Detection and evaluation of signals associated with exposure to individual and combination of medications in pregnancy: a signal detection study protocol.

INTRODUCTION: Considering the high prevalence of polypharmacy in pregnant women and the knowledge gap in the risk-benefit safety profile of their often-complex treatment plan, more research is needed to optimise prescribing. In this study, we aim to detect adverse and protective effect signals of exposure to individual and pairwise combinations of medications during pregnancy. METHODS AND ANALYSIS: Using a range of real-world data sources from the UK, we aim to conduct a pharmacovigilance study to assess the safety of medications prescribed during the preconception period (3 months prior to conception) and first trimester of pregnancy. Women aged between 15 and 49 years with a record of pregnancy within the Clinical Practice Research Datalink (CPRD) Pregnancy Register, the Welsh Secure Anonymised Information Linkage (SAIL), the Scottish Morbidity Record (SMR) data sets and the Northern Ireland Maternity System (NIMATS) will be included. A series of case control studies will be conducted to estimate measures of disproportionality, detecting signals of association between a range of pregnancy outcomes and exposure to individual and combinations of medications. A multidisciplinary expert team will be invited to a signal detection workshop. By employing a structured framework, signals will be transparently assessed by each member of the team using a questionnaire appraising the signals on aspects of temporality, selection, time and measurement-related biases and confounding by underlying disease or comedications. Through group discussion, the expert team will reach consensus on each of the medication exposure-outcome signal, thereby excluding spurious signals, leaving signals suggestive of causal associations for further evaluation. ETHICS AND DISSEMINATION: Ethical approval has been obtained from the Independent Scientific Advisory Committee, SAIL Information Governance Review Panel, University of St. Andrews Teaching and Research Ethics Committee and Office for Research Ethics Committees Northern Ireland (ORECNI) for access and use of CPRD, SAIL, SMR and NIMATS data, respectively.

COVID-19 vaccination in pregnancy: the impact of multimorbidity and smoking status on vaccine hesitancy, a cohort study of 25,111 women in Wales, UK.

BACKGROUND: Multimorbidity, smoking status, and pregnancy are identified as three risk factors associated with more severe outcomes following a SARS-CoV-2 infection, thus vaccination uptake is crucial for pregnant women living with multimorbidity and a history of smoking. This study aimed to examine the impact of multimorbidity, smoking status, and demographics (age, ethnic group, area of deprivation) on vaccine hesitancy among pregnant women in Wales using electronic health records (EHR) linkage. METHODS: This cohort study utilised routinely collected, individual-level, anonymised population-scale linked data within the Secure Anonymised Information Linkage (SAIL) Databank. Pregnant women were identified from 13th April 2021 to 31st December 2021. Survival analysis was employed to examine and compare the length of time to vaccination uptake in pregnancy by considering multimorbidity, smoking status, as well as depression, diabetes, asthma, and cardiovascular conditions independently. The study also assessed the variation in uptake by multimorbidity, smoking status, and demographics, both jointly and separately for the independent conditions, using hazard ratios (HR) derived from the Cox regression model. RESULTS: Within the population cohort, 8,203 (32.7%) received at least one dose of the COVID-19 vaccine during pregnancy, with 8,572 (34.1%) remaining unvaccinated throughout the follow-up period, and 8,336 (33.2%) receiving the vaccine postpartum. Women aged 30 years or older were more likely to have the vaccine in pregnancy. Those who had depression were slightly but significantly more likely to have the vaccine compared to those without depression (HR = 1.08, 95% CI 1.03 to 1.14, p = 0.002). Women living with multimorbidity were 1.12 times more likely to have the vaccine compared to those living without multimorbidity (HR = 1.12, 95% CI 1.04 to 1.19, p = 0.001). Vaccine uptakes were significantly lower among both current smokers and former smokers compared to never smokers (HR = 0.87, 95% CI 0.81 to 0.94, p < 0.001 and HR = 0.92, 95% CI 0.85 to 0.98, p = 0.015 respectively). Uptake was also lower among those living in the most deprived areas compared to those living in the most affluent areas (HR = 0.89, 95% CI 0.83 to 0.96, p = 0.002). CONCLUSION: Younger women, living without multimorbidity, current and former smokers, and those living in the more deprived areas are less likely to have the vaccine, thus, a targeted approach to vaccinations may be required for these groups. Pregnant individuals living with multimorbidity exhibit a slight but statistically significant reduction in vaccine hesitancy towards COVID-19 during pregnancy.

Maternal multimorbidity and preterm birth in Scotland: an observational record-linkage study.

