Data Resource Profile: Clinical Practice Research Datalink (CPRD)

Metformin and risk of adverse pregnancy outcomes among pregnant women with gestational diabetes in the United Kingdom: A population-based cohort study

AIMS

Metformin is increasingly used off-label as the treatment of gestational diabetes (GDM). Our objective was to determine if metformin versus insulin initiation is associated with the adverse pregnancy outcomes.

MATERIALS AND METHODS

We conducted a retrospective cohort study using data from the Clinical Practice Research Datalink, its pregnancy register, and Hospital Episode Statistics from 1998 to 2018. We included pregnancies of women who initiated metformin or insulin between 20 weeks gestation and pregnancy end. The primary outcome was a composite outcome of large for gestational age (LGA) and macrosomia. The secondary outcomes included small for gestational age (SGA), preterm birth, caesarean delivery, and hypertensive disorders during pregnancy (HDP). Inverse probability weighted-Cox proportional hazards models were to estimate adjusted hazard ratios (HRs) and 95% confidence intervals (CI), comparing those who initiated metformin versus insulin at cohort entry, accounting for baseline covariates.

RESULTS

Our cohort included pregnancies of 1297 women initiating metformin and of 895 women initiating insulin. Compared to insulin initiation, metformin initiation was associated with a decreased risk of LGA or macrosomia (HR 0.64, 95% CI 0.49, 0.78), Caesarean delivery (HR 0.83, 95% CI 0.69, 0.98), and preterm birth (HR 0.83, 95% CI 0.58, 1.08). The HRs for HDP and SGA were 0.92 (95% CI 0.57, 1.27) and 1.33 (95% CI 0.67, 2.00), respectively.

CONCLUSIONS

Our study suggests that, compared to initiating insulin, initiating metformin is associated with decreased risks of adverse pregnancy outcomes among women with GDM. These findings provide important real-world evidence regarding the use of metformin for GDM.

A UK population-based case-control study of blood tests before cancer diagnosis in patients with non-specific abdominal symptoms

BACKGROUND

Abnormal results in commonly used primary care blood tests could be early markers of cancer in patients presenting with non-specific abdominal symptoms.

METHODS

Using linked data from the UK Clinical Practice Research Datalink (CPRD) and national cancer registry we compared blood test use and abnormal results from the 24-months pre-diagnosis in 10,575 cancer patients (any site), and 52,875 matched-controls aged ≥30 presenting, with abdominal pain or bloating to primary care.

RESULTS

Cancer patients had two-fold increased odds of having a blood test (odds ratio(OR):1.51-2.29) and 2-3-fold increased odds of having an abnormal blood test result (OR:2.42-3.30) in the year pre-diagnosis compared to controls. Raised inflammatory markers were the most common abnormality (74-79% of tested cases). Rates of blood test use and abnormal results progressively increased from 7 months pre-diagnosis in cancer patients, with relatively small corresponding increases in symptomatic controls. In cancer patients, the largest increases from baseline were raised platelets in males with abdominal pain (increased 33-fold), raised white blood cell count in males with abdominal bloating (increased 37-fold) and low albumin in females with either symptom (increased 22-41 fold).

CONCLUSIONS

Common blood test abnormalities are early signals of cancer in some individuals with non-specific abdominal symptoms and could support expedited cancer diagnosis.

Burden of vascular risk factors by age, sex, ethnicity and deprivation in young adults with and without newly diagnosed type 2 diabetes

AIMS

Do associations between age at diagnosis of type 2 diabetes and vascular risk factors vary by ethnicity and deprivation?

METHODS

Utilising the Clinical Practice Research Datalink, we matched 16-50-year-old individuals with newly diagnosed type 2 diabetes to ∼10 individuals without using sex, age and primary care practice. Differences in BMI, obesity, LDL-cholesterol, HbA1c, and hypertension between individuals with vs without type 2 diabetes across sex, age, ethnicity and deprivation quintiles were explored using generalised linear models.

RESULTS

We included 108,061 individuals (45.6% women) with newly diagnosed type 2 diabetes and 829,946 controls. BMI, obesity, LDL-cholesterol, and hypertension were higher in individuals with vs without type 2 diabetes. Across both sexes, all ethnic groups and deprivation quintiles, these differences were larger with an earlier age, particularly for BMI and obesity. Association between age and HbA1c were variable across subgroups. Differences in BMI, obesity, and hypertension (individuals with vs without diabetes) were largest in White individuals and with less deprivation.

CONCLUSIONS

The increased vascular risk phenotype associated with an earlier age of diagnosis of type 2 diabetes was consistent across ethnic and deprivation groups. Population-based strategies are needed to address the risk associated with early-onset type 2 diabetes, especially weight-management-based strategies.

Prevalence and outcomes of patients taking oral corticosteroids for over 1 month undergoing major surgery in England 2010-2020

INTRODUCTION

Approximately 1% of the UK population is prescribed oral corticosteroids at any one time. It is not known how many of these patients present for major surgery. We aimed to establish the prevalence, characteristics and outcomes of patients taking oral corticosteroids.

METHODS

We identified patients aged > 18 y undergoing major surgery between 1 April 2010 and 31 March 2020 from Hospital Episode Statistics with linked Clinical Practice Research Datalink data and the Office for National Statistics Mortality register in England. Prescribing data were used to define three sets of patients: 'low-dose' - taking ≤ 7.5 mg oral prednisolone equivalents per day for at least 28/91 days before surgery; 'high-dose' - taking > 7.5 mg oral prednisolone equivalents per day for at least 28/ 91 days before surgery; and a 'no-steroids' group. We used ≤ 7.5 mg of prednisolone equivalents per day as our threshold, as this would likely exclude almost all patients who were taking corticosteroids as replacement for absolute adrenal/pituitary deficiency.

RESULTS

We identified 1,999,326 adult patients for inclusion in the dataset: 1,929,291 (96.5%) in the no-steroids; 63,353 (3.2%) in the low-dose group; and 6682 (0.3%) in the high-dose group. Median (IQR [range]) duration of hospital stay increased with increasing dose of corticosteroid (no-steroid 3 (0-14 [0-14,739]); low-dose 5 (1-26 [1-8079]); and high-dose 7 (2-28 [0-6956]) days). Mortality after the index surgery was 1.5%, 3.8% and 8.9% at 30 days and 5.5%, 11.6% and 39.9% at 1 year for no-steroids, low-dose and high-dose groups, respectively.

CONCLUSION

Around 1 in 29 patients undergoing major surgery are taking oral corticosteroids for > 28 days in the 3 months before major surgery. Their outcomes are poor and warrant highlighting within care pathways to aid risk prediction and mitigation.

The Stockholm early detection of cancer study (STEADY-CAN): rationale, design, data collection, and baseline characteristics for 2.7 million participants

The Stockholm Early Detection of Cancer Study (STEADY-CAN) cohort was established to investigate strategies for early cancer detection in a population-based context within Stockholm County, the capital region of Sweden. Utilising real-world data to explore cancer-related healthcare patterns and outcomes, the cohort links extensive clinical and laboratory data from both inpatient and outpatient care in the region. The dataset includes demographic information, detailed diagnostic codes, laboratory results, prescribed medications, and healthcare utilisation data. Since its inception, STEADY-CAN has collected longitudinal data on 2,732,005 individuals aged ≥ 18 years old living in or having access to health care in Stockholm County during the years 2011-2021. Focusing on cancer, the cohort includes 140,042 (5.1%) individuals with incident cancer and a control group of 2,591,963 (94.9%) cancer-free individuals. The cohort's diverse adult population enables robust analyses of early symptom detection, incidental findings, and the impact of comorbidities on cancer diagnoses. Utilizing the wide range of available laboratory data and clinical variables allow for advanced statistical analyses and adjustments for important confounding factors. The cohort's primary focus is to improve understanding of the early diagnostic phase of cancer, offering a crucial resource for studying cancer detection in clinical practice. Its comprehensive data collection provides unique opportunities for research into comorbidities and cancer outcomes, making the cohort a useful resource for ongoing cancer surveillance and public health strategies. The present study gives a detailed description of the rationale for creating the STEADY-CAN cohort, its design, the data collection procedure, and baseline characteristics of collected data.

Open challenges and opportunities in federated foundation models towards biomedical healthcare

This survey explores the transformative impact of foundation models (FMs) in artificial intelligence, focusing on their integration with federated learning (FL) in biomedical research. Foundation models such as ChatGPT, LLaMa, and CLIP, which are trained on vast datasets through methods including unsupervised pretraining, self-supervised learning, instructed fine-tuning, and reinforcement learning from human feedback, represent significant advancements in machine learning. These models, with their ability to generate coherent text and realistic images, are crucial for biomedical applications that require processing diverse data forms such as clinical reports, diagnostic images, and multimodal patient interactions. The incorporation of FL with these sophisticated models presents a promising strategy to harness their analytical power while safeguarding the privacy of sensitive medical data. This approach not only enhances the capabilities of FMs in medical diagnostics and personalized treatment but also addresses critical concerns about data privacy and security in healthcare. This survey reviews the current applications of FMs in federated settings, underscores the challenges, and identifies future research directions including scaling FMs, managing data diversity, and enhancing communication efficiency within FL frameworks. The objective is to encourage further research into the combined potential of FMs and FL, laying the groundwork for healthcare innovations.

Going beyond randomised controlled trials to assess treatment effect heterogeneity across target populations

Methods have been developed for transporting evidence from randomised controlled trials (RCTs) to target populations. However, these approaches allow only for differences in characteristics observed in the RCT and real-world data (overt heterogeneity). These approaches do not recognise heterogeneity of treatment effects (HTE) according to unmeasured characteristics (essential heterogeneity). We use a target trial design and apply a local instrumental variable (LIV) approach to electronic health records from the Clinical Practice Research Datalink, and examine both forms of heterogeneity in assessing the comparative effectiveness of two second-line treatments for type 2 diabetes mellitus. We first estimate individualised estimates of HTE across the entire target population defined by applying eligibility criteria from national guidelines (n = 13,240) within an overall target trial framework. We define a subpopulation who meet a published RCT's eligibility criteria ('RCT-eligible', n = 6497), and a subpopulation who do not ('RCT-ineligible', n = 6743). We compare average treatment effects for pre-specified subgroups within the RCT-eligible subpopulation, the RCT-ineligible subpopulation, and within the overall target population. We find differences across these subpopulations in the magnitude of subgroup-level treatment effects, but that the direction of estimated effects is stable. Our results highlight that LIV methods can provide useful evidence about treatment effect heterogeneity including for those subpopulations excluded from RCTs.

