Publications
scientifiques

OBJECTIVE: Systemic autoimmune rheumatic diseases (SARD) are associated with an increased risk of premature cardiovascular disease (CVD) and all-cause mortality. We examined the potential survival benefit of statin use among patients with SARD in a general population setting. METHODS: We conducted an incident user cohort study using a UK general population database. Our population included patients with a SARD as determined by Read code diagnoses of systemic lupus erythematosus, systemic sclerosis, Sjögren syndrome, dermatomyositis, polymyositis, mixed connective tissue disease, Behçet disease, or antineutrophil cytoplasmic antibodies-associated vasculitis between January 1, 2000, and December 31, 2014. We compared propensity score-matched cohorts of statin initiators and noninitiators within 1-year cohort accrual blocks to account for potential confounders, including disease duration, body mass index, lifestyle factors, comorbidities, and medication use. RESULTS: Of 2305 statin initiators, 298 died during the followup period (mean 5.1 yrs), whereas among 2305 propensity score-matched noninitiators, 338 died during the followup period (mean 4.8 yrs). This corresponded to mortality rates of 25.4/1000 and 30.3/1000 person-years, respectively. Statin initiation was associated with reduced all-cause mortality (HR 0.84, 95% CI 0.72-0.98). When we compared the unmatched cohorts, the statin initiators (n = 2863) showed increased mortality (HR 1.85, 95% CI 1.58-2.16) compared with noninitiators (n = 2863 randomly selected within 1-year cohort accrual blocks) because of confounding by indication. CONCLUSION: In this general population-based study, statin initiation was shown to reduce overall mortality in patients with SARD after adjusting for relevant determinates of CVD risk.

Observational studies which evaluate effectiveness are often viewed with skepticism owing to the fact that patients are not randomized to treatment, meaning that results are more prone to bias. Therefore, randomized controlled trials remain the gold standard for evaluating treatment effectiveness. However, it is not always possible to conduct randomized trials. This may be due to financial constraints, for example, in identifying funding for a randomized trial for medicines that have already gained market authorization. There can also be challenges with recruitment, for example, of people with rare conditions or in hard-to-reach population subgroups. This is why observational studies are still needed. In this manuscript, we discuss how researchers can mitigate the risk of bias in the most common type of observational study design for evaluation of treatment effectiveness, the cohort study. We outline some key issues that warrant careful consideration at the outset when the question is being developed and the cohort study is being designed. We focus our discussion on the importance of deciding when to start follow-up in a study, choosing a comparator, managing confounding and measuring outcomes. We also illustrate the application of these considerations in a more detailed case study based on an examination of comparative effectiveness of two antidiabetic treatments using data collected during routine clinical practice.

OBJECTIVE: To determine the incidence of and risk factors for development of celiac disease (CD) in individuals with type 1 diabetes. METHODS: Cohort study using The Health Improvement Network (THIN), a UK primary care database of >13 million people. Individuals with incident type 1 diabetes diagnosed at 1 to 35 years of age between 1995 and 2015 with no previous diagnosis of CD were included. Cox regression was used to identify risk factors for CD, including age at diabetes diagnosis and sex, while adjusting for year of diagnosis to control for potential rising incidence in CD over time. RESULTS: Subjects (n = 9180; 43% female) had a median observation time of 5.1 years (interquartile range 2.0-10.1). CD was diagnosed in 196 (2%) during follow up. Median time to diagnosis was 2.1 years, but 25% were diagnosed more than 5 years after diabetes diagnosis. Incidence (per 10 000 person-years) was greater in females (43.0 [95% confidence interval [CI] 35.2-52.0]) vs males (26.8 [95% CI 21.5-32.9]). In multivariable Cox regression stratified by childhood- vs young adult-onset diabetes, younger age at diabetes diagnosis within childhood (hazard ratio [HR] 0.91 [95% CI 0.88-0.94]) and female sex among the adult-onset diabetes group (HR 3.19 [95% CI 1.39-7.34]) were associated with greater risk of CD. CONCLUSIONS: As expected, incidence of CD was higher in individuals with childhood-onset diabetes vs those with adult-onset diabetes. However, individuals with diabetes are at risk of developing CD throughout childhood and adulthood, and prolonged screening after diagnosis may be warranted. Prospective studies are needed in order to guide risk-stratified approaches to screening.

