- Email: [email protected]
A Cost-effectiveness comparison of bimatoprost versus latanoprost in patients with glaucoma or ocular hypertension
SURVEY OF OPHTHALMOLOGY
VOLUME 49 • SUPPLEMENT 1 • MARCH 2004
A Cost-effectiveness Comparison of Bimatoprost Versus Latanoprost in Patients with Glaucoma or Ocular Hypertension John G. Walt, MBA, and Jeffrey T. Lee, PharmD, FCCP Global Health Outcomes Strategy and Research, Allergan, Inc., Irvine, CA
Abstract. To compare the cost-effectiveness of bimatoprost versus latanoprost for the treatment of glaucoma and ocular hypertension, a cost-effectiveness algorithm was developed to estimate the annual costs and cost per treatment success (cost-effectiveness) for bimatoprost 0.03% QD compared with latanoprost 0.005% QD. Medication costs for the model were abstracted from published sources. Clinical assumptions for the model were based on the treatment success rates from one 3-month controlled clinical trial comparing bimatoprost versus latanoprost (n ⫽ 119 and 113, respectively), and another 6-month controlled clinical trial comparing bimatoprost versus latanoprost (n ⫽ 133 and 136, respectively). Treatment success was defined as the percentage of patients achieving various target intraocular pressures. A larger percentage of patients achieved low target IOPs on bimatoprost than on latanoprost. The cost per treatment success for patients who started treatment on bimatoprost monotherapy was less than for patients started on latanoprost, despite the fact that the estimated yearly costs were similar for these drugs. In this model bimatoprost was found to be more cost-effective than latanoprost, given the average wholesale price of these medications and the assumptions of the costeffectiveness algorithm about clinical success at the 3- and 6-month decision points. (Surv Ophthalmol 49(Suppl 1):S36–S44, 2004. 쑖 2004 Elsevier Inc. All rights reserved.) Key words. economics
bimatoprost
•
cost-effectiveness
•
latanopost
•
pharmacoeconomics
•
health
reimbursement formularies.7 These organizations recognize that cost-effectiveness (the cost to reach a particular desired outcome) may be a more relevant measure of cost than drug average wholesale price (AWP) alone. Cost-effectiveness analysis is a widely used method to compare the costs and clinical outcomes of competing treatment options. Cost-effectiveness analysis provides an estimate of the costs incurred to achieve a particular outcome. It is measured by dividing a therapy’s total cost by its therapeutic effectiveness, which might be cure rate, remission rate, or some
Pharmacoeconomics is the study of the various economic costs associated with prescribing a given drug. In order to improve their decision-making processes, managed care organizations (MCOs) request pharmacoeconomic data. In the U.S., some MCOs now require pharmacoeconomic data to be included with requests to add a medication to their formularies. The Academy of Managed Care Pharmacy Formulary Submission Guide requires pharmacoeconomic data as well. Canada, Australia, New Zealand, and many European countries require pharmacoeconomic evaluations before drugs may be placed on public S36
쑖 2004 by Elsevier Inc. All rights reserved.
0039-6257/04/$–see front matter doi:10.1016/j.survophthal.2003.12.018
S37
MEDICAL MANAGEMENT OF GLAUCOMA
other end point depending on the drug and disease involved.7 A cost-effectiveness comparison of two hypothetical drugs is shown in Fig. 1: If drug A (AWP US$150) and drug B (AWP US$100) each induced a 100% success rate, the cost-effectiveness would be equal to the original cost. If drug A’s success rate were 100%, its cost-effectiveness would be US$150/ 100% or US$150 for one success. In contrast, if drug B’s success rate were only 50%, then the cost-effectiveness for drug B would be US$100/50% or US$200 for one success. Although a preferred drug in most scenarios should possess the best possible clinical efficacy, its cost-effectiveness (cost per treatment success) should also be as low as possible. In addition, the incremental cost effectiveness ratio (ICER), which is the cost to achieve one additional treatment success, should also be as low as possible. ICER is commonly used by MCOs to evaluate the relative merits of substituting a new treatment for an existing standard treatment7; it should show that a preferred drug is more effective and either less expensive than, or at an acceptable additional cost to, the comparator. Bimatoprost and latanoprost are functional analogs of PGF2α that have demonstrated superior efficacy to timolol in lowering IOP in patients with glaucoma or ocular hypertension (OHT).2,5 Several short-term studies indicated that bimatoprost lowered IOP numerically more than latanoprost did,3,4,11 although not all differences were statistically significant. A controlled 6-month trial demonstrated that bimatoprost lowered IOP significantly more than latanoprost at every measurement at every visit,10 thus supporting the results from the shorter term studies. In a cost analysis, Mick et al found a mean drop volume for bimatoprost of .0297 ml and .0324 ml with latanoprost; mean bottle volume was 3.06 ml for
bimatoprost (22.4% overfill) and 2.98 ml for latanoprost (19.2% overfill).9 Given that the AWP of bimatoprost and latanoprost are similar (US$55.79 and US$55.60, respectively, for 2.5 ml bottles),13 the purpose of this economic analysis was to determine the relative cost-effectiveness of bimatoprost versus latanoprost.
Methods This cost-evaluation model was built from a computer-based algorithm that used two types of data inputs: 1) costs for bimatoprost and latanoprost based on published sources, and 2) evidence-based clinical assumptions about treatment effectiveness. SOURCES OF COST INFORMATION
The average wholesale cost of every drug available by prescription in the United States is published periodically.13 The algorithms used in this study were based on 2003 costs (Table 1). SOURCES OF CLINICAL EFFECTIVENESS DATA
In this cost-effectiveness algorithm, a hypothetical patient is prescribed either bimatoprost or latanoprost monotherapy at baseline. The clinician will decide whether to maintain the treatment or change it at certain decision points (i.e., after 3 months and 6 months of treatment). The treatment decisions are based on evidence of each drug’s clinical effectiveness derived retrospectively from two randomized clinical trials that compared bimatoprost and latanoprost in patients with glaucoma and OHT, one of 3months’ duration,4 and one of 6-months’ duration.10 PATIENT POPULATIONS
Eligibility criteria for both studies included 1) diagnosis of OHT or glaucoma; and 2) IOP ⱖ22 and ⱕ34 mm Hg in at least one eye. Exclusion criteria included 1) any condition or medication that would put a patient at risk during the course of the study or interfere with interpretation of the study results; 2) any
TABLE 1
Cost Inputs
Fig. 1. Cost-effectiveness calculation. The cost-effectiveness of a hypothetical drug A is compared with drug B under two different assumptions about treatment success rates (see accompanying text).