Cancer incidence and mortality among patients with new-onset atrial fibrillation: A population-based matched cohort study

BACKGROUND

Understanding the risk of cancer after the diagnosis of another condition can present opportunities for earlier diagnosis. We examined the risk of cancer diagnosis conditional on prior diagnosis of atrial fibrillation (AF).

METHODS

Linked electronic health records were used to identify patients aged ≥18 with new-onset AF and age-sex-matched controls. Cumulative incidence of and mortality from cancer (overall and cancer-site specific) within three months, three months to five years and beyond five years from diagnosis of AF were examined. Findings were further validated using Mendelian randomisation (MR).

RESULTS

The cohort included 117,173 patients with new-onset AF and 117,173 matched controls (median age 78). In the first three months, 2.2% of AF patients were diagnosed with cancer vs. 0.47% in controls (relative risk: 4.7 [95%CI 4.2-5.4] in men and 4.4 [95%CI 3.8-5.0] in women). Nearly 80% of cancers related to thoracic or abdominal organs. Differences in cumulative incidence were only evident in women between three months and five years (subdistribution hazard ratio=1.1 [95%CI 1.01-1.12]) and absent in all patients beyond five years. MR analysis did not support the presence of a causal association between AF and major cancer subtypes.

CONCLUSION

There is a large short-term increase in cancer incidence and mortality following new-onset AF. The findings may reflect incidental identification of AF or paraneoplastic manifestation. New-onset AF confers high short-term risk of cancer diagnosis, at levels comparable with symptomatic risk threshold mandating urgent assessment for suspected cancer.

Treatment for osteoporosis and risk of osteonecrosis of the jaw among female patients in the United Kingdom Clinical Practice Research Datalink

Osteonecrosis of the jaw (ONJ) is an adverse effect of antiresorptives. Among female patients treated for osteoporosis, ONJ risk was threefold higher after 2-3 years of treatment and eightfold after 10 years compared with past use. Absolute risks remained low (~ 0.05% after 5 years) and diminished after discontinuation.

PURPOSE

Osteonecrosis of the jaw (ONJ) is a rare adverse effect of antiresorptive drug use; however, the magnitude of risk in osteoporosis patients has not been clearly described.

METHODS

We conducted a cohort study among cancer-free female patients aged 40-89 with, or at risk for, osteoporosis in United Kingdom Clinical Practice Research Datalink (CPRD) Aurum. We followed patients from first osteoporosis treatment until first of osteonecrosis diagnosis, age 90, record end, or other prespecified censoring event, and accumulated person-time by osteoporosis treatment. ONJ cases were selected from CPRD Aurum and linked Hospital Episode Statistics data using an algorithm and manual review. We estimated incidence rates (IR) of ONJ by current treatment type and post discontinuation. We conducted a nested case-control analysis to further describe risk by cumulative dose and duration of antiresorptive therapies.

RESULTS

Among 467,654 eligible patients, there were 208 ONJ cases. IR among patients currently treated with antiresorptives (primarily alendronate) was 1.2 (95% confidence interval [CI] 1.0-1.4) per 10,000 person-years. Compared with past use of antiresorptives, odds ratios of ONJ were 3.0 (95% CI 1.5-5.7) after 2-3 years of treatment and 8.1 (95% CI 4.4-15) after 10 years. However, absolute risks remained low (~ 0.05% after 5 years and ~ 0.18% after 10 years) and elevated risks diminished to near zero within 6 to 9 months of discontinuation.

CONCLUSION

Risk of ONJ increased after 2-3 years of treatment with antiresorptives; however, the absolute risk was low and returned to baseline shortly after treatment discontinuation.

Identifying and Predicting Risk for Hospital Admission among Patients with Parkinsonism

BACKGROUND

Patients with parkinsonism are more likely than age-matched controls to be admitted to hospital. It may be possible to reduce the cost and negative impact by identifying patients at highest risk and intervening to reduce hospital-related costs. Predictive models have been developed in nonparkinsonism populations.

OBJECTIVES

The aims were to (1) describe the reasons for admission, (2) describe the rates of hospital admission/emergency department attendance over time, and (3) use routine data to risk stratify unplanned hospital attendance in people with parkinsonism.

METHODS

This retrospective cohort study used Clinical Practice Research Datalink GOLD, a large UK primary care database, linked to hospital admission and emergency department attendance data. The primary diagnoses for nonelective admissions were categorized, and the frequencies were compared between parkinsonism cases and matched controls. Multilevel logistic and negative binomial regression models were used to estimate the risk of any and multiple admissions, respectively, for patients with parkinsonism.

RESULTS

There were 9189 patients with parkinsonism and 45,390 controls. The odds of emergency admission more than doubled from 2010 to 2019 (odds ratio [OR] 2.33; 95% confidence interval [CI] 1.96, 2.76; P-value for trend <0.001). Pneumonia was the most common reason for admission among cases, followed by urinary tract infection. Increasing age, multimorbidity, parkinsonism duration, deprivation, and care home residence increased the odds of admission. Rural location was associated with reduced OR for admission (OR 0.79; 95% CI 0.70, 0.89).

CONCLUSIONS

Our risk stratification tool may enable empirical targeting of interventions to reduce admission risk for parkinsonism patients.

Incidence trends in ischaemic and non-ischaemic heart failure in people with and without type 2 diabetes, 2000-2019: An observational study in England

AIM

To investigate trends in ischaemic and non-ischaemic heart failure (HF) in adults with type 2 diabetes and without diabetes between 1st January 2000 and 31st December 2019 in England.

METHODS

We used the Clinical Practice Research Datalink datasets, linked to the Hospital Episode Statistics and Office for National Statistics to estimate sex-specific crude and age-standardised rates of incident ischaemic and non-ischaemic HF up to 10 years per calendar year of diabetes diagnosis and diabetes status.

RESULTS

In a cohort of 735,810 individuals, 5,073 ischaemic (2,038 in people with type 2 diabetes and 3,035 in those without) and 16,501 non-ischaemic (6,358 and 10,143, respectively) HF events were recorded during a median follow-up of 10 years. From 2000 to 2004 to 2005-2009, the age-standardised rates of ischaemic HF marginally declined, while rates remained stable for non-ischaemic HF and were consistently higher for non-ischaemic than ischaemic HF, regardless of diabetes status or sex. Adjusted incidence rate ratios demonstrated negligible impact on trends after accounting for differences in demographics, comorbidities and medications.

CONCLUSIONS

Improving HF prevention and management strategies remains crucial to decrease the risk of HF in the general population and reduce the persistent risk-gap associated with type 2 diabetes in England.

Multimorbidity in elderly patients with or without T2DM: A real-world cross-sectional analysis based on primary care and hospitalisation data

Background

Shanghai's high level of ageing has given rise to a considerable number of elderly patients with type 2 diabetes mellitus (T2DM) who are confronted with the challenge of managing multimorbidity. We aimed to determine the prevalence of multimorbidity in elderly T2DM patients in a representative Pudong New Area community and critically evaluate current guidelines' inclusiveness in addressing major comorbidities.

Methods

Through the Shanghai Health Cloud platform, we extracted medical records of residents in the Huamu community (Pudong New Area, Shanghai) to screen elderly patients with at least three outpatient visits or one hospitalisation per year between 2019 and 2022. According to International Classification of Disease, 10th edition codes and personal identification number, we identified the status of T2DM and 12 other common chronic diseases, matched T2DM patients and non-T2DM patients 1:1 by age and gender, and then calculated the prevalence of multimorbidity status and annual prevalence of each comorbidity. We analysed associations between T2DM and specific chronic diseases using logistic regression models.

Results

More than 90% of elderly T2DM patients had at least one additional chronic disease. Multimorbidity was more frequent in women and older patients. Hyperlipidemia, hypertension, and ischaemic heart disease were the most prevalent comorbidities. The diagnosis of T2DM was significantly associated with both cardiovascular-kidney-metabolic and neuropsychiatric diseases. In addition, a higher prevalence and risk of chronic obstructive pulmonary disease (COPD) were consistently detected in elderly patients with T2DM, regardless of age and gender.

Conclusions

Multimorbidity in elderly patients with T2DM needs broader acknowledgement. Current guidelines focus more on cardiovascular-kidney-metabolic and neuropsychiatric diseases with inadequate guidance on COPD management. Hence, the pleiotropic effects of glucose-lowering drugs on COPD should be further investigated to optimise the comprehensive management strategy for this population.

Impact of the COVID-19 pandemic on primary care for hypertension in the UK: a population-based cohort study

OBJECTIVES

To describe the impact of the COVID-19 pandemic on hypertension diagnosis and management in UK primary care.

DESIGN

Population-based cohort study.

SETTING

Over 2000 general practices across the UK contributing to the Clinical Practice Research Datalink.

PARTICIPANTS

A cohort of 23 076 390 patients over 18 years of age and registered with their general practice for at least 1 year between 2011 and 2022, who did not have a previous diagnosis of hypertension. From these patients, a subcohort of 712 461 patients diagnosed with hypertension between 2011 and 2022 was selected.

PRIMARY AND SECONDARY OUTCOME MEASURES

Coprimary outcomes included rates of hypertension diagnosis and rates of antihypertensive treatment initiation, treatment change and blood pressure measurement in patients newly diagnosed with hypertension.

RESULTS

In April 2020, the first month of lockdown, incident hypertension diagnosis rates fell by 65% (95% CI 64% to 67%) compared with historical trends and remained depressed until November 2021, leading to 51 000 fewer diagnoses than expected by March 2022. However, by March 2022, there were 2.6% fewer diagnoses than expected in Scotland, compared with 20%-30% fewer in other UK Nations. Rates of treatment initiation and change fell by 47% (95% CI 43% to 51%) and 36% (95% CI 33% to 38%), respectively, in April 2020. However, initiation rates rebounded above expectations and remained elevated until March 2022. Blood pressure measurements fell by 69% (95% CI 65% to 72%) in April 2020, recovering in February 2021.

CONCLUSIONS

Hypertension diagnosis and management in UK primary care were significantly disrupted during the COVID-19 pandemic. Future studies should investigate the potential clinical implications for the cardiovascular health of the UK population.

Prescribing of antipsychotics for people diagnosed with severe mental illness in UK primary care 2000-2019: 20-year investigation of who receives treatment, with which agents and at what doses

BACKGROUND

Contemporary data relating to antipsychotic prescribing in UK primary care for patients diagnosed with severe mental illness (SMI) are lacking.

AIMS

To describe contemporary patterns of antipsychotic prescribing in UK primary care for patients diagnosed with SMI.

