Scientific
publications

The ADO (age, dyspnoea, airflow obstruction) score predicts 3-year overall mortality among chronic obstructive pulmonary disease (COPD) patients. Information on the changes in COPD prognostic scores is sparse and it is unclear if the ADO score should be measured serially. We followed 4804 UK COPD patients with three or more ADO measurements from The Health Improvement Network (2005-2014) in a retrospective open cohort design. Patient's ADO scores were calculated once per year unless an obstruction or dyspnoea measurement was missing. Cox regression models assessed the independent role of serial ADO scores on mortality. The associations between baseline patient characteristics and long-term change in ADO scores were assessed using linear mixed effect models. Fewer than 7% of patients had worsened ( increased) by ≥1 point per year after a median follow-up of 4.4 years. There was strong evidence that patients with more rapid worsening in ADO scores had increased mortality (hazard ratio 2.00 (95% CI 1.59-2.52) per 1 point increase in ADO per year). More rapid ADO score worsening was seen among current smokers (rate difference 0.059 (95% CI 0.031-0.087); p=0.001) and ex-smokers (0.028 (95% CI 0.003-0.054); p=0.032) and patients with depression (0.038 (95% CI 0.005-0.071); p=0.022), while overweight (-0.0347 (95% CI -0.0544- -0.0150); p=0.001) and obese (-0.0412 (95% CI -0.0625- -0.0198); p<0.001) patients had a less rapid ADO score worsening. Serial assessment of the ADO score can identify patients with worsening disease and update their prognosis, especially for patients who smoke, are depressed or have lower body mass index.

Background Domestic abuse (DA) against women is a global public health problem. Although the possible health burden could be substantial, the associations between DA and subsequent cardiometabolic disease (cardiovascular disease, hypertension, and type 2 diabetes mellitus) and all-cause mortality are poorly understood. Methods and Results This retrospective cohort study consisted of UK-based primary care patients between January 1, 1995, to December 1, 2017. Overall, 18 547 women exposed to DA were matched to 72 231 unexposed women by age and lifestyle factors. The main outcomes, presented as adjusted incidence rate ratios (IRRs), were the risk of developing cardiovascular disease, hypertension, type 2 diabetes mellitus, and all-cause mortality. In total, 181 exposed women experienced a cardiovascular disease event compared with 644 of the unexposed control group, relating to an increased adjusted IRR of 1.31 (95% CI, 1.11-1.55; =0.001). There was also an increased risk of subsequent type 2 diabetes mellitus (adjusted IRR: 1.51; 95% CI, 1.30-1.76; <0.001) and all-cause mortality (adjusted IRR: 1.44; 95% CI, 1.24-1.67; <0.001) following exposure to DA. This observation was not seen with hypertension (adjusted IRR: 0.99; 95% CI, 0.88-1.12; =0.873). Conclusions There is an increased risk of subsequent cardiovascular disease, type 2 diabetes mellitus, and all-cause mortality in female survivors of DA. However, there is no association with the development of hypertension in this group, in keeping with previous literature. Considering the high prevalence of DA, clinicians should be made aware of the disproportionally increased risk and thus are encouraged to manage modifiable risk factors actively in this group.

Autism spectrum disorder (ASD) is a lifelong disorder. In the UK, risperidone is the only psychotropic medication approved for the management of the behavioural symptoms that may accompany autism. This is a population-based study aimed to provide an evaluation of the changing trend in the incidence and prevalence of ASD and to analyse the pattern of psychotropic medication prescribing in the UK. 20,194 patients with ASD were identified. The prevalence increased 3.3-fold from 0.109 per 100 persons in 2009 to 0.355 per 100 persons in 2016. Approximately one-third of the identified cohort was prescribed at least one psychotropic medication. Although the medications approved to manage the symptoms of ASD are limited, the prescribing of such medications is increasing.

