Clinical, economic, and humanistic burden of asthma in Canada: a systematic review

Background Asthma, one of the most common chronic respiratory diseases, affects about 3 million Canadians. The objective of this study is to provide a comprehensive evaluation of the published literature that reports on the clinical, economic, and humanistic burden of asthma in Canada. Methods A search of the PubMed, EMBASE, and EMCare databases was conducted to identify original research published between 2000 and 2011 on the burden of asthma in Canada. Controlled vocabulary with “asthma” as the main search concept was used. Searches were limited to articles written in English, involving human subjects and restricted to Canada. Articles were selected for inclusion based on predefined criteria like appropriate study design, disease state, and outcome measures. Key data elements, including year and type of research, number of study subjects, characteristics of study population, outcomes evaluated, results, and overall conclusions of the study, were abstracted and tabulated. Results Thirty-three of the 570 articles identified by the clinical and economic burden literature searches and 14 of the 309 articles identified by the humanistic burden literature searches met the requirements for inclusion in this review. The included studies highlighted the significant clinical burden of asthma and show high rates of healthcare resource utilization among asthma patients (hospitalizations, ED, physician visits, and prescription medication use). The economic burden is also high, with direct costs ranging from an average annual cost of $366 to $647 per patient and a total annual population-level cost ranging from ~ $46 million in British Columbia to ~ $141 million in Ontario. Indirect costs due to time loss from work, productivity loss, and functional impairment increase the overall burden. Although there is limited research on the humanistic burden of asthma, studies show a high (31%-50%) prevalence of psychological distress and diminished QoL among asthma patients relative to subjects without asthma. Conclusions As new therapies for asthma become available, economic evaluations and assessment of clinical and humanistic burden will become increasingly important. This report provides a comprehensive resource for health technology assessment that will assist decision making on asthma treatment selection and management guidelines in Canada.


Background
Asthma, an inflammatory disorder of the airways [1], accounts for roughly 80% of cases of chronic respiratory disease in Canada [2]. It affects more than 3 million Canadians and roughly 235 million people worldwide [3,4]. According to Statistics Canada,8.5% of the population aged 12 and older has been diagnosed with asthma [5]. Its prevalence in this country has been increasing over the last 20 years [3]. Worldwide, asthma prevalence rates have been rising on average by 50% every decade [3]. Notably, asthma is the leading cause of hospital admissions in the overall Canadian population [3,6], the leading cause of absenteeism from school, and the third leading cause of work loss [3]. Each year, there are 146,000 emergency room visits due to asthma attacks in Canada [3]. Asthma is also a major cause of hospitalization [7] among the estimated 13% of Canadian children who suffer from the disease [8].
High prevalence in conjunction with significant asthmarelated morbidity leads to a heavy clinico-economic and humanistic burden of asthma in Canada [9,10]. Healthcare utilization and costs are even higher when management and control of the disease are suboptimal [11]. The direct and indirect costs associated with asthma are expected to rank among the highest for chronic diseases due to the significant healthcare utilization associated with the disease [9] and asthma's detrimental impact on physical, emotional, social, and professional lives of sufferers [12].
This systematic review is the first to consolidate and summarize the literature (from 2000-2011) encompassing not only the clinical and economic, but also the humanistic burden of asthma in Canada. It, thus, provides a holistic overview of the weight this disease poses to the healthcare system, patients and society. Specifically, this systematic literature review unveils the direct and indirect costs of asthma per patient, the key drivers of healthcare resource utilization, and the humanistic impact of asthma on patients' quality of life (QoL), which cannot be inferred from clinical measures [13]. This information, consolidated in a single review, can be of value to payers, policy makers and healthcare providers in making decisions pertaining to the management and treatment of asthma.

