Regression Models For Binary Outcomes Research Articles

Efficacy and Safety of Allopurinol and Febuxostat in Patients With Gout and CKD: Subgroup Analysis of the STOP Gout Trial

Rationale & ObjectiveWe conducted a prespecified examination of the efficacy and safety of allopurinol and febuxostat administered using a treat-to-target strategy in trial participants with chronic kidney disease (CKD). Study DesignPrespecified sub cohort analysis of a randomized controlled trial. Setting& Participants: A sub study of the STOP Gout trial in participants with CKD. CKD was defined as an eGFR 30-59 mL/min/1.73 m2 at baseline. ExposureTrial participants with CKD and gout and serum urate (sUA) concentration ≥6.8 mg/dL were randomized 1:1 to receive allopurinol or febuxostat. Urate lowering therapy (ULT) was titrated during weeks 0-24 to achieve a goal sUA of <6.0 mg/dl (<5.0 mg/dl with tophi) (Phase 1) and maintained during weeks 25-48 (Phase 2). Gout flare was assessed between weeks 49-72 (Phase 3). OutcomeGout flare between weeks 49-72 (Phase 3) was the primary outcome. Secondary outcomes included sUA goal achievement and ULT dosing at end of Phase 2, and serious adverse events (SAEs). Analytical ApproachOutcomes between treatment groups were compared using logistic regression models for binary outcomes, and Poisson regression for flare rates. Multivariable models were subsequently used, adjusting for factors identified to be imbalanced by treatment arm. Results351 of 940 participants (37.3%) had CKD; 277 were assessed for the primary outcome. Fewer patients randomized to allopurinol had a flare during phase 3 (32% vs 45%; p=0.02) despite similar attainment of sUA goal (79% vs. 81%; p=0.6) by the end of Phase 2. Acute kidney injury (AKI) was more common in participants with stage 3 CKD randomized to allopurinol compared to febuxostat. LimitationsLimited power to assess infrequent safety events, largely male, older population. ConclusionsAllopurinol and febuxostat are similarly efficacious and well-tolerated in the treatment of gout in people with CKD when used in a treat-to-target regimen.

Robit regression in Stata

Logistic and probit models are the most popular regression models for binary outcomes. A simple robust alternative is the robit model, which replaces the underlying normal distribution in the probit model with a Student’s t distribution. The heavier tails of the t distribution (compared with the normal distribution) mean that model outliers are less influential. Robit regression models can be fit as generalized linear models with the link function defined as the inverse cumulative t distribution function with a specified number of degrees of freedom; they have been advocated as being particularly suitable for estimating inverse-probability weights and propensity scoring more generally. Here we describe a new command, robit, that implements robit regression in Stata.

A Bayesian (meta-)regression model for treatment effects on the risk difference scale.

In clinical settings, the absolute risk reduction due to treatment that can be expected in a particular patient is of key interest. However, logistic regression, the default regression model for trials with a binary outcome, produces estimates of the effect of treatment measured as a difference in log odds. We explored options to estimate treatment effects directly as a difference in risk, specifically in the network meta-analysis setting. We propose a novel Bayesian (meta-)regression model for binary outcomes on the additive risk scale. The model allows treatment effects, covariate effects, interactions and variance parameters to be estimated directly on the linear scale of clinical interest. We compared effect estimates from this model to (1) a previously proposed additive risk model by Warn, Thompson and Spiegelhalter ("WTS model") and (2) backtransforming the predictions from a logistic model to the natural scale after regression. The models were compared in a network meta-analysis of 20 hepatitis C trials, as well as in the analysis of simulated single trial settings. The resulting estimates diverged, in particular for small sample sizes or true risks close to 0% or 100%. Researchers should be aware that modelling untransformed risk can yield very different results from default logistic models. The treatment effect in participants with such extreme predicted risks weighed more heavily on the overall treatment effect estimate from our proposed model compared to the WTS model. In our network meta-analysis, this sensitivity of our proposed model was needed to detect all information in the data.

