Prenatal exposure to particulate matter and infant birth outcomes: Evidence from a population-wide database.

There are growing concerns about the impact of pollution on maternal and infant health. Despite an extensive correlational literature, observational studies which adopt methods that seek to address potential biases due to unmeasured confounders draw mixed conclusions. Using a population database of births in Northern Ireland (NI) linked to localized geographic information on pollution in mothers' postcodes (zipcodes) of residence during pregnancy, we examine whether prenatal exposure to PM2.5 is associated with a comprehensive range of birth outcomes, including placental health. Overall, we find little evidence that particulate matter is related to infant outcomes at the pollution levels experienced in NI, once we implement a mother fixed effects approach that accounts for time-invariant factors. This contrasts with strong associations in models that adjust for observed confounders but without fixed effects. While reducing ambient air pollution remains an urgent public health priority globally, our results imply that further improvements in short-run levels of prenatal PM2.5 exposure in a relatively low-pollution, higher-income country context, are unlikely to impact on birth outcomes at the population level.

Exploring the associations between auditory hallucinations and psychopathological experiences in 10,933 patient narratives: moving beyond diagnostic categories and surveys.

BACKGROUND: Previous research suggests that auditory hallucinations are prevalent within both the clinical and general populations. Yet, we know little about how these phenomena are associated with other psychopathology symptoms and experiences. The current study aids investigations towards preventing, predicting and more effectively responding to such distressing occurrences. There have been substantial efforts in the literature to propose models of auditory hallucination and attempts to verify them. However, many of these studies used survey methods that restrict the person's responses to a set of pre-defined criteria or experiences and do not allow exploration of potential important other symptoms beyond them. This is the first study to explore the correlates of auditory hallucination using a qualitative dataset consisting of unrestricted responses of patients about their lived experiences with mental illness. METHOD: The study used a dataset consisting of 10,933 narratives from patients diagnosed with mental illnesses. For analysis, the study used correlation on the text-based data. This approach is an alternative to the knowledge-based approach where experts manually read the narratives and infer the rules and relationships from the dataset. RESULT: This study found at least 8 correlates of auditory hallucination (small correlation coefficients), with the unusual ones being "pain." The study also found that auditory hallucinations were independent of obsessive thoughts and compulsive behaviours, and dissociation, in contrast with the literature. CONCLUSION: This study presents an innovative approach to explore the possible associations between symptoms without the restrictions of (or outside the confines of) traditional diagnostic categories. The study exemplified this by finding the correlates of auditory hallucination. However, any other symptom or experience of interest can be studied similarly. Potential future directions of these findings are discussed in the context of mental healthcare screening and treatment.

What can we learn about the psychiatric diagnostic categories by analysing patients' lived experiences with Machine-Learning?

BACKGROUND: To deliver appropriate mental healthcare interventions and support, it is imperative to be able to distinguish one person from the other. The current classification of mental illness (e.g., DSM) is unable to do that well, indicating the problem of diagnostic heterogeneity between disorders (i.e., the disorder categories have many common symptoms). As a result, the same person might be diagnosed with two different disorders by two independent clinicians. We argue that this problem might have resulted because these disorders were created by a group of humans (APA taskforce members) who relied on more intuition and consensus than data. Literature suggests that human-led decisions are prone to biases, group-thinking, and other factors (such as financial conflict of interest) that can enormously influence creating diagnostic and treatment guidelines. Therefore, in this study, we inquire that if we prevent such human intervention (and thereby their associated biases) and use Artificial Intelligence (A.I.) to form those disorder structures from the data (patient-reported symptoms) directly, then can we come up with homogenous clusters or categories (representing disorders/syndromes: a group of co-occurring symptoms) that are adequately distinguishable from each other for them to be clinically useful. Additionally, we inquired how these A.I.-created categories differ (or are similar) from human-created categories. Finally, to the best of our knowledge, this is the first study, that demonstrated how to use narrative qualitative data from patients with psychopathology and group their experiences using an A.I. Therefore, the current study also attempts to serve as a proof-of-concept. METHOD: We used secondary data scraped from online communities and consisting of 10,933 patients' narratives about their lived experiences. These patients were diagnosed with one or more DSM diagnoses for mental illness. Using Natural Language Processing techniques, we converted the text data into a numeric form. We then used an Unsupervised Machine Learning algorithm called K-Means Clustering to group/cluster the symptoms.  RESULTS: Using the data mining approach, the A.I. found four categories/clusters formed from the data. We presented ten symptoms or experiences under each cluster to demonstrate the practicality of application and understanding. We also identified the transdiagnostic factors and symptoms that were unique to each of these four clusters. We explored the extent of similarities between these clusters and studied the difference in data density in them. Finally, we reported the silhouette score of + 0.046, indicating that the clusters are poorly distinguishable from each other (i.e., they have high overlapping symptoms). DISCUSSION: We infer that whether humans attempt to categorise mental illnesses or an A.I., the result is that the categories of mental disorders will not be unique enough to be able to distinguish one service seeker from another. Therefore, the categorical approach of diagnosing mental disorders can be argued to fall short of its purpose. We need to search for a classification system beyond the categorical approaches even if there are secondary merits (such as ease of communication and black-and-white (binary) decision making). However, using our A.I. based data mining approach had several meritorious findings. For example, we found that some symptoms are more exclusive or unique to one cluster. In contrast, others are shared by most other clusters (i.e., identification of transdiagnostic experiences). Such differences are interesting objects of inquiry for future studies. For example, in clear contrast to the traditional diagnostic systems, while some experiences, such as auditory hallucinations, are present in all four clusters, others, such as trouble with eating, are exclusive to one cluster (representing a syndrome: a group of co-occurring symptoms). We argue that trans-diagnostic conditions (e.g., auditory hallucinations) might be prime targets for symptom-level interventions. For syndrome-level grouping and intervention, however, we argue that exclusive symptoms are the main targets. CONCLUSION: Categorical approach to mental disorders is not a way forward because the categories are not unique enough and have several shared symptoms. We argue that the same symptoms can be present in more than one syndrome, although dimensionally different. However, we need additional studies to test this hypothesis. Future directions and implications were discussed.

