A scoping review of asthma and machine learning.

OBJECTIVE: The objective of this study was to determine the extent of machine learning (ML) application in asthma research and to identify research gaps while mapping the existing literature. DATA SOURCES: We conducted a scoping review. PubMed, ProQuest, and Embase Scopus databases were searched with an end date of September 18, 2020. STUDY SELECTION: DistillerSR was used for data management. Inclusion criteria were an asthma focus, human participants, ML techniques, and written in English. Exclusion criteria were abstract only, simulation-based, not human based, or were reviews or commentaries. Descriptive statistics were presented. RESULTS: A total of 6,317 potential articles were found. After removing duplicates, and reviewing the titles and abstracts, 102 articles were included for the full text analysis. Asthma episode prediction (24.5%), asthma phenotype classification (16.7%), and genetic profiling of asthma (12.7%) were the top three study topics. Cohort (52.9%), cross-sectional (20.6%), and case-control studies (11.8%) were the study designs most frequently used. Regarding the ML techniques, 34.3% of the studies used more than one technique. Neural networks, clustering, and random forests were the most common ML techniques used where they were used in 20.6%, 18.6%, and 17.6% of studies, respectively. Very few studies considered location of residence (i.e. urban or rural status). CONCLUSIONS: The use of ML in asthma studies has been increasing with most of this focused on the three major topics (>50%). Future research using ML could focus on gaps such as a broader range of study topics and focus on its use in additional populations (e.g. location of residence).Supplemental data for this article is available online at http://dx.doi.org/ .

Atopy risk among school-aged children in relation to early exposures to a farm environment: A systematic review.

BACKGROUND AND OBJECTIVES: Childhood atopy is a complex condition with both a genetic and an environmental component. This systematic review will explore the current understanding of the importance of early life exposures to a farm in the development of atopy measured by objective markers of skin prick testing, and specific IgE measurements in school age children. METHODS: A systematic review was performed. RESULTS: Among 7285 references identified, 14 studies met the inclusion criteria (13 cross-sectional studies and 1 case-control study). The results were fairly consistent in that early farm-related exposures can protect children from becoming atopic at school age. In general, there was heterogeneity in the assessment of outcomes and exposures. CONCLUSIONS: Early-life farm exposures are associated with a protective effect on childhood atopy as assessed by objective markers. Future work should focus on understanding specific farm exposures that may important in these associations between atopy and farm exposures in children.

Personal and early life factors associated with new-onset asthma, remission, and persistence of asthma in a 2-year follow-up of schoolchildren.

OBJECTIVE: The natural course of childhood asthma, after its onset, is characterized by periods of persistence, relapse and remission. To investigate personal and early life factors associated with new-onset asthma, persistence and remission among children. METHODS: The study was conducted in the province of Saskatchewan, Canada. Children in grades Kindergarten to Grade 8 (ages 5-14 years) participated in a cross-sectional study in 2013. In 2015, we approached those who gave consent in 2013 to be re-contacted, creating a prospective cohort. Data were collected using questionnaires in both years. Participants in 2013 who also had data in 2015 (25%: n = 324/1,348) had their asthma status reclassified and longitudinal descriptors were applied: "no asthma", "new-onset asthma", "persistent" or "remission". Personal and early life factors associations with asthma outcomes in 2015 were evaluated. RESULTS: Among those without asthma in 2013 (n = 245), the incidence of new-onset asthma in 2015 was 7.2%. Among those with asthma in 2013 (n = 79), 47.1% had remission and 52.9% had persistent asthma in 2015. Parental history of asthma (adjusted odds ratio (aOR): 4.99; 95% confidence interval (CI): 1.88-28.27), early life respiratory infection (aOR: 1.92; 95%CI: 1.47-7.88), early life allergy [aOR: 6.39; 95%CI: 1.34-30.58) and early life infection (aOR: 4.99; 95%CI: 1.19-20.93) were associated with new onset asthma. Similarly, while parental history of asthma (aOR: 1.13; 95%CI: 0.29-4.34), early life respiratory infection (aOR: 2.71; 95%CI: 0.70-10.45), and early life ear infection (aOR: 1.34; 95%CI: 0.36-5.05) were also positively association with persistent asthma, the associations were not statistically significant. CONCLUSION: Parental history of asthma, early life respiratory infection and allergy might not only influence the onset of childhood asthma but also be associated with asthma persistence.

