AI for social good: unlocking the opportunity for positive impact

Advances in machine learning (ML) and artificial intelligence (AI) present an opportunity to build better tools and solutions to help address some of the world's most pressing challenges, and deliver positive social impact in accordance with the priorities outlined in the United Nations' 17 Sustainable Development Goals (SDGs). The AI for Social Good (AI4SG) movement aims to establish interdisciplinary partnerships centred around AI applications towards SDGs. We provide a set of guidelines for establishing successful long-term collaborations between AI researchers and application-domain experts, relate them to existing AI4SG projects and identify key opportunities for future AI applications targeted towards social good.

Lung function trajectories from pre-school age to adulthood and their associations with early life factors: a retrospective analysis of three population-based birth cohort studies

BACKGROUND

Maximal lung function in early adulthood is an important determinant of mortality and COPD. We investigated whether distinct trajectories of lung function are present during childhood and whether these extend to adulthood and infancy.

METHODS

To ascertain trajectories of FEV₁, we studied two population-based birth cohorts (MAAS and ALSPAC) with repeat spirometry from childhood into early adulthood (1046 participants from 5-16 years and 1390 participants from 8-24 years). We used a third cohort (PIAF) with repeat lung function measures in infancy (V'_maxFRC) and childhood (FEV₁; 196 participants from 1 month to 18 years of age) to investigate whether these childhood trajectories extend from early life. We identified trajectories using latent profile modelling. We created an allele score to investigate genetic associations of trajectories, and constructed a multivariable model to identify their early-life predictors.

FINDINGS

We identified four childhood FEV₁ trajectories: persistently high, normal, below average, and persistently low. The persistently low trajectory (129 [5%] of 2436 participants) was associated with persistent wheezing and asthma throughout follow-up. In genetic analysis, compared with the normal trajectory, the pooled relative risk ratio per allele was 0·96 (95% CI 0·92-1·01; p=0·13) for persistently high, 1·01 (0·99-1·02; p=0·49) for below average, and 1·05 (0·98-1·13; p=0·13) for persistently low. Most children in the low V'_maxFRC trajectory in infancy did not progress to the low FEV₁ trajectory in childhood. Early-life factors associated with the persistently low trajectory included recurrent wheeze with severe wheezing exacerbations, early allergic sensitisation, and tobacco smoke exposure.

INTERPRETATION

Reduction of childhood smoke exposure and minimisation of the risk of early-life sensitisation and wheezing exacerbations might reduce the risk of diminished lung function in early adulthood.

FUNDING

None.

Relation between circulating CC16 concentrations, lung function, and development of chronic obstructive pulmonary disease across the lifespan: a prospective study

BACKGROUND

Low concentrations of the anti-inflammatory protein CC16 (approved symbol SCGB1A1) in serum have been associated with accelerated decline in forced expiratory volume in 1 s (FEV1) in patients with chronic obstructive pulmonary disease (COPD). We investigated whether low circulating CC16 concentrations precede lung function deficits and incidence of COPD in the general population.

METHODS

We assessed longitudinal data on CC16 concentrations in serum and associations with decline in FEV1 and incidence of airflow limitation for adults who were free from COPD at baseline in the population-based Tucson Epidemiological Study of Airway Obstructive Disease ([TESAOD] n=960, mean follow-up 14 years), European Community Respiratory Health Survey ([ECRHS-Sp] n=514, 11 years), and Swiss Cohort Study on Air Pollution and Lung Diseases in Adults ([SAPALDIA] n=167, 8 years) studies. Additionally, we measured circulating CC16 concentrations in samples from children aged 4-6 years in the Tucson Children's Respiratory Study (n=427), UK Manchester Asthma and Allergy Study (n=481), and the Swedish Barn/children, Allergy, Milieu, Stockholm, Epidemiological survey (n=231) birth cohorts to assess whether low CC16 concentrations in childhood were predictive for subsequent lung function.

FINDINGS

After adjustment for sex, age, height, smoking status and intensity, pack-years, asthma, and FEV1 at baseline, we found an inverse association between CC16 concentration and decline in FEV1 in adults in TESAOD (4·4 mL/year additional FEV1 decline for each SD decrease in baseline CC16 concentration, p=0·0014) and ECRHS-Sp (2·4 mL/year, p=0·023); the effect in SAPALDIA was marginal (4·5 mL/year, p=0·052). Low CC16 concentration at baseline was also associated with increased risk of incident stage 2 airflow limitation (ratio of FEV1 to forced expiratory volume [FEV1/FVC] less than 70% plus FEV1 % predicted less than 80%) in TESAOD and ECRHS-Sp. In children, the lowest tertile of CC16 concentrations was associated with a subsequent FEV1 deficit of 68 mL up to age 16 years (p=0·0001), which was confirmed in children who had never smoked by age 16 years (-71 mL, p<0·0001).