METHOD

Cohort study of patients with an SMI diagnosis (i.e. schizophrenia, bipolar disorder, other non-organic psychoses) first recorded in primary care between 2000 and 2017 derived from Clinical Practice Research Datalink. Patients were considered exposed to antipsychotics if prescribed at least one antipsychotic in primary care between 2000 and 2019. We compared characteristics of patients prescribed and not prescribed antipsychotics; calculated annual prevalence rates for antipsychotic prescribing; and computed average daily antipsychotic doses stratified by patient characteristics.

RESULTS

Of 309 378 patients first diagnosed with an SMI in primary care between 2000 and 2017, 212,618 (68.7%) were prescribed an antipsychotic between 2000 and 2019. Antipsychotic prescribing prevalence was 426 (95% CI, 420-433) per 1000 patients in the year 2000, reaching a peak of 550 (547-553) in 2016, decreasing to 470 (468-473) in 2019. The proportion prescribed antipsychotics was higher among patients diagnosed with schizophrenia (81.0%) than with bipolar disorder (64.6%) and other non-organic psychoses (65.7%). Olanzapine, quetiapine, risperidone and aripiprazole accounted for 78.8% of all antipsychotic prescriptions. Higher mean olanzapine equivalent total daily doses were prescribed to patients with the following characteristics: schizophrenia diagnosis, ethnic minority status, male gender, younger age and greater relative deprivation.

CONCLUSIONS

Antipsychotic prescribing is dominated by olanzapine, quetiapine, risperidone and aripiprazole. We identified potential disparities in both the receipt and prescribed doses of antipsychotics across subgroups. To inform efforts to optimise prescribing and ensure equity of care, further research is needed to understand why certain groups are prescribed higher doses and are more likely to be treated with long-acting injectable antipsychotics compared with others.

Obstructive sleep apnoea syndrome and future risk of dementia among individuals managed in UK general practice

BACKGROUND

Obstructive sleep apnoea syndrome (OSAS) has been recognised as a potential risk factor for cognitive decline, yet its precise relationship with dementia remains uncertain. This study aimed to determine the risk of dementia among individuals with and without OSAS.

METHODS

Data derived from 2.3 million adults (aged ≥18 years) were extracted from the Clinical Practice Research Datalink (2000-2022), a nationally representative primary care electronic health records database in the UK. 193 600 individuals with OSAS were propensity score-matched to 536 701 individuals without OSAS. Cox proportional hazard regression models were applied to quantify the risk of developing all-cause dementia, vascular dementia and Alzheimer's disease between individuals with and without OSAS.

RESULTS

With a median follow-up of 4.0 (IQR 1.8-7.5) years, 2802 and 6211 individuals developed all-cause dementia in those with and without OSAS, corresponding to crude incidence rates of 2.47 and 2.34 per 1000 person-years, respectively. The presence of OSAS was associated with higher risks of all-cause dementia (adjusted HR (aHR) 1.12, 95% CI 1.07 to 1.17), vascular dementia (1.29, 95% CI 1.19 to 1.41) and unchanged risk of Alzheimer's disease (1.07, 95% CI 0.99 to 1.16). Individuals with OSAS who had received continuous positive airway pressure (CPAP) treatment exhibited a similar risk of all-cause dementia as their matched counterparts (0.99, 95% CI 0.74 to 1.32).

CONCLUSION

OSAS is associated with higher risks of all-cause dementia and some of its subtypes. Further investigation is needed to investigate the clinical benefits of screening for cognitive impairment in people with OSAS and to further evaluate the impact of CPAP on cognitive decline and dementia risk.

Long-term exposure to ambient air pollution and risk of lung cancer - A comparative analysis of incidence and mortality in four administrative cohorts in the ELAPSE study

BACKGROUND

Studies have linked air pollution to lung cancer incidence and mortality, but few have compared these associations, which may differ due to cancer survival variations. We aimed to evaluate the association between long-term air pollution exposure and lung cancer incidence and compare findings with previous lung cancer mortality analyses within the same cohorts.

METHODS

We analyzed four population-based administrative cohorts in Denmark (2000-2015), England (2011-2017), Norway (2001-2016) and Rome (2001-2015). We assessed residential exposure to annual average fine particulate matter (PM2.5), nitrogen dioxide (NO₂), black carbon (BC), and warm-season ozone (O3) using Europe-wide land use regression models. We used Cox proportional hazard models to evaluate cohort-specific hazard ratios (HRs) and 95% confidence intervals (CIs) for lung cancer incidence identified using hospital admission records (English and Roman cohorts) or cancer registries (Danish and Norwegian cohorts). We evaluated the associations at low exposure levels using subset analyses and natural cubic splines. Cohort-specific HRs were pooled using random-effects meta-analyses, separately for incidence and mortality.

RESULTS

Over 93,733,929 person-years of follow-up, 111,949 incident lung cancer cases occurred. Incident lung cancer was positively associated with PM2.5, NO2 and BC, and negatively associated with O3. The negative O3 association became positive after adjustment for NO2. Associations were almost identical or slightly stronger for lung cancer incidence than mortality in the same cohorts, with respective meta-analytic HRs (95% CIs) of 1.14 (1.06, 1.22) and 1.12 (1.02, 1.22) per 5 μg/m3 increase in PM2.5, and 1.10 (1.04, 1.16) and 1.09 (1.02, 1.16) per 10 μg/m3 increase in NO2. Positive associations persisted for both incidence and mortality at low pollution levels with similar magnitude.

CONCLUSIONS

We found similarly elevated risks of lung cancer incidence and mortality in association with residential exposure to PM2.5, NO2 and BC in meta-analyses of four European administrative cohorts, which persisted at low pollution levels.

Underlying disease risk among patients with fatigue: a population-based cohort study in primary care

BACKGROUND

Presenting to primary care with fatigue is associated with a wide range of conditions, including cancer, although their relative likelihood is unknown.

AIM

To quantify associations between new-onset fatigue presentation and subsequent diagnosis of various diseases, including cancer.

DESIGN AND SETTING

A cohort study of patients presenting in English primary care with new-onset fatigue during 2007-2017 (the fatigue group) compared with patients who presented without fatigue (the non-fatigue group), using Clinical Practice Research Datalink data linked to hospital episodes and national cancer registration data.

METHOD

The excess short-term incidence of 237 diseases in patients who presented with fatigue compared with those who did not present with fatigue is described. Disease-specific 12-month risk by sex was modelled and the age-adjusted risk calculated.

RESULTS

The study included 304 914 people in the fatigue group and 423 671 in the non-fatigue group. In total, 127 of 237 diseases studied were more common in men who presented with fatigue than in men who did not, and 151 were more common in women who presented with fatigue. Diseases that were most strongly associated with fatigue included: depression; respiratory tract infections; insomnia and sleep disturbances; and hypo/hyperthyroidism (women only). By age 80 years, cancer was the third most common disease and had the fourth highest absolute excess risk in men who presented with fatigue (fatigue group: 7.01%, 95% confidence interval [CI] = 6.54 to 7.51; non-fatigue group: 3.36%, 95% CI = 3.08 to 3.67; absolute excess risk 3.65%). In women, cancer remained relatively infrequent; by age 80 years it had the thirteenth highest excess risk in patients who presented with fatigue.

CONCLUSION

This study ranked the likelihood of possible diagnoses in patients who presented with fatigue, to inform diagnostic guidelines and doctors' decisions. Age-specific findings support recommendations to prioritise cancer investigation in older men (aged ≥70 years) with fatigue, but not in women at any age, based solely on the presence of fatigue.

The effect of workload on primary care doctors on referral rates and prescription patterns: evidence from English NHS

This paper investigates the impact of workload pressure on primary care outcomes using a unique dataset from English general practices. Leveraging the absence of General Practitioner (GP) colleagues as an instrumental variable, we find that increased workload leads to an increase in prescription rates of antibiotics as well as in the share of assessment referrals. On the other hand, the quantity and frequency of psychotropics decreases. When there is an absence, workload is intensified mostly on GP partners, and the mode of consultation shifts toward remote interactions as a response to higher workload pressure. The effects are more pronounced for patients above 65 years-old and those in Short-staffed practices. Our study sheds light on the intricate relationship between workload pressure and patient care decisions in primary care settings.

Long-term trends in English general practice consultation rates from 1995 to 2021: a retrospective analysis of two electronic health records databases

OBJECTIVES

Although primary care is central to healthcare provision, inconsistent methods and data sources mean that relatively little is known about long-term trends in general practice consultation rates. We aimed to explore long-term trends in English general practice consultation rates using two electronic health records databases, Clinical Practice Research Datalink Gold and Aurum, from 1995 to 2021.

METHODS

Consultations were identified and classified using a set of rules and code lists. Consultation rates were calculated as the ratio of the count of consultations and the registration duration. We used negative binomial regression to model the relationship between the number of consultations and the patient's age, sex and year from each database. These models are then used to estimate annual crude consultation rates for England.

RESULTS

The number of general practices in the Gold database decreased from 346 (1995) to 41 (2021), while in Aurum, it increased from 555 (1995) to 1347 (2021). In Gold, the average number of consultations per person-year increased from 2.91 in 1995 to 5.12 in 2012, then fell to 4.59 in 2019 and to 4.12 in 2021. In Aurum, average consultations per person-year rose from 2.17 (1995) to 4.89 (2012), then fell to 4.76 (2019) and rose again to 5.40 (2021). Half of the total increase in consultations from Aurum was due to a new consultation code, for 'AccurX consultations', an electronic messaging system for communication with patients.

CONCLUSION

Trends in general practice consultation showed three distinct epochs: rising from 1995 to 2012, falling from 2012 to 2019 and rising in Aurum but falling in Gold from 2019 to 2021. Consultation rates in Gold were higher than Aurum until the inclusion of a new consultation code in Aurum in 2019, which underscores the need for operational definitions of a consultation.

Cardiorenal effects of angiotensin-converting enzyme inhibitors and angiotensin receptor blockers among people underrepresented in trials: analysis of routinely collected data with emulation of a reference trial (ONTARGET)

Cardiovascular disease is a leading cause of death globally. Angiotensin-converting enzyme inhibitors (ACEi) and angiotensin receptor blockers (ARB), compared in the ONTARGET trial (Ongoing Telmisartan Alone and in Combination with Ramipril Global Endpoint Trial), each prevent cardiovascular disease. However, trial results may not be generalizable, and their effectiveness in underrepresented groups is unclear. Using trial emulation methods within routine-care data to validate findings, we explored the generalizability of ONTARGET results. For people prescribed an ACEi/ARB in the UK Clinical Practice Research Datalink GOLD dataset from January 1, 2001, to July 31, 2019, we applied trial criteria and propensity-score methods to create an ONTARGET trial-eligible cohort. Comparing ARB with ACEi, we estimated hazard ratios for the primary composite trial outcome (cardiovascular death, myocardial infarction, stroke, or hospitalization for heart failure) and secondary outcomes. Because the prespecified criteria were met, confirming trial emulation, we then explored treatment heterogeneity among 3 trial-underrepresented subgroups: females, persons aged ≥75 years, and those with chronic kidney disease. In the trial-eligible population (n = 137 155), results for the primary outcome demonstrated similar effects of ARB and ACEi (hazard ratio = 0.97; 95% CI, 0.93-1.01), meeting the prespecified validation criteria. When extending this outcome to trial-underrepresented groups, similar treatment effects were observed by sex, age, and chronic kidney disease. This suggests that ONTARGET trial findings are generalizable to trial-underrepresented subgroups. This article is part of a Special Collection on Pharmacoepidemiology.