The course of asthma: A population-based 10-year study examining asthma remission in children diagnosed with asthma in preschool.

BACKGROUND: The natural course of asthma may differ depending on the age of onset. OBJECTIVE: To investigate predictors of asthma remission with a focus on the age of asthma onset. METHODS: The study was a retrospective birth cohort of children with asthma in Saskatchewan, Canada. Using the Saskatchewan Ministry of Health databases, we identified children with a diagnosis of asthma in the first 6 years of life and who had at least 10 years of follow-up after diagnosis (n = 22 563). Of these, we included 6393 children either with persistent asthma (≥1 physician visit or hospitalization for asthma [PVHA] during each year of follow-up) and those who had remission (had PVHA in the first year after diagnosis but at some point during the follow-up no longer received PVHA until end of the study). We used survival analysis to examine associations between remission and age of asthma onset. RESULTS: Of the study participants, 87.2% had early-onset (≤3 years) and 12.8% had late-onset (4-6 years) asthma. Over the 10-years of follow-up, the rate of asthma remission was 37 per 100 person-years. Early-onset asthma (hazard ratio [HR] = 1.10; 95%confidence interval [CI]: 1.01-1.20), being female (HR = 1.12; 95%CI: 1.07-1.18), living in a rural (HR = 1.20; 95%CI: 1.14-1.27) and medium urban (HR = 1.16; 95%CI: 1.08-1.26) location were positively associated with remission while history of atopy decreased likelihood of remission (HR = 0.73; 95%CI: 0.54-0.97). CONCLUSION: Most children with asthma experienced remission, especially those with the onset of symptoms within the first 3 years of life.

Are Dietary Factors Associated with Lung Function in Canadian Adults?

Purpose: The purpose of this study was to explore the impact of dietary factors and biomarkers on lung function among Canadian adults (18-79 years). Methods: Our data source was the Canadian Health Measures Survey Cycle-3, which included 3397 adults. The household and clinic questionnaires and physical measures were used to assess individual dietary factors, modified Mediterranean Diet Scores, and biomarkers. Results: The overall mean percent predicted values for FVC and FEV1 were 97% and 95%, respectively. While somewhat inconsistent between outcomes, higher lung function was associated with bean, grain, milk, fruit, and vegetable consumption, whereas lower lung function was associated with egg and potato consumption. Among the biomarkers, vitamin D, chloride, total serum protein, and red blood cell folate were associated with higher lung function, whereas C-reactive protein and vitamin B12 was associated with lower lung function. Conclusion: Our study provides support for an association between some dietary factors and lung function, though not entirely consistent between a specific dietary factor and the outcomes studied (FVC, FEV1, FVC/FEV1, and FEF25%-75%). The associations between a specific biomarker and lung function were more consistent (i.e., observed with a larger number of lung function outcomes) than were the dietary factors.

Incidence of Daytime Sleepiness and Associated Factors in Two First Nations Communities in Saskatchewan, Canada.

Excessive daytime sleepiness (EDS) is the tendency to sleep at inappropriate times during the day. It can interfere with day-to-day activities and lead to several health issues. The objective of this study was to investigate the association between income, housing conditions, and incidence of EDS in adults living in two Cree First Nations communities. The data for this study involved 317 individuals aged 18 years and older who participated in baseline and follow-up evaluations (after four years) of the First Nations Lung Health Project, which was conducted in Saskatchewan in 2012-2013 and 2016. Both at baseline and follow-up survey after four years, an Epworth Sleepiness Scale (ESS) score >10 was considered to be abnormal. Logistic regression models were used to assess relationships between abnormal ESS and covariates at baseline. In 2016, 7.6% (24/317) of the participants reported an ESS >10 with the mean being 12.8 ± 2.0. For the same group, the mean ESS at baseline was 6.9 ± 2.2. The incidence of subjective EDS based on the ESS >10 was estimated at 7.6% over four years. This study showed an association between incidence of subjective EDS and less money left over at end of the month, having a house in need of repairs, having water or dampness in the past 12 months, and damage caused by dampness.

The association between endotoxin and beta-(1 → 3)-D-glucan in house dust with asthma severity among schoolchildren.