BACKGROUND: Multimorbidity is common in women across the life course. Preterm birth is the single biggest cause of neonatal mortality and morbidity. We aim to estimate the prevalence of multimorbidity in pregnant women and to examine the association between maternal multimorbidity and PTB. METHODS: This is a retrospective cohort study using electronic health records from the Scottish Morbidity Records. All pregnancies among women aged 15 to 49 with a conception date between 1 January 2014 and 31 December 2018 were included. Multimorbidity was defined as the presence of two or more pre-existing long-term physical or mental health conditions, and complex multimorbidity as the presence of four or more. It was calculated at the time of conception using a predefined list of 79 conditions published by the MuM-PreDiCT consortium. PTB was defined as babies born alive between 24 and less than 37 completed weeks of gestation. We used Generalised Estimating Equations adjusted for maternal age, socioeconomic status, number of previous pregnancies, BMI, and smoking history to estimate the effect of maternal pre-existing multimorbidity. Absolut rates are reported in the results and tables, whilst Odds Ratios (ORs) are adjusted (aOR). RESULTS: Thirty thousand five hundred fifty-seven singleton births from 27,711 pregnant women were included in the analysis. The prevalence of pre-existing multimorbidity and complex multimorbidity was 16.8% (95% CI: 16.4-17.2) and 3.6% (95% CI: 3.3-3.8), respectively. The prevalence of multimorbidity in the youngest age group was 10.2%(95% CI: 8.8-11.6), while in those 40 to 44, it was 21.4% (95% CI: 18.4-24.4), and in the 45 to 49 age group, it was 20% (95% CI: 8.9-31.1). In women without multimorbidity, the prevalence of PTB was 6.7%; it was 11.6% in women with multimorbidity and 15.6% in women with complex multimorbidity. After adjusting for maternal age, socioeconomic status, number of previous pregnancies, Body Mass Index (BMI), and smoking, multimorbidity was associated with higher odds of PTB (aOR = 1.64, 95% CI: 1.48-1.82). CONCLUSIONS: Multimorbidity at the time of conception was present in one in six women and was associated with an increased risk of preterm birth. Multimorbidity presents a significant health burden to women and their offspring. Routine and comprehensive evaluation of women with multimorbidity before and during pregnancy is urgently needed.

Insights from linking police domestic abuse data and health data in South Wales, UK: a linked routine data analysis using decision tree classification.

BACKGROUND: Exposure to domestic abuse can lead to long-term negative impacts on the victim's physical and psychological wellbeing. The 1998 Crime and Disorder Act requires agencies to collaborate on crime reduction strategies, including data sharing. Although data sharing is feasible for individuals, rarely are whole-agency data linked. This study aimed to examine the knowledge obtained by integrating information from police and health-care datasets through data linkage and analyse associated risk factor clusters. METHODS: This retrospective cohort study analyses data from residents of South Wales who were victims of domestic abuse resulting in a Public Protection Notification (PPN) submission between Aug 12, 2015 and March 31, 2020. The study links these data with the victims' health records, collated within the Secure Anonymised Information Linkage databank, to examine factors associated with the outcome of an Emergency Department attendance, emergency hospital admission, or death within 12 months of the PPN submission. To assess the time to outcome for domestic abuse victims after the index PPN submission, we used Kaplan-Meier survival analysis. We used multivariable Cox regression models to identify which factors contributed the highest risk of experiencing an outcome after the index PPN submission. Finally, we created decision trees to describe specific groups of individuals who are at risk of experiencing a domestic abuse incident and subsequent outcome. FINDINGS: After excluding individuals with multiple PPN records, duplicates, and records with a poor matching score or missing fields, the resulting clean dataset consisted of 8709 domestic abuse victims, of whom 6257 (71·8%) were female. Within a year of a domestic abuse incident, 3650 (41·9%) individuals had an outcome. Factors associated with experiencing an outcome within 12 months of the PPN included younger victim age (hazard ratio 1·183 [95% CI 1·053-1·329], p=0·0048), further PPN submissions after the initial referral (1·383 [1·295-1·476]; p<0·0001), injury at the scene (1·484 [1·368-1·609]; p<0·0001), assessed high risk (1·600 [1·444-1·773]; p<0·0001), referral to other agencies (1·518 [1·358-1·697]; p<0·0001), history of violence (1·229 [1·134-1·333]; p<0·0001), attempted strangulation (1·311 [1·148-1·497]; p<0·0001), and pregnancy (1·372 [1·142-1·648]; p=0·0007). Health-care data before the index PPN established that previous Emergency Department and hospital admissions, smoking, smoking cessation advice, obstetric codes, and prescription of antidepressants and antibiotics were associated with having a future outcome following a domestic abuse incident. INTERPRETATION: The results indicate that vulnerable individuals are detectable in multiple datasets before and after involvement of the police. Operationalising these findings could reduce police callouts and future Emergency Department or hospital admissions, and improve outcomes for those who are vulnerable. Strategies include querying previous Emergency Department and hospital admissions, giving a high-risk assessment for a pregnant victim, and facilitating data linkage to identify vulnerable individuals. FUNDING: National Institute for Health Research.

Weighting of risk factors for low birth weight: a linked routine data cohort study in Wales, UK.