Lithium Use During Pregnancy in 14 Countries

Importance

In pregnancy, the benefits of lithium treatment for relapse prevention in psychiatric conditions must be weighed against potential teratogenic effects. Currently, there is a paucity of information on how and when lithium is used by pregnant women.

Objective

To examine lithium use in the perinatal period.

Design, Setting, and Participants

This cohort study used individual-level data of pregnancies from January 1, 2000, to December 31, 2021, in Australia, Denmark, Finland, Germany, Hong Kong, Iceland, Israel, New Zealand, Norway, South Korea, Sweden, Taiwan, the UK, and 2 cohorts in the US. Analyses were performed from September 1 to November 30, 2023.

Exposures

The prevalence of lithium use as the proportion of pregnancies with at least 1 prescription fill or prescription within 3 months before pregnancy until childbirth was estimated using a common protocol. Lithium use during pregnancy by trimester and in the 3 months before and after pregnancy was examined.

Main Outcomes and Measures

Comparison of prevalence between the first and last 3-year periods of available data.

Results

Among 21 659 454 pregnancies from all collaborating sites, the prevalence of lithium use ranged from 0.07 per 1000 pregnancies in Hong Kong to 1.56 per 1000 in the US publicly insured population. Lithium use increased per 1000 pregnancies in 10 populations (Australia [0.60 to 0.74], Denmark [0.09 to 0.51], Finland [0.10 to 0.29], Iceland [0.24 to 0.99], Israel [0.25 to 0.37], Norway [0.24 to 0.47], South Korea [0.30 to 0.44], Sweden [0.42 to 1.07], the UK [0.07 to 0.10], and Taiwan [0.15 to 0.19]), remained stable in 4 populations (Germany [0.17 to 0.16], Hong Kong [0.06 to 0.06], and the publicly [1.50 to 1.34] and commercially [0.38 to 0.36] insured US populations), and decreased in 1 population (New Zealand [0.54 to 0.39]). Use of lithium decreased with each trimester of pregnancy, while prevalence of postpartum use was similar to prepregnancy levels. The proportion of lithium use in the second trimester compared with the prepregnancy period ranged from 2% in South Korea to 80% in Denmark.

Conclusions and Relevance

Prevalence of lithium use in pregnant women over the past 2 decades varied markedly between populations. Patterns of use before, during, and after pregnancy suggest that many women discontinued lithium use during pregnancy and reinitiated treatment after childbirth, with large variations between countries. These findings underscore the need for internationally harmonized guidelines, specifically for psychiatric conditions among pregnant women that may benefit from lithium treatment.

Dicloxacillin is an inducer of intestinal P-glycoprotein but neither dicloxacillin nor flucloxacillin increases the risk of stroke/systemic embolism in direct oral anticoagulant users

AIM

We aimed to assess if dicloxacillin/flucloxacillin reduces the therapeutic efficacy of direct oral anticoagulants (DOACs) and the underlying molecular mechanism.

METHODS

In a randomized, crossover study, we assessed whether dicloxacillin reduces oral absorption of drugs through P-glycoprotein (P-gp) during 10 and 28 days of treatment. To study the impact of dicloxacillin/flucloxacillin on intestinal and hepatic expression of P-gp in vitro, we usd LS174T cells and 3D spheroids of primary human hepatocytes. Finally, we used nationwide Danish health registries and the UK's Clinical Practice Research Datalink to estimate hazard ratios (HRs) for the risk of stroke and systemic embolism following dicloxacillin/flucloxacillin exposure among DOAC users, using phenoxymethylpenicillin and amoxicillin as active comparators.

RESULTS

Dicloxacillin reduced the area under the curve of dabigatran to a geometric mean ratio 10 days of 0.67 (95% confidence interval [CI]: 0.42-1.1) and geometric mean ratio 28 days of 0.72 (95% CI: 0.39-1.4), suggesting reduced oral absorption via increased P-gp expression. In vitro, dicloxacillin raised P-gp expression in both intestinal and liver cells, while flucloxacillin only affected liver cells. In the pharmacoepidemiologic study, dicloxacillin and flucloxacillin were not associated with increased risk of stroke/systemic embolism (dicloxacillin vs. phenoxymethylpenicillin HR: 0.93, 95% CI: 0.72-1.2; flucloxacillin vs. amoxicillin HR: 0.89, 95% CI: 0.51-1.5).

CONCLUSIONS

Dicloxacillin increases expression of intestinal P-gp, leading to reduced oral absorption of dabigatran. However, concomitant use of dicloxacillin/flucloxacillin was not associated with stroke and systemic embolism among DOAC users, suggesting no clinical impact from the drug-drug interaction between dicloxacillin/flucloxacillin and DOACs.

External validation of a prognostic model to improve prediction of psychosis: a retrospective cohort study in primary care

BACKGROUND

Early detection could reduce the duration of untreated psychosis. GPs are a vital part of the psychosis care pathway, but find it difficult to detect the early features. An accurate risk prediction tool, P Risk, was developed to detect these.

AIM

To externally validate P Risk.

DESIGN AND SETTING

This retrospective cohort study used a validation dataset of 1 647 934 UK Clinical Practice Research Datalink (CPRD) primary care records linked to secondary care records.

METHOD

The same predictors (age; sex; ethnicity; social deprivation; consultations for suicidal behaviour, depression/anxiety, and substance misuse; history of consultations for suicidal behaviour; smoking history; substance misuse; prescribed medications for depression/anxiety/post-traumatic stress disorder/obsessive compulsive disorder; and total number of consultations) were used as for the development of P Risk. Predictive risk, sensitivity, specificity, and likelihood ratios were calculated for various risk thresholds. Discrimination (Harrell's C-index) and calibration were calculated. Results were compared between the development (CPRD GOLD) and validation (CPRD Aurum) datasets.

RESULTS

Psychosis risk increased with values of the P Risk prognostic index. Incidence was highest in younger age groups and, in the main, higher in males. Harrell's C was 0.79 (95% confidence interval = 0.78 to 0.79) in the validation dataset and 0.77 in the development dataset. A risk threshold of 1.0% gave sensitivity of 65.9% and specificity of 86.6%.

CONCLUSION

Further testing is required, but P Risk has the potential to be used in primary care to detect future risk of psychosis.

Federated learning of medical concepts embedding using BEHRT

Objectives

Electronic health record data is often considered sensitive medical information. Therefore, the EHR data from different medical centers often cannot be shared, making it difficult to create prediction models using multicenter EHR data, which is essential for such models' robustness and generalizability. Federated learning (FL) is an algorithmic approach that allows learning a shared model using data in multiple locations without the need to store all data in a single central place. Our study aims to evaluate an FL approach using the BEHRT model for predictive tasks on EHR data, focusing on next visit prediction.

Materials and Methods

We propose an FL approach for learning medical concepts embedding. This pretrained model can be used for fine-tuning for specific downstream tasks. Our approach is based on an embedding model like BEHRT, a deep neural sequence transduction model for EHR. We train using FL, both the masked language modeling (MLM) and the next visit downstream model.

Results

We demonstrate our approach on the MIMIC-IV dataset. We compare the performance of a model trained with FL to one trained on centralized data, observing a difference in average precision ranging from 0% to 3% (absolute), depending on the length of the patients' visit history. Moreover, our approach improves average precision by 4%-10% (absolute) compared to local models. In addition, we show the importance of the usage of pretrained MLM for the next visit diagnoses prediction task.

Discussion and Conclusion

We find that our FL approach reaches very close to the performance of a centralized model, and it outperforms local models in terms of average precision. We also show that pretrained MLM improves the model's average precision performance in the next visit diagnoses prediction task, compared to an MLM without pretraining.

Maternal and pregnancy factors contributing to the association between area deprivation and infant mortality in England: a retrospective cohort study

Background

Socioeconomic inequality in infant mortality in the UK is rising. This study aims to identify contributory maternal and pregnancy factors that can explain the known association between area deprivation and infant mortality.

Methods

A cohort study was conducted using Clinical Practice Research Datalink (CPRD) primary care data between 2004 and 2019 linked to the Index of Multiple Deprivation (IMD), and infant mortality from the Office for National Statistics death data. Potential maternal and pregnancy contributory factors included: maternal age, prior maternal health conditions, pregnancy lifestyle factors and complications, use of medications during pregnancy, and characteristics of birth. Counterfactual-based decomposition analysis was used to quantify the relative importance of equalising these factors to reduce inequalities in infant mortality.

Findings

A total of 392,606 mother-child dyads were included in this study. The overall risk of infant mortality was greatest for individuals in the most deprived quintile (risk ratio 2.13 [95% CI 1.58-2.90]; risk difference 6.6 [3.8-8.8] per 10,000 live births) compared with the least deprived. Four contributory factors were identified as potentially important: preterm birth (Proportion eliminated [PE] 15.25% [95% CI 9.44-24.12%]), smoking during pregnancy (PE 13.61% [95% CI 3.96-80.97%]), maternal age <20 years at childbirth (PE 10.52% [95% CI 2.93-21.35%]) and maternal depression (PE 9.13% [95% CI 4.47-14.93%]). These collectively accounted for more than one-third of the socioeconomic inequality in mortality.

Interpretation

Multifactorial interventions targeting maternal mental health, smoking, teenage pregnancy and preterm birth may mitigate a proportion of the effects of socioeconomic inequality but targeting these, alone, will not stem the rise in infant mortality. Structural efforts to reduce socioeconomic inequalities will also be required to prevent these excess infant deaths.

Funding

UK Medical Research Council, Scottish Chief Scientist Office, Wellcome Trust.

Use of heart failure medical therapy before and after a cancer diagnosis: A longitudinal study

AIMS

We aim to evaluate change in the use of prognostic guideline-directed medical therapies (GDMTs) for heart failure (HF) before and after a cancer diagnosis as well as the matched non-cancer controls, including renin-angiotensin-system inhibitors (RASIs), beta-blockers, and mineralocorticoid receptor antagonists (MRAs).