BACKGROUND: Asthma severity can be affected by microbial exposures. However, less is known about the specific indoor agents aggravating the disease in children. We examined the associations between indoor endotoxin and beta-(1 → 3)-D-glucan exposures and asthma severity in children with asthma. METHODS: A clinical cross-sectional study of schoolchildren (aged 7-17 years) was conducted in the province of Saskatchewan, Canada. Children with asthma (n = 116) were identified from 335 participants using a combination of survey responses and objective clinical assessments. We then ascertained asthma severity based on recommended guidelines (continuous daytime asthma symptoms, frequent nighttime asthma symptoms, and ≤ 60% predicted FEV1). Levels of indoor endotoxin and beta-(1 → 3)-D-glucan were measured in dust samples obtained from play area floors and child's mattresses. RESULTS: The study population of 116 children with asthma was comprised of 75.9% mild asthma and 24.1% moderate/severe asthma. Higher mattress endotoxin concentration was associated with increased odds of moderate/severe asthma [adjusted odds ratio (aOR) = 11.40, 95% confidence interval (CI): 1.45-89.43] while higher beta-(1 → 3)-D-glucan concentration (aOR = 0.16, 95% CI: 0.03-0.89) and load (aOR = 0.10, 95% CI: 0.02-0.72) in play areas were inversely associated with moderate/severe asthma. Furthermore, higher mattress endotoxin concentration was associated with lower FVC (p = 0.01) and FEV1 (p = 0.03). These associations were not seen for beta-(1 → 3)-D-glucan. CONCLUSION: Our results showed differential effects of microbial exposures on childhood asthma severity and further highlight domestic endotoxin exposure effects on respiratory health outcomes in children with asthma.

The association between endotoxin in house dust with atopy and exercise-induced bronchospasm in children with asthma.

BACKGROUND: Studies have reported protective and adverse associations between microbial exposure and childhood asthma. However, among children with asthma the relationships between endotoxin and exercise-induced bronchospasm (EIB) is less clear. OBJECTIVE: We investigated the association between exposure to endotoxin in house dust with atopy and EIB in children with asthma. METHODS: A cross-sectional survey was conducted among schoolchildren (aged 7-17 years) in the province of Saskatchewan, Canada. A subpopulation with asthma (n = 116) were identified from 335 participants using a validated asthma algorithm. We determined atopy among the asthma subpopulation by skin prick testing (SPT) while EIB was evaluated using exercise challenge testing (ECT). Dust samples were collected from mattress and play area floors, and endotoxin was measured in dust extracts. Logistic regression analyses were used to explore associations between endotoxin with atopy and EIB. RESULTS: Among the 116 children with asthma, 99 completed SPT and all had completed ECT. Of these, 71/99 (71.7%) were atopic and 26/116 (22.4%) had EIB. Exposure to high play area endotoxin concentration [adjusted odds ratio (aOR) = 0.15, 95% CI: 0.03-0.85] and load (aOR = 0.11, 95% CI: 0.02-0.73) were negatively associated with atopy. In contrast, EIB was positively associated with high mattress endotoxin concentration (aOR = 6.01, 95% CI: 1.20-30.13). CONCLUSION: Indoor microbial endotoxin exposure has varied associations with atopy and exercise-induced bronchospasm among children with asthma.

Asthma diagnosis among children along an urban-rural gradient.

OBJECTIVE: Studies have reported lower asthma prevalence in rural compared to urban areas. While environmental factors have mostly been implicated for these differences, the lower asthma prevalence could also be linked to asthma under-diagnosis in rural children. We investigate if rural children experience under-diagnosis of asthma more compared to urban children. METHODS: In 2013, we conducted a cross-sectional survey of schoolchildren across an urban-rural gradient in Saskatchewan, Canada. The participants formed sampling frame for future studies. In 2015, we approached those who gave consent in 2013 for further testing, repeated the survey, and conducted clinical testing. Based on survey responses, children were classified into "no asthma," "at-risk-for-asthma," and "diagnosed asthma." We then classified asthma status as either "no asthma" or "probable asthma" based on a validated asthma algorithm. RESULTS: The study population of 335 schoolchildren (aged 7-17 years) comprised of 73.4% from large urban, 13.7% from small urban, and 12.8% from rural areas. Proportion with report of physician-diagnosed asthma was 28.5% (Large urban), 34.8% (Small urban), and 20.9% (Rural). Mean percent predicted FEV1 and FEF25%-75% were lower in rural compared to small urban and large urban children (p < 0.05). Among those not classified as with "diagnosed asthma" by the survey, the algorithm further identified presence of asthma in 5.5% large urban, 8.1% small urban, and 18.8% rural children (p = 0.03). CONCLUSION: The study revealed evidence of asthma underdiagnosis in rural areas and further supports the use of objective measures in addition to symptoms history when investigating asthma across urban-rural gradients.

