Linda Nab

Electronic health records (EHRs) and other administrative health data are increasingly used in research to generate evidence on the effectiveness, safety, and utilisation of medical products and services, and to inform public health guidance and policy. Reproducibility is a fundamental step for research credibility and promotes trust in evidence generated from EHRs. At present, ensuring research using EHRs is reproducible can be challenging for researchers. Research software platforms can provide technical solutions to enhance the reproducibility of research conducted using EHRs. In response to the COVID‐19 pandemic, we developed the secure, transparent, analytic open‐source software platform OpenSAFELY designed with reproducible research in mind. OpenSAFELY mitigates common barriers to reproducible research by: standardising key workflows around data preparation; removing barriers to code‐sharing in secure analysis environments; enforcing public sharing of programming code and codelists; ensuring the same computational environment is used everywhere; integrating new and existing tools that encourage and enable the use of reproducible working practices; and providing an audit trail for all code that is run against the real data to increase transparency. This paper describes OpenSAFELY's reproducibility‐by‐design approach in detail.

Background The COVID-19 pandemic disrupted healthcare delivery, including difficulty accessing in-person care, which may have increased the need for strong pharmacological pain relief. Methods With NHS England approval, we used routine clinical data from >20 million general practice adult patients in OpenSAFELY-TPP. Using interrupted time series analysis, we quantified prevalent and new opioid prescribing prior to the COVID-19 pandemic (January 2018-February 2020), and during lockdown (March 2020-March 2021) and recovery periods (April 2021-June 2022), overall and stratified by demographics (age, sex, deprivation, ethnicity, geographic region) and to people in care homes. Outcomes The median number of people prescribed an opioid per month was 50.9 per 1000 patients prior to the pandemic. There was little change in prevalent prescribing during the pandemic, except for a temporary increase in March 2020. We observed a 9.8% (95%CI -14.5%, -6.5%) reduction in new opioid prescribing from March 2020, sustained to June 2022 for all demographic groups except people 80+ years. Among care home residents, in April 2020 new opioid prescribing increased by 112.5% (95%CI 92.2%, 134.9%) and parenteral opioid prescribing increased by 186.3% (95%CI 153.1%, 223.9%). Interpretation New opioid prescribing increased among older people and care home residents, likely reflecting use to treat end-of-life COVID-19 symptoms, but decreased among most other groups. Further research is needed to understand what is driving the reduction in new opioid prescribing and its relation to changes to health care provision during the pandemic. Funding The OpenSAFELY Platform is supported by grants from the Wellcome Trust (222097/Z/20/Z) and MRC (MR/V015737/1, MC_PC_20059, MR/W016729/1). In addition, development of OpenSAFELY has been funded by the Longitudinal Health and Wellbeing strand of the National Core Studies programme (MC_PC_20030: MC_PC_20059), the NIHR funded CONVALESCENCE programme (COV-LT-0009), NIHR (NIHR135559, COV-LT2-0073), and the Data and Connectivity National Core Study funded by UK Research and Innovation (MC_PC_20058) and Health Data Research UK (HDRUK2021.000). The views expressed are those of the authors and not necessarily those of the NIHR, NHS England, UK Health Security Agency (UKHSA) or the Department of Health and Social Care. Evidence before this study We searched Pubmed for publications between 1 March 2020 and 8 January 2023 using the following search terms: (“COVID-19” OR “SARS-CoV-2”) AND (“United Kingdom” OR “England” OR “Britain” OR “Scotland” OR “Wales”) AND (“opioid”). We also searched the reference list of relevant articles. We included research studies (excluding conference abstracts and editorials) that quantified opioid prescribing or use in the United Kingdom during the COVID-19 pandemic. Studies focussed solely on opioid substitution therapy for treatment of opioid use disorder were excluded. We identified four studies. One described opioid use among a cohort of people on a waiting list for hip or knee arthroplasty in Scotland (n=548) and found higher rates of long-term opioid use during the COVID-19 pandemic compared with historical controls. The second study quantified changes in opioid prescribing using English aggregate prescription data. This study found no changes in opioid prescribing after the start of the COVID-19 pandemic. The third study of 1.3 million people with rheumatic and musculoskeletal diseases found a decrease in new opioid users among people with certain conditions, but not in the number of overall prescriptions. The last study of 34,711 people newly diagnosed with cancer and 30,256 who died of cancer in Wales found increases in strong opioid prescribing in both populations. Added value of this study This is the largest study (>20 million patients) of opioid prescribing during the COVID-19 pandemic in a representative sample of the population of England. We used person-level data to quantify changes in the number of people prescribed opioids and identified that prevalent opioid prescribing changed little, with the exception of a temporary increase at the start of the first lockdown. However, we also identified meaningful reductions in new opioid prescribing. While our findings confirm previous studies quantifying variation in opioid prescribing by sex, ethnicity, region and deprivation, we showed that changes to new prescribing during the COVID-19 pandemic were experienced approximately similarly across these subgroups. The exceptions were older people and people in care homes. The latter group experienced substantial increases in new opioid prescribing (especially parenteral opioids, which are used in palliative care) coinciding with periods of greatest COVID-19 morbidity and mortality. Implications of all the available evidence The COVID-19 pandemic resulted in substantial disruptions to the healthcare system. Despite concerns that difficulty or delays in providing care during the pandemic may have led to shifts from non-pharmacological treatments to greater opioid prescribing, we observed no increases in prescribing prevalence in most demographic groups in England. The one major exception is people residing in care homes, where the observed prescribing patterns suggest use to treat end of life symptoms, consistent with best practice. However, our findings do not preclude increased prescribing in high risk subgroups, such as people on procedure waiting lists. Further research to quantify changes in this population is warranted.