METHODS AND RESULTS

We conducted a longitudinal study in patients with HF in the UK Clinical Practice Research Datalink between 2005 and 2021. We selected patients with probable HF with reduced ejection fraction (HFrEF) based on diagnostic and prescription records. We described the longitudinal trends in the use and dosing of GDMTs before and after receiving an incident cancer diagnosis. HF patients with cancer were matched with a 1:1 ratio to HF patients without cancer to investigate the association between cancer diagnosis and treatment adherence, persistence, initiation, and dose titration as odds ratios (ORs) with 95% confidence intervals (CIs) using multivariable logistic regression models. Of 8504 eligible HFrEF patients with incident cancer, 4890 were matched to controls without cancer. The mean age was 75.7 (±8.4) years and 73.9% were male. In the 12 months following a cancer diagnosis, patients experienced reductions in the use and dosing of GDMT. Compared with the non-cancer controls, patients with cancer had higher risks for poor adherence for all three medication classes (RASIs: OR = 1.51, 95% CI = 1.35-1.68; beta-blockers: OR = 1.22, 95% CI = 1.08-1.37; MRAs: OR = 1.31, 95% CI = 1.08-1.59) and poor persistence (RASIs: OR = 2.04, 95% CI = 1.75-2.37; beta-blockers: OR = 1.35, 95% CI = 1.12-1.63; MRAs: OR = 1.49, 95% CI = 1.16-1.93), and higher risks for dose down-titration for RASIs (OR = 1.69, 95% CI = 1.40-2.04) and beta-blockers (OR = 1.31, 95% CI = 1.05-1.62). Cancer diagnosis was not associated with treatment initiation or dose up-titration. Event rates for HF hospitalization and mortality were higher in patients with poor adherence or persistence to GDMTs.

CONCLUSIONS

Following a cancer diagnosis, patients with HFrEF were more likely to have reduced use of GDMTs for HF.

Are patients with newly diagnosed frozen shoulder more likely to be diagnosed with type 2 diabetes? A cohort study in UK electronic health records

AIM

To estimate the association between newly diagnosed frozen shoulder and a subsequent diagnosis of type 2 diabetes in primary care.

METHODS

We conducted an age-, gender- and practice-matched cohort study in UK primary care electronic medical records containing 31 226 adults diagnosed with frozen shoulder, matched to 31 226 without frozen shoulder. Patients with pre-existing diabetes were excluded. Variables were identified using established Read codes. A hazard ratio (HR) for the association between incident frozen shoulder and a subsequent type 2 diabetes diagnosis was estimated using shared frailty Cox regression, adjusted for age and gender. To determine whether the association could be explained by increased testing for type 2 diabetes based on other risk factors, a secondary analysis involved re-running the Cox model adjusting for the mean number of consultations per year, hyperlipidaemia, hypertension, obesity, thyroid dysfunction, ethnicity, deprivation, age, and gender.

RESULTS

Participants with frozen shoulder were more likely to be diagnosed with type 2 diabetes (1559 out of 31 226 patients [5%]) than participants without frozen shoulder (88 out of 31 226 patients [0.28%]). The HR for a diagnosis of type 2 diabetes in participants with frozen shoulder versus people without frozen shoulder was 19.4 (95% confidence interval [CI] 15.6-24.0). The secondary analysis, adjusting for other factors, produced similar results: HR 20.0 (95% CI 16.0-25.0).

CONCLUSIONS

People who have been newly diagnosed with frozen shoulder are more likely to be diagnosed with type 2 diabetes in the following 15.8 years. The value of screening patients presenting with frozen shoulder for type 2 diabetes at presentation, alongside more established risk factors, should be considered in future research.

Opportunities and challenges for identifying undiagnosed Rare Disease patients through analysis of primary care records: long QT syndrome as a test case

BACKGROUND

Patients with rare genetic diseases frequently experience significant diagnostic delays. Routinely collected data in the electronic health record (EHR) may be used to help identify patients at risk of undiagnosed conditions. Long QT syndrome (LQTS) is a rare inherited cardiac condition associated with significant morbidity and premature mortality. In this study, we examine LQTS as an exemplar disease to assess if clinical features recorded in the primary care EHR can be used to develop and validate a predictive model to aid earlier detection.

METHODS

1495 patients with an LQTS diagnostic code and 7475 propensity-score matched controls were identified from 10.5 million patients' electronic primary care records in the UK's Clinical Practice Research Datalink (CPRD). Associated clinical features recorded before diagnosis (with p < 0.05) were incorporated into a multivariable logistic regression model, the final model was determined by backwards regression and validated by bootstrapping to determine model optimism.

RESULTS

The mean age at LQTS diagnosis was 58.4 (SD 19.41). 18 features were included in the final model. Discriminative accuracy, assessed by area under the curve (AUC), was 0.74, (95% CI 0.73, 0.75) (optimism 6%). Features occurring at significantly greater frequency before diagnosis included: epilepsy, palpitations, syncope, collapse, mitral valve disease and irritable bowel syndrome.

CONCLUSION

This study demonstrates the potential to develop primary care prediction models for rare conditions, like LQTS, in routine primary care records and highlights key considerations including disease suitability, finding an appropriate linked dataset, the need for accurate case ascertainment and utilising an approach to modelling suitable for rare events.

Estimating the healthcare burden of Prurigo Nodularis in England: a CPRD database study

Purpose: Prurigo nodularis (PN) is a skin disease characterized by intensely itchy skin nodules and is associated with a significant healthcare resource utilization (HCRU). This study aimed to estimate the HCRU of patients in England with PN overall and moderate-to-severe PN (MSPN) in particular.

Methods: This retrospective cohort study used data from the Clinical Practice Research Datalink and Hospital Episode Statistics in England. Patients with Mild PN (MiPN) were matched to patients with MSPN by age and gender for the primary analysis. Patients were enrolled in the study between 1st April 2007 and 1st March 2019. All-cause HCRU was calculated, including primary and secondary care contacts and costs (cost-year 2022).

Results: Of 23,522 identified patients, 8,933 met the inclusion criteria, with a primary matched cohort of 2,479 PN patients. During follow up, the matched cohort's primary care visits were 21.27 per patient year (PPY) for MSPN group and 11.35 PPY for MiPN group. Any outpatient visits were 10.72 PPY and 4.87 PPY in MSPN and MiPN groups, respectively. Outpatient dermatology visits were 1.96 PPY and 1.14 PPY in MSPN and MiPN groups, respectively.

Conclusion: PN, especially MSPN, has a high HCRU burden in England, highlighting the need for new and improved disease management treatments.

Differences in recording of cancer diagnosis between datasets in England: A population-based study of linked cancer registration, hospital, and primary care data

BACKGROUND

Differences in the recording of cancer case status and diagnosis date have been observed between cancer registry (CR) - the reference standard - and electronic health records (EHRs); such differences may affect estimates of cancer risk or misclassify diagnostic pathways. This study aims to quantify differences in recording of case status and date of cancer diagnosis between cancer registry and EHRs.

METHODS

Linked primary care (Clinical Practice Research Datalink (CPRD)), secondary care (Hospital Episode Statistics (HES)) and national Cancer Registry (CR) data, were used to identify 14,301 patients with a recorded diagnosis of brain, colon, lung, ovarian, or pancreatic cancer between 1999 and 2018. Agreement in case status between datasets, differences in recorded diagnosis dates, and change in agreement over time were investigated for each cancer site.

RESULTS

Between 84 % (ovary) to 92 % (colon) of diagnoses in cancer registry were also recorded in combined CPRD-HES data. Agreement with cancer registry was slightly lower in HES (78 % (ovary) to 86 % (colon)) and CPRD (61 % (ovary, pancreas) to 72 % (brain)). The proportion of CPRD-HES diagnoses confirmed in CR varied by cancer site (50 % (brain) to 86 % (lung)). Agreement between CR and HES was relatively stable within cancer sites over time. Concordance between CR and CPRD was more heterogeneous between cancer sites and over time. Best agreement in diagnosis date was observed between CR and HES (median difference 0 or 1 days, all cancer sites).

CONCLUSION

Agreement between CR and EHR data is heterogeneous across cancer sites. Concordance does not appear to have improved over time. Combined data from primary and secondary care may be sufficient to approximate case status in CR in some circumstances, but the date we consider to represent the diagnosis may impact study outcomes.

Incidence of Side Effects Associated With Acetaminophen in People Aged 65 Years or More: A Prospective Cohort Study Using Data From the Clinical Practice Research Datalink

OBJECTIVE

The main objective of this study is to examine the safety of oral acetaminophen at its therapeutic dose in adults aged ≥65 years.

METHODS

This population-based cohort study used the Clinical Practice Research Datalink-Gold data. Participants were aged ≥65 years registered with a UK general practice for at least 12 months between 1998 and 2018. Acetaminophen exposure was defined as at least two acetaminophen prescriptions within six months of the first acetaminophen prescription, the first prescription date being the index date. Acetaminophen nonexposure was defined as the absence of two acetaminophen prescriptions within six months over the study period. We calculated propensity score (PS) for acetaminophen prescription and undertook inverse probability treatment weighting using PS and PS-matched analyses to account for confounding. Missing data were handled using multiple imputation. The adjusted hazard ratio (aHR) and 95% confidence interval (95% CI) were calculated using the Cox proportional hazards regression model.

RESULTS

In total, 180,483 acetaminophen exposed and 402,478 unexposed participants were included in this study. Acetaminophen exposure was associated with an increased risk of perforation or ulceration or bleeding (aHR 1.24; 95% CI 1.16-1.34), uncomplicated peptic ulcers (aHR 1.20; 95% CI 1.10-1.31), lower gastrointestinal bleeding (aHR 1.36; 95% CI 1.29-1.46), heart failure (aHR 1.09; 95% CI 1.06-1.13), hypertension (aHR 1.07; 95% CI 1.04-1.11), and chronic kidney disease (aHR 1.19; 95% CI 1.13-1.24).

CONCLUSION

Despite its perceived safety, acetaminophen is associated with several serious complications. Given its minimal analgesic effectiveness, acetaminophen as the first-line oral analgesic option for long-term conditions in older people requires careful reconsideration.

Epidemiology of Childhood Asthma in the UK

Purpose

Global prevalence of pediatric asthma and associated morbidity and mortality has continuously increased. Asthma is the most common chronic illness in children in the UK; however, recent epidemiology data are lacking. This analysis describes the overall prevalence and burden of illness of asthma in children.

