Cost-Effectiveness of Osteo-Odonto Keratoprosthesis in Singapore

Purpose

To determine the long-term cost-effectiveness of osteo-odonto keratoprosthesis (OOKP) relative to no treatment among patients with end-stage corneal and ocular surface diseases in Singapore.

Design

Cost-effectiveness analysis based on data from a retrospective cohort study.

Methods

From a health system perspective, we calculated the incremental cost-effectiveness ratio of OOKP treatment relative to no treatment over a 30-year horizon, based on data from a cohort of 23 patients who underwent OOKP surgery between 2004 and 2009 at Singapore National Eye Centre. Preoperative and postoperative vision-related quality-of-life values were estimated from patients’ visual outcomes and were used to calculate the gain in quality-adjusted life years (QALYs) resulting from OOKP treatment. Unsubsidized costs for surgery, consultations, examinations, medications, follow-up visits, and treatments for complications were retrieved from patients’ bills to estimate the total costs associated with OOKP treatment. Sensitivity analyses were conducted to test the robustness of the model.

Results

Over a 30-year period, OOKP treatment, compared with no treatment, improved QALYs by 3.991 among patients with end-stage corneal and ocular surface diseases at an additional cost of S$67 840 (US$55 150), resulting in an incremental cost-effectiveness ratio of S$17 000/QALY (US$13 820/QALY).

Conclusions

Based on commonly cited cost-effectiveness benchmarks, the OOKP is a cost-effective treatment for patients with end-stage corneal and ocular surface diseases.

Stevens-Johnson syndrome, chemical burns, and thermal burns can all result in end-stage corneal disease and a poor ocular surface. This will manifest clinically with patients having poor vision and pain from ocular surface inflammation that can affect their daily functioning and quality of life. Patients with severe end-stage corneal disease caused by ocular surface failure typically have a dry and keratinized ocular surface; therefore, traditional corneal grafts, even if combined with limbal stem cell reconstruction, carry a poor prognosis.

In such patients, keratoprostheses are an option. Possible keratoprostheses include the type I and type II Boston (Massachusetts Eye and Ear Infirmary, Boston, Massachusetts, USA; PMMA disc, J.G. Machine Company, Woburn, Massachusetts, USA), AlphaCor (Argus Biomedical Pty. Ltd., Perth, Australia), Pintucci (developed by Stefano Pintucci of Rome, Italy; no longer available), Seoul-type (Seon-Kyoung Company, Seoul, South Korea), and the osteo-odonto keratoprosthesis (OOKP [PMMA optical cylinder, Schalcon, Rome, Italy]). Unlike other synthetic keratoprostheses, the OOKP aims to replace the natural cornea by using an autologous bony lamina to provide a skirt material for a central optical cylinder. The type II Boston, Pintucci, and Seoul-type keratoprostheses were designed to be used in severe end-stage corneal disease and have similar indications to the OOKP. However, clinical outcomes suggest that the OOKP remains the best option in such patients. A systematic review of OOKP outcomes revealed that the average anatomic survival rate was 87.8% at 5 years and 81% at 20 years, with visual acuity more than 6/18 achieved in 52% of the operated eyes.

Although the OOKP improves vision and quality of life (QoL), it is expensive. Few studies have examined the cost-effectiveness of the keratoprostheses. To our knowledge, only 2 cost-effectiveness analyses have been performed to evaluate the type I and type II Boston keratoprostheses based on data from a medical center in the United States, and the cost-effectiveness of the OOKP remains unknown. We present a cost-effectiveness analysis of the OOKP compared with no treatment from the health system perspective based on data from the Singapore National Eye Centre.

Patients and Procedures

The study was approved by the Singhealth Centralised Institutional Review Board. A retrospective cohort of 23 patients who underwent OOKP surgery between 2004 and 2009 at Singapore National Eye Centre and had at least 2 years of follow-up were identified. The average age of the 23 patients at the time of surgery was 36 ± 13 years, and the most common indications for surgery were Stevens-Johnson syndrome (11 patients) and chemical injury (11 patients). All patients had bilateral blindness, and the eye with better visual acuity potential was chosen for surgery. OOKP staged surgery was performed as described previously. Briefly, stage 1a involved covering the ocular surface of the eye with a full-thickness buccal mucosa graft. In stage 1b, a single rooted tooth and surrounding intact alveolar bone were harvested. The optical cylinder was cemented into the dentine and implanted into the cheek. Stage 2 occurred 2 to 4 months later, when the cornea and anterior segment contents were removed and the implant was sutured under the buccal mucosa, allowing the optical cylinder to protrude through.

Decision Analysis

A decision tree model was used to calculate the long-term cost-effectiveness of OOKP surgery in comparison with no treatment. Based on 2-year follow-up data and long-term outcomes reported by another study on OOKP, the model was populated for a 30-year time horizon, roughly representing the average remaining life expectancy of patients in our cohort. The remaining life expectancy was calculated based on the conservative assumption that OOKP treatment does not prolong life and the evidence suggesting that the life expectancy of a blind person is two thirds that of the general population, which is 82 years in Singapore.

The effectiveness of OOKP treatment was quantified in quality-adjusted life years (QALYs), defined as remaining life years after adjusting for expected QoL. Vision-related QoL was estimated from best spectacle-corrected visual acuity (BSCVA). For patients whose visual acuity was measured using a Snellen chart, QoL was calculated using an empirical formula derived by Lansingh and Carter, which quantified the relationship between BSCVA and QoL based on results from time-tradeoff experiments. The formula is:

y = − 0.0479 x 3 + 0.191 x 2 − 0.4233 x + 0.9128
where y is the vision-related QoL and x is the visual acuity in logarithm of the minimal angle of resolution (logMAR) units. The predicted QoL was adjusted further for 2 factors. First, OOKP devices give patients a smaller visual field, which can compromise QoL independently from visual acuity. To account for this, we downwardly adjusted the QoL based on a patient’s postoperative visual field following the empirical correlation between QoL and visual field obtained by Freeman and associates. Second, postoperative visual acuity is likely to deteriorate in the long term. Hence in our base case analysis, we modeled a 0.02 increase in logMAR each year after OOKP surgery based on limited long-term data from another study.

For patients whose visual acuity was classified as counting fingers, hand movements, light perception, and no light perception, their QoL values were assigned to be 0.479, 0.372, 0.3, and 0.26, respectively, based on existing studies that directly survey the QoL of these patients using a time-tradeoff approach or on empirical studies that estimate the Snellen ratio equivalent of the semiquantitative scales.

The improvement in vision-related QoL resulting from OOKP treatment was calculated as the difference between the postoperative vision-related QoL and preoperative vision-related QoL. The incremental QALYs over a 30-year horizon were calculated using the following formula:

Δ QALY = ∑ t = 1 30 Δ Q o L t ( 1 + d ) t − 1