INTERPRETATION

Low concentrations of CC16 in serum are associated with reduced lung function in childhood, accelerated lung function decline in adulthood, and development of moderate airflow limitation in the general adult population.

FUNDING

National Heart, Lung, and Blood Institute and European Union Seventh Framework Programme.

Peanut allergy: effect of environmental peanut exposure in children with filaggrin loss-of-function mutations

BACKGROUND

Filaggrin (FLG) loss-of-function mutations lead to an impaired skin barrier associated with peanut allergy. Household peanut consumption is associated with peanut allergy, and peanut allergen in household dust correlates with household peanut consumption.

OBJECTIVE

We sought to determine whether environmental peanut exposure increases the odds of peanut allergy and whether FLG mutations modulate these odds.

METHODS

Exposure to peanut antigen in dust within the first year of life was measured in a population-based birth cohort. Peanut sensitization and peanut allergy (defined by using oral food challenges or component-resolved diagnostics [CRD]) were assessed at 8 and 11 years. Genotyping was performed for 6 FLG mutations.

RESULTS

After adjustment for infantile atopic dermatitis and preceding egg skin prick test (SPT) sensitization, we found a strong and significant interaction between natural log (ln [loge]) peanut dust levels and FLG mutations on peanut sensitization and peanut allergy. Among children with FLG mutations, for each ln unit increase in the house dust peanut protein level, there was a more than 6-fold increased odds of peanut SPT sensitization, CRD sensitization, or both in children at ages 8 years, 11 years, or both and a greater than 3-fold increased odds of peanut allergy compared with odds seen in children with wild-type FLG. There was no significant effect of exposure in children without FLG mutations. In children carrying an FLG mutation, the threshold level for peanut SPT sensitization was 0.92 μg of peanut protein per gram (95% CI, 0.70-1.22 μg/g), that for CRD sensitization was 1.03 μg/g (95% CI, 0.90-1.82 μg/g), and that for peanut allergy was 1.17 μg/g (95% CI, 0.01-163.83 μg/g).

CONCLUSION

Early-life environmental peanut exposure is associated with an increased risk of peanut sensitization and allergy in children who carry an FLG mutation. These data support the hypothesis that peanut allergy develops through transcutaneous sensitization in children with an impaired skin barrier.

Developmental profiles of eczema, wheeze, and rhinitis: two population-based birth cohort studies

BACKGROUND

The term "atopic march" has been used to imply a natural progression of a cascade of symptoms from eczema to asthma and rhinitis through childhood. We hypothesize that this expression does not adequately describe the natural history of eczema, wheeze, and rhinitis during childhood. We propose that this paradigm arose from cross-sectional analyses of longitudinal studies, and may reflect a population pattern that may not predominate at the individual level.

METHODS AND FINDINGS

Data from 9,801 children in two population-based birth cohorts were used to determine individual profiles of eczema, wheeze, and rhinitis and whether the manifestations of these symptoms followed an atopic march pattern. Children were assessed at ages 1, 3, 5, 8, and 11 y. We used Bayesian machine learning methods to identify distinct latent classes based on individual profiles of eczema, wheeze, and rhinitis. This approach allowed us to identify groups of children with similar patterns of eczema, wheeze, and rhinitis over time. Using a latent disease profile model, the data were best described by eight latent classes: no disease (51.3%), atopic march (3.1%), persistent eczema and wheeze (2.7%), persistent eczema with later-onset rhinitis (4.7%), persistent wheeze with later-onset rhinitis (5.7%), transient wheeze (7.7%), eczema only (15.3%), and rhinitis only (9.6%). When latent variable modelling was carried out separately for the two cohorts, similar results were obtained. Highly concordant patterns of sensitisation were associated with different profiles of eczema, rhinitis, and wheeze. The main limitation of this study was the difference in wording of the questions used to ascertain the presence of eczema, wheeze, and rhinitis in the two cohorts.

CONCLUSIONS

The developmental profiles of eczema, wheeze, and rhinitis are heterogeneous; only a small proportion of children (∼ 7% of those with symptoms) follow trajectory profiles resembling the atopic march. Please see later in the article for the Editors' Summary.

