Representativeness and optimal use of body mass index (BMI) in the UK Clinical Practice Research Datalink (CPRD)

Prevalence of COVID-19-related risk factors and risk of severe influenza outcomes in cancer survivors: A matched cohort study using linked English electronic health records data

BACKGROUND

People with active cancer are recognised as at risk of COVID-19 complications, but it is unclear whether the much larger population of cancer survivors is at elevated risk. We aimed to address this by comparing cancer survivors and cancer-free controls for (i) prevalence of comorbidities considered risk factors for COVID-19; and (ii) risk of severe influenza, as a marker of susceptibility to severe outcomes from epidemic respiratory viruses.

METHODS

We included survivors (≥1 year) of the 20 most common cancers, and age, sex and general practice-matched cancer-free controls, derived from English primary care data linked to cancer registrations, hospital admissions and death registrations. Comorbidity prevalences were calculated 1 and 5 years from cancer diagnosis. Risk of hospitalisation or death due to influenza was compared using Cox models adjusted for baseline demographics and comorbidities.

FINDINGS

108,215 cancer survivors and 523,541 cancer-free controls were included. Cancer survivors had more diabetes, asthma, other respiratory, cardiac, neurological, renal, and liver diseases, and less obesity, compared with controls, but there was variation by cancer site. There were 205 influenza hospitalisations/deaths, with cancer survivors at higher risk than controls (adjusted HR 2.78, 95% CI 2.04-3.80). Haematological cancer survivors had large elevated risks persisting for >10 years (HR overall 15.17, 7.84-29.35; HR >10 years from cancer diagnosis 10.06, 2.47-40.93). Survivors of other cancers had evidence of raised risk up to 5 years from cancer diagnosis only (HR >5 years 2.22, 1.31-3.74).

INTERPRETATION

Risks of severe COVID-19 outcomes are likely to be elevated in cancer survivors. This should be taken into account in policies targeted at clinical risk groups, and vaccination for both influenza, and, when available, COVID-19, should be encouraged in cancer survivors.

Estimated impact of the COVID-19 pandemic on cancer services and excess 1-year mortality in people with cancer and multimorbidity: near real-time data on cancer care, cancer deaths and a population-based cohort study

OBJECTIVES

To estimate the impact of the COVID-19 pandemic on cancer care services and overall (direct and indirect) excess deaths in people with cancer.

METHODS

We employed near real-time weekly data on cancer care to determine the adverse effect of the pandemic on cancer services. We also used these data, together with national death registrations until June 2020 to model deaths, in excess of background (pre-COVID-19) mortality, in people with cancer. Background mortality risks for 24 cancers with and without COVID-19-relevant comorbidities were obtained from population-based primary care cohort (Clinical Practice Research Datalink) on 3 862 012 adults in England.

RESULTS

Declines in urgent referrals (median=-70.4%) and chemotherapy attendances (median=-41.5%) to a nadir (lowest point) in the pandemic were observed. By 31 May, these declines have only partially recovered; urgent referrals (median=-44.5%) and chemotherapy attendances (median=-31.2%). There were short-term excess death registrations for cancer (without COVID-19), with peak relative risk (RR) of 1.17 at week ending on 3 April. The peak RR for all-cause deaths was 2.1 from week ending on 17 April. Based on these findings and recent literature, we modelled 40% and 80% of cancer patients being affected by the pandemic in the long-term. At 40% affected, we estimated 1-year total (direct and indirect) excess deaths in people with cancer as between 7165 and 17 910, using RRs of 1.2 and 1.5, respectively, where 78% of excess deaths occured in patients with ≥1 comorbidity.

CONCLUSIONS

Dramatic reductions were detected in the demand for, and supply of, cancer services which have not fully recovered with lockdown easing. These may contribute, over a 1-year time horizon, to substantial excess mortality among people with cancer and multimorbidity. It is urgent to understand how the recovery of general practitioner, oncology and other hospital services might best mitigate these long-term excess mortality risks.

Validation of cardiovascular outcomes and risk factors in the Clinical Practice Research Datalink in the United Kingdom

Purpose

Strategies to identify and validate acute myocardial infarction (AMI) and stroke in primary‐care electronic records may impact effect measures, but to an unknown extent. Additionally, the validity of cardiovascular risk factors that could act as confounders in studies on those endpoints has not been thoroughly assessed in the United Kingdom Clinical Practice Research Datalink's (CPRD's) GOLD database. We explored the validity of algorithms to identify cardiovascular outcomes and risk factors and evaluated different outcome‐identification strategies using these algorithms for estimation of adjusted incidence rate ratios (IRRs).

Methods

First, we identified AMI, stroke, smoking, obesity, and menopausal status in a cohort treated for overactive bladder by applying computerized algorithms to primary care medical records (2004–2012). We validated these cardiovascular outcomes and risk factors with physician questionnaires (gold standard for this analysis). Second, we estimated IRRs for AMI and stroke using algorithm–identified and questionnaire–confirmed cases, comparing these with IRRs from cases identified through linkage with hospitalization/mortality data (best estimate).

Results

For AMI, the algorithm's positive predictive value (PPV) was >90%. Initial algorithms for stroke performed less well because of inclusion of codes for prevalent stroke; algorithm refinement increased PPV to 80% but decreased sensitivity by 20%. Algorithms for smoking and obesity were considered valid. IRRs based on questionnaire‐confirmed cases only were closer to IRRs estimated from hospitalization/mortality data than IRRs from algorithm‐identified cases.

Conclusions

AMI, stroke, smoking, obesity, and postmenopausal status can be accurately identified in CPRD. Physician questionnaire–validated AMI and stroke cases yield IRRs closest to the best estimate.

Use of Hydrochlorothiazide and Risk of Melanoma and Nonmelanoma Skin Cancer

INTRODUCTION

There are concerns that hydrochlorothiazide may increase the risk of incident nonmelanoma (cutaneous squamous cell carcinoma [cSCC], basal cell carcinoma [BCC]) and melanoma skin cancer, with regulatory agencies and societies calling for additional studies.

METHODS

We conducted a propensity score-matched population-based cohort study using the United Kingdom Clinical Practice Research Datalink. A total of 20,513 new users of hydrochlorothiazide were propensity score matched, in a 1:1 ratio, to new users of other thiazide diuretics between January 1, 1988 and March 31, 2018, with follow-up until March 31, 2019. Cox proportional hazards models were used to estimate hazard ratios (HRs) with 95% confidence intervals (CIs) for cSCC, BCC, and melanoma, comparing use of hydrochlorothiazide with use of other thiazide diuretics overall, by cumulative duration of use, and cumulative dose.

RESULTS

After an 8.6-year median follow-up, hydrochlorothiazide was associated with an increased risk of cSCC (HR 1.50, 95% CI 1.06-2.11). HRs increased with cumulative duration of use, with evidence of an association after 5-10 years (HR 2.10, 95% CI 1.20-3.67) and highest after > 10 years (HR 3.70, 95% CI 1.77-7.73). Similarly, HRs increased with cumulative dose, with higher estimates for ≥ 100,000 mg (HR 4.96, 95% CI 2.51-9.81). In contrast, hydrochlorothiazide was not associated with an increased risk of BCC (HR 1.01, 95% CI 0.91-1.13) or melanoma (HR 0.82, 95% CI 0.63-1.08), with no evidence of duration- or dose-response relationships.

CONCLUSIONS

Use of hydrochlorothiazide was associated with an increased risk of cSCC and with evidence of a duration- and dose-response relationship. In contrast, no association was observed for BCC or melanoma.

Determinants of referral of women with urinary incontinence to specialist services: a national cohort study using primary care data from the UK

BACKGROUND

Female urinary incontinence is underdiagnosed and undertreated in primary care. There is little evidence on factors that determine whether women with urinary incontinence are referred to specialist services. This study aimed to investigate characteristics associated with referrals from primary to specialist secondary care for urinary incontinence.

METHODS

We carried out a cohort study, using primary care data from over 600 general practices contributing to the Clinical Practice Research Datalink (CPRD) in the United Kingdom. We used multi-level logistic regression to estimate adjusted odds ratios (aOR) that reflect the impact of patient and GP practice-level characteristics on referrals to specialist services in secondary care within 30 days of a urinary incontinence diagnosis. All women aged ≥18 years newly diagnosed with urinary incontinence between 1 April 2004 and 31 March 2013 were included. One-year referral was estimated with death as competing event.

RESULTS

Of the 104,466 included women (median age: 58 years), 28,476 (27.3%) were referred within 30 days. Referral rates decreased with age (aOR 0.34, 95% CI 0.31-0.37, comparing women aged ≥80 with those aged 40-49 years) and was lower among women who were severely obese (aOR 0.84, 95% CI 0.78-0.90), smokers (aOR 0.94, 95% CI 0.90-0.98), women from a minority-ethnic backgrounds (aOR 0.76, 95% CI 0.65-0.89 comparing Asian with white women), women with pelvic organ prolapse (aOR 0.77, 95% CI 0.68-0.87), and women in Scotland (aOR 0.60, 95% CI 0.46-0.78, comparing women in Scotland and England). One-year referral rate was 34.0% and the pattern of associations with patient characteristics was almost the same as for 30-day referrals.

CONCLUSIONS

About one in four women with urinary incontinence were referred to specialist secondary care services within one month after a UI diagnosis and one in three within one year. Referral rates decreased with age which confirms concerns that older women with UI are less likely to receive care according to existing clinical guidelines. Referral rates were also lower in women from minority-ethnic backgrounds. These finding may reflect clinicians' beliefs about the appropriateness of referral, differences in women's preferences for treatment, or other factors leading to inequities in referral for urinary incontinence.

Stopping renin-angiotensin system blockers after acute kidney injury and risk of adverse outcomes: parallel population-based cohort studies in English and Swedish routine care

BACKGROUND

The safety of restarting angiotensin-converting enzyme inhibitors (ACEI) or angiotensin II receptor blockers (ARB) after acute kidney injury (AKI) is unclear. There is concern that previous users do not restart ACEI/ARB despite ongoing indications. We sought to determine the risk of adverse events after an episode of AKI, comparing prior ACEI/ARB users who stop treatment to those who continue.

METHODS

We conducted two parallel cohort studies in English and Swedish primary and secondary care, 2006-2016. We used multivariable Cox regression to estimate hazard ratios (HR) for hospital admission with heart failure (primary analysis), AKI, stroke, or death within 2 years after hospital discharge following a first AKI episode. We compared risks of admission between people who stopped ACEI/ARB treatment to those who were prescribed ACEI/ARB within 30 days of AKI discharge. We undertook sensitivity analyses, including propensity score-matched samples, to explore the robustness of our results.