Methods

This was a retrospective, longitudinal, database analysis using the Clinical Practice Research Datalink database. Primary care records of 19,330 patients (6-11 years) between January 1 and December 31, 2017, were analyzed. Asthma prevalence was assessed by severity (as described by Global Initiative for Asthma 2017 guidelines), and symptoms, comorbidities, and treatments were compared between asthma patients and matched non-asthmatic controls. Results are presented descriptively; logistic regression analyses were performed for asthma symptoms.

Results

The estimated prevalence of pediatric asthma was 6.5% (95% CI: 6.4-6.5) in the UK (mild: 74.2%; moderate: 15.0%; severe: 10.8%). All patients with moderate or severe asthma and 72.5% of patients with mild asthma were prescribed drug therapy. Most patients with moderate or severe asthma were prescribed a short-acting β2-agonist (94.9% and 96.0%, respectively), compared with 69.2% of mild asthma patients. Daytime symptoms were reported by 78.1% in those with severe asthma; 34.9% reported night-time symptoms and 30.8% reported an impact on usual activities. Asthma patients had a higher baseline prevalence of comorbidities compared with non-asthmatic controls, notably atopic dermatitis (47.8% in severe asthma versus 20.8% in controls) and allergic rhinitis (13.3% in severe asthma versus 2.0% in controls).

Conclusion

This analysis confirmed that asthma remains a common morbidity among children in the UK. Increasing asthma severity was associated with worsening symptoms, and asthma patients had significantly more comorbidities compared with non-asthmatic controls.

Oral Corticosteroid-Related Healthcare Resource Utilization and Associated Costs in Patients with COPD

INTRODUCTION

Oral corticosteroids (OCS) are used to manage chronic obstructive pulmonary disease (COPD) exacerbations but are associated with adverse outcomes that may increase healthcare resource utilization and costs. We compared attendance/costs associated with OCS-related adverse outcomes in patients who ever used OCS versus those who never used OCS and examined associations between cumulative OCS exposure and attendance/costs.

METHODS

This direct matched observational cohort study used the UK Clinical Practice Research Datalink GOLD database (data range 1987-2019). Patients with a COPD diagnosis on/after April 1, 2003, and Hospital Episode Statistics linkage were included. Emergency room, specialist or primary care outpatient, and inpatient attendance were analyzed. Costs, estimated using Health and Social Care 2019 and National Health Service Reference Costs 2019-2020 reports, were adjusted for sex, age, exacerbation number, and inhaler type used in the 12 months before index date.

RESULTS

The OCS cohort had higher annualized disease-specific (excluding respiratory) total attendance/costs versus the non-OCS cohort (adjusted incidence rate ratio [aIRR] with 95% confidence intervals [CIs]) ranging from 37% (1.37 [1.31, 1.43]) for emergency room attendances to 149% (2.49 [2.36, 2.63]) for specialist consultations. Disease-specific (excluding respiratory) attendance/costs increased in a positive dose-response relationship for most attendance categories versus the < 0.5 g reference dose. For the 0.5 to < 1.0 g cumulative dose category, the greatest increases in disease-specific (excluding respiratory) attendance/costs occurred for primary care consultations (aIRR [95% CI] 1.38 [1.32, 1.44]). For the ≥ 10 g cumulative dose category, the greatest increases were observed for primary care consultations (aIRR [95% CI] 2.83 [2.66, 3.00]), non-elective long stays (≥ 2 days; 2.54 [2.15, 2.99]), and non-elective short stays (≤ 1 day; 2.51 [2.12, 2.98]). Similar findings were observed for all-cause attendance/costs.

CONCLUSION

Among patients with COPD, OCS-related adverse outcomes were associated with higher attendance and costs, with a positive dose-response relationship. A graphical abstract is available with this article.

Risk of dementia associated with anticholinergic drugs for overactive bladder in adults aged ≥55 years: nested case-control study

ABSTRACT

Objective

To investigate whether different anticholinergic drug treatments for overactive bladder have differential risks for incident dementia, in a large representative population of older adults in England.

Design

Nested case-control study.

Setting

General practices in England providing data to the Clinical Practice Research Datalink (CPRD) GOLD database, with linked patient admission records from secondary care (Hospital Episode Statistics), 1 January 2006 and 16 February 2022.

Participants

170 742 patients aged ≥55 years, with a first reported diagnosis of dementia during the study period, matched by age, sex, and general practice with 804 385 individuals without dementia (controls).

Interventions

Cumulative drug use (defined using total standardised daily dose) of different anticholinergic drugs used for the treatment of an overactive bladder, and a non-anticholinergic drug, mirabegron, in the period 3-16 years before a diagnosis of dementia (or equivalent date in matched controls).

Main outcome measures

Odds ratios for onset of dementia associated with the different anticholinergic drugs used for the treatment of an overactive bladder, adjusted for sociodemographic characteristics, clinical comorbidities, and use of other anticholinergic drug treatments.

Results

The study population comprised 62.6% women, and median age was 83 (interquartile range 77-87) years. 15 418 (9.0%) patients with dementia and 63 369 (7.9%) controls without dementia had used anticholinergic drugs for the treatment of an overactive bladder in the 3-16 years before diagnosis (or equivalent date for controls). The adjusted odds ratio for dementia associated with the use of any anticholinergic drug used to treat an overactive bladder was 1.18 (95% confidence interval (CI) 1.16 to 1.20), and was higher in men (1.22, 1.18 to 1.26) than women (1.16, 1.13 to 1.19). The risk of dementia was substantially increased with the use of oxybutynin hydrochloride (adjusted odds ratio 1.31, 95% CI 1.21 to 1.42 and 1.28, 1.15 to 1.43 for use of 366-1095 and >1095 total standardised daily doses, respectively), solifenacin succinate (1.18, 1.09 to 1.27 and 1.29, 1.19 to 1.39), and tolterodine tartrate (1.27, 1.19 to 1.37 and 1.25, 1.17 to 1.34). No significant increases in the risk of dementia associated with darifenacin, fesoterodine fumarate, flavoxate hydrochloride, propiverine hydrochloride, and trospium chloride were found. The association between mirabegron, a non-anticholinergic drug, and dementia was variable across the dose categories and might be caused by previous use of anticholinergic drugs for the treatment of an overactive bladder in these individuals.

Conclusions

Of the different anticholinergic drugs used to treat an overactive bladder, oxybutynin hydrochloride, solifenacin succinate, and tolterodine tartrate were found to be most strongly associated with the risk of dementia in older adults. This finding emphasises the need for clinicians to take into account the possible long term risks and consequences of the available treatment options for an overactive bladder in older adults, and to consider prescribing alternative treatments that might be associated with a lower risk of dementia.

The risk of acute infection in association with first ever diagnosed depression: a cohort study

PURPOSE

To assess the risk of acute infections in patients with first ever diagnosed depression compared to patients with no diagnosed depression in a primary-care database.

METHODS

We conducted a cohort study using the UK CPRD GOLD database (2000-2019). We identified patients aged 18 years or older with a recorded Read code for depression (cohort entry date) and compared them to patients with no Read codes for depression using risk set sampling. Comparison groups were frequency-matched on age and sex, and comparison patients were required to have ≥ 1 general practitioner (GP) contact within 14 days before cohort entry. The primary outcome was a composite of outpatient diagnosed acute infections, including respiratory, gastrointestinal, urogenital infections and septicemia) within the two-years after cohort entry. We applied propensity score fine stratification and estimated incidence rates and IR ratios (IRR) using negative binomial regression.

RESULTS

In a weighted population of 285,922 patients with diagnosed depression and 285,921 comparison patients, the IR of acute infections was 97.3/1000 person-years (py) in patients with and 83.7/1000 py in patients with no diagnosed depression. The weighted IRR of acute infection was 1.18 (95% CI 1.16-1.20) comparing those with and with no diagnosed depression. Excluding patients with baseline comorbidities yielded an IRR even closer to the null: 1.07 (95% CI, 1.04-1.09).

CONCLUSIONS

Our results suggest that patients with diagnosed depression are not at a meaningfully increased risk of acute infections compared to patients with no diagnosed depression. Slightly increased overall relative risks of infections can be explained by residual differences in health care utilization and by the severity of comorbidities.

Implementation and Effectiveness of Guideline-Recommended Clinical Activities for Children With Asthma: Population-Based Cohort

BACKGROUND

Guidelines advise minimizing asthma exacerbation risk, which is achieved partially through good clinical practice activities, including scheduled asthma reviews, inhaler technique checks, and asthma management plans. We assessed how frequently these activities are provided and how effective they are in clinical practice.

RESEARCH QUESTION

Do guideline-recommended activities such as asthma reviews, inhaler technique checks, and asthma management plans prevent asthma exacerbations?

STUDY DESIGN AND METHODS

This retrospective chart review used United Kingdom primary care medical records between 2004 and 2021, linked to hospital records. Children were eligible from asthma diagnosis until age 16 years. Annual implementation of asthma review, inhaler technique check, and asthma management plan was measured. Risk factors for these activities not being undertaken were determined by using multivariable logistic regression. Self-controlled case series was adopted to assess the effectiveness of each activity over 12 months; this was divided into two 6-month periods.

RESULTS

A total of 126,483 children were eligible; 30% to 45% received each annual activity, and 8% received all 3 together. Risk factors for not receiving activities included younger age, more socioeconomic deprivation, and higher or no BMI measurement. Management plans and asthma reviews, as standalone activities, were associated with an approximately 15% exacerbation reduction over 12 months and 8% over 6 months, respectively (asthma management plan, n = 4,624; 0-6 months [incidence rate ratio (95% CI)]: 0.87 [0.79-0.96]; 6-12 months: 0.83 [0.73-0.95]; asthma review, n = 6,948; 0-6 months: 0.92 [0.85-0.99]; 6-12 months: 0.93 [0.83-1.03]). Standalone inhaler technique checks were not associated with exacerbations. Provision of all activities together was associated with an approximately 30% exacerbation reduction over 12 months (n = 3,643; 0-6 months: incidence rate ratio, 0.76 [0.68-0.85]; 6-12 months: incidence rate ratio, 0.69 [95% CI, 0.60-0.81]).

INTERPRETATION

Most children in the United Kingdom do not receive the guideline-recommended activities to monitor their asthma. This study suggests that these activities, if implemented, are effective in clinical practice and maximally effective when combined in the same visit.

Antidepressant prescribing inequalities in people with comorbid depression and type 2 diabetes: A UK primary care electronic health record study

AIMS

To compare the likelihood of being prescribed an antidepressant in depressed individuals with and without type 2 diabetes.

