Mark D. Atkinson

The incidence of acute pancreatitis: impact of social deprivation, alcohol consumption, seasonal and demographic factors

The incidence of acute pancreatitis has increased sharply in many European countries and the USA in recent years.

No Increased Rate of Acute Myocardial Infarction or Stroke Among Patients with Ankylosing Spondylitis—A Retrospective Cohort Study Using Routine Data

OBJECTIVES To examine if people with ankylosing spondylitis (AS) are at higher risk of acute myocardial infarction (MI) or stroke compared to those without AS. METHODS Primary care records were linked with all hospital admissions and deaths caused by MI or stroke in Wales for the years 1999-2010. The linked data were then stratified by AS diagnosis and survival analysis was used to obtain the incidence rate of MI and separately cerebrovascular disease (CVD)/stroke. Cox regression was used to adjust for gender and age. Logistic regression was used to examine prevalence of diabetes, hypertension, or hyperlipidemia for those with AS compared to those without. RESULTS There were 1686 AS patients (75.9% male, average age 46.1 years) compared to 1,206,621 controls (48.9% male, average age 35.9 years). Age- and gender-adjusted hazard ratios for MI were 1.28 (95% CI: 0.93 to 1.74) P = 0.12, and for CVD/stroke 1.0 (95% CI: 0.73 to 1.39) P = 0.9, in AS compared to controls. The prevalence of diabetes and hypertension, but not hyperlipidemia/hypercholesterolemia, was higher in AS. CONCLUSIONS There is no increase in the MI or CVD/stroke rates in patients with AS compared to those without AS, despite higher rates of hypertension, which may be related to nonsteroidal anti-inflammatory drug use.

Incidence of Campylobacter and Salmonella infections following first prescription for PPI– a cohort study using routine data.

OBJECTIVES:To examine the incidence of Campylobacter and Salmonella infection in patients prescribed proton pump inhibitors (PPIs) compared with controls.METHODS:Retrospective cohort study using anonymous general practitioner (GP) data. Anonymised individual-level records from the Secure Anonymised Information Linkage (SAIL) system between 1990 and 2010 in Wales were selected. Data were available from 1,913,925 individuals including 358,938 prescribed a PPI. The main outcome measures examined included incidence of Campylobacter or Salmonella infection following a prescription for PPI.RESULTS:The rate of Campylobacter and Salmonella infections was already at 3.1–6.9 times that of non-PPI patients even before PPI prescription. The PPI group had an increased hazard rate of infection (after prescription for PPI) of 1.46 for Campylobacter and 1.2 for Salmonella, compared with baseline. However, the non-PPI patients also had an increased hazard ratio with time. In fact, the ratio of events in the PPI group compared with the non-PPI group using the prior event rate ratio was 1.17 (95% CI 0.74–1.61) for Campylobacter and 1.00 (0.5–1.5) for Salmonella.CONCLUSIONS:People who go on to be prescribed PPIs have a greater underlying risk of gastrointestinal (GI) infection beforehand and they have a higher prevalence of risk factors before PPI prescription. The rate of diagnosis of infection is increasing with time regardless of PPI use, and there is no evidence that PPI is associated with an increase in diagnosed GI infection. It is likely that factors associated with the demographic profile of the patient are the main contributors to increased rate of GI infection for patients prescribed PPIs.

Obesity in pregnancy: a retrospective prevalence-based study on health service utilisation and costs on the NHS.

Objective To estimate the direct healthcare cost of being overweight or obese throughout pregnancy to the National Health Service in Wales. Design Retrospective prevalence-based study. Setting Combined linked anonymised electronic datasets gathered on a cohort of women enrolled on the Growing Up in Wales: Environments for Healthy Living (EHL) study. Women were categorised into two groups: normal body mass index (BMI; n=260) and overweight/obese (BMI>25; n=224). Participants 484 singleton pregnancies with available health service records and an antenatal BMI. Primary outcome measure Total health service utilisation (comprising all general practitioner visits and prescribed medications, inpatient admissions and outpatient visits) and direct healthcare costs for providing these services in the year 2011–2012. Costs are calculated as cost of mother (no infant costs are included) and are related to health service usage throughout pregnancy and 2 months following delivery. Results There was a strong association between healthcare usage cost and BMI (p<0.001). Adjusting for maternal age, parity, ethnicity and comorbidity, mean total costs were 23% higher among overweight women (rate ratios (RR) 1.23, 95% CI 1.230 to 1.233) and 37% higher among obese women (RR 1.39, 95% CI 1.38 to 1.39) compared with women with normal weight. Adjusting for smoking, consumption of alcohol, or the presence of any comorbidities did not materially affect the results. The total mean cost estimates were £3546.3 for normal weight, £4244.4 for overweight and £4717.64 for obese women. Conclusions Increased health service usage and healthcare costs during pregnancy are associated with increasing maternal BMI; this was apparent across all health services considered within this study. Interventions costing less than £1171.34 per person could be cost-effective if they reduce healthcare usage among obese pregnant women to levels equivalent to that of normal weight women.