OBJECTIVE: To investigate trends in the incidence of imported malaria in the UK between 2005 and 2016. DESIGN: Analysis of longitudinal electronic health records (EHRs) in The Health Improvement Network (THIN) primary care database. SETTING: UK primary care. PARTICIPANTS: In total, we examined 12,349,003 individuals aged 0 to 99 years. OUTCOME MEASURE: The rate of malaria recordings in THIN was calculated per year between 2005 and 2016. Rate ratios exploring differences by age, sex, location of general practice, socioeconomic status and ethnicity were estimated using multivariable Poisson regression. RESULTS: A total of 1,474 individuals with a first diagnosis of malaria were identified in THIN between 2005 and 2016. The incidence of recorded malaria followed a decreasing trend dropping from a rate of 3.33 in 2005 to 1.36 cases per 100,000 person years at risk in 2016. Multivariable Poisson regression showed that adults of working age (20 to 69 years), men, those registered with a general practice in London, higher social deprivation and non-white ethnicity were associated with higher rates of malaria recordings. CONCLUSION: There has been a decrease in the number of malaria recordings in UK primary care over the past decade. This decrease exceeds the rate of decline reported in national surveillance data; however there are similar associations with age, sex and deprivation. Improved geographic information on the distribution of cases and the potential for automation of case identification suggests that EHRs could provide a complementary role for investigating malaria trends over time.

OBJECTIVE: To evaluate the patient pathways and associated health outcomes, resource use and corresponding costs attributable to managing unhealed surgical wounds in clinical practice, from initial presentation in the community in the UK. METHODS: This was a retrospective cohort analysis of the records of 707 patients in The Health Improvement Network (THIN) database whose wound failed to heal within 4 weeks of their surgery. Patients' characteristics, wound-related health outcomes and healthcare resource use were quantified, and the total National Health Service (NHS) cost of patient management was estimated at 2015/2016 prices. RESULTS: Inconsistent terminology was used in describing the wounds. 83% of all wounds healed within 12 months from onset of community management, ranging from 86% to 74% of wounds arising from planned and emergency procedures, respectively. Mean time to healing was 4 months per patient. Patients were predominantly managed in the community by nurses and only around a half of all patients who still had a wound at 3 months were recorded as having had a follow-up visit with their surgeon. Up to 68% of all wounds may have been clinically infected at the time of presentation, and 23% of patients subsequently developed a putative wound infection a mean 4 months after initial presentation. Mean NHS cost of wound care over 12 months was £7300 per wound, ranging from £6000 to £13 700 per healed and unhealed wound, respectively. Additionally, the mean NHS cost of managing a wound without any evidence of infection was ~£2000 and the conflated cost of managing a wound with a putative infection ranged from £5000 to £11 200. CONCLUSION: Surgeons are unlikely to be fully aware of the problems surrounding unhealed surgical wounds once patients are discharged into the community, due to inconsistent recording in patients' records coupled with the low rate of follow-up appointments. These findings offer the best evidence available with which to inform policy and budgetary decisions pertaining to managing unhealed surgical wounds in the community.

