Survey-based epidemiologic studies of asthma from around the world have shown similar trends over time. To et al. (2012) published World Health Survey (WHS) data from 2002–2003 surveys across 70 countries showing an estimated asthma prevalence of 4.5% of adults (18–45 years old) with wide variation depending on geographic area1. The CDC reported in 2015 that 7.6% of people over 18 years of age, 18.4 million, and 8.4% of children had a diagnosis of asthma, making it 8.2% of the entire US population2. The prevalence of asthma in the US increased from 2001 to 2010 to 8.4%, and then stabilized at 8.2% in 20152,3. Australia saw similar trends of increases in the prevalence of asthma in children between 1982 and 1992, with decreases or stabilization from 1992 to 200217. A Statistics Canada Health Survey showed that in 2014, 8.1% of Canadians aged 12 and older, 2.4 million people, had been diagnosed with asthma by a health professional in the community. Using telephone surveys of a sample of the population, Statistics Canada estimates that 274 661 people were living with asthma in Alberta in 2014, which translates to approximately 8.1% of the population that was 12 years old and older at that time4. Using health databases, we have identified that amongst Albertans of all ages, there were 240 777 (11.9%) females and 239 560 (11.4%) males living with asthma in 2014 (Table S1).
Although there are no recent publications on the prevalence of asthma in Alberta, there are several publications reporting on emergency care for asthma in Alberta that make use of databases. A study on the rates of asthma presentations in ED from 1999 to 2011 showed that the number of ED presentations has been declining during that period10. Rosychuk 2010 showed that during a similar period (1999–2005) there were higher rates of ED visits in non-urban areas of Alberta compared to Edmonton and Calgary, with crude rates in 2004/05 of 7.9/1000, 6.5/1000 and 15.4/1000 in the Edmonton, Calgary and non-urban areas, respectively. These trends were stable over the study period11. Smaller and older, survey-based studies tended to focus on specific populations. The prevalence of asthma in 2001 in children aged 5 to 19 years old in two cities in Alberta Red Deer and Medicine Hat showed the prevalence of asthma was higher in Medicine Hat (17.0%) than in Red Deer (12.8%)18. These estimates were based on surveys sent out to students at several schools in each city and were followed up by telephone surveys.
A great example of the use of databases to inform more accurate estimates of incidence and prevalence is seen in Ontario, where OASIS has reported on asthma epidemiology since 2003, with a variety of topics. Specifically, in 2010 Gershon et al. showed that from 1996 to 2005 the asthma prevalence in Ontario increased from 8.5% in 1996 to 13.3% in 2005. Asthma incidence rates increased in children by 30.0% and were relatively stable in adults as all-cause mortality decreased. Asthma prevalence in Ontario, had increased partly due to the increase in childhood asthma incidence9. The same group calculated the lifetime risk of physician-diagnosed asthma was 33.9% in Ontario in 201019 and forecasted that in 2022 there will be more than 71 000 new cases of asthma and more than 1.9 million individuals living with asthma20. More recently, the group reported the asthma prevalence in 2016 as 15.5% and incidence 0.25%. We show that Alberta has (Table 1, S2 and S4) similar trends as Ontario, with decreasing incidence, but a steady increase in prevalence. Fewer patients were being diagnosed with asthma yearly, but the prevalence continued to steadily increase, although at a slower rate. Our data closely resembles that presented by OASIS. We have shown that in Alberta, while asthma incidence has decreased from 1.5% in 1995 to 0.7% in 2015, prevalence increased by 0.25–2% each year. We have used similar methods and the same definition of asthma, therefore, comparison with the Ontario data provides important information on asthma in two separate jurisdictions in Canada.
Several studies around the world have estimated mortality rates in patients with asthma and have shown that mortality directly due to asthma is not reported frequently, but the risk of death from any cause in patients with asthma is higher than in those without. A recent European study used data from Netherlands, Italy, UK, Denmark and Spain from 2008 to 2013, and evaluated 586 436 adult asthma patients, estimating that the age and sex standardized all-cause mortality rate was 5.2 to 9.5/1000 person-years in asthma. Mortality rates were higher in the first month following a severe asthma exacerbation and decreased thereafter; thus, showing that asthma is an important factor leading to increased mortality21. Similarly, 54 320 adult subjects were examined in seven independent cross-sectional population surveys repeated every five years between 1982 and 2012 in Finland. All-cause mortality decreased between 1982 and 2015, though mortality in asthmatic subjects compared with non-asthmatics was higher from all causes22. Another study from Finland looked at a smaller number of patients over the age of 30. This was a combination of surveys and database use. This study is important as it had a control non-asthma group. Over 15.6 years, there were 221 deaths among 1 052 asthma patients and 335 deaths among 1 889 non-asthma patients. Asthma was associated with increased all-cause mortality (adjusted HR 1.25; 95% CI 1.05–1.49, P = 0.011)23.