Methods
We conducted a search of the PubMed, EMBASE, and EMCare databases to identify original research (crosssectional, observational, or longitudinal studies on the burden-of-illness and cost-of-illness) published from 2000 to 2011 on the burden of asthma in Canada. Review articles, letters, editorials, commentaries, studies reporting summaries of meeting proceedings or conferences, abstracts or posters presented at scientific meetings, and studies assessing the efficacy or effectiveness of specific interventions were not included. The time frame was selected to reflect more recent developments in the treatment and management of asthma in Canada. Each search was conducted using controlled vocabulary and key words, with "asthma" as the main search concept. Search terms included "Canada," "cost of illness," "hospitalization," "utilization," "burden of illness," "quality of life," "sickness impact profile," and "healthcare cost." Appendix shows the detailed search strategies for each topic area. Searches were limited to articles published in English and studies involving humans. Studies were restricted to Canada. Titles and abstracts of articles identified were carefully screened in the initial review for relevance to the topic. At the second review, articles were selected for inclusion based on predefined acceptance criteria, which included relevant patient population (ie, adults/children diagnosed with asthma) and appropriate study design and outcome measures (patient-and population-level). Two independent reviewers determined whether studies met the inclusion criteria, and discrepancies between reviewer decisions were resolved in consensus.
Reasons for study exclusions were recorded. For articles that met predefined inclusion/exclusion criteria, the quality of the studies was assessed using methodological checklists provided in the NICE Guidelines Manual [14] and the STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) guidelines [15,16]. Key data elements were abstracted and tabulated in summary tables: year and type of study, number of study subjects, asthma definition, characteristics of study population, outcomes evaluated, results, and overall conclusions of the study.
Reported costs were inflated to 2011 Canadian dollars (CAD) using the Consumer Price Index from Statistics Canada [17]. Figure 1 depicts the step-by-step study selection process. The MEDLINE, EMBASE, and EMCare database searches yielded 320 citations, 230 citations, and 20 citations, respectively.

Results
In the first-level selection process (based on the information presented in the article abstracts) for the clinicoeconomic burden, 503 of the 570 citations were rejected: 174 reported inappropriate outcomes (i.e., outcomes that were not aligned with the outcomes of interest), 150 due to inappropriate disease state (eg, the studies focused on other chronic respiratory diseases or included only a small number of the subjects with asthma), and 91 due to inappropriate study design. Other reasons for rejection during the first-level selection process are shown in Figure 1. Of the 67 full-text articles retrieved for potential inclusion, 34 were excluded during the second-level selection process (28 due to inappropriate outcomes). Thus, 33 articles fulfilled all criteria and were included in the clinico-economic burden review ( Figure 1).
After duplicates were removed, 309 studies were identified by the humanistic burden literature searches from the 3 databases. Of these, 288 studies were excluded during the first-level selection for inappropriate disease state (n = 44), inappropriate outcome measure (n = 60), inappropriate study design (n = 96), jurisdiction (n = 9), inappropriate patient population (n = 14), treatment comparator (n = 26), because data could not be extracted in the required format (n = 38), or because they were duplicate studies (n = 1).Twenty-one studies were selected for potential inclusion in the review. During the secondlevel selection, full-text articles were reviewed and a further 7 were excluded for inappropriate outcome measure (n = 1), study design (n = 2) or jurisdiction (n = 4). Fourteen articles fulfilled all criteria and were included in the humanistic burden review (Figure 1). Table 1 depicts the quality assessment of the articles on clinical, economic, and humanistic burden using STROBE tools, and Table 2 summarizes quality assessment of the articles on clinical burden using the NICE RCT assessment tool.