Predictors of Physical Activity Levels in Community-Dwelling Older Adults: A Multivariate Approach Based on a Socio-Ecological Framework.

The purpose of this study was to explore cross-sectional relationships between self-reported physical activity (PA) and personal, social, and environmental factors in community-dwelling adults aged 50 years and older. Accounting for clustering by neighborhood, generalized estimating equations were used to examine associations between selected correlates and the Physical Activity Scale for the Elderly (PASE) score while adjusting for confounders. Data for 601 participants were analyzed: 79% female, 37% married, mean age 76.8 (± 8.7) years, mean PASE score 112.6 (± 64.8). Age, living in seniors' housing, using nursing/home care services, receiving encouragement to be active, and having benches available in the neighborhood were inversely associated with PASE. Self-efficacy, SF-12 score, PA barriers, social support, and the presence of trails showed positive associations. Several personal, social, and environmental factors associated with PA were identified. The inverse association between PA and living in seniors' housing units should be considered when developing PA programs for older adults.

Bronchitis and Its Associated Risk Factors in First Nations Children.

Respiratory diseases, such as bronchitis and pneumonia, are common in First Nations children in Canada. The objectives are to determine prevalence and associated risk factors of bronchitis in children 6-17 years old residing in two reserve communities. The cross-sectional study was conducted in 2013 and children from two First Nations reserve communities participated. The outcome was ever presence/absence of bronchitis. Logistic regression analysis was conducted to examine the relationship between bronchitis and the individual and environmental factors. A total of 351 First Nations children participated in the study. The prevalence of bronchitis was 17.9%. While 86.6% had at least one parent who smoked, smoking inside home was 43.9%. Signs of mold and mildew in homes were high. Prevalence of houses with any damage caused by dampness was 42.2%, with 44.2% of homes showing signs of mold or mildew. Significant predictors of increased risk of bronchitis were: being obese; having respiratory allergies; exposed to parental cigarette smoking; and signs of mold and mildew in the home. There are several modifiable risk factors that should be considered when examining preventive interventions for bronchitis including obesity, smoking exposure, and home mold or dampness.

An international comparison of asthma, wheeze, and breathing medication use among children.

BACKGROUND: There is variation in childhood asthma between countries with typically higher prevalence in "Westernized" nations. We compared asthma, respiratory symptoms, and medication prevalence in Eastern and Central European regions and Canada. METHODS: We conducted a cross-sectional survey study of children (5-15 years) from one urban centre in each of Canada, Belarus, Poland, Republic of Georgia (Adjara), Republic of Macedonia, and Ukraine. Surveys were distributed through randomly selected schools to parents (2013-2015). RESULTS: The prevalence of asthma differed by country from 20.6% in Canada to 1.5% in Ukraine (p < 0.001). This association remained after confounder adjustment. Except for Canada (58.7%) and Poland (42.5%), less than 10% of children with a history of wheeze had a diagnosis of asthma. Regardless of country, more than 50% of children with a diagnosis of asthma used breathing medications in the past year. Finally, except for Georgia (12.1%), all countries had a prevalence of ever wheeze above 20% (23.8% in Poland to 30.9% in Macedonia). CONCLUSIONS: Despite large differences in asthma prevalence, respiratory morbidity was more comparable suggesting asthma prevalence may be underestimated. Further validation of asthma diagnosis is needed. It is important to promote best diagnostic practices among first contact physicians.

Prevalence and associated risk factors of chronic bronchitis in First Nations people.