RESULTS

In England, we included 7303 people with AKI hospitalisation following recent ACEI/ARB therapy for the primary analysis. Four thousand three (55%) were classified as stopping ACEI/ARB based on no prescription within 30 days of discharge. In Sweden, we included 1790 people, of whom 1235 (69%) stopped treatment. In England, no differences were seen in subsequent risk of heart failure (HR 1.10; 95% confidence intervals (CI) 0.93-1.30), AKI (HR 0.90; 95% CI 0.77-1.05), or stroke (HR 0.99; 95% CI 0.71-1.38), but there was an increased risk of death (HR 1.27; 95% CI 1.15-1.41) in those who stopped ACEI/ARB compared to those who continued. Results were similar in Sweden: no differences were seen in risk of heart failure (HR 0.91; 95% CI 0.73-1.13) or AKI (HR 0.81; 95% CI 0.54-1.21). However, no increased risk of death was seen (HR 0.94; 95% CI 0.78-1.13) and stroke was less common in people who stopped ACEI/ARB (HR 0.56; 95% CI 0.34-0.93). Results were similar across all sensitivity analyses.

CONCLUSIONS

Previous ACEI/ARB users who continued treatment after an episode of AKI did not have an increased risk of heart failure or subsequent AKI compared to those who stopped the drugs.

Rheumatic Conditions as Risk Factors for Self-Harm: A Retrospective Cohort Study

OBJECTIVE

To examine the risk of self-harm in rheumatic conditions.

METHODS

We conducted a retrospective cohort study using data from the Clinical Practice Research Datalink. Patients with ankylosing spondylitis, fibromyalgia, osteoarthritis, or rheumatoid arthritis were identified from 1990 to 2016 and matched to patients without these conditions. Incident self-harm was defined by medical record codes following a rheumatic diagnosis. Incidence rates (per 10,000 person-years) were reported for each condition, both overall and year-on-year (2000-2016). Cox regression analysis determined risk (hazard ratio [HR] and 95% confidence interval [95% CI]) of self-harm for each rheumatic cohort compared to the matched unexposed cohort. Initial crude analysis was subsequently adjusted and stratified by age and sex. Due to nonproportionality over time, osteoarthritis was also stratified by disease duration (<1 year, ≥1 to <5 years, ≥5 to <10 years, and ≥10 years).

RESULTS

The incidence of self-harm was highest in patients with fibromyalgia (HR 25.12 [95% CI 22.45-28.11] per 10,000 person-years) and lowest for osteoarthritis (HR 6.48 [95% CI 6.20-6.76]). There was a crude association with each rheumatic condition and self-harm, except for ankylosing spondylitis. Although attenuated, these associations remained after adjustment for fibromyalgia (HR 2.06 [95% CI 1.60-2.65]), rheumatoid arthritis (HR 1.59 [95% CI 1.20-2.11]), and osteoarthritis (1 to <5 years HR 1.12 [95% CI 1.01-1.24]; ≥5 to <10 years HR 1.35 [95% CI 1.18-1.54]). Age and sex were weak effect modifiers for these associations.

CONCLUSION

Primary care patients with fibromyalgia, osteoarthritis, or rheumatoid arthritis (but not ankylosing spondylitis) are at increased risk of self-harm compared to people without these rheumatic conditions. Clinicians need to be aware of the potential for self-harm in patients with rheumatic conditions (particularly fibromyalgia), explore mood and risk with them, and offer appropriate support and management.

Preschool Obesity Is Associated With an Increased Risk of Childhood Fracture: A Longitudinal Cohort Study of 466,997 Children and Up to 11 Years of Follow-up in Catalonia, Spain

This study aimed to determine if having an overweight or obese range body mass index (BMI) at time of beginning school is associated with increased fracture incidence in childhood. A dynamic cohort was created from children presenting for routine preschool primary care screening, collected in the Information System for Research in Primary Care (SIDIAP) platform in Catalonia, Spain. Data were collected from 296 primary care centers representing 74% of the regional pediatric population. A total of 466,997 children (48.6% female) with a validated weight and height measurement within routine health care screening at age 4 years (±6 months) between 2006 and 2013 were included, and followed up to the age of 15, migration out of region, death, or until December 31, 2016. BMI was calculated at age 4 years and classified using WHO growth tables, and fractures were identified using previously validated ICD10 codes in electronic primary care records, divided by anatomical location. Actuarial lifetables were used to calculate cumulative incidence. Cox regression was used to investigate the association of BMI category and fracture risk with adjustment for socioeconomic status, age, sex, and nationality. Median follow-up was 4.90 years (interquartile range [IQR] 2.50 to 7.61). Cumulative incidence of any fracture during childhood was 9.20% (95% confidence interval [CI] 3.79% to 14.61%) for underweight, 10.06% (9.82% to 10.29%) for normal weight, 11.28% (10.22% to 12.35%) for overweight children, and 13.05% (10.69% to 15.41%) for children with obesity. Compared with children of normal range weight, having an overweight and obese range BMI was associated with an excess risk of lower limb fracture (adjusted hazard ratio [HR] = 1.42 [1.26 to 1.59]; 1.74 [1.46 to 2.06], respectively) and upper limb fracture (adjusted HR = 1.10 [1.03 to 1.17]; 1.19 [1.07 to 1.31]). Overall, preschool children with an overweight or obese range BMI had increased incidence of upper and lower limb fractures in childhood compared with contemporaries of normal weight. © 2020 The Authors. Journal of Bone and Mineral Research published by American Society for Bone and Mineral Research.

Incretin-based drugs and risk of lung cancer among individuals with type 2 diabetes

AIM

To assess whether dipeptidyl peptidase-4 inhibitors and glucagon-like peptide-1 receptor agonists are associated with an increased lung cancer risk among individuals with type 2 diabetes.

METHODS

We conducted a population-based cohort study using the UK Clinical Practice Research Datalink. We identified 130 340 individuals newly treated with antidiabetes drugs between January 2007 and March 2017, with follow-up until March 2018. We used a time-varying approach to model use of dipeptidyl peptidase-4 inhibitors and glucagon-like peptide-1 receptor agonists compared with use of other second- or third-line antidiabetes drugs. We used Cox proportional hazards models to estimate the adjusted hazard ratios, with 95% CIs, of incident lung cancer associated with use of dipeptidyl peptidase-4 inhibitors and glucagon-like peptide-1 receptor agonists, separately, by cumulative duration of use, and by time since initiation.

RESULTS

A total of 790 individuals were newly diagnosed with lung cancer (median follow-up 4.6 years, incidence rate 1.5/1000 person-years, 95% CI 1.4-1.6). Compared with use of second-/third-line drugs, use of dipeptidyl peptidase-4 inhibitors and glucagon-like peptide-1 receptor agonists was not associated with an increased lung cancer risk (hazard ratio 1.07, 95% CI 0.87-1.32, and hazard ratio 1.02, 95% CI 0.68-1.54, respectively). There was no evidence of duration-response relationships.

CONCLUSIONS

In individuals with type 2 diabetes, use of incretin-based drugs was not associated with increased lung cancer risk.

Incidence of Depression and First-Line Antidepressant Therapy in People with Obesity and Depression in Primary Care

OBJECTIVE

The aim of this study was to describe the age- and gender-specific incidence of depression, the dose-response relationship between BMI and risk of depression (Cox proportional hazards), and antidepressant drug prescribing in adults with overweight or obesity.

METHODS

A retrospective electronic health record study using the Clinical Practice Research Datalink was conducted to identify adults with overweight and obesity (≥ 18 years) with incident depression (no prior depression diagnosis in their records), followed up from 2000 to 2019.

RESULTS

Among 519,513 adults, incidence of depression was 9.2 per 1,000 person-years and was higher in women and in 40- to 59-year-old men who had severe obesity. Compared with having overweight, the hazard of depression increased with each BMI category as follows: 1.13 (30-34 kg/m² ; 95% CI: 1.10-1.16), 1.34 (35-39 kg/m² ; 1.29-1.40), 1.51 (40-44 kg/m² ; 1.41-1.61), and 1.67 (45-49 kg/m² ; 1.48-1.87), attenuating at BMI 50+ kg/m² (1.54; 2.91-1.84). Antidepressants were prescribed as first-line therapy in two-thirds (66.3%) of cases. Prescriptions for fluoxetine reduced over time (20.4% [2000]; 8.8% [2018]), and prescriptions for sertraline increased (4.3% [2000]; 38.9% [2018]).

CONCLUSIONS

We recommend guidance on antidepressant drug prescribing and specific services for people with obesity and depression that address both symptoms and behaviors.

Can the Use of Bayesian Analysis Methods Correct for Incompleteness in Electronic Health Records Diagnosis Data? Development of a Novel Method Using Simulated and Real-Life Clinical Data

Background: Patient health information is collected routinely in electronic health records (EHRs) and used for research purposes, however, many health conditions are known to be under-diagnosed or under-recorded in EHRs. In research, missing diagnoses result in under-ascertainment of true cases, which attenuates estimated associations between variables and results in a bias toward the null. Bayesian approaches allow the specification of prior information to the model, such as the likely rates of missingness in the data. This paper describes a Bayesian analysis approach which aimed to reduce attenuation of associations in EHR studies focussed on conditions characterized by under-diagnosis.

Methods: Study 1: We created synthetic data, produced to mimic structured EHR data where diagnoses were under-recorded. We fitted logistic regression (LR) models with and without Bayesian priors representing rates of misclassification in the data. We examined the LR parameters estimated by models with and without priors. Study 2: We used EHR data from UK primary care in a case-control design with dementia as the outcome. We fitted LR models examining risk factors for dementia, with and without generic prior information on misclassification rates. We examined LR parameters estimated by models with and without the priors, and estimated classification accuracy using Area Under the Receiver Operating Characteristic.

Results: Study 1: In synthetic data, estimates of LR parameters were much closer to the true parameter values when Bayesian priors were added to the model; with no priors, parameters were substantially attenuated by under-diagnosis. Study 2: The Bayesian approach ran well on real life clinic data from UK primary care, with the addition of prior information increasing LR parameter values in all cases. In multivariate regression models, Bayesian methods showed no improvement in classification accuracy over traditional LR.

Conclusions: The Bayesian approach showed promise but had implementation challenges in real clinical data: prior information on rates of misclassification was difficult to find. Our simple model made a number of assumptions, such as diagnoses being missing at random. Further development is needed to integrate the method into studies using real-life EHR data. Our findings nevertheless highlight the importance of developing methods to address missing diagnoses in EHR data.

Interstitial lung disease is a risk factor for ischaemic heart disease and myocardial infarction

Objectives

Despite many shared risk factors and pathophysiological pathways, the risk of ischaemic heart disease (IHD) and myocardial infarction (MI) in interstitial lung disease (ILD) remains poorly understood. This lack of data could be preventing patients who may benefit from screening for these cardiovascular diseases from receiving it.

Methods

A population-based cohort study used electronic patient records from the Clinical Practice Research Datalink and linked Hospital Episode Statistics to identify 68 572 patients (11 688 ILD exposed (mean follow-up: 3.8 years); 56 884 unexposed controls (mean follow-up: 4.0 years), with 349 067 person-years of follow-up. ILD-exposed patients (pulmonary sarcoidosis (PS) or idiopathic pulmonary fibrosis (PF)) were matched (by age, sex, registered general practice and available follow-up time) to patients without ILD or IHD/MI. Rates of incident MI and IHD were estimated. HRs were modelled using multivariable Cox proportional hazards regression accounting for potential confounders.