Clinical burden studies Overview
Of the 33 studies meeting all criteria for inclusion, 23 contained clinical burden data only, 7 had information on both clinical and economic burden of asthma, and 3 had data on the economic burden of asthma only.
Of the 30 studies on clinical burden, 1 was a casecontrol, 22 were cohort, and 7 were cross-sectional studies. Characteristics of studies reporting on clinical burden are shown in Table 3.
Most studies clearly reported the study design (97%), setting (100%), participants (87%), and statistical methods employed (70%). However, less than half reported on potential sources of bias and confounding factors or how missing data was handled. Furthermore, less than half of the studies reported on how loss to followup was addressed in both the methods and results sections, or how sensitivity analyses were conducted. Main results for outcomes data were appropriately reported in 97 % of the clinical burden studies, and more than 90% met the STROBE criteria for appropriate quality discussion. Most (77%) gave the source of study funding and the roles of the funders. (Tables 1 and 2).
Studies employed a variety of definitions for asthma, including ICD codes, physician visits and/or hospitalizations for asthma (based on billing codes), asthma medication prescriptions filled, and patient self-report. We report the definitions used, but these definitions were not reconciled in this review. When asthma was defined by the presence of ICD codes, it was considered to be narrowly defined, whereas a broad asthma definition included visits for an asthma-related diagnosis and asthma-related hospitalizations among the discharge diagnoses. Table 4 provides an overview of hospitalization rates for adult and pediatric patients with asthma in Canada. Reported rates of hospitalization for asthma varied widely according to age, geographic region, gender, and asthma medication use. In a large cohort study spanning over 20 years, Suissa et al. [41] obtained data from the Saskatchewan Health databases on asthma patients from that province aged 5-44 between 1975 and 1991 and found that the overall rate of asthma hospitalization was 42 per 1000 asthma patients per year in patients with at least 1 year of follow-up. The rate was higher (48 per 1000) in patients receiving at least 3 anti-asthma medication prescriptions in any 1 year. During the variable follow-up period (up to 4 years), regular use of inhaled corticosteroids (ICS) was associated with a 31%       reduction in the rate of hospital admissions for asthma and a 39% reduction in the rate of readmissions for the cohort with more severe asthma who had been previously hospitalized for the condition during the 1-year baseline period. The study investigators concluded that their findings emphasize the importance of regular use of inhaled corticosteroids to avoid hospitalizations. In a retrospective cross-sectional study of asthma patients aged 5-54 years using health databases in Saskatchewan, Klomp et al. [36] found that, in 2002-03 and 2003-04, the hospitalization rate for asthma was 10.9 per 1000 patients per year.

Key findings on clinical burden Hospitalizations
Agha et al. [27], using data on hospital admissions from the Dischrage Abstract Database of the Canadian Institute for Health Information, reported 8,583 asthma hospitalizations among 255,284 pediatric patients (a rate of 33.6 in 1000 patients) born between 1993 and 2000 in Toronto.
A significantly lower rate was reported in Canada by Seung et al. [42], who cited figures reported by the Public Health Agency of Canada of 143 asthma-related hospitalizations per 100,000 adult and pediatric patients, or 1.43 in 1000, in 1998 (with an additional 3.7 per 1000, many of whom had underlying asthma, hospitalized for influenza/pneumonia).
Higher rates were reported for hospital admissions of patients who initially presented to the emergency department (ED). Lougheed et al. [28] reported that 6.9% of adults and 10.8% of children who presented to the ED with asthma were admitted to the hospital.
According to the results of a study based on interviews with parents, 25% of the pediatric study population (124 of 490 patients) had been hospitalized for asthma in the previous 12 months [21]. In a large study utilizing data from Quebec administrative databases, children aged 5 to 15 years with at least 1 exacerbation in the year prior to treatment initiation with ICS or leukotriene receptor antagonists (LTRA) had higher rates of hospitalizations than those with no exacerbation in the previous year (0.03 vs. 0.005 hospitalizations per patient per year in the ICS group and 0.06 vs. 0.003 per patient per year in the LTRA group) [43]. The proportion of prescribed days covered was significantly higher in the LTRA group than in the ICS group (52% vs. 34%) [43].
In a study of all Ontario babies born during the year 1994 who were diagnosed with asthma before their sixth birthday, there was a decreasing trend in hospitalization rates with age, from 86.7 per 1000 patients per year in the 0 to 2 years age group to 27.3 per 1000 patients for those aged 3 to 5 years and 10.9 per 1000 for those aged 6 to 9 years. These investigators also found that children with persistent asthma had more than one and a half times higher hospitalization rates compared with patients whose asthma was in remission (63 per 1000 patients vs. 39 per 1000 patients per year) [26].
In another Ontario-based study that examined asthma seasonality and hospitalizations by gender and age group over a 12-year period, results of spectral analysis revealed that hospitalization rates for children with asthma were highest in September and October each year across the 12-year period, with a 2 to 3-times higher rate of hospitalizations in boys (180 per 100,000) than in girls under the age of 9 years [20]. However, among children older than 9 years, female hospitalizations exceeded those of males [20].
The large variations in reported rates of hospitalizations may be due to variations in ED visit rates and/or hospital admission percentages [28]. Hospital admissions appear to follow a bimodal age distribution pattern, with the very young and the elderly more likely to be admitted [28]. Other factors that can drive up rates of hospitalization in particular regions or among specific populations are higher disease prevalence, greater disease severity, multiple comorbidities, and barriers to care associated with socioeconomic status [27].

