Nd:YAG laser capsulotomy rates in the Netherlands: practice variation and association with physician practice styles

Nd:YAG laser capsulotomy rates within 1 year after cataract surgery vary significantly across clinics in the Netherlands and correlate with physician practice styles known to affect posterior capsule opacification. Purpose: To determine the practice variation in the rate of Nd:YAG laser capsulotomy within 1 year after cataract surgery and to identify possible associations with physician practice styles. Setting: All hospitals and private clinics in the Netherlands. Design: Retrospective observational study. Methods: In the national medical claims database, we identified all laser capsulotomies performed in the Netherlands within a year after cataract surgery in the years 2016 and 2017. Centers with the lowest and highest percentages of Nd:YAG laser capsulotomies were interviewed on their physician practice styles related to the development of posterior capsule opacification. Results: The incidence of Nd:YAG laser capsulotomy varied between 1.2% and 26.0% in 2016 (median 5.0%) and between 0.9% and 22.7% in 2017 (median 5.0%). The rate of capsulotomy was highly consistent over time for each center (Pearson correlation coefficient, 0.89, P < .001). In general, ophthalmology centers with a high rate of Nd:YAG laser capsulotomy more often did not (routinely) polish the posterior lens capsule, performed cortex removal with coaxial irrigation/aspiration (I/A, instead of bimanual), and more often used hydrophilic intraocular lenses (IOLs) (compared with only using hydrophobic IOLs). Conclusions: We found a significant practice variation in performing Nd:YAG laser capsulotomy within 1 year after cataract surgery in the Netherlands. Routinely polishing the posterior capsule, using bimanual I/A, and the use of hydrophobic IOLs are associated with a lower incidence in Nd:YAG laser capsulotomy. Incorporating these practice styles may lower the practice variation and thus prevent added medical burden for the patient and decrease costs.

P osterior capsule opacification (PCO) may develop after cataract surgery because of proliferation and migration of lens epithelial cells that remained in the capsular bag with surgery. 1 This may lead to a reduction in visual function. 2 PCO is treated by Nd:YAG laser capsulotomy. Complications after Nd:YAG laser capsulotomy are rare. Nonetheless, cystoid macular oedema, increased intraocular pressure, retinal detachment, damage to the intraocular lens (IOL) (pits or cracks), IOL displacement, and uveitis have been reported to occur with Nd:YAG laser capsulotomy. [3][4][5] The reported incidence of Nd:YAG laser capsulotomy after cataract surgery is highly variable. Overall, it increases with time after surgery and depends on various factors, such as the type of IOL that was implanted, age of the patient, whether a complication occurred during cataract surgery, and the presence of comorbidity such as diabetes. [6][7][8][9][10][11] Several surgical factors have been shown to reduce the development of PCO, such as a good cortical clean-up, polishing of the posterior capsule (PC), and performing a capsulorhexis that is smaller than the optic diameter. 11 To what extent the incidence of Nd:YAG laser capsulotomy may vary between different clinics and how it relates to physician practice styles has not yet been studied. The aim of this study, therefore, was to assess the incidence (and its variation) in treatment of PCO in all clinics in the Netherlands that perform cataract surgery using medical claims data. In addition, we investigated physician practice styles that might influence the development of PCO.

Data Description
For our study, we used Dutch medical claims data for cataract surgery and Nd:YAG laser capsulotomy. In the Netherlands, every person who lives or works in the Netherlands is legally obligated to have standard health insurance to cover the costs of, for example, hospital treatment and prescription medication. All claims of delivered care from healthcare institutions and processed prescription medication to healthcare insurance companies in the Netherlands are collected by the Dutch Business Intelligence Centre for Health Care (Vektis). Demographic characteristics such as age are also available. Since cataract surgery and Nd:YAG laser capsulotomy are reimbursed treatments in the Netherlands, all medical claims for both procedures are present in the Vektis database.
We selected all patients aged 55 years or older with a diagnosis of cataract, a medical claim for noncomplex cataract surgery, and a medical claim for a Nd:YAG laser capsulotomy within 365 days after the last surgical procedure. We included only patients of 55 years or older because younger patients are at higher risk for developing PCO after cataract surgery. 12 The age of patients is computed in the Vektis database as their age on July 1 to comply with the General Data Protection Regulation. In the Netherlands, a medical claim can be made for noncomplex and complex cataract surgery. Complex cataract surgery is defined as an extracapsular cataract extraction with implantation of an IOL using nonstandard techniques or materials (such as a pupil expansion device, a capsular tension ring, and capsule stains). In this study, we excluded complex cataract cases because it may be either more difficult to remove all lens epithelial cells (eg, in the case of a mature [white] cataract) or the risk for developing PCO is higher (eg, in the case of a history of uveitis). 13 The search was performed for both the year 2016 and the year 2017 (the year represents the year in which the cataract surgery was performed).
