Inequalities in the management of diabetic kidney disease in UK primary care: A cross-sectional analysis of a large primary care database.

AIMS: To determine differences in the management of diabetic kidney disease (DKD) relevant to patient sex, ethnicity and socio-economic group in UK primary care. METHODS: A cross-sectional analysis as of January 1, 2019 was undertaken using the IQVIA Medical Research Data dataset, to determine the proportion of people with DKD managed in accordance with national guidelines, stratified by demographics. Robust Poisson regression models were used to calculate adjusted risk ratios (aRR) adjusting for age, sex, ethnicity and social deprivation. RESULTS: Of the 2.3 million participants, 161,278 had type 1 or 2 diabetes, of which 32,905 had DKD. Of people with DKD, 60% had albumin creatinine ratio (ACR) measured, 64% achieved blood pressure (BP, <140/90 mmHg) target, 58% achieved glycosylated haemoglobin (HbA1c, <58 mmol/mol) target, 68% prescribed renin-angiotensin-aldosterone system (RAAS) inhibitor in the previous year. Compared to men, women were less likely to have creatinine: aRR 0.99 (95% CI 0.98-0.99), ACR: aRR 0.94 (0.92-0.96), BP: aRR 0.98 (0.97-0.99), HbA1c : aRR 0.99 (0.98-0.99) and serum cholesterol: aRR 0.97 (0.96-0.98) measured; achieve BP: aRR 0.95 (0.94-0.98) or total cholesterol (<5 mmol/L) targets: aRR 0.86 (0.84-0.87); or be prescribed RAAS inhibitors: aRR 0.92 (0.90-0.94) or statins: aRR 0.94 (0.92-0.95). Compared to the least deprived areas, people from the most deprived areas were less likely to have BP measurements: aRR 0.98 (0.96-0.99); achieve BP: aRR 0.91 (0.8-0.95) or HbA1c : aRR 0.88 (0.85-0.92) targets, or be prescribed RAAS inhibitors: aRR 0.91 (0.87-0.95). Compared to people of white ethnicity; those of black ethnicity were less likely to be prescribed statins aRR 0.91 (0.85-0.97). CONCLUSIONS: There are unmet needs and inequalities in the management of DKD in the UK. Addressing these could reduce the increasing human and societal cost of managing DKD.

Patterns of initial and first-intensifying antidiabetic drug utilization among patients with type 2 diabetes mellitus in Scotland, 2010-2020: A retrospective population-based cohort study.

AIM: To evaluate the utilization and prescribing patterns of antidiabetic drugs (ADDs) for patients with type 2 diabetes mellitus (T2DM) at treatment initiation and first intensification. METHODS: A retrospective cohort study was performed using linked routinely collected data of patients with T2DM who received ADDs between January 2010 and December 2020 in Scotland. The prescribing patterns were quantified using frequency/percentages, absolute/relative change, and trend tests. RESULTS: Overall, 145 909 new ADD users were identified, with approximately 91% (N = 132 382) of patients receiving a single ADD at first treatment initiation. Metformin was the most often prescribed monotherapy (N = 118 737, 89.69%). A total of 50 731 patients (39.40%) who were started on metformin (N = 46 730/118 737, 39.36%) or sulphonylurea (SU; N = 4001/10 029, 39.89%) monotherapy had their treatment intensified with one or more additional ADD. Most initial-metformin (45 963/46 730; 98.36%) and initial-SU users (3894/4001; 97.33%) who added further drugs were intensified with single ADDs. SUs (22 197/45 963; 48.29%) were the most common first-intensifying monotherapy after initial metformin use, but these were replaced by sodium-glucose cotransporter-2 (SGLT2) inhibitors in 2019 (SGLT2 inhibitors: 2039/6065, 33.62% vs. SUs: 1924/6065, 31.72%). Metformin was the most frequently added monotherapy to initial SU use (2924/3894, 75.09%). Although the majority of patients received a single ADD, the use of combination therapy significantly increased over time. Nevertheless, there was a significant increasing trend towards prescribing the newer ADD classes (SGLT2 inhibitors, dipeptidyl peptidase-4 inhibitors) as monotherapy or in combination compared with the older ones (SUs, insulin, thiazolidinediones) at both drug initiation and first intensification. CONCLUSIONS: An overall increasing trend in prescribing the newer ADD classes compared to older ADDs was observed. However, metformin remained the most commonly prescribed first-line ADD, while SGLT2 inhibitors replaced SUs as the most common add-on therapy to initial metformin use in 2019.

