Exploring the relationships between pathogen-specific prenatal infections requiring inpatient admission and domains of offspring behaviour at age 5

BACKGROUND

Research exploring the relationship between prenatal infection and child behavioural outcomes would benefit from further studies utilising full-population samples with the scale to investigate specific infections and to employ robust designs. We tested the association among several common infections requiring inpatient admission during and after pregnancy with a range of childhood behavioural outcomes, to determine whether any negative impact was specific to the period of foetal development.

METHODS

The sample included all mother-offspring pairs from the Australian state of New South Wales (NSW) for whom the child commenced their first year of full-time schooling in 2009 (~age 5 years; n = 77,302 offspring), with records linked across four health administrative data sets including the NSW perinatal data collection (PDC), the NSW admitted patient data collection (APDC) and the NSW component of the 2009 Australian Early Development Census (AEDC). Multivariable linear regression was used to test associations between a number of infections requiring inpatient admission during and after pregnancy with a range of teacher assessed behavioural outcomes.

RESULTS

Associations specific to the prenatal period were only found for streptococcus A although this would need to be reproduced in external samples given the low prevalence. Otherwise, 12 out of 15 selected infections either showed no association prenatally or also demonstrated associations in the 12 months after pregnancy. For example, prenatal hepatitis C, influenza and urinary E. coli infections were associated with lower scores of several domains of childhood behaviour, but even stronger associations were found when these same maternal infections occurred after pregnancy.

CONCLUSIONS

The prenatal infections we tested appeared not to impact childhood behaviour by altering foetal neurodevelopment. Rather, the strong associations we found among infections occurring during and after pregnancy point to either residual socioeconomic/lifestyle factors or a shared familial/genetic liability between infections and behavioural problems.

The Massachusetts public health data warehouse and the opioid epidemic: A qualitative study of perceived strengths and limitations for advancing research

•

The Massachusetts Public Health Data Warehouse is a public health innovation.

•

We assessed the utility of this big data resource for research on the opioid epidemic.

•

Big data have many advantages and limitations for opioid epidemic research.

•

Findings can help to maximize the advantages of big data and avoid inappropriate use.

•

Lessons learned can aid other states to establish big data for public health.

Due to the opioid overdose epidemic, Massachusetts created a Public Health Data Warehouse, encompassing individually-linked administrative data on most of the population as provided by more than 20 systems. As others seek to assemble and mine big data on opioid use, there is a need to consider its research utility. To identify perceived strengths and limitations of administrative big data, we collected qualitative data in 2019 from 39 stakeholders with knowledge of the Massachusetts Public Health Data Warehouse. Perceived strengths included the ability to: (1) detect new and clinically significant relationships; (2) observe treatments and services across institutional boundaries, broadening understanding of risk and protective factors, treatment outcomes, and intervention effectiveness; (3) use geographic-specific lenses for community-level health; (4) conduct rigorous “real-world” research; and (5) generate impactful findings that legitimize the scope and impacts of the opioid epidemic and answer urgent questions. Limitations included: (1) oversimplified information and imprecise measures; (2) data access and analysis challenges; (3) static records and substantial lag times; and (4) blind spots that bias or confound results, mask upstream or root causes, and contribute to incomplete understanding. Using administrative big data to conduct research on the opioid epidemic offers advantages but also has limitations which, if unrecognized, may undermine its utility. Findings can help researchers to capitalize on the advantages of big data, and avoid inappropriate uses, and aid states that are assembling big data to guide public health practice and policy.

Predicting neonatal respiratory distress syndrome and hypoglycaemia prior to discharge: Leveraging health administrative data and machine learning

OBJECTIVES

A major challenge for hospitals and clinicians is the early identification of neonates at risk of developing adverse conditions. We develop a model based on routinely collected administrative data, which accurately predicts two common disorders among early term and preterm (<39 weeks) neonates prior to discharge.

STUDY DESIGN

The data included all inpatient live births born prior to 39 weeks (n = 154,755) occurring in the Australian state of Queensland between January 2009 and December 2015. Predictor variables included all maternal data captured in administrative records from the beginning of gestation up to, and including, the delivery, as well as neonatal data recorded at the delivery. Gradient boosted trees were used to predict neonatal respiratory distress syndrome and hypoglycaemia prior to discharge, with model performance benchmarked against a logistic regression models.

RESULTS

The gradient boosted trees model achieved very high discrimination for respiratory distress syndrome [AUC = 0.923, 95% CI (0.917, 0.928)] and good discrimination for hypoglycaemia [AUC = 0.832, 95% CI (0.827, 0.837)] in the validation data, as well as outperforming the logistic regression models.