Lumigan Xalatan Lumigan 2nd Drug Xalatan 2nd Drug Initial Doctor Visit Subsequent Doctor Visits
US$55.79 US$55.60 US$20.57 US$20.57 US$119.00 US$59.00
The Average Wholesale Price of all drugs used in the cost-effectiveness model.12,13
S38
Surv Ophthalmol 49 (Suppl 1) March 2004
sensitivity or contraindication to the study medications; and 3) women who were pregnant, nursing, or of childbearing potential who were not using a reliable form of contraception. INTERVENTIONS
In both studies, patients self-administered either bimatoprost 0.03% or latanoprost 0.005% into both eyes once daily in the evening. OUTCOME MEASURES
Because the functional analogs of PGF2α are relatively new medications, there is not much data on how well they limit progression to blindness, the outcome measure of ultimate interest, so we must use a proxy measure. In both the Gandolfi4 and Noecker10 studies, IOP was measured at several times of the day at each follow-up visit. The proxy measure relevant to the current cost evaluation was the percentage of patients achieving each specific target IOP at the 12 pm measurement at the end of the treatment period. Mean IOP and diurnal IOP were not used in this model as these data do not reflect which patients were successful and which were not, which are the fundamental factors needed for decision algorithms in pharmacoeconomic models. ANALYSIS
The efficacy analysis for both studies was based on the intent-to-treat population with last observation carried forward. The frequency distributions of patients who had achieved desirable target IOP levels were compared between groups within each study. A comparison was made for each target IOP using either Pearson’s chi square test4 or the Fisher exact test.10 Cost-effectiveness Algorithm The cost-effectiveness treatment algorithm for this study was developed by a panel of eight glaucoma experts.15 According to the algorithm, the physician will see a patient during a comprehensive new patient visit (long visit) at baseline, during which a target IOP will be established and patients will be assigned to either bimatoprost monotherapy or latanoprost monotherapy in equal numbers. An established patient follow-up visit (short visit) is scheduled 1 month after baseline for both patients. After the 1-month visit, the cost-effectiveness algorithm has two decision points. A short visit is built into the model 3 months after initiating either bimatoprost or latanoprost therapy in order to determine tolerability and initial response. At the month-3 decision point, the bimatoprost- and latanoprost-treated patients would achieve
WALT AND LEE
clinically relevant low target pressures according to the distribution of targets achieved at the month-3 visit in the study published by Gandolfi et al.4 If the patient has achieved the predetermined target pressure on bimatoprost or latanoprost monotherapy, the model assumes that the patient will continue to achieve the target pressure consistently, will remain on the assigned monotherapy without requiring additional glaucoma medications, and will be scheduled for the next physician visit at month 9. With this scenario, the costs include one medication for 12 months, a long physician visit at baseline, and short visits at 1, 3, and 9 months (Fig. 2). If the patient does not achieve the pre-established target IOP at month 3 on bimatoprost or latanoprost monotherapy, the physician will prescribe adjunctive medication. The patient will be seen 1, 3, and 9 months after initiating dual therapy (i.e., 4, 6, and 12 months after beginning the initial treatment). All patients with an inadequate response at month 3 will achieve target IOPs with dual therapy within 6 months. With this scenario, the costs include the initial medication for 12 months, the adjunctive medication for 9 months, and the baseline, 1-, 3-, 4-, 6-, and 12-month office visits (Fig. 2). Estimated Yearly Costs Each comprehensive, new patient visit costs US$119 and each established patient visit costs US$59, in accordance with the 2003 Physicians Fee & Coding Guide,12 a standardized reference used in U.S. pharmacoeconomic studies (Table 1). The monthly cost of adjunctive medication (for patients who do not achieve target IOP on monotherapy) varies depending on the treatment comparison. The model assumes that 50% of patients who need adjunctive medication receive generic timolol, 30% receive brimonidine, 10% receive the fixed combination of timolol/dorzolamide, and 10% receive dorzolamide (source: Scott-Levin marketing data, 2003). The model calculates a weighted average of the four possible adjunctive medications in the costeffectiveness calculations. Average monthly medication costs are US$55.79 for bimatoprost, US$55.60 for latanoprost, US$36.00 for brimonidine, US$5.00 for timolol, US$47.00 for timolol/dorzolamide, and US$27.00 for dorzolamide.13 Based on the distribution described above, the weighted average cost of an adjunctive medication is US$20.57 per month (Table 1). This average is used in all costeffectiveness calculations. The estimated yearly costs to achieve a specific IOP are calculated from the treatment model (Fig. 2): 12 months of monotherapy (upper path) or 3 months of monotherapy followed by 9 months of dual therapy
MEDICAL MANAGEMENT OF GLAUCOMA
S39 Fig. 2. Treatment algorithm. All patients are seen during a comprehensive new patient visit at baseline, and at a short visit at months 1 and 3. The monotherapy arm includes a short visit at month 9. The dual therapy arm includes short visits at months 4, 6, and 12.
(lower path). For purposes of this calculation, the success rate is defined as the percentage of patients who achieve a specific target IOP. Yearly drug costs were calculated for all target pressures from 13–20 mm Hg. If the target IOP is 17 mm Hg, for example, the formula to calculate yearly drug costs is (Success Rate to Achieve IOP of 17 mm Hg × Drug Cost of Upper Path) ⫹([1 ⫺ Success Rate] × Drug Cost of Lower Path) ⫽Cost ($)/patient/year to reach/maintain a specific IOP For this calculation, the percentage of patients who reached specific target IOPs was derived from a 3month comparison of bimatoprost with latanoprost (Fig. 3).4 A similarly structured formula is used to calculate yearly costs for medical visits, and the drug and visit costs are added to calculate the total yearly cost to reach that target IOP. The cost-effectiveness (cost per treatment success) was calculated by dividing the estimated yearly total costs for each drug by the treatment success rate for each target IOP. For this calculation, the treatment success rates were derived from the distribution of patients achieving specific IOPs seen in a 6-month comparison of bimatoprost and latanoprost (Fig. 4),10 because this study had the longest available follow-up period for a direct, head-to-head comparison of these two drugs. The same efficacy values were used for the 12-month calculations, supported by evidence that bimatoprost or latanoprost efficacy did not vary substantially from 6 to 12 months of followup when compared to timolol.2,5 A trial that directly compared bimatoprost, latanoprost, and travoprost was examined as a potential data source, but the
range of individual target pressures needed to construct the decision tree was not reported.11 To calculate the weighted target cost-effectiveness, the costs of the two treatment paths for each target IOP were weighted using the distribution of commonly established target IOPs derived from a 361-patient, nationwide, internet-based research survey (Walt et al, Cost-effectiveness of bimatoprost versus latanoprost plus adjunctive products for glaucoma treatment, ARVO, 2003). Participating physicians were from 32 centers throughout the United
Fig. 3. Patients who achieved target pressures at 12 pm, month 3. More patients achieved each target IOP with bimatoprost than with latanoprost. The differences between the treatment groups were statistically significant for target pressures of 13, 14, and 15 mm Hg (P ⱕ .049). Reprinted from Gandolfi et al4 with permission of Advanced Therapy.).
S40
Surv Ophthalmol 49 (Suppl 1) March 2004
WALT AND LEE
The yearly costs to reach each individual IOP were averaged over all patients to calculate the weighted target cost-effectiveness. The incremental cost-effectiveness ratio (ICER) for each target IOP was calculated using the following formula: (Total Yearly Costs of Total Yearly Costs ⫺ Bimatoprost of Latanoprost) (Success Rate of Success Rate Bimatoprost ⫺ of Latanoprost)
Results Fig. 4. Patients who achieved target pressures at 12 pm, month 6. A greater percentage of bimatoprost-treated patients achieved target pressures between 13 mm Hg and 20 mm Hg than did latanoprost-treated patients. For example, 36% of bimatoprost patients achieved pressures of 15 mm Hg or lower (at 12 pm) compared with 22% of latanoprost patients. The differences between the treatment groups were statistically significant for target pressures of 13–17, 19, and 20 mm Hg (p ⱕ 0.044). (Reprinted from Noecker et al10 with permission of American Journal of Ophthalmology.).