OBJECTIVE: Globally, 20 million children are born with a birth weight below 2500 g every year, which is considered as a low birthweight (LBW) baby. This study investigates the contribution of modifiable risk factors in a nationally representative Welsh e-cohort of children and their mothers to inform opportunities to reduce LBW prevalence. DESIGN: A longitudinal cohort study based on anonymously linked, routinely collected multiple administrative data sets. PARTICIPANTS: The cohort, (N=693 377) comprising of children born between 1 January 1998 and 31 December 2018 in Wales, was selected from the National Community Child Health Database. OUTCOME MEASURES: The risk factors associated with a binary LBW (outcome) variable were investigated with multivariable logistic regression (MLR) and decision tree (DT) models. RESULTS: The MLR model showed that non-singleton children had the highest risk of LBW (adjusted OR 21.74 (95% CI 21.09 to 22.40)), followed by pregnancy interval less than 1 year (2.92 (95% CI 2.70 to 3.15)), maternal physical and mental health conditions including diabetes (2.03 (1.81 to 2.28)), anaemia (1.26 (95% CI 1.16 to 1.36)), depression (1.58 (95% CI 1.43 to 1.75)), serious mental illness (1.46 (95% CI 1.04 to 2.05)), anxiety (1.22 (95% CI 1.08 to 1.38)) and use of antidepressant medication during pregnancy (1.92 (95% CI 1.20 to 3.07)). Additional maternal risk factors include smoking (1.80 (95% CI 1.76 to 1.84)), alcohol-related hospital admission (1.60 (95% CI 1.30 to 1.97)), substance misuse (1.35 (95% CI 1.29 to 1.41)) and evidence of domestic abuse (1.98 (95% CI 1.39 to 2.81)). Living in less deprived area has lower risk of LBW (0.70 (95% CI 0.67 to 0.72)). The most important risk factors from the DT models include maternal factors such as smoking, maternal weight, substance misuse record, maternal age along with deprivation-Welsh Index of Multiple Deprivation score, pregnancy interval and birth order of the child. CONCLUSION: Resources to reduce the prevalence of LBW should focus on improving maternal health, reducing preterm births, increasing awareness of what is a sufficient pregnancy interval, and to provide adequate support for mothers' mental health and well-being.

Polypharmacy during pregnancy and associated risk factors: a retrospective analysis of 577 medication exposures among 1.5 million pregnancies in the UK, 2000-2019.

BACKGROUND: The number of medications prescribed during pregnancy has increased over the past few decades. Few studies have described the prevalence of multiple medication use among pregnant women. This study aims to describe the overall prevalence over the last two decades among all pregnant women and those with multimorbidity and to identify risk factors for polypharmacy in pregnancy. METHODS: A retrospective cohort study was conducted between 2000 and 2019 using the Clinical Practice Research Datalink (CPRD) pregnancy register. Prescription records for 577 medication categories were obtained. Prevalence estimates for polypharmacy (ranging from 2+ to 11+ medications) were presented along with the medications commonly prescribed individually and in pairs during the first trimester and the entire pregnancy period. Logistic regression models were performed to identify risk factors for polypharmacy. RESULTS: During the first trimester (812,354 pregnancies), the prevalence of polypharmacy ranged from 24.6% (2+ medications) to 0.1% (11+ medications). During the entire pregnancy period (774,247 pregnancies), the prevalence ranged from 58.7 to 1.4%. Broad-spectrum penicillin (6.6%), compound analgesics (4.5%) and treatment of candidiasis (4.3%) were commonly prescribed. Pairs of medication prescribed to manage different long-term conditions commonly included selective beta 2 agonists or selective serotonin re-uptake inhibitors (SSRIs). Risk factors for being prescribed 2+ medications during the first trimester of pregnancy include being overweight or obese [aOR: 1.16 (1.14-1.18) and 1.55 (1.53-1.57)], belonging to an ethnic minority group [aOR: 2.40 (2.33-2.47), 1.71 (1.65-1.76), 1.41 (1.35-1.47) and 1.39 (1.30-1.49) among women from South Asian, Black, other and mixed ethnicities compared to white women] and smoking or previously smoking [aOR: 1.19 (1.18-1.20) and 1.05 (1.03-1.06)]. Higher and lower age, higher gravidity, increasing number of comorbidities and increasing level of deprivation were also associated with increased odds of polypharmacy. CONCLUSIONS: The prevalence of polypharmacy during pregnancy has increased over the past two decades and is particularly high in younger and older women; women with high BMI, smokers and ex-smokers; and women with multimorbidity, higher gravidity and higher levels of deprivation. Well-conducted pharmaco-epidemiological research is needed to understand the effects of multiple medication use on the developing foetus.

Machine Learning in Colorectal Cancer Risk Prediction from Routinely Collected Data: A Review.