OBJECTIVE/BACKGROUND: In July 2013, new UK guidelines recommended that all patients with symptomatic varicose veins (VV) be referred to a specialist vascular service for consideration of superficial venous intervention (SVI). In the UK, general practitioners (GPs) in primary care control access to publicly funded vascular services provided through the National Health Service. GP awareness and concordance with Clinical Guideline (CG)168 recommendations is vital if patients with VV are to receive evidence-based treatment in line with national recommendations. The aim was to assess the UK-wide impact of new guidelines on GP management of VV using a large database of electronic GP records. METHODS: An eligible population of patients aged ≥ 18 years was analysed over two 18-month periods, before and after guideline publication. Those with a new diagnosis of VV in each time period were analysed in terms of demographics, specialist referral, compression hosiery prescriptions, and recorded SVI. RESULTS: Analysis included approximately two million patients from 285 GP practices. Before and after CG168 cohorts were well matched. Study populations included 13,014 patients before and 12,466 patients after guideline publication. There was an increase in specialist referrals from 24% (n = 3173) to 28% (n = 3457) (Cox hazard ratio [HR] 1.15, 95% confidence interval [CI] 1.09-1.20; p < .001). Median time to referral was 1.5 days. Prescribed compression hosiery declined from 20% (n = 2558) before the new guidelines to 18% (n = 2292) after the new guidelines (HR 0.93, 95% CI 0.88-0.98; p = .008). There were similar increases in proportions recorded as having SVI, from 3.6% (n = 469) before the new guidelines to 4.2% (n = 526) after the new guidelines (HR 1.16, 95% CI 1.02-1.31; p = .023). There was a statistically significant increase in endothermal ablation after CG168. In Cox models, age, sex, Townsend quintile, and body mass index were significantly related to the chance of referral and SVI. CONCLUSION: Encouragingly, following publication of National Institute for Health and Care Excellence CG168, there has been a statistically significant improvement in the management of VV in primary care in line with the CG recommendations.

A matched cohort study was conducted to determine the incidence of falls in patients following a diagnosis of COPD using a UK primary care database. 44 400 patients with COPD and 175 545 non-COPD subjects were identified. The incidence rate of fall per 1000 person-years in patients with COPD was higher (44.9; 95% CI 44.1 to 45.8) compared with non-COPD subjects (24.1; 95% CI 23.8 to 24.5) (P<0.0001). Patients with COPD were 55% more likely to have an incident record of fall than non-COPD subjects (adjusted HR, 1.55; 95% CI 1.50 to 1.59). The greater falls risk in patients with COPD needs consideration and modifiable factors addressed.

BACKGROUND: Use of selective COX-2 non-steroidal anti-inflammatory drugs (NSAIDs) (coxibs) has been associated with an increased risk of acute myocardial infarction (AMI). However, the risk of AMI has only been studied for very few NSAIDs that are frequently used. OBJECTIVES: To estimate the risk of AMI for individual NSAIDs. METHODS: A nested case-control study was performed from a cohort of new NSAID users ≥18 years (1999-2011) matching cases to a maximum of 100 controls on database, sex, age, and calendar time. Data were retrieved from six healthcare databases. Adjusted odds ratios (ORs) of current use of individual NSAIDs compared to past use were estimated per database. Pooling was done by two-stage pooling using a random effects model (ORmeta) and by one-stage pooling (ORpool). RESULTS: Among 8.5 million new NSAID users, 79,553 AMI cases were identified. The risk was elevated for current use of ketorolac (ORmeta 2.06;95%CI 1.83-2.32, ORpool 1.80; 1.49-2.18) followed, in descending order of point estimate, by indometacin, etoricoxib, rofecoxib, diclofenac, fixed combination of diclofenac with misoprostol, piroxicam, ibuprofen, naproxen, celecoxib, meloxicam, nimesulide and ketoprofen (ORmeta 1.12; 1.03-1.22, ORpool 1.00;0.86-1.16). Higher doses showed higher risk estimates than lower doses. CONCLUSIONS: The relative risk estimates of AMI differed slightly between 28 individual NSAIDs. The relative risk was highest for ketorolac and was correlated with COX-2 potency, but not restricted to coxibs.

