A spatial analysis of asthma prevalence in Ontario.

Crighton, Eric J. ; Feng, Jing ; Gershon, Andrea 等

Asthma is the most common chronic respiratory disease in Canada, representing a significant burden on the individual and society in terms of reduced productivity and increased demands on the health care system. (1) While recent Ontario evidence suggests that asthma incidence is decreasing as disease management and education improves, given the absence of a cure, prevalence continues to increase. (2) Despite the significant burden of asthma, we still know very little about how prevalence varies geographically in Canada and (for the purposes of this study) specifically in Ontario. Prevention programs and health care policies developed based on assumptions of spatial homogeneity may contradict evidence (3-5) and therefore may not lead to improved health outcomes. Research aimed at understanding how asthma prevalence varies geographically is critical for informing the effective allocation of scarce public health resources and can provide clues to understanding determinants of the disease.

Research has shown that factors related to asthma are highly spatially variable in Canadian contexts. Lajoie et al. (6) found strong regional variability in asthma-related emergency department (ED) visits in Quebec, with the highest rates being in urban areas and areas of lower socio-economic status. Significant regional variability was also identified by Lougheed (7) in an analysis comparing hospitalizations to ED visits across Ontario hospital catchment areas. In both studies, the authors acknowledge that outcomes such as ED visits do not reflect asthma prevalence but rather access to community health care and public health services. Studies using survey-derived self-reported measures are also commonly used to make geographic comparisons of asthma prevalence. (3,8,9) These studies, however, are often limited by inadequate sample sizes for detailed analyses, or are highly localized or population-specific in their focus.

There has been a growing interest in developing and mapping asthma prevalence at the population level using administrative health data to improve disease surveillance, inform resource allocation and provide insight into disease aetiology. In contrast to survey data, population-based studies have the potential to yield results that are more generalizable to the population at large. However, reliable prevalence estimates require data that represent more than just one type of health service (e.g., ED visits). One such study conducted by To et al. (5) examined asthma prevalence across Ontario's 14 LHINs, estimated using hospitalization, emergency department visit and physician visit data. Findings revealed a 1.6-fold variation between the lowest rates (in the North) and the highest (in the South). Although limited by the use of large areal units, the findings identified a need for further research aimed at better understanding the geographies of Ontario's asthma burden. The current study attempts to address this need by exploring spatial patterns of asthma prevalence in the province by age and sex using a population-based validated asthma registry at a more refined geographic scale.

METHODS

We conducted a retrospective, population-based ecological-level study to assess spatial patterns of asthma prevalence in Ontario over a 5-year period (fiscal years 2002 to 2006) at the sub-Local Health Integration Network (subLHIN) level (version 9). For the purpose of public health and health care planning, Ontario is divided into 14 LHINs which are subdivided into subLHINs (n=141) for more refined local planning (see Figure 1).

Asthma cases were identified through the Ontario Asthma Surveillance Information System Database (OASIS), a validated registry of all Ontario residents with asthma. The registry was generated using: the Ontario Health Insurance Plan (OHIP) which contains information on all fee-for-service billings for physician services as well as emergency department visits, including diagnosis; the Canadian Institute for Health Information Discharge Abstract Database which records the diagnoses for all patients discharged from acute care hospitals; and the Registered Persons Database (RPDB) which contains information on date of birth, sex and location of residence. Databases were linked using encrypted Ontario health insurance numbers. Cases are defined as anyone with at least two asthma-related physician visits within two consecutive years and/or at least one asthma hospitalization since April 1, 1991. This definition, described previously, (2,10,11) yielded 89% sensitivity and 72% specificity in children (<18 years of age) and 84% sensitivity and 76% specificity in adults (>17 years of age). Cases have accumulated since April 1, 1996, and once entered into the database, remain there for as long as the individual is alive and lives in the province. This is consistent with evidence indicating that once diagnosed, asthma may remit but will not resolve. (2,12,13)

To ensure reliable prevalence estimates, average rates were calculated over the study period. Population estimates for all subLHINs (n=141) were not available from any one source, therefore estimates were made using a combination of 2006 weighted population data from the RPDB and 2006 population data from the Ministry of Health and Long-Term Care. While OASIS data go back to 1991, analysis was limited to the 5-year period (fiscal years 2002-2006) to ensure consistent boundary definitions. Descriptive statistics for age- and sex-specific asthma prevalence (unadjusted) were calculated at the provincial and subLHIN level. Statistics include the coefficient of variation (CV), defined as the standard deviation expressed as a percentage of the mean. Using the 2006 Ontario population, rates were then age and sex standardized (direct method) using 10 age categories: 0-4; 5-9; 10-14; 15-19; 20-29; 30-39; 40-49; 50-59; 60-69 and 70+). Following this, comparative morbidity figures (CMFs) were calculated. (14) The CMF is a ratio between the observed directly standardized morbidity rate in a given subLHIN and the expected provincial rate. A CMF value <1 indicates that the rate is below the provincial average and a value >1 indicates that it is above. CMF confidence intervals (at 95%) were calculated (gamma method) and mapped.