CONCLUSION

Our study suggests that routinely collected health data have the potential to play an important role in assisting clinicians to identify neonates at risk of developing selected disorders shortly after birth. Despite achieving high levels of discrimination, many issues remain before such models can be implemented in practice, which we discuss in relation to our findings.

Predicting postpartum psychiatric admission using a machine learning approach

AIMS

The accurate identification of mothers at risk of postpartum psychiatric admission would allow for preventive intervention or more timely admission. We developed a prediction model to identify women at risk of postpartum psychiatric admission.

METHODS

Data included administrative health data of all inpatient live births in the Australian state of Queensland between January 2009 and October 2014. Analyses were restricted to mothers with one or more indicator of mental health problems during pregnancy (n = 75,054 births). The predictors included all maternal data up to and including the delivery, and neonatal data recorded at delivery. We used multiple machine learning methods to predict hospital admission in the 12 months following delivery in which the primary diagnosis was recorded as an ICD-10 psychotic, bipolar or depressive disorders.

RESULTS

The boosted trees algorithm produced the best performing model, predicting postpartum psychiatric admission in the validation data with good discrimination [AUC = 0.80; 95% CI = (0.76, 0.83)] and achieving good calibration. This model outperformed benchmark logistic regression model and an elastic net model. In addition to indicators of maternal metal health history, maternal and neonatal anthropometric measures and social/lifestyle factors were strong predictors.

CONCLUSION

Our results indicate the potential of a big data approach when aiming to identify mothers at risk of postpartum psychiatric admission. Mothers at risk could be followed-up and supported after neonatal discharge to either remove the need for admission or facilitate more timely admission.

Inter- and intra-provincial variation in screen-detected breast cancer across five Canadian provinces: a CanIMPACT study

OBJECTIVE

Breast cancer screening aims to identify cancers in early stages when prognosis is better and treatments less invasive. We describe inter- and intra-provincial variation in the percentage of screen-detected cases under publicly funded healthcare systems and factors related to having screen- vs non-screen-detected breast cancer across five Canadian provinces.

METHODS

Women aged 40+ diagnosed with incident breast cancer from 2007 to 2012 in five Canadian provinces were identified from their respective provincial cancer registries. Standardized provincial datasets were created linking screening, health administrative, and claims data. Province-specific logistic regression models were used to evaluate the association of demographic and healthcare utilization factors in each province with the odds of screen-detected cancer.

RESULTS

There was significant inter- and intra-provincial variation by age. Screen detection ranged from 42% to 52% in ages 50-69 but women aged 50-59 had approximately 4-8% lower screen detection than those aged 60-69 in all provinces. Screening associations with income quintile and rurality varied across provinces. Those least likely to be screen-detected within a province were consistently in the lowest income quintile; OR ranged from 0.62-0.89 relative to highest income quintile/urban patients aged 50-69. Lack of visits to primary care 30 months prior to diagnosis was also consistently associated with lower odds of screen detection (OR range, 0.37-0.76).

CONCLUSION

Breast cancer screen detection rates in the Canadian provinces examined are relatively high. Associations with income-rurality indicate a need for greater attention and/or targeted outreach to specific communities and/or provincial regions to improve access to breast cancer screening services intra-provincially.

The impact of child maltreatment on the risk of deliberate self-harm among adolescents: A population-wide cohort study using linked administrative records

Adolescents exposed to maltreatment have an elevated risk of deliberate self-harm (DSH). The aim of this study was to investigate longitudinally the effects of the number, timing, and type of maltreatment allegations on adolescent risk of having a DSH-related hospital admission, using linked data in Western Australia. A total of 351,372 children born between 1986 and 2000 were followed from birth up to the year 2010. Cox regression models were utilized, while controlling for a range of psychosocial covariates. Compared to children without allegations of maltreatment, children with unsubstantiated allegations only (aHR=1.04, 95%CI: 1.00-1.08, p<.01) and children with a substantiated allegation (aHR=1.10, 95%CI: 1.06-1.15, p<.001) all had significantly increased risk of DSH in adolescence. Among children with a substantiated allegation of maltreatment, the greater the number of allegations, the longer the exposure to maltreatment, and the more types of maltreatment experienced by a child, the higher the child's risk of DSH. However, this dose-response pattern was not found among children with unsubstantiated allegations only. This study calls for the early identification of children who are vulnerable to maltreatment, the better identification of the duration and severity of maltreatment experiences, and the provision of continued care and support, to reduce the child's DSH risk in adolescence.