Results

ILD was independently associated with IHD (HR 1.85, 95% CI 1.56 to 2.18) and MI (HR 1.74, 95% CI 1.44 to 2.11). In all disease categories, risk of both IHD and MI peaked between ages 60 and 69 years, except for the risk of MI in PS which was greatest <50 years. Men with PF were at greatest risk of IHD, while women with PF were at greatest risk of MI.

Conclusions

ILD, particularly PF, is independently associated with MI and IHD after adjustment for established cardiovascular risk factors. Our results suggest clinicians should prioritise targeted assessment of cardiovascular risk in patients with ILD, particularly those aged 60–69 years. Further research is needed to understand the impact of such an approach to risk management.

Data-driven discovery of changes in clinical code usage over time: a case-study on changes in cardiovascular disease recording in two English electronic health records databases (2001-2015)

OBJECTIVES

To demonstrate how data-driven variability methods can be used to identify changes in disease recording in two English electronic health records databases between 2001 and 2015.

DESIGN

Repeated cross-sectional analysis that applied data-driven temporal variability methods to assess month-by-month changes in routinely collected medical data. A measure of difference between months was calculated based on joint distributions of age, gender, socioeconomic status and recorded cardiovascular diseases. Distances between months were used to identify temporal trends in data recording.

SETTING

400 English primary care practices from the Clinical Practice Research Datalink (CPRD GOLD) and 451 hospital providers from the Hospital Episode Statistics (HES).

MAIN OUTCOMES

The proportion of patients (CPRD GOLD) and hospital admissions (HES) with a recorded cardiovascular disease (CPRD GOLD: coronary heart disease, heart failure, peripheral arterial disease, stroke; HES: International Classification of Disease codes I20-I69/G45).

RESULTS

Both databases showed gradual changes in cardiovascular disease recording between 2001 and 2008. The recorded prevalence of included cardiovascular diseases in CPRD GOLD increased by 47%-62%, which partially reversed after 2008. For hospital records in HES, there was a relative decrease in angina pectoris (-34.4%) and unspecified stroke (-42.3%) over the same time period, with a concomitant increase in chronic coronary heart disease (+14.3%). Multiple abrupt changes in the use of myocardial infarction codes in hospital were found in March/April 2010, 2012 and 2014, possibly linked to updates of clinical coding guidelines.

CONCLUSIONS

Identified temporal variability could be related to potentially non-medical causes such as updated coding guidelines. These artificial changes may introduce temporal correlation among diagnoses inferred from routine data, violating the assumptions of frequently used statistical methods. Temporal variability measures provide an objective and robust technique to identify, and subsequently account for, those changes in electronic health records studies without any prior knowledge of the data collection process.

Atopic Eczema in Adulthood and Risk of Depression and Anxiety: A Population-Based Cohort Study

BACKGROUND

Atopic eczema is a common and debilitating condition associated with depression and anxiety, but the nature of this association remains unclear.

OBJECTIVE

To explore the temporal relationship between atopic eczema and new depression/anxiety.

METHODS

This matched cohort study used routinely collected data from the UK Clinical Practice Research Datalink, linked to hospital admissions data. We identified adults with atopic eczema (1998-2016) using a validated algorithm, and up to 5 individuals without atopic eczema matched on date of diagnosis, age, sex, and general practice. We estimated the hazard ratio (HR) for new depression/anxiety using stratified Cox regression to account for age, sex, calendar period, Index of Multiple Deprivation, glucocorticoid treatment, obesity, smoking, and harmful alcohol use.

RESULTS

We identified 526,808 adults with atopic eczema who were matched to 2,569,030 without. Atopic eczema was associated with increased incidence of new depression (HR, 1.14; 99% CI, 1.12-1.16) and anxiety (HR, 1.17; 99% CI, 1.14-1.19). We observed a stronger effect of atopic eczema on depression with increasing atopic eczema severity (HR [99% CI] compared with no atopic eczema: mild, 1.10 [1.08-1.13]; moderate, 1.19 [1.15-1.23]; and severe, 1.26 [1.17-1.37]). A dose-response association, however, was less apparent for new anxiety diagnosis (HR [99% CI] compared with no atopic eczema: mild, 1.14 [1.11-1.18]; moderate, 1.21 [1.17-1.26]; and severe, 1.15; [1.05-1.25]).

CONCLUSIONS

Adults with atopic eczema are more likely to develop new depression and anxiety. For depression, we observed a dose-response relationship with atopic eczema severity.

Characteristics of patients with body mass index recorded within the Kent Integrated Dataset (KID)

Background

Obesity is a significant health issue and key public health priority. This study explored body mass index (BMI) recording in general practice within the Kent Integrated Dataset.

Methods

Using a sample aged 18–100 years, resident within Kent, who were alive and currently registered to a Kent general practice as of 6 August 2018 within the Kent Integrated Dataset. We identified the latest BMI from event records between 2015/2016 and 2017/2018. Recording was evaluated by sex, age, deprivation, hypertension, serious mental illness and multimorbidity.

Results

Between 2015/2016 and 2017/2018 using the sample of 1 154 652 persons, BMI was recorded for 43.7% of the sample. Multiple logistic regression showed that BMI recording was higher in females, the middle age bands, persons living in the most deprived areas and within persons who were hypertensive, had serious mental illness or were multimorbid.

Conclusions

Findings were aligned to previous research using nationally representative samples. Completeness of recording varied by age, sex, deprivation and comorbidity. Recording within general practice was aligned to chronic disease management. From a prevention perspective, earlier assessment and intervention for the management of excess weight within primary care may be an opportunity for avoiding increases in BMI trajectory. There may also be merit in recognising that the external disease agents that influence obesity can be controlled or reduced (obesogenic environment) from a national policy perspective. Such a perspective may also help reduce stigmatisation and the pressure around arguments that centre on personal responsibility for obesity.

Use of oral anticoagulants in older people with atrial fibrillation in UK general practice: protocol for a cohort study using the Clinical Practice Research Datalink (CPRD) database

INTRODUCTION

Warfarin has frequently been underused in older people for stroke prevention in atrial fibrillation (AF). Direct oral anticoagulants (DOACs) entered the UK market from 2008 and have been recommended as an alternative to warfarin. This study aimed to describe any changes in the prescribing of oral anticoagulants (OACs) to people aged ≥75 years in UK general practice before and after the introduction of DOACs, to examine differences in patient characteristics which may influence prescribers' decisions regarding anticoagulation, to evaluate the time people stay on OACs and switching between OACs.

METHODS AND ANALYSIS

A retrospective cohort study design will be used. Patients with a diagnosis of AF will be identified from the Clinical Practice Research Datalink (CPRD). The study period will run from 1 January 2003 to 27 December 2017. Patients enter the cohort at the latest date of the start of the study period, first AF diagnosis, 75th birthday or a year from when they started to contribute research standard data. Follow-up continues until they leave the practice, death, the date the practice stops contributing research standard data or the end of the study period (27 December 2017). Exposure to OACs will be defined as ≥1 prescription issued for an OAC of interest during the study period. Patients issued an OAC in the year preceding study entry will be defined as 'prevalent users'. Patients starting on an OAC during the study period will be defined as 'incident users'. Incidence and prevalence of OAC prescribing, patient demographics and characteristics will be described during three time periods: 2003-2007, 2008-2012 and 2013-2017. Persistence (defined as the time from initiation to discontinuation of medication) with and switching between different OACs will be described.

ETHICS AND DISSEMINATION

The protocol for this study was approved by the CPRD Independent Scientific Advisory Committee. The results will be disseminated in a peer-reviewed journal and at conferences.

TRIAL REGISTRATION NUMBER

EUPAS29923.

UK phenomics platform for developing and validating electronic health record phenotypes: CALIBER

OBJECTIVE

Electronic health records (EHRs) are a rich source of information on human diseases, but the information is variably structured, fragmented, curated using different coding systems, and collected for purposes other than medical research. We describe an approach for developing, validating, and sharing reproducible phenotypes from national structured EHR in the United Kingdom with applications for translational research.

MATERIALS AND METHODS

We implemented a rule-based phenotyping framework, with up to 6 approaches of validation. We applied our framework to a sample of 15 million individuals in a national EHR data source (population-based primary care, all ages) linked to hospitalization and death records in England. Data comprised continuous measurements (for example, blood pressure; medication information; coded diagnoses, symptoms, procedures, and referrals), recorded using 5 controlled clinical terminologies: (1) read (primary care, subset of SNOMED-CT [Systematized Nomenclature of Medicine Clinical Terms]), (2) International Classification of Diseases-Ninth Revision and Tenth Revision (secondary care diagnoses and cause of mortality), (3) Office of Population Censuses and Surveys Classification of Surgical Operations and Procedures, Fourth Revision (hospital surgical procedures), and (4) DM+D prescription codes.

RESULTS

Using the CALIBER phenotyping framework, we created algorithms for 51 diseases, syndromes, biomarkers, and lifestyle risk factors and provide up to 6 validation approaches. The EHR phenotypes are curated in the open-access CALIBER Portal (https://www.caliberresearch.org/portal) and have been used by 40 national and international research groups in 60 peer-reviewed publications.

CONCLUSIONS

We describe a UK EHR phenomics approach within the CALIBER EHR data platform with initial evidence of validity and use, as an important step toward international use of UK EHR data for health research.

Determinants and extent of weight recording in UK primary care: an analysis of 5 million adults' electronic health records from 2000 to 2017

BACKGROUND

Excess weight and unexpected weight loss are associated with multiple disease states and increased morbidity and mortality, but weight measurement is not routine in many primary care settings. The aim of this study was to characterise who has had their weight recorded in UK primary care, how frequently, by whom and in relation to which clinical events, symptoms and diagnoses.

METHODS

A longitudinal analysis of UK primary care electronic health records (EHR) data from 2000 to 2017. Descriptive statistics were used to summarise weight recording in terms of patient sociodemographic characteristics, health professional encounters, clinical events, symptoms and diagnoses. Negative binomial regression was used to model the likelihood of having a weight record each year, and Cox regression to the likelihood of repeated weight recording.

RESULTS

A total of 14,049,871 weight records were identified in the EHR of 4,918,746 patients during the study period, representing 26,998,591 person-years of observation. Around a third of patients had a weight record each year. Forty-nine percent of weight records were repeated within a year with an average time to a repeat weight record of 1.92 years. Weight records were most often taken by nursing staff (38-42%) and GPs (37-39%) as part of a routine clinical care, such as chronic disease reviews (16%), medication reviews (6-8%) and health checks (6-7%), or were associated with consultations for contraception (5-8%), respiratory disease (5%) and obesity (1%). Patient characteristics independently associated with an increased likelihood of weight recording were as follows: female sex, younger and older adults, non-drinkers, ex-smokers, low or high BMI, being more deprived, diagnosed with a greater number of comorbidities and consulting more frequently. The effect of policy-level incentives to record weight did not appear to be sustained after they were removed.