Rates of death from asthma in the United States increased from 0.8/100 000 in 1977 to 2.1 in 1994 but decreased to 1.6/100 000 in 2000. Rates had been higher for women than men but have overall decreased for both24. The CDC published that asthma deaths further decreased from 15 per million in 2001 to 11 per million in 2015. Adults were more likely than children to die from asthma. The asthma death rate was highest among the 65 years and older age group compared with all other age groups2. In Ontario in 2008, the asthma population had higher all-cause mortality compared with the general population (rate ratio, 1.3) and total mortality in those with asthma was fourfold higher than mortality directly due to asthma25. All-cause mortality rates in asthma patients have decreased substantially over the past 20 years as data from Ontario and now Alberta have shown. Compared with the general population however, the asthma population has a higher all-cause mortality and is more likely to die from comorbid conditions. In Alberta, the asthma population had a mortality rate 22% higher than the non-asthma population in 2015, adding to the increasing evidence that an asthma diagnosis is associated with higher mortality.
Strengths and limitations of our study:
As research in the field progresses, multiple asthma phenotypes have been identified which add to the difficulty of the diagnosis. Because it is a heterogeneous, chronic disease and requires long-term follow-up for diagnosis, it is also difficult to study its epidemiology. Socioeconomic factors that may impair access to healthcare, the association and overlap with COPD, multiple phenotypes and difficulty in predicting disease courses are some of the barriers to the accurate diagnosis of asthma. Diagnostic code usage could vary among practitioners. An asthma presentation may be coded as bronchitis, dyspnea, respiratory distress, etc. Objective assessment with pulmonary function testing and spirometry are not used for every patient that is diagnosed with asthma as noted by the Ontario group and by the Respiratory Health – Strategic Clinical Network of Alberta (authors of this study are member of the reporting group)26. Physicians have identified difficulty in continuity of care as a problem in the diagnosis of asthma. Less than half of the patients with new physician-diagnosed asthma in Ontario received objective pulmonary function testing around the time of diagnosis27.
Given the heterogeneity in epidemiological studies regarding definitions, study methods, geographic areas, age groups and population subsections studied, it is not surprising that there have been conflicting reports of the incidence and prevalence of asthma28. Accuracy begins with the clinical asthma diagnosis, but according to a recent publication as many as 33% of asthma patients may be misdiagnosed as having asthma29. Misdiagnosis is likely multifactorial and includes lack of pulmonary function testing, as well as difficulty establishing continuity of care26,27. Furthermore, the definition of asthma in population studies has varied widely from survey questions of a prior diagnosis, current or past symptoms, one or several ED visits, and multiple health claims for asthma30. This leads to heterogeneous data that cannot be compared or used to estimate changes in incidence and prevalence over time.
Brogger et al. in 2004 showed that survey-based retrospective estimates of trends in asthma incidence are likely to be severely biased by differential recall5. In recent years, there has been increasing use of database research to improve the accuracy of estimates and to follow trends over time in a more efficacious and reliable way. For database type studies, OASIS has shown that the algorithm of two or more ambulatory care visits and/or one or more hospitalization(s) for asthma in 2 years allows for accurate statistical investigations in both adults and children7,8. A large, single, publicly funded healthcare system such as Alberta Health, facilitates database research. This improves the accuracy of the data we present over studies based on survey methods or using smaller populations.
Sources of error in the data collection include different access to healthcare for various groups including lower socio-economic, first nations, remote geographic areas, and different age groups. There are several other potential sources of error in our estimates. The data was collected starting in 1983 and if a patient with asthma had not met that diagnosis within the past 2 years then they were not included in the prevalence and may either become part of the incidence and then prevalence or may never be picked up. With a long period of latency, certain cases of childhood asthma that becomes symptomatic in adulthood again, may be difficult to ascertain by the duration of the study. In our study, the accuracy in increased by a long look-back period that started in 1983. Other challenges include immigration and emigration as people with asthma may come into the province and leave, leading to cases that are missed. It is very likely that here are also cases of overdiagnosis, especially in the pediatric population. Starting in the early to mid-1990’s there was an increase in asthma awareness which may have influenced the increase in diagnosis and therefore incidence31. Indeed, recommendations of the Canadian Asthma Consensus Report of 1999 may have also impacted physician awareness and diagnosis rates32. The 2003 Canadian Asthma Consensus Guidelines Executive Summary highlighted the importance of early diagnosis of asthma and included criteria supporting a diagnosis of asthma in preschool children33.
Practical applications of epidemiologic studies:
There is need for a unified, accurate system for epidemiological research that can improve resource allocation and therefore clinical care and outcomes. If we can harness data from all provinces and territories and more accurately describe the epidemiology of asthma across the country, we will be better able to fill in gaps in health care resource allocation. We hope this study is the starting point of future studies that can analyze important topics such as medication use, pulmonary function testing and cost to the healthcare system.
Having accurate measurements of the state of a disease fosters the allocation of resources in a responsible manner. Over the past several years, Alberta’s annual health care budgets have been over 20 billion dollars, about 43% of the total annual budget. Strategizing healthcare funding and investments can lead to more efficient and effective overall spending. For example, an accurate asthma diagnosis with pulmonary function testing and then proper treatment and general practitioner follow-up can lead to less ED visits, hospitalizations, time off-work, impaired lung function and mortality which could translate into cost savings in the long term. Descriptive epidemiological studies that can accurately identify the burden of chronic disease such as asthma are important for disease monitoring over time, resource allocation and measuring the effectiveness of interventions34.