Use this URL to cite or link to this record in EThOS:
Title: The development of a decision support system for the diagnosis of chronic idiopathic facial pain
Author: Chalidapongse, Premthip
ISNI:       0000 0001 3526 0804
Awarding Body: University of London
Current Institution: University College London (University of London)
Date of Award: 2004
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
The aim of the research was to develop (a) a well structured electronic medical record for a decision support system, and (b) logical algorithms for the diagnosis of Chronic Idiopathic Facial Pain (CIFP) and for educating trainees. This project started by validating the paper-based Facial Pain Proforma (FPP) with a panel of 3 experts. The FPP received a top grade consensus for history and examination. However, family relationships were considered too intrusive by one pain specialist and one clinical psychologist. A retrospective survey of 93 free hand pain histories taken by pain specialists (31 records), oral and maxillofacial registrars (12 records), senior house officers (31 records), and postgraduate students (19 records) were compared to the FPP. This revealed illegible data with many omissions. Medically trained surgeons produced good medical and examination data but overlooked important pain related and psychosocial data. Postgraduate students were often patient-led. A computerised FPP was developed as an electronic medical record - the Electronic Eastman Pain Proforma (EEPP) - using relational database software (Microsoft Access 97). The EEPP was validated for acceptability by clinicians and patients and compared to the free hand history (FH), and the FPP, (119 patients including 40 FH, 46 FPP, and 33 EEPP). Use of the EEPP did not diminish doctor-patient relationship. EEPP's history taking took 22 minutes compared to FPP (18 minutes) and FH (13 minutes). The average rating for EEPP was 2.8 out of 4. The design interface was rated as good. The clinicians were supportive for the concept of an electronic medical record. "Hand-crafted decision trees" were constructed by using expert knowledge and transcribed into "Diagnostic Rules". Machine learning technique were also used to induce comparable diagnostic trees from patient data (n=280). 5-fold cross validation of two induced decision trees showed diagnostic accuracy of 88% and 86%, with reasonable comprehensibility and high discriminative performance. The hand-crafted decision trees were validated using the same data. The resulting accuracy was 85% but comprehensibility was better than that of the induced decision trees. This work strongly supports the development and use of electronic medical records and a diagnostic decision tree system for clinical use.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available