In Portugal, the eXpeRience registry recruited a total of 62 patients with uncontrolled persistent allergic asthma from 20 participating centers. Response to omalizumab treatment, as measured by the physician’s Global Evaluation of Treatment Effectiveness (GETE) is available at approximately 16 weeks after initiation of omalizumab. Further data on patients’ clinical symptoms, asthma control, lung function, exacerbations, and healthcare utilization are available retrospectively for up to 12 months before and prospectively at 16 weeks and possibly 8, 12, 18, and 24 months after omalizumab initiation.17–21 Information on omalizumab and concomitant medication use is available only prospectively in the registry. We performed a pragmatic cost-effectiveness analysis based on data from the Portuguese registry participants, comparing the costs and effectiveness of two treatment alternatives for the management of severe allergic asthma: omalizumab plus standard of care and standard of care alone. A scheme representative of the analysis is presented in Fig. 1.

Figure 1.

Scheme of pragmatic cost-effectiveness analysis of omalizumanb.

(0.26MB).

For the omalizumab plus standard of care treatment arm, the total number of exacerbations, clinical episodes (emergency room visits, hospitalizations, and unscheduled doctor visits), and days of absence (work and/or school) observed during the prospective phase of the study, as well as the matching total amount of concomitant medication (standard of care), were available. The patient-year was calculated by the sum of observation time of each patient. GETE response was considered as observed after approximately 16 weeks of omalizumab initiation, considering a five-point scale: 1-excellent, 2-good, 3-moderate, 4-poor, and 5-worsening.18

For the standard of care treatment arm, to provide a term of comparison, corresponding retrospective data for each individual participant from the 12-month period before initiation of omalizumab was considered. During this period, patients were treated solely with standard of care. Retrospective data was extrapolated to a period equivalent to each participant’s prospective period of the study.

As an example, consider a patient with 5 asthma exacerbations in the pre-omalizumab treatment period (12-months) and a 24-month omalizumab treatment period. In this study for the standard of care treatment arm, a period of 24 months with 10 exacerbations was considered (i.e., the estimated incidence rate in the pre-treatment period was 5 exacerbations/year and applying this rate to a period of time equivalent to the time period of the omalizumab arm, 24 months, gives 10 exacerbations). In a similar way to this example, the incidence rates, based on pre-treatment data, of the other indicators and applied to a period equivalent to the period of treatment with omalizumab were estimated by patient, for the standard of care treatment arm.

Since basal GETE does not exist, it was assumed that without any further change in treatment (e.g. continuing standard of care) the patient´s health would worsen, so counterfactual GETE response on the standard of care at approximately 16 weeks was rated to be ‘5-worsening’ for all participants.

In the absence of retrospective information on standard of care medication, concomitant medication recorded between baseline and the 16-week visit was conservatively used as a proxy.

Missing prospective data up to the final follow-up of each patient were imputed using a last observation carried forward (LOCF) approach, where deemed suitable. No data imputation was performed for retrospective missing data.

The number of exacerbations (severe and non-severe), the number of clinical episodes, the number of days absent from work and/or school, and GETE response to therapy were considered as effectiveness outcomes. As cost indicators, both direct and indirect costs were considered, costs are presented in euros, updated to 2017 values.23 Direct costs relate to the cost of omalizumab, standard of care, and clinical episodes (emergency room visits, hospitalizations, and unscheduled doctor visits). Indirect costs relate to the societal cost of work absenteeism.

Unit costs for clinical episodes and drugs (Table 1) were taken from official sources within the Portuguese Health Authority.24–26 The cost per hospitalization was determined as a weighted average of the costs of All Patient Refined Diagnosis Related Group numbers APR-DRG 141 (asthma, severity grades: 3 and 4) and APR-DRG 130 (respiratory system diagnosis with ventilator support 96+ hrs, severity grades: 1–4),25 assuming that 7% of asthma related hospitalizations require mechanical ventilation.27 Costs of emergency room visits and unscheduled doctor visits were obtained from the analytical accounting database for hospitals of the Portuguese National Health Service.24 Drug costs were taken from the official drug retail price list.26

The indirect cost of a day of work absenteeism was based on the average monthly gross earnings of employees in Portugal 1097є in 201528), augmented by the employer contribution to Social Security (23.75%29) and the estimated cost to a firm of missed work (33%30).

Final results are presented in terms of incremental cost-effectiveness ratios (ICER), representing the additional cost of obtaining one additional unit of effectiveness. Since no imputation of missing data was performed, the ICER related to each effectiveness outcome was calculated based only on participants for whom the necessary retrospective and prospective data was available to calculate the impact of omalizumab on both effectiveness and costs. For the GETE response to therapy, costs were calculated only up to the corresponding visit.

