Table 1Summary of the Manufacturer’s Economic Submission
View in own window
| Drug Product | onabotulinumtoxinA (Botox) |
| Study Question | From the perspective of the public payer in Canada, what is the cost-utility of onabotulinumtoxinA as compared with best supportive care (BSC) for the prophylaxis of headache in adults who have failed three or more oral prophylactic medications?
The manufacturer also undertook a scenario analysis of onabotulinumtoxinA in the full Health Canada indication as per CDR guidelines. |
| Type of Economic Evaluation | Cost-utility analysis |
| Target Population | Adults with CM, ≥ 15 HDPM with headache lasting 4 hours a day or longer) who have failed (i.e., lack of efficacy, intolerance or clinical contraindication) 3 or more prior oral prophylactic medications
As indicated above, the manufacturer undertook a scenario analysis of the full population — CM without treatment failure. |
| Treatment | onabotulinumtoxinA |
| Outcome | QALY |
| Comparators | BSC (appears to be placebo + acute treatments) |
| Perspective | Public payer |
| Time Horizon | 3 years |
| Manufacturer’s Results (Base Case) | $28,940 per QALY (full indication population)
$25,470 per QALY (subpopulation) |
| Key Limitations and CDR Estimate(s) |
The key limitation of the manufacturer’s economic submission is whether the submitted model presents a good representation of the chronic nature of the disease and expected treatment. The modelling of CM is based on assumptions regarding patients with EM continuing treatment with onabotulinumtoxinA, lack of clarity around discontinuation of treatment, use of a “30% reduction in headache days” stopping rule, and the choice of a 3-year time horizon for a chronic condition. The costs associated with physician visits, drug administration, and drug acquisition for onabotulinumtoxinA were likely underestimated in the economic model, which bias the results in favour of onabotulinumtoxinA. Given the limitations identified, the likely ICUR is uncertain. When accounting for more likely cost inputs, CDR calculated ICURs in the range of $42,000 to $47,000 per QALY.
|
BSC = best supportive care; CDR = CADTH Common Drug Review; CM = chronic migraine; EM = episodic migraine; HDPM = headache days per month; ICUR = incremental cost-utility ratio; QALY = quality-adjusted life-year.
Background
OnabotulinumtoxinA (Botox) was submitted to CDR for the prophylaxis of headaches in adults with chronic migraine (CM, ≥ 15 days per month with headache lasting four hours a day or longer). The manufacturer has requested listing in a subpopulation of patients that have failed to respond to treatment with three or more oral prophylactic agents. OnabotulinumtoxinA is administered as a minimum of 31 injections and a maximum of 39 injections of five units (U) per injection to the head and neck every 12 weeks (total: 155 U to 195 U per administration). The submitted price is listed at $3.57 per U, with the pack sizes 50 U, 100 U, and 200 U. Given that the vials cannot be reused1 (therefore incurring wastage), the cost per administration is $714.
OnabotulinumtoxinA was previously reviewed by CDR in 2012 for the treatment of urinary incontinence due to neurogenic detrusor overactivity resulting from neurogenic bladder associated with multiple sclerosis (MS) or subcervical spinal cord injury (SCI). CDEC recommended that onabotulinumtoxinA be listed with conditions.2
Summary of Economic Analysis
The manufacturer submitted a cost-utility analysis (CUA) of onabotulinumtoxinA compared with best supportive care (BSC) for the prophylaxis of headache in adults who have failed three or more prior oral prophylactic medications; this represents a subpopulation of the Health Canada (HC) indication, as the HC indication does not limit use to patients who have failed prior therapy. The analysis was based on a Markov model with seven health states, six based on the number of headache days experienced per 28-day period (zero to three days, four to nine days, 10 to 14 days, 15 to 19 days, 20 to 23 days, and 24 days or more) and one discontinuation state. Efficacy data were derived from pooling data from the PREEMPT studies (PREEMPT-1 and PREEMPT-2). Patients entered the model at one of the three health states defined as CM (≥ 15 days headache days per month [HDPM]) and transitioned to the other health states based upon patient-level data from the PREEMPT studies. The manufacturer captured treatment costs associated with onabotulinumtoxinA and BSC, as well as the costs of medical resource utilization. The manufacturer included a scenario analysis from the societal perspective, taking into account lost productivity. Utility values for each health state were obtained through mapping quality-of-life data captured in the clinical trials to the EuroQol 5 dimensions (EQ-5D) utility instrument. Scenario analyses using survey-generated EQ-5D utility values were also undertaken. The time horizon for the analysis was set at three years, with a cycle length of 12 weeks.
Results of Manufacturer’s Analysis
The result of the manufacturer’s analysis of the full HC population was an ICUR of $28,940 per QALY gained for onabotulinumtoxinA compared with BSC. For the analysis of the requested subpopulation (patients who failed to respond to treatment with three or more oral prophylactic agents), the ICUR was $25,470 per QALY gained.
Interpretations and Key Limitations
The key limitation of the manufacturer’s submitted economic evaluation is whether the submitted economic evaluation presents a good representation of the chronic nature of the condition and expected treatment, where key assumptions include:
The model included health states that represent episodic migraine (EM, < 15 HDPM), a population for which onabotulinumtoxinA is not indicated. EM reflects a distinct clinical entity different from CM. Patients improving to episodic states continued to receive treatment, incurring the costs of treatment and the clinical benefits. The effects of including patients no longer in CM could overestimate the benefits of onabotulinumtoxinA.
The choice of a 30% stopping rule is arbitrary. Improvements of 25%, 50%, and 75% were captured in the clinical trials, as opposed to 30%. Treatment guidelines and CDR clinical guidance has indicated that a 50% reduction or return to the EM health state is the clinical goal of treatment (no statistically significant difference in PREEMPT-1).
CM, defined as ≥ 15 HDPM with headache lasting 4 hours a day or longer, is currently a long-term condition; however, a three-year time horizon was used in the economic model. The manufacturer stated that onabotulinumtoxinA is a preventive therapy, and continued treatment is needed for most patients in order to maintain treatment response and benefit, similar to any other preventive medication.
3 This leads to the assumption that a longer time horizon may be more appropriate, where patients are expected continue on onabotulinumtoxinA indefinitely. Long-term use of onabotulinumtoxinA has not been studied, so it is unclear whether treatment effects would be maintained over time.
The costs associated with physician visits, drug administration, and drug acquisition for onabotulinumtoxinA were likely underestimated in the economic model.
It is also unclear whether patients who “discontinued” treatment entered an absorbing state or were permitted to cycle back into the model. How recurrence of chronic states is included was not clearly detailed.
Results of CADTH Common Drug Review Analysis
The manufacturer made several assumptions regarding the structure of the model which did not permit modification for reanalyses that would provide a more likely estimate of the cost-effectiveness of onabotulinumtoxinA. Assumptions regarding the cost of physician visits, the cost of administration, and the cost of drug acquisition for onabotulinumtoxinA were considered — all of which increased the ICUR between 10% and 28% individually from the manufacturer’s base-case results for both the full and restricted populations, and between 63% and 65% when considered collectively ($42,000 to $47,000 per QALY).
Issues for Consideration
There is an unmet clinical need in patients who have failed prior oral prophylactic therapies, as no other treatments are currently indicated.
There is the potential for use beyond the current indication being reviewed, as onabotulinumtoxinA is approved for eight clinical indications for use.
Conclusions
Given the limitations of the model structure, a full exploration of CDR identified limitations was not possible. Consequently, there is some uncertainty regarding the cost-effectiveness of onabotulinumtoxinA.
