Development of the Research Center for Medical Information Systems at Ben-Gurion University in collaboration with the Soroka Medical Center will allow doctors to make decisions that are not based on intuition
Dr. Erez Shalom's doctoral research, under the direction of Prof. Yuval Shahar, from the Department of Information Systems Engineering at Ben-Gurion University of the Negev, carefully investigated for the first time the effects of the use of a computerized medical decision support system on the actual decisions of doctors during the management of medical care over time.
The research of Dr. Erez Shalom and Prof. Shahar was done in collaboration with a number of researchers from the university: Prof. Eitan Lonnenfeld, from the Faculty of Health Sciences at Ben-Gurion University of the Negev, Chairman of the Division of Obstetrics and Gynecology at the Soroka University Medical Center, and Dr. Israel Permet, From the Department of Industry and Management at Ben-Gurion University of the Negev. The article was recently published in the leading journal IJMI and describes the "Picard" medical decision support system developed at the Research Center for Medical Information Systems at Ben-Gurion University of the Negev (headed by Prof. Shahar) and how it may be integrated into the work of doctors.
As part of the study, the researchers electronically represented the guideline for the treatment of "preeclampsia/eclampsia" of the American Society of Gynecology and Obstetrics, ACOG, which serves as the basis for treatment in most obstetrics departments, including those at the "Soroka" University Medical Center. The electronic representation of the guideline is preserved as part of the knowledge base of the digital medical guideline library (DeGeL) developed at the research center.
The premise of the study is that providing treatment recommendations as part of a computerized system to support medical decisions based on evidence-based medical guidelines (evidence base medicine) will dramatically change doctors' responsiveness to medical guideline recommendations, increase the correctness of their actions relative to the guideline, and prevent unnecessary actions.
The system was evaluated by 36 doctors from the women's and midwives division at the "Soroka" medical center (24 interns and 12 specialists), where each doctor made 60 different decisions within the framework of six complicated treatment scenarios over time for preeclampsia. These scenarios were defined by a senior team of doctors from the division. Since this is currently an experimental program, which of course cannot yet be run on female patients at the medical center, the researchers created for the computerized operation, with the help of special simulation software developed at the research center, a series of complete medical records throughout several months of pregnancy, which simulated real patients in a very tangible way. The doctors reviewed the patients' data on a computer and made decisions in a way that was very similar to their usual way of working. Half of the doctors' decisions were made manually, without the help of the decision support system, with the exception of reviewing an electronic record with a complete image of the patient and the fetal development chart; In the second half of the decisions, the same doctors were additionally assisted by the therapeutic recommendations that came from the "Picard" system. The doctor could accept or reject any recommendation and even ask for an explanation. A senior physician who did not participate in the trial also served as a judge of the quality of each decision, with the ACOG guideline serving as the ultimate quality standard,
(The Gold Standard)
The results showed that the doctors' compliance with the system's instructions increased from 41% when they did not use the "Picard" system, to 93% when they were exposed to the system's recommendations. The percentage of doctors' actions that were assessed as both correct and necessary based on the ACOG guideline increased from 27% without the use of "Picard", to 91% when the doctors received recommendations from the system.
At the same time, the rate of actions that are not absolute mistakes, but are not necessary based on the guideline) dropped from 68% without using the "Picard" system, to only 3% in decisions that used it - a fact that has considerable economic significance. (Note: A recent study at the Johns Hopkins University Medical Center found that 66% of the requests for the expensive troponin enzyme tests, which are relevant to the diagnosis of a heart attack, were unnecessary. It is possible, therefore, that 2/3 of the tests and other operations which are performed by doctors, are unnecessary - perhaps due to defensive medicine or ambiguity as to the relevant guideline for each case that must be followed). Likewise, no difference was found in the quality of decisions between the different types of doctors in terms of experience.
The evaluation of the "Picard" system showed that systems such as these have the potential to reduce the number of medical errors, reduce medical expenses, and raise the quality of the clinical decision to a uniformly high level, regardless of the professional experience of the attending physician, the type of decision, or the relevant medical guideline. Therefore, the researchers expect that in the future, as more studies and evaluations are added that will show the contribution of such systems to doctors, these systems will enter the medical environment as a necessary standard to improve the quality of medical decision-making. This trend can already be seen with the new law of US President Barack Obama, which requires the implementation of standard medical information systems of the Ministry of Health in order to demonstrate meaningful use of the technology and receive participation in its ownership (meaningful use).
Furthermore: these days the "Picard" system is used as a central component of the seventh research and development project of the European Union (7FP), "MOBIGUIDE" which is led by Dr. Shalom and Prof. Shahar on behalf of Ben-Gurion University. This research develops and examines A system for monitoring at home, by sensors that are on the patient's body or in his home and using the patient's mobile phone. The monitoring is carried out for chronic patients in general, and in particular for patients suffering from atrial fibrillation, and women with gestational diabetes or hypertension. Picard sends alerts if necessary to the patient or to his doctors, and provides personalized recommendations, based on the landing of the appropriate medical hotline, to the patient's mobile phone or the doctor's computer. The result is treatment that is tailored to the patient, and changes if necessary according to the patient's clinical, physical, mental and socioeconomic status, and in accordance to the patient's preferences in terms of treatment, and his physical environment.
It should be noted that these days a field trial was started in the Mobiguide project in 2 hospitals in Europe: in Barcelona, Spain at the Sabadell Hospital in the treatment of patients suffering from gestational diabetes; and in Pavia, Italy at the Fondazione Salvatore Maugeri Hospital in the treatment of patients suffering from cardiac arrhythmia of the atrial fibrillation type
