Introduction

Type I Boston keratoprosthesis (KPro) has recently been shown to be a highly cost-effective medical intervention at 16 140 $/quality-adjusted life years (QALYs).1 However, not all corneal diseases are amenable to or appropriate for type I implantation. Patients with autoimmune diseases, such as Stevens–Johnson syndrome/toxic epidermal necrolysis and mucous membrane pemphigoid have severe ocular surface diseases where destruction of the corneal epithelial stem cells located at the corneal limbus results in conjunctival invasion, corneal neovascularization, chronic inflammation, and stromal scarring.2 In end-stage cases, the ocular surface becomes completely dry and the fornices become obliterated. Corneal transplantation and type I KPro, in such cases, are almost inevitably associated with a poor prognosis.3, 4 Patients' quality of life is appreciably affected, often experiencing considerable, if not complete, loss of vision and numerous comorbidities. Because of the destructive nature of these conditions, there have been few reports on the use of KPro surgery for the treatment of autoimmune diseases.5, 6, 7, 8 In contrast, several manuscripts have described the modified osteoodontokeratoprosthesis as efficacious in this patient population.9, 10, 11 For many, this represents the gold standard for severe ocular surface disease. Nevertheless, the morbidity and postoperative complications associated with this lengthy, two-stage procedure,11 and the possibility for a more simplistic and pragmatic KPro cannot be ignored, especially as recent modifications to the type II design and postoperative management of patients have led to improved clinical outcomes.12, 13

To determine cost effectiveness by way of cost-utility analysis (CUA), the perceived value of an intervention or health state is taken into account. Utilities on a scale from 0 to 1 are generated, most commonly by the time trade-off approach,14, 15 and these are used to determine overall benefit. Benefit is measured in terms of QALYs.

To our knowledge, no CUA has been performed on the use of the type II Boston KPro. It is the intention of this study to objectively assess the (1) comparative effectiveness (gain in QALYs) and (2) average cost effectiveness (compared with no further treatment or current visual state) of the type II Boston KPro procedure for the treatment of severe corneal disease and blindness.

Patients and methods

Boston KPro type II

Type II Boston KPro was developed at the Massachusetts Eye and Ear Infirmary.16 As seen in Figures 1 and 2, it is a collar button shaped polymethylmethacrylate device consisting of two curved plates that is implanted through the eyelids in severe dry eye and cicatricial diseases. Type II Boston KPro is performed far less than type I, and is reserved for near-hopeless cases with severe destruction of the ocular surface.

Figure 1
figure 1

The type II Boston keratoprosthesis. The front plate of the keratoprosthesis is shown assembled ex vivo.

Full size image
Figure 2
figure 2

The clinical appearance of the type II Boston keratoprosthesis. The front plate of the keratoprosthesis is shown implanted through the eyelid of a patient with a history of Steven–Johnson's Syndrome.

Full size image

Patients

In a retrospective cohort study chart review, 29 patients who underwent type II KPro surgery at the Massachusetts Eye and Ear Infirmary between the years 2000 and 2009 were identified. Patients had to have a minimum of 5 years of visual acuity follow-up data recorded. Patients receiving type II KPro were in the worst prognostic groups, such as autoimmune diseases and chemical burns. A total of 11 patients had a minimum of 5 years of follow-up and were included in the study.

Patient characteristics with underlying diagnoses before KPro surgery are given in Table 1. A complete ophthalmic exam was performed before KPro surgery. Median preoperative best-corrected visual acuity (BCVA) in the treated eye was logMAR 2.3±0.7 (Snellen equivalent HM). Visual acuity values were normally distributed per Shapiro–Wilk testing. We certify that all the applicable institutional and governmental regulations concerning the ethical use of human volunteers were followed during this research.

Table 1 Characteristics of type II Boston KPro patients included in this study
Full size table

Decision analysis

Average cost effectiveness was determined, comparing type II KPro surgery with no further intervention (the current visual state). Complications and additional procedures were incorporated into an expected-value decision tree. A list of the most common postoperative complications and treatments are given in Table 2. To make cost-utility calculations, several model assumptions were made.

Table 2 Postoperative complications and procedures over 5 years for type II Boston KPro recipients in this study
Full size table

In our sample, 5-year anatomical retention was 72.7%. (1) The time frame chosen for the CUA was 5 years because of presumed quiescence of disease at this postoperative time; (2) a yearly KPro implant survival probability of 93.5% was conservatively interpolated from the data; (3) BCVA preoperatively and at 5 years postoperatively was utilized to calculate incremental utilities; (4) the average incremental utility of the 5-year cohort represented that of the entire sample; (5) the mean patient age was 66, and we postulated that this was representative of this population.