Electronic health records (EHRs) and other administrative health data are increasingly used in research to generate evidence on the effectiveness, safety, and utilisation of medical products and services, and to inform public health guidance and policy. Reproducibility is a fundamental step for research credibility and promotes trust in evidence generated from EHRs. At present, ensuring research using EHRs is reproducible can be challenging for researchers. Research software platforms can provide technical solutions to enhance the reproducibility of research conducted using EHRs. In response to the COVID-19 pandemic, we developed the secure, transparent, analytic open-source software platform OpenSAFELY designed with reproducible research in mind. OpenSAFELY mitigates common barriers to reproducible research by: standardising key workflows around data preparation; removing barriers to code-sharing in secure analysis environments; enforcing public sharing of programming code and codelists; ensuring the same computational environment is used everywhere; integrating new and existing tools that encourage and enable the use of reproducible working practices; and providing an audit trail for all code that is run against the real data to increase transparency. This paper describes OpenSAFELY’s reproducibility-by-design approach in detail.

Background The effectiveness of COVID-19 monoclonal antibody and antiviral therapies against severe COVID-19 outcomes is unclear. Initial benefit was shown in unvaccinated patients and before the Omicron variant emerged. We used the OpenSAFELY platform to emulate target trials to estimate the effectiveness of sotrovimab or molnupiravir, versus no treatment. Methods With the approval of NHS England, we derived population-based cohorts of non-hospitalised high-risk individuals in England testing positive for SARS-CoV-2 during periods of dominance of the BA.1 (16/12/2021-10/02/2022) and BA.2 (11/02/2022-21/05/2022) Omicron sublineages. We used the clone-censor-weight approach to estimate the effect of treatment with sotrovimab or molnupiravir initiated within 5 days after positive test versus no treatment. Hazard ratios (HR) for COVID-19 hospitalisation or death within 28 days were estimated using weighted Cox models. Results Of the 35,856 [BA.1 period] and 39,192 [BA.2 period] patients, 1,830 [BA.1] and 1,242 [BA.2] were treated with molnupiravir and 2,244 [BA.1] and 4,164 [BA.2] with sotrovimab. The estimated HRs for molnupiravir versus untreated were 1.00 (95%CI: 0.81;1.22) [BA.1] and 1.22 (0.96;1.56) [BA.2]; corresponding HRs for sotrovimab versus untreated were 0.76 (0.66;0.89) [BA.1] and 0.92 (0.79;1.06) [BA.2]. Interpretation Compared with no treatment, sotrovimab was associated with reduced risk of adverse outcomes after COVID-19 in the BA.1 period, but there was weaker evidence of benefit in the BA2 period. Molnupiravir was not associated with reduced risk in either period. Funding UKRI, Wellcome Trust, MRC, NIHR and HDRUK.