Sodium-glucose cotransporter-2 inhibitors and the risk of gout in patients with type 2 diabetes mellitus: A propensity-score-matched, new-user design study with an active comparator using the IQVIA Medical Research Data UK database.

AIM: To conduct a pharmacoepidemiological study to explore the association between sodium-glucose cotransporter-2 (SGLT2) inhibitors and gout in patients with type 2 diabetes mellitus (T2DM). MATERIALS AND METHODS: A retrospective open cohort study using the IQVIA Medical Research Data UK database was performed between November 1, 2012 and December 31, 2018, estimating the risk of gout in patients with T2DM who were new users of SGLT2 inhibitors, compared to propensity-score-matched new users of dipeptidyl peptidase-4 (DPP-4) inhibitors. RESULTS: A total of 85 incident cases of gout were recorded over 30 389 person-years of observation in 13 617 new users of SGLT2 inhibitors and 29 426 new users of DPP-4 inhibitors. Crude incidence rates (IRs) per 1000 person-years were 2.90 and 2.47 for new users of SGLT2 inhibitors and DPP-4 inhibitors, respectively. The unadjusted hazard ratio (HR) was 1.18 (95% confidence interval [CI] 0.76-1.83). The adjusted HR was 1.20 (95% CI 0.77-1.86). In the at-treatment analysis, crude IRs per 1000 person-years were found to be 2.68 and 2.53 for SGLT2 inhibitor and DPP-4 inhibitor users, respectively. In the adjusted model, the adjusted HR was 1.3 (95% CI 0.90-2.29). Sensitivity analyses did not change the findings. CONCLUSIONS: In this nationwide study, no difference in the incidence of gout was documented in patients treated with SGLT2 inhibitors compared to DPP-4 inhibitor users. This neutral finding remained consistent in sensitivity analyses.

Sodium-glucose co-transporter-2 inhibitors and susceptibility to COVID-19: A population-based retrospective cohort study.

Sodium-glucose co-transporter-2 (SGLT2) inhibitors are widely prescribed in people with type 2 diabetes. We aimed to investigate whether SGLT2 inhibitor prescription is associated with COVID-19, when compared with an active comparator. We performed a propensity-score-matched cohort study with active comparators and a negative control outcome in a large UK-based primary care dataset. Participants prescribed SGLT2 inhibitors (n = 9948) and a comparator group prescribed dipeptidyl peptidase-4 (DPP-4) inhibitors (n = 14 917) were followed up from January 30 to July 27, 2020. The primary outcome was confirmed or clinically suspected COVID-19. The incidence rate of COVID-19 was 19.7/1000 person-years among users of SGLT2 inhibitors and 24.7/1000 person-years among propensity-score-matched users of DPP-4 inhibitors. The adjusted hazard ratio was 0.92 (95% confidence interval 0.66 to 1.29), and there was no evidence of residual confounding in the negative control analysis. We did not observe an increased risk of COVID-19 in primary care amongst those prescribed SGLT2 inhibitors compared to DPP-4 inhibitors, suggesting that clinicians may safely use these agents in the everyday care of people with type 2 diabetes during the COVID-19 pandemic.

Renin-angiotensin system inhibitors and susceptibility to COVID-19 in patients with hypertension: a propensity score-matched cohort study in primary care.

INTRODUCTION: Renin-angiotensin system (RAS) inhibitors have been postulated to influence susceptibility to Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2). This study investigated whether there is an association between their prescription and the incidence of COVID-19 and all-cause mortality. METHODS: We conducted a propensity-score matched cohort study comparing the incidence of COVID-19 among patients with hypertension prescribed angiotensin-converting enzyme I (ACE) inhibitors or angiotensin II type-1 receptor blockers (ARBs) to those treated with calcium channel blockers (CCBs) in a large UK-based primary care database (The Health Improvement Network). We estimated crude incidence rates for confirmed/suspected COVID-19 in each drug exposure group. We used Cox proportional hazards models to produce adjusted hazard ratios for COVID-19. We assessed all-cause mortality as a secondary outcome. RESULTS: The incidence rate of COVID-19 among users of ACE inhibitors and CCBs was 9.3 per 1000 person-years (83 of 18,895 users [0.44%]) and 9.5 per 1000 person-years (85 of 18,895 [0.45%]), respectively. The adjusted hazard ratio was 0.92 (95% CI 0.68 to 1.26). The incidence rate among users of ARBs was 15.8 per 1000 person-years (79 out of 10,623 users [0.74%]). The adjusted hazard ratio was 1.38 (95% CI 0.98 to 1.95). There were no significant associations between use of RAS inhibitors and all-cause mortality. CONCLUSION: Use of ACE inhibitors was not associated with the risk of COVID-19 whereas use of ARBs was associated with a statistically non-significant increase compared to the use of CCBs. However, no significant associations were observed between prescription of either ACE inhibitors or ARBs and all-cause mortality.