ED visits
The number of asthma emergency visits varied by age, type of treatment, social status, and living area (urban/ non-urban). Table 5 summarizes ranges and mean numbers of annual ED visits for asthma, as reported in the included studies. According to several studies, both children and adults with asthma averaged less than 1 ED visit per patient per year [21,23,28]. ED visit rates were significantly higher in women than in men and, overall, the rate of ED visits increased with age [28].
In a study investigating the impact of appropriate use (according to the 1999 Canadian asthma consensus report and the National Heart, Lung and Blood Institute Guidelines for the Diagnosis and Management of Asthma) and compliance with asthma medications in adults, the rate of ED visits for asthma was twice as high for patients not using asthma medication appropriately (high-dose SABA plus low-dose ICS) than for those using it appropriately (lowdose SABA plus high-dose ICS) [38].
Rosychuk et al. [19] examined trends in asthmarelated ED visits by more than 45 Sin et al. [30,44] reported that elderly asthmatic patients using ICS post-discharge from hospital were 29% less likely to be readmitted to hospital for asthma and 39% less likely to experience all-cause mortality compared with those who did not receive ICS post-discharge over a 1-year follow-up period. When age, sex, comorbidity, and use of other antiasthma medications were controlled for, ICS use was associated with a 32% relative rate reduction for recurrent hospitalization or all-cause mortality (95% CI 23%-39%). Among patients who received at least 1 prescription for ICS within 1 year prior to the index hospitalization, the use of ICS 90 days postdischarge was associated with a 41% decrease in recurrent asthma-related hospitalizations or deaths compared with non-use of ICS (95% CI 32%-49%).
Sin et al. [32] also reported on ED visits in children born in Alberta between 1985 and 1988, stratified by SES, and found that very poor children were 23% more likely to have had an ED visit for asthma compared with children from non-poor families (RR 1.23; 95% CI 1.14 -1.33). Very poor children had a similar risk of having an asthma-related ED visit as poor children (RR 0.97; 95% CI 0.91 -1.04).

Physician visits
Studies that reported rates of asthma-related physician visits are summarized in Table 6. In a population-based study evaluating 14-year trends in Manitoba in utilization of physician resources for asthma, Dik et al. [29] found that, between the period 1985-1988 and 1994-1998, the greatest increases in prevalence and incidence of physician visits for asthma occurred in the youngest age groups, while in adults the prevalence and incidence changed little with time. However, the average rate of physician visits for asthma decreased from 1.66 visits per patient-year in 1985-1988 to 1.40 in 1989-1993, and further to 1.16 visits per patient-year in 1994-1998.
More former or current smokers than non-smokers visited their physician [35], as did patients with workrelated asthma vs. non-work-related asthma [25] and patients inappropriately using their asthma medication ICS=inhaled corticosteroid, LTRA=leukotriene receptor antagonist, NMU= non-major urban areas, NWRA=non-work-related asthma, SABA=short-acting β-agonist, WRA=work-related asthma. [38]. Among elderly patients, the rate of physician visits for asthma was not influenced by treatment with ICS [44].
Children in an Ontario-based study who were born in 1994 and diagnosed with asthma before age 6, and whose asthma persisted until age 11 (as determined by the presence of claims for physician and/or hospital visits between the ages of 6 and 11), had a higher rate of physician visits than those in remission (60 vs. 46.9 visits per 100 patients per year) [26].