To formally test for clusters of high or low CMFs (i.e., "hot spots" and "cold spots"), Local Indicator of Spatial Autocorrelation (LISA) analyses were conducted. (15) Significant spatial autocorrelation indicates a regular pattern in data over space such that a value at a given location depends on, and is similar to, a value of defined spatial neighbours. Global indicators of spatial autocorrelation such as the Moran's I statistic (16) are commonly used to assess this but do not detect localized patterns. The LISA allows for the decomposition of the global indicator into the contribution of each individual observation. A positive value indicates clustering of similarly high values or low values. Neighbour relationships were defined using a queen's contiguity method, expressed in a row-standardized spatial weights matrix. (16) To test for significant departures from zero autocorrelation, a Monte Carlo permutation approach was used (999 permutations at [alpha]<0.05) and a Bonferroni correction was applied. Analysis was carried out in SAS version 9.2 (SAS Institute Inc., Cary, NC) and GeoDa (v.0.9.5.1). (17)

[FIGURE 1 OMITTED]

RESULTS

There were 1,601,353 individuals identified as having asthma in Ontario over the study period (Table 1). The overall prevalence rate was 12.93%, with age-adjusted rates being slightly higher for females than for males at 13.54% and 12.31%, respectively. The highest rates were among males under 20 years, ranging from 18.70% (0-4 years) to 28.02% (10-14 years). The lowest rates are seen among working-age adults where rates were higher for females than for males (e.g., 11.95% vs. 6.99%, respectively, for 50-59 year age group). When disaggregated by subLHIN, the coefficient of variation (CV) shows considerable subLHIN variability that is most pronounced in the youngest age group (e.g., CV=34.02% for females aged 0-4 years) and the oldest age group (e.g., CV=34.41% for males age 70+ years).

Figure 2 shows CMF maps and LISA results for the total population and by sex. Similar patterns of significantly high (p<0.05) CMFs are seen across subLHINs in Eastern Ontario (Champlain LHIN) and near Toronto, where rates are more than 1.2 times the provincial average. The lowest CMFs are seen in the far north and far south of the province, where rates are approximately 1.3 times below the provincial average. Hot spots were detected for both males and females in central Ontario (Central East LHIN) (Figure 2). Among females, one additional cluster was detected to the south of Ottawa (Champlain LHIN). Among males, two clusters are seen in the suburban and rural areas to the west and north of Toronto. No hot spots were detected for the total population. Several cold spots were identified for all three groups, including one in the far north and several in Southern Ontario.

Figure 3 shows CMF maps and the results of LISA analyses on asthma prevalence for three age groups. One age group was chosen to illustrate each of the following demographics: children, working-age adults, and older adults. For each of the age groups, significantly high (p<0.05) CMFs are seen in the east of the province (Champlain LHIN) and north and east of Toronto. Notable in the 30-39 year age group is the large number of significantly high CMF values relative to the 60-69 year age group. Again, the lowest CMFs are found among subLHINs in the North (i.e., North East and North West LHINs) and in the South (e.g., Waterloo Wellington LHIN and Hamilton Niagara Haldimand Brant LHIN (HNHB)). Significant hot spots were detected across all age groups (Figure 3), although there is little overlap. For the 10-14 year age group, there are four small clusters spread out across Southern Ontario: one near Fort Erie and Port Colburn (HNHB LHIN); a second near Peterborough (Central East LHIN); a third in the suburbs around Toronto; and a fourth near Sarnia (Erie St. Claire LHIN). In the 30-39 year age group, there are two large clusters centered in eastern Ontario: one southeast of Ottawa (Champlain LHIN), and a large one centered on Peterborough (Central East LHIN). Among 60-69 year olds, two clusters were identified: one large cluster in the suburbs around Toronto, and a second in Cornwall (Champlain LHIN). Several cold spots are common to all age groups: one centred on Hamilton (HNHB LHIN); a second in the North (North East LHIN); and a third near London (South West LHIN). An additional cold spot centered on Kingston (South East LHIN) is seen in the 60-69 year age group.