Utility assessment

The time trade-off method was used for patient-based utility assessment in the CUA. The mean (±SD) preoperative utility value was 0.391 (±0.1.36), increasing to 0.568 (±0.224) at 5 years postoperatively. Decreases in vision were accounted for by incorporating negative utilities into the mean incremental utility calculation. The mean incremental utility in our study population at 5 years was 0.177. The total QALY gain (comparative effectiveness) was also determined by multiplying the years of utility gain by years of benefit duration and comparing it with the preoperative utility (quality of life) state.

Identification of costs

The costs for the KPro type II device, surgical procedure, hospitalization, and follow-up management were obtained from the Massachusetts Eye and Ear Infirmary CPT diagnosis and procedure codes. The CUA was conducted based on the Medicare reimbursement rates in 2009. Ophthalmologic visits were conducted every 3 months for the first year, followed by every 6 months over the second year, and then annually. Maintenance costs included medications and patient travel. An overview of all costs is given in Table 3.

Table 3 2010 Medicare reimbursement schedule associated with Boston Kpro type II
Full size table

The expected-value decision tree is illustrated in Figure 3. Future costs, such as follow-up visits and possible complications requiring procedures, are weighted based on normal clinical practice and rates observed in our sample (refer to Table 2). Total costs were calculated by the summation of initial costs with appropriately discounted future costs. Discounted future costs included inevitable costs and the average of weighted probable/possible future costs.

Figure 3
figure 3

Expected value decision tree for the type II Boston keratoprosthesis—preoperative, perioperative, and postoperative management. Decision tree is broken down into initial, inevitable, probable, and possible costs. Costs are reported on the basis of the 2010 allowable reimbursement from Medicare. Probability of occurrence is displayed below the decision/cost item. A single asterisk indicates costs that require appropriate discounting. Discounting can occur over some or all of the years being assessed in this study, depending on need and utilization of the service. Furthermore, in some instances, the cost changes after the initial year of use.

Full size image

Discounting

There is a consensus that both utilities and costs should be discounted in health care economics analyses. We decided that the commonly used 3% discount rate was appropriate.17

Calculation of QALYs

The QALYs for the 5-year period was calculated using the following formula:18, 19

where t is the yearly transplant survival rate (0.935), u is the average incremental utility (0.177), and dq is the discounting rate for QALYs (3%).

Calculation of cost

The equation for the total discounted cost associated with KPro surgery is:

An initial cost that was incurred at or immediately before or following the time of surgery was not discounted. Costs paid for over the initial year alone were discounted accordingly as were costs paid for throughout the entire time period. In the equation, x represents the year of follow-up and dc is the discounted rate for costs (3%).

Sensitivity analysis

The model was assessed using a univariate sensitivity analysis (Table 4.). The relevant parameters included utility value, retention rate, discounting rate for QALYs, and discounting rate for costs. Each parameter was varied at fixed intervals individually.

Table 4 Univariate sensitivity analysis of the cost-utility analysis for type II Boston Kpro
Full size table

Results

Median preoperative BCVA in the treated eye was logMAR 2.3±0.7 (Snellen equivalent HM). At 5 years postoperatively, the median BCVA increased to logMAR 1.30±1.17 (Snellen equivalent of 20/400). A total discounted incremental QALY gain of 0.668 was obtained for type II KPro. This correlates with a conferred QALY gain (or improvement in quality of life) of 8.7% for the average patient. The total discounted cost associated with this utility equaled $42 215. Using the current parameters, the cost utility of KPro from third-party insurer (Medicare) perspective was 63 196 $/QALY. The univariate sensitivity analysis resulted in a range of incremental cost-effectiveness ratios from 52 078 to 83 871 $/QALY.

Discussion

As noted in the paper by Ament et al., describing the cost effectiveness of type I Boston KPro, the commonly cited guideline considers interventions costing below 20 000 $/QALY as highly cost effective and interventions costing more than 100 000 $/QALY as not cost effective.20 The UK National Institute for Health and Clinical Excellence (NICE) uses 60 000 $/QALY to define cost-effective treatments.21 It is nevertheless recognized that these benchmarks and the unit $/QALY, as a measure of value in medicine, are inherently limited. Indeed, insurance companies and national health boards often rebuff reimbursement below these guidelines, fund beyond them, or develop novel pricing arrangements to expand access of otherwise less cost effective interventions.22 Various cost-effective values ($/QALY) for several medical interventions are illustrated in Table 5.