OBJECTIVE: To test the hypothesis that measles infection increases the incidence of non-measles infectious diseases over a prolonged period of time. DESIGN: A population-based matched cohort study. DATA SOURCES: This study examined children aged 1-15 years in The Health Improvement Network UK general practice medical records database. Participants included 2228 patients diagnosed with measles between 1990 and 2014, which were matched on age, sex, general practitioner practice and calendar year with 19 930 children without measles. All controls had received at least one measles vaccination. Children with a history of immune-compromising conditions or with immune-suppressive treatment were excluded. PRIMARY OUTCOME MEASURES: Incidence rate ratio (IRR) of infections, anti-infective prescriptions and all-cause hospitalisations following measles in predetermined periods using multivariate analysis to adjust for confounding variables. RESULTS: In children with measles, the incidence rate for non-measles infectious disease was significantly increased in each time period assessed up to 5 years postmeasles: 43% in the first month (IRR: 1.43; 95% CI 1.22 to 1.68), 22% from month one to the first year (IRR: 1.22; 95% CI 1.14 to 1.31), 10% from year 1 to 2.5 years (IRR: 1.10; 95% CI 1.02 to 1.19) and 15% (IRR: 1.15; 95% CI 1.06 to 1.25) in years 2.5 to 5 years of follow-up. Children with measles were more than three times as likely to receive an anti-infective prescription in the first month and 15%-24% more likely between the first month and 5 years. The rate of hospitalisation in children with measles was increased only in the month following diagnosis but not thereafter (IRR: 2.83; 95% CI 1.72 to 4.67). CONCLUSION: Following measles, children had increased rates of diagnosed infections, requiring increased prescribing of antimicrobial therapies. This population-based matched cohort study supports the hypothesis that measles has a prolonged impact on host resistance to non-measles infectious diseases.

OBJECTIVES: Data derived from primary care electronic medical records (EMRs) are being used for research and surveillance. Case definitions are required to identify patients with specific conditions in EMR data with a degree of accuracy. The purpose of this study is to identify and provide a summary of case definitions that have been validated in primary care EMR data. MATERIALS AND METHODS: We searched MEDLINE and Embase (from inception to June 2016) to identify studies that describe case definitions for clinical conditions in EMR data and report on the performance metrics of these definitions. RESULTS: We identified 40 studies reporting on case definitions for 47 unique clinical conditions. The studies used combinations of International Classification of Disease version 9 (ICD-9) codes, Read codes, laboratory values, and medications in their algorithms. The most common validation metric reported was positive predictive value, with inconsistent reporting of sensitivity and specificity. DISCUSSION: This review describes validated case definitions derived in primary care EMR data, which can be used to understand disease patterns and prevalence among primary care populations. Limitations include incomplete reporting of performance metrics and uncertainty regarding performance of case definitions across different EMR databases and countries. CONCLUSION: Our review found a significant number of validated case definitions with good performance for use in primary care EMR data. These could be applied to other EMR databases in similar contexts and may enable better disease surveillance when using clinical EMR data. Consistent reporting across validation studies using EMR data would facilitate comparison across studies. SYSTEMATIC REVIEW REGISTRATION: PROSPERO CRD42016040020 (submitted June 8, 2016, and last revised June 14, 2016).

OBJECTIVES: Survival in infants with gastroschisis is increasing although little is known about early childhood morbidity. In the context of a hypothesized link between the gastrointestinal (GI) tract and immune function, this study explores rates of GI and respiratory infections in children with gastroschisis. METHODS: We conducted a population-based retrospective cohort study using data from the Health Improvement Network, a large database of UK primary care medical records. We identified children born from 1990 to 2013, and extracted follow-up data to their fifth birthday. We calculate incidence rates (IR) of GI and respiratory tract infections, overall and stratified by age, sex, socioeconomic status, and gestational age at birth, and compared these between children with and without gastroschisis by calculating adjusted incidence rate ratios (aIRR). RESULTS: Children with gastroschisis had a 65% higher IR of GI infection compared to children without (aIRR 1.65, 95% confidence interval [CI] 1.37-1.99, P < 0.001). Children with gastroschisis had a 27% higher IR of all respiratory tract infections (aIRR 1.27, 95% CI 1.12-1.44, P < 0.001) and more than 2-fold increase in lower respiratory tract infections compared to children without the condition (aIRR 2.15, 95% CI 1.69-2.74, P < 0.001). CONCLUSIONS: Children born with gastroschisis have a significantly higher incidence of GI and respiratory tract infections compared to children without gastroschisis. This association requires further investigations but could be related to the neonatal care they receive such as delayed enteral feeding or frequent antibiotic courses altering the gut microbiome and developing immune system.