DISCUSSION

Using population-level data for the province of Ontario, this study examined spatial patterns of asthma prevalence by age and sex. Before discussing the findings, a few limitations of the study should be addressed. First, this is a descriptive study and as such does not address factors that may explain the spatial patterns identified. Second, the asthma definition used in the OASIS database does not reflect cases where no physician diagnosis has occurred, and as a result less-severe cases might get missed where primary care services are lacking. Also missing from the database are individuals who sought treatment out of province, First Nations populations treated on reserves, and uninsured individuals. Third, COPD (chronic obstructive pulmonary disease) is commonly misdiagnosed in elderly populations as asthma, (18) thereby potentially inflating prevalence rates for the oldest age groups in our study. Finally, the OASIS database only begins in 1991 and as such does not include individuals diagnosed before this year if they have not received treatments since.

Overall prevalence (Table 1) is consistent with other studies using OASIS data but considerably higher than estimates from studies using survey data. (19-21) For example, Garner and Kohen (20) reporting on data from the National Longitudinal Survey of Children and Youth identified 13.4% of children under 12 years as having asthma compared to 19.5% here (data not shown). A partial explanation for this may relate to recall bias in survey data. It can be expected that recently diagnosed incident cases may be well approximated using surveys whereas more remotely diagnosed prevalent cases are missed. (19) Similarly, those with well-controlled asthma who have not experienced symptoms for a lengthy duration may be under-reported. As such, caution should be taken in making direct comparisons between survey and health administrative data. (2,22) While absolute prevalence rates may differ, the relative age and sex patterns here are fairly consistent with other Canadian studies. (2,20,22,23)

This research shows relatively little overlap between asthma hot spots by age group, suggesting that different spatial processes are at play. For example, hot spots in the 10-14 year age group (Figure 3) are centered in areas known for industrial developments and air pollution, including Sarnia (South West LHIN) and Port Colborne (HNHB LHIN). (24) Air pollution is identified in the literature as an important determinant of asthma, particularly among children. (25,26) On the other hand, the clusters of high CMFs among the 30-39 year age group are centered on rural and less environmentally challenged areas of Central Ontario. It could be hypothesized that occupational or natural environmental factors (i.e., allergens) may be playing a role. (27) Unlike the high CMF clusters, the spatial pattern of cold spots is fairly consistent across groups, with cold spots occurring both in urban/suburban environments in and around Toronto and Hamilton, and in rural areas, including Perth and Huron subLHINs (South West LHIN) and the Cochrane subLHIN (North East LHIN). Low CMFs in the North and rural South may be due in part to poor access to primary health care. (28) In Toronto and Hamilton, cold spots could be expected to reflect the presence of various "protective" factors, including socio-economic or lifestyle factors. (3,29,30) Further research is required to understand these potential relationships.

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

This research demonstrates the utility of OASIS data for asthma surveillance, in which spatial surveillance is a key component. Results demonstrate that there are marked age- and sex-specific spatial patterns of asthma prevalence across the province. Understanding where the needs are greatest and among whom can be expected to help inform more effective public health and health care programming and resource allocation. These results will also inform a future study examining relationships between spatial patterns of asthma prevalence and potential disease determinants, including air pollution, allergens, socio-economic conditions, occupation and health care.

Acknowledgements: Funding for this study was provided by the Government of Ontario and AllerGen NCE Inc. Population-based data were provided by the Institute for Clinical Evaluative Sciences (ICES). The sponsors/funders had no influence on design and conduct of the study; collection, management, analysis and interpretation of the data; or preparation, review and approval of the manuscript. The opinions, results and conclusions are those of the authors and no endorsement by the Government of Ontario, AllerGen NCE Inc. or ICES is intended or should be inferred.

Conflict of Interest: None to declare.

REFERENCES

(1.) Public Health Agency of Canada. Life and breath: Respiratory disease in Canada. Ottawa, ON: Government of Canada, 2007.

(2.) Gershon AS, Guan J, Wang C, To T. Trends in asthma prevalence and incidence in Ontario, Canada, 1996-2005: A population study. Am J Epidemiol 2010;172(6):728-36.

(3.) Crighton EJ, Wilson K, Senecal S. The relationship between socio-economic and geographic factors and asthma among Canada's Aboriginal populations. Int J Circumpolar Health 2010;69(2):138-50.