Medication prescriptions
Lynd et al. [40] reported that 27% of patients receive oral corticosteroids, 17% use no ICS, 47% receive less than 4 ICS canisters per year, 29% use 5 to 12 canisters, and 8% use more than 12 ICS canisters per year. Based on available data, children received more prescriptions per patient per year than adults [11,21,38]. Patients with inappropriate use of asthma medications (ie, those who were non-adherent to guidelines recommended in the 1999 Canadian asthma consensus report and the National Heart, Lung and Blood Institute Guidelines for the Diagnosis and Management of Asthma) received more than double the number of prescriptions per patient per year (mean [SD] 7.5 [4.9]) compared with those who used asthma medication appropriately (mean [SD] 3.3 [1.9]) [38].

Economic burden studies Overview
Ten studies evaluated the economic burden of asthma in Canada (5 cohort studies, 4 cross-sectional, and 1 economic analysis). Costs in the economic analysis were calculated for 1,350,871 persons, based on the 1998/1999 estimate that 57% of 2,389,085 persons aged ≥4 years had uncontrolled asthma.
More than 80% of these studies met the STROBE criteria for appropriate quality discussion. Most studies clearly reported the study design (90%), setting (100%), participants (90%), and statistical methods employed (70%). However, less than half reported on potential sources of bias and confounding factors or how missing data was handled and how loss to follow-up was addressed in both the methods and results sections or sensitivity analyses conducted. Most studies (60%) gave the source of funding and the role of the funders for their study (Table 1).
Asthma cases were identified using ICD codes or clearly stated diagnosis, retrospective physician visits, hospitalizations for asthma, and/or asthma medication prescriptions filled or patient self-report of asthma diagnosis or symptoms. There was available evidence on both the direct and indirect components of the economic burden of asthma in Canada. The overall burden varied based on whether studies reported costs from the perspective of an individual patient with asthma or costs at the population level. Few Canadian studies reported a cost per episode of acute asthma, and no studies reported the cost per patients overall. Five studies reported data on the direct costs of asthma at the patient-level. Three of these studies reported asthma costs per asthma patient [10,24,37], while 2 studies reported asthma costs per acute asthma episode [34,42]. Three studies reported population-level direct costs for asthma [10,24,42]. Study characteristics are presented in Table 7.

Key findings on economic burden
All costs reported in this section are in 2011 Canadian dollars.

Patient-level direct costs
Based on data from administrative databases in British Columbia, average total annual direct cost estimates in the general population ranged from $366.17 to $490.88 per asthma patient (Table 8) [10].
Ungar and colleagues [33] estimated the total cost of asthma in children aged 0-14 years in Ontario to be $883.48 per child from the healthcare perspective (Table 8) During the period from 1996 -2000, average hospitalization costs ranged from $67.90 to $136.87 per patient per year in the general population in British Columbia (aged 5 to 55 years), depending on the definition used to categorize asthma-related hospitalizations [10]. The estimated average annual hospitalization cost for asthma in children was $682.21 per patient in Ontario [33].
Sadatsafavi et al. [10] reported that ED visits made by asthma patients in the general population could cost the healthcare system anywhere between $66.35 and $122.09 per visit, depending on the asthma definition used. The reported range of average costs for ED visit per acute asthma episode was $209.48 to $274.48 [10,33]. Ungar and colleagues [33] estimated the average annual cost for ED visits in children to be $15.68.
The average costs for physician visits per acute asthma episode were estimated to range from $31.72 [34] in an economic modeling study using prospectively collected resource utilization data (9/1/94 to 8/31/95) from hospital emergency department visitors to $31.91 in the economic analysis by Seung and Mittmann [42]. Average costs of $98.02 and $70.57 annually per pediatric patient were reported for family physician and specialist visits, respectively, in the prospective study by Ungar et al. [33]. Although the cost per respiratory specialist visit was higher than the cost per family physician visit ($105.40 vs. $51.40 for the first visit and $23.10 vs. $16.25 for an additional visit), nearly twice as many patients (271, or 80%) reported visiting a family physician, at an average annual use of 3.6, compared with a respiratory specialist (141, or 42% of patients), at an average annual use of 2.1 [33]. A study conducted in Ontario demonstrated that outpatient claim costs for persons with asthma exceeded those for persons without asthma by about $200 per person per year [2].
With regard to asthma medication prescriptions, the administrative database study from British Columbia estimated the average annual cost for asthma medication in the general population to be $231.92 per patient [10].  [10,33].

