Diagnostic support systems
Editor–The recent editorial (Clin Med August 2011 pp 310–11) and article (Clin Med August 2011 pp 317–21) on misdiagnosis suggest that one of the most effective ways to improve the quality of diagnosis is to formulate a list of differential diagnoses and continually re-evaluate it throughout the management process. As junior doctors, we are usually the first and sometimes only port of call for patients, particularly on medical wards at night and our diagnostic ability is limited by our lack of clinical knowledge and experience.
In these instances the use of diagnostic support systems may be able to help us. In our hospital we are trailing the ISABEL system.1 By entering the demographic and clinical features highlighted in the published case study, one of the conditions redflagged by the system is diverticular disease of the colon (Fig 1). The use of ISABEL might have alerted the clinicians to the correct diagnosis far earlier in the course of the patient's illness.
Such systems are not intended to replace a clinician's judgement but are reference tools, which compliment our knowledge. They can suggest diagnose that may be highly appropriate but seldom considered or recognised, particularly by an inexperienced junior doctor.
While we should all try harder to be better diagnosticians, mindful of the traps and pitfalls along the diagnostic journey, diagnostic support systems such as Isabel may prove to be an effective way to reduce diagnostic error.
Footnotes
Please submit letters for the editor's consideration within three weeks of receipt of Clinical Medicine. Letters should ideally be limited to 350 words, and sent by email to: clinicalmedicine{at}rcplondon.ac.uk
- © 2012 Royal College of Physicians
References
- ↵
- Isabel Healthcare
Diagnostic support systems
We thank Dr Luo and Dr Payne for their contribution to the discussion on improving the quality of diagnosis. In his editorial, Dr Scarpello stressed the importance of differential diagnosis whereas in our paper we were concerned primarily with stage 1–the initial thoughts and investigations–and how this stage may subsequently influence the progress of a case.
The ISABEL system is well known to our team. In 1999, the parents of Isabel funded the system. Isabel, their three-year-old daughter, survived a cardiac arrest and spent a month in intensive care at St Mary's Hospital (now Imperial NHS Trust), as a result of a previously unrecognised, life-threatening complication of chickenpox (necrotising fasciitis). ISABEL is perhaps the most sophisticated of the clinical decision support systems (CDSSs), developed over the past 50 years, with a database of more than 11,000 diagnoses and 4,000 drugs. It uses natural language processing software to search its database of medical textbooks and journals and it has a ‘knowledge window’ providing a wealth of information on any chosen condition.1
However, in real life, all the information required for the diagnosis may not be immediately available. Moreover, a definitive correct diagnosis is often not needed to initiate an appropriate work-up or treatment. In the case described the admitting doctors were not aware of the tender pelvic mass, but they recognised the evidence of sepsis–and if this situation had been managed efficiently the septic focus would have been found within 24 hours of admission.
Junior doctors may perceive a CDSS as potentially helpful but may have insufficient knowledge and experience to judge the appropriateness of the many suggestions it offers. Moreover, it is difficult to see how using a CDSS might be integrated into a busy clinician's workload. In the future this may become easier if the system could obtain its data from an electronic medical record.
Meanwhile, perhaps the main value of ISABEL is its potential to ameliorate cognitive tendencies to ‘premature closure’ and ‘diagnostic momentum’ and to improve diagnostic accuracy in complex cases that are difficult to unravel.
Footnotes
Please submit letters for the editor's consideration within three weeks of receipt of Clinical Medicine. Letters should ideally be limited to 350 words, and sent by email to: clinicalmedicine{at}rcplondon.ac.uk
- © 2012 Royal College of Physicians
Article Tools
Citation Manager Formats
Related Articles
- No related articles found.
Cited By...
- No citing articles found.