(4.) Chen E, Johanson MD, Thillaiampalam S, Sambell C. Asthma. Health Rep 2005;16(2):4.

(5.) To T, Gershon A, Tassoudji M, Guan J, Wang C, Estrabillo E, et al. The Burden of Asthma in Ontario. Toronto, ON: Institute for Clinical Evaluative Sciences (ICES), 2006.

(6.) Lajoie P, Laberge A, Lebel G, Boulet LP, Demers M, Mercier P, et al. Cartography of emergency department visits for asthma--targeting high-morbidity populations. Can Respir J 2004;11(6):427-33.

(7.) Lougheed MD, Garvey N, Chapman KR, Cicutto L, Dales R, Day AG, et al. The Ontario Asthma Regional Variation Study: Emergency department visit rates and the relation to hospitalization rates. Chest 2006;129(4):909-17.

(8.) Wang HY, Pizzichini MM, Becker AB, Duncan JM, Ferguson AC, Greene JM, et al. Disparate geographic prevalences of asthma, allergic rhinoconjunctivitis and atopic eczema among adolescents in five Canadian cities. Pediatr Allergy Immunol 2010;21(5):867-77.

(9.) Sahsuvaroglu T, Jerrett M, Sears MR, McConnell R, Finkelstein N, Arain A, et al. Spatial analysis of air pollution and childhood asthma in Hamilton, Canada: Comparing exposure methods in sensitive subgroups. Environ Health 2009;8:14.

(10.) To T, Dell S, Dick PT, Cicutto L, Harris JK, MacLusky IB, et al. Case verification of children with asthma in ontario. Pediatr Allergy Immunol 2006;17(1):69-76.

(11.) Gershon AS, Wang C, Guan J, Vasilevska-Ristovska J, Cicutto L, To T. Identifying patients with physician-diagnosed asthma in health administrative databases. Can Respir J 2009;16(6):183-88.

(12.) Stern DA, Morgan WJ, Halonen M, Wright AL, Martinez FD. Wheezing and bronchial hyper-responsiveness in early childhood as predictors of newly diagnosed asthma in early adulthood: A longitudinal birth-cohort study. Lancet 2008;372(9643):1058-64.

(13.) Gershon AS, Guan J, Victor JC, Wang C, To T. The course of asthma activity: A population study. J Allergy Clin Immunol 2012;129(3):679-86.

(14.) Breslow N, Day N. Statistical Methods in Cancer Research. Lyon, France: IARC Scientific Publications, 1987.

(15.) Anselin L. Local indicators of spatial association--LISA. GeogrAnal 1995;27:93-115.

(16.) Baily T, Gatrell A. Interactive Spatial Data Analysis. Harlow, England: Prentice Hall, 1995.

(17.) Anselin L, Syabri I, Kho Y. GeoDa: An introduction to spatial data analysis. Geogr Anal 2006;38:5-22.

(18.) Tinkelman DG, Price DB, Nordyke RJ, Halbert RJ. Misdiagnosis of COPD and asthma in primary care patients 40 years of age and over. J Asthma 2006;43(1):75-80.

(19.) von Hertzen L, Haahtela T. Signs of reversing trends in prevalence of asthma. Allergy 2005;60(3):283-92.

(20.) Garner R, Kohen D. Changes in the prevalence of asthma among Canadian children. Health Rep 2008;19(2):45-50.

(21.) Dales RE, Raizenne M, El-Saadany S, Brook J, Burnett R. Prevalence of childhood asthma across Canada. Int J Epidemiol 1994;23(4):775-81.

(22.) Huzel L, Roos LL, Anthonisen NR, Manfreda J. Diagnosing asthma: The fit between survey and administrative database. Can Respir J2002;9(6):407-12.

(23.) Crighton EJ, Mamdani MM, Upshur RE. A population based time series analysis of asthma hospitalisations in Ontario, Canada: 1988 to 2000. BMC Health SerrvRes 2001;1(1):7.

(24.) Ministry of the Environment. Air Quality in Ontario: Report for 2010. Toronto: Ministry of the Environment, 2012.

(25.) Price K, Plante C, Goudreau S, Pascua E, Perron S, Smargiassi A. Risk of childhood asthma prevalence attributable to residential proximity to major roads in Montreal, Canada. Can J Public Health 2012;103(2):113-18.