AIMS: Identification, a priori, of those at high risk of progression from pre-diabetes to diabetes may enable targeted delivery of interventional programmes while avoiding the burden of prevention and treatment in those at low risk. We studied whether the use of a machine-learning model can improve the prediction of incident diabetes utilizing patient data from electronic medical records. METHODS: A machine-learning model predicting the progression from pre-diabetes to diabetes was developed using a gradient boosted trees model. The model was trained on data from The Health Improvement Network (THIN) database cohort, internally validated on THIN data not used for training, and externally validated on the Canadian AppleTree and the Israeli Maccabi Health Services (MHS) data sets. The model's predictive ability was compared with that of a logistic-regression model within each data set. RESULTS: A cohort of 852 454 individuals with pre-diabetes (glucose ≥ 100 mg/dL and/or HbA1c ≥ 5.7) was used for model training including 4.9 million time points using 900 features. The full model was eventually implemented using 69 variables, generated from 11 basic signals. The machine-learning model demonstrated superiority over the logistic-regression model, which was maintained at all sensitivity levels - comparing AUC [95% CI] between the models; in the THIN data set (0.865 [0.860,0.869] vs 0.778 [0.773,0.784] P < .05), the AppleTree data set (0.907 [0.896, 0.919] vs 0.880 [0.867, 0.894] P < .05) and the MHS data set (0.925 [0.923, 0.927] vs 0.876 [0.872, 0.879] P < .05). CONCLUSIONS: Machine-learning models preserve their performance across populations in diabetes prediction, and can be integrated into large clinical systems, leading to judicious selection of persons for interventional programmes.

CONTEXT: Mortality and infection-related hospital admissions are increased in patients with primary adrenal insufficiency (PAI). However, the risk of primary care-managed infections in patients with PAI is unknown. OBJECTIVE: To estimate infection risk in PAI due to Addison's disease (AD) and congenital adrenal hyperplasia (CAH) in a primary care setting. DESIGN: Retrospective cohort study using UK data collected from 1995 to 2018. MAIN OUTCOME MEASURES: Incidence of lower respiratory tract infections (LRTIs), urinary tract infections (UTIs), gastrointestinal infections (GIIs), and prescription counts of antimicrobials in adult PAI patients compared to unexposed controls. RESULTS: A diagnosis of PAI was established in 1580 AD patients (mean age 51.7 years) and 602 CAH patients (mean age 35.4 years). All AD patients and 42% of CAH patients were prescribed glucocorticoids, most frequently hydrocortisone in AD (82%) and prednisolone in CAH (50%). AD and CAH patients exposed to glucocorticoids, but not CAH patients without glucocorticoid treatment, had a significantly increased risk of LRTIs (adjusted incidence rate ratio AD 2.11 [95% confidence interval (CI) 1.64-2.69], CAH 3.23 [95% CI 1.21-8.61]), UTIs (AD 1.51 [95% CI 1.29-1.77], CAH 2.20 [95% CI 1.43-3.34]), and GIIs (AD 3.80 [95% CI 2.99-4.84], CAH 1.93 [95% CI 1.06-3.52]). This was mirrored by increased prescription of antibiotics (AD 1.73 [95% CI 1.69-1.77], CAH 1.77 [95% CI 1.66-1.89]) and antifungals (AD 1.89 [95% CI 1.74-2.05], CAH 1.91 [95% CI 1.50-2.43]). CONCLUSIONS: There is an increased risk of infections and antimicrobial use in PAI in the primary care setting at least partially linked to glucocorticoid treatment. Future studies will need to address whether more physiological glucocorticoid replacement modes could reduce this risk.

OBJECTIVE: Tramadol has been widely used among patients with osteoarthritis (OA); however, there is paucity of information on its cardiovascular risk. We aimed to examine the association of tramadol with risk of myocardial infarction (MI) among patients with OA. DESIGN: Among OA patients aged 50-90 years without history of MI, cancer, or opioid use disorder in The Health Improvement Network database in the United Kingdom (2000-2016), three sequential propensity-score matched cohort studies were assembled, i.e., (1) patients who initiated tramadol or naproxen (negative comparator); (2) patients who initiated tramadol or diclofenac (positive comparator); and (3) patients who initiated tramadol or codeine (a commonly used weak opioid). The outcome was incident MI over six-months. RESULTS: Among tramadol and naproxen initiators (n = 33,024 in each cohort), 77 (4.8/1000 person-years) and 46 (2.8/1000 person-years) incident MI occurred, respectively. The rate difference (RD) and hazard ratios (HR) for incident MI with tramadol initiation were 1.9 (95% confidence interval [CI] 0.6 to 2.3)/1000 person-years and 1.68 (95% CI 1.16 to 2.41) relative to naproxen initiation, respectively. Among tramadol and diclofenac initiators (n = 18,662 in each cohort), 58 (6.4/1000 person-years) and 47 (5.1/1000 person-years) incident MIs occurred, respectively. The corresponding RD and HR for incident MI were 1.2 (95%CI -2.1 to 14.1)/1000 person-years and 1.24 (95%CI 0.84 to 1.82), respectively. Among tramadol and codeine initiators (n = 42,722 in each cohort), 127 (6.1/1000 person-years) and 103 (5.0/1000 person-years) incident MI occurred, respectively, and the corresponding RD and HR were 1.1 (95%CI:-0.3 to 2.5)/1000 person-years and 1.23 (95%CI:0.95 to 1.60), respectively. CONCLUSIONS: In this population-based cohort of patients with OA, the six-month risk of MI among initiators of tramadol was higher than that of naproxen, but comparable to, if not lower than, those of diclofenac or codeine.