Visit-to-visit HbA1c variability and systolic blood pressure (SBP) variability are significantly and additively associated with mortality in individuals with type 1 diabetes: An observational study.

AIM: To investigate the relationship between variability in both visit-to-visit HbA1c and SBP and mortality in individuals with type 1 diabetes. METHODS: The Scottish Care Information (SCI) Diabetes dataset was used to identify 5952 individuals with type 1 diabetes for inclusion in this observational study. The SCI-Diabetes dataset allowed access to blood pressure values, HbA1c readings, demographic information and mortality rates for all study participants. Participants were dichotomized to above and below median values for both HbA1c coefficient of variation (CV) and SBP CV, thus dividing participants into 4 cohorts for survival analysis. Survival analysis was carried out over 1430 days. A Cox proportional hazard model was used to allow comparison of mortality between the 4 cohorts. RESULTS: Of the 5952 patients, death occurred in 416. CV for both HbA1c and SBP were significantly associated with mortality. The median values for HbA1c CV and SBP CV were 8.0 and 8.1, respectively. The hazard ratio for high HbA1c CV only (P = .0015) was 1.78 ± 0.36. The hazard ratio for high SBP CV only (P = .0018) was 1.69 ± 0.33. The hazard ratio for both high HbA1c CV and high SBP CV (P < .00001) was 2.37 ± 0.32. CONCLUSIONS: Our findings demonstrate that variability of both HbA1c and SBP is significantly and additively associated with mortality in individuals with type 1 diabetes. The variability of these parameters might be useful for risk stratification and is a potential target for future interventional studies.

Socioeconomic Deprivation and the Risk of Sight-Threatening Diabetic Retinopathy: A Population-Based Cohort Study in the U.K.

OBJECTIVE: To evaluate the associations between socioeconomic deprivation and sight-threatening diabetic retinopathy (STDR) in individuals with type 1 diabetes (T1D) and type 2 diabetes (T2D). RESEARCH DESIGN AND METHODS: Data from 175,628 individuals with diabetes in the Health Improvement Network were used to assess the risk of STDR across Townsend Deprivation Index quantiles using Cox proportional hazard regression. RESULTS: Among individuals with T1D, the risk of STDR was three times higher (adjusted hazard ratio [aHR] 2.67, 95% CI 1.05-7.78) in the most deprived quintile compared with the least deprived quintile. In T2D, the most deprived quintile had a 28% higher risk (aHR 1.28; 95% CI 1.15-1.43) than the least deprived quintile. CONCLUSIONS: Increasing socioeconomic deprivation is associated with a higher risk of developing STDR in people with diabetes. This underscores persistent health disparities linked to poverty, even within a country offering free universal health care. Further research is needed to address health equity concerns in socioeconomically deprived regions.

The optimal second-line therapy for older adults with type 2 diabetes mellitus: protocol for a systematic review and network meta-analysis using individual participant data (IPD).

BACKGROUND: Due to increasing life expectancy, almost half of people with type 2 diabetes are aged 65 years or over worldwide. When metformin alone does not control blood sugar, the choice of which second-line therapy to prescribe next is not clear from currently available evidence. The existence of frailty and comorbidities in older adults further increases the complexity of medical decision-making. As only a relatively small proportion of trials report results separately for older adults, the relative efficacy and safety of second-line therapies in older adults with type 2 diabetes mellitus are unknown and require further investigation. This individual participant data (IPD) network meta-analysis evaluates the relative efficacy and safety of second-line therapies on their own or in combination in older adults with type 2 diabetes mellitus. METHODS: All relevant published and unpublished trials will be identified. Studies published prior to 2015 will be identified from two previous comprehensive aggregate data network meta-analyses. Searches will be conducted in CENTRAL, MEDLINE, and EMBASE from 1st January 2015 onwards, and in clinicaltrials.gov from inception. Randomised controlled trials with at least 100 estimated older adults (≥ 65 years) receiving at least 24 weeks of intervention that assess the effects of glucose-lowering drugs on mortality, glycemia, vascular and other comorbidities outcomes, and quality of life will be eligible. The screening and data extraction process will be conducted independently by two researchers. The quality of studies will be assessed using the Cochrane risk of bias tool 2. Anonymised IPD of all eligible trials will be requested via clinical trial portals or by contacting the principal investigators or sponsors. Received data will be reanalysed where necessary to standardise outcome metrics. Network meta-analyses will be performed to determine the relative effectiveness of therapies. DISCUSSION: With the increasing number of older adults with type 2 diabetes worldwide, an IPD network meta-analysis using data from all eligible trials will provide new insights into the optimal choices of second-line antidiabetic drugs to improve patient management and reduce unnecessary adverse events and the subsequent risk of comorbidities in older adults. SYSTEMATIC REVIEW REGISTRATION: PROSPERO CRD42021272686.