States, and were selected from a list of ophthalmologists who were top prescribers of glaucoma medications. Physicians were asked what patients they wanted to reach specific target IOPs. The responses of all participants (Fig. 5) were summarized to develop a weighting factor for each level of target IOP.
AVERAGE EXPECTED ANNUAL COSTS
Sample cost-effectiveness calculations for target IOPs of ⱕ17 mm Hg are highlighted, because the Advanced Glaucoma Intervention Study1 demonstrated that patients whose pressures were consistently below 18 mm Hg did not progress over a 6-year period. The average expected annual costs for an individual bimatoprost-treated patient to achieve a target pressure, would be US$669 for 12 months of medication and US$296 for one baseline and three short medical visits (Table 2). For a patient who failed to achieve the target pressure, the costs for an individual patient would be US$855 for medications (bimatoprost during months 1 through 12 plus an adjunctive medication for months 3 through 12) and US$414 for medical visits. The costs for an individual latanoprost-treated patient would be essentially the same (Table 2). The weighted average annual costs to treat a population of glaucoma patients takes into account the percentage of patients who achieve a given target weighted against the percentage of patients who fail to achieve that target. For example, to treat a population to a target pressure of 15 mm Hg or lower takes into account that 29% of bimatoprost patients achieve such a target after 3 months of treatment, while only 14% of latanoprost patients do.4 Thus, TABLE 2
Individual Annual Costs Fig. 5. Commonly established target IOPs derived from a 361-patient, nationwide, internet-based naturalistic effectiveness trial. Participating physicians were randomly selected from a population of glaucoma specialists from 32 centers throughout the United States. Physicians were asked what percentage of patients they wanted to reach specific target IOPs. Glaucoma specialists establish target pressures between 15 and 17 mm Hg for half of their patients. Other targets are chosen less frequently. (Source: Walt, personal communication).
Drug, Medical Visits, Total, US$ US$ US$ Bimatoprost Drug success @ 3 months Drug added @ 3 months Latanoprost Drug success @ 3 months Drug added @ 3 months
669 855
296 414
965 1,269
667 852
296 414
963 1,266
Calculations are based on costs to treat an individual patient.
S41
MEDICAL MANAGEMENT OF GLAUCOMA TABLE 3
Weighted Average Expected Annual Costs Total Costs, Percentage Product, US$ Patients, % US$ Bimatoprost Drug success @ 3 months 965 Drug added @ 3 months 1269 Total Average Cost Latanoprost Drug success @ 3 months 963 Drug added @ 3 months 1266 Total Average Cost
29 71
280 901 1181
14 86
135 1089 1224
Calculations are based on costs to treat a population of patients.
the weighted average expected annual costs for a bimatoprost-treated patient who achieved a target IOP of 15 mm Hg would be 29% times the US$965 cost to achieve that target plus 71% times the US$1,269 cost of failure to achieve that target (Table 3). Similarly, for latanoprost-treated patients, the weighted average expected annual costs would be 14% times the US$963 cost of success plus 86% times the US$1,266 cost of treatment failure (Table 3). COST-EFFECTIVENESS
For all target pressures in both scenarios (monotherapy with bimatoprost or with latanoprost for 12 months, or monotherapy with bimatoprost or with latanoprost for 3 months and dual therapy for 9 months), bimatoprost cost less per treatment success
than latanoprost (i.e., bimatoprost was more costeffective than latanoprost). The difference in costeffectiveness was most notable at the lowest target pressures (Fig. 6). At a target IOP of 13 mm Hg, for example, cost-effectiveness (cost per treatment success) was US$8,845 with bimatoprost and US$20,954 with latanoprost (Table 4). This contrasts with more similar cost-effectiveness between the two drugs at a target IOP of 20 mm Hg; for example, cost-effectiveness was US$1,104 with bimatoprost and US$1,288 with latanoprost (Table 4). The difference in cost-effectiveness between bimatoprost and latanoprost for a target pressure of ⱕ20 mm Hg is much smaller than with the lower targets, but it could represent substantial differences overall to an MCO that serves many glaucoma patients. Overall, at every target pressure, in this model bimatoprost costs approximately the same as and is more effective than latanoprost; in pharmacoeconomic terms, bimatoprost dominates latanoprost (Table 5). The cost-effectiveness findings were sensitive to changes in the price of products and to changes in treatment success rates. For example, the sensitivity analysis indicated that the treatment success rate for latanoprost would have to be at least 15–133% higher than the observed rate for it to be as costeffective as bimatoprost. Given that the AWP of latanoprost is only US$0.19 less than for bimatoprost, the cost-effectiveness of each drug would be closer in value only if the two drugs were equally efficacious.
Fig. 6. Comparative projected 1-year cost per treatment success.
694 701 311 317 1,005 1,018 1,104 1,288 (1005/91%) (1018/79%) Bimatoprost cost/success is 14% less than latanoprost
TABLE 5
Incremental Cost-effectiveness Ratio (ICER) Bimatoprost Relative to Latanoprost Target IOP
ICER, US$
747 760 346 355 1,093 1,115 1,681 2,477 (1093/65%) (1115/45%) Bimatoprost cost/success is 32% less than latanoprost
ⱕ13 ⱕ14 ⱕ15 ⱕ16 ⱕ17 ⱕ18 ⱕ19 ⱕ20 Weighted Average
⫺237 ⫺311 ⫺309 ⫺216 ⫺110 ⫺68 ⫺117 ⫺107 ⫺188
Discussion
801 826 380 397 1,181 1,224 3,280 5,563 (1181/36%) (1224/22%) Bimatoprost cost/success is 41% less than latanoprost Includes success and failure. a
36 6
Effectiveness in reaching low pressure, %10 Average expected cost/patient/yeara, US$ Drug Medical Total expected cost Cost effectiveness, US$ (Cost/treatment success)
14
WALT AND LEE
836 847 402 410 1,238 1,257 8,845 20,954 (1238/14%) (1257/6%) Bimatoprost cost/success is 58% less than latanoprost
91 45 65 22
Bimatoprost Bimatoprost
ⱕ13
Latanoprost
Bimatoprost
ⱕ15
Latanoprost
Target Pressure (mm Hg)
Comparative Cost-effectiveness for Selected Target Pressures
TABLE 4
ⱕ17
Latanoprost
Bimatoprost
ⱕ20
79
Surv Ophthalmol 49 (Suppl 1) March 2004 Latanoprost
S42
The most important result of the present analysis was that in this model bimatoprost is more cost-effective than latanoprost over a range of clinically relevant target pressures. Several recent studies have reinforced the evidence that the patients with the lowest IOPs experience the least deterioration in visual field over time, whether the initial diagnosis is OHT or glaucoma. In the Early Manifest Glaucoma Trial (EMGT),8 62% (78/126 ) of the untreated controls experienced deterioration of visual field, compared with 45% (58/ 129) of the treatment group. The Ocular Hypertension Treatment Study6 followed 1,636 participants randomized to observation or topical glaucoma medication. The trial tested the hypothesis that medical treatment of patients with OHT would lead to a lower incidence of primary open angle glaucoma (POAG). A target IOP of 24 mm Hg (minimum reduction of 20% from baseline) was established for patients who were randomized to treatment. After 60 months of observation, the treated group achieved a 22.5% reduction in IOP, compared with a 4.0% reduction in the untreated group. After 60 months of observation, the cumulative probability of developing POAG was 4.4% in the treated group, compared with 9.5% in the observation group. A recent study has demonstrated that the annual direct costs for glaucoma resources, including physician visits, glaucoma surgeries, and medication use, increased as the severity of glaucoma increased (Lee et al, A multi-center, retrospective study of resource utilization associated with severity of disease in glaucoma, AAO, 2002). Using a six-stage glaucoma staging system, the average direct cost of glaucoma treatment for patients with early stage disease was US$523/patient/year. The average cost for patients with end-stage disease was four times higher