Embolization therapy for type 2 endoleaks after endovascular aortic aneurysm repair: imaging-based predictive factors and clinical outcomes on long-term follow-up

To evaluate the technical, radiological, and clinical outcomes after type 2 endoleak (T2EL) embolization in patients with a growing aneurysm sac after endovascular aortic aneurysm repair (EVAR). Additionally, to determine clinical and imaging-based factors for outcome prediction after embolization of a T2EL. A single-institution, retrospective analysis was performed of 60 patients who underwent a T2EL embolization procedure between September 2005 and August 2016 to treat a growing aneurysm sac diameter following EVAR. The patients' electronic medical records and all available pre- and post-embolization imaging were reviewed. Statistical analysis methods included logistic regression models for binary outcomes, proportional odds models for ordinal outcomes, and linear regression models for continuous outcomes. The Kaplan-Meier method was used to estimate the overall survival probability. Technical, radiological, and clinical success rates after T2EL embolization were 95% (n = 57), 26.7% (n = 16), and 76.7% (n = 46), respectively. Persistent aneurysm sac expansion was found in 31 patients (51.7%). Unsharp or blurred T2EL delineation on pre-interventional computed tomography (CT) was a predictive factor for a post-embolization persistent visible endoleak and persistent growth of the aneurysm sac (P = 0.025). Median survival after T2EL embolization was 5.35 years, with no difference observed between patients with persistent sac expansion compared with patients with stable or decreased sac diameter. Progression of the aneurysm sac diameter was observed in half the study patients, despite technically successful T2EL embolization. Unsharp or blurred T2EL delineation on pre-interventional CT seemed to be an imaging-based predictor for a persistent T2EL and progressive aneurysm sac growth after embolization.

Revisiting IgG Antibody Reactivity to Epstein-Barr Virus in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome and Its Potential Application to Disease Diagnosis.

Infections by the Epstein-Barr virus (EBV) are often at the disease onset of patients suffering from Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS). However, serological analyses of these infections remain inconclusive when comparing patients with healthy controls (HCs). In particular, it is unclear if certain EBV-derived antigens eliciting antibody responses have a biomarker potential for disease diagnosis. With this purpose, we re-analyzed a previously published microarray data on the IgG antibody responses against 3,054 EBV-related antigens in 92 patients with ME/CFS and 50 HCs. This re-analysis consisted of constructing different regression models for binary outcomes with the ability to classify patients and HCs. In these models, we tested for a possible interaction of different antibodies with age and gender. When analyzing the whole data set, there were no antibody responses that could distinguish patients from healthy controls. A similar finding was obtained when comparing patients with non-infectious or unknown disease trigger with healthy controls. However, when data analysis was restricted to the comparison between HCs and patients with a putative infection at their disease onset, we could identify stronger antibody responses against two candidate antigens (EBNA4_0529 and EBNA6_0070). Using antibody responses to these two antigens together with age and gender, the final classification model had an estimated sensitivity and specificity of 0.833 and 0.720, respectively. This reliable case-control discrimination suggested the use of the antibody levels related to these candidate viral epitopes as biomarkers for disease diagnosis in this subgroup of patients. To confirm this finding, a follow-up study will be conducted in a separate cohort of patients.