Practical Guide to Glucocorticoid Induced Hyperglycaemia and Diabetes.

Glucocorticoids, also known as steroids, are a class of anti-inflammatory drugs utilised widely in clinical practice for a variety of conditions. They are associated with a range of side effects including abnormalities of glucose metabolism. Multiple guidelines have been published to illustrate best management of glucocorticoid-induced hyperglycaemia and diabetes in a variety of settings. This article discusses current best clinical practice including diagnosis, investigations and ongoing management of glucocorticoid-induced dysglycaemia in both in- and outpatient settings.

Associations of antidiabetic drugs with diabetic retinopathy in people with type 2 diabetes: an umbrella review and meta-analysis.

Background: Diabetic retinopathy (DR) is the most frequent complication of type 2 diabetes and remains the leading cause of preventable blindness. Current clinical decisions regarding the administration of antidiabetic drugs do not sufficiently incorporate the risk of DR due to the inconclusive evidence from preceding meta-analyses. This umbrella review aimed to systematically evaluate the effects of antidiabetic drugs on DR in people with type 2 diabetes. Methods: A systematic literature search was undertaken in Medline, Embase, and the Cochrane Library (from inception till 17th May 2022) without language restrictions to identify systematic reviews and meta-analyses of randomized controlled trials or longitudinal studies that examined the association between antidiabetic drugs and DR in people with type 2 diabetes. Two authors independently extracted data and assessed the quality of included studies using the AMSTAR-2 (A MeaSurement Tool to Assess Systematic Reviews) checklist, and evidence assessment was performed using the GRADE (Grading of recommendations, Assessment, Development and Evaluation). Random-effects models were applied to calculate relative risk (RR) or odds ratios (OR) with 95% confidence intervals (CI). This study was registered with PROSPERO (CRD42022332052). Results: With trial evidence from 11 systematic reviews and meta-analyses, we found that the use of glucagon-like peptide-1 receptor agonists (GLP-1 RA), sodium-glucose cotransporter-2 inhibitors (SGLT-2i), or dipeptidyl peptidase-4 inhibitors (DPP-4i) was not statistically associated with the risk of DR, compared to either placebo (RR: GLP-1 RA, 0.98, 0.89-1.08; SGLT-2i, 1.00, 95% CI 0.79-1.27; DPP-4i, 1.17, 0.99-1.39) or other antidiabetic drugs. Compared to other antidiabetic drugs, meglitinides (0.34, 0.01-8.25), SGLT-2i (0.73, 0.10-5.16), thiazolidinediones (0.92, 0.67-1.26), metformin (1.15, 0.81-1.63), sulphonylureas (1.24, 0.93-1.65), and acarbose (4.21, 0.44-40.43) were not statistically associated with the risk of DR. With evidence from longitudinal studies only, insulin was found to have a higher risk of DR than other antidiabetic drugs (OR: 2.47, 95% CI: 2.04-2.99). Conclusion: Our results indicate that antidiabetic drugs are generally safe to prescribe regarding the risk of DR among people with type 2 diabetes. Further robust and large-scale trials investigating the effects of insulin, meglitinides, and acarbose on DR are warranted. Systematic review registration: https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=332052, identifier CRD42022332052.

Machine learning of COVID-19 clinical data identifies population structures with therapeutic potential.

Clinical outcomes for patients with COVID-19 are heterogeneous and there is interest in defining subgroups for prognostic modeling and development of treatment algorithms. We obtained 28 demographic and laboratory variables in patients admitted to hospital with COVID-19. These comprised a training cohort (n = 6099) and two validation cohorts during the first and second waves of the pandemic (n = 996; n = 1011). Uniform manifold approximation and projection (UMAP) dimension reduction and Gaussian mixture model (GMM) analysis was used to define patient clusters. 29 clusters were defined in the training cohort and associated with markedly different mortality rates, which were predictive within confirmation datasets. Deconvolution of clinical features within clusters identified unexpected relationships between variables. Integration of large datasets using UMAP-assisted clustering can therefore identify patient subgroups with prognostic information and uncovers unexpected interactions between clinical variables. This application of machine learning represents a powerful approach for delineating disease pathogenesis and potential therapeutic interventions.