CONCLUSION

Weight recording is not a routine activity in UK primary care. It is recorded for around a third of patients each year and is repeated on average every 2 years for these patients. It is more common in females with higher BMI and in those with comorbidity. Incentive payments and their removal appear to be associated with increases and decreases in weight recording.

Completeness and validity of alcohol recording in general practice within the UK: a cross-sectional study

BACKGROUND

Since 2010 the National Institute for Health and Care Excellence has recommended screening adults for excessive alcohol consumption to try and help prevent alcohol-use disorders. Little is known about the extent to which these recommendations are followed, and the resulting completeness and validity of alcohol-related data recording in primary care.

OBJECTIVE

To investigate the completeness and accuracy of recording of alcohol use within primary care records in the UK.

DESIGN AND SETTING

Cross-sectional study in the Clinical Practice Research Datalink.

PARTICIPANTS

We included all adult patients registered on 1^st January 2018 with ≥1 year of follow-up.

PRIMARY AND SECONDARY OUTCOME MEASURES

We calculated prevalence of alcohol consumption recording overall and within patient groups. We then validated alcohol consumption data against recommended screening tools (Alcohol Use Disorders Identification Test (AUDIT)) as the gold standard. We also calculated how prevalence of alcohol recording changed over the preceding decade.

RESULTS

In 2018, among 1.8 million registered adult patients, just over half (51.9%) had a record for a code related to alcohol in the previous 5 years. Recording of alcohol consumption was more common among women, older people, ex-smokers and those from more deprived areas, who were overweight/obese, or with comorbidities. A quarter of patients had units per week recorded in the last 5 years, but <10% had an AUDIT or Fast Alcohol Screening Test (FAST) alcohol screening test score. The recorded alcohol measures corresponded to results from gold standard AUDIT scores. The distribution of consumption among current drinkers was similar to the Health Survey for England.

CONCLUSIONS

Half of adults in UK primary care have no recorded alcohol consumption data. When consumption is recorded, we have demonstrated internal and external validity of the data, suggesting greater recording may help identify opportunities for interventions to reduce harms.

Real-World Associations of Renin-Angiotensin-Aldosterone System Inhibitor Dose, Hyperkalemia, and Adverse Clinical Outcomes in a Cohort of Patients With New-Onset Chronic Kidney Disease or Heart Failure in the United Kingdom

Background

Dosing of renin–angiotensin–aldosterone system inhibitors (RAASi) may be modified to manage associated hyperkalemia risk; however, this approach could adversely affect cardiorenal outcomes. This study investigated real‐world associations of RAASi dose, hyperkalemia, and adverse clinical outcomes in a large cohort of UK cardiorenal patients.

Methods and Results

This observational study included RAASi‐prescribed patients with new‐onset chronic kidney disease (n=100 572) or heart failure (n=13 113) first recorded between January 2006 and December 2015 in Clinical Practice Research Datalink and linked Hospital Episode Statistics databases. Odds ratios associating hyperkalemia and RAASi dose modification were estimated using logistic generalized estimating equations with normal (<5.0 mmol/L) serum potassium level as the reference category. Patients with serum potassium ≥5.0 mmol/L had higher risk of RAASi down‐titration (adjusted odds ratios, chronic kidney disease: 1.79 [95% CI, 1.64–1.96]; heart failure: 1.33 [95% CI, 1.08–1.62]). Poisson models were used to estimate adjusted incident rate ratios of adverse outcomes based on total RAASi exposure (<50% and ≥50% of the guideline‐recommended RAASi dose). Incidence of major adverse cardiac events and mortality was consistently higher in the lower dose group (adjusted incident rate ratios: chronic kidney disease: 5.60 [95% CI, 5.29–5.93] for mortality and 1.60 [95% CI, 1.55–1.66] for nonfatal major adverse cardiac events; heart failure: 7.34 [95% CI, 6.35–8.48] for mortality and 1.85 [95% CI, 1.71–1.99] for major adverse cardiac events).

Conclusions

The results of this real‐world analysis highlight the potential negative impact of suboptimal RAASi dosing and the need for strategies that allow patients to be maintained on appropriate therapy, avoiding RAASi dose modification or discontinuation.

Noncommunicable disease among adults with cerebral palsy: A matched cohort study

OBJECTIVE

To compare the incidence of noncommunicable diseases between adults with and without cerebral palsy (CP).

METHODS

A cohort study was conducted using primary care data from the Clinical Practice Research Datalink. Cox models, stratified by matched set and adjusted for potential confounders, were fitted to compare the risk of any noncommunicable disease, cancer, cardiovascular disease, type 2 diabetes mellitus, and respiratory disease between adults with and without CP.

RESULTS

The analysis included 1,705 adults with CP and 5,115 age-, sex-, and general practice-matched adults without CP. There was evidence from adjusted analyses that adults with CP had 75% increased risk of developing any noncommunicable disease compared to adults without CP (hazard ratio [HR] 1.75, 95% confidence interval [CI] 1.58-1.94). Specifically, they had increased risk of cardiovascular disease (HR 1.76, 95% CI 1.48-2.11) and respiratory disease (HR 2.61, 95% CI 2.14-3.19). There was no evidence of increased risk of cancer or type 2 diabetes mellitus.

CONCLUSIONS

Adults with CP had increased risk of noncommunicable disease, specifically cardiovascular and respiratory disease. These findings highlight the need for clinical vigilance regarding identification of noncommunicable disease in people with CP and further research into the etiology and management of noncommunicable disease in this population.

The comparative effectiveness of fourth-line drugs in resistant hypertension: An application in electronic health record data

PURPOSE

To examine the utility of electronic health records from a routine care setting in assessing comparative effectiveness of fourth-line anti-hypertensive drugs to treat resistant hypertension.

METHODS

We conducted a cohort study using the Clinical Practice Research Datalink: a repository of electronic health records from UK primary care. We identified patients newly prescribed fourth-line anti-hypertensive drugs (aldosterone antagonist , beta-blocker, or alpha-blocker). Using propensity score-adjusted Cox proportional hazards models, we compared the incidence of the primary outcome (composite of all-cause mortality, stroke, and myocardial infarction) between patients on different fourth-line drugs. AA was the reference drug in all comparisons. Secondary outcomes were individual components of the primary outcome, blood pressure changes, and heart failure. We used a negative control outcome, Herpes Zoster, to detect unmeasured confounding.

RESULTS

Overall, 8639 patients were included. In propensity score-adjusted analyses, the hazard ratio for the primary outcome was 0.81 (95% CI, 0.55-1.19) for beta-blockers and 0.68 (95% CI, 0.46-0.96) for alpha-blockers versus AA. Findings for individual cardiovascular outcomes trended in a more plausible direction, albeit imprecise. A trend for a protective effect for Herpes Zoster across both comparisons was seen.

CONCLUSIONS

A higher rate of all-cause death in the AA group was likely due to unmeasured confounding in our analysis of the composite primary outcome, supported by our negative outcome analysis. Results for cardiovascular outcomes were plausible, but imprecise due to small cohort sizes and a low number of observed outcomes.

Descriptive epidemiology of hip and knee replacement in rheumatoid arthritis: An analysis of UK electronic medical records

OBJECTIVE

To provide descriptive data on rates of total hip replacement (THR) and total knee replacement (TKR) within a large RA cohort and describe variation in risk.

METHODS

Incident RA patients (1995 to 2014) were identified from the Clinical Practice Research Datalink (CPRD). First subsequent occurrence of THR and TKR were identified (analysed separately) and incidence rates calculated, stratified by sex, age, BMI, geographic region, and quintiles of the index of multiple deprivation (IMD) score.

RESULTS

There were 27,607 RA patients included, with a total of 1,028 THRs (mean age at surgery: 68.4 years) and 1,366 TKRs (mean age at surgery: 67.6 years), at an overall incidence rate per 1,000 person-years (PYs) [95% CI] of 6.38 [6.00-6.78] and 8.57 [8.12-9.04], respectively. TKR incidence was similar by gender but THR rates were higher in females than males. Rates of TKR but not THR rose according to BMI. An increasing trend was observed in rates of both outcomes according to age (although not ≥75) but of decreasing rates according to socio-economic deprivation. There was some evidence for regional variation in TKR. The 10-year cumulative incidence was 5.2% [4.9, 5.6] and 7.0% [6.6, 7.4] for THR and TKR, respectively.

CONCLUSION

We provide generalizable estimates of THR and TKR incidence in the UK RA patient population and note variation across several key variables. Increased BMI was associated with a large increase in TKR but not THR incidence. Increased deprivation was associated with a downward trend in rates of THR and TKR.

Does the ethnic density effect extend to obesity? A cross-sectional study of 415 166 adults in east London

OBJECTIVES

To examine the prevalence of obesity by ethnic group and to examine the association between ethnic density and obesity prevalence.

DESIGN AND SETTING

Cross-sectional study utilising electronic primary care records of 128 practices in a multiethnic population of east London.

PARTICIPANTS

Electronic primary care records of 415 166 adults with a body mass index recorded in the previous 3 years.

OUTCOME MEASURES

(1) Odds of obesity for different ethnic groups compared with white British. (2) Prevalence of obesity associated with each 10% increase in own-group ethnic density, by ethnic group.

RESULTS

Using multilevel logistic regression models, we find that compared with white British/Irish males, the odds of obesity were significantly higher among black ethnic groups and significantly lower among Asian and white other groups. Among females, all ethnic groups except Chinese and white other were at increased odds of obesity compared with white British/Irish. There was no association between increasing ethnic density and obesity prevalence, except among black Africans and Indian females. A 10% increase in black ethnic density was associated with a 15% increase in odds of obesity among black African males (95% CI 1.07 to 1.24) and 18% among black African females (95% CI 1.08 to 1.30). Among Indian females, a 10% increase in Indian ethnic density was associated with a 7% decrease in odds of obesity (95% CI 0.88 to 0.99).

CONCLUSION

Wider environmental factors play a greater role in determining obesity than the ethnic composition of the area for most ethnic groups. Further research is needed to understand the mechanism through which increasing ethnic density is associated with increased odds of obesity among black Africans and decreased odds of obesity among Indian females.