Overlapping group screening for detection of gene-environment interactions with application to TCGA high-dimensional survival genomic data

BackgroundIn the context of biomedical and epidemiological research, gene-environment (G-E) interaction is of great significance to the etiology and progression of many complex diseases. In high-dimensional genetic data, two general models, marginal and joint models, are proposed to identify important interaction factors. Most existing approaches for identifying G-E interactions are limited owing to the lack of robustness to outliers/contamination in response and predictor data. In particular, right-censored survival outcomes make the associated feature screening even challenging. In this article, we utilize the overlapping group screening (OGS) approach to select important G-E interactions related to clinical survival outcomes by incorporating the gene pathway information under a joint modeling framework.ResultsSimulation studies under various scenarios are carried out to compare the performances of our proposed method with some commonly used methods. In the real data applications, we use our proposed method to identify G-E interactions related to the clinical survival outcomes of patients with head and neck squamous cell carcinoma, and esophageal carcinoma in The Cancer Genome Atlas clinical survival genetic data, and further establish corresponding survival prediction models. Both simulation and real data studies show that our method performs well and outperforms existing methods in the G-E interaction selection, effect estimation, and survival prediction accuracy.ConclusionsThe OGS approach is useful for selecting important environmental factors, genes and G-E interactions in the ultra-high dimensional feature space. The prediction ability of OGS with the Lasso penalty is better than existing methods. The same idea of the OGS approach can apply to other outcome models, such as the proportional odds survival time model, the logistic regression model for binary outcomes, and the multinomial logistic regression model for multi-class outcomes.

Determinant factors behind changes in health-seeking behaviour before and after implementation of universal health coverage in Indonesia

BackgroundThe health insurance system in Indonesia was transformed in 2014 to achieve universal health coverage (UHC). The effective implementation of essential primary health services through UHC has resulted in efficient healthcare utilisation, which is reflected in the health-seeking behaviour of the community. Our study aimed to examine the changes in health-seeking behaviour before and after the implementation of UHC in Indonesia and to identify what factors determine these changes.MethodsWe conducted a retrospective cohort study using the recall method and data collected through questionnaire-based interviews in Bandung, Indonesia. We used a two-step sampling technique—randomised sampling and purposive sampling, and a total of 579 respondents with acute or chronic episodes were recruited. {chi }^{2} tests were used to identify the association between factors. Difference in difference model and a logistic regression model for binary outcomes were used to estimate the effect of the implementation of UHC on the health-seeking behaviour.ResultsUtilisation of public health facilities increased significantly after implementation of UHC, from 34.9% to 65.4% among the respondents with acute episodes and 33.7% to 65.8% among those with chronic episodes. The odds of respondents going to health facilities when they developed an acute episode increased after the implementation of UHC (OR = 1.22, p = 0.05; AOR = 1.42, p < 0.001). For respondents experiencing chronic episodes, the implementation of UHC increased the odds ratio (OR = 1.74, p < 0.001; AOR = 1.64, p < 0.001) that they would use health facilities. Five years after the implementation of UHC, we still found respondents who did not have health insurance (26 and 19 respondents among those with acute episode and chronic episode, respectively).ConclusionsThe effect of the implementation of UHC seemed greater for those experiencing chronic episodes than for those with an acute episode. Although the implementation of UHC has improved utilisation of public health facilities, the presence of people who are not covered by health insurance is a potential problem that could threaten future improvements in healthcare access and utilisation.

A joint fairness model with applications to risk predictions for underrepresented populations.

In data collection for predictive modeling, underrepresentation of certain groups, based on gender, race/ethnicity, or age, may yield less accurate predictions for these groups. Recently, this issue of fairness in predictions has attracted significant attention, as data-driven models are increasingly utilized to perform crucial decision-making tasks. Existing methods to achieve fairness in the machine learning literature typically build a single prediction model in a manner that encourages fair prediction performance for all groups. These approaches have two major limitations: (i) fairness is often achieved by compromising accuracy for some groups; (ii) the underlying relationship between dependent and independent variables may not be the same across groups. We propose a joint fairness model (JFM) approach for logistic regression models for binary outcomes that estimates group-specific classifiers using a joint modeling objective function that incorporates fairness criteria for prediction. We introduce an accelerated smoothing proximal gradient algorithm to solve the convex objective function, and present the key asymptotic properties of the JFM estimates. Through simulations, we demonstrate the efficacy of the JFM in achieving good prediction performance and across-group parity, in comparison with the single fairness model, group-separate model, and group-ignorant model, especially when the minority group's sample size is small. Finally, we demonstrate the utility of the JFM method in a real-world example to obtain fair risk predictions for underrepresented older patients diagnosed with coronavirus disease 2019 (COVID-19).