Diabetic Foot Risk Classification at the Time of Type 2 Diabetes Diagnosis and Subsequent Risk of Mortality: A Population-Based Cohort Study.

Aim: We aimed to compare the mortality of individuals at low, moderate, and high risk of diabetic foot disease (DFD) in the context of newly diagnosed type 2 diabetes, before developing active diabetic foot problem. Methods: This was a population-based cohort study of adults with newly diagnosed type 2 diabetes utilizing IQVIA Medical Research Data. The outcome was all-cause mortality among individuals with low, moderate, and high risk of DFD, and also in those with no record of foot assessment and those who declined foot examination. Results: Of 225,787 individuals with newly diagnosed type 2 diabetes, 34,061 (15.1%) died during the study period from January 1, 2000 to December 31, 2019. Moderate risk and high risk of DFD were associated with increased mortality risk compared to low risk of DFD (adjusted hazard ratio [aHR] 1.50, 95% CI 1.42, 1.58; aHR 2.01, 95% CI 1.84, 2.20, respectively). Individuals who declined foot examination or who had no record also had increased mortality risk of 75% and 25% vs. those at low risk of DFD, respectively (aHR 1.75, 95% CI 1.51, 2.04; aHR 1.25, 95% CI 1.20, 1.30). Conclusion: Individuals with new-onset type 2 diabetes who had moderate to high risk of DFD were more likely to die compared to those at low risk of DFD. The associations between declined foot examination and absence of foot examinations, and increased risk of mortality further highlight the importance of assessing foot risk as it identifies not only patients at risk of diabetic foot ulceration but also mortality.

Development and external validation of prognostic models for COVID-19 to support risk stratification in secondary care.

OBJECTIVES: Existing UK prognostic models for patients admitted to the hospital with COVID-19 are limited by reliance on comorbidities, which are under-recorded in secondary care, and lack of imaging data among the candidate predictors. Our aims were to develop and externally validate novel prognostic models for adverse outcomes (death and intensive therapy unit (ITU) admission) in UK secondary care and externally validate the existing 4C score. DESIGN: Candidate predictors included demographic variables, symptoms, physiological measures, imaging and laboratory tests. Final models used logistic regression with stepwise selection. SETTING: Model development was performed in data from University Hospitals Birmingham (UHB). External validation was performed in the CovidCollab dataset. PARTICIPANTS: Patients with COVID-19 admitted to UHB January-August 2020 were included. MAIN OUTCOME MEASURES: Death and ITU admission within 28 days of admission. RESULTS: 1040 patients with COVID-19 were included in the derivation cohort; 288 (28%) died and 183 (18%) were admitted to ITU within 28 days of admission. Area under the receiver operating characteristic curve (AUROC) for mortality was 0.791 (95% CI 0.761 to 0.822) in UHB and 0.767 (95% CI 0.754 to 0.780) in CovidCollab; AUROC for ITU admission was 0.906 (95% CI 0.883 to 0.929) in UHB and 0.811 (95% CI 0.795 to 0.828) in CovidCollab. Models showed good calibration. Addition of comorbidities to candidate predictors did not improve model performance. AUROC for the International Severe Acute Respiratory and Emerging Infection Consortium 4C score in the UHB dataset was 0.753 (95% CI 0.720 to 0.785). CONCLUSIONS: The novel prognostic models showed good discrimination and calibration in derivation and external validation datasets, and performed at least as well as the existing 4C score using only routinely collected patient information. The models can be integrated into electronic medical records systems to calculate each individual patient's probability of death or ITU admission at the time of hospital admission. Implementation of the models and clinical utility should be evaluated.

The clinical profile and associated mortality in people with and without diabetes with Coronavirus disease 2019 on admission to acute hospital services.