Impacts of the Primary School Free Breakfast Initiative on socio-economic inequalities in breakfast consumption among 9–11-year-old schoolchildren in Wales

Objectives Universal interventions may widen or narrow inequalities if disproportionately effective among higher or lower socio-economic groups. The present paper examines impacts of the Primary School Free Breakfast Initiative in Wales on inequalities in childrens dietary behaviours and cognitive functioning. Design Cluster-randomised controlled trial. Responses were linked to free school meal (FSM) entitlement via the Secure Anonymised Information Linkage databank. Impacts on inequalities were evaluated using weighted school-level regression models with interaction terms for intervention × whole-school percentage FSM entitlement and intervention × aggregated individual FSM entitlement. Individual-level regression models included interaction terms for intervention × individual FSM entitlement. Setting Fifty-five intervention and fifty-six wait-list control primary schools. Subjects Approximately 4500 children completed measures of dietary behaviours and cognitive tests at baseline and 12-month follow-up. Results School-level models indicated that children in intervention schools ate a greater number of healthy items for breakfast than children in control schools (b = 0·25; 95 % CI 0·07, 0·44), with larger increases observed in more deprived schools (interaction term b = 1·76; 95 % CI 0·36, 3·16). An interaction between intervention and household-level deprivation was not significant. Despite no main effects on breakfast skipping, a significant interaction was observed, indicating declines in breakfast skipping in more deprived schools (interaction term b = −0·07; 95 % CI −0·15, −0·00) and households (OR = 0·67; 95 % CI 0·46, 0·98). No significant influence on inequality was observed for the remaining outcomes. Conclusions Universal breakfast provision may reduce socio-economic inequalities in consumption of healthy breakfast items and breakfast skipping. There was no evidence of intervention-generated inequalities in any outcomes.

Grain quality in the Broadbalk Wheat Experiment and the winter North Atlantic Oscillation

SUMMARY Previous work has shown that the national average quality of the UK wheat crop from 1974 to 1999 was associated with the preceding winter North Atlantic Oscillation (NAO). The association of the winter NAO with the grain quality measure, specific weight, was shown to be mediated by sunshine duration during grain filling and unconditional wet day probability during grain ripening (the probability of a wet day following either a dry or a wet day). The present study tests the hypothesis that the association between specific weight and the winter NAO can be detected in data from 158 years of the Broadbalk Wheat Experiment at Rothamsted in south-east England. Specific weight from the Broadbalk Experiment responded to sunshine duration during grain filling and unconditional wet day probability during grain ripening in a similar way to the national average data. An association with the winter NAO was found in the Broadbalk data from 1956 to 2001, but not in the previous 112 years (1844–1955). This finding is consistent with other work showing significant correlations between the winter NAO and summer climate only in recent decades. It is concluded that the association between wheat quality and the NAO is a recent phenomenon.

Protocol for a population-based Ankylosing Spondylitis (PAS) cohort in Wales

BackgroundTo develop a population-based cohort of people with ankylosing spondylitis (AS) in Wales using (1) secondary care clinical datasets, (2) patient-derived questionnaire data and (3) routinely-collected information in order to examine disease history and the health economic cost of AS.MethodsThis data model will include and link (1) secondary care clinician datasets (i.e. electronic patient notes from the rheumatologist) (2) patient completed questionnaires (giving information on disease activity, medication, function, quality of life, work limitations and health service utilisation) and (3) a broad range of routinely collected data (including; GP records, in-patient hospital admission data, emergency department data, laboratory/pathology data and social services databases). The protocol involves the use of a unique and powerful data linkage system which allows datasets to be interlinked and to complement each other.DiscussionThis cohort can integrate patient supplied, primary and secondary care data into a unified data model. This can be used to study a range of issues such as; the true economic costs to the health care system and the patient, factors associated with the development of severe disease, long term adverse events of new and existing medication and to understand the disease history of this condition. It will benefit patients, clinicians and health care managers. This study forms a pilot project for the use of routine data/patient data linked cohorts for other chronic conditions.

Algorithms for the Capture and Adjudication of Prevalent and Incident Diabetes in UK Biobank.