BACKGROUND: The right dataset is essential to obtain the right insights in data science; therefore, it is important for data scientists to have a good understanding of the availability of relevant datasets as well as the content, structure, and existing analyses of these datasets. While a number of efforts are underway to integrate the large amount and variety of datasets, the lack of an information resource that focuses on specific needs of target users of datasets has existed as a problem for years. To address this gap, we have developed a Dataset Information Resource (DIR), using a user-oriented approach, which gathers relevant dataset knowledge for specific user types. In the present version, we specifically address the challenges of entry-level data scientists in learning to identify, understand, and analyze major datasets in healthcare. We emphasize that the DIR does not contain actual data from the datasets but aims to provide comprehensive knowledge about the datasets and their analyses. METHODS: The DIR leverages Semantic Web technologies and the W3C Dataset Description Profile as the standard for knowledge integration and representation. To extract tailored knowledge for target users, we have developed methods for manual extractions from dataset documentations as well as semi-automatic extractions from related publications, using natural language processing (NLP)-based approaches. A semantic query component is available for knowledge retrieval, and a parameterized question-answering functionality is provided to facilitate the ease of search. RESULTS: The DIR prototype is composed of four major components-dataset metadata and related knowledge, search modules, question answering for frequently-asked questions, and blogs. The current implementation includes information on 12 commonly used large and complex healthcare datasets. The initial usage evaluation based on health informatics novices indicates that the DIR is helpful and beginner-friendly. CONCLUSIONS: We have developed a novel user-oriented DIR that provides dataset knowledge specialized for target user groups. Knowledge about datasets is effectively represented in the Semantic Web. At this initial stage, the DIR has already been able to provide sophisticated and relevant knowledge of 12 datasets to help entry health informacians learn healthcare data analysis using suitable datasets. Further development of both content and function levels is underway.

IMPORTANCE: Clinicians are often cautious about use of allopurinol in patients with gout when renal function declines. OBJECTIVE: To assess the association of allopurinol use in gout with the risk of developing chronic kidney disease stage 3 or higher. DESIGN, SETTING, AND PARTICIPANTS: A time-stratified propensity score-matched, population-based, prospective cohort study of individuals with newly diagnosed gout who initiated allopurinol (≥300 mg/d) compared with those who did not initiate allopurinol, using the Health Improvement Network (THIN), a United Kingdom general practitioner electronic health records database, was carried out. The data were analyzed using Cox proportional hazards regression. Among adults aged 18 to 89 years with newly diagnosed gout, we propensity score matched 4760 initiators of allopurinol (≥300 mg/d) to the same number of noninitiators of allopurinol, excluding those with chronic kidney disease stage 3 or higher or urate-lowering therapy use before their gout diagnosis. EXPOSURES: Allopurinol initiation at a dose of 300 mg or more per day. MAIN OUTCOMES AND MEASURES: Development of chronic kidney disease stage 3 or higher. RESULTS: Of the 4760 allopurinol initiators (3975 men, 785 women) and same number of noninitiators (3971 men, 789 women), 579 and 623, respectively, developed chronic kidney disease stage 3 or higher, with a mean follow-up time of 5 and 4 years, mean age of 57 years, and mean body mass index (calculated as weight in kilograms divided by height in meters squared) of 30 for both groups. Use of allopurinol of at least 300 mg/d was associated with lower risk of developing chronic kidney disease stage 3 or higher compared with nonusers, with a hazard ratio (HR) of 0.87 (95% CI, 0.77-0.97). Allopurinol initiation at less than 300 mg/d was not associated with renal function decline (HR, 1.00; 95% CI, 0.91-1.09). CONCLUSIONS AND RELEVANCE: In this large cohort, allopurinol initiation of at least 300 mg/d was associated with a lower risk of renal function deterioration. Because allopurinol does not appear to be associated with renal function decline, clinicians should consider evaluating other potential causes when patients with gout experience renal function decline.