METHODS

We performed a matched cohort study using primary care record data from the UK Clinical Practice Research Datalink. We used multivariable logistic regression to compare antidepressant prescribing during the first five years of starting oral antidiabetic medication to a comparison group without type 2 diabetes, matched based on GP practice, age and sex. We performed subgroup analyses stratified by sex, age and ethnicity.

RESULTS

People with type 2 diabetes and depression were 75% less likely to be prescribed an antidepressant compared to people with depression alone (odds ratio (OR) 0.25, 95% confidence interval (CI) 0.25 to 0.26). This difference was greater in males (OR 0.23, 95% CI, 0.22 to 0.24), people older than 56 years (OR 0.23, 95% CI, 0.22 to 0.24), or from a minoritised ethnic background (Asian OR 0.14, 95% CI 0.12-0.14; Black OR 0.12, 95% CI 0.09-0.14).

CONCLUSIONS

There may be inequalities in access to antidepressant treatment for people with type 2 diabetes, particularly those who are male, older or from minoritised ethnic backgrounds.

Investigating the Potential Short-term Adverse Effects of the Quadrivalent Human Papillomavirus Vaccine: A Novel Regression Discontinuity Analysis

BACKGROUND

Human papillomavirus (HPV) vaccination has been offered in over a hundred countries worldwide (including the United Kingdom, since September 2008). Controversy around adverse effects persists, with inconsistent evidence from follow-up of randomized controlled trials and confounding by indication limiting the conclusions drawn from larger-scale observational studies. This study aims to estimate the association between receiving a quadrivalent HPV vaccine and the reporting of short-term adverse effects and to demonstrate the utility of regression discontinuity design for examining side effects in routine data.

METHODS

We applied a novel regression discontinuity approach to a retrospective population-based cohort using primary care data from the UK Clinical Practice Research Datalink linked to hospital and social deprivation data. We examined the new onset of gastrointestinal, neuromuscular, pain, and headache/migraine symptoms using READ and International Classification of Diseases, tenth revision diagnostic codes. For each year between 2012 and 2017, we compared girls in school year 8 (born July/August) who were eligible to receive the vaccine with girls in year 7 (born September/October) who were not eligible.

RESULTS

Of the 21,853 adolescent girls in the cohort, 10,881 (50%) were eligible for HPV vaccination. There was no evidence of increased new gastrointestinal symptoms (adjusted odds ratio [OR]: 0.99; 95% confidence interval [CI]: 0.85, 1.15), headache/migraine symptoms (OR: 0.84; 95% CI: 0.70, 1.01), or pain symptoms (OR: 1.05; 95% CI: 0.95, 1.16) when comparing those eligible and ineligible for HPV vaccination.

CONCLUSION

This study found no evidence that HPV vaccination eligibility is associated with reporting short-term adverse effects among adolescent girls.

Reliable generation of privacy-preserving synthetic electronic health record time series via diffusion models

OBJECTIVE

Electronic health records (EHRs) are rich sources of patient-level data, offering valuable resources for medical data analysis. However, privacy concerns often restrict access to EHRs, hindering downstream analysis. Current EHR deidentification methods are flawed and can lead to potential privacy leakage. Additionally, existing publicly available EHR databases are limited, preventing the advancement of medical research using EHR. This study aims to overcome these challenges by generating realistic and privacy-preserving synthetic EHRs time series efficiently.

MATERIALS AND METHODS

We introduce a new method for generating diverse and realistic synthetic EHR time series data using denoizing diffusion probabilistic models. We conducted experiments on 6 databases: Medical Information Mart for Intensive Care III and IV, the eICU Collaborative Research Database (eICU), and non-EHR datasets on Stocks and Energy. We compared our proposed method with 8 existing methods.

RESULTS

Our results demonstrate that our approach significantly outperforms all existing methods in terms of data fidelity while requiring less training effort. Additionally, data generated by our method yield a lower discriminative accuracy compared to other baseline methods, indicating the proposed method can generate data with less privacy risk.

DISCUSSION

The proposed model utilizes a mixed diffusion process to generate realistic synthetic EHR samples that protect patient privacy. This method could be useful in tackling data availability issues in the field of healthcare by reducing barrier to EHR access and supporting research in machine learning for health.

CONCLUSION

The proposed diffusion model-based method can reliably and efficiently generate synthetic EHR time series, which facilitates the downstream medical data analysis. Our numerical results show the superiority of the proposed method over all other existing methods.

Glucosamine and Cancer Incidence in Osteoarthritis: A Prevalent New-User Cohort Design

OBJECTIVE

Observational studies have associated glucosamine, used to treat joint pain and osteoarthritis, with reductions in cancer incidence, although their study design was affected by selection bias. We assessed this association using a study design that mitigates this selection bias.

METHODS

We used the UK Clinical Practice Research Datalink to identify a cohort of patients diagnosed with osteoarthritis during 1995 through 2017. The prevalent new-user cohort design was employed to match glucosamine initiators with non-users on time-conditional propensity scores, who were observed until cancer incidence. Hazard ratios (HRs) and 95% confidence intervals (CIs) of cancer incidence were estimated to compare glucosamine initiators with non-users.

RESULTS

The study cohort of patients with osteoarthritis included 20,541 glucosamine initiators who were matched to 20,541 non-users. Over an average follow-up of eight years, the overall incidence rate of any cancer was 16.4 per 1,000 per year. The HR of any cancer incidence with glucosamine treatment was 0.97 (95% CI 0.91-1.02) compared with non-users. For lung cancer, the HR with glucosamine treatment was 0.99 (95% CI 0.83-1.18), whereas it was 1.11 (95% CI 0.93-1.33) for colorectal cancer, 1.07 (95% CI 0.93-1.23) for breast cancer in women, and 1.03 (95% CI 0.88-1.22) for prostate cancer.

CONCLUSION

In this large, real-world study of patients with osteoarthritis, designed to emulate a trial, treatment with glucosamine did not reduce the incidence of cancer. This finding reinforces that previous studies, not based on glucosamine initiators, were affected by selection bias. Our study does not support the prescription of glucosamine to prevent cancer in patients with osteoarthritis.

No trend to rising rates: A review of Parkinson's prevalence studies in the United Kingdom

BACKGROUND

Of the neurodegenerative diseases, Parkinson's disease is recognised to have the fastest growing prevalence. It is unclear whether this is due to the ageing global population alone, with several environmental factors increasingly implicated in changing prevalence rates. Large data sets have been used nationally and globally to help predict future disease burden. However, the reliability of such sources is yet unknown for Parkinson's disease.

SUMMARY

This review discusses the methods used in all published UK prevalence studies conducted to date. Direct comparison between prevalence figures obtained from the 10 to discussed prevalence studies is precluded due to differences in methodology for case ascertainment and diagnosis. Age adjusted estimates vary from 105/100,000 to 168/100,000.

KEY MESSAGES

These studies demonstrate no overall trend in changing prevalence figures between 1961 and 2007. No difference in prevalence trends were seen for those living in rural or urban areas. Differences between ethnic groups, for example, remains an under explored area.

Dependent censoring bias assessment using inverse probability of censoring weights: Type 2 diabetes mellitus risk in patients initiating bisoprolol versus other antihypertensives in a Clinical Practice Research Datalink cohort study

Dependent censoring involves a preferential attrition of a subgroup of interest; occurring in survival analysis, it may impact interpretation by introducing a selection bias. To assess the potential bias in a comparison of bisoprolol to other antihypertensives in terms of Type 2 diabetes mellitus (T2DM) incidence, inverse probability of censoring weights (IPCW) was used. It was further used to contextualize results obtained through competing risks analysis. Two estimands were considered to assess T2DM incidence while accounting for deviations from the initial antihypertensive monotherapy (DFM). A hypothetical estimand using IPCW, treating DFM as censoring, was interpreted together with a 'while-on-treatment' estimand, treating DFM as a competing risk. We illustrated our application with a cohort study based on Clinical Practice Research Datalink (CPRD) including 267,352 patients with newly diagnosed arterial hypertension between 2000 and 2017, initiating antihypertensive monotherapy among bisoprolol, other beta-blockers, renin-angiotensin system drugs (ACEi/ARB), diuretics and calcium-channel blockers. A mild dependent censoring process was hypothesized, leading to slight overestimation of T2DM incidence. Although subject to some limitations, a nonsignificant trend toward an excess of risk associated with ACEi/ARB was yielded consistently by IPCW and competing risks analyses. Conversely, in comparisons of bisoprolol versus either diuretics, other beta-blockers or calcium channel blockers, no significant differences or critical dependent censoring impact were found. Concurrent use of complementary estimands allowed formulating a refined interpretation of our findings: though not significant, the trend toward an excess of T2DM risk associated with a ACEi/ARB monotherapy compared with bisoprolol is likely not originating only from the minor dependent censoring. Reassessing identical estimands in other cohorts would provide insights to corroborate or refute this result.

Incidence of post-acute COVID-19 symptoms across healthcare settings in seven countries: an international retrospective cohort study using routinely-collected data

Background

The World Health Organisation (WHO) has identified a range of symptomatic manifestations to aid in the clinical diagnosis of post-COVID conditions, herein referred to as post-acute COVID-19 symptoms. We conducted an international network cohort study to estimate the burden of these symptoms in North American, European, and Asian populations.

Methods

A federated analysis was conducted including 10 databases from the United Kingdom, Netherlands, Norway, Estonia, Spain, France, South Korea, and the United States, between September 1st 2020 and latest data availability (which varied from December 31st 2021 to February 28th 2023), covering primary and secondary care, nationwide registries, and claims data, all mapped to the Observational Medical Outcomes Partnership Common Data Model (OMOP CDM). We defined two cohorts for the main analyses: a SARS-CoV-2 infection cohort [positive polymerase chain reaction (PCR) or rapid lateral flow test (LFT) result or clinical COVID-19 diagnosis] and a general population cohort. Individuals with less than 365 days of prior history or 120 days of follow-up were excluded. We estimated incidence rates (IRs) of the 25 WHO-proposed post-acute COVID-19 symptoms, considering symptoms that occurred ≥90 and ≤365 days after index date, excluding individuals with the respective symptoms 180 days prior to the index event. Stratified analyses were conducted by age and sex. Incidence rate ratios (IRRs) were calculated comparing rates in the infected cohort versus the general population. Results from the different databases were combined using random-effects meta-analyses.