Objectives UK Biobank is a UK-wide cohort of 502,655 people aged 40–69, recruited from National Health Service registrants between 2006–10, with healthcare data linkage. Type 2 diabetes is a key exposure and outcome. We developed algorithms to define prevalent and incident diabetes for UK Biobank. The algorithms will be implemented by UK Biobank and their results made available to researchers on request. Methods We used UK Biobank self-reported medical history and medication to assign prevalent diabetes and type, and tested this against linked primary and secondary care data in Welsh UK Biobank participants. Additionally, we derived and tested algorithms for incident diabetes using linked primary and secondary care data in the English Clinical Practice Research Datalink, and ran these on secondary care data in UK Biobank. Results and Significance For prevalent diabetes, 0.001% and 0.002% of people classified as “diabetes unlikely” in UK Biobank had evidence of diabetes in their primary or secondary care record respectively. Of those classified as “probable” type 2 diabetes, 75% and 96% had specific type 2 diabetes codes in their primary and secondary care records. For incidence, 95% of people with the type 2 diabetes-specific C10F Read code in primary care had corroborative evidence of diabetes from medications, blood testing or diabetes specific process of care codes. Only 41% of people identified with type 2 diabetes in primary care had secondary care evidence of type 2 diabetes. In contrast, of incident cases using ICD-10 type 2 diabetes specific codes in secondary care, 77% had corroborative evidence of diabetes in primary care. We suggest our definition of prevalent diabetes from UK Biobank baseline data has external validity, and recommend that specific primary care Read codes should be used for incident diabetes to ensure precision. Secondary care data should be used for incident diabetes with caution, as around half of all cases are missed, and a quarter have no corroborative evidence of diabetes in primary care.

Defining Disease Phenotypes in Primary Care Electronic Health Records by a Machine Learning Approach: A Case Study in Identifying Rheumatoid Arthritis.

Objectives 1) To use data-driven method to examine clinical codes (risk factors) of a medical condition in primary care electronic health records (EHRs) that can accurately predict a diagnosis of the condition in secondary care EHRs. 2) To develop and validate a disease phenotyping algorithm for rheumatoid arthritis using primary care EHRs. Methods This study linked routine primary and secondary care EHRs in Wales, UK. A machine learning based scheme was used to identify patients with rheumatoid arthritis from primary care EHRs via the following steps: i) selection of variables by comparing relative frequencies of Read codes in the primary care dataset associated with disease case compared to non-disease control (disease/non-disease based on the secondary care diagnosis); ii) reduction of predictors/associated variables using a Random Forest method, iii) induction of decision rules from decision tree model. The proposed method was then extensively validated on an independent dataset, and compared for performance with two existing deterministic algorithms for RA which had been developed using expert clinical knowledge. Results Primary care EHRs were available for 2,238,360 patients over the age of 16 and of these 20,667 were also linked in the secondary care rheumatology clinical system. In the linked dataset, 900 predictors (out of a total of 43,100 variables) in the primary care record were discovered more frequently in those with versus those without RA. These variables were reduced to 37 groups of related clinical codes, which were used to develop a decision tree model. The final algorithm identified 8 predictors related to diagnostic codes for RA, medication codes, such as those for disease modifying anti-rheumatic drugs, and absence of alternative diagnoses such as psoriatic arthritis. The proposed data-driven method performed as well as the expert clinical knowledge based methods. Conclusion Data-driven scheme, such as ensemble machine learning methods, has the potential of identifying the most informative predictors in a cost-effective and rapid way to accurately and reliably classify rheumatoid arthritis or other complex medical conditions in primary care EHRs.

Association of Diabetes in Pregnancy with Child Weight at Birth, Age 12 Months and 5 Years – A Population-Based Electronic Cohort Study

Background This study examines the effect of diabetes in pregnancy on offspring weight at birth and ages 1 and 5 years. Methods A population-based electronic cohort study using routinely collected linked healthcare data. Electronic medical records provided maternal diabetes status and offspring weight at birth and ages 1 and 5 years (n = 147,773 mother child pairs). Logistic regression models were used to obtain odds ratios to describe the association between maternal diabetes status and offspring size, adjusted for maternal pre-pregnancy weight, age and smoking status. Findings We identified 1,250 (0.9%) pregnancies with existing diabetes (27.8% with type 1 diabetes), 1,358 with gestational diabetes (0.9%) and 635 (0.4%) who developed diabetes post-pregnancy. Children whose mothers had existing diabetes were less likely to be large at 12 months (OR: 0.7 (95%CI: 0.6, 0.8)) than those without diabetes. Maternal diabetes was associated with high weight at age 5 years in children whose mothers had a high pre-pregnancy weight tertile (gestational diabetes, (OR:2.1 (95%CI:1.25–3.6)), existing diabetes (OR:1.3 (95%CI:1.0 to 1.6)). Conclusion The prevention of childhood obesity should focus on mothers with diabetes with a high maternal pre-pregnancy weight. We found little evidence that diabetes in pregnancy leads to long term obesity ‘programming’.