S43
MEDICAL MANAGEMENT OF GLAUCOMA
(US$2,200/patient/year). By choosing the medication that is the most cost-effective method of achieving low target pressures, the physician can both increase the potential for saving the patient’s vision and decrease the lifetime costs of treatment. The results of the present analyses suggest that the most medically effective option for glaucoma patients that need to be maintained at very low IOPs can also be the most cost-effective option when measured in terms of cost per treatment success. For example, at a target pressure of ⱕ13 mm Hg, bimatoprost costs approximately 58% less per treatment success than does latanoprost. In this model, data from two different clinical trials were used to estimate what percentages of patients would reach clinically relevant target pressures. In the 3- and 6-month direct comparisons of bimatoprost and latanoprost from which the success rates were derived,4,10 a larger percentage of patients on bimatoprost reached low target pressures. In addition, an unpublished study was used to provide a weighting factor for the calculation of the weighted target cost-effectiveness (Walt et al, ISPOR, 2003). Pharmacoeconomic studies are considered to be especially rigorous if they derive treatment success rate estimates from more than one study, especially when results from the two studies are consistent with each other.7,14 One limitation of this cost-effectiveness analysis is that this cost-effectiveness algorithm is predicated on a 1-year treatment period, whereas glaucoma is a chronic disease requiring long-term treatment. This may limit the generalizability of results. This restriction, however, was necessary because trials about bimatoprost or latanoprost that extend to longer than12 months do not include the range of individual target pressures on which to base additional treatment-success decision points. Another reason for choosing a 12-month model is that the use of a longer interval would involve including costs for potential surgical interventions. The assumption that medication alone will be tried for a year may be violated at the lowest target pressures (12–14 mm Hg) because advanced glaucoma patients who are assigned to these targets are likely to be treated with medication for several months and then be scheduled for surgery. Modeling surgical costs would add costs to the failure arms of the decision tree, and would complicate both treatment arms proportionately. Therefore, surgical costs were not considered. The latter factor, however, may not be of prime importance, because there is evidence that the frequency of surgical interventions for glaucoma are dropping in the United States, perhaps as a result of the availability of functional analogs of PGF2α
that are highly effective in decreasing IOP.16 In addition, the cost-effectiveness measures based on this algorithm stay stable only so long as the AWPs of bimatoprost and latanoprost remain fixed. Another limitation of this study is that the cost of potential side effects was not factored into the model, although a larger percentage of bimatoprost patients than of latanoprost patients develop such adverse reactions as hyperemia, pruritus, or allergy.10,11 Patients who develop side effects would require more physician visits than patients who don’t. However, it is hard to assign costs to adverse events because patients and physicians outside the context of a controlled clinical trial behave more variably than they do under the trial environment where all physician visits and drugs are paid for.
Conclusions Evidence from recent clinical trials supports the need to lower IOP in glaucoma patients as low as is safely possible, and ocular hypotensive lipids are the newest chemical entities that can help attain that goal. Based on retrospective efficacy data from two clinical studies that directly compared bimatoprost and latanoprost, and on a simple pharmacoeconomic economic model decision-tree model, bimatoprost is a cost-effective alternative to latanoprost in treating patients with glaucoma.
Method of Literature Search The National Library of Medicine database was searched using the PubMed engine (www.ncbi.nlm. nih.gov/entrez/query.fcgi). All available years (1966–2003) were included. Foreign language articles were ignored. Keywords used in the search: cost and cost analysis; glaucoma/economics/therapy; health care costs; health resources/utilization; health services research; insurance, health reimbursement/economics; intraocular pressure; econometric models; ocular hypertension.
References 1. The AGIS Investigators: The advanced glaucoma intervention study (AGIS) 7. The relationship between control of intraocular pressure and visual field deterioration. Am J Ophthalmol 130:429–40, 2000 2. Camras CB, Alm A, Watson P, Stjernschantz J: Latanoprost, a prostaglandin analog, for glaucoma therapy. Efficacy and safety after 1 year of treatment in 198 patients. Latanoprost Study Groups. Ophthalmology 103:1916–24, 1996 3. DuBiner H, Cooke D, Dirks M: Efficacy and safety of bimatoprost in patients with elevated intraocular pressure: a 30day comparison with latanoprost. Surv Ophthalmol 45(Suppl 4):S353–60, 2001 4. Gandolfi S, Simmons ST, Sturm R: Three-month comparison of bimatoprost and latanoprost in patients with glaucoma and ocular hypertension. Adv Ther 18:110–21, 2001
S44 5.
6.
7. 8. 9. 10.
11.
Surv Ophthalmol 49 (Suppl 1) March 2004 Higginbotham EJ, Schuman JS, Goldberg I: One-year, randomized study comparing bimatoprost and timolol in glaucoma and ocular hypertension. Arch Ophthalmol 120: 1286–93, 2002 Kass MA, Heuer DK, Higginbotham EJ: The Ocular Hypertension Treatment Study: a randomized trial determines that topical ocular hypotensive medication delays or prevents the onset of primary open-angle glaucoma. Arch Ophthalmol 120:701–713; discussion 829–30, 2002 Kobelt G: Health Economics: An Introduction to Economic Evaluation. London: Office of Health Economics, 2002 Leske MC, Heijl A, Hussein M: Factors for glaucoma progression and the effect of treatment: the early manifest glaucoma trial. Arch Ophthalmol 121:48–56, 2003 Mick AB, Gonzalez S, Dunbar MT, McSoley JJ: A Cost analysis of the prostaglandin analogs. Optometry 73:614–9, 2002 Noecker RS, Dirks MS, Choplin NT: A six-month randomized clinical trial comparing the intraocular pressure-lowering efficacy of bimatoprost and latanoprost in patients with ocular hypertension or glaucoma. Am J Ophthalmol 135:55–63, 2003 Parrish RK, Palmberg P, Sheu WP: A comparison of latanoprost, bimatoprost, and travoprost in patients with elevated
WALT AND LEE
12. 13. 14. 15. 16.
intraocular pressure: a 12-week, randomized, masked-evaluator multicenter study. Am J Ophthalmol 135:688–703, 2003 Physicians Fee & Coding Guide. Augusta,GA: MAG Mutual HealthCare Consultants, 2003 PriceAlert, The Official Guide to Average Wholesale Pricing. Indianapolis: Facts & Comparisons, 2003 Rittenhouse B: Uses of models in economic evaluations of medicines and other health technologies. London: Office of Health Economics, 1996 Serle J, Cantor L, Gross R: Best practice treatment algorithm for primary open-angle glaucoma: implications for U.S. ophthalmology practice. Manag Care Interface 15:37–48, 2002 Strutton DR, Walt JG: Trends in glaucoma surgery before and after the introduction of new topical glaucoma pharmacotherapies. J Glaucoma (in press)
These data were presented at the Annual Meeting of Managed Care Pharmacy; October 17–20, 2001; Dallas, TX. John Walt and Jeffrey Lee are employees of Allergan, Inc. Reprint address: John Walt, MBA, Global Pharmacoeconomic Strategy and Research, Allergan Inc, 2525 Dupont Drive, Irvine, CA, 92612-1599.