A univariate sensitivity analysis was performed for GETE response and cost parameters (hospitalization costs, emergency room visit costs, unscheduled doctor visit costs, and work absenteeism cost). For univariate sensitivity analysis of GETE was considered a baseline GETE response rated with 3 – “moderate” and 4 – “poor” instead of “5-worsening” for all participants. The different costs of hospitalization, varying according to the different severity grades of asthma and respiratory system diagnosis with ventilator support 96+ hrs and for the remind costs: emergency room visit costs, unscheduled doctor visit costs, and work absenteeism cost the sensitivity analysis was performed changing the costs according to the 95% confidence interval limits.

The analysis was performed with Microsoft Excel (2016).

Retrospective and prospective information on exacerbations was available for a total of 58 patients (Table 3). Under omalizumab plus standard of care, 151 exacerbations occurred for a total follow-up of 103 patient-years, leading to a rate of 1.5 exacerbations per patient-year. Continuing the pre-omalizumab pattern of exacerbations, during the same total follow-up, would have led to an estimated 845 exacerbations under standard of care alone, corresponding to a rate of 8.2 exacerbations per patient-year. This implies an 82.1% reduction in the incidence of exacerbations for omalizumab plus standard of care as compared to standard of care alone. Similar results were obtained for the remaining event-related effectiveness outcomes (Table 3).

For GETE response to therapy, prospective data was available for 58 of the 62 patents (Fig. 2) after a median follow-up of 17 weeks. Considering the previously introduced five-point scale, average GETE response for omalizumab plus standard of care was estimated at 2.3, corresponding to a good-to-moderate response. With the counterfactual average GETE response on standard of care of ‘5-worsening’ for all participants, this implies a 54.1% reduction in GETE-score.

Figure 2.

Distribution of GETE response to omalizumab plus standard of care for the Portuguese patients in the eXpeRience registry.

(0.07MB).

For 54 Portuguese patients enrolled in the eXpeRience registry (Table 3), sufficient information on clinical episodes was available to estimate the impact of add-on therapy with omalizumab on direct costs. As can be seen from Table 4, over the considered time-horizon average 22 months, an estimated additional annual per-patient cost of 14,932є due to omalizumab therapy is expected to be off-set by a saving of 1844є in the remaining direct costs (85.3% of which due to avoided hospitalizations and emergency room visits). This leads to a net increase of 13,088є in annual per-person direct costs.

The impact on indirect costs due to the introduction of omalizumab could only be estimated on the basis of 22 patients for whom data on work absenteeism was available (Table 3). Given the 72.3% reduction in missed days of work, for these patients, a reduction in indirect costs of 818є per person-year was estimated (Table 5).

As described earlier, ICERs were estimated only for patients for whom both the effectiveness outcome and a full cost estimation was possible.

Of the 58 patients for whom effectiveness estimates are available in terms of the number of exacerbations (Table 3), 5 patients were not included in the direct cost estimate due to a lack of information about clinical episodes. For the remaining 53 patients, cost-effectiveness results are presented in Table 6. Over an average follow-up of 1.8 years, a total of 559 exacerbations were estimated to have been avoided with omalizumab plus standard of care over standard of care alone. This comes to an estimated increase in total direct costs of 1,253,490є, resulting in an ICER of 2244є per exacerbation avoided.

With the exception of GETE response to therapy, for the remaining effectiveness outcomes a similar analysis was performed (Table 6), leading to ICERs of 2592є per clinical episode avoided (54 patients), 1058є per day of work absenteeism avoided (21 patients), and 1096є per day of school absenteeism avoided (5 patients). For work absenteeism, the only outcome for which indirect costs were considered, the increase in total costs of 475,273є is the result of an increase of 502,538є in total direct costs, off-set by a decrease in total indirect costs of 27,265є.

For GETE response to therapy, with costs calculated only up to the corresponding response evaluation visit, over an average follow-up of 19 weeks (0.4 years), treating patients with omalizumab plus standard of care as compared to standard of care alone is estimated to lead to an ICER of 1750є per unit decrease in GETE classification.

The univariate sensitivity analysis shows that considering the different costs of hospitalization, the ICER varies between 2208є and 2258є per exacerbation avoided. Changing the cost of work absenteeism (cost per absent day) according to the 95% confidence interval limits, the ICER per avoided day of work absenteeism varies between 1055є and 1062є (Table 7). Finally, considering a GETE response rated with 3 – “moderate” and 4 – “poor” instead of “5-worsening” for all participants for the counterfactual standard of care generates ICER values of 7104є and 2808є per unit decrease in GETE classification, respectively (Table 8).