Findings

3,019,408 individuals were included in the infection cohort. 1,585,160 of them were female and 1,434,248 of them male. 929,351,505 individuals were included in the general population group. 461,195,036 of them were female and 466,022,004 of them male. The 1-year IR of any post-acute COVID-19 symptom in the COVID-19 infection cohort varied significantly across databases, from 4.4 (95% CI 3.8-5.1) per 100 person-years to 103.9 (95% CI 103.2-104.7). The five most common symptoms were joint pain (from 1.6 (95% CI 1.3-1.9) to 14.3 (95% CI 14.1-14.6)), abdominal pain (from 0.3 (95% CI 0.1-0.5) to 9.9 (95% CI 9.7-10.1)), gastrointestinal issues (from 0.6 (95% CI 0.4-0.9) to 13.3 (95% CI 13.1-13.6)), cough (from 0.3 (95% CI 0.2-0.5) to 9.1 (95% CI 8.9-9.3)), and anxiety (from 0.8 (95% CI 0.6-1.2) to 11.4 (95% CI 11.2-11.6)); whereas muscle spasms (from 0.01 (95% CI 0.008-0.2) to 1.7 (95% CI 1.6-1.8)), pins and needles (from 0.05 (95% CI 0.03-0.0.9) to 1.5 (95% CI 1.4-1.6)), memory issues (from 0.03 (95% CI 0.02-0.06) to 0.8 (95% CI 0.7-0.8)), cognitive dysfunction (from 0.007 (95% CI 0.004-0.01) to 0.6 (95% CI 0.4-0.8)), and altered smell and/or taste (from 0.04 (95% CI 0.03-0.04) to 0.7 (95% CI 0.6-0.8)) were least common. Incidence rates of any post-acute COVID-19 symptoms generally increased with age, with certain symptoms peaking in middle-aged adults (anxiety, depressive disorders, headache, altered smell and taste) and others in pre-school children (gastrointestinal issues and cough). Females had higher incidence rates for most symptoms. Based on the random-effects model, the infected cohort had a higher incidence of any post-acute COVID-19 symptom than the general population, with a meta-analytic incidence rate ratio (meta-IRR) of 1.4 (1-2). A similar pattern was seen for all individual symptoms. The highest meta-IRRs were depressive disorder, 2.6 (1.7-3.9); anxiety, 2.3 (1.4-3.8); allergy, 2.1 (1.7-2.8) and sleep disorders, 2.1 (1.5-2.6). The meta-IRR for altered smell and/or taste was 1.9 (1.3-2.8).

Interpretation

Post-acute COVID-19 symptoms, as listed by the WHO, were commonly observed following COVID-19 infection. However, even after standardising research methods, there was significant heterogeneity in the incidence rates from different healthcare settings and geographical locations. This is the first international study of the epidemiology of post-acute COVID-19 symptoms using the WHO-listed symptoms. Its findings contibute to understand the epidemiology of this condition from a multinational approach. Limitations of this study include the lack of consensus of the post-acute COVID-19 definition, as well as the difficulty to capture the impact on daily life of the post-acute COVID-19 symptoms in the available datasets.

Funding

This work has been funded by the European Health Data Evidence Network (EHDEN) through an Evidence Generation Fund Grant and by the National Institute for Health and Care Research (NIHR) Oxford Biomedical Research Centre (BRC).

Healthcare utilisation and associated costs for methadone versus buprenorphine recipients: Examination of interlinked primary and secondary care electronic health records in England

INTRODUCTION

More evidence for patterns of healthcare utilisation and associated costs among people receiving opioid agonist therapy (OAT) is needed. We investigated primary and secondary healthcare usage and costs among methadone and buprenorphine recipients in England.

METHODS

We conducted a cohort study using the Clinical Practice Research Datalink GOLD and Aurum databases of patients who were prescribed OAT between 1 January 2007 and 31 July 2019. The cohort was linked to Hospital Episode Statistics admitted patient care, outpatient and emergency department data, neighbourhood- and practice-level Index of Multiple Deprivation quintiles and mortality records. Negative binomial regression models were applied to estimate weighted rate ratios (wRR) of healthcare utilisation. Total and mean costs were calculated using Unit Costs of Health and Social Care and the National Healthcare Service Payment by Results National Tariffs.

RESULTS

Among 12,639 patients observed over 39,016 person-years, we found higher rate of hospital admissions (wRR 1.18; 1.08-1.28) among methadone compared with buprenorphine recipients. The commonest hospital discharge diagnoses among methadone patients were infectious diseases (19.2%), mental and behavioural disorders (17.0%) and drug-related poisoning (16.5%); the three commonest among buprenorphine patients were mental and behavioural diseases (21.5%), endocrine (13.8%) and genitourinary system diseases (13.1%). Methadone patients had similar mean costs compared with buprenorphine patients (cost difference: £539.01; 432.11-1006.69).

DISCUSSION AND CONCLUSIONS

Differences in healthcare utilisation frequency for methadone versus buprenorphine recipients were observed. The differences in associated costs were mainly driven by hospital admissions. These findings offer valuable insights for optimising care strategies and resource allocation for OAT recipients.

The association of type 2 diabetes-related characteristics with fracture risk at different sites

AIM

To determine the association of diabetes-related characteristics with fractures at different sites in individuals with type 2 diabetes (T2D).

MATERIALS AND METHODS

We conducted a cohort study using the Clinical Practice Research Datalink (CPRD) GOLD. Patients aged over 30 years with T2D were identified within the CPRD. Patients were followed from the start of diabetes treatment until the end of data collection, death, or the occurrence of a fracture. Cox proportional hazards models were used to estimate the hazard ratios for the association of the individual characteristics (diabetes duration, glycated haemoglobin [HbA1c] level, and microvascular complications) with fracture risk, adjusted for demographics, comorbidities and comedication.

RESULTS

A diabetes duration of >10 years was associated with an increased risk of any fracture and major osteoporotic fractures (MOFs), while a diabetes duration of >8 years was associated with an increased hip fracture risk, compared to a duration <2 years. An HbA1c level <6% was associated with an increased fracture risk compared to HbA1c values of 6% to <7%. The presence of one or two microvascular complications was associated with an increased risk of any fracture and MOFs and the presence of two microvascular complications was associated with an increased hip fracture risk, compared to no microvascular complications.

CONCLUSION

In conclusion, our study shows that a diabetes duration of 10 years or more, strict glycaemic control resulting in HbA1c levels below 6%, and/or the presence of at least one microvascular complication increased the risk of any fracture, hip fractures, MOFs, and humerus fractures, but not ankle, scapula or skull fractures.

Trajectories and management of vascular risk following the diagnosis of multiple sclerosis: A population-based matched cohort study between 1987 and 2018 in England

BACKGROUND

People with multiple sclerosis (PwMS) have an increased cardiovascular and cerebrovascular disease burden, but this could be mitigated by vascular risk factor management.

OBJECTIVES

We compared the trajectories of vascular risk factors, vascular comorbidities and clinical management in PwMS against the general population post-MS diagnosis while controlling for frailty.

METHODS

Retrospective longitudinal analysis using English data from the Clinical Practice Research Datalink between 1987 and 2018 comprising PwMS matched with up to six controls without MS by age, sex and general practice.

RESULTS

We compared 12,251 PwMS with 72,572 matched controls; 3.8% of PwMS had mild-moderate frailty, 1.2% more than matched controls. Compared to controls, PwMS had an elevated incidence of Type 2 diabetes (HR 1.18, 95% CI (1.04, 1.34)), and starting antihypertensive medications (HR 1.40, 95% CI (1.33, 1.47)). Among those with hypertension at baseline, blood pressure trajectories did not differ between PwMS and controls. PwMS had increased rates of meeting targets for hypertension management (HR 1.25, 95% CI (1.12, 1.41)).

CONCLUSION

The observation that PwMS with hypertension are more likely to meet treatment targets than matched controls is encouraging, but the elevated rates of vascular comorbidities suggest that tighter vascular management may be needed in this population.

The incidence of lung cancer amongst primary care chest radiograph referrals-an evaluation of national and local datasets within the United Kingdom

OBJECTIVES

To determine the incidence of lung cancer amongst primary care referrals for investigation with a chest radiograph (CXR).

METHODS

Retrospective evaluation of datasets from the national Clinical Practice Research Datalink (CPRD) and from a single large regional centre. Data were extracted for cohorts of consecutive adults aged over 40 years for whom a CXR had been performed between 2016 and 2018. Using cancer registry data, the incidence of lung cancer within a 2 years of the CXR referral and the variations with age, gender, and smoking status were evaluated.

RESULTS

A total of 291 294 CXR events were evaluated from the combined datasets. The incidence of lung cancer amongst primary care CXR referrals was 1.4% in CPRD with a consistent correlation with increasing age and smoking status. The incidence of lung cancer within two-years of the CXR varied between 0.03% (95%CI, 0.0-0.1) amongst never smokers aged 40-45 years to 4.8% (95%CI, 4.2-5.5) amongst current-smokers aged 70-75 years. The findings were similar for the single large centre data, although cancer incidence was higher.

CONCLUSIONS

A simple estimation and stratification of the risk of lung cancer amongst primary care referrals for investigation with a CXR is possible using age and smoking status.

ADVANCES IN KNOWLEDGE

This is the first estimate of the incidence of lung cancer amongst primary care CXR referrals and a demonstration of how the demographic information contained within a request could be used to optimize investigations and interpret test results.

Delirium risk and mortality in people with pre-existing severe mental illness: a retrospective cohort study using linked datasets in England

BACKGROUND

Delirium is a severe neuropsychiatric syndrome caused by physical illness, associated with high mortality. Understanding risk factors for delirium is key to targeting prevention and screening. Whether severe mental illness (SMI) predisposes people to delirium is not known. We aimed to establish whether pre-existing SMI diagnosis is associated with higher risk of delirium diagnosis and mortality following delirium diagnosis.

METHODS

A retrospective cohort and nested case-control study using linked primary and secondary healthcare databases from 2000-2017. We identified people diagnosed with SMI, matched to non-SMI comparators. We compared incidence of delirium diagnoses between people with SMI diagnoses and comparators, and between SMI subtypes; schizophrenia, bipolar disorder and 'other psychosis'. We compared 30-day mortality following a hospitalisation involving delirium between people with SMI diagnoses and comparators, and between SMI subtypes.

RESULTS

We identified 20 566 people with SMI diagnoses, matched to 71 374 comparators. Risk of delirium diagnosis was higher for all SMI subtypes, with a higher risk conferred by SMI in the under 65-year group, (aHR:7.65, 95% CI 5.45-10.7, ⩾65-year group: aHR:3.35, 95% CI 2.77-4.05). Compared to people without SMI, people with an SMI diagnosis overall had no difference in 30-day mortality following a hospitalisation involving delirium (OR:0.66, 95% CI 0.38-1.14).

CONCLUSIONS

We found an association between SMI and delirium diagnoses. People with SMI may be more vulnerable to delirium when in hospital than people without SMI. There are limitations to using electronic healthcare records and further prospective study is needed to confirm these findings.