Artificial intelligence in the diagnosis of low-back pain and sciatica
- Mathew, B., Norris, D., Hendry, D., & Waddell, G.
- Spine, 13, 168-172
- An online version of the paper can be found here
- Want a summary of academic papers with alpha? Check out our free Academic Alpha Database!
In a prospective trial of 200 patients with low-back pain or sciatica, the diagnostic performance of a computer was compared with that of a clinician in a variety of clinical settings. The results indicate that artificial Intelligence techniques can be used for the differential diagnosis of low-back disorders, can outperform clinicians, and can be used to develop better methods of human differential diagnosis.
The sample contained 200 patients, with 50 in each of the four diagnostic categories:
- simple low-back pain
- root pain
- spinal pathology
- abnormal illness behavior.
These patients were analyzed by doctors, which were classified into 3 groups:
- Full assessment: hospital doctors with access to special investigations;
- Clinical assessment: family doctors based in the community;
- Systematic assessment: limited to a set of symptoms only.
The computer system used 2 diagnostic methods based on Fuzzy Logic:
- No-dialogue diagnosis: Data is entered into the computer and the diagnosis is computed;
- Dialogue diagnosis: The computer asks questions of the patient until it is confident it has a diagnosis.
Some insight into the algo:
The table below highlights how the computer dynamically develops a checklist approach to determining a diagnosis for a particular category of back pain.
Here is the finding:
- The computer outperformed the clinicians.
Still Trust your Doctor?
- Join thousands of other readers and subscribe to our blog.
- This site provides NO information on our value ETFs or our momentum ETFs. Please refer to this site.