Abstract
Statement of Purpose: To enable GP’s and nurse practitioners to refer quickly, appropriately and effectively, at the first diagnosis of lower back pain.
Methods and result: Patients at the clinic, who were assessed and diagnosed by a GP/Nurse Practitioner, were asked to fill in an electronic questionnaire. The results of these were passed into a database where they could be accessed by an expert system program. This system used rule based logic and data mining concepts to assess the questionnaire answers, and presented a diagnosis of the patient. To determine system accuracy diagnoses were compared with actual diagnoses from the GP/Nurse Practitioner.
Conclusion: This project has been beset by several problems. The first being that a lack of patients with certain diagnoses has meant that it has been impossible to test the system with the more rare back problems.
The system accuracy is too low for the system to currently be of any use. This project is ongoing, the accuracy has improved significantly over the past year and we expect the improvement to continue next year. However, we have identified some problems in improving the accuracy. It has been noticed that there is a certain apathy present in some patients completing the questionnaires, resulting in less than accurate answers. Also the system can only produce one diagnosis. Patients with two back problems will get an incorrect diagnosis from the system.
Correspondence should be addressed to the editorial secretary: Dr Charles Pither, c/o British Orthopaedic Society, Royal College of Surgeons, 35-43 Lincoln’s Inn Fields, London WC2A 3PN.
