An Analysis Of Patient Flow Using Markov Chains: A Case Of Kapsabet County Referral Hospital

ABSTRACT

Hospital is an essential welfare of society. It provides management of illnesses through

treatment and prevention interventions by medical and health professionals. Due to growing

population and rise in chronic diseases, there is an increased demand for health care

services. This causes congestion and overcrowding in most hospitals. Hospital overcrowding

is a major problem to patients, hospital administration and to the general health

workers. Hospitals in many jurisdictions struggle to reduce congestion and improve the

patient flow across the continuum of care.

In this project report, we have developed an objective patient flow assessment through

an analysis of Markov chains using weekly data from Kapsabet county referral hospital

to check on the patient’s flow at the hospital.

We used weekly data to construct transition matrices for each day in a week to portray the

weekly routine amount of the patients in the hospital using the states; high, medium,low

and very low. Steady state transition matrices were also computed for each day of the

week to reflect the future flow for each week.

It was found that the patient flow had some pattern observed through the steady states.

The probability of patient flow being high tends to be up on Mondays with probability

of 0.57, medium on Tuesdays to Thursdays with probabilities of being high ranging from

0.36 on Tuesdays and 0.3 on Thursdays, on Fridays is when the steady states of being

high starting to decrease upto Sunday with steady state probabilities of 0.22 on Friday,

0.17 on Saturday and 0.12 on Sunday.

Through the analysis of patient flow using Markov chains, we have identified some pattern

of how the patient flow throughout the week. Generally through this study, the patient

flow congestion can be easily understood, handled and hence controlled