(26.) Lavigne L, Villeneuve PJ, Cakmak S. Air pollution and emergency department visits for asthma in Windsor, Canada. Can J Public Health 2012;103(1):4-8.

(27.) Mapp CE, Boschetto P, Maestrelli P, Fabbri LM. Occupational asthma. Am J Respir Crit Care Med 2005;172(3):280-305.

(28.) Hay C, Pacey M, Bains N, Ardal S. Understanding the unattached population in Ontario: Evidence from the Primary Care Access Survey (PCAS). Healthcare Policy 2010;6(2):33-47.

(29.) Babey SH, Hastert TA, Meng YY, Brown ER. Low-income Californians bear unequal burden of asthma. Policy Brief UCLA Cent Health Policy Res 2007;(PB2007-1):1-7.

(30.) Gwynn RC. Risk factors for asthma in US adults: Results from the 2000 Behavioral Risk Factor Surveillance System. J Asthma 2004;41(1):91-98.

Received: May 4, 2012

Accepted: July 18, 2012

Eric J. Crighton, PhD, [1,2] Jing Feng, MSc, [2] Andrea Gershon, MD, MSc, [3] Jun Guan, MSc, [3] Teresa To, PhD [3,4]

Author Affiliations

[1.] Director, Environment and Health Analysis Laboratory (HEALab), Department of Geography, University of Ottawa, Ottawa, ON

[2.] Department of Geography, University of Ottawa, Ottawa, ON

[3.] Institute for Clinical Evaluative Sciences, Toronto, ON

[4.] Child Health Evaluative Sciences, The Hospital for Sick Children, Toronto, ON

Correspondence: Eric Crighton, Director, Environment and Health Analysis Laboratory (HEALab), Department of Geography, 60 University Private, Room 06, University of Ottawa, Ottawa, ON K1N 7Z5, Tel: 613-562-5800, ext. 1065, Fax: 613562-5145, E-mail: eric.crighton@uottawa.ca

Table 1. Asthma Prevalence (Unadjusted) and Variability in Ontario
by Age and Sex for Fiscal Years 2002 to 2006

Age Group   Sex          Province Level *
(years)

Count     Prevalence
                                      Rate (%)

0-4         Male          65,815       18.70
            Female        40,729       12.15
5-9         Male          105,168      26.95
            Female        71,079       18.99
10-14       Male          119,396      28.02
            Female        86,244       21.14
15-19       Male          87,723       20.23
            Female        71,943       17.52
20-29       Male          90,252       10.79
            Female        110,845      13.50
30-39       Male          68,000        7.27
            Female        107,589      11.56
40-49       Male          73,123        7.15
            Female        119,051      11.69
50-59       Male          53,694        6.99
            Female        94,291       11.95
60-69       Male          39,306        8.17
            Female        63,567       12.33
70+         Male          50,582       10.74
            Female        82,964       12.49
All ages    Male          753,055      12.31
            Female        848,298      13.54
Total       Both sexes   1,601,353     12.93

Age Group   Sex                       SubLHIN Levelt
(years)

Median         Range        CV ([double
                         Prevalence                    dagger]) (%)

0-4         Male           16.40        7.28-47.39        32.31
            Female         10.64        2.86-26.65        34.02
5-9         Male           24.45        9.61-52.71        24.57
            Female         17.81        5.80-38.69        27.64
10-14       Male           27.33        7.12-47.54        20.84
            Female         20.62        5.53-41.48        24.93
15-19       Male           20.25        3.43-31.89        20.30
            Female         17.85        3.13-29.68        21.85
20-29       Male           11.19        1.94-16.79        20.89
            Female         14.82        4.15-24.67        24.15
30-39       Male            7.41        2.02-11.66        20.19
            Female         12.31        6.03-21.28        22.08
40-49       Male            6.94        1.85-16.90        24.65
            Female         11.82        4.41-29.71        23.04
50-59       Male            6.80        2.66-21.73        27.70
            Female         11.90        6.26-33.83        24.41
60-69       Male            7.97        2.63-18.94        27.67
            Female         12.13        5.52-35.51        27.53
70+         Male           10.17        3.06-40.18        34.41
            Female         11.97        5.65-45.88        32.99
All ages    Male           12.23        4.20-23.99        20.04
            Female         13.77        4.84-29.08        19.54
Total       Both sexes     12.94        4.51-26.66        19.37

* Counts and rates based on aggregate province-level data.

([dagger]) Median, range and CV calculated using ecological
subLHIN level data (n=141).

([double dagger]) CV=Coefficient of variation.