VOLUME 49 • SUPPLEMENT 1 • MARCH 2004
A Cost-effectiveness Comparison of Bimatoprost Versus Latanoprost in Patients with Glaucoma or Ocular Hypertension John G. Walt, MBA, and Jeffrey T. Lee, PharmD, FCCP Global Health Outcomes Strategy and Research, Allergan, Inc., Irvine, CA
Abstract. To compare the cost-effectiveness of bimatoprost versus latanoprost for the treatment of glaucoma and ocular hypertension, a cost-effectiveness algorithm was developed to estimate the annual costs and cost per treatment success (cost-effectiveness) for bimatoprost 0.03% QD compared with latanoprost 0.005% QD. Medication costs for the model were abstracted from published sources. Clinical assumptions for the model were based on the treatment success rates from one 3-month controlled clinical trial comparing bimatoprost versus latanoprost (n ⫽ 119 and 113, respectively), and another 6-month controlled clinical trial comparing bimatoprost versus latanoprost (n ⫽ 133 and 136, respectively). Treatment success was defined as the percentage of patients achieving various target intraocular pressures. A larger percentage of patients achieved low target IOPs on bimatoprost than on latanoprost. The cost per treatment success for patients who started treatment on bimatoprost monotherapy was less than for patients started on latanoprost, despite the fact that the estimated yearly costs were similar for these drugs. In this model bimatoprost was found to be more cost-effective than latanoprost, given the average wholesale price of these medications and the assumptions of the costeffectiveness algorithm about clinical success at the 3- and 6-month decision points. (Surv Ophthalmol 49(Suppl 1):S36–S44, 2004. 쑖 2004 Elsevier Inc. All rights reserved.) Key words. economics
bimatoprost
•
cost-effectiveness
•
latanopost
•
pharmacoeconomics
•
health
reimbursement formularies.7 These organizations recognize that cost-effectiveness (the cost to reach a particular desired outcome) may be a more relevant measure of cost than drug average wholesale price (AWP) alone. Cost-effectiveness analysis is a widely used method to compare the costs and clinical outcomes of competing treatment options. Cost-effectiveness analysis provides an estimate of the costs incurred to achieve a particular outcome. It is measured by dividing a therapy’s total cost by its therapeutic effectiveness, which might be cure rate, remission rate, or some
Pharmacoeconomics is the study of the various economic costs associated with prescribing a given drug. In order to improve their decision-making processes, managed care organizations (MCOs) request pharmacoeconomic data. In the U.S., some MCOs now require pharmacoeconomic data to be included with requests to add a medication to their formularies. The Academy of Managed Care Pharmacy Formulary Submission Guide requires pharmacoeconomic data as well. Canada, Australia, New Zealand, and many European countries require pharmacoeconomic evaluations before drugs may be placed on public S36
쑖 2004 by Elsevier Inc. All rights reserved.
0039-6257/04/$–see front matter doi:10.1016/j.survophthal.2003.12.018
S37
MEDICAL MANAGEMENT OF GLAUCOMA
other end point depending on the drug and disease involved.7 A cost-effectiveness comparison of two hypothetical drugs is shown in Fig. 1: If drug A (AWP US$150) and drug B (AWP US$100) each induced a 100% success rate, the cost-effectiveness would be equal to the original cost. If drug A’s success rate were 100%, its cost-effectiveness would be US$150/ 100% or US$150 for one success. In contrast, if drug B’s success rate were only 50%, then the cost-effectiveness for drug B would be US$100/50% or US$200 for one success. Although a preferred drug in most scenarios should possess the best possible clinical efficacy, its cost-effectiveness (cost per treatment success) should also be as low as possible. In addition, the incremental cost effectiveness ratio (ICER), which is the cost to achieve one additional treatment success, should also be as low as possible. ICER is commonly used by MCOs to evaluate the relative merits of substituting a new treatment for an existing standard treatment7; it should show that a preferred drug is more effective and either less expensive than, or at an acceptable additional cost to, the comparator. Bimatoprost and latanoprost are functional analogs of PGF2α that have demonstrated superior efficacy to timolol in lowering IOP in patients with glaucoma or ocular hypertension (OHT).2,5 Several short-term studies indicated that bimatoprost lowered IOP numerically more than latanoprost did,3,4,11 although not all differences were statistically significant. A controlled 6-month trial demonstrated that bimatoprost lowered IOP significantly more than latanoprost at every measurement at every visit,10 thus supporting the results from the shorter term studies. In a cost analysis, Mick et al found a mean drop volume for bimatoprost of .0297 ml and .0324 ml with latanoprost; mean bottle volume was 3.06 ml for
bimatoprost (22.4% overfill) and 2.98 ml for latanoprost (19.2% overfill).9 Given that the AWP of bimatoprost and latanoprost are similar (US$55.79 and US$55.60, respectively, for 2.5 ml bottles),13 the purpose of this economic analysis was to determine the relative cost-effectiveness of bimatoprost versus latanoprost.
Methods This cost-evaluation model was built from a computer-based algorithm that used two types of data inputs: 1) costs for bimatoprost and latanoprost based on published sources, and 2) evidence-based clinical assumptions about treatment effectiveness. SOURCES OF COST INFORMATION
The average wholesale cost of every drug available by prescription in the United States is published periodically.13 The algorithms used in this study were based on 2003 costs (Table 1). SOURCES OF CLINICAL EFFECTIVENESS DATA
In this cost-effectiveness algorithm, a hypothetical patient is prescribed either bimatoprost or latanoprost monotherapy at baseline. The clinician will decide whether to maintain the treatment or change it at certain decision points (i.e., after 3 months and 6 months of treatment). The treatment decisions are based on evidence of each drug’s clinical effectiveness derived retrospectively from two randomized clinical trials that compared bimatoprost and latanoprost in patients with glaucoma and OHT, one of 3months’ duration,4 and one of 6-months’ duration.10 PATIENT POPULATIONS
Eligibility criteria for both studies included 1) diagnosis of OHT or glaucoma; and 2) IOP ⱖ22 and ⱕ34 mm Hg in at least one eye. Exclusion criteria included 1) any condition or medication that would put a patient at risk during the course of the study or interfere with interpretation of the study results; 2) any
TABLE 1
Cost Inputs
Fig. 1. Cost-effectiveness calculation. The cost-effectiveness of a hypothetical drug A is compared with drug B under two different assumptions about treatment success rates (see accompanying text).