INTRODUCTION: To assess if in adults with COVID-19, whether those with diabetes and complications (DM+C) present with a more severe clinical profile and if that relates to increased mortality, compared to those with diabetes with no complications (DM-NC) and those without diabetes. METHODS: Service-level data was used from 996 adults with laboratory confirmed COVID-19 who presented to the Queen Elizabeth Hospital Birmingham, UK, from March to June 2020. All individuals were categorized into DM+C, DM-NC, and non-diabetes groups. Physiological and laboratory measurements in the first 5 days after admission were collated and compared among groups. Cox proportional hazards regression models were used to evaluate associations between diabetes status and the risk of mortality. RESULTS: Among the 996 individuals, 104 (10.4%) were DM+C, 295 (29.6%) DM-NC and 597 (59.9%) non-diabetes. There were 309 (31.0%) in-hospital deaths documented, 40 (4.0% of total cohort) were DM+C, 99 (9.9%) DM-NC and 170 (17.0%) non-diabetes. Individuals with DM+C were more likely to present with high anion gap/metabolic acidosis, features of renal impairment, and low albumin/lymphocyte count than those with DM-NC or those without diabetes. There was no significant difference in mortality rates among the groups: compared to individuals without diabetes, the adjusted HRs were 1.39 (95% CI 0.95-2.03, p = 0.093) and 1.18 (95% CI 0.90-1.54, p = 0.226) in DM+C and DM-C, respectively. CONCLUSIONS: Those with COVID-19 and DM+C presented with a more severe clinical and biochemical profile, but this did not associate with increased mortality in this study.

Predicting Clinical Events Based on Raw Text: From Bag-of-Words to Attention-Based Transformers.

Identifying which patients are at higher risks of dying or being re-admitted often happens to be resource- and life- saving, thus is a very important and challenging task for healthcare text analytics. While many successful approaches exist to predict such clinical events based on categorical and numerical variables, a large amount of health records exists in the format of raw text such as clinical notes or discharge summaries. However, the text-analytics models applied to free-form natural language found in those notes are lagging behind the break-throughs happening in the other domains and remain to be primarily based on older bag-of-words technologies. As a result, they rarely reach the accuracy level acceptable for the clinicians. In spite of their success in other domains, the superiority of deep neural approaches over classical bags of words for this task has not yet been convincingly demonstrated. Also, while some successful experiments have been reported, the most recent break-throughs due to the pre-trained language models have not yet made their ways into the medical domain. Using a publicly available healthcare dataset, we have explored several classification models to predict patients' re-admission or a fatality based on their discharge summaries and established that 1) The performance of the neural models used in our experiments convincingly exceeds those based on bag-of-words by several percentage points as measured by the standard metrics. 2) This allows us to achieve the accuracy typically acceptable by the clinicians as of practical use (area under the ROC curve above 0.70) for the majority of our prediction targets. 3) While the pre-trained attention-based transformer performed only on par with the model that averages word embeddings when applied to full length discharge summaries, the transformer still handles shorter text segments substantially better, at times with the margin of 0.04 in the area under the ROC curve. Thus, our findings extend the success of pre-trained language models reported in other domains to the task of clinical event prediction, and likely to other text-classification tasks in the healthcare analytics domain. 4) We suggest several models to overcome the transformers' major drawback (their input size limitation), and confirm that this is crucial to achieve their top performance. Our modifications are domain agnostic, and thus can be applied in other applications where the text inputs exceed 200 words. 5) We have successfully demonstrated how non-text attributes (such as patient age, demographics, type of admission etc.) can be combined with text to gain additional improvements for several prediction targets. We include extensive ablation studies showing the impact of the training size, and highlighting the tradeoffs between the performance and the resources needed.

Social Deprivation and Incident Diabetes-Related Foot Disease in Patients With Type 2 Diabetes: A Population-Based Cohort Study.

OBJECTIVE: To investigate the relationship between social deprivation and incident diabetes-related foot disease (DFD) in newly diagnosed patients with type 2 diabetes. RESEARCH DESIGN AND METHODS: A population-based open retrospective cohort study using The Health Improvement Network (1 January 2005 to 31 December 2019) was conducted. Patients with type 2 diabetes free of DFD at baseline were stratified by Townsend deprivation index, and risk of developing DFD was calculated. DFD was defined as a composite of foot ulcer (FU), Charcot arthropathy, lower-limb amputation (LLA), peripheral neuropathy (PN), peripheral vascular disease (PVD), and gangrene. RESULTS: A total of 176,359 patients were eligible (56% men; mean age 62.9 [SD 13.1] years). After excluding 26,094 patients with DFD before/within 15 months of type 2 diabetes diagnosis, DFD incidentally developed in 12.1% of the study population over 3.27 years (interquartile range 1.41-5.96). Patients in the most deprived Townsend quintile had increased risk of DFD compared with those in the least deprived (adjusted hazard ratio [aHR] 1.22; 95% CI 1.16-1.29) after adjusting for sex, age at type 2 diabetes diagnosis, ethnicity, smoking, BMI, HbA1c, cardiovascular disease, hypertension, retinopathy, estimated glomerular filtration rate, insulin, glucose/lipid-lowering medication, and baseline foot risk. Patients in the most deprived Townsend quintile had higher risk of PN (aHR 1.18; 95% CI 1.11-1.25), FU (aHR 1.44; 95% CI 1.17-1.77), PVD (aHR 1.40; 95% CI 1.28-1.53), LLA (aHR 1.75; 95% CI 1.08-2.83), and gangrene (aHR 8.49; 95% CI 1.01-71.58) compared with those in the least. CONCLUSIONS: Social deprivation is an independent risk factor for the development of DFD, PN, FU, PVD, LLA, and gangrene in newly diagnosed patients with type 2 diabetes. Considering the high individual and economic burdens of DFD, strategies targeting patients in socially deprived areas are needed to reduce health inequalities.