Characterizing wheeze phenotypes to identify endotypes of childhood asthma, and the implications for future management

It is now a commonly held view that asthma is not a single disease, but rather a set of heterogeneous diseases sharing common symptoms. One of the major challenges in treating asthma is understanding these different asthma phenotypes and their underlying biological mechanisms. This review gives an epidemiological perspective of our current understanding of the different phenotypes that develop from birth to childhood that come under the umbrella term 'asthma'. The review focuses mainly on publications from longitudinal birth cohort studies where the natural history of asthma symptoms is observed over time in the whole population. Identifying distinct pathophysiological mechanisms for these different phenotypes will potentially elucidate different asthma endotypes, ultimately leading to more effective treatment and management strategies.

Trajectories of lung function during childhood

RATIONALE

Developmental patterns of lung function during childhood may have major implications for our understanding of the pathogenesis of respiratory disease throughout life.

OBJECTIVES

To explore longitudinal trajectories of lung function during childhood and factors associated with lung function decline.

METHODS

In a population-based birth cohort, specific airway resistance (sRaw) was assessed at age 3 (n = 560), 5 (n = 829), 8 (n = 786), and 11 years (n = 644). Based on prospective data (questionnaires, skin tests, IgE), children were assigned to wheeze phenotypes (no wheezing, transient, late-onset, and persistent) and atopy phenotypes (no atopy, dust mite, non-dust mite, multiple early, and multiple late). We used longitudinal linear mixed models to determine predictors of change in sRaw over time.

MEASUREMENTS AND MAIN RESULTS

Contrary to the assumption that sRaw is independent of age and sex, boys had higher sRaw than girls (mean difference, 0.080; 95% confidence interval [CI], 0.049-0.111; P < 0.001) and a higher rate of increase over time. For girls, sRaw increased by 0.017 kPa ⋅ s(-1) per year (95% CI, 0.011-0.023). In boys this increase was significantly greater (P = 0.012; mean between-sex difference, 0.011 kPa ⋅ s(-1); 95% CI, 0.003-0.019). Children with persistent wheeze (but not other wheeze phenotypes) had a significantly greater rate of deterioration in sRaw over time compared with never wheezers (P = 0.009). Similarly, children with multiple early, but not other atopy phenotypes had significantly poorer lung function than those without atopy (mean difference, 0.116 kPa ⋅ s(-1); 95% CI, 0.065-0.168; P < 0.001). sRaw increased progressively with the increasing number of asthma exacerbations.

CONCLUSIONS

Children with persistent wheeze, frequent asthma exacerbations, and multiple early atopy have diminished lung function throughout childhood, and are at risk of a progressive loss of lung function from age 3 to 11 years. These effects are more marked in boys.

Joint modeling of parentally reported and physician-confirmed wheeze identifies children with persistent troublesome wheezing

BACKGROUND

Previous studies have suggested the presence of different childhood wheeze phenotypes through statistical modeling based on parentally reported wheezing.

OBJECTIVE

We sought to investigate whether joint modeling of observations from both medical records and parental reports helps to more accurately define wheezing disorders during childhood and whether incorporating information from medical records better characterizes severity.

METHODS

In a population-based birth cohort (n = 1184), we analyzed data from 2 sources (parentally reported current wheeze at 4 follow-ups and physician-confirmed wheeze from medical records in each year from birth to age 8 years) to determine classes of children who differ in wheeze trajectories. We tested the validity of these classes by examining their relationships with objective outcomes (lung function, airway hyperreactivity, and atopy), asthma medication, and severe exacerbations.

RESULTS

Longitudinal latent class modeling identified a 5-class model that best described the data. We assigned classes as follows: no wheezing (53.3%), transient early wheeze (13.7%), late-onset wheeze (16.7%), persistent controlled wheeze (13.1%), and persistent troublesome wheeze (PTW; 3.2%). Longitudinal trajectories of atopy and lung function differed significantly between classes. Patients in the PTW class had diminished lung function and more hyperreactive airways compared with all other classes. We observed striking differences in exacerbations, hospitalizations, and unscheduled visits, all of which were markedly higher in patients in the PTW class compared with those in the other classes. For example, the risk of exacerbation was much higher in patients in the PTW class compared with patients with persistent controlled wheeze (odds ratio [OR], 3.58; 95% CI, 1.27-10.09), late-onset wheeze (OR, 15.92; 95% CI, 5.61-45.15), and transient early wheeze (OR, 12.24; 95% CI, 4.28-35.03).

CONCLUSION

We identified a novel group of children with persistent troublesome wheezing, who have markedly different outcomes compared with persistent wheezers with controlled disease.