BACKGROUND: Inadequate housing, low family income, household smoking, personal smoking status, and poor schooling are some of the conditions that have been significantly associated with the prevalence and incidence of chronic bronchitis. The aim of the current study was to determine the prevalence of chronic bronchitis (CB) and associated risk factors among First Nations people. METHODS: An interviewer-administered survey was conducted as part of the First Nations Lung Health Project in 2012 and 2013 with 874 individuals from 406 households in two First Nations communities located in the province of Saskatchewan, Canada. The questionnaire collected information on individual and contextual determinants of health and a history of ever diagnosed with CB (outcome variable) from the two communities participating in the First Nations Lung Health Project. Clustering effect within households was adjusted using Generalized Estimating Equations. RESULTS: The prevalence of CB was 8.9% and 6.8% among residents (18 years and older) of community A and community B respectively and was not significantly different. CB prevalence was positively associated with odour or musty smell of mildew/mould in the house [OR adj (95% CI) = 2.33 (1.21, 4.50)], allergy to house dust [3.49 (1.75, 6.97)], an air conditioner in home [2.33 (1.18, 4.24)], and increasing age [0.99 (0.33, 2.95), 4.26 (1.74, 10.41), 6.08 (2.58, 14.33)]. An interaction exposure to environmental tobacco smoke in the house*body mass index showed that exposure to household smoke increased the risk of CB for overweight and obese participants (borderline). Some of the variables of interest were not significantly associated with the prevalence of CB in multivariable analysis, possibly due to small numbers. CONCLUSIONS: Our results suggest that significant determinants of CB were: increasing age; odour or musty smell of mildew/mould in the house; allergy to house dust; and, body mass index. Modifiable risk factors identified were: (i) community level-housing conditions (such as mould or mildew in home, exposure to environmental tobacco smoke in house); and, (ii) policy level-remediation of mould, and obesity. TRIAL REGISTRATION: Not applicable.

Childhood asthma, asthma severity indicators, and related conditions along an urban-rural gradient: a cross-sectional study.

BACKGROUND: Asthma prevalence is generally lower in rural locations with some indication of an urban-rural gradient. However, among children with asthma, certain rural exposures thought to protect against the development of asthma could aggravate the condition. We examined childhood asthma prevalence and related conditions along an urban-rural gradient and also examined the characteristics of those with asthma along the urban-rural gradient. METHODS: In 2013 we completed a cross-sectional survey of 3509 children aged 5-14 years living in various population densities of Saskatchewan, Canada. Location of dwelling was identified as belonging to one of the following population densities: large urban region (approximately 200,000), small urban (approximately 35,000), or rural (small town of <1,500 or farm dweller). Physician-diagnosed asthma and asthma-related symptoms were ascertained from responses in the parental-completed questionnaires. RESULTS: Of the study population, 69% lived in a large urban region, 11% lived in a small urban centre and 20% were rural dwellers. Overall, asthma prevalence was 19.6% with differences in asthma prevalence with differences between locations (large urban = 20.7%; small urban = 21.5%; rural = 15.1%; p = 0.003). After adjustment for potential confounders, the association between location of dwelling and asthma remained significant. Despite a lower prevalence of asthma in the rural area, the prevalence and risk of ever wheeze and having more than 3 wheezing episodes in the past 12 months among those who reported asthma, was higher in rural locations after adjustment for potential confounders. CONCLUSIONS: The results of this study support the evidence of a difference in childhood asthma prevalence between urban and rural locations and that once a child has asthma, certain rural exposures may aggravate the disease.

Prevalence, risk factors, and clinical outcomes of atopic and nonatopic asthma among rural children.

BACKGROUND: Because of time and cost constraints, objective classification of atopic and nonatopic asthma has been limited in large epidemiologic studies. However, as we try to better understand exposure-outcome associations and ensure appropriate treatment of asthma, it is important to focus on phenotype-defined asthma classification. OBJECTIVE: To compare atopic and nonatopic asthma in rural children with regard to risk factors and clinical outcomes. METHODS: We conducted a cross-sectional study in rural Saskatchewan, Canada, in 2011. Parents of 6- to 14-year-old children completed a health and exposure survey. Skin prick tests were completed in a subsample of 529 children. Asthma was based physician diagnosis. Asthma status was defined as no asthma, nonatopic asthma, and atopic asthma. RESULTS: Asthma prevalence was 14.7% of which 32.1% of cases were atopic. After adjustment, early respiratory illness and a family history of asthma were predictors of childhood asthma, regardless of atopic status (P < .05). Being overweight and having a dog in the home were associated with an increased risk of nonatopic asthma (P < .05). A mother with a history of smoking increased the risk of atopic asthma (P = .01). Compared with those with nonatopic asthma, in the past 12 months, children with atopic asthma were more likely to report a sneezy, runny, or blocked nose or have shortness of breath (odds ratio >2), whereas those with nonatopic asthma were more likely to have parents who missed work (odd ratio >3). Those with nonatopic asthma had significantly lower forced expiratory volume in 1 second compared w2ith those with atopic asthma. CONCLUSION: Exposures may contribute differentially to atopic and nonatopic asthma and result in differential clinical presentation or burden. The study of these characteristics is important for etiologic understanding and management decisions.

