed using simulation models developed for each programme to obtain the better estimate of the average waiting time of a patient per year. With the estimates given by the simulation models, the best serving programme for each blood type patient is selected, declared and recommended.

Neural networks prediction of preterm delivery with first trimester bleeding.

OBJECTIVE: This paper illustrates a retrospective study of the outcome of those pregnancies that continued from an initial episode of bleeding in first trimester. METHODS: Neural networks is used for the prediction of preterm delivery, using various inputs such as the age of women, gestational age when the bleeding occurred, duration of the bleeding days, amount of bleeding, number of episodes, presence or absence of hematoma and placentation position. RESULTS: The success rate of prediction obtained using the feed forward backpropogation network is 70%. Hence, this model can be used to identify women at the risk of premature delivery for planning antenatal care and clinical interventions in pregnancy.