The inclusion of machine-learning-derived models in systematic reviews of risk prediction models for colorectal cancer is rare. Whilst such reviews have highlighted methodological issues and limited performance of the models included, it is unclear why machine-learning-derived models are absent and whether such models suffer similar methodological problems. This scoping review aims to identify machine-learning models, assess their methodology, and compare their performance with that found in previous reviews. A literature search of four databases was performed for colorectal cancer prediction and prognosis model publications that included at least one machine-learning model. A total of 14 publications were identified for inclusion in the scoping review. Data was extracted using an adapted CHARM checklist against which the models were benchmarked. The review found similar methodological problems with machine-learning models to that observed in systematic reviews for non-machine-learning models, although model performance was better. The inclusion of machine-learning models in systematic reviews is required, as they offer improved performance despite similar methodological omissions; however, to achieve this the methodological issues that affect many prediction models need to be addressed.

Real-world use of an etanercept biosimilar including selective versus automatic substitution in inflammatory arthritis patients: a UK-based electronic health records study.

Objective: Biosimilars are approved as an alternative treatment to their originators. We compared the clinical outcomes of etanercept (ETN) biosimilar compared with ETN originator in real-world practice, from two local health boards in Wales with different policies on switching: automatic vs selective. Methods: Data from the Secure Anonymised Information Linkage (SAIL) databank in Wales were used to create a retrospective cohort study using linked primary and secondary care data. Patients aged ≥18 years with diagnosis codes for RA, PsA or AS were included. Outcomes included treatment failure and DAS-28 score (for RA). The local health board with a policy of automatic switching (i.e. clinician/nurse involvement not mandated) is labelled as automatic switch area, and the other, which required clinician/nurse supervision, as selective switch. Results: Of 8925 individuals with inflammatory arthritis, 13.3% (365) received ETN biosimilar and 31.5% (863) ETN originator. The treatment discontinuation rate was similar for ETN biosimilar and originator by Kaplan-Meier analysis. More biosimilar failure patients were treated in the automatic switch area (15 vs 4.8%). In the automatic switch area, 28.8% (75 of 260) of patients switched automatically from ETN originator to biosimilar compared with 10.5% (11 of 105) in the selective switch area. ETN biosimilar reduced DAS-28 by 1.6 ± 1.8 in the selective switch area vs 0.4 ± 0.6 in the automatic switch area. Conclusion: The ETN biosimilar was well tolerated. Fewer people were switched using selective policy, but this was associated with lower failure rates. Automatic switch policy led to more patients being switched and did not lead to significant worsening of disease.

Shielding reduced incidence of COVID-19 in patients with inflammatory arthritis but vulnerability is associated with increased mortality.

OBJECTIVES: Investigate whether individuals with inflammatory arthritis (IA), their treatments and shielding status affect the risk of adverse outcomes from COVID-19 for the entire population of Wales, UK. METHODS: Retrospective, population-based cohort study using linked, anonymized electronic health data from SAIL Databank, including primary/secondary care, rheumatology, Office for National Statistics Mortality and COVID-19 laboratory data. Individuals aged 18 years and over testing positive for COVID-19 between March 2020 and May 2021 with READ Codes present for rheumatoid arthritis, psoriatic arthritis and ankylosing spondylitis formed the study cases. RESULTS: A total of 1966 people with IA and 166 602 without tested positive for COVID-19. The incidence rate was 3.5% (1966/56 914) in IA, vs 6% in the general population (166 602/2 760 442), (difference: 2.5%, 95% CI: 2.4%, 2.7%, P ≤0.001). In an adjusted Cox proportional hazard model, IA was not associated with higher mortality (HR: 0.56, 95% CI: 0.18, 1.64, P=0.286). Significant risk factors included shielding (HR: 1.52, 95% CI: 1.40, 1.64, P ≤0.001), hospitalization for previous infections (HR: 1.20, 95% CI: 1.12, 1.28, P ≤0.001), hospitalizations one year pre-pandemic (HR: 1.34, 95% CI: 1.25, 1.44, P ≤0.001) and glucocorticoid use (HR: 1.17, 95% CI: 1.09, 1.25, P ≤0.001). CONCLUSIONS: Individuals with IA had a lower incidence of COVID-19, probably due to shielding. IA was not associated with increased mortality following COVID-19 infection; being vulnerable (shielded), comorbidities and other factors were associated with increased risk. These key risk factors can identify individuals with IA at greater risk from COVID-19 and advised to shield during high community prevalence.

Impact of first UK COVID-19 lockdown on hospital admissions: Interrupted time series study of 32 million people.