Measuring diagnostic heterogeneity using text-mining of the lived experiences of patients.

BACKGROUND: The diagnostic system is fundamental to any health discipline, including mental health, as it defines mental illness and helps inform possible treatment and prognosis. Thus, the procedure to estimate the reliability of such a system is of utmost importance. The current ways of measuring the reliability of the diagnostic system have limitations. In this study, we propose an alternative approach for verifying and measuring the reliability of the existing system. METHODS: We perform Jaccard's similarity index analysis between first person accounts of patients with the same disorder (in this case Major Depressive Disorder) and between those who received a diagnosis of a different disorder (in this case Bulimia Nervosa) to demonstrate that narratives, when suitably processed, are a rich source of data for this purpose. We then analyse 228 narratives of lived experiences from patients with mental disorders, using Python code script, to demonstrate that patients with the same diagnosis have very different illness experiences. RESULTS: The results demonstrate that narratives are a statistically viable data resource which can distinguish between patients who receive different diagnostic labels. However, the similarity coefficients between 99.98% of narrative pairs, including for those with similar diagnoses, are low (< 0.3), indicating diagnostic Heterogeneity. CONCLUSIONS: The current study proposes an alternative approach to measuring diagnostic Heterogeneity of the categorical taxonomic systems (e.g. the Diagnostic and Statistical Manual, DSM). In doing so, we demonstrate the high Heterogeneity and limited reliability of the existing system using patients' written narratives of their illness experiences as the only data source. Potential applications of these outputs are discussed in the context of healthcare management and mental health research.

Is casual employment in Australia bad for workers' health?

OBJECTIVES: This paper assessed the impact of working in casual employment, compared with permanent employment, on eight health attributes that make up the 36-Item Short Form (SF-36) Health Survey, separately by sex. The mental health impacts of casual jobs with irregular hours over which the worker reports limited control were also investigated. METHODS: Longitudinal data from the Household, Income and Labour Dynamics in Australia Survey, over the period 2001-2018, were used to investigate the relationship between the eight SF-36 subscales and workers' employment contract type. Individual, household and job characteristic confounders were included in dynamic panel data regression models with correlated random effects. RESULTS: For both men and women, health outcomes for casual workers were no worse than for permanent workers for any of the eight SF-36 health attributes. For some health attributes, scores for casual workers were higher (ie, better) than for permanent workers (role physical: men: ß=1.15, 95% CI 0.09 to 2.20, women: ß=1.79, 95% CI 0.79 to 2.80; bodily pain: women: ß=0.90, 95% CI 0.25 to 1.54; vitality: women: ß=0.65, 95% CI 0.13 to 1.18; social functioning: men: ß=1.00, 95% CI 0.28 to 1.73); role emotional: men: ß=1.81, 95% CI 0.73 to 2.89, women: ß=1.24, 95% CI 0.24 to 2.24). Among women (but not men), mental health and role emotional scores were lower for irregular casual workers than for regular permanent workers but not statistically significantly so. CONCLUSIONS: This study found no evidence that casual employment in Australia is detrimental to self-assessed worker health.

Estimating the impact of disability onset on employment.

This paper examines the impact of disability onset on the probability of employment using an underexplored longitudinal dataset for Britain. It contrasts estimates based on a control group drawn from those not experiencing disability onset - a common approach in the literature - with estimates based on a control group drawn from those who do experience disability onset, but one year after the treatment group. Compared to the non-disabled control group, the control group of later-onsetters is likely to be more similar to the treatment group in terms of unobservables, with the resulting estimates therefore more plausibly interpreted as causal. Using this control group we estimate that the probability of employment falls by 11 percentage points in the year of disability onset. The equivalent estimate using the control group drawn from those not experiencing onset is about fifty percent larger. The employment effects of disability onset are also shown to be larger for those with lower qualification levels, consistent with weaker attachment to the labour market.

Evaluation of the Effectiveness and Cost-Effectiveness of Personalized Surveillance After Colorectal Adenomatous Polypectomy.

