Introduction
Cataract, the clouding or loss of transparency of the eye’s natural lens is the leading cause for blindness worldwide (Abraham et al.
2006). Prevalence rates of age-related cataract as the most common form considerably increases from about 30% for the 60–69-year-old population up to more than 60% for those older than 70 years (Prokofyeva et al.
2013). Up to now, surgical cataract extraction is the sole effective treatment to restore visual function and prevent blindness (Prokofyeva et al.
2013). With about 850,000 up to 1,000,000 yearly cases, cataract extraction is one of the most frequently performed surgeries in Germany, representing a high economic burden to the healthcare system (AQUA - Institut für angewandte Qualitätsförderung und Forschung im Gesundheitswesen GmbH
2010). During the procedure, the affected lens is extracted and replaced by an IOL implant. IOLs are available in four optic materials, i.e. polymethylmethacrylate (PMMA), high water content hydrophilic acrylic, low water content hydrophobic acrylic and hydrophobic silicone. Up to now, hydrophobic acrylic is the most frequently implemented IOL material (Lundström et al.
2012).
Clinical and economic impact of complications after cataract surgery
Risks of cataract surgery itself (e.g. posterior capsular rupture, zonular dehiscence or suprachoroidal hemorrhage) are generally low (AQUA - Institut für angewandte Qualitätsförderung und Forschung im Gesundheitswesen GmbH
2010; Chan et al.
2010; Lundström et al.
2012). Postsurgical complications such as changes in intraocular pressure, inflammations (i.e. uveitis, endophthalmitis) or a corneal edema occur sometimes (<0.4%) within a short time frame after surgery (Lundström et al.
2012). Also there is a risk of 0.9% for a retinal detachment in the first 4 years (Chan et al.
2010); therefore, a treatment of glaucoma, a vitreoretinal intervention or an explantation with secondary implantation of an IOL could rarely occur. The by-farmost frequent long-term complication after cataract surgery is a PCO, which usually occurs a few weeks up to several years after cataract surgery (Findl et al.
2007). PCO results in a decreased visual acuity, impairing the patient. In addition, PCO causes impaired contrast sensitivity and glare disability (Nibourg et al.
2015). A development of PCO within a maximum of 1 year after cataract extraction is estimated at 4.2% (Greenberg et al.
2011). Within a 2–4 years’ time frame, an incidence rate of 22.8% (Auffarth et al.
2004) to 38.5% (Fong et al.
2014) has been reported in cataract patients. This so-called secondary cataract can be treated with neodymium-doped yttrium-aluminum-garnet (Nd:YAG) laser capsulotomy, which is generally safe and the standard treatment procedure up to now (Karahan et al.
2014a); however, the Nd:YAG laser treatment of PCO itself can cause secondary effects. Transient increase of an intraocular pressure is observed in 5% of all cases and a decrease in visual acuity in about 4% (Boureau et al.
2009a). Other effects (glaucoma, cystoid macular edema or detachment of the retina) appear in less than 1.5% of laser-treated patients (Karahan et al.
2014b). The main complication is the formation of Elschnig pearls, which affects about 47.60% of the treated patients (Boureau et al.
2009a). On average, a secondary capsulotomy is necessary for every fourth patient (Kato et al.
1997).
Postsurgical complications cause a considerably part of the total costs of cataract treatment (Smith et al.
2005; Boureau et al.
2009a). In the US health system Nd:YAG laser capsulotomy accounts for $500,000,000 annually and is ranked second place in cost statistics for health interventions following cataract surgery itself (Menapace
2007). Billotte and Berdeaux (
2004) estimated that in the long-term,
1 up to 11,500 adverse events based on 400,000 cataract extractions could be avoided if Nd:YAG laser capsulotomy rates could be reduced. To develop effective strategies for PCO prevention, several influencing factors such as material and design of the implanted lens, surgical technique (Pandey et al.
2004), as well as pharmacological management during and post operation are discussed in the literature (Dewey
2006; Wormstone et al.
2009; Chandler et al.
2015; Nibourg et al.
2015). Comparing the available types of IOL materials, PCO development seems to be less likely after hydrophobic acrylic lense implantation compared to hydrophilic acrylic or silicone lens implants (Auffarth et al.
2004; Vasavada et al.
2011; Li et al.
2013; Sundelin et al.
2014). A further factor for the development of PCO seems to be the design of the lens, in favor of sharp-edged compared to round-edged IOLs (Findl et al.
2007; Mencucci et al.
2015).