Eldercare services for people with and without a dementia diagnosis: an analysis of Swedish registry data

BackgroundThe growing number of people living with dementia (PlwD) implies an increase in the demand for eldercare services in Sweden like in many other countries. Few studies have analyzed the use of eldercare services for PlwD. The aim of the present study is to investigate the association between demographic factors (age, sex, cohabiting status) and the use of municipal eldercare services (including both home care and residential care) for older adults with dementia compared to older adults without dementia in Sweden.MethodsThis study used several nationwide Swedish registers targeting all individuals aged 65 and above living in Sweden in 2014 and still alive 31st of March 2015 (n = 2,004,409). The primary outcomes variables were different types of eldercare service, and all participants were clustered based on age, sex, cohabiting status, and dementia diagnosis. In addition to descriptive statistics, we performed multivariate logistic regression models for binary outcomes and linear regression models for continuous outcomes.ResultsResults showed that (1) older age is a significantly strong predictor for the use of eldercare services, although PlwD start using eldercare at an earlier age compared with people without dementia; (2) women tend to receive more eldercare services than men, especially in older age, although men with dementia who live alone are more likely than women living alone to receive eldercare; (3) having a dementia diagnosis is a strong predictor for receiving eldercare. However, it was also found that a substantial proportion of men and women with dementia did not receive any eldercare services.ConclusionsWe found that people with a dementia diagnosis use more as well as start to use eldercare services at an earlier age than people without dementia. However, further research is needed to investigate why a substantial part of people with a dementia diagnosis does not have any eldercare at all and what the policy implications of this might be.

Accuracy Measure of Separate and Joint Modelling for a Correlated Binary Outcome: The Case Study of Mother Education and Immunization in Bangladesh

Joint modelling is a statistical approach that is used to analyze correlated data when two or more outcome variables are correlated. By joint modeling, we refer to the simultaneous analysis of two or more different response variables from the same individual. But in a separate model, it is unable to measure the effect of covariate simultaneously. This article focuses on separate and joint modelling for correlated discrete data, including logistic regression models for binary outcomes. Since most of the women are illiterate in Bangladesh and most of the people are living in urban areas, as a result, most of the women are not aware of immunization. But an educated mother is always aware of her child's health which is dependent on immunization. Therefore, mother education and immunization are interdependent. We jointly address the effect of maternal education and immunization. Joint modeling of these two outcomes is appropriate because mother education helps raise awareness of the child's health and immunization is the prevention of various diseases for the child's health. We also identified factors influencing maternal education and immunization among women in Bangladesh. By jointly modelling we found the correlation between maternal education and immunization and the most important contributing factor. The joint model removes a less significant impact of covariates as opposed to separate models. These findings further suggested that the simultaneous impact of correlated outcomes can be adequately addressed between different responses, which is overestimated or underestimated when examined separately.

Mapping chronic disease prevalence based on medication use and socio-demographic variables: an application of LASSO on administrative data sources in healthcare in the Netherlands