Background: Uncontrolled infection and lockdown measures introduced in response have resulted in an unprecedented challenge for health systems internationally. Whether such unprecedented impact was due to lockdown itself and recedes when such measures are lifted is unclear. We assessed the short- and medium-term impacts of the first lockdown measures on hospital care for tracer non-COVID-19 conditions in England, Scotland and Wales across diseases, sexes, and socioeconomic and ethnic groups. Methods: We used OpenSAFELY (for England), EAVEII (Scotland), and SAIL Databank (Wales) to extract weekly hospital admission rates for cancer, cardiovascular and respiratory conditions (excluding COVID-19) from the pre-pandemic period until 25/10/2020 and conducted a controlled interrupted time series analysis. We undertook stratified analyses and assessed admission rates over seven months during which lockdown restrictions were gradually lifted. Findings: Our combined dataset included 32 million people who contributed over 74 million person-years. Admission rates for all three conditions fell by 34.2% (Confidence Interval (CI): -43.0, -25.3) in England, 20.9% (CI: -27.8, -14.1) in Scotland, and 24.7% (CI: -36.7, -12.7) in Wales, with falls across every stratum considered. In all three nations, cancer-related admissions fell the most while respiratory-related admissions fell the least (e.g., rates fell by 40.5% (CI: -47.4, -33.6), 21.9% (CI: -35.4, -8.4), and 19.0% (CI: -30.6, -7.4) in England for cancer, cardiovascular-related, and respiratory-related admissions respectively). Unscheduled admissions rates fell more in the most than the least deprived quintile across all three nations. Some ethnic minority groups experienced greater falls in admissions (e.g., in England, unscheduled admissions fell by 9.5% (CI: -20.2, 1.2) for Whites, but 44.3% (CI: -71.0, -17.6), 34.6% (CI: -63.8, -5.3), and 25.6% (CI: -45.0, -6.3) for Mixed, Other and Black ethnic groups respectively). Despite easing of restrictions, the overall admission rates remained lower in England, Scotland, and Wales by 20.8%, 21.6%, and 22.0%, respectively when compared to the same period (August-September) during the pre-pandemic years. This corresponds to a reduction of 26.2, 23.8 and 30.2 admissions per 100,000 people in England, Scotland, and Wales respectively. Interpretation: Hospital care for non-COVID diseases fell substantially across England, Scotland, and Wales during the first lockdown, with reductions persisting for at least six months. The most deprived and minority ethnic groups were impacted more severely. Funding: This work was funded by the Medical Research Council as part of the Lifelong Health and Wellbeing study as part of National Core Studies (MC_PC_20030). SVK acknowledges funding from the Medical Research Council (MC_UU_00022/2), and the Scottish Government Chief Scientist Office (SPHSU17). EAVE II is funded by the Medical Research Council (MR/R008345/1) with the support of BREATHE - The Health Data Research Hub for Respiratory Health (MC_PC_19004), which is funded through the UK Research and Innovation Industrial Strategy Challenge Fund and delivered through Health Data Research UK. BG has received research funding from the NHS National Institute for Health Research (NIHR), the Wellcome Trust, Health Data Research UK, Asthma UK, the British Lung Foundation, and the Longitudinal Health and Wellbeing strand of the National Core Studies programme.

Health checks for adults with intellectual disability and association with survival rates: a linked electronic records matched cohort study in Wales, UK.

OBJECTIVE: To examine if mortality rates are lower in people with intellectual disability who have had a health check compared with those who have not had health checks. SETTING: General practice records of 26 954 people with an intellectual disability in Wales between 2005-2017, of which 7650 (28.4%) with a health check were matched 1:2 with those without a health check. PRIMARY OUTCOME MEASURE: Office of National Statistics mortality data; a Cox regression was utilised to examine time to death adjusted for comorbidities and gender. RESULTS: Patients who had a health check were stratified by those who (1) had a confirmed health check, that is, Read Code for a health check (n=7650 (28.4 %)) and (2) had no evidence of receiving a health check in their medical record. Patients with a health check were matched for age at time of health check with two people who did not have a health check. The health check was associated with improved survival for those with autism or Down's Syndrome (HR 0.58 (95% CI 0.37 to 0.91) and HR 0.76 (95% CI 0.64 to 0.91), respectively). There was no evidence of improved survival for those diagnosed with diabetes or cancer. The people who had a health check were more likely to be older, have epilepsy and less likely to have autism or Down's syndrome. CONCLUSIONS: Health checks are likely to influence survival if started before a person is diagnosed with a chronic condition, especially for people with autism or Down's syndrome.

Epidemiology of pre-existing multimorbidity in pregnant women in the UK in 2018: a population-based cross-sectional study.

