Discussion
The total socioeconomic costs of CF in Canada from four perspectives (healthcare system, individual/caregiver, variable (ie, medicines) and society) across three domains (direct, indirect, intangible) were estimated by triangulating data sources. These estimates provide a comprehensive evaluation of the total socioeconomic burden of CF in Canada prior to the widespread introduction of highly effective modulator therapies. A comprehensive summary of the total socioeconomic burden, as well as the direct and indirect costs incurred by individuals and their families, will help to inform and support advocacy efforts and programme development aimed at provincial, territorial and federal governments, as well as insurers and group benefit plans. From a policy perspective, the socioeconomic cost estimates will help inform economic evaluations of existing and novel therapeutics.
There have been at least 40 CF cost-of-illness studies published to date, including four systematic reviews.6 29–31 Direct comparisons with previous literature are challenging since the perspective (healthcare system, individual, society) and specific domains and items included in each of the studies varied considerably.6 Direct out-of-pocket costs of CF to individuals and caregivers have been infrequently reported. Since 1990, there have been five Canadian studies published; three reported direct healthcare costs, one included indirect caregiver costs and one included both direct and indirect costs. The total direct healthcare costs estimated (including medications) in the current study ($C47 330/person/year in 2021 Canadian dollars) are consistent with a recently published study which reported the total direct healthcare costs in the province of British Columbia, including inpatient, outpatient and medication costs (adjusted to $C46 000/person/year in 2021 Canadian dollars) (7). The indirect costs of CF estimated in the current study ($C29 694/person/year in 2021 Canadian dollars) were similar to a 2006 study by Guerriere et al (adjusted to $C30 392/person/year in 2021 Canadian dollars).32 The indirect costs in the Guerriere et al, study represented 72% of the cost of CF, whereas in the current study indirect costs represented 33% of the total cost. The difference likely reflects that Guerriere et al, only included ambulatory care as direct costs. Interestingly, the average number of sick days reported by the BoD survey respondents was similar to the general Canadian population (mean of 10.3 days33). There have been a limited number of studies investigating the socioeconomic burden of rare diseases in Canada. The identified studies were heterogeneous in terms of perspectives and domains making comparisons with CF difficult. For example, a study of the economic costs of 14 inherited retinal conditions in Canada (~20 000–80 000 people) found per person costs to range from $C78 197 to $C319 281 per year, which included the direct and indirect costs as well as the cost of loss of well-being measured using disability-adjusted life-years.34
Medications accounted for half of the direct CF costs, which is consistent with previous literature.35 In Canada, medically necessary healthcare services (eg, clinic visits and hospitalisations) are covered by the single payer (ie, government). The coverage of prescription medications varies by province and individual’s private drug plan coverage. For many people with CF in Canada, medications are paid for out of pocket, or with high co-pays and deductibles. To cover the costs of medicines, many Canadians had supplemental health benefits, either from an employer or a private provider that cover a proportion or all the non-essential healthcare costs (including medications). Some provinces also have programmes that help to cover the costs of medications specifically for CF. The high costs of medications likely contributed to the fact that 19% of respondents reported skipping a medication because they could not afford it. This finding is consistent with other recent Canadian surveys of the general population.36 37 Affordability of medicines will further be exacerbated for individuals with lifetime caps on private benefits insurance, which will be exceeded after a few years of CFTR modulator coverage.
Using CFQ-R responses from the BoD survey and applying an algorithm to estimate EQ-5D,25 we report on quality of life using an established patient-reported measure. The average EQ-5D was 0.74 which was lower than what is expected in health (a score of 1) but higher than in the study by Acaster et al, which reported an EQ-5D of 0.67.25 The differences may be due to the reporting of EQ-5D as a median in our study and mean by Acaster et al.25 Further, the quality-of-life score was relatively consistent across disease severity, suggesting that the BoD survey respondents may be healthier overall.
Strengths and limitations
The comprehensive list of CF-related expenses was codeveloped with the CF community after systematic review of existing CF cost of illness studies and literature for other rare diseases. Linkage of the BoD survey with the CCFR reduced participant burden and improved coverage of items that could be included. The BoD survey respondents (8% of all people with CF) were generally representative of the Canadian population in terms of socioeconomic status (eg, annual household income).38 The BoD survey respondents represented a slightly older, with more female respondents, and more severe disease group compared with the CCFR, which may have led to an overestimate of some items. To mitigate this, BoD survey responses were cross referenced with either CCFR data or vendor websites and where possible disease severity stratum costs were applied.
The average cost of each medication was applied to all people taking that medication, which was appropriate for many, but not all medicines. Stratified costs, especially by disease severity, often included a small number of observations for each medication. The list of possible expenses incurred by the healthcare system or individuals with CF and their caregivers may be missing important items, or items that are needed by a smaller subset of the population not captured in the survey. Further, we were not able to quantify a cost associated with presenteeism (lower work productivity due to disease or caregiving responsibilities). Finally, the CCFR data were taken from the year 2019 to represent a time before the widespread availability of elexacaftor-tezacaftor-ivacaftor and before the COVID-19 pandemic, whereas the survey was conducted in 2021, during a time when elexacaftor-tezacaftor-ivacaftor was available to some people through the compassionate use programme. The choice of different sources was purposeful and pragmatic to reduce recall bias for the survey. Recall bias may have randomly affected the reported costs. Nonetheless, we cannot exclude the possible influences of either early use of elexacaftor-tezacaftor-ivacaftor or the pandemic on our estimates. These limitations may explain why the EQ-5D scores were higher in the severe group compared with previous literature.
The costs of CFTR modulators (which can be in excess of $C250 000 per person per year) were excluded from the medication costs as our primary aim was to characterise the burden of CF prior to the availability of highly effective modulators. A growing number of studies have demonstrated substantive reductions in hospitalisations rates and medication costs after the introduction of elexacaftor-tezacaftor-ivacaftor use.39–41 At the time of the survey, 15% of people with CF were prescribed a CFTR modulators, and less than 1% were prescribed elexacaftor-tezacaftor-ivacaftor. The majority of this group was prescribed lumacaftor-ivacaftor or tezacaftor-ivacaftor which only showed modest improvements in outcomes and minimal changes to healthcare utilisations.42 This comprehensive summary of the pre-elexacaftor-tezacaftor-ivacaftor socioeconomic costs of CF will allow for direct comparisons across all domains in the future.
Indirect costs included the time individuals with CF or their caregivers spent on treatment, attending clinical visits and during hospitalisations. We applied the human capital approach to estimate indirect costs. There are other approaches (eg, friction cost method) and informal caregiver ‘spillover’ effects that were not considered and may have produced different estimates.43–46 Common informal care activities are reported (online supplemental table 4) but not included in the indirect caregiver costs. In Chevreul et al, informal caregiving costs determined using the proxy good method attributed between 17% and 77% of the total cost of CF across eight European countries.47 Wherein Angelis et al also used the proxy good method and attributed 45% of the total costs to informal caregiving.48 Therefore, the omission of informal care may have substantially underestimated the total costs of CF in Canada.47 49 Direct comparisons between the studies are challenging because of the different approaches used to define each domain and estimate costs.
Intangible costs were estimated using the CFQ-R questionnaire for people with CF over the age of 14 years. Caregivers of people with CF experience unique challenges, which may impact quality of life as well. The BoD survey did not include an assessment of the quality of life of caregivers, which will be important to include in future studies.
Conclusion
This socioeconomic burden study represents a community-oriented approach to estimating the total costs of living with CF in Canada, prior to the widespread availably of highly effective CFTR modulators. The framework applied serves as a benchmark for post-CFTR modulator studies.