The present analysis builds on the results of a preliminary study (Kossack et al.
2016). It assesses the impact of two different IOL material types (hydrophobic and hydrophilic acrylate) on the development of PCO rates and other complications related to cataract surgery combined with the associated costs in a 4-year follow-up based on a random sampling from SHI claims data. We performed a retrospective analysis to compare:
1.
The incidence of PCO and Nd:YAG laser capsulotomy after cataract extraction in relation to hydrophobic and hydrophilic acrylic IOL implantation in current German practice
2.
The associated costs of Nd:YAG laser capsulotomy due to PCO after cataract extraction from a SHI perspective
Study design/material and methods
Data and study population
Anonymized claims data were provided by the Institut für angewandte Gesundheitsforschung (InGef). The InGef research database covers approximately 6.7 million insured persons from different German SHIs, mainly company health insurance funds. The external validity of this database compared to German population data has been shown previously (Andersohn and Walker
2016). For the present study, a sample of approximately 4 million insured persons served as study population. This sample is representative for the German population with regard to age and sex for the year 2013. The database includes demographic information, diagnoses, utilization of ambulatory services, hospitalizations and reimbursed drugs as well as remedies and aids on a patient individual level. In Germany, SHI reimbursement of ambulatory services is regulated by the German National Ambulatory Evaluation Scheme (EBM). Utilization of outpatient services can thus be identified by the invoiced fee schedule position numbers (GOP). For historic reasons only, in the region of Bavaria these position numbers differ for the type of implanted IOL, enabling us to differentiate between an implantation of hydrophobic (GOP: 96104A) and hydrophilic (GOP: 96104B) acrylic IOLs. All patient-level data in the InGef database is de-identified to comply with German data protection regulations and German Federal Law; hence, approval of an institutional review board or ethics committee was not required.
The two groups under comparison were selected by a stepwise approach. First, we identified individuals who had been living in Bavaria without interruption and were continuously SHI insured within our study period 2009–2014. Relevant patients were identified with a diagnosis code of cataract according to the International Classification of Diseases, Tenth Revision (ICD-10) codes (cataracta senilis: H25, other cataract: H26, diabetic cataract/cataract in other endocrine, nutritional and metabolic diseases/cataract in other diseases classified elsewhere: H28, without other disorders of lens in diseases classified elsewhere: H28.8) and a surgical cataract extraction with an acrylic IOL implant in 2010 (index period). Patients who received both types of acrylic implants as well as patients who already have had an IOL implant before 2010 were excluded from the analysis. In a second step, we divided this sample into two subsamples depending on whether a hydrophobic or hydrophilic acrylic IOL had been implanted after cataract extraction.
Group comparison
By comparing the two study groups, we analyzed the impact of the implanted IOL material (hydrophobic acrylic versus hydrophilic acrylic) on the development of PCO rates in a 4-year follow-up. The incidence of PCO was identified in patients who underwent capsulotomy, which was identified by documented ICPM (International Classification of Procedures in Medicine) codes for the procedures of laser capsule polishing (5–142.0), surgical capsulotomy (5–142.1), Nd:YAG laser capsulotomy (5–142.2) or surgical posterior capsule polishing (5–142.3). To eliminate potential confounders, we additionally considered the following postoperative adverse effects as control variables: changes of intraocular pressure by treatment of glaucoma (procedure codes: 5–131, 5–132, 5–133, 5–134), a retinal detachment by vitreoretinal intervention (procedure codes: 5–158, 5–159) and an explantation of the newly implanted IOL with implantation of a next IOL—explantation with secondary implantation of an IOL (procedure codes: 5–146, 5–147.2, 5–147.3).
In order to ensure comparability of the study populations we analyzed sociodemographic structures as well as their respective medical history within 12 months prior to the cataract extraction. We considered the most relevant comorbidities as well as relevant drugs with a potential impact on PCO development/postoperative complications. Relevant comorbidities include retinal detachment and tear (H33), glaucoma (H40), diabetes mellitus (E10–E14) and hypertension (I10–I15; Pham et al.
2004). Certain medical treatments, systemic as well as topical, can have an influence on the proliferation of cells (Guo and DiPietro
2010; Noon et al.