BACKGROUND: Although maternal death is rare in the United Kingdom, 90% of these women had multiple health/social problems. This study aims to estimate the prevalence of pre-existing multimorbidity (two or more long-term physical or mental health conditions) in pregnant women in the United Kingdom (England, Northern Ireland, Wales and Scotland). STUDY DESIGN: Pregnant women aged 15-49 years with a conception date 1/1/2018 to 31/12/2018 were included in this population-based cross-sectional study, using routine healthcare datasets from primary care: Clinical Practice Research Datalink (CPRD, United Kingdom, n = 37,641) and Secure Anonymized Information Linkage databank (SAIL, Wales, n = 27,782), and secondary care: Scottish Morbidity Records with linked community prescribing data (SMR, Tayside and Fife, n = 6099). Pre-existing multimorbidity preconception was defined from 79 long-term health conditions prioritised through a workshop with patient representatives and clinicians. RESULTS: The prevalence of multimorbidity was 44.2% (95% CI 43.7-44.7%), 46.2% (45.6-46.8%) and 19.8% (18.8-20.8%) in CPRD, SAIL and SMR respectively. When limited to health conditions that were active in the year before pregnancy, the prevalence of multimorbidity was still high (24.2% [23.8-24.6%], 23.5% [23.0-24.0%] and 17.0% [16.0 to 17.9%] in the respective datasets). Mental health conditions were highly prevalent and involved 70% of multimorbidity CPRD: multimorbidity with ≥one mental health condition/s 31.3% [30.8-31.8%]). After adjusting for age, ethnicity, gravidity, index of multiple deprivation, body mass index and smoking, logistic regression showed that pregnant women with multimorbidity were more likely to be older (CPRD England, adjusted OR 1.81 [95% CI 1.04-3.17] 45-49 years vs 15-19 years), multigravid (1.68 [1.50-1.89] gravidity ≥ five vs one), have raised body mass index (1.59 [1.44-1.76], body mass index 30+ vs body mass index 18.5-24.9) and smoked preconception (1.61 [1.46-1.77) vs non-smoker). CONCLUSION: Multimorbidity is prevalent in pregnant women in the United Kingdom, they are more likely to be older, multigravid, have raised body mass index and smoked preconception. Secondary care and community prescribing dataset may only capture the severe spectrum of health conditions. Research is needed urgently to quantify the consequences of maternal multimorbidity for both mothers and children.

Predicting Hospital Readmission for Campylobacteriosis from Electronic Health Records: A Machine Learning and Text Mining Perspective.

(1) Background: This study investigates influential risk factors for predicting 30-day readmission to hospital for Campylobacter infections (CI). (2) Methods: We linked general practitioner and hospital admission records of 13,006 patients with CI in Wales (1990-2015). An approach called TF-zR (term frequency-zRelevance) technique was presented to evaluates how relevant a clinical term is to a patient in a cohort characterized by coded health records. The zR is a supervised term-weighting metric to assign weight to a term based on relative frequencies of the term across different classes. Cost-sensitive classifier with swarm optimization and weighted subset learning was integrated to identify influential clinical signals as predictors and optimal model for readmission prediction. (3) Results: From a pool of up to 17,506 variables, 33 most predictive factors were identified, including age, gender, Townsend deprivation quintiles, comorbidities, medications, and procedures. The predictive model predicted readmission with 73% sensitivity and 54% specificity. Variables associated with readmission included male gender, recurrent tonsillitis, non-healing open wounds, operation for in-gown toenails. Cystitis, paracetamol/codeine use, age (21-25), and heliclear triple pack use, were associated with a lower risk of readmission. (4) Conclusions: This study gives a profile of clustered variables that are predictive of readmission associated with campylobacteriosis.

Biologic use in psoriatic arthritis and ankylosing spondylitis patients: a descriptive epidemiological study using linked, routine data in Wales, UK.

OBJECTIVES: PsA and AS are chronic diseases associated with significant morbidities. National and international management guidelines include treatment with biologic therapies to improve outcomes and quality of life. There are limited real-world data on the patients' journey from symptom onset to diagnosis and treatment in the UK. We use real-life, linked health data to explore patient pathways and the impact of biologics on patient outcomes. METHODS: Data from the Secure Anonymised Information Linkage databank in Wales were used to assess diagnosis and treatment of patients ≥18 years of age with at least one International Classification of Diseases, Tenth Revision code present for PsA/AS in rheumatology clinic data and at least one Read code present in primary care records. We investigated the use of biologics while exploring demographics, comorbidities and surgical procedures of 641 AS patients and 1312 PsA patients. RESULTS: AS patients were significantly younger at diagnosis and were predominantly male. The average time from presenting symptoms to diagnosis of AS and PsA was 7.9 (s.d. 5.5) and 9.3 (s.d. 5.5) years, respectively. The proportion of patients receiving biologic treatment was significantly higher in AS (46%) compared with PsA patients (28.8%); of these, 23.1% of AS and 22.2% of PsA patients stopped/switched a biologic. There was a significant reduction in primary care involvement, sick notes and disability living allowance for both AS and PsA patients following biologic initiation. CONCLUSION: This real-world descriptive study confirms that patients treated with biologics have reduced disability and time off work despite being initiated ∼13 years after the first symptoms and 6 years after diagnosis.

Predictors of objectively measured physical activity in 12-month-old infants: A study of linked birth cohort data with electronic health records.