Population-level direct costs
The 1998-1999 healthcare cost for asthmatic children in Ontario ($120 million, or $227.1 -$640.3 per child per year, depending on age group) was considerably higher than the total asthma cost for the general population (all ages) of British Columbia during the period 1996 -2000 (~$41.8 million, or $331 per patient per year) (Table 9) [10,24]. Based on data from administrative healthcare databases (April 1996 through March 2000), the total annual population-level asthma cost estimates in the general population in British Columbia ranged from~$46.3 million to $62.0 million , depending on the definition of asthma used [10]. Between~8.5 million and~17.2 million of that was spent on asthma-related hospitalizations,~8.4 million to~15.5 million on physician/ED visits, and~15.4 million to~29.3 million on asthma medications [10]. Medication costs represented the bulk (63.9%) of the total cost, hospitalizations/ED visits comprised 17.8%, and physician visits accounted for 18.3% of the total cost.

Patient-level indirect costs
About 50% of children missed 1-3 days of school (47.6% in the group with low-severity asthma, 53.9% in those with moderate severity, and 50.6% in the severe asthma group), and 5.7% of the low severity, 5.3% of the moderate severity, and 9.1% of the severe asthma patients were absent for 7 or more days [47].
Malo et al. [46] evaluated a random sample of 8 to 10 accepted claims for occupational asthma per year from 1988 to 2002 in Quebec and found that the mean cost of compensation for loss of income (CLI) across the 15 years (not accounting for inflation) was $72,500 (median $40,700) and the mean cost of compensation for functional impairment (CFI) was $11,700 (median $7,600). Median CLI costs were significantly higher in men than women (69.9 vs 13.1), in workers

Population-level indirect costs
In a population of~1.5 million working-age individuals in Alberta with an asthma prevalence of 8.5%, the number of asthma-related productivity lost work days ranged from 441,728 to 533,363 in 1 year, at a cost of $78.1 to $94.4 million in lost productivity [48]. Ungar et al. [49] reported productivity loss days (PLD) without reporting actual indirect costs. They found that annual PLD varied from 12 in employed persons to 20 in students, 22 in homemakers, retirees and the unemployed, and 49 in disability pensioners. Annual PLDs increased with increasing disease severity.

Humanistic burden studies Overview
Fourteen articles reporting results from 13 studies were retained for inclusion out of the 309 studies identified by the humanistic burden literature search. Two were cohort studies, 8 were cross-sectional, and 4 studies were RCTs. Only 1 of these was a pediatric study [50], which assessed the impact of asthma medication on children using the 3-domain Pediatric Asthma Quality of Life Questionnaire (PAQLQ). No studies were identified that reported utilities or QoL from a caregiver perspective. QoL assessments focused on subgroups of the asthma population, and studies had small numbers of participants. A variety of definitions were used to define asthma including clinical diagnosis, presence of symptoms, and positive inhalation tests. Characteristics of studies reporting on humanistic burden are detailed in Table 10.
Overall, most studies on humanistic burden met good reporting quality standards in accordance with STROBE criteria (Table 1). However, less than half of the studies reported how missing data and loss to follow-up was handled or sensitivity or other analyses performed. Most studies also met the STROBE criteria for appropriate quality discussion (80%) and reported information on study funding (80%).