AIM: To evaluate the impact of the introduction and withdrawal of financial incentives on alcohol screening and brief advice delivery in English primary care. DESIGN: Interrupted time-series using data from The Health Improvement Network (THIN) database. Data were split into three periods: (1) before the introduction of financial incentives (1 January 2006-31 March 2008); (2) during the implementation of financial incentives (1 April 2008-31 March 2015); and (3) after the withdrawal of financial incentives (1 April 2015-31 December 2016). Segmented regression models were fitted, with slope and step change coefficients at both intervention points. SETTING: England. PARTICIPANTS: Newly registered patients (16+) in 500 primary care practices for 2006-16 (n = 4 278 723). MEASUREMENTS: The outcome measures were percentage of patients each month who: (1) were screened for alcohol use; (2) screened positive for higher-risk drinking; and (3) were reported as having received brief advice on alcohol consumption. FINDINGS: There was no significant change in the percentage of newly registered patients who were screened for alcohol use when financial incentives were introduced. However, the percentage fell (P < 0.001) immediately when incentives were withdrawn, and fell by a further 2.96 [95% confidence interval (CI) = 2.21-3.70] patients per 1000 each month thereafter. After the introduction of incentives, there was an immediate increase of 9.05 (95% CI = 3.87-14.23) per 1000 patients screening positive for higher-risk drinking, but no significant further change over time. Withdrawal of financial incentives was associated with an immediate fall in screen-positive rates of 29.96 (95% CI = 19.56-40.35) per 1000 patients, followed by a rise each month thereafter of 2.14 (95% CI = 1.51-2.77) per 1000. Screen-positive patients recorded as receiving alcohol brief advice increased by 20.15 (95% CI = 12.30-28.00) per 1000 following the introduction of financial incentives, and continued to increase by 0.39 (95% CI = 0.26-0.53) per 1000 monthly until withdrawal. At this point, delivery of brief advice fell by 18.33 (95% CI = 11.97-24.69) per 1000 patients and continued to fall by a further 0.70 (95% CI = 0.28-1.12) per 1000 per month. CONCLUSIONS: Removing a financial incentive for alcohol prevention in English primary care was associated with an immediate and sustained reduction in the rate of screening for alcohol use and brief advice provision. This contrasts with no, or limited, increase in screening and brief advice delivery rates following the introduction of the scheme.

Effective glycemic control is the cornerstone of successful type 2 diabetes management. However, many patients fail to reach glycemic control targets, and therapeutic inertia (failure to intensify therapy to address poor glycemic control in a timely manner) has been widely reported. The aim of the present study was to evaluate the economic burden associated with diabetes-related complications due to poor glycemic control for patients with type 2 diabetes in the UK. A validated long-term model of type 2 diabetes (IQVIA CORE Diabetes Model) was used to project cost outcomes for a UK population with type 2 diabetes, based on data from The Health Improvement Network primary care database, at different levels of glycemic control. Costs associated with diabetes-related complications were accounted in 2017 Pounds Sterling (GBP). Complication costs were estimated for populations achieving different glycated hemoglobin (HbA1c) targets, in a number of delayed treatment intensification scenarios, and across a range of time horizons. For patients with an HbA1c level of 8.2% (66 mmol/mol), 7 years in poor control could increase mean costs associated with diabetes-related complications by over GBP 690 per patient and lead to costs of over GBP 1,500 in lost workplace productivity compared with achieving good glycemic control (HbA1c 7.0%, 53 mmol/mol) over a 10-year time horizon. Based on published estimates of the proportion of type 2 diabetes patients failing to meet glycemic targets in the UK, this corresponds to an additional economic burden of ∼GBP 2,600 million (complication costs plus lost productivity costs). The economic burden of poor glycemic control in type 2 diabetes in the UK is substantial. Efforts to avoid therapeutic inertia could substantially reduce diabetes-related complication costs even in the short-term.