In the Vektis database, we thus selected patients for whom a medical claim was made, with a diagnosis code of 554 (cataract), and with a care product code (in Dutch: diagnose-behandel combinatie zorgproduct) 070401 008 (inpatient noncomplex cataract surgery) or 070401 009 (outpatient noncomplex cataract surgery). Then, we linked these patients to the centers where they had been operated for their cataracts. Some centers had merged in 2017; for ease of comparison, we combined their Nd:YAG-laser capsulotomy rates in 2016. We excluded centers that performed less than 10 cataract surgeries in either 2016 or 2017. Then, we selected the patients with a care activity code (in Dutch: zorgactiviteit) 030898 (treatment of intraocular diseases with a YAG laser) within (and including) 365 days after the date of most recent cataract surgery (based on the care product code of the medical claim for cataract surgery). In the search for this care activity code, we looked for all care activities in the care product group 070401 (diseases of the lens).

Interviews With Healthcare Providers
A questionnaire was constructed to assess physician practice styles that are known to, or potentially could, affect the rate of PCO and the rate of Nd:YAG laser capsulotomy (Table 1). We identified the 5 ophthalmology centers with the highest percentage of Nd:YAG laser capsulotomies and the 4 with the lowest percentage of capsulotomies in 2017. The (representatives of the) ophthalmologists in these practices were invited for an interview. Two of the practices with a high laser capsulotomy incidence did not want to participate in the interview for reasons not specified. Therefore, we invited 2 other practices with the then highest percentage of Nd:YAG laser capsulotomies. The reason for the interview was provided in an invitation letter. The general results of our findings about the practice variation and the methods that were used were also provided. The questionnaire was added to the invitation with the request to prepare answers to the various questions upfront. Representatives of Zorgverzekeraars Nederland visited all 9 centers. One of the authors was present at all the interviews. During the interview, the questions regarding the physician practice styles were discussed.

RESULTS
We analyzed data from 91 ophthalmology centers that performed cataract surgery in the Netherlands in both the years 2016 and 2017. The number of patients who were included were 114 624 in 2016 and 115 556 in 2017. The incidence of performing a Nd:YAG laser capsulotomy in a patient within a year after cataract surgery in these centers ranged between 1.2% and 26.0% in 2016 (median 5.0%, 25th percentile 3.1%, and 75th percentile 7.4%), and between 0.9% and 22.7% in 2017 (median 5.0%, 25th percentile 3.2%, and 75th percentile 6.7%). Figure 1 shows the percentages of Nd:YAG laser capsulotomy for each center in the year 2016 vs the year 2017. It demonstrates the large variation in the incidence of PCO treatment between the centers. In addition, there is a high correlation in the percentage of patients in whom Nd: YAG laser capsulotomy was performed in a center in 2016 and in 2017. The Pearson correlation coefficient was 0.89 (P < .001). Table 2 presents the responses from the ophthalmology centers to our inquiry on their preferred practice patterns for cataract surgery and Nd:YAG laser capsulotomy. Practice patterns are in bold and bold-italic text relative to the presumed risk for PCO development: regular text indicates a lower risk, bold relates to a medium risk, and bold-italic indicates a higher risk for PCO.
In general, centers with a high rate of Nd:YAG laser capsulotomy more often used coaxial irrigation/aspiration (I/A) for aspiration of cortex (instead of bimanual I/A), more frequently polished the PC sometimes (instead of always), and more often used hydrophilic IOLs or a combination of hydrophilic and hydrophobic IOLs (compared with only using hydrophobic IOLs).
The ophthalmology center with the highest rate of laser capsulotomy made a capsulorhexis size between 5.0 mm and 6.0 mm (instead of always smaller than the IOL optic). None of the practices reported using FLACS for cataract surgery. All practices used vision complaints in combination with confirmation by slitlamp microscopy to make an indication for Nd:YAG laser capsulotomy. In most of the centers, the indication was made by an ophthalmologist. Only in 1 center, this was performed by an optometrist; this was a practice with a higher rate of laser capsulotomy. Most centers had 4 or fewer cataract surgeons; only 1 had more than 5 cataract surgeons. This was a center with a relatively high percentage of laser capsulotomies. Two of the centers trained residents in cataract surgery: 1 was a practice with a low rate of laser capsulotomy and 1 with a high rate.