Lumigan Xalatan Lumigan 2nd Drug Xalatan 2nd Drug Initial Doctor Visit Subsequent Doctor Visits
US$55.79 US$55.60 US$20.57 US$20.57 US$119.00 US$59.00
The Average Wholesale Price of all drugs used in the cost-effectiveness model.12,13
S38
Surv Ophthalmol 49 (Suppl 1) March 2004
sensitivity or contraindication to the study medications; and 3) women who were pregnant, nursing, or of childbearing potential who were not using a reliable form of contraception. INTERVENTIONS
In both studies, patients self-administered either bimatoprost 0.03% or latanoprost 0.005% into both eyes once daily in the evening. OUTCOME MEASURES
Because the functional analogs of PGF2α are relatively new medications, there is not much data on how well they limit progression to blindness, the outcome measure of ultimate interest, so we must use a proxy measure. In both the Gandolfi4 and Noecker10 studies, IOP was measured at several times of the day at each follow-up visit. The proxy measure relevant to the current cost evaluation was the percentage of patients achieving each specific target IOP at the 12 pm measurement at the end of the treatment period. Mean IOP and diurnal IOP were not used in this model as these data do not reflect which patients were successful and which were not, which are the fundamental factors needed for decision algorithms in pharmacoeconomic models. ANALYSIS
The efficacy analysis for both studies was based on the intent-to-treat population with last observation carried forward. The frequency distributions of patients who had achieved desirable target IOP levels were compared between groups within each study. A comparison was made for each target IOP using either Pearson’s chi square test4 or the Fisher exact test.10 Cost-effectiveness Algorithm The cost-effectiveness treatment algorithm for this study was developed by a panel of eight glaucoma experts.15 According to the algorithm, the physician will see a patient during a comprehensive new patient visit (long visit) at baseline, during which a target IOP will be established and patients will be assigned to either bimatoprost monotherapy or latanoprost monotherapy in equal numbers. An established patient follow-up visit (short visit) is scheduled 1 month after baseline for both patients. After the 1-month visit, the cost-effectiveness algorithm has two decision points. A short visit is built into the model 3 months after initiating either bimatoprost or latanoprost therapy in order to determine tolerability and initial response. At the month-3 decision point, the bimatoprost- and latanoprost-treated patients would achieve
WALT AND LEE
clinically relevant low target pressures according to the distribution of targets achieved at the month-3 visit in the study published by Gandolfi et al.4 If the patient has achieved the predetermined target pressure on bimatoprost or latanoprost monotherapy, the model assumes that the patient will continue to achieve the target pressure consistently, will remain on the assigned monotherapy without requiring additional glaucoma medications, and will be scheduled for the next physician visit at month 9. With this scenario, the costs include one medication for 12 months, a long physician visit at baseline, and short visits at 1, 3, and 9 months (Fig. 2). If the patient does not achieve the pre-established target IOP at month 3 on bimatoprost or latanoprost monotherapy, the physician will prescribe adjunctive medication. The patient will be seen 1, 3, and 9 months after initiating dual therapy (i.e., 4, 6, and 12 months after beginning the initial treatment). All patients with an inadequate response at month 3 will achieve target IOPs with dual therapy within 6 months. With this scenario, the costs include the initial medication for 12 months, the adjunctive medication for 9 months, and the baseline, 1-, 3-, 4-, 6-, and 12-month office visits (Fig. 2). Estimated Yearly Costs Each comprehensive, new patient visit costs US$119 and each established patient visit costs US$59, in accordance with the 2003 Physicians Fee & Coding Guide,12 a standardized reference used in U.S. pharmacoeconomic studies (Table 1). The monthly cost of adjunctive medication (for patients who do not achieve target IOP on monotherapy) varies depending on the treatment comparison. The model assumes that 50% of patients who need adjunctive medication receive generic timolol, 30% receive brimonidine, 10% receive the fixed combination of timolol/dorzolamide, and 10% receive dorzolamide (source: Scott-Levin marketing data, 2003). The model calculates a weighted average of the four possible adjunctive medications in the costeffectiveness calculations. Average monthly medication costs are US$55.79 for bimatoprost, US$55.60 for latanoprost, US$36.00 for brimonidine, US$5.00 for timolol, US$47.00 for timolol/dorzolamide, and US$27.00 for dorzolamide.13 Based on the distribution described above, the weighted average cost of an adjunctive medication is US$20.57 per month (Table 1). This average is used in all costeffectiveness calculations. The estimated yearly costs to achieve a specific IOP are calculated from the treatment model (Fig. 2): 12 months of monotherapy (upper path) or 3 months of monotherapy followed by 9 months of dual therapy
MEDICAL MANAGEMENT OF GLAUCOMA
S39 Fig. 2. Treatment algorithm. All patients are seen during a comprehensive new patient visit at baseline, and at a short visit at months 1 and 3. The monotherapy arm includes a short visit at month 9. The dual therapy arm includes short visits at months 4, 6, and 12.
(lower path). For purposes of this calculation, the success rate is defined as the percentage of patients who achieve a specific target IOP. Yearly drug costs were calculated for all target pressures from 13–20 mm Hg. If the target IOP is 17 mm Hg, for example, the formula to calculate yearly drug costs is (Success Rate to Achieve IOP of 17 mm Hg × Drug Cost of Upper Path) ⫹([1 ⫺ Success Rate] × Drug Cost of Lower Path) ⫽Cost ($)/patient/year to reach/maintain a specific IOP For this calculation, the percentage of patients who reached specific target IOPs was derived from a 3month comparison of bimatoprost with latanoprost (Fig. 3).4 A similarly structured formula is used to calculate yearly costs for medical visits, and the drug and visit costs are added to calculate the total yearly cost to reach that target IOP. The cost-effectiveness (cost per treatment success) was calculated by dividing the estimated yearly total costs for each drug by the treatment success rate for each target IOP. For this calculation, the treatment success rates were derived from the distribution of patients achieving specific IOPs seen in a 6-month comparison of bimatoprost and latanoprost (Fig. 4),10 because this study had the longest available follow-up period for a direct, head-to-head comparison of these two drugs. The same efficacy values were used for the 12-month calculations, supported by evidence that bimatoprost or latanoprost efficacy did not vary substantially from 6 to 12 months of followup when compared to timolol.2,5 A trial that directly compared bimatoprost, latanoprost, and travoprost was examined as a potential data source, but the
range of individual target pressures needed to construct the decision tree was not reported.11 To calculate the weighted target cost-effectiveness, the costs of the two treatment paths for each target IOP were weighted using the distribution of commonly established target IOPs derived from a 361-patient, nationwide, internet-based research survey (Walt et al, Cost-effectiveness of bimatoprost versus latanoprost plus adjunctive products for glaucoma treatment, ARVO, 2003). Participating physicians were from 32 centers throughout the United
Fig. 3. Patients who achieved target pressures at 12 pm, month 3. More patients achieved each target IOP with bimatoprost than with latanoprost. The differences between the treatment groups were statistically significant for target pressures of 13, 14, and 15 mm Hg (P ⱕ .049). Reprinted from Gandolfi et al4 with permission of Advanced Therapy.).
S40
Surv Ophthalmol 49 (Suppl 1) March 2004
WALT AND LEE
The yearly costs to reach each individual IOP were averaged over all patients to calculate the weighted target cost-effectiveness. The incremental cost-effectiveness ratio (ICER) for each target IOP was calculated using the following formula: (Total Yearly Costs of Total Yearly Costs ⫺ Bimatoprost of Latanoprost) (Success Rate of Success Rate Bimatoprost ⫺ of Latanoprost)
Results Fig. 4. Patients who achieved target pressures at 12 pm, month 6. A greater percentage of bimatoprost-treated patients achieved target pressures between 13 mm Hg and 20 mm Hg than did latanoprost-treated patients. For example, 36% of bimatoprost patients achieved pressures of 15 mm Hg or lower (at 12 pm) compared with 22% of latanoprost patients. The differences between the treatment groups were statistically significant for target pressures of 13–17, 19, and 20 mm Hg (p ⱕ 0.044). (Reprinted from Noecker et al10 with permission of American Journal of Ophthalmology.).
States, and were selected from a list of ophthalmologists who were top prescribers of glaucoma medications. Physicians were asked what patients they wanted to reach specific target IOPs. The responses of all participants (Fig. 5) were summarized to develop a weighting factor for each level of target IOP.
