Limits of refractive change from OrthoK – Part 2

The amount of refractive correction that can be achieved by OrthoK has become such a contentious issue that it warrants more detailed explanation than I provided in my first blog on this topic. Welcome to the world of statistics - but don’t let this put you off, this is basic stats that you will be already familiar with in one way or another. The only thing that needs to be understood to follow my view on this topic is a normal distribution, which is how a measurement will naturally vary across a sample of a population. In an ideal world we would measure the entire population but for obvious reasons, we can’t. Instead, members of a given population are randomly selected to form a cohort, that is then measured. This range of measurements is then accepted as applying to the overall population.

A familiar example is a height, where we will all know people that are shorter and taller than us. If for example, you were to walk around an Australian town and measure the height of every person you bumped into, your measurements would be clumped around 175.6cm for men and 161.8cm for women, which is the average height of Australians while I am writing this book. You will also find Australians who are either over 215cm or under 140cm, but these will be few and far between. Plotted out you will find that they follow a bell curve or normal distribution. Measurements of most biological systems tend to follow normal distributions, so we should expect the same bell-curved distribution for maximum possible change from OrthoK lens wear until shown otherwise. Which I can’t see being done in a hurry as we would need to find a sufficiently sized cohort of people who are prepared to have their refractive error corrected to the maximum, and hence in some cases overcorrected, just to find the normal distribution of maximum possible OrthoK change.

Like everyone else, I can only guess how this normal distribution of maximum OrthoK change would look. I suspect it will be a steep, rather than well-spread distribution, otherwise, there would be ample evidence of successful fits to high prescriptions. Instead, we tend to see only one or two images of topography maps from successful high correction fits make the conference circuits, and these always display small treatment zones. This is a difficult concept to get your head around, but if we are saying that -4.50D is the average maximum possible change from normal OrthoK lens designs, and maximum possible change is normally distributed in people, then only 50% of people can achieve correction up to -4.50D (see image below). As we move to higher corrections the percentage able to achieve a successful outcome will reduce. Following this model through would indicate that correction of up to -5.50D is only possible in 20% of the population, or in other words, only 1 in 5 would achieve a successful outcome, with ever higher attempted corrections leading to progressively less likelihood of success. The only way for this not to be the case is for maximum possible OrthoK change not to follow a normal distribution, which in a biological system is highly unlikely. In a normally distributed system, the only way to increase the likelihood of success with higher refractive corrections is to accept a higher average maximum possible correction.

Normal distribution of OK change_Fig 1_2

We could argue about this topic until the cows come home or until we run out of beer, so I’ll get to my point, which is aimed more at helping the less experienced OrthoK fitter. If we lower the maximum possible change that we wish to target, by accepting a normal distribution model there will be more people that will successfully achieve the targeted correction. If we accept -4.50D as the average maximum possible correction, and we follow the same distribution as described above, then 80% of people would be successful with up to -3.50D correction, so 4 in 5 which is monumentally more acceptable than the hypothetical 1 in 5 successes for -5.50D targeted correction. Reducing your maximum targeted correction expectations further still and even more of the population would achieve a successful outcome.

Limitations of this explanation

Please accept all of what I have just described as a hypothetical model. The figures I suggest for average maximum possible correction are purely illustrative - as described above, the population maximum would be difficult to determine. The only factor that I suggest you accept until proven otherwise, is that maximum possible change follows a normal distribution. For this reason alone, you will reach an increasing percentage of successful outcomes as you reduce the maximum refractive error that you attempt to correct.

About Paul

Dr Paul Gifford is a co-founder of Eyefit, an information resource to assist contact lens practitioners in all modes of practice. Learn more about him here.