The Arnold Mühren of Refractive Surgery?*

How did you get started in ophthalmic laser research?

I started in refractive surgery with John Marshall, who got me interested in excimer laser surgery and connected me with an experimental excimer laser group run by Ian Constable. While other clinics were buying early excimer lasers, Ian said, “For that money, I can create a research lab and build my own, and it will be better than anything on the market.” I joined that lab, working with Paul van Saarloos, in 1990.

We were trying to develop a way of using the excimer laser to drill a small, accurate hole through the sclera for glaucoma filtration surgery. We created a system and took it to Phase I trials, but ultimately it wasn’t a good solution for glaucoma drainage – the diameter required for ideal flow resistance was only 50 µm, and a single cell could block the pathway.

I also had the opportunity to work closely with Graham Barrett, a pioneer in cornea, cataract and refractive surgery. In combination with the excimer laser experience, it was an ideal platform from which to pursue further subspecialist training at Moorfields Eye Hospital, where I was appointed as a consultant in 1998.

And then you started developing refractive surgery nomograms?

In 1995, I visited Robert Maloney, who used linear regression analysis to improve refractive surgery targeting. Linear regression describes the difference between what you attempt and what you actually achieve. In an ideal world, that line is a 45° slope – what you attempt is exactly what you achieve. But in reality, there’s often a systematic difference; the slope isn’t exactly 45° if you’re over- or undercorrecting, and if the regression line doesn’t cross through zero, you need to account for that difference as well to achieve your ideal results. Then, if you’re talking about laser surgery, you can change the energy or number of pulses you deliver.

When we started using the excimer laser a lot, we wanted consider influences besides spherical equivalent. The solution was multiple regression analysis, which can tell you whether a number of clinically plausible influences on your outcome are actually affecting it and weight each influence so you can factor that into your targeting adjustment.

How can you quantify what  patients think?

Many people are now interested in patient-reported outcomes – they’re looking for questionnaires that let you determine whether one surgical technique is better-accepted by patients than another.

Implantable collamer lens (ICL) surgery is a good example of the way this can be used. I’ve been implanting ICLs for over a decade; they cover an important group of younger patients who are out of range for laser treatment. The striking thing about people with a higher refractive error is the transformational effect ICL surgery has on their quality of life. We wanted to find a way of putting a number on this.

We gave quality-of-life questionnaires to a group of ICL recipients before and after surgery, and to a control group of happy contact lens wearers (not considering refractive surgery) with a similar level of myopia. The ICL recipients scored higher on almost every question – they were able to engage in sporting activity more easily, and were less concerned about their eye health and their vision in general.

Another important questionnaire has helped us determine which multifocal implant to use in cataract surgery and refractive lens exchange; you don’t get something for nothing optically, and spectacle-free vision often comes with a penalty in terms of visual quality or unwanted visual side effects. The Quality of Vision questionnaire is a way of quantifying this.

We’ve found patient-reported outcomes very useful for determining appropriate treatments, and we’re currently working on statistical techniques like principal component analysis to simplify the questionnaires and methods of online administration.

Are meta-analyses like Cochrane reviews also helpful in comparing techniques?

They can be useful, but it’s important to understand that they have the same time limitations as randomized trials. So they’re good for evaluating short-term outcomes, but less effective for long-term data. In comparing LASIK with PRK, the Cochrane methodology works well, because outcomes stabilize quickly and there are no real long-term safety concerns. Comparisons between lasers and phakic intraocular lenses are more problematic, as it takes several years to evaluate the safety of implantation adequately.

Cochrane reviews are great for the right applications, but limited if you’re looking at long-term outcomes. You only really find out about those from registry data and population studies. So randomized trials are not the only data that matters. All types of research are important in guiding clinical decisions, and I think registry work in particular has an important role to play in the future of refractive surgery.

*Bruce sees his varied research career in the mold of the classic Dutch midfielder from Bobby Robson’s glory days at Ipswich Town: weaving all the elements of the game together.