Key findings on humanistic burden
Depression and anxiety were prevalent among asthma patients and were associated with worse asthma control and quality of life (QoL) [52]. Yacoub et al. reported a 50% prevalence of anxiety and/or depression among 40 subjects with occupational asthma [56]. In a study conducted by Lavoie et al., 31% of 504 adults with physician-diagnosed asthma met the diagnostic criteria for 1 or more psychiatric diagnoses [57]. A study specifically looking at occupational asthma also found that psychological distress and psychiatric disorders including depression, anxiety, and dysthymia were associated with impaired QoL [63].
As one would expect, QoL became progressively worse as disease severity increased [54,62]. Furthermore, QoL was lower in asthma patients who had at least 1 other chronic disease compared to those who had no other chronic disease [54].
A study of 504 consecutive adults with physiciandiagnosed asthma reported that depressive and anxiety disorders were both independently associated with decreased health-related QoL (as measured by AQLQ scores), but only depressive disorders were independently associated with worse asthma control (as measured by ACQ scores) [57]. Interestingly, having both depressive and anxiety disorders did not increase the risk for worse asthma control or decreased QoL [57]. According to the study authors, this finding suggests that there is no incremental risk associated with having both a depressive disorder and an anxiety disorder on asthma control and QoL. The researchers also noted that the lack of an independent association between anxiety disorders and asthma control may be due to the fact that patients with anxiety disorder are more inclined to selfmonitor their symptoms, and are thus more likely than depressed patients to detect asthma symptoms and seek intervention.
Lavoie et al. [58] studied the association between clinical measures of asthma morbidity and body mass index (BMI), and found that patients with higher BMI scores had worse asthma control and poorer QoL (i.e., higher ACQ and lower AQLQ scores), independent of age, gender, and asthma severity. However, BMI was not associated with asthma severity.