2013), i.e., pharmacological interventions during or after surgery with the goal of depleting or inhibiting regeneration of remaining lens epithelial cells. Systemic drugs with a potential impact on complications such as cytostatic drugs, immunosuppressant drugs like corticosteroids and selective serotonin reuptake inhibitors (SSRI), a group of drugs which is mainly prescribed in case of depression, were identified by assessing the prescriptions based on ATC codes for pharmaceutical treatment, i.e., cytostatic drugs (ATC L01 and PZN 9999092, 2566881), Cortisone (ATC: H02, S01BA, S01BB, S01BX, S01CA, S01CB) and SSRI (ATC: N06AB02–10). To control for the severity of a patient’s comorbidities, we considered the Charlson comorbidity index (Charlson et al.
1987), which predicts the 1-year mortality.
We used Fisher’s exact test and Welch’s t-test to identify significant group differences and Bonferroni correction to avoid a false rejection of the null-hypothesis due to multiple testing in the same sample.
2 To assess the impact of the risk factors on the dependent variable PCO incidence we used a weighted multivariate logistic regression. We included age at cataract surgery, gender, the implanted IOL type and comorbidities as well as prescribed medication. Patients with a contralateral IOL implantation during the follow-up period got a higher weight in the regression analysis to counterbalance the higher risk of PCO caused by the second IOL-implantation. All analyses were performed in SAS version 9.2 using a proc. logistic model with effect coding. Statistical significance was set at
p < 0.05.
Costs of complications
The economic analysis compared direct costs over a period of 48 months after cataract surgery from an SHI payer perspective. Follow-up capsulotomy after IOL implantation is usually performed in outpatient settings in Germany. To calculate costs, we thus valued these interventions by means of the EBM of 2014. The number of follow-up Nd:YAG laser capsulotomies was calculated by summing up the respective invoiced EBM codes 31341
3 per patient in each group when the relevant procedures were coded. We additionally identified all billing positions that are related to PCO treatment. These include postoperative monitoring (EBM code 31501) and postoperative services either based on the surgeon’s referral (EBM code 31724) or performed by the operating surgeon him/herself (EBM code 31725). Since we used the most recent available version of the scheme, we did not discount costs in the analysis.
Discussion
This retrospective analysis was performed to compare the incidence of PCO measured by Nd:YAG laser capsulotomy or surgical intervention after implantation of either hydrophobic or hydrophilic IOLs following cataract extraction. Additionally, other postoperative adverse events such as glaucoma, vitreoretinal interventions and the explantation and replacement of the IOL were considered as control variables. The focus of the economic analysis was to assess the follow-up costs related to the different implanted IOL materials and the potentially following complications from an SHI perspective.
Most of the available evidence to compare IOL-types is provided by retrospective cohort studies or randomized controlled trials (RCT) with small sample sizes under artificial controlled study conditions mainly with short follow-up times. Under these conditions, hydrophobic acrylic lens material (compared to hydrophilic IOLs) was found to be related to a lower risk of PCO development measured by the rate of Nd:YAG laser capsulotomy compared to hydrophilic acrylic IOL material (34% vs. 49%,
p = 0.04; Schriefl et al.
2015).
The meta-analysis of Li et al. (
2013) finds a relative risk of 6.96 for Nd:YAG laser capsulotomy 2 years after cataract surgery, when hydrophilic IOLs are compared with hydrophobic IOLs. Several retrospective studies come to similar conclusions. Gauthier, Lafuma, Laurendeau and Berdeaux (
2010) found that the rate of Nd:YAG laser capsulotomy after bilateral hydrophobic IOL implantation was 8.8%, whereas it reached 37.5% after bilateral hydrophilic IOL implantation in a 2-year follow-up of 312 eyes. Boureau et al. (
2009b) found, that 45.4% of patients with hydrophilic implants had undergone Nd:YAG laser capsulotomy, while the amount was significantly lower (
p < 0.001) if a hydrophobic IOL was implanted (13%/23.4%). However, the summarized cohort or RCT to compare rates of Nd:YAG laser capsulotomy in relation to different IOL had small sample sizes (675 patients on the average) and follow-up periods of 1–3 years after the first IOL implantation (Auffarth et al.
2004; Boureau et al.
2009b; Gauthier et al.
2010; Vasavada et al.
2011; Li et al.
2013). Many scientific conclusions were made accordingly (e.g. as Auffarth et al. (
2004) that state, incidence rates of Nd:YAG laser capsulotomy do not considerably decrease over a 3-year follow up period).
In contrast, the present study assessed a population of 3,025 patients in a 4-year-follow-up under real world conditions. Accordingly, our results add to the scientific knowledge of Nd:YAG laser capsulotomies within longer time frames and with a large study population under real world conditions.
