Search

BackgroundLong COVID describes new or persistent symptoms at least 4 weeks after onset of acute COVID-19. Clinical codes to describe this phenomenon were recently created.AimTo describe the use of long-COVID codes, and variation of use by general practice, demographic variables, and over time.Design and settingPopulation-based cohort study in English primary care.MethodWorking on behalf of NHS England, OpenSAFELY data were used encompassing 96% of the English population between 1 February 2020 and 25 May 2021. The proportion of people with a recorded code for long COVID was measured overall and by demographic factors, electronic health record software system (EMIS or TPP), and week.ResultsLong COVID was recorded for 23 273 people. Coding was unevenly distributed among practices, with 26.7% of practices having never used the codes. Regional variation ranged between 20.3 per 100 000 people for East of England (95% confidence interval [CI] = 19.3 to 21.4) and 55.6 per 100 000 people in London (95% CI = 54.1 to 57.1). Coding was higher among females (52.1, 95% CI = 51.3 to 52.9) than males (28.1, 95% CI = 27.5 to 28.7), and higher among practices using EMIS (53.7, 95% CI = 52.9 to 54.4) than those using TPP (20.9, 95% CI = 20.3 to 21.4).ConclusionCurrent recording of long COVID in primary care is very low, and variable between practices. This may reflect patients not presenting; clinicians and patients holding different diagnostic thresholds; or challenges with the design and communication of diagnostic codes. Increased awareness of diagnostic codes is recommended to facilitate research and planning of services, and also surveys with qualitative work to better evaluate clinicians’ understanding of the diagnosis.

BackgroundSince 2014 English national guidance recommends ‘high-intensity’ statins, reducing low-density lipoprotein (LDL) cholesterol by ≥40%.AimTo describe trends and variation in low-/medium-intensity statin prescribing and assess the feasibility of rapid prescribing behaviour change.Design and settingA retrospective cohort study using OpenPrescribing data from all 8142 standard NHS general practices in England from August 2010 to March 2019.MethodStatins were categorised as high- or low-/medium-intensity using two different thresholds, and the proportion prescribed below these thresholds was calculated. The authors plotted trends and geographical variation, carried out mixed-effects logistic regression to identify practice characteristics associated with breaching of guidance, and used indicator saturation to identify sudden prescribing changes.ResultsThe proportion of statins prescribed below the recommended 40% LDL-lowering threshold has decreased gradually from 80% in 2011/2012 to 45% in 2019; the proportion below a pragmatic 37% threshold decreased from 30% to 18% in 2019. Guidance from 2014 had minimal impact on trends. Wide variation was found between practices (interdecile ranges 20% to 85% and 10% to 30% respectively in 2018). Regression identified no strong associations with breaching of guidance. Indicator saturation identified several practices exhibiting sudden changes towards greater guideline compliance.ConclusionBreaches of guidance on choice of statin remain common, with substantial variation between practices. Some have implemented rapid change, indicating the feasibility of rapid prescribing behaviour change. This article discusses the potential for a national strategic approach, using data and evidence to optimise care, including targeted education alongside audit and feedback to outliers through services such as OpenPrescribing.

Background On December 8th 2020, NHS England administered the first COVID-19 vaccination as part of an ambitious vaccination programme during a global health emergency. Aims To develop a framework for detailed near-real-time monitoring of COVID-19 vaccine roll-out; to describe trends and variation in coverage by geographic area, and between key clinical and demographic patient groups. Methods Working on behalf of NHS England we used routine clinical data from 23.4 million patients to conduct a retrospective cohort study of comprehensive electronic health record data in NHS England, using the OpenSAFELY-TPP platform which covers approximately 40% of the general population in England with weekly data updates. We developed algorithms to identify key demographic and clinical sub-groups within this population and generated descriptive statistics on proportion of eligible patients receiving the vaccine among key Joint Committee on Vaccination and Immunisation (JCVI) target groups. Results Between December 8th and January 13th 961,580 people out of 23.4m in our dataset received a COVID-19 vaccine. Of 1,160,062 patients aged 80 or over and not living in a care home (currently targeted by JCVI) 476,375 had been vaccinated in total (41.1%). We observed a substantial divergence in vaccination by ethnicity within this group (White 42.5% vaccinated, Black 20.5%) and across rankings of deprivation (least deprived 44.7%, most deprived 37.9%). Patients with pre-existing medical conditions were equally likely, or more likely, to have received a vaccine across most co-morbidity groups with two exceptions: severe mental illness (30.3% vaccinated) and learning disability (28.1%). We identify substantial variation in vaccination among the over-80s between Sustainability and Transformation Partnerships (STPs; Range 12%-74%); lower vaccination rates among ethnic minority and deprived groups was observed in most but not all STPs. In the 70-79 age cohort 74,108 people (3.6%) had been vaccinated. 378,921 vaccine recipients under 70 and not identifiably resident in a care home were presumed to be health or social care workers; 32,174 recipients were identified as older aged care home residents (33.2% coverage). Of all those vaccinated, 169,472 had received a second dose (17.6%). Conclusions The NHS in England has rapidly delivered mass vaccination. We were able to deploy a data monitoring framework across small clinical subgroups using linked patient-level NHS data on 23.4 million people with very short delays from vaccine administration to completed analysis. Targeted activity may be needed to address lower vaccination rates observed among certain key groups: ethnic minorities, people living in areas of higher deprivation, and those with severe mental illness or learning disabilities. However we note that this data is only from the first preliminary weeks of the vaccination programme. Variation in vaccination coverage between groups and regions will have many complex drivers, the figures presented in this manuscript require thoughtful interpretation to support a rapidly evolving NHS vaccination campaign; we are sharing local level data with national and regional NHS teams on request.