Lifetime risk of developing colorectal cancer is 5%, and 5-year survival at early stage is 92%. Individuals with precancerous lesions removed at primary screening are typically recommended surveillance colonoscopy. Because greater benefits are anticipated for those with higher risk of colorectal cancer, scope for risk-specific surveillance recommendations exists. This review assesses published cost-effectiveness estimates of postpolypectomy surveillance to consider the potential for personalized recommendations by risk group. Meta-analyses of incidence of advanced neoplasia postpolypectomy for low-risk cases were comparable to those without adenoma, with both rates under the lifetime risk of 5%. This group may not benefit from intensive surveillance, which risks unnecessary harm and inefficient use of often scarce colonoscopy capacity. Therefore, greater personalization through deintensified strategies for low-risk individuals could be beneficial. The potential for noninvasive testing, such as fecal immunochemical tests, combined with primary prevention or chemoprevention may reserve colonoscopy for targeted use in personalized risk-stratified surveillance. This review appraised evidence supporting a program of personalized surveillance in patients with colorectal adenoma according to risk group and compared the effectiveness of surveillance colonoscopy with alternative prevention strategies. It assessed trade-offs among costs, benefits, and adverse effects that must be considered in a decision to adopt or reject personalized surveillance.

Cross-section and panel estimates of peer effects in early adolescent cannabis use: With a little help from my 'friends once removed'.

Peer effects in adolescent cannabis are difficult to estimate, due in part to the lack of appropriate data on behaviour and social ties. This paper exploits survey data that have many desirable properties and have not previously been used for this purpose. The data set, collected from teenagers in three annual waves from 2002 to 2004 contains longitudinal information about friendship networks within schools (N = 5020). We exploit these data on network structure to estimate peer effects on adolescents from their nominated friends within school using two alternative approaches to identification. First, we present a cross-sectional instrumental variable (IV) estimate of peer effects that exploits network structure at the second degree, i.e. using information on friends of friends who are not themselves ego's friends to instrument for the cannabis use of friends. Second, we present an individual fixed effects estimate of peer effects using the full longitudinal structure of the data. Both innovations allow a greater degree of control for correlated effects than is commonly the case in the substance-use peer effects literature, improving our chances of obtaining estimates of peer effects than can be plausibly interpreted as causal. Both estimates suggest positive peer effects of non-trivial magnitude, although the IV estimate is imprecise. Furthermore, when we specify identical models with behaviour and characteristics of randomly selected school peers in place of friends', we find effectively zero effect from these 'placebo' peers, lending credence to our main estimates. We conclude that cross-sectional data can be used to estimate plausible positive peer effects on cannabis use where network structure information is available and appropriately exploited.

From substance use to homelessness or vice versa?

Homelessness is associated with substance use, but whether substance use precedes or follows homelessness is unclear. We investigate the nature of the relationship between homelessness and substance use using data from the unique Australian panel dataset Journeys Home collected in 4 surveys over the period from October 2011 to May 2013. Our data refer to 1325 individuals who were homeless or at risk of becoming homeless. We investigate dynamics in homelessness and substance use over the survey period. We find that the two are closely related: homeless individuals are more likely to be substance users and substance users are more likely to be homeless. These relationships, however, are predominantly driven by observed and unobserved individual characteristics which cause individuals to be both more likely to be homeless and to be substance users. Once we take these personal characteristics into account it seems that homelessness does not affect substance use, although we cannot rule out that alcohol use increases the probability that an individual becomes homeless. These overall relationships also hide some interesting heterogeneity by 'type' of homelessness.

Estimates of peer effects in adolescent smoking across twenty six European countries.

Although it is widely believed that one of the key factors influencing whether an adolescent smokes or not is the smoking behaviour of his or her peers, empirical evidence on the magnitude of such peer effects, and even on their existence, is mixed. This existing evidence comes from a range of studies using a variety of country-specific data sources and a variety of identification strategies. This paper exploits a rich source of individual level, school-based, survey data on adolescent substance use across countries--the 2007 European Schools Survey Project on Alcohol and Other Drugs--to provide estimates of peer effects between classmates in adolescent smoking for 75,000 individuals across 26 European countries, using the same methods in each case. The results suggest statistically significant peer effects in almost all cases. These peer effects estimates are large: on average across countries, the probability that a 'typical' adolescent smokes increases by between .31 and .38 percentage points for a one percentage point increase in the proportion of classmates that smoke. Further, estimated peer effects in adolescent smoking are stronger intra-gender than inter-gender. They also vary across countries: in Belgium, for example, a one percentage point increase in reference group smoking is associated with a .16 to .27 percentage point increase in own smoking probability; in The Netherlands the corresponding increase is between .42 and .59 percentage points.

Choice of contracts in the British National Health Service: an empirical study.

Following major reforms of the British National Health Service (NHS) in 1990, the roles of purchasing and providing health services were separated, with the relationship between purchasers and providers governed by contracts. Using a mixed multinomial logit analysis, we show how this policy shift led to a selection of contracts that is consistent with the predictions of a simple model, based on contract theory, in which the characteristics of the health services being purchased and of the contracting parties influence the choice of contract form. The paper thus provides evidence in support of the practical relevance of theory in understanding health care market reform.