BackgroundPolicymakers generally lack sufficiently detailed health information to develop localized health policy plans. Chronic disease prevalence mapping is difficult as accurate direct sources are often lacking. Improvement is possible by adding extra information such as medication use and demographic information to identify disease. The aim of the current study was to obtain small geographic area prevalence estimates for four common chronic diseases by modelling based on medication use and socio-economic variables and next to investigate regional patterns of disease.MethodsAdministrative hospital records and general practitioner registry data were linked to medication use and socio-economic characteristics. The training set (n = 707,021) contained GP diagnosis and/or hospital admission diagnosis as the standard for disease prevalence. For the entire Dutch population (n = 16,777,888), all information except GP diagnosis and hospital admission was available. LASSO regression models for binary outcomes were used to select variables strongly associated with disease. Dutch municipality (non-)standardized prevalence estimates for stroke, CHD, COPD and diabetes were then based on averages of predicted probabilities for each individual inhabitant.ResultsAdding medication use data as a predictor substantially improved model performance. Estimates at the municipality level performed best for diabetes with a weighted percentage error (WPE) of 6.8%, and worst for COPD (WPE 14.5%)Disease prevalence showed clear regional patterns, also after standardization for age.ConclusionAdding medication use as an indicator of disease prevalence next to socio-economic variables substantially improved estimates at the municipality level. The resulting individual disease probabilities could be aggregated into any desired regional level and provide a useful tool to identify regional patterns and inform local policy.

Trimester effects of source-specific PM10 on birth weight outcomes in the Avon Longitudinal Study of Parents and Children (ALSPAC)

BackgroundEvidence suggests that exposure to particulate matter with aerodynamic diameter less than 10 μm (PM10) is associated with reduced birth weight, but information is limited on the sources of PM10 and exposure misclassification from assigning exposures to place of residence at birth.MethodsTrimester and source-specific PM10 exposures (PM10 from road source, local non-road source, and total source) in pregnancy were estimated using dispersion models and a full maternal residential history for 12,020 births from the Avon longitudinal study of parents and children (ALSPAC) cohort in 1990–1992 in the Bristol area. Information on birth outcomes were obtained from birth records. Maternal sociodemographic and lifestyle factors were obtained from questionnaires. We used linear regression models for continuous outcomes (birth weight, head circumference (HC), and birth length (BL) and logistic regression models for binary outcomes (preterm birth (PTB), term low birth weight (TLBW) and small for gestational age (SGA)). Sensitivity analysis was performed using multiple imputation for missing covariate data.ResultsAfter adjustment, interquartile range increases in source specific PM10 from traffic were associated with 17 to 18% increased odds of TLBW in all pregnancy periods. We also found odds of TLBW increased by 40% (OR: 1.40, 95%CI: 1.12, 1.75) and odds of SGA increased by 18% (OR: 1.18, 95%CI: 1.05, 1.32) per IQR (6.54 μg/m3) increase of total PM10 exposure in the third trimester.ConclusionThis study adds to evidence that maternal PM10 exposures affect birth weight, with particular concern in relation to exposures to PM10 from road transport sources; results for total PM10 suggest greatest effect in the third trimester. Effect size estimates relate to exposures in the 1990s and are higher than those for recent studies – this may relate to reduced exposure misclassification through use of full residential history information, changes in air pollution toxicity over time and/or residual confounding.

Are changes in positive mental health associated with increased likelihood of depression over a two year period? A test of the mental health promotion and protection hypotheses

This paper investigates the mental health promotion and protection (MHPP) model of reducing depression. Data are from the Clergy Health Initiative Longitudinal Survey of United Methodist ministers in North Carolina that included the Mental Health Continuum Short Form (MHC-SF) for positive mental health and the Patient Health Questionnaire (PHQ-9) for depression in 2014 and 2016 (N = 955). The promotion hypothesis predicts reduced risk of depression in 2016 among clergy whose mental health increased to flourishing and the increased risk of depression in 2016 for clergy who stayed not flourishing. The protection hypothesis predicts increased risk of depression in 2016 for clergy who were flourishing in 2014 but went down to ‘not flourishing’ in 2016. The reference group is clergy who stayed flourishing. We used modified Poisson regression models for binary outcomes to estimate Prevalence Ratios (PR) and to estimate Incidence Rate Ratios (IRR) of depression in 2016 associated with changes in mental health status. Results support both hypotheses. Compared to clergy who stayed flourishing, clergy who improved to flourishing were as likely, while clergy who stayed not flourishing were nearly seven times more likely, to have depression in 2016. Clergy who declined to not flourishing were six times more likely to have depression in 2016 compared to those who stayed flourishing. Similar patterns were observed when the sample was restricted to clergy without depression in 2014. These findings suggest focusing on MHPP as a complementary approach to treatment to reduce the incidence, prevalence and burden of depression.