Background The COVID-19 pandemic has disrupted healthcare activity globally. The NHS in England stopped most non-urgent work by March 2020, but later recommended that services should be restored to near-normal levels before winter where possible. The authors are developing the OpenSAFELY NHS Service Restoration Observatory , using data to describe changes in service activity during COVID-19, and reviewing signals for action with commissioners, researchers and clinicians. Here we report phase one: generating, managing, and describing the data. Objective To describe the volume and variation of coded clinical activity in English primary care across 23.8 million patients’ records, taking respiratory disease and laboratory procedures as key examples. Methods Working on behalf of NHS England we developed an open source software framework for data management and analysis to describe trends and variation in clinical activity across primary care EHR data on 23.8 million patients; and conducted a population cohort-based study to describe activity using CTV3 coding hierarchy and keyword searches from January 2019-September 2020. Results Much activity recorded in general practice declined to some extent during the pandemic, but largely recovered by September 2020, with some exceptions. There was a large drop in coded activity for commonly used laboratory tests, with broad recovery to pre-pandemic levels by September. One exception was blood coagulation tests such as International Normalised Ratio (INR), with a smaller reduction (median tests per 1000 patients in 2020: February 8.0; April 6.2; September 7.0). The overall pattern of recording for respiratory symptoms was less affected, following an expected seasonal pattern and classified as “no change” from the previous year. Respiratory tract infections exhibited a sustained drop compared with pre-pandemic levels, not returning to pre-pandemic levels by September 2020. Various COVID-19 codes increased through the period. We observed a small decline associated with high level codes for long-term respiratory conditions such as chronic obstructive pulmonary disease (COPD) and asthma. Asthma annual reviews experienced a small drop but since recovered, while COPD annual reviews remain below baseline. Conclusions We successfully delivered an open source software framework to describe trends and variation in clinical activity across an unprecedented scale of primary care data. The COVD-19 pandemic led to a substantial change in healthcare activity. Most laboratory tests showed substantial reduction, largely recovering to near-normal levels by September 2020, with some important tests less affected. Records of respiratory infections decreased with the exception of codes related to COVID-19, whilst activity of other respiratory disease codes was mixed. We are expanding the NHS Service Restoration Observatory in collaboration with clinicians, commissioners and researchers and welcome feedback.

BackgroundAntidepressants are commonly prescribed. There are clear national guidelines in relation to treatment sequencing. This study examines trends and variation in antidepressant prescribing across English primary care.AimTo examine trends and variation in antidepressant prescribing in England, with a focus on: monoamine oxidase inhibitors (MAOIs); paroxetine; and dosulepin and trimipramine.Design & settingRetrospective longitudinal study using national and practice-level data on antidepressant items prescribed per year (1998–2018) and per month (2010–2019).MethodClass- and drug-specific proportions were calculated at national and practice levels. Descriptive statistics were generated, percentile charts and maps were plotted, and logistic regression analysis was conducted.ResultsAntidepressant prescriptions more than tripled between 1998 and 2018, from 377 items per 1000 population to 1266 per 1000. MAOI prescribing fell substantially, from 0.7% of all antidepressant items in 1998 to 0.1% in 2018. There was marked variation between practices in past year prescribing of paroxetine (median practice proportion [MPP] = 1.7%, interdecile range [IDR] = 2.6%) and dosulepin (MPP = 0.7%, IDR = 1.8%), but less for trimipramine (MPP = 0%, IDR = 0.2%).ConclusionRapid growth and substantial variation in antidepressant prescribing behaviour was found between practices. The causes could be explored using mixed-methods research. Interventions to reduce prescribing of specific antidepressants, such as dosulepin, could include review prompts, alerts at the time of prescribing, and clinician feedback through tools like OpenPrescribing.net.

BackgroundElectronic health record (EHR) systems are used by clinicians to record patients’ medical information, and support clinical activities such as prescribing. In England, healthcare professionals are advised to ‘prescribe generically’ because generic drugs are usually cheaper than branded alternatives, and have fixed reimbursement costs. ‘Ghost-branded generics’ are a new category of medicines savings, caused by prescribers specifying a manufacturer for a generic product, often resulting in a higher reimbursement price compared with the true generic.AimTo describe time trends and practice factors associated with excess medication costs from ghost-branded generic prescribing.Design and settingRetrospective cohort study of English GP prescribing data and EHR deployment data.MethodA retrospective cohort study was conducted, based on data from the OpenPrescribing.net database from May 2013 to May 2019. Total spending on ghost-branded generics across England was calculated, and excess spend on ghost-branded generics calculated as a percentage of all spending on generics for every CCG and general practice in England, for every month in the study period.ResultsThere were 31.8 million ghost-branded generic items and £9.5 million excess cost in 2018, compared with 7.45 million ghost-branded generic items and £1.3 million excess cost in 2014. Most excess costs were associated with one EHR, SystmOne, and it was identified that SystmOne offered ghost-branded generic options as the default. After informing the vendor, the authors monitored for subsequent change in costs, and report a rapid decrease in ghost-branded generic expenditure.ConclusionA design choice in a commonly used EHR has led to £9.5 million in avoidable excess prescribing costs for the NHS in 1 year. Notifying the vendor led to a change in user interface and a rapid, substantial spend reduction. This finding illustrates that EHR user interface design has a substantial impact on the quality, safety, and cost-effectiveness of clinical practice; this should be a priority for quantitative research.