DISCUSSION
We found a large variation in the treatment rate for PCO within a year after cataract surgery between the various ophthalmology practices in the Netherlands. The range in Nd: YAG laser capsulotomy at 1 year after cataract surgery was 1.2% to 26.0% in 2016 and 0.9% to 22.7% in 2017. The rate was remarkably consistent for each center in the years 2016 and 2017. Javitt et al. also used medical claims data to investigate the rate of Nd:YAG laser capsulotomy. 14 In the period 1986 to 1987, they found a variation between 10% and 24% across different states in the United States. However, they did not look for any differences between individual clinics. 14 The (median) incidence of Nd:YAG laser capsulotomy that we found (5% in both 2016 and 2017) is lower than the incidence of 11.8% at 1 year as reported in a systematic review of studies published between 1979 and 1996. 6 A more recent study, however, reporting on data from a single large site from 2007 to 2016 with implantation of a hydrophobic IOL, reported a significantly lower incidence of 1.2% (95% CI, 1.0%-1.4%) at 1 year after cataract surgery. 9 For longer periods of time after cataract surgery, incidences have been recently reported to vary between 2.4% and 31.1% at 3 years postoperatively and between 5.8% and 19.3% at 5 years, depending on the type of IOL that was implanted. 7-10 These latter data are more similar to our findings, both in amount and in variability, and may be more comparable with our study because they also involved various types of IOLs.
A strength of this study is that we used Dutch medical claims data. Since all claims for a reimbursement by any healthcare provider in the Netherlands are collected by Vektis, we were able to identify all cataract procedures and all Nd: YAG laser capsulotomies. This even allowed us to link any laser capsulotomy to a cataract procedure if it were performed in a different clinic. It does not, however, contain information on the laterality of the procedure. Therefore, it may have been that some cases that we identified in fact had a laser capsulotomy in the fellow eye to the one that had undergone cataract surgery. Nonetheless, we feel that this will have To what extent do you and/or your colleagues use FLACS? 6. Which criteria are used (in your practice) to make the indication for laser capsulotomy? 7. Who (in your practice) makes the indication for treating posterior capsule opacification? 8. How many ophthalmologists (in your practice) perform cataract surgery? 9. Do you train ophthalmology residents in cataract surgery (in your practice)?  I/A = irrigation/aspiration of cortical remnants; PC = posterior capsule The number in the first column represents the order in which the centers had been interviewed influenced our findings only slightly because many patients in the Netherlands undergo cataract surgery in both eyes within a year. In addition, if a patient were to present with clinically significant PCO in 1 eye and cataract in the fellow eye, the PCO is usually treated first. In addition, the above is true for all centers that we analyzed, both with a low and with a high percentage of laser capsulotomy. There is conflicting evidence regarding the influence of diabetes mellitus on the development of PCO. Some studies report a decrease in PCO, others an increase in PCO depending on the duration of diabetes itself, and still others show a significantly higher chance of PCO and Nd:YAGcapsulotomy in all patients with diabetes. [15][16][17][18][19] We currently did not exclude patients with diabetes from our study because it would have excluded a significant number of patients.
Based on Vektis data, we determined the prevalence of patients with diabetes type I and type II in the centers that we interviewed. In the centers with the lower rates of Nd:YAG laser capsulotomy, the prevalence of diabetes in 2016 varied between 0.5% and 1.4% for type I and between 16.9% and 27.9% for type II. For 2017, these rates varied between 0.7% and 1.7% for type I and between 12.2% and 27.3% for type II. Similarly, the percentage of diabetes in centers with the higher rates of laser capsulotomy in 2016 ranged between 0.1% and 4.8% for type I and between 13.4% and 22.5% for type II. For 2017, these rates varied between 0.4% and 3.6% for type I and between 10.8% and 23.6% for type II. We therefore feel that this will not have significantly affected our results.
(Part of) the variability that we found in the rate of Nd: YAG laser capsulotomy may be due to differences in the indication for the treatment of PCO in the various centers. It may be that centers with higher Nd:YAG laser capsulotomy rates had a lower threshold for performing a laser capsulotomy. The reverse may be true for centers with lower rates of laser capsulotomy. It is important to note that, during the interviews with the health insurers, all (representatives of the) ophthalmologists were convinced that they delivered the best possible care for their patients, regardless of their rate of Nd: YAG laser capsulotomy. A possible explanation for differences in the threshold for laser capsulotomy may be the assessment of the severity of PCO at the slitlamp. This assessment should be performed with retroillumination because one can then assess the extent to which light is refracted and scattered by the PC. 2 By contrast, observing light that is reflected by the PC (eg, by using direct illumination) may reveal fine details of any opacifications but provides less information on its effect on visual function and may lead to nonbeneficial Nd:YAG laser capsulotomies. 2 In case of only little PCO, measuring the amount of straylight in the eye may assist in deciding whether a laser capsulotomy is indicated. 2 Various factors have been reported to reduce the development of PCO. 11 These include a good cortical clean-up, polishing of the PC, implantation of the IOL in the bag, performing a capsulorhexis that is smaller than the optic diameter (thus allowing a tight adhesion between the capsule and the IOL), use of hydrophobic IOLs (instead of hydrophilic ones), and use of an IOL with a square optic geometry creating an effective migration barrier. 7,10,11 In our study, we looked at several of these factors and found a relationship between treating PCO with Nd:YAG laser capsulotomy and using coaxial I/A for cortical clean-up, not (routinely) polishing the PC, and using hydrophilic IOLs.