Semiparametric isotonic regression analysis for risk assessment under nested case-control and case-cohort designs.

Two-phase sampling designs, including nested case-control and case-cohort designs, are frequently utilized in large cohort studies involving expensive biomarkers. To analyze data from two-phase designs with a binary outcome, parametric models such as logistic regression are often adopted. However, when the model assumptions are not valid, parametric models may lead to biased estimation and risk evaluation. In this paper, we propose a robust semiparametric regression model for binary outcomes and an easy-to-implement computational procedure that combines the pool-adjacent violators algorithm with inverse probability weighting. The asymptotic properties are established, including consistency and the convergence rate. Simulation studies show that the proposed method performs well and is more robust than logistic regression methods. We demonstrate the application of the proposed method to real data from the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial.

Self-harm in midlife: analysis using data from the Multicentre Study of Self-harm in England.

In England suicide rates are highest in midlife (defined as age 40-59). Despite a strong link with suicide there has been little focus on self-harm in this age group.AimTo describe characteristics and treatment needs of people in midlife who present to hospital following self-harm. Data from the Multicentre Study of Self-harm in England were used to examine rates over time and characteristics of men and women who self-harm in midlife. Data (2000-2013) were collected via specialist assessments or hospital records. Trends were assessed by negative binomial regression models. Comparative analysis used logistic regression models for binary outcomes. Repetition and suicide mortality were assessed by Cox proportional hazards models. A quarter of self-harm presentations were made by people in midlife (n = 24 599, 26%). Incidence rates increased over time in men, especially after 2008 (incidence rate ratio [IRR] 1.07, 95% CI 1.02-1.12, P &lt; 0.01), and were positively correlated with national suicide incidence rates (r = 0.52, P = 0.05). Rates in women remained relatively stable (IRR 1.00, 95% CI 1.00-1.02, P = 0.39) and were not correlated with suicide. Alcohol use, unemployment, housing and financial factors were more common in men; whereas indicators of poor mental health were more common in women. In men and women 12-month repetition was 25%, and during follow-up 2.8% of men and 1.2% of women died by suicide. Self-harm in midlife represents a key target for intervention. Addressing underlying issues, alcohol use and economic factors may help prevent further self-harm and suicide.Declaration of interestK.H. and N.K. are members of the Department of Health's National Suicide Prevention Advisory Group. N.K. chaired the National Institute for Health and Care Excellence (NICE) guideline development group for the longer-term management of self-harm and the NICE Topic Expert Group which developed the quality standards for self-harm services. N.K. also chairs the NICE guideline committee for the management of depression. All other authors declare no conflict of interest.

Using Predictions and Marginal Effects to Compare Groups in Regression Models for Binary Outcomes

Methods for group comparisons using predicted probabilities and marginal effects on probabilities are developed for regression models for binary outcomes. Unlike approaches based on the comparison of regression coefficients across groups, the methods we propose are unaffected by the scalar identification of the coefficients and are expressed in the natural metric of the outcome probability. While we develop our approach using binary logit with two groups, we consider how our interpretive framework can be used with a broad class of regression models and can be extended to any number of groups.