BACKGROUND: Antipsychotics are often prescribed for long-term periods, however, most evidence of their impact on body weight comes from short-term clinical trials. Particularly, impact associated with dosage has been barely studied. AIMS: The aim of this study was to describe the short- and long-term change in body weight of people initiated on high or low doses of the three most commonly prescribed second-generation antipsychotics. METHODS: Retrospective cohorts of individuals with a diagnosed psychotic disorder observed from 2005 to 2015 in the UK primary care. The exposure was the first prescription of olanzapine, quetiapine or risperidone. The main outcome was change in body weight four years before and four years after initiation of antipsychotic treatment, stratified on sex and 'low' or 'high' dose. RESULTS: In total, 22,306 women and 16,559 men were observed. Olanzapine treatment was associated with the highest change in weight, with higher doses resulting in more weight gain. After 4 years, given a high dose of olanzapine (> 5 mg), women gained on average +6.1 kg; whereas given a low dose (⩽ 5 mg), they gained +4.4 kg. During the first six weeks of olanzapine treatment, they gained on average +3.2 kg on high dose and +1.9 kg on low dose. The trends were similar for men. Individuals prescribed risperidone and quetiapine experienced less weight gain in both the short- and long-term. CONCLUSIONS: Olanzapine treatment was associated with the highest increase in weight. Higher doses were associated with more weight gain. Doctors should prescribe the lowest effective dose to balance mental-health benefits, weight gain and other adverse effects.

PURPOSE: To test hypotheses that minority ethnic people with dementia in the UK receive fewer anti-dementia drugs and more psychotropic and anticholinergic drugs associated with harms. PATIENTS AND METHODS: We analyzed UK primary care electronic health records from The Health Improvement Network (THIN) database (2014-2016), comparing psychotropic drug prescribing initiation and duration between people with dementia from White, Black, and Asian ethnic groups. We repeated analyses in people (aged 50+) without dementia, to explore whether any differences found reflected prescribing patterns in the general older population, or were specific to dementia. RESULTS: We included 53,718 people with and 1,648,889 people without dementia. Among people with dementia, compared to White ethnic groups, Asian people were less likely to be prescribed anti-dementia drugs when they were potentially indicated (adjusted prevalence rate ratio 0.86 (95% Confidence Interval 0.76-0.98)), and received them for on average 15 days/year less. Compared to White groups, Asian and Black individuals with dementia were no more likely to take an antipsychotic drug, but those that had were prescribed them for 17 and 27 days/year more, respectively (190.8 (179.6-199.1) and 200.7 (191.1-206.5) days). Black people were less likely to be prescribed anxiolytics/hypnotics (0.60 (0.44-0.8)), but the duration these drugs were prescribed was similar across ethnic groups. Asian people were more likely to be prescribed anticholinergic drugs (1.43 (1.19-1.73)), in analyses unadjusted for cardiovascular comorbidities. Among people without dementia, those in the Asian and Black ethnic groups were less likely to be prescribed psychotropic drugs, relative to people from White groups. CONCLUSION: Among people with dementia, Asian groups received less potentially beneficial symptomatic treatments, and Asian and Black groups were prescribed antipsychotic drugs for longer than White ethnic groups. Our findings may indicate care inequalities.