Changing prevalence of current asthma and inhaled corticosteroid treatment in the UK: population-based cohort 2006-2016

Asthma is the most common respiratory disorder in the UK, yet we have incomplete knowledge on the prevalence of current disease, treatment and exacerbations.We used UK electronic healthcare records, 2006-2016, to estimate the prevalence of current asthma by year, sex and age (<5, 5-11, 12-17, 18-24, 25-54 and ≥55 years), and the proportion prescribed inhaled corticosteroids (ICS) and additional asthma therapy, treated for exacerbations and other asthma care markers.Overall current asthma prevalence was 6.5% in 2016 (7.2% in 2006). Prevalence fell in those aged <45 years. The lowest prevalence and largest absolute decrease was in children aged <5 years. In 2016, 80% of current asthma patients were managed on ICS (65% in 2006); this increase occurred in all age groups, primarily due to an increase in low-dose ICS. During this time there was an increase in all age groups in the proportion prescribed additional asthma therapy, treated for an exacerbation within primary care and given an annual asthma review or management plan. Hospitalised exacerbations showed minimal change over time.Asthma remains highly prevalent and a significant healthcare burden. In those with a diagnosis, there was an increase in ICS prescriptions and treatment of exacerbations across all age groups. This may reflect a trend towards more aggressive asthma management within primary care. An apparent decline in prevalence was observed in those aged <45 years, particularly in children aged <5 years.

Body mass index and Hodgkin's lymphoma: UK population-based cohort study of 5.8 million individuals

Previous epidemiological studies describe a positive association between body mass index (BMI) and Hodgkin’s lymphoma, mainly in obese vs. normal weight individuals. We examined the shape of this relationship in individuals aged 16 years or older, using primary care data from the United Kingdom’s Clinical Practice Research Datalink. Cox models were fitted with linear, non-linear (spline) and categorical BMI. Models were adjusted for potential confounders and effect modification was investigated. Five point eight two million patients were included, 927 of whom developed Hodgkin’s lymphoma during 41.6 million years of follow-up. Each 5 kg/m² increase in BMI was associated with a 10% increase in Hodgkin’s lymphoma (95% confidence intervals: 2–19). Analysis of non-linearity suggested a J-shaped association with incidence increasing with BMI above 24.2 kg/m². Seven point four per cent of adult Hodgkin’s lymphoma cases were estimated to be attributable to excess weight. Our findings suggest a pattern of increasing risk beyond the World Health Organisation healthy weight category in the general population.

Health indicator recording in UK primary care electronic health records: key implications for handling missing data

Background

Clinical databases are increasingly used for health research; many of them capture information on common health indicators including height, weight, blood pressure, cholesterol level, smoking status, and alcohol consumption. However, these are often not recorded on a regular basis; missing data are ubiquitous. We described the recording of health indicators in UK primary care and evaluated key implications for handling missing data.

Methods

We examined the recording of health indicators in The Health Improvement Network (THIN) UK primary care database over time, by demographic variables (age and sex) and chronic diseases (diabetes, myocardial infarction, and stroke). Using weight as an example, we fitted linear and logistic regression models to examine the associations of weight measurements and the probability of having weight recorded with individuals’ demographic characteristics and chronic diseases.

Results

In total, 6,345,851 individuals aged 18–99 years contributed data to THIN between 2000 and 2015. Women aged 18–65 years were more likely than men of the same age to have health indicators recorded; this gap narrowed after age 65. About 60–80% of individuals had their height, weight, blood pressure, smoking status, and alcohol consumption recorded during the first year of registration. In the years following registration, these proportions fell to 10%–40%. Individuals with chronic diseases were more likely to have health indicators recorded, particularly after the introduction of a General Practitioner incentive scheme. Individuals’ demographic characteristics and chronic diseases were associated with both observed weight measurements and missingness in weight.

Conclusion

Missing data in common health indicators will affect statistical analysis in health research studies. A single analysis of primary care data using the available information alone may be misleading. Multiple imputation of missing values accounting for demographic characteristics and disease status is recommended but should be considered and implemented carefully. Sensitivity analysis exploring alternative assumptions for missing data should also be evaluated.

Identifying Positive Adaptive Pathways in Low-Income Families in Singapore: Protocol for Sequential, Longitudinal Mixed-Methods Design

Background

This study aims to examine the adaptive process of children and mothers from multistressed low-income families in Singapore. It aims to bridge the knowledge gap left by existing poverty studies, which are predominately risk focused. Through a sequential longitudinal mixed-methods design, we will differentiate children and mothers who demonstrate varied social, developmental, and mental health trajectories of outcomes. Through utilizing the Latent Growth Curve Model (LGCM), we aim to detect the development and changes of the positive Family Agency and adaptive capacities of these families over time. The construct of Family Agency is underpinned by the theoretical guidance from the Social Relational Theory, which examines child agency, parent agency, relational agency, and the interactions among these members. It is hypothesized that positive Family Agency within low-income families may lead to better outcomes. The key research questions include whether the extent of positive Family Agency mediates the relationship among financial stress, resource utilization, home environment, and parental stress.

Objective

The study elucidates the Family Agency construct through interviews with mother-child dyads. It also aims to understand how financial stress and resources are differentially related to home environment, parent stress, and parent and child outcomes.

Methods

In phase 1, 60 mother-child dyads from families receiving government financial assistance and with children aged between 7 and 12 years will be recruited. In-depth interviews will be conducted separately with mothers and children. On the basis of 120 interviews, a measurement for the construct of Family Agency will be developed and will be pilot tested. In phase 2a, a longitudinal survey will be conducted over 3 time points from 800 mother-child dyads. The 3 waves of survey results will be analyzed by LGCM to identify the trajectories of adaptation pathways of these low-income families. In addition, 10 focus groups with up to 15 participants in each will be conducted to validate the LGCM results.

Results

This project is funded by the Social Science Research Thematic Grant (Singapore). The recruitment of 60 mother-child dyads has been achieved. Data collection will commence once the amendment to the protocol has been approved by the Institutional Review Board. Analysis of phase 1 data will be completed by the end of the first quarter of 2019, and the first set of results is expected to be submitted for publication by the second quarter of 2019. Phase 2 implementation will commence in the second quarter of 2019, and the project end date is in May 2021.

Conclusions

Findings from this study can potentially inform social policy and programs as it refines the understanding of low-income families by distinguishing trajectories of adaptive capacities so that policies and interventions can be targeted in enhancing the adaptive pathways of low-income families with children.

International Registered Report Identifier (IRRID)

PRR1-10.2196/11629

The internal validation of weight and weight change coding using weight measurement data within the UK primary care Electronic Health Record

PURPOSE

To use recorded weight values to internally validate weight status and weight change coding in the primary care Electronic Health Record (EHR).

PATIENTS AND METHODS

We included adult patients with weight-related Read codes recorded in the UK's Clinical Practice Research Datalink EHR between 2000 and 2017. Weight status codes were compared to weight values recorded on the same day and positive predictive values (PPVs) were calculated for commonly used codes. Weight change codes were validated using three methods: the percentage (%) difference in kilograms at the time of the code and 1) the previous weight measurement, 2) the weight predicted using linear regression, and 3) the historic mean weight. Weight change codes were validated if estimates were consistent across two out of three methods.

RESULTS

A total of 8,108,481 weight codes were recorded in 1,000,002 patients' EHR. Twice as many were recorded in females (n=5,208,593, 64%). The mean body mass index for "overweight" codes ranged from 31.9 kg/m2 to 46.9 kg/m2 and from 17.4 kg/m2 to 19.2 kg/m2 for "underweight" codes. PPVs for the most commonly used weight status codes ranged from 81.3% (80%-82.5%) to 99.3% (99.2%-99.4%). Across the estimation methods, and using only validated weight change codes, mean weight loss ranged from - 5.2% (SD 5.8%) to -7.9% (SD 7.3%) and mean weight gain from 4.2 % (SD 5.5%) to 7.9 % (SD 8.2%). The previous and predicted weight methods were most consistent.

CONCLUSION

We have developed an internationally applicable methodology to internally validate weight-related EHR coding by using available weight measurement data. We demonstrate the UK Read codes that can be confidently used to classify weight status and weight change in the absence of weight values. We provide the first evidence from primary care that a Read code for unexpected weight loss represents a mean loss of ≥ 5 % in a 6-month period, which was broadly consistent across age groups and gender.

Association of BMI with overall and cause-specific mortality: a population-based cohort study of 3·6 million adults in the UK

BACKGROUND

BMI is known to be strongly associated with all-cause mortality, but few studies have been large enough to reliably examine associations between BMI and a comprehensive range of cause-specific mortality outcomes.

METHODS

In this population-based cohort study, we used UK primary care data from the Clinical Practice Research Datalink (CPRD) linked to national mortality registration data and fitted adjusted Cox regression models to examine associations between BMI and all-cause mortality, and between BMI and a comprehensive range of cause-specific mortality outcomes (recorded by International Classification of Diseases, 10th revision [ICD-10] codes). We included all individuals with BMI data collected at age 16 years and older and with subsequent follow-up time available. Follow-up began at whichever was the latest of: start of CPRD research-standard follow up, the 5-year anniversary of the first BMI record, or on Jan 1, 1998 (start date for death registration data); follow-up ended at death or on March 8, 2016. Fully adjusted models were stratified by sex and adjusted for baseline age, smoking, alcohol use, diabetes, index of multiple deprivation, and calendar period. Models were fitted in both never-smokers only and the full study population. We also did an extensive range of sensitivity analyses. The expected age of death for men and women aged 40 years at baseline, by BMI category, was estimated from a Poisson model including BMI, age, and sex.

FINDINGS

3 632 674 people were included in the full study population; the following results are from the analysis of never-smokers, which comprised 1 969 648 people and 188 057 deaths. BMI had a J-shaped association with overall mortality; the estimated hazard ratio per 5 kg/m2 increase in BMI was 0·81 (95% CI 0·80-0·82) below 25 kg/m2 and 1·21 (1·20-1·22) above this point. BMI was associated with all cause of death categories except for transport-related accidents, but the shape of the association varied. Most causes, including cancer, cardiovascular diseases, and respiratory diseases, had a J-shaped association with BMI, with lowest risk occurring in the range 21-25 kg/m2. For mental and behavioural, neurological, and accidental (non-transport-related) causes, BMI was inversely associated with mortality up to 24-27 kg/m2, with little association at higher BMIs; for deaths from self-harm or interpersonal violence, an inverse linear association was observed. Associations between BMI and mortality were stronger at younger ages than at older ages, and the BMI associated with lowest mortality risk was higher in older individuals than in younger individuals. Compared with individuals of healthy weight (BMI 18·5-24·9 kg/m2), life expectancy from age 40 years was 4·2 years shorter in obese (BMI ≥30·0 kg/m2) men and 3·5 years shorter in obese women, and 4·3 years shorter in underweight (BMI <18·5 kg/m2) men and 4·5 years shorter in underweight women. When smokers were included in analyses, results for most causes of death were broadly similar, although marginally stronger associations were seen among people with lower BMI, suggesting slight residual confounding by smoking.

INTERPRETATION

BMI had J-shaped associations with overall mortality and most specific causes of death; for mental and behavioural, neurological, and external causes, lower BMI was associated with increased mortality risk.

FUNDING

Wellcome Trust.

Recording of weight in electronic health records: an observational study in general practice

Background

Routine weight recording in electronic health records (EHRs) could assist general practitioners (GPs) in the identification, prevention, and management of overweight patients. However, the extent to which weight management is embedded in general practice in the Netherlands has not been investigated. The purpose of this study was to evaluate the frequency of weight recording in general practice in the Netherlands for patients who self-reported as being overweight. The specific objectives of this study were to assess whether weight recording varied according to patient characteristics, and to determine the frequency of weight recording over time for patients with and without a chronic condition related to being overweight.