Actual Body Weight and the Parent's Perspective of Child's Body Weight among Rural Canadian Children.

The prevalence of being overweight during childhood continues to increase in the USA and Canada and children living in rural areas are more at risk than their urban counterparts. The objectives of this study were to evaluate how well the parent's perception of their child's weight status correlated with objectively measured weight status among a group of rural children and to identify predictors of inaccurate parental perceptions of child's weight status. Participants were children from the Saskatchewan Rural Health Study conducted in 2010. Self-administered questionnaires were distributed through rural schools to parents of children in grades one to eight. Parents reported their child's height and weight and rated their child's weight status (underweight, just about the right weight, or overweight). Standardized body mass index (BMI) categories were calculated for clinically measured height and weight and for parental report of height and weight for 584 children. Logistic regression analysis was performed to identify predictors of misclassification of the parent's perception of child's weight status adjusting for potential confounders. Clinically measured overweight was much higher (26.5%) compared to parental perceived overweight (7.9%). The misclassification of the child's BMI was more likely to occur if the child was a boy (odds ratio (OR) = 1.58) or non-Caucasian (OR = 2.03). Overweight was high in this group of rural children and parental perception of weight status underestimated the actual weight status of overweight school-age children. Parental reporting of child weight status has implications for public health policy and prevention strategies. Future research should focus on assessing longitudinal effects of parental misperceptions of child's weight status.

Ear Infection and Its Associated Risk Factors in First Nations and Rural School-Aged Canadian Children.

Background. Ear infections in children are a major health problem and may be associated with hearing impairment and delayed language development. Objective. To determine the prevalence and the associated risk factors of ear infections in children 6-17 years old residing on two reserves and rural areas in the province of Saskatchewan. Methodology. Data were provided from two rural cross-sectional children studies. Outcome variable of interest was presence/absence of an ear infection. Logistic regression analysis was conducted to examine the relationship between ear infection and the other covariates. Results. The prevalence of ear infection was 57.8% for rural Caucasian children and 43.6% for First Nations children living on-reserve. First Nations children had a lower risk of ear infection. Ear infection prevalence was positively associated with younger age; first born in the family; self-reported physician-diagnosed tonsillitis; self-reported physician-diagnosed asthma; and any respiratory related allergy. Protective effect of breastfeeding longer than three months was observed on the prevalence of ear infection. Conclusions. While ear infection is a prevalent condition of childhood, First Nations children were less likely to have a history of ear infections when compared to their rural Caucasian counterparts.

Early farm residency and prevalence of asthma and hay fever in adults.

BACKGROUND: Asthma and hay fever have been found to be both positively and negatively associated with farming lifestyles in adulthood. Lack of congruency may depend upon early life exposure. OBJECTIVE: To assess the importance of different periods of farm residency for asthma and hay fever in an adult Canadian population. METHODS: We conducted a questionnaire survey in rural Saskatchewan, Canada. We assessed a history of asthma and hay fever with five categories of farm residency that were mutually exclusive: first year of life only, currently living on a farm, both first year of life and currently living on a farm, other farm living, and no farm living. Generalized estimating equations were used to adjust for clustering effects of adults within households. RESULTS: Of the 7148 responding, 30.6% had an early farm living experience only, 34.4% had both early and current farm living experiences, while 17.4% had never lived on a farm. The overall prevalence of ever asthma and hay fever was 8.6% and 12.3%, respectively, and was higher in women. Sex modified the associations between ever asthma and hay fever with farm residency variables whereby women had a decreased risk for both asthma [adjusted odds ratio (ORadj): 0.67, 95% confidence interval (CI):0.47-0.96] and hay fever (ORadj: 0.60, 95% CI: 0.44-0.83) with an early farm exposure only. Men currently living on a farm without an early farm exposure had an increased risk for ever asthma (ORadj: 1.82, 95% CI: 1.02-3.24). CONCLUSION: Farm residency in the first year of life shows a protective effect for adult asthma and hay fever that appears to differ by sex.