AVERAGE EXPECTED ANNUAL COSTS
Sample cost-effectiveness calculations for target IOPs of ⱕ17 mm Hg are highlighted, because the Advanced Glaucoma Intervention Study1 demonstrated that patients whose pressures were consistently below 18 mm Hg did not progress over a 6-year period. The average expected annual costs for an individual bimatoprost-treated patient to achieve a target pressure, would be US$669 for 12 months of medication and US$296 for one baseline and three short medical visits (Table 2). For a patient who failed to achieve the target pressure, the costs for an individual patient would be US$855 for medications (bimatoprost during months 1 through 12 plus an adjunctive medication for months 3 through 12) and US$414 for medical visits. The costs for an individual latanoprost-treated patient would be essentially the same (Table 2). The weighted average annual costs to treat a population of glaucoma patients takes into account the percentage of patients who achieve a given target weighted against the percentage of patients who fail to achieve that target. For example, to treat a population to a target pressure of 15 mm Hg or lower takes into account that 29% of bimatoprost patients achieve such a target after 3 months of treatment, while only 14% of latanoprost patients do.4 Thus, TABLE 2
Individual Annual Costs Fig. 5. Commonly established target IOPs derived from a 361-patient, nationwide, internet-based naturalistic effectiveness trial. Participating physicians were randomly selected from a population of glaucoma specialists from 32 centers throughout the United States. Physicians were asked what percentage of patients they wanted to reach specific target IOPs. Glaucoma specialists establish target pressures between 15 and 17 mm Hg for half of their patients. Other targets are chosen less frequently. (Source: Walt, personal communication).
Drug, Medical Visits, Total, US$ US$ US$ Bimatoprost Drug success @ 3 months Drug added @ 3 months Latanoprost Drug success @ 3 months Drug added @ 3 months
669 855
296 414
965 1,269
667 852
296 414
963 1,266
Calculations are based on costs to treat an individual patient.
S41
MEDICAL MANAGEMENT OF GLAUCOMA TABLE 3
Weighted Average Expected Annual Costs Total Costs, Percentage Product, US$ Patients, % US$ Bimatoprost Drug success @ 3 months 965 Drug added @ 3 months 1269 Total Average Cost Latanoprost Drug success @ 3 months 963 Drug added @ 3 months 1266 Total Average Cost
29 71
280 901 1181
14 86
135 1089 1224
Calculations are based on costs to treat a population of patients.
the weighted average expected annual costs for a bimatoprost-treated patient who achieved a target IOP of 15 mm Hg would be 29% times the US$965 cost to achieve that target plus 71% times the US$1,269 cost of failure to achieve that target (Table 3). Similarly, for latanoprost-treated patients, the weighted average expected annual costs would be 14% times the US$963 cost of success plus 86% times the US$1,266 cost of treatment failure (Table 3). COST-EFFECTIVENESS
For all target pressures in both scenarios (monotherapy with bimatoprost or with latanoprost for 12 months, or monotherapy with bimatoprost or with latanoprost for 3 months and dual therapy for 9 months), bimatoprost cost less per treatment success
than latanoprost (i.e., bimatoprost was more costeffective than latanoprost). The difference in costeffectiveness was most notable at the lowest target pressures (Fig. 6). At a target IOP of 13 mm Hg, for example, cost-effectiveness (cost per treatment success) was US$8,845 with bimatoprost and US$20,954 with latanoprost (Table 4). This contrasts with more similar cost-effectiveness between the two drugs at a target IOP of 20 mm Hg; for example, cost-effectiveness was US$1,104 with bimatoprost and US$1,288 with latanoprost (Table 4). The difference in cost-effectiveness between bimatoprost and latanoprost for a target pressure of ⱕ20 mm Hg is much smaller than with the lower targets, but it could represent substantial differences overall to an MCO that serves many glaucoma patients. Overall, at every target pressure, in this model bimatoprost costs approximately the same as and is more effective than latanoprost; in pharmacoeconomic terms, bimatoprost dominates latanoprost (Table 5). The cost-effectiveness findings were sensitive to changes in the price of products and to changes in treatment success rates. For example, the sensitivity analysis indicated that the treatment success rate for latanoprost would have to be at least 15–133% higher than the observed rate for it to be as costeffective as bimatoprost. Given that the AWP of latanoprost is only US$0.19 less than for bimatoprost, the cost-effectiveness of each drug would be closer in value only if the two drugs were equally efficacious.
Fig. 6. Comparative projected 1-year cost per treatment success.
694 701 311 317 1,005 1,018 1,104 1,288 (1005/91%) (1018/79%) Bimatoprost cost/success is 14% less than latanoprost
TABLE 5
Incremental Cost-effectiveness Ratio (ICER) Bimatoprost Relative to Latanoprost Target IOP
ICER, US$
747 760 346 355 1,093 1,115 1,681 2,477 (1093/65%) (1115/45%) Bimatoprost cost/success is 32% less than latanoprost
ⱕ13 ⱕ14 ⱕ15 ⱕ16 ⱕ17 ⱕ18 ⱕ19 ⱕ20 Weighted Average
⫺237 ⫺311 ⫺309 ⫺216 ⫺110 ⫺68 ⫺117 ⫺107 ⫺188
Discussion
801 826 380 397 1,181 1,224 3,280 5,563 (1181/36%) (1224/22%) Bimatoprost cost/success is 41% less than latanoprost Includes success and failure. a
36 6
Effectiveness in reaching low pressure, %10 Average expected cost/patient/yeara, US$ Drug Medical Total expected cost Cost effectiveness, US$ (Cost/treatment success)
14
WALT AND LEE
836 847 402 410 1,238 1,257 8,845 20,954 (1238/14%) (1257/6%) Bimatoprost cost/success is 58% less than latanoprost
91 45 65 22
Bimatoprost Bimatoprost
ⱕ13
Latanoprost
Bimatoprost
ⱕ15
Latanoprost
Target Pressure (mm Hg)
Comparative Cost-effectiveness for Selected Target Pressures
TABLE 4
ⱕ17
Latanoprost
Bimatoprost
ⱕ20
79
Surv Ophthalmol 49 (Suppl 1) March 2004 Latanoprost
S42
The most important result of the present analysis was that in this model bimatoprost is more cost-effective than latanoprost over a range of clinically relevant target pressures. Several recent studies have reinforced the evidence that the patients with the lowest IOPs experience the least deterioration in visual field over time, whether the initial diagnosis is OHT or glaucoma. In the Early Manifest Glaucoma Trial (EMGT),8 62% (78/126 ) of the untreated controls experienced deterioration of visual field, compared with 45% (58/ 129) of the treatment group. The Ocular Hypertension Treatment Study6 followed 1,636 participants randomized to observation or topical glaucoma medication. The trial tested the hypothesis that medical treatment of patients with OHT would lead to a lower incidence of primary open angle glaucoma (POAG). A target IOP of 24 mm Hg (minimum reduction of 20% from baseline) was established for patients who were randomized to treatment. After 60 months of observation, the treated group achieved a 22.5% reduction in IOP, compared with a 4.0% reduction in the untreated group. After 60 months of observation, the cumulative probability of developing POAG was 4.4% in the treated group, compared with 9.5% in the observation group. A recent study has demonstrated that the annual direct costs for glaucoma resources, including physician visits, glaucoma surgeries, and medication use, increased as the severity of glaucoma increased (Lee et al, A multi-center, retrospective study of resource utilization associated with severity of disease in glaucoma, AAO, 2002). Using a six-stage glaucoma staging system, the average direct cost of glaucoma treatment for patients with early stage disease was US$523/patient/year. The average cost for patients with end-stage disease was four times higher