Our finding that coaxial I/A (instead of bimanual I/A) was used more frequently in centers with a higher rate of Nd:YAG laser capsulotomy may reflect that bimanual I/A might make it easier to reach all areas of the capsular bag, thus allowing a better clean-up of any remaining lens cells. We only found 1 published study that investigated the incidence of PCO after using either bimanual or coaxial I/A. 20 However, in this nonrandomized and relatively small study on 60 eyes from 39 patients, only 1 eye developed PCO and no statistically significant difference in PCO formation was found. 20 The importance of polishing the PC clearly stands out in our study. All 3 centers with the highest percentage of PCO treatment did not routinely polish the PC. Polishing the PC has been shown to reduce the incidence of PCO because it removes lens epithelial cells that otherwise might remain in between the IOL and the PC thus more easily inducing PCO. 11 Two centers with a high rate of Nd:YAG laser capsulotomy (9.7% and 11.4%) used hydrophilic IOLs exclusively for implantation. A third center with a higher rate (15.9%) used both hydrophobic and hydrophilic IOLs (apart from other predisposing practice styles). This is in line with previous findings and suggests that the use of hydrophilic IOLs increases the rate of development of PCO. 7,10 Other practice patterns that we investigated did not provide a clear difference between centers with a high or a low rate of Nd: YAG laser capsulotomy. Only 1 center reported to be using a capsulorhexis size between 5 mm and 6 mm. This was the center with the highest rate in laser capsulotomy. This clinic also showed other physician practice styles related to an increased risk for PCO. None of the centers that we currently interviewed used FLACS (which only a few clinics in the Netherlands use anyway). We also found no clear evidence that the type of eyecare provider who makes the indication for Nd: YAG laser capsulotomy significantly influenced its rate. The only center in which someone else other than the ophthalmologist made the indication for laser capsulotomy did, in fact, have the highest incidence of laser capsulotomy. However, multiple other factors that may influence the development of PCO were also present. Similarly, only 1 center had more than 5 surgeons performing surgery. In addition, although this center did show a higher incidence of Nd:YAG laser capsulotomy, it also had other risk factors for PCO (most notably the use of hydrophilic IOLs). Finally, training of residents in cataract surgery did not seem to have an impact on the rate of Nd:YAG laser capsulotomy. Our hypothesis was that residents, being less experienced than more established surgeons, might have less control over correct sizing of the capsulorhexis and might not be able to effectively polish the PC. Whether this effect is truly absent, or that the number of surgeries by trainees was too small to be able to detect a difference, needs further research.
We analyzed a subset of ophthalmology centers (9 of the 91) for their practice pattern styles. It may, of course, be that (some of) these practices have similar practice pattern styles without an increased laser capsulotomy rate. In this regard, we would also like to emphasize that the correlation we found between the identified physician practice styles and the rate of Nd:YAG laser capsulotomy does not necessarily equal causation (ie, there may not be a causal relationship). It may be that there are factors, even including financial ones, that influence both a physician's practice style and their indication for a Nd:YAG laser capsulotomy.
In conclusion, we showed that the rate of Nd:YAG laser capsulotomy for treating PCO within 1 year after routine cataract surgery in the Netherlands varies significantly. We identified differences in physician practice styles between the various ophthalmology centers that are known to influence the development of PCO. Specifically, 3 physician practice styles seemed to predispose for a higher treatment of PCO: (1) not (routinely) polishing the posterior lens capsule, (2) using coaxial I/A (instead of bimanual I/A) for cortex removal, and (3) implanting hydrophilic IOLs (instead of using only hydrophobic ones). We would like to emphasize that the indication for Nd:YAG laser capsulotomy should be based on subjective symptoms, parameters of visual function, and slitlamp examination with retroillumination. Adjusting these practice styles may help to decrease the practice variation in Nd:YAG laser capsulotomy that we found and thus prevent added medical burden for the patient and decrease costs.