BACKGROUND: Physical activity (PA) levels are associated with long-term health, and levels of PA when young are predictive of adult activity levels. OBJECTIVES: This study examines factors associated with PA levels in 12-month infants. METHOD: One hundred forty-one mother-infant pairs were recruited via a longitudinal birth cohort study (April 2010 to March 2013). The PA level was collected using accelerometers and linked to postnatal notes and electronic medical records via the Secure Anonymised Information Linkage databank. Univariable and multivariable linear regressions were used to examine the factors associated with PA levels. RESULTS: Using univariable analysis, higher PA was associated with the following (P value less than 0.05): being male, larger infant size, healthy maternal blood pressure levels, full-term gestation period, higher consumption of vegetables (infant), lower consumption of juice (infant), low consumption of adult crisps (infant), longer breastfeeding duration, and more movement during sleep (infant) but fewer night wakings. Combined into a multivariable regression model (R2  = 0.654), all factors remained significant, showing lower PA levels were associated with female gender, smaller infant, preterm birth, higher maternal blood pressure, low vegetable consumption, high crisp consumption, and less night movement. CONCLUSION: The PA levels of infants were strongly associated with both gestational and postnatal environmental factors. Healthy behaviours appear to cluster, and a healthy diet was associated with a more active infant. Boys were substantially more active than girls, even at age 12 months. These findings can help inform interventions to promote healthier lives for infants and to understand the determinants of their PA levels.

Cardiovascular risk factors predicting cardiac events are different in patients with rheumatoid arthritis, psoriatic arthritis, and psoriasis.

OBJECTIVES: Increased cardiovascular risk in rheumatoid arthritis (RA) is well established. Examining traditional cardiovascular risk factors alone underestimates cardiovascular risk in RA. Systematic inflammation, measured by erythrocyte sedimentation rate or C-reactive protein is also a major risk factor. However, the contribution of traditional cardiovascular risk factors (such as obesity and hyperlipidaemia) compared to inflammation is uncertain in psoriatic arthritis (PsA) and RA. We examine the incidence of major adverse cardiac events (MACE) among patients with RA, PsA psoriasis, and controls adjusting for risk factors, inflammation and disease modifying anti-rheumatic drug treatment, to better define cardiovascular risk. METHODS: Using the Secure Anonymised Information Linkage databank, comprising routinely collected Welsh health data from 1999 to 2013, the incidence and first occurrence of a MACE in individuals with RA (n = 8650), PsA (n = 2128) and psoriasis (n = 24,630) compared to controls (n = 11,87,706) was investigated. RESULTS: Traditional cardiovascular risk factors are higher in RA, PsA and psoriasis than controls. After adjusting for these factors, additional cardiovascular risk was only significantly increased in female RA patients (HR = 1.3; 95% CI: 1.0-1.7; p = 0.05) and psoriasis (HR = 1.2; 95% CI: 1.0-1.4; p = 0.02) but not statistically significant for PsA (HR = 1.5; 95% CI: 0.9-2.5; p = 0.13). ESR and CRP were increased in patients with RA but not in patients with psoriasis. CONCLUSION: Additional increased cardiovascular risk was observed in female RA and psoriasis but not PsA. Systematic inflammation is higher in RA but not psoriasis, indicating that there are varying mediators of cardiovascular risk across these conditions.

Active children through individual vouchers - evaluation (ACTIVE): protocol for a mixed method randomised control trial to increase physical activity levels in teenagers.

BACKGROUND: Many teenagers are insufficiently active despite the health benefits of physical activity (PA). There is strong evidence to show that inactivity and low fitness levels increase the risk of non-communicable diseases such as coronary heart disease (CHD), type 2 diabetes and breast and colon cancers (Lee et al. Lancet 380:219-29, 2012). A major barrier facing adolescents is accessibility (e.g. cost and lack of local facilities). The ACTIVE project aims to tackle this barrier through a multi-faceted intervention, giving teenagers vouchers to spend on activities of their choice and empowering young people to improve their fitness and PA levels. DESIGN: ACTIVE is a mixed methods randomised control trial in 7 secondary schools in Swansea, South Wales. Quantitative and qualitative measures including PA (cooper run test (CRT), accelerometery over 7 days), cardiovascular (CV) measures (blood pressure, pulse wave analysis) and focus groups will be undertaken at 4 separate time points (baseline, 6 months,12 months and follow-up at 18 months). Intervention schools will receive a multi-component intervention involving 12 months of £20 vouchers to spend on physical activities of their choice, a peer mentor scheme and opportunities to attend advocacy meetings. Control schools are encouraged to continue usual practice. The primary aim is to examine the effect of the intervention in improving cardiovascular fitness. DISCUSSION: This paper describes the protocol for the ACTIVE randomised control trial, which aims to increase fitness, physical activity and socialisation of teenagers in Swansea, UK via a voucher scheme combined with peer mentoring. Results can contribute to the evidence base on teenage physical activity and, if effective, the intervention has the potential to inform future physical activity interventions and policy. TRIAL REGISTRATION: ISRCTN75594310 (Assigned 06/03/2017).

