Data mining can help epilepsy treatment

A study in the US has shown that computer software has proved to be useful in estimating the prevalence and mortality rates of patients with epilepsy, and that healthcare professionals can use such programs to work out where best to distribute resources.

The paper, published in the journal Epilepsia, used data mining and analysis of trends over periods of one, three and five years, concentrating on one area in New Mexico. The best computer model the researchers ran had an accuracy level of 90%.

"Healthcare administrators in managed care organisations could benefit from an accurate method of assessing the distribution of epilepsy patients, so they can allocate resources for the health care needs of this patient population," the study authors wrote.

"It is possible in managed care organisations to identify prevalent and incident cases from existing management information systems data for monitoring disease burden and developing disease management programmes," E Wayne Holden, lead author of the study, said.

Although epidemiological software for analysing disease data exists, the study authors say their methods brought into account other risk factors, such as seizures. The ethnic make-up of the area was also taken into account using software that estimated ethnic origins through each person’s surname.

The researchers stressed that mortality rates from epilepsy are generally difficult to judge, partly because of the flow of people moving in and out of the area, and partly because deaths from epilepsy are often complicated by other illnesses.

Further benefits of using systems like these include low cost, as the data is already held within the organisation, and the wider implications for research into the illness.

The researchers conclude: "Epidemiologic research and surveillance are important to access the public health burden of epilepsy, to provide accurate information to assist in policy development, to ensure necessary services for those with epilepsy."