A considerable part of cataract-associated costs results from complications of the surgery and its treatments as well as related secondary adverse events; however, economic analysis from a SHI perspective has been very limited up to now. Based on a retrospective, multicenter study including 767 eyes treated, Boureau et al. (
2009a,
b) calculated a model of patients’ lifetime costs from a French SHI perspective for populations that either had hydrophobic or hydrophilic acrylic IOL implanted. Direct and indirect cost estimates were derived from official French documents, published literature and expert declarations. The authors found significantly higher adjusted risk ratios of undergoing an Nd:YAG laser capsulotomy (RR: 5.1;
p < 0.0001) for hydrophilic lenses compared to the hydrophobic IOL. For a linear extrapolation of Nd:YAG laser capsulotomy rates from 5 to 26 years of follow-up, total costs of treatment and management of complications per patient
7 were more than 3-times higher for patients with a hydrophilic lens implant (268.90 €) compared to the best performing hydrophobic implant (84.78 €) in the study. Indirect costs, associated to the risk of blindness, account for a considerable amount of about 42% of total costs. In the budget analyses, savings of 21.936.621 € from a French SHI perspective were estimated if all patients undergoing cataract extraction in 2005 switched to the better-performing IOL material (Boureau et al.
2009a). Due to the differences in national economic regulation, a generalizability of the results is however limited; moreover, the modeling approach is susceptible to uncertainties with regard to variable quality of the information considered. Complication rates have not been collected from medical records or claims data but estimated based on the literature, which might cause some imprecision. Smith et al. (
2005) compared cost-effectiveness ratios, i.e. costs per patient successfully treated without Nd:YAG laser capsulotomy, of four types of IOL material in a retrospective cross-country study. Their results show that hydrophobic IOLs seem to be more cost-effective than other materials in most countries. Costs were estimated on the basis on official documents and the mean total costs per successfully treated patient and IOL-type were computed and compared. However, the differences in regulatory frameworks and reimbursements between countries influence the economic estimates and limit the comparability of the results.
Comparing the follow-up costs associated with PCO treatment by Nd:YAG laser capsulotomy and post-operative monitoring and services, our results point in the same direction as previous studies (Smith et al.
2005; Boureau et al.
2009a). The difference in average costs per patient indicates the superiority of hydrophobic lens material from an economic point of view.
Our results add to the scarce health economic evidence related to treatment of complications after cataract surgery regarding two main aspects. Firstly, in contrast to available research (Smith et al.
2005; Boureau et al.
2009a) our health economic analysis is based on longitudinal claims data. To our knowledge, there is no study that compares different IOL material types combined with a cost analysis based on real world claims data. Results from real world economic analysis is particularly important for payers (i.e. statutory health insurance in Germany) to make value-basedcoverage and re-imbursementdecisions (Garrison et al.
2007). The database we used is characterized by a good accordance with the German population regarding sociodemographic indicators, morbidity, mortality and medication use. Thus, generalizability of the results is superior compared to studies including records from selected health care providers. Persistence with the database is high (78.5% of the insurants can be observed from 2009 to 2013); hence, longitudinal analysis is very reliable (Andersohn and Walker
2016). In contrast to retrospective multicenter post-test studies that account for most of the outlined evidence of IOL-related outcomes research, SHI claims data reflect daily practice and cover the whole spectrum of treatments and prescriptions by different health care providers (Pigeot et al.
2008). It guarantees 100% study participation proportions, which reduces potential selection biases; thus, it is possible to derive reliable actual SHI cost estimates.
This study has some limitations that are related to the structure and characteristics of claims data. Claims data are collected for billing purposes and occasionally there are inconsistencies and implausible values (Pigeot et al.
2008). The choice of potential confounders is limited to the available variables and a prospective design is not possible (Pigeot et al.
2008). Moreover, problems regarding the validity of diagnoses and procedure codes as well as coding errors are well-known problems in terms of billing data (Swart
2014). A differentiated coding for hydrophobic and hydrophilic acrylic IOL implantation is to our knowledge limited in Germany to the federal state of Bavaria, which may limit the generalizability of our results. However, the German SHI benefits for cataract surgery, pre-and post-operative care including treatment of complications are comparable to other federal states and related to the national reimbursement catalog for public medical services (EBM catalog). In addition, no restrictions exist regarding which IOL material should be implanted. In all German federal states, cataract surgeons can select the IOL material they prefer and coverage is granted by public reimbursement; therefore, there is no reason to assume that treatment and associated effects vary considerably in other federal states.