Methods

Baseline data from the Occupational and Environmental Health Cohort Study (2012) were combined with data from EHRs of general practices (2012–2015). Data concerned 3446 self-reported overweight patients who visited their GP in 2012, and 1516 patients who visited their GP every year between 2012 and 2015. Logistic multilevel regression analyses were performed to identify associations between patient characteristics and weight recording.

Results

In 2012, weight was recorded in the EHRs of a quarter of patients who self-reported as being overweight. Greater age, lower education level, higher self-reported body mass index, and the presence of diabetes mellitus, chronic obstructive pulmonary disease, and/or cardiovascular disorders were associated with higher rates of weight recording. The strongest association was found for diabetes mellitus (adjusted OR = 10.3; 95% CI [7.3, 14.5]). Between 2012 and 2015, 90% of patients with diabetes mellitus had at least one weight measurement recorded in their EHR. In the group of patients without a chronic condition related to being overweight, this percentage was 33%.

Conclusions

Weight was frequently recorded for overweight patients with a chronic condition, for whom regular weight measurement is recommended in clinical guidelines, and for which weight recording is a performance indicator as part of the payment system. For younger patients and those without a chronic condition related to being overweight, weight was less frequently recorded. For these patients, routine recording of weight in EHRs deserves more attention, with the aim to support early recognition and treatment of overweight.

Electronic supplementary material

The online version of this article (10.1186/s12875-018-0863-x) contains supplementary material, which is available to authorized users.

Machine learning models in electronic health records can outperform conventional survival models for predicting patient mortality in coronary artery disease

Prognostic modelling is important in clinical practice and epidemiology for patient management and research. Electronic health records (EHR) provide large quantities of data for such models, but conventional epidemiological approaches require significant researcher time to implement. Expert selection of variables, fine-tuning of variable transformations and interactions, and imputing missing values are time-consuming and could bias subsequent analysis, particularly given that missingness in EHR is both high, and may carry meaning. Using a cohort of 80,000 patients from the CALIBER programme, we compared traditional modelling and machine-learning approaches in EHR. First, we used Cox models and random survival forests with and without imputation on 27 expert-selected, preprocessed variables to predict all-cause mortality. We then used Cox models, random forests and elastic net regression on an extended dataset with 586 variables to build prognostic models and identify novel prognostic factors without prior expert input. We observed that data-driven models used on an extended dataset can outperform conventional models for prognosis, without data preprocessing or imputing missing values. An elastic net Cox regression based with 586 unimputed variables with continuous values discretised achieved a C-index of 0.801 (bootstrapped 95% CI 0.799 to 0.802), compared to 0.793 (0.791 to 0.794) for a traditional Cox model comprising 27 expert-selected variables with imputation for missing values. We also found that data-driven models allow identification of novel prognostic variables; that the absence of values for particular variables carries meaning, and can have significant implications for prognosis; and that variables often have a nonlinear association with mortality, which discretised Cox models and random forests can elucidate. This demonstrates that machine-learning approaches applied to raw EHR data can be used to build models for use in research and clinical practice, and identify novel predictive variables and their effects to inform future research.

Acute Pericarditis and Cancer Risk: A Matched Cohort Study Using Linked UK Primary and Secondary Care Data

Background We aimed to examine whether acute pericarditis is an indicator of undetected cancer and identify patient-level factors associated with high cancer risk among patients presenting with pericarditis. Methods and Results A population-based matched cohort study was conducted using primary care data from the UK Clinical Practice Research Datalink linked to Hospital Episode Statistics. Patients with acute pericarditis (n=6530) were matched to a comparison cohort (n=26 111) on age, sex, calendar time, and general practice. We estimated cumulative cancer incidences, and calculated hazard ratios using Cox regression. Effect modification by patients' characteristics and lifestyle factors was examined, and we fitted a parsimonious model to evaluate absolute excess risk of later cancer among pericarditis patients by key patient-level factors. We identified 728 and 1379 incidents of cancer among pericarditis patients and the comparison cohort (median follow-up, 2.8 and 3.5 years). Pericarditis was associated with an elevated subsequent risk of any cancer (hazard ratio=3.03; 95% confidence interval, 2.74-3.36). The association was particularly pronounced 0 to 3 months after pericarditis (hazard ratio=23.56; 95% confidence interval, 18.00-30.83), but a more-modest association remained thereafter (hazard ratio=1.95; 95% confidence interval, 1.48-2.57 after 3-12 months, and hazard ratio=1.40; 95% confidence interval, 1.21-1.62 after >12 month). Older individuals hospitalized with pericarditis and with combinations of obesity and smoking were at the highest excess risk of having a cancer diagnosis 3 to 12 months later, reaching 4.8%. Conclusions Occult cancers may be going undiagnosed during the acute episode of pericarditis. Patients presenting with pericarditis and combinations of older age, obesity, smoking, and a need for hospitalization might warrant targeted investigations for cancer.

Weight loss as a predictor of cancer in primary care: a systematic review and meta-analysis

BACKGROUND

Weight loss is a non-specific cancer symptom for which there are no clinical guidelines about investigation in primary care.

AIM

To summarise the available evidence on weight loss as a clinical feature of cancer in patients presenting to primary care.

DESIGN AND SETTING

A diagnostic test accuracy review and meta-analysis.

METHOD

Studies reporting 2 × 2 diagnostic accuracy data for weight loss (index test) in adults presenting to primary care and a subsequent diagnosis of cancer (reference standard) were included. QUADAS-2 was used to assess study quality. Sensitivity, specificity, positive likelihood ratios, and positive predictive values were calculated, and a bivariate meta-analysis performed.

RESULTS

A total of 25 studies were included, with 23 (92%) using primary care records. Of these, 20 (80%) defined weight loss as a physician's coding of the symptom; the remainder collected data directly. One defined unexplained weight loss using objective measurements. Positive associations between weight loss and cancer were found for 10 cancer sites: prostate, colorectal, lung, gastro-oesophageal, pancreatic, non-Hodgkin's lymphoma, ovarian, myeloma, renal tract, and biliary tree. Sensitivity ranged from 2% to 47%, and specificity from 92% to 99%, across cancer sites. The positive predictive value for cancer in male and female patients with weight loss for all age groups ≥60 years exceeded the 3% risk threshold that current UK guidance proposes for further investigation.

CONCLUSION

A primary care clinician's decision to code for weight loss is highly predictive of cancer. For such patients, urgent referral pathways are justified to investigate for cancer across multiple sites.

Generalizing Randomized Clinical Trial Results: Implementation and Challenges Related to Missing Data in the Target Population

Statins are indicated in patients with elevated levels of high-sensitivity C-reactive protein and normal low-density lipoprotein cholesterol based on results of the multicountry trial, Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) (2003-2008), but the benefit in real-world populations remains unknown. We sought to generalize JUPITER results to trial-eligible population using data from the UK Clinical Practice Research Datalink (CPRD), 2001-2014. We multiply imputed missing baseline characteristics for the CPRD population and selected the trial-eligible participants as the target population based on observed and imputed values. Trial participants were weighted to be representative of the CPRD population (n = 383,418) based on individual predicted probability of selection into the trial. Trial participants were also standardized to the CPRD population without missing values (n = 2,677). In JUPITER, rosuvastatin reduced cardiovascular risk with a 3-year risk difference of -2.0% (95% confidence interval (CI): -2.9, -1.1). The rosuvastatin effect was muted in the first 2 years but remained strong at 3 years after standardizing to the imputed CPRD population (3-year risk difference = -2.7%; 95% CI: -5.8, 0.4) and the CPRD population without missing data (3-year risk difference = -1.7%; 95% CI: -3.5, 0.1). The study serves as an illustration of possible approaches to understanding generalizability of trials using real-world databases given limitations due to missing data on inclusion/exclusion criteria.

Trimethoprim use for urinary tract infection and risk of adverse outcomes in older patients: cohort study

OBJECTIVE

To determine if trimethoprim use for urinary tract infection (UTI) is associated with an increased risk of acute kidney injury, hyperkalaemia, or sudden death in the general population.

DESIGN

Cohort study.

SETTING

UK electronic primary care records from practices contributing to the Clinical Practice Research Datalink linked to the Hospital Episode Statistics database.

PARTICIPANTS

Adults aged 65 and over with a prescription for trimethoprim, amoxicillin, cefalexin, ciprofloxacin, or nitrofurantoin prescribed up to three days after a primary care diagnosis of UTI between April 1997 and September 2015.

MAIN OUTCOME MEASURES

The outcomes were acute kidney injury, hyperkalaemia, and death within 14 days of a UTI treated with antibiotics.

RESULTS

Among a cohort of 1 191 905 patients aged 65 and over, 178 238 individuals were identified with at least one UTI treated with antibiotics, comprising a total of 422 514 episodes of UTIs treated with antibiotics. The odds of acute kidney injury in the 14 days following antibiotic initiation were higher following trimethoprim (adjusted odds ratio 1.72, 95% confidence interval 1.31 to 2.24) and ciprofloxacin (1.48, 1.03 to 2.13) compared with amoxicillin. The odds of hyperkalaemia in the 14 days following antibiotic initiation were only higher following trimethoprim (2.27, 1.49 to 3.45) compared with amoxicillin. However, the odds of death within the 14 days following antibiotic initiation were not higher with trimethoprim than with amoxicillin: in the whole population the adjusted odds ratio was 0.90 (95% confidence interval 0.76 to 1.07) while among users of renin-angiotensin system blockers the odds of death within 14 days of antibiotic initiation was 1.12 (0.80 to 1.57). The results suggest that, for 1000 UTIs treated with antibiotics among people 65 and over, treatment with trimethoprim instead of amoxicillin would result in one to two additional cases of hyperkalaemia and two admissions with acute kidney injury, regardless of renin-angiotensin system blockade. However, for people taking renin-angiotensin system blockers and spironolactone treatment with trimethoprim instead of amoxicillin there were 18 additional cases of hyperkalaemia and 11 admissions with acute kidney injury.

CONCLUSION

Trimethoprim is associated with a greater risk of acute kidney injury and hyperkalaemia compared with other antibiotics used to treat UTIs, but not a greater risk of death. The relative risk increase is similar across population groups, but the higher baseline risk among those taking renin-angiotensin system blockers and potassium-sparing diuretics translates into higher absolute risks of acute kidney injury and hyperkalaemia in these groups.

Acute kidney injury and infections in patients taking antihypertensive drugs: a self-controlled case series analysis

Background

The relative risk of acute kidney injury (AKI) following different infections, and whether angiotensin-converting enzyme inhibitors (ACEIs)/angiotensin receptor blockers (ARBs) modify the risk, is unclear. We aimed to determine the risks of hospital admission with AKI following infections (urinary tract infection [UTI], lower respiratory tract infection [LRTI], and gastroenteritis) among users of antihypertensive drugs.