Increased COVID-19 infections in women with polycystic ovary syndrome: a population-based study.

OBJECTIVE: Several recent observational studies have linked metabolic comorbidities to an increased risk from COVID-19. Here we investigated whether women with PCOS are at an increased risk of COVID-19 infection. DESIGN: Population-based closed cohort study between 31 January 2020 and 22 July 2020 in the setting of a UK primary care database (The Health Improvement Network, THIN). METHODS: The main outcome was the incidence of COVID-19 coded as suspected or confirmed by the primary care provider. We used Cox proportional hazards regression model with stepwise inclusion of explanatory variables (age, BMI, impaired glucose regulation, androgen excess, anovulation, vitamin D deficiency, hypertension, and cardiovascular disease) to provide unadjusted and adjusted hazard risks (HR) of COVID-19 infection among women with PCOS compared to women without PCOS. RESULTS: We identified 21 292 women with a coded diagnosis of PCO/PCOS and randomly selected 78 310 aged and general practice matched control women. The crude COVID-19 incidence was 18.1 and 11.9 per 1000 person-years among women with and without PCOS, respectively. Age-adjusted Cox regression analysis suggested a 51% higher risk of COVID-19 among women with PCOS compared to women without PCOS (HR: 1.51 (95% CI: 1.27-1.80), P < 0.001). After adjusting for age and BMI, HR reduced to 1.36 (1.14-1.63)], P = 0.001. In the fully adjusted model, women with PCOS had a 28% increased risk of COVID-19 (aHR: 1.28 (1.05-1.56), P = 0.015). CONCLUSION: Women with PCOS are at an increased risk of COVID-19 infection and should be specifically encouraged to adhere to infection control measures during the COVID-19 pandemic. SIGNIFICANCE STATEMENT: Women with polycystic ovary syndrome (PCOS) have an increased risk of cardio-metabolic disease, which have been identified as a risk factor for COVID-19. To investigate whether the increased metabolic risk in PCOS translates into an increased risk of COVID-19 infection, we carried out a population-based closed cohort study in the UK during its first wave of the SARS-CoV-2 pandemic (January to July 2020), including 21 292 women with PCOS and 78 310 controls matched for sex, age and general practice location. Results revealed a 52% increased risk of COVID-19 infection in women with PCOS, which remained increased at 28% above controls after adjustment for age, BMI, impaired glucose regulation and other explanatory variables.

Nonsteroidal Antiinflammatory Drugs and Susceptibility to COVID-19.

OBJECTIVE: To identify whether active use of nonsteroidal antiinflammatory drugs (NSAIDs) increases susceptibility to developing suspected or confirmed coronavirus disease 2019 (COVID-19) compared to the use of other common analgesics. METHODS: We performed a propensity score-matched cohort study with active comparators, using a large UK primary care data set. The cohort consisted of adult patients age ≥18 years with osteoarthritis (OA) who were followed up from January 30 to July 31, 2020. Patients prescribed an NSAID (excluding topical preparations) were compared to those prescribed either co-codamol (paracetamol and codeine) or co-dydramol (paracetamol and dihydrocodeine). A total of 13,202 patients prescribed NSAIDs were identified, compared to 12,457 patients prescribed the comparator drugs. The primary outcome measure was the documentation of suspected or confirmed COVID-19, and the secondary outcome measure was all-cause mortality. RESULTS: During follow-up, the incidence rates of suspected/confirmed COVID-19 were 15.4 and 19.9 per 1,000 person-years in the NSAID-exposed group and comparator group, respectively. Adjusted hazard ratios for suspected or confirmed COVID-19 among the unmatched and propensity score-matched OA cohorts, using data from clinical consultations in primary care settings, were 0.82 (95% confidence interval [95% CI] 0.62-1.10) and 0.79 (95% CI 0.57-1.11), respectively, and adjusted hazard ratios for the risk of all-cause mortality were 0.97 (95% CI 0.75-1.27) and 0.85 (95% CI 0.61-1.20), respectively. There was no effect modification by age or sex. CONCLUSION: No increase in the risk of suspected or confirmed COVID-19 or mortality was observed among patients with OA in a primary care setting who were prescribed NSAIDs as compared to those who received comparator drugs. These results are reassuring and suggest that in the absence of acute illness, NSAIDs can be safely prescribed during the ongoing pandemic.