Cognitive Change and Driving Behavior among Older Drivers

Per vehicle miles traveled, older adults have a high fatal crash rate. One factor is dementia. This study aims to assess how differences in cognition affect driving behavior among older drivers. We analyzed data from the Health and Retirement Study. Our study used cognition, demographics, and driving behavior from 2006 to 2014 for respondents aged 65 and above. Three levels of driving behavior were measured: whether the individual could drive, whether they had driven in the past month, and whether they drove long distances. Cognitive function was measured through the Telephone Interview for Cognitive Status. Additionally, individuals were coded as having no diagnosis of dementia, a diagnosis within 2 years, or a diagnosis more than 2 years previously. We estimated the likelihood of each driving behavior in association with cognition using a modified Poisson regression model for binary outcomes. Among respondents ( N = 16,061), 79% could drive. Of these, 93% had driven in the past month, and of these, 64% drove long distances. Compared with no impairment, mild impairment was associated with a significant 12% decrease in probability of being able to drive, an 8% decrease in driving within the past month, and a 24% decrease in driving long distances. The decrease was larger among those with severe impairment. Results were in a similar direction and strength comparing individuals without dementia with individuals 0 to 2 years after diagnosis, and to more than 2 years after diagnosis. A strong positive association exists between lower cognition and lower driving exposure.

Calidad del empleo y migración interna en Colombia en 2015

En el presente estudio se calculó un Índice Compuesto de Calidad del Empleo (ICCE) para los de-partamentos de Colombia, distinguiendo entre zonas urbanas y rurales. Para ello, se utilizaron datos de la Gran Encuesta Integrada de Hogares (GEIH) de los años 2014 y 2015. Posteriormente, se em-plearon modelos de regresión de respuesta binaria (logit y probit) para determinar cómo los subín-dices referentes a las dimensiones del ICCE explican la decisión de migrar de la PEA de Colombia a través del país. Los resultados indican que el empleo de calidad predomina en las zonas urbanas, destacándose el liderazgo de Bogotá y Medellín. Por otro lado, los datos obtenidos señalan que tanto el ICCE como sus componentes fueron significativos al momento de explicar los patrones migratorios internos de Colombia.

Bounds on Concordance-Based Validation Statistics in Regression Models for Binary Responses

Association measures based on concordance, such as Kendall’s tau, Somers’ delta or Goodman and Kruskal’s gamma are often used to measure explained variations in regression models for binary outcomes. As responses only assume values in {0, 1}, these association measures are constrained, which makes their interpretation more difficult as a relatively small value may in fact strongly support the fitted model. In this paper, we derive the set of attainable values for concordance-based association measures in this setting so that the closeness to the best-possible fit can be properly assessed.

The employment history of people with and without asthma

Introduction People with asthma may have a greater likelihood of leaving employment due to ill health, spending more time in unemployment and changing jobs more frequently because of their disease. Some studies have reported that people with childhood asthma were disadvantaged in later employment. The aim of this analysis is to investigate the effects of asthma on employment history from early-to mid-adult life (age 42) in a British birth cohort. Methods Members of the National Child Development Study (NCDS) cohort with information on asthma and an employment history from ages 16 to 42 were included (n=9881). Asthma was defined as childhood asthma (0-16), severe childhood asthma, current asthma at age 16 and adult onset asthma. Outcomes were ever having a period out of employment, total periods of employment and unemployment and total time out of employment. The associations between asthma and the outcomes were assessed separately for men and women using Mann-Whitney or Kruskal Wallis tests, with significant results explored using a logistic regression model for binary outcomes and a negative binomial regression model (with a zero inflated model for outcomes with large numbers of zeros) for count data. Results By age 42, 43% of the cohort had experienced at least one period out of employment and the median number of employment periods was 6, with women reporting on average one more job by age 42 than men. Women reporting adult onset asthma were likely to have had more periods of employment when compared to women who never had asthma (p=0.003). For both men and women with adult onset asthma there was a doubling of the risk of ever being out of employment due to ill health compared to those without asthma (men OR 1.94 [95% CI 1.24-3.04], women OR 2.00 [95% CI 1.43-2.80]). Conclusion There is evidence that adult onset (but not childhood) asthma affects employment history, particularly for women. It is not known whether this reflects the experience of more recent cohorts.