OBJECTIVES: We investigated combined hormonal contraceptives (CHC) prescribing patterns (focusing on combined oral contraceptives; COC) in three countries (Netherlands, Denmark, United Kingdom) in a time period preceding and in a time period following the European Commission's decision to update product information, and we estimated changes in incidence of venous thromboembolism (VTE) between the two periods. STUDY DESIGN: We conducted a drug utilization analysis and a cohort study using routinely collected data. We calculated number, proportion and incidence rate of new users, switchers, and stoppers of COC in both time periods. VTE incidence was calculated in new users of COC and in all women aged 18-49 years. RESULTS: In all countries, the largest proportion (> 75%) of new users used COC containing levonorgestrel, norethisterone, or norgestimate, (i.e., indicated by European Medicines Agency (EMA) as the safest preparations) in both time periods. Switching did not demonstrate a clear pattern towards these types of COC and distribution of stoppers was similar in both time periods. While the proportion of new users initiating COC containing levonorgestrel, norethisterone, or norgestimate increased slightly, this did not translate to a decrease in the overall VTE incidence. CONCLUSION: All three countries had the greatest proportion of women initiating a COC containing levonorgestrel, norethisterone, or norgestimate, and this proportion increased in the period after the European Commission decision albeit the increase was small due to the high percentage of use before the decision. This did not translate into a measureable change in the incidence of VTE. IMPLICATIONS: Both before and after the European Commission's decision, the largest proportion of new users started with combined oral contraceptives containing levonorgestrel, norethisterone, or norgestimate. Earlier studies had already indicated an increased risk of VTE associated with COC containing other progestogens compared with these preparations, so it is possible that physicians were already preferentially prescribing COC containing levonorgestrel, norethisterone, or norgestimate to new users.

Using primary care electronic medical records (the United Kingdom Health Improvement Network Database 2003-2015), we examined the control of cardiovascular risk factors in the first year after diagnosis in British adults with diabetes mellitus. Among 292 170 individuals with diabetes receiving frequent outpatient management (median of 16 primary care visits in the prior year), control of cardiovascular risk factors a median of 354 days after diagnosis was suboptimal: 14.7% had HbA1C < 7%, SBP < 140 mm Hg, LDL cholesterol ≤1.8 mmol/L or taking a statin, and were nonsmokers (the proportion dropped to 7.5% if the SBP target was defined as <130 mm Hg). While 90.4% had an LDL cholesterol ≤1.8 mmol/L or were taking a statin, and 86.0% were nonsmokers, only 52.0% had HbA1C < 7% and 53.1% had SBP < 140 mm Hg (29.8% had SBP < 130 mm Hg) despite 71.4% taking antihypertensive agents. Thus, there is still a need for quality improvement strategies that target all atherosclerotic risk factors in individuals with diabetes and not just glycaemic control.

BACKGROUND: Despite simpler regimens than vitamin K antagonists (VKAs) for stroke prevention in atrial fibrillation (AF), adherence (taking drugs as prescribed) and persistence (continuation of drugs) to direct oral anticoagulants are suboptimal, yet understudied in electronic health records (EHRs). OBJECTIVE: We investigated (1) time trends at individual and system levels, and (2) the risk factors for and associations between adherence and persistence. METHODS: In UK primary care EHR (The Health Information Network 2011-2016), we investigated adherence and persistence at 1 year for oral anticoagulants (OACs) in adults with incident AF. Baseline characteristics were analysed by OAC and adherence/persistence status. Risk factors for non-adherence and non-persistence were assessed using Cox and logistic regression. Patterns of adherence and persistence were analysed. RESULTS: Among 36 652 individuals with incident AF, cardiovascular comorbidities (median CHA DS VASc[Congestive heart failure, Hypertension, Age≥75 years, Diabetes mellitus, Stroke, Vascular disease, Age 65-74 years, Sex category] 3) and polypharmacy (median number of drugs 6) were common. Adherence was 55.2% (95% CI 54.6 to 55.7), 51.2% (95% CI 50.6 to 51.8), 66.5% (95% CI 63.7 to 69.2), 63.1% (95% CI 61.8 to 64.4) and 64.7% (95% CI 63.2 to 66.1) for all OACs, VKA, dabigatran, rivaroxaban and apixaban. One-year persistence was 65.9% (95% CI 65.4 to 66.5), 63.4% (95% CI 62.8 to 64.0), 61.4% (95% CI 58.3 to 64.2), 72.3% (95% CI 70.9 to 73.7) and 78.7% (95% CI 77.1 to 80.1) for all OACs, VKA, dabigatran, rivaroxaban and apixaban. Risk of non-adherence and non-persistence increased over time at individual and system levels. Increasing comorbidity was associated with reduced risk of non-adherence and non-persistence across all OACs. Overall rates of 'primary non-adherence' (stopping after first prescription), 'non-adherent non-persistence' and 'persistent adherence' were 3.5%, 26.5% and 40.2%, differing across OACs. CONCLUSIONS: Adherence and persistence to OACs are low at 1 year with heterogeneity across drugs and over time at individual and system levels. Better understanding of contributory factors will inform interventions to improve adherence and persistence across OACs in individuals and populations.