Methods

We used UK electronic health records from practices contributing to the Clinical Practice Research Datalink linked to the Hospital Episode Statistics database. We identified adults initiating ACEIs/ARBs or alternative antihypertensive therapy (β-blockers, calcium channel blockers, or thiazide diuretics) between April 1997 and March 2014 with at least 1 year of primary care registration prior to first prescription, who had a hospital admission for AKI, and who had a primary care record for incident UTI, LRTI, or gastroenteritis. We used a self-controlled case series design to calculate age-adjusted incidence rate ratios (IRRs) for AKI during risk periods following acute infection relative to noninfected periods (baseline).

Results

We identified 10,219 eligible new users of ACEIs/ARBs or other antihypertensives with an AKI record. Among these, 2,012 had at least one record for a UTI during follow-up, 2,831 had a record for LRTI, and 651 had a record for gastroenteritis. AKI risk was higher following infection than in baseline noninfectious periods. The rate ratio was highest following gastroenteritis: for the period 1–7 days postinfection, the IRR for AKI following gastroenteritis was 43.4 (95% CI=34.0–55.5), compared with 6.0 following LRTI (95% CI=5.0–7.3), and 9.3 following UTI (95% CI=7.8–11.2). Increased risks were similar for different antihypertensives.

Conclusion

Acute infections are associated with substantially increased transient AKI risk among antihypertensive users, with the highest risk after gastroenteritis. The increase in relative risk is not greater among users of ACEIs/ARBs compared with users of other antihypertensives.

ADHD in the United Kingdom: Regional and Socioeconomic Variations in Incidence Rates Amongst Children and Adolescents (2004-2013)

OBJECTIVE

To describe the incidence and distribution of ADHD within the United Kingdom, and to examine whether there was any association between ADHD incidence and socioeconomic deprivation.

METHOD

The study used data from the Clinical Practice Research Datalink (CPRD). Patients diagnosed with ADHD before the age of 19 between January 1, 2004 and December 31, 2013 were stratified according to the region in which their general practice was based. Practice Index of Multiple Deprivation (IMD) score was used as a surrogate measure of patients' deprivation status.

RESULTS

ADHD incidence was relatively stable between 2004 and 2013, but peaked in the last 2 years studied. Statistically significant ( p ≤ .05) differences in incidence were observed between U.K. regions. In almost every year studied, incidence rates were highest among the most deprived patients and lowest among the least deprived patients.

CONCLUSION

In the United Kingdom, ADHD may be associated with socioeconomic deprivation.

Epidemiology of adult overweight recording and management by UK GPs: a systematic review

Background

Primary care guidelines for managing adult overweight/obesity recommend routine measurement of body mass index (BMI) and the offer of weight management interventions. Many studies state that this is rarely done, but the extent to which overweight/obesity is recognised, considered, and documented in routine care has not been determined.

Aim

To identify the epidemiology of adult overweight documentation and management by UK GPs.

Design and setting

A systematic review of studies since 2006 from eight electronic databases and grey literature.

Method

Included studies measured the proportion of adult patients with documented BMI or weight loss intervention offers in routine primary care in the UK. A narrative synthesis reports the prevalence and pattern of the outcomes.

Results

In total, 2845 articles were identified, and seven were included; four with UK-wide data and three with regional-level data. The proportion of patients with a documented BMI was 58–79% (28–37% within a year). For overweight/obese patients alone, 43–52% had a recent BMI record, and 15–42% had a documented intervention offer. BMI documentation was positively associated with older age, female sex, higher BMI, coexistent chronic disease, and higher deprivation.

Conclusion

BMI is under-recorded and weight loss interventions are under-referred for primary care adult patients in the UK despite the obesity register in the Quality and Outcomes Framework (QOF). The review identified likely underserved groups such as younger males and otherwise healthy overweight/obese individuals to whom attention should now be directed. The proposed amendment to the obesity register QOF could prompt improvements but has not been adopted for 2017.

Economic Evaluations Alongside Efficient Study Designs Using Large Observational Datasets: the PLEASANT Trial Case Study

BACKGROUND

Large observational datasets such as Clinical Practice Research Datalink (CPRD) provide opportunities to conduct clinical studies and economic evaluations with efficient designs.

OBJECTIVES

Our objectives were to report the economic evaluation methodology for a cluster randomised controlled trial (RCT) of a UK NHS-delivered public health intervention for children with asthma that was evaluated using CPRD and describe the impact of this methodology on results.

METHODS

CPRD identified eligible patients using predefined asthma diagnostic codes and captured 1-year pre- and post-intervention healthcare contacts (August 2012 to July 2014). Quality-adjusted life-years (QALYs) 4 months post-intervention were estimated by assigning utility values to exacerbation-related contacts; a systematic review identified these utility values because preference-based outcome measures were not collected. Bootstrapped costs were evaluated 12 months post-intervention, both with 1-year regression-based baseline adjustment (BA) and without BA (observed).

RESULTS

Of 12,179 patients recruited, 8190 (intervention 3641; control 4549) were evaluated in the primary analysis, which included patients who received the protocol-defined intervention and for whom CPRD data were available. The intervention's per-patient incremental QALY loss was 0.00017 (bias-corrected and accelerated 95% confidence intervals [BCa 95% CI] -0.00051 to 0.00018) and cost savings were £14.74 (observed; BCa 95% CI -75.86 to 45.19) or £36.07 (BA; BCa 95% CI -77.11 to 9.67), respectively. The probability of cost savings was much higher when accounting for BA versus observed costs due to baseline cost differences between trial arms (96.3 vs. 67.3%, respectively).

CONCLUSION

Economic evaluations using data from a large observational database without any primary data collection is feasible, informative and potentially efficient. Clinical Trials Registration Number: ISRCTN03000938.

Body Mass Index and Incident Type 1 and Type 2 Diabetes in Children and Young Adults: A Retrospective Cohort Study

Context:

Little is known about the association between obesity and temporal trends in the incidence of diabetes in children and young adults.

Objective:

We examined the recent incidence of types 1 and 2 diabetes in relation to a high body mass index (BMI) in UK children and young adults.

Design:

Cohort and nested case-control.

Setting:

A total of 375 general practices that contribute to the UK Clinical Practice Research Datalink (CPRD).

Participants:

A total of 369,362 participants aged 2 to 15 years at BMI measurement in CPRD from 1994 to 2013.

Intervention:

None.

Main outcome measures:

Incident type 1 diabetes (T1D) and type 2 diabetes (T2D) diagnoses up to age 25 years.

Results:

A total of 654 incident cases of T2D and 1318 T1D cases were found. The incidence of T2D per 100,000 persons annually increased from 6.4 in 1994 to 1998 to 33.2 in 2009 to 2013; and that for T1D increased from 38.2 to 52.1 per 100,000 persons during the same period. The incidence of T2D increased in both overweight (85th to 95th percentile for age- and sex-specific BMI; P = 0.01) and obese (≥95th percentile; P < 0.01) individuals from 1994 to 2013. Obese individuals, who constituted 47.1% of T2D cases, had a markedly greater risk of incident T2D [odds ratio, 3.75; 95% confidence interval (CI), 3.07 to 4.57], with an incidence rate ratio of 4.33 (95% CI, 3.68 to 5.08) compared with the normal BMI category. No positive linear association was found between obesity (greater BMI) and incident T1D cases.

Conclusions:

Increasing obesity has contributed to the increasing incidence of T2D but not T1D among UK children and young adults, with a fourfold greater risk of developing T2D in obese individuals.

Can machine-learning improve cardiovascular risk prediction using routine clinical data?

BACKGROUND

Current approaches to predict cardiovascular risk fail to identify many people who would benefit from preventive treatment, while others receive unnecessary intervention. Machine-learning offers opportunity to improve accuracy by exploiting complex interactions between risk factors. We assessed whether machine-learning can improve cardiovascular risk prediction.

METHODS

Prospective cohort study using routine clinical data of 378,256 patients from UK family practices, free from cardiovascular disease at outset. Four machine-learning algorithms (random forest, logistic regression, gradient boosting machines, neural networks) were compared to an established algorithm (American College of Cardiology guidelines) to predict first cardiovascular event over 10-years. Predictive accuracy was assessed by area under the 'receiver operating curve' (AUC); and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) to predict 7.5% cardiovascular risk (threshold for initiating statins).

FINDINGS

24,970 incident cardiovascular events (6.6%) occurred. Compared to the established risk prediction algorithm (AUC 0.728, 95% CI 0.723-0.735), machine-learning algorithms improved prediction: random forest +1.7% (AUC 0.745, 95% CI 0.739-0.750), logistic regression +3.2% (AUC 0.760, 95% CI 0.755-0.766), gradient boosting +3.3% (AUC 0.761, 95% CI 0.755-0.766), neural networks +3.6% (AUC 0.764, 95% CI 0.759-0.769). The highest achieving (neural networks) algorithm predicted 4,998/7,404 cases (sensitivity 67.5%, PPV 18.4%) and 53,458/75,585 non-cases (specificity 70.7%, NPV 95.7%), correctly predicting 355 (+7.6%) more patients who developed cardiovascular disease compared to the established algorithm.

CONCLUSIONS

Machine-learning significantly improves accuracy of cardiovascular risk prediction, increasing the number of patients identified who could benefit from preventive treatment, while avoiding unnecessary treatment of others.

Venous thromboembolism in adults screened for sickle cell trait: a population-based cohort study with nested case-control analysis

OBJECTIVE

To determine whether sickle cell carriers ('sickle cell trait') have an increased risk of venous thromboembolism (VTE).

DESIGN

Cohort study with nested case-control analysis.

SETTING

General population with data from 609 UK general practices in the Clinical Practice Research Datalink (CPRD).

PARTICIPANTS

All individuals registered with a CPRD general practice between 1998 and 2013, with a medical record of screening for sickle cell between 18 and 75 years of age.

MAIN OUTCOMES MEASURES

Incidence of VTE per 10 000 person-years (PY) among sickle cell carriers and non-carriers; and adjusted OR for VTE among sickle cell carriers compared with non-carriers.

RESULTS

We included 30 424 individuals screened for sickle cell, with a follow-up time of 179 503 PY, identifying 55 VTEs in 6758 sickle cell carriers and 125 VTEs in 23 666 non-carriers. VTE incidence among sickle cell carriers (14.9/10 000 PY; 95% CI 11.4 to 19.4) was significantly higher than non-carriers (8.8/10 000 PY; 95% CI 7.4 to 10.4). Restricting analysis to confirmed non-carriers was non-significant, but performed on a small sample. In the case-control analysis (180 cases matched to 1775 controls by age and gender), sickle cell carriers remained at increased risk of VTE after adjusting for body mass index, pregnancy, smoking status and ethnicity (OR 1.78, 95% CI 1.18 to 2.69, p=0.006), with the greatest risk for pulmonary embolism (PE) (OR 2.27, 95% CI 1.17 to 4.39, p=0.011).