Discussion
This review is the first to summarize the literature encompassing not only the clinical and economic burden of asthma, but also the humanistic burden of asthma in Canada. This systematic review confirms that the burden associated with asthma is substantial, and will undoubtedly become more pronounced as the  asthma prevalence increases in Canada. The asthma burden as it is known today can likely be decreased by the development and implementation of innovative treatment strategies in the management of this disease. A considerable body of literature was included in this systematic review (33 articles for the clinical and economic burden and 14 for the humanistic burden).The reviewed literature suggested that the healthcare resource utilization in asthma varied greatly in Canada by age group and type of treatment used. The substantial clinical burden was reflected by high rates of hospitalizations, ED and physician visits, and medication use. Lower rates of ED visits and hospitalizations, as well as reduced deaths, were observed among ICS users compared with non-users (except among the elderly), but these reductions were not as pronounced in patients who had experienced recent asthma exacerbations.
We collected evidence on both the direct and indirect components of the economic burden of asthma in Canada. The overall burden varied based on whether studies reported costs from the perspective of an individual patient with asthma or costs at the population level. Reported estimates for patient-level total direct costs, inflated to 2011 Canadian dollars, ranged from $99.33 per patient in a cohort of children aged 0-14 years in Ontario (May -October 1995) [33] to $646.95 per patient in a cohort of children aged 0-9 years, also in Ontario (1994/1995) [24]. Reported estimates for population-level total direct costs, inflated to 2011 Canadian dollars, ranged from $46,285,583 for patients aged 5-55 years in British Columbia (April 1996 -March 2000) [10] to $140,882,165 for patients aged 0-9 years in Ontario (1994-1998) [24]. Few Canadian studies reported a cost per episode of acute asthma.
Fourteen studies assessed the impact of asthma on the QoL of patients; however, only 4 reported on QoL of children with asthma, which represents a significant knowledge gap. For the most part, QoL assessments focused on subgroups of the asthma population and studies had small numbers of participants. Asthma was associated with depression and/or anxiety in several studies.
As noted above, these research studies vary considerably in terms of geographic region of study, characteristics of patient populations, study methodologies, and definitions of asthma used, which presents a significant challenge in drawing definitive conclusions from our study. Furthermore, unique findings reported in single studies have yet to be confirmed or refuted by subsequent research. Thus, in our review, results are presented as reported, but no consensus can be reached on the rates of resource utilization among asthmatic patients, asthma-related costs, or the degree of QoL impairment among individuals with asthma. Our study suggests that there is a significant knowledge gap in understanding the comprehensive burden of asthma across Canada.
Nevertheless, the high rates of healthcare resource utilization observed among patients with asthma during this review revealed only the tip of the iceberg. The economic burden is noteworthy, with direct costsparticularly those related to hospitalizations and physician/ED visitsrepresenting the highest proportion of asthmarelated costs. The indirect costs mainly due to time loss from work, productivity loss, functional impairment and caregiver time also add to this significant burden. Although there is a paucity of research on the humanistic burden of asthma in Canada, the few studies included in this review indicate that QoL is unquestionably diminished in asthmatic patients and that there is a high prevalence of psychological distress and psychiatric disorders among patients with asthma. Notable knowledge gaps on the humanistic burden of asthma are the lack of QoL assessments in children and caregivers, as well as quantifying the asthma-attributable burden in this patient population.
This systematic review provides a holistic overview of the burden of asthma in Canada, detailing the direct and indirect costs, the key drivers of healthcare resource utilization, and the impact of asthma on patients' quality of life -information that cannot be inferred from clinical measures. This information can be of value to payers, policy makers and healthcare providers in making decisions pertaining to the management and treatment of asthma.
For example, knowing that depression is often associated with asthma and that its severity and asthma control are intertwined, it might be useful to have psychologists/psychiatric healthcare professionals on the disease management team from the time of asthma diagnosis. Also, findings that BMI levels and asthma control and QoL are related, can lead to adding interventional measures to the treatment strategy.
As far as treatment options go, the use of inhaled corticosteroids was noted in many of the reviewed articles to be associated with lower rates of ED visits and hospitalizations; therefore recommending the appropriate use of medications (low-dose SABA plus high-dose ICS) should be emphasized.
More research in Canada is needed to add to the holistic picture of the impact of this disorder on the lives of patients, their families, and caregivers. Furthermore, much remains to be learned about the optimal use of the currently available treatments, how to combine them for maximal benefit, and how to incorporate new drugs in development into existing treatment regimens.

Limitations
All literature reviews are limited by the publication bias of the articles that are available. We acknowledge the fact that studies identifying a significant burden of asthma may likely be published than the ones reporting a low burden. The articles in this review are limited to the English language, and publication constraints were placed on articles identified by the search with studies limited to those published since 2000. Spatial restrictions were also applied, limiting studies to Canada. Studies employed a variety of defining criteria for asthma (from patient selfreport to ICD-codes, from physician-recorded diagnosis to discharge diagnosis combined with medication use), and these definitions were not reconciled in this review. This may have led to underreporting or overreporting of certain outcomes. Results were analyzed as reported, but direct comparisons between studies are lacking, due to the high heterogeneity of methodological approaches. In spite of these limitations, this review was systematic in nature and summarizes all available and relevant data published since 2000, thus providing a better understanding of the literature with respect to the clinical, economic, and humanistic burden of asthma.

Conclusions
The information contained within this study provides a comprehensive overview of the burden of asthma in Canada. Moreover, our study identifies several key knowledge gaps in understanding this area. As new therapies for asthma become available, health technology assessments will become increasingly important not only as it pertain to amendments to clinical practice guidelines but also with regard to formulating reimbursement decisions. Our study summarizes information that can prove important for physicians, healthcare authorities, and government officials involved in the treatment selection and development of disease management guidelines for asthma.