Development of an algorithm for determining smoking status and behaviour over the life course from UK electronic primary care records.

BACKGROUND: Patients' smoking status is routinely collected by General Practitioners (GP) in UK primary health care. There is an abundance of Read codes pertaining to smoking, including those relating to smoking cessation therapy, prescription, and administration codes, in addition to the more regularly employed smoking status codes. Large databases of primary care data are increasingly used for epidemiological analysis; smoking status is an important covariate in many such analyses. However, the variable definition is rarely documented in the literature. METHODS: The Secure Anonymised Information Linkage (SAIL) databank is a repository for a national collection of person-based anonymised health and socio-economic administrative data in Wales, UK. An exploration of GP smoking status data from the SAIL databank was carried out to explore the range of codes available and how they could be used in the identification of different categories of smokers, ex-smokers and never smokers. An algorithm was developed which addresses inconsistencies and changes in smoking status recording across the life course and compared with recorded smoking status as recorded in the Welsh Health Survey (WHS), 2013 and 2014 at individual level. However, the WHS could not be regarded as a "gold standard" for validation. RESULTS: There were 6836 individuals in the linked dataset. Missing data were more common in GP records (6%) than in WHS (1.1%). Our algorithm assigns ex-smoker status to 34% of never-smokers, and detects 30% more smokers than are declared in the WHS data. When distinguishing between current smokers and non-smokers, the similarity between the WHS and GP data using the nearest date of comparison was κ = 0.78. When temporal conflicts had been accounted for, the similarity was κ = 0.64, showing the importance of addressing conflicts. CONCLUSIONS: We present an algorithm for the identification of a patient's smoking status using GP self-reported data. We have included sufficient details to allow others to replicate this work, thus increasing the standards of documentation within this research area and assessment of smoking status in routine data.

Severe flare as a predictor of poor outcome in ankylosing spondylitis: a cohort study using questionnaire and routine data linkage.

OBJECTIVE: To explore severe flare and constant disease pattern (no periods of remission) in AS as predictors of poor outcomes [impaired function, unemployment/early retirement, work impairment, anti-TNF, surgery, frequent general practitioner (GP) visits, depression and anxiety]. METHODS: Three hundred and forty-eight AS patients completed questionnaires about their experience with disease flares. Questionnaire data were linked to electronic medical records to examine visits to GPs and hospital admission data. Outcomes were stratified in two ways: self-reported experience of severe flare and constant disease pattern using the flare illustration tool. RESULTS: The majority of patients (72%, 208/289) experienced flare pre-diagnosis. Severe flares were reported by 58% (202/348) of participants (self-report); of these, 195 responded about earliest flares and 69% (135/195) of severe flare patients experienced flares pre-diagnosis. Patients who self-reported severe flares had worse function, disease activity, work impairment and symptoms of anxiety and depression, were less likely to be employed and had more GP encounters per year compared with those who never reported severe flares. Participants who reported constant unremitting disease on the flare illustration tool had worse disease activity, impaired function and work impairment and were more likely to smoke compared with those with intermittent disease. Analysis showed a relationship between self-report of severe flare and subsequent depression, impaired function, increased disease activity and work limitations. CONCLUSION: Severe flare is associated with poor outcomes such as work impairment and impaired function. The onset of severe flare early in the disease course may be a risk factor for later poor outcome and this group could benefit from targeted early aggressive treatment to improve prognosis.

Physical activity and excess weight in pregnancy have independent and unique effects on delivery and perinatal outcomes.

BACKGROUND: This study examines the effect of low daily physical activity levels and overweight/obesity in pregnancy on delivery and perinatal outcomes. METHODS: A prospective cohort study combining manually collected postnatal notes with anonymised data linkage. A total of 466 women sampled from the Growing Up in Wales: Environments for Healthy Living study. Women completed a questionnaire and were included in the study if they had an available Body mass index (BMI) (collected at 12 weeks gestation from antenatal records) and/or a physical activity score during pregnancy (7-day Actigraph reading). The full statistical model included the following potential confounding factors: maternal age, parity and smoking status. Main outcome measures included induction rates, duration of labour, mode of delivery, infant health and duration of hospital stay. FINDINGS: Mothers with lower physical activity levels were more likely to have an instrumental delivery (including forceps, ventouse and elective and emergency caesarean) in comparison to mothers with higher activity levels (adjusted OR:1.72(95%CI: 1.05 to 2.9)). Overweight/obese mothers were more likely to require an induction (adjusted OR:1.93 (95%CI 1.14 to 3.26), have a macrosomic baby (adjusted OR:1.96 (95%CI 1.08 to 3.56) and a longer hospital stay after delivery (adjusted OR:2.69 (95%CI 1.11 to 6.47). CONCLUSIONS: The type of delivery was associated with maternal physical activity level and not BMI. Perinatal outcomes (large for gestational age only) were determined by maternal BMI.