The assessment of patients to decide on the likelihood of them either having or likely to develop chronic open angle glaucoma (COAG) must take into account a great many risk factors and clinical measurements and signs. For this reason, protocols are in place to minimise the ever-present risk of false positive and negatives in the referral process from primary to secondary care. Never has this been more important than in these cash-strapped times.
As we discussed recently (Optician 27.04.18), national guidelines exist, variously in the regions, to ensure that initial assessment is in some way standardised. There are also a range of further refinement or double check processes, again varying between regions, that are aimed at reducing unnecessary referral, though do, however, introduce a second tier of assessment that is not ideal for many cases.
Algorithmic analysis
In theory, it is possible to know the link between specific measurements, such as tonometry for example, and the chance of developing COAG. The West Kent Clinical Commissioning Group Community Ophthalmology Team (COT) offers a referral refinement service whereby any suspect patients are assessed by a trained specialist optometrist who undertake the tests listed in table 1.
Table 1
- Goldmann applanation tonometry
- Disc assessment (after dilation) of:
- Vertical disc size
- Vertical CD ratio
- Inter-eye difference in vertical disc size and CD ratio
- Contact pachymetry
- SITA fast 20-4 Humphrey visual fields analysis
- Gonioscopy of anterior angle (to exclude other cause of glaucoma)
Now it is also well known that some people are more at risk than others of COAG and this needs to be factored in before a referral made. For example, COAG is highly unlikely in a 20-year-old. Risk factors are listed in table 2.
Table 2: Risk factors to be considered for COAG
- Age
- Gender
- Ethnicity
- IOP level
- IOP difference level
- Disc features
- Corneal thickness value
- Level of threshold field loss
Now, if value or the numerical data measurements were assumed, based on epidemiological studies of COAG, it should be possible to calculate likely risk of any one individual for whom all these factors and measurements were known of having or is likely to develop COAG. Referral could then be generated without the subjective decision-making of the specialist optometrist. This is the thinking behind a recent paper1 published in Nature.
Bayesian learning scheme
Bayes’ theorem predicts that high first visit discharge rates will occur for relatively rare diseases like chronic open angle glaucoma (COAG), even when the sensitivity and specificity of screening tests are high. Application of Bayes’ involves estimating the probability of an outcome by multiplying an initial estimate, based on prevalence of that outcome, by likelihood ratios derived from the sensitivity and specificity of each diagnostic test carried out as outlined in table 1.
The study aimed to see how accurately Bayes’ could predict clinical decisions made by the specialist optometrists in the COT. Decisions made by Bayes’ were compared to those of the specialist optometrists. Results were initially expressed in the form of 10 separate confusion matrices, one for each cross-validation run. These matrices simplified side-by-side comparisons of the Bayes’ and specialist optometrists’ decisions to discharge, follow-up or refer. Accuracy was expressed as the percentage of cases for which Bayes’ matched decisions made by specialist optometrists.
Conclusion
The results from this study were eye catching. Outcomes of cross-validation, expressed as means and standard deviations, showed the accuracy of Bayes’ was high (95%, 2.0%) but that it falsely discharged (3.4%, 1.6%) or referred (3.1%, 1.5%) some cases. The authors therefore concluded that the results indicate that Bayes’ has the potential to augment the decisions of specialist optometrists.
Optometrist roles have developed from data capture, indeed many of our test results are from procedures undertaken by trained and supervised support staff. Decision making upon these data is still the realm of the clinician, but where there are a great many influences upon a likelihood of positive diagnosis, we should embrace the potential for accurate algorithmic analysis in supporting and complementing that decision. After all, despite the harbingers of doom suggesting machine intelligence threatens our future roles, there is still the important matter of patient interaction and communication which is imperative of the suggested management strategy is to be complied with.
Reference
1 JC Gurney, E Ansari, D Harle, N O’Kane, RV Sagar and MC M. Dunne. Application of Bayes’ to the prediction of referral decisions made by specialist optometrists in relation to chronic open angle glaucoma. Eye, https://doi.org/10.1038/s41433-018-0023-5
Do you welcome the integration of automated decision-making into optometry clinics? bill.harvey@markallengroup.com