CONCLUSIONS

Although absolute numbers are small, in a general population screened for sickle cell, carriers have a higher incidence and risk of VTE, particularly PE, than non-carriers. Clinicians should be aware of this elevated risk in the clinical care of sickle cell carriers, or when discussing carrier screening, and explicitly attend to modifiable risk factors for VTE in these individuals. More complete primary care coding of carrier status could improve analysis.

Serum creatinine elevation after renin-angiotensin system blockade and long term cardiorenal risks: cohort study

Objective To examine long term cardiorenal outcomes associated with increased concentrations of creatinine after the start of angiotensin converting enzyme inhibitor/angiotensin receptor blocker treatment.Design Population based cohort study using electronic health records from the Clinical Practice Research Datalink and Hospital Episode Statistics.Setting UK primary care, 1997-2014.Participants Patients starting treatment with angiotensin converting enzyme inhibitors or angiotensin receptor blockers (n=122 363).Main outcome measures Poisson regression was used to compare rates of end stage renal disease, myocardial infarction, heart failure, and death among patients with creatinine increases of 30% or more after starting treatment against those without such increases, and for each 10% increase in creatinine. Analyses were adjusted for age, sex, calendar period, socioeconomic status, lifestyle factors, chronic kidney disease, diabetes, cardiovascular comorbidities, and use of other antihypertensive drugs and non-steroidal anti-inflammatory drugs.Results Among the 2078 (1.7%) patients with creatinine increases of 30% or more, a higher proportion were female, were elderly, had cardiorenal comorbidity, and used non-steroidal anti-inflammatory drugs, loop diuretics, or potassium sparing diuretics. Creatinine increases of 30% or more were associated with an increased adjusted incidence rate ratio for all outcomes, compared with increases of less than 30%: 3.43 (95% confidence interval 2.40 to 4.91) for end stage renal disease, 1.46 (1.16 to 1.84) for myocardial infarction, 1.37 (1.14 to 1.65) for heart failure, and 1.84 (1.65 to 2.05) for death. The detailed categorisation of increases in creatinine concentrations (<10%, 10-19%, 20-29%, 30-39%, and ≥40%) showed a graduated relation for all outcomes (all P values for trends <0.001). Notably, creatinine increases of less than 30% were also associated with increased incidence rate ratios for all outcomes, including death (1.15 (1.09 to 1.22) for increases of 10-19% and 1.35 (1.23 to 1.49) for increases of 20-29%, using <10% as reference). Results were consistent across calendar periods, across subgroups of patients, and among continuing users.Conclusions Increases in creatinine after the start of angiotensin converting enzyme inhibitor/angiotensin receptor blocker treatment were associated with adverse cardiorenal outcomes in a graduated relation, even below the guideline recommended threshold of a 30% increase for stopping treatment.

Body-mass index and risk of advanced chronic kidney disease: Prospective analyses from a primary care cohort of 1.4 million adults in England

BACKGROUND

It is uncertain whether being overweight, but not obese, is associated with advanced chronic kidney disease (CKD) and how the size and shape of associations between body-mass index (BMI) and advanced CKD differs among different types of people.

METHODS

We used Clinical Practice Research Datalink records (2000-2014) with linkage to English secondary care and mortality data to identify a prospective cohort with at least one BMI measure. Cox models adjusted for age, sex, smoking and social deprivation and subgroup analyses by diabetes, hypertension and prior cardiovascular disease assessed relationships between BMI and CKD stages 4-5 and end-stage renal disease (ESRD).

FINDINGS

1,405,016 adults aged 20-79 with mean BMI 27.4kg/m2 (SD 5.6) were followed for 7.5 years. Compared to a BMI of 20 to <25kg/m2, higher BMI was associated with a progressively increased risk of CKD stages 4-5 (hazard ratio 1.34, 95% CI 1.30-1.38 for BMI 25 to <30kg/m2; 1.94, 1.87-2.01 for BMI 30 to <35kg/m2; and 3.10, 2.95-3.25 for BMI ≥35kg/m2). The association between BMI and ESRD was shallower and reversed at low BMI. Current smoking, prior diabetes, hypertension or cardiovascular disease all increased risk of CKD, but the relative strength and shape of BMI-CKD associations, which were generally log-linear above a BMI of 25kg/m2, were similar among those with and without these risk factors. There was direct evidence that being overweight was associated with increased risk of CKD stages 4-5 in these subgroups. Assuming causality, since 2000 an estimated 39% (36-42%) of advanced CKD in women and 26% (22-30%) in men aged 40-79 resulted from being overweight or obese.

CONCLUSIONS

This study provides direct evidence that being overweight increases risk of advanced CKD, that being obese substantially increases such risk, and that this remains true for those with and without diabetes, hypertension or cardiovascular disease. Strategies to reduce weight among those who are overweight, as well as those who are obese may reduce CKD risk, with each unit reduction in BMI yielding similar relative reductions in risk.

Do pre-diagnosis primary care consultation patterns explain deprivation-specific differences in net survival among women with breast cancer? An examination of individually-linked data from the UK West Midlands cancer registry, national screening programme and Clinical Practice Research Datalink

BACKGROUND

In England and Wales breast cancer survival is higher among more affluent women. Our aim was to investigate the potential of pre-diagnostic factors for explaining deprivation-related differences in survival.

METHODS

Individually-linked data from women aged 50-70 in the West Midlands region of England, diagnosed with breast cancer 1989-2006 and continuously eligible for screening, was retrieved from the cancer registry, screening service and Clinical Practice Research Datalink. Follow-up was to the end of July 2012. Deprivation was measured at small area level, based on the quintiles of the income domain of the English indices of deprivation. Consultation rates per woman per week, time from last breast-related GP consultation to diagnosis, and from diagnosis to first surgery were calculated. We estimated net survival using the non-parametric Pohar-Perme estimator.

RESULTS

The rate of primary care consultations was similar during the 18 months prior to diagnosis in each deprivation group for breast and non-breast symptoms. Survival was lower for more deprived women from 4 years after diagnosis. Lower net survival was associated with more advanced extent of disease and being non-screen-detected. There was a persistent trend of lower net survival for more deprived women, irrespective of the woman's obesity, alcohol, smoking or comorbidity status. There was no significant variation in time from last breast symptom to diagnosis by deprivation. However, women in more deprived categories experienced significantly longer periods between cancer diagnosis and first surgery (mean = 21.5 vs. 28.4 days, p = 0.03). Those whose surgery occurred more than 12 weeks following their cancer diagnosis had substantially lower net survival.

CONCLUSIONS

Our data suggest that although more deprived women with breast cancer display lifestyle factors associated with poorer outcomes, their consultation frequency, comorbidities and the breast cancer symptoms they present with are similar. We found weak evidence of extended times to surgical treatment among most deprived women who were not screen-detected but who presented with symptoms in primary care, which suggests that treatment delay may play a role. Further investigation of interrelationships between these variables within a larger dataset is warranted.

Adherence to guidelines for creatinine and potassium monitoring and discontinuation following renin-angiotensin system blockade: a UK general practice-based cohort study

OBJECTIVES

To examine adherence to serum creatinine and potassium monitoring and discontinuation guidelines following initiation of treatment with ACE inhibitors (ACEI) or angiotensin receptor blockers (ARBs); and whether high-risk patients are monitored.

DESIGN

A general practice-based cohort study using electronic health records from the UK Clinical Practice Research Datalink and Hospital Episode Statistics.

SETTING

UK primary care, 2004-2014.

SUBJECTS

223 814 new ACEI/ARB users.

MAIN OUTCOME MEASURES

Proportion of patients with renal function monitoring before and after ACEI/ARB initiation; creatinine increase ≥30% or potassium levels >6 mmol/L at first follow-up monitoring; and treatment discontinuation after such changes. Using logistic regression models, we also examined patient characteristics associated with these biochemical changes, and with follow-up monitoring within the guideline recommendation of 2 weeks after treatment initiation.

RESULTS

10% of patients had neither baseline nor follow-up monitoring of creatinine within 12 months before and 2 months after initiation of an ACEI/ARB, 28% had monitoring only at baseline, 15% only at follow-up, and 47% both at baseline and follow-up. The median period between the most recent baseline monitoring and drug initiation was 40 days (IQR 12-125 days). 34% of patients had baseline creatinine monitoring within 1 month before initiating therapy, but <10% also had the guideline-recommended follow-up test recorded within 2 weeks. Among patients experiencing a creatinine increase ≥30% (n=567, 1.2%) or potassium level >6 mmol/L (n=191, 0.4%), 80% continued treatment. Although patients with prior myocardial infarction, hypertension or baseline potassium >5 mmol/L were at high risk of ≥30% increase in creatinine after ACEI/ARB initiation, there was no evidence that they were more frequently monitored.

CONCLUSIONS

Only one-tenth of patients initiating ACEI/ARB therapy receive the guideline-recommended creatinine monitoring. Moreover, the vast majority of the patients fulfilling postinitiation discontinuation criteria for creatinine and potassium increases continue on treatment.

Prescription of renin-angiotensin system blockers and risk of acute kidney injury: a population-based cohort study

OBJECTIVE

To investigate whether there is an association between use of ACE inhibitors (ACEI) and angiotensin receptor blockers (ARB) and risk of acute kidney injury (AKI).

STUDY DESIGN

We conducted a new-user cohort study of the rate of AKI among users of common antihypertensives.

SETTING

UK primary care practices contributing to the Clinical Practice Research Datalink (CPRD) eligible for linkage to hospital records data from the Hospital Episode Statistics (HES) database between April 1997 and March 2014.

PARTICIPANTS

New users of antihypertensives: ACEI/ARB, β-blockers, calcium channel blockers and thiazide diuretics.

OUTCOMES

The outcome was first episode of AKI. We estimated incidence rate ratio (RR) for AKI during time exposed to ACEI/ARB compared to time unexposed, adjusting for age, sex, comorbidities, use of other antihypertensive drugs and calendar period using Poisson regression. Covariates were time updated.

RESULTS

Among 570 445 participants, 303 761 were prescribed ACEI/ARB with a mean follow-up of 4.1 years. The adjusted RR of AKI during time exposed to ACEI/ARB compared to time unexposed was 1.12 (95% CI 1.07 to 1.17). This relative risk varied depending on absolute risk of AKI, with lower or no increased relative risk from the drugs among those at greatest absolute risk. For example, among people with stage 4 chronic kidney disease (who had 6.69 (95% CI 5.57 to 8.03) times higher rate of AKI compared to those without chronic kidney disease), the adjusted RR of AKI during time exposed to ACEI/ARB compared to time unexposed was 0.66 (95% CI 0.44 to 0.97) in contrast to 1.17 (95% CI 1.09 to 1.25) among people without chronic kidney disease.

CONCLUSIONS

Treatment with ACEI/ARB is associated with only a small increase in AKI risk while individual patient characteristics are much more strongly associated with the rate of AKI. The degree of increased risk varies between patient groups.