Association of Metformin with Susceptibility to COVID-19 in People with Type 2 Diabetes.

OBJECTIVE: Diabetes has emerged as an important risk factor for mortality from COVID-19. Metformin, the most commonly prescribed glucose-lowering agent, has been proposed to influence susceptibility to and outcomes of COVID-19 via multiple mechanisms. We investigated whether, in patients with diabetes, metformin is associated with susceptibility to COVID-19 and its outcomes. RESEARCH DESIGN AND METHODS: We performed a propensity score-matched cohort study with active comparators using a large UK primary care dataset. Adults with type 2 diabetes patients and a current prescription for metformin and other glucose-lowering agents (MF+) were compared to those with a current prescription for glucose-lowering agents that did not include metformin (MF-). Outcomes were confirmed COVID-19, suspected/confirmed COVID-19, and associated mortality. A negative control outcome analysis (back pain) was also performed. RESULTS: There were 29 558 and 10 271 patients in the MF+ and MF- groups, respectively, who met the inclusion criteria. In the propensity score-matched analysis, the adjusted hazard ratios for suspected/confirmed COVID-19, confirmed COVID-19, and COVID-19-related mortality were 0.85 (95% CI 0.67, 1.08), 0.80 (95% CI 0.49, 1.30), and 0.87 (95% CI 0.34, 2.20) respectively. The negative outcome control analysis did not suggest unobserved confounding. CONCLUSION: Current prescription of metformin was not associated with the risk of COVID-19 or COVID-19-related mortality. It is safe to continue prescribing metformin to improve glycemic control in patients with.

Obstructive Sleep Apnea, a Risk Factor for Cardiovascular and Microvascular Disease in Patients With Type 2 Diabetes: Findings From a Population-Based Cohort Study.

OBJECTIVE: To determine the risk of cardiovascular disease (CVD), microvascular complications, and mortality in patients with type 2 diabetes who subsequently develop obstructive sleep apnea (OSA) compared with patients with type 2 diabetes without a diagnosis of OSA. RESEARCH DESIGN AND METHODS: This age-, sex-, BMI-, and diabetes duration-matched cohort study used data from a U.K. primary care database from 1 January 2005 to 17 January 2018. Participants aged ≥16 years with type 2 diabetes were included. Exposed participants were those who developed OSA after their diabetes diagnosis; unexposed participants were those without diagnosed OSA. Outcomes were composite CVD (ischemic heart disease [IHD], stroke/transient ischemic attack [TIA], heart failure [HF]), peripheral vascular disease (PVD), atrial fibrillation (AF), peripheral neuropathy (PN), diabetes-related foot disease (DFD), referable retinopathy, chronic kidney disease (CKD), and all-cause mortality. The same outcomes were explored in patients with preexisting OSA before a diagnosis of type 2 diabetes versus diabetes without diagnosed OSA. RESULTS: A total of 3,667 exposed participants and 10,450 matched control participants were included. Adjusted hazard ratios for the outcomes were as follows: composite CVD 1.54 (95% CI 1.32, 1.79), IHD 1.55 (1.26, 1.90), HF 1.67 (1.35, 2.06), stroke/TIA 1.57 (1.27, 1.94), PVD 1.10 (0.91, 1.32), AF 1.53 (1.28, 1.83), PN 1.32 (1.14, 1.51), DFD 1.42 (1.16, 1.74), referable retinopathy 0.99 (0.82, 1.21), CKD (stage 3-5) 1.18 (1.02, 1.36), albuminuria 1.11 (1.01, 1.22), and all-cause mortality 1.24 (1.10, 1.40). In the prevalent OSA cohort, the results were similar, but some associations were not observed. CONCLUSIONS: Patients with type 2 diabetes who develop OSA are at increased risk of CVD, AF, PN, DFD, CKD, and all-cause mortality compared with patients without diagnosed OSA. Patients with type 2 diabetes who develop OSA are a high-risk population, and strategies to detect OSA and prevent cardiovascular and microvascular complications should be implemented.