Table 5 Cost utility of various medical interventions in the US, adjusted to 2010 US dollars23
Full size table

In this analysis, only patients with 5 years of follow-up data were included. Although the 5-year sample was small, it was determined that 2- to 3-year follow-up was insufficient for this population. Based on anecdotal evidence, severe complications remain a concern well after the 2-year postoperative period in autoimmune patients undergoing type II KPro surgery. This is unlike type I KPro, in which visual gains can ostensibly be maintained almost indefinitely. Despite this, it is important to note that those patients considered eligible for type II KPro implantation typically experience severe, debilitating sequellae of their underlying disease process, and may perceive a limited and transient 2-year improvement, irrespective of the complication rates and risks, as significant, thereby possessing inherent utility. Indeed, 16 patients had complete 2 years of follow-up and, on average, improved from HM vision to 20/70. The average incremental utility increase for this cohort was exceedingly high at 0.278. This, when compared with a 0.177 average incremental utility increase in our 5-year cohort, represents a 57% increase in average utility change. Reassessing 2-year costs and conducting the appropriate cost-utility calculations yield a cost-effective value of 31 719 $/QALY for this 2-year sample. These markedly disparate values illustrate that a dramatic change occurs in type II KPro population after the 2-year postoperative period. Although physiologic and pathologic processes need to be further elucidated, it is noteworthy that the cost-effective curve, as it relates to visual acuity, appears to be non-linear. Worsening vision and the associated substantial decrease in incremental utility disproportionately affect the cost-effective calculation as compared with only moderate increases in costs. Furthermore, vision changes from HM to 20/400, for example, are associated with a greater utility than, say, improvement from 20/200 to 20/20.

Despite the limited sample, the original 5-year study cohort demonstrated remarkable improvement in median visual acuity within 5 years from HM to 20/400 for a moderately cost-effective value of 63 196 $/QALY. Though optimistic for suggesting possible expansion of type II KPro for the management of patients with severe autoimmune ophthalmic conditions, the complexities of postoperative management should not be underestimated. Consequently, the limitations of this analysis require proper evaluation. Patients undergoing type II KPro surgery must commit to a twice-daily regimen of antibiotic eye drops for life, sometimes take toxic and expensive systemic immunosuppressive drugs, be willing to accept the cosmetic burden of the device, and be prepared for life-long follow-up with an experienced type II KPro surgeon. These requirements may be burdensome for some patients and therefore represent disutilities of the device not accounted for in this retrospective analysis. It is possible that the 0.668 QALYs that we observed would decrease in light of these disutilities and that the ultimate cost effectiveness of the device would be diminished. The lack of a validated testing mechanism to assess patient satisfaction and preferences postoperatively limits this and other cost-effective analyses.

Additional limitations of this analysis were its restriction to one location and the lack of available cost-utility literature to directly compare type II KPro with the modified OOKP—the latter being, arguably, the accepted standard of care outside the USA for this patient population. Nonetheless, the authors felt that astype II KPro is performed so infrequently and without adequate standardization outside of MEEI that including external type II KPro data into this analysis would have injected incalculable variation. With respect to the OOKP, only one study noted the cost of rehabilitating end-stage ocular surface disease with OOKP surgery (13 661 pounds or 21 786 USD).38 A case report by Geerling et al.9 did demonstrate an overall cost savings with the OOKP of ∼7400 USD over 2 years. However, both of these only examined costs and savings of the OOKP, and did not include formal cost-utility analyses. Review of the literature suggests that no such analysis exists, making appropriate comparisons impossible at this time. Given the lack of comparative data and a large multicentered sample, external validity is limited, and it is difficult to determine whether the cost effectiveness of type II KPro will decrease or increase ,as it becomes more utilized in the clinical setting.

Excluded from this cost analysis is a consideration of the costs incurred before type II KPro implantation. Of the 11 patients included in this analysis, 4 underwent previous penetrating keratoplasty, 3 underwent type I KPro implantation once, and 1 underwent type I KPro implantation twice, all of which failed. Additionally, seven patients had glaucoma valves in place at the time of surgery. In contrast, expensive systemic immunosuppressive therapies were assumed to be continued for the entire 5 years and may represent an overestimation of costs. Furthermore, a significant source of additional cost for type II KPro lies in the high probability of revision and/or replacement. This could be greatly mitigated by improving device design and biocompatibility, and exemplifies the need to revisit this analysis in the future. Taking the risks, complications, and morbidity associated with type II KPro into account, efforts to refer patients who are less likely to benefit from traditional corneal transplantation or type I KPro may be helpful in decreasing both the ultimate personal and societal cost of surgical intervention in these patients, as evidenced by the multiple, failed procedures that occur in this population.