S43
MEDICAL MANAGEMENT OF GLAUCOMA
(US$2,200/patient/year). By choosing the medication that is the most cost-effective method of achieving low target pressures, the physician can both increase the potential for saving the patient’s vision and decrease the lifetime costs of treatment. The results of the present analyses suggest that the most medically effective option for glaucoma patients that need to be maintained at very low IOPs can also be the most cost-effective option when measured in terms of cost per treatment success. For example, at a target pressure of ⱕ13 mm Hg, bimatoprost costs approximately 58% less per treatment success than does latanoprost. In this model, data from two different clinical trials were used to estimate what percentages of patients would reach clinically relevant target pressures. In the 3- and 6-month direct comparisons of bimatoprost and latanoprost from which the success rates were derived,4,10 a larger percentage of patients on bimatoprost reached low target pressures. In addition, an unpublished study was used to provide a weighting factor for the calculation of the weighted target cost-effectiveness (Walt et al, ISPOR, 2003). Pharmacoeconomic studies are considered to be especially rigorous if they derive treatment success rate estimates from more than one study, especially when results from the two studies are consistent with each other.7,14 One limitation of this cost-effectiveness analysis is that this cost-effectiveness algorithm is predicated on a 1-year treatment period, whereas glaucoma is a chronic disease requiring long-term treatment. This may limit the generalizability of results. This restriction, however, was necessary because trials about bimatoprost or latanoprost that extend to longer than12 months do not include the range of individual target pressures on which to base additional treatment-success decision points. Another reason for choosing a 12-month model is that the use of a longer interval would involve including costs for potential surgical interventions. The assumption that medication alone will be tried for a year may be violated at the lowest target pressures (12–14 mm Hg) because advanced glaucoma patients who are assigned to these targets are likely to be treated with medication for several months and then be scheduled for surgery. Modeling surgical costs would add costs to the failure arms of the decision tree, and would complicate both treatment arms proportionately. Therefore, surgical costs were not considered. The latter factor, however, may not be of prime importance, because there is evidence that the frequency of surgical interventions for glaucoma are dropping in the United States, perhaps as a result of the availability of functional analogs of PGF2α
that are highly effective in decreasing IOP.16 In addition, the cost-effectiveness measures based on this algorithm stay stable only so long as the AWPs of bimatoprost and latanoprost remain fixed. Another limitation of this study is that the cost of potential side effects was not factored into the model, although a larger percentage of bimatoprost patients than of latanoprost patients develop such adverse reactions as hyperemia, pruritus, or allergy.10,11 Patients who develop side effects would require more physician visits than patients who don’t. However, it is hard to assign costs to adverse events because patients and physicians outside the context of a controlled clinical trial behave more variably than they do under the trial environment where all physician visits and drugs are paid for.
Conclusions Evidence from recent clinical trials supports the need to lower IOP in glaucoma patients as low as is safely possible, and ocular hypotensive lipids are the newest chemical entities that can help attain that goal. Based on retrospective efficacy data from two clinical studies that directly compared bimatoprost and latanoprost, and on a simple pharmacoeconomic economic model decision-tree model, bimatoprost is a cost-effective alternative to latanoprost in treating patients with glaucoma.
Method of Literature Search The National Library of Medicine database was searched using the PubMed engine (www.ncbi.nlm. nih.gov/entrez/query.fcgi). All available years (1966–2003) were included. Foreign language articles were ignored. Keywords used in the search: cost and cost analysis; glaucoma/economics/therapy; health care costs; health resources/utilization; health services research; insurance, health reimbursement/economics; intraocular pressure; econometric models; ocular hypertension.
References 1. The AGIS Investigators: The advanced glaucoma intervention study (AGIS) 7. The relationship between control of intraocular pressure and visual field deterioration. Am J Ophthalmol 130:429–40, 2000 2. Camras CB, Alm A, Watson P, Stjernschantz J: Latanoprost, a prostaglandin analog, for glaucoma therapy. Efficacy and safety after 1 year of treatment in 198 patients. Latanoprost Study Groups. Ophthalmology 103:1916–24, 1996 3. DuBiner H, Cooke D, Dirks M: Efficacy and safety of bimatoprost in patients with elevated intraocular pressure: a 30day comparison with latanoprost. Surv Ophthalmol 45(Suppl 4):S353–60, 2001 4. Gandolfi S, Simmons ST, Sturm R: Three-month comparison of bimatoprost and latanoprost in patients with glaucoma and ocular hypertension. Adv Ther 18:110–21, 2001
S44 5.
6.
7. 8. 9. 10.
11.
Surv Ophthalmol 49 (Suppl 1) March 2004 Higginbotham EJ, Schuman JS, Goldberg I: One-year, randomized study comparing bimatoprost and timolol in glaucoma and ocular hypertension. Arch Ophthalmol 120: 1286–93, 2002 Kass MA, Heuer DK, Higginbotham EJ: The Ocular Hypertension Treatment Study: a randomized trial determines that topical ocular hypotensive medication delays or prevents the onset of primary open-angle glaucoma. Arch Ophthalmol 120:701–713; discussion 829–30, 2002 Kobelt G: Health Economics: An Introduction to Economic Evaluation. London: Office of Health Economics, 2002 Leske MC, Heijl A, Hussein M: Factors for glaucoma progression and the effect of treatment: the early manifest glaucoma trial. Arch Ophthalmol 121:48–56, 2003 Mick AB, Gonzalez S, Dunbar MT, McSoley JJ: A Cost analysis of the prostaglandin analogs. Optometry 73:614–9, 2002 Noecker RS, Dirks MS, Choplin NT: A six-month randomized clinical trial comparing the intraocular pressure-lowering efficacy of bimatoprost and latanoprost in patients with ocular hypertension or glaucoma. Am J Ophthalmol 135:55–63, 2003 Parrish RK, Palmberg P, Sheu WP: A comparison of latanoprost, bimatoprost, and travoprost in patients with elevated
WALT AND LEE
12. 13. 14. 15. 16.
intraocular pressure: a 12-week, randomized, masked-evaluator multicenter study. Am J Ophthalmol 135:688–703, 2003 Physicians Fee & Coding Guide. Augusta,GA: MAG Mutual HealthCare Consultants, 2003 PriceAlert, The Official Guide to Average Wholesale Pricing. Indianapolis: Facts & Comparisons, 2003 Rittenhouse B: Uses of models in economic evaluations of medicines and other health technologies. London: Office of Health Economics, 1996 Serle J, Cantor L, Gross R: Best practice treatment algorithm for primary open-angle glaucoma: implications for U.S. ophthalmology practice. Manag Care Interface 15:37–48, 2002 Strutton DR, Walt JG: Trends in glaucoma surgery before and after the introduction of new topical glaucoma pharmacotherapies. J Glaucoma (in press)
These data were presented at the Annual Meeting of Managed Care Pharmacy; October 17–20, 2001; Dallas, TX. John Walt and Jeffrey Lee are employees of Allergan, Inc. Reprint address: John Walt, MBA, Global Pharmacoeconomic Strategy and Research, Allergan Inc, 2525 Dupont Drive, Irvine, CA, 92612-1599.