Objectives To quantify in absolute and relative terms how population-level COVID-19 death rates have changed in demographic and clinical subgroups. Design Retrospective cohort study on behalf of NHS England. Setting Linked primary care and death registry data from the OpenSAFELY-TPP platform, covering the first three pandemic waves in England (wave 1: March 23 to May 30, 2020; wave 2: September 7, 2020 to April 24, 2021; and wave 3, delta: May 28 to December 14, 2021). Participants In total, 18.7, 18.8, and 18.7 million adults were included for waves 1, 2, and 3 respectively. Main outcome measures COVID-19-related mortality based on linked death registry records. Results The crude absolute COVID-19-related death rate per 1,000 person-years decreased from 4.48 in wave 1 (95%CI 4.41;4.55), to 2.70 in wave 2 (95%CI 2.67;2.73), to 0.64 in wave 3 (95%CI 0.63;0.66). The absolute death rate decreased by 90% between waves 1 and 3 in patients aged 80+, but by only 20% in patients aged 18-39. This higher proportional reduction in age- and sex-standardised death rates was also seen for other groups, such as neurological disease, learning disability and severe mental illness. Conversely, standardised death rates in transplant recipients stayed constant across successive waves at 10 per 1,000 person-years. There was also only a small decrease in death rates between waves in people with kidney disease, haematological malignancies or conditions associated with immunosuppression. Consequently, the relative hazard of COVID-19-related death decreased over time for some variables (e.g. age), remained similar for some (e.g. sex, ethnicity), and increased for others (e.g. transplant). Conclusions COVID-19 death rates decreased over the first three pandemic waves. An especially large decrease was seen in older age groups and people with neurological disease, learning disability or severe mental illness. Some demographic inequalities in death rates persisted over time. Groups more likely to experience impaired vaccine effectiveness did not see the same benefit in COVID-19 mortality reduction.

Background: Kidney disease is a significant risk factor for COVID-19-related mortality. Achieving high COVID-19 vaccine coverage among people with kidney disease is therefore a public health priority. Methods: With the approval of NHS England, we performed a retrospective cohort study using the OpenSAFELY-TPP platform. Individual-level routine clinical data from 24 million people in England were included. A cohort of individuals with stage 3–5 chronic kidney disease (CKD) or receiving renal replacement therapy (RRT) at the start of the COVID-19 vaccine roll-out was identified based on evidence of reduced estimated glomerular filtration rate or inclusion in the UK Renal Registry. Individual-level factors associated with vaccine uptake were explored via Cox proportional hazards models. Results: 948,845 people with stage 3–5 CKD or receiving RRT were included. Cumulative vaccine coverage as of 11 ^th May 2022 was 97.5%, 97.0%, and 93.5% for doses 1, 2, and 3, respectively, and 61.1% among individuals with one or more indications for receipt of a fourth dose. Delayed 3-dose vaccine uptake was associated with non-White ethnicity, social deprivation, and severe mental illness – associations that were consistent across CKD stages and in RRT recipients. Similar associations were observed for 4-dose uptake, which was also delayed among care home residents. Conclusion: Although high primary and booster dose coverage has been achieved among people with kidney disease in England, key disparities in vaccine uptake remain across demographic groups. Identifying how to address these disparities remains a priority to reduce the risk of severe disease in this vulnerable patient group.

Background: The UK COVID-19 vaccination programme delivered its first “booster” doses in September 2021, initially in groups at high risk of severe disease then across the adult population. The BNT162b2 Pfizer-BioNTech vaccine was used initially, with Moderna mRNA-1273 subsequently also used. Methods: We used the OpenSAFELY-TPP database, covering 40% of English primary care practices and linked to national coronavirus surveillance, hospital episodes, and death registry data, to estimate the effectiveness of boosting with BNT162b2 compared with no boosting in eligible adults who had received two primary course vaccine doses between 16 September and 16 December 2021 when the Delta variant of SARS-CoV-2 was dominant. Follow up was for up to 10 weeks. Each booster recipient was matched with an unboosted control on factors relating to booster priority status and prior immunisation. Additional factors were adjusted for in Cox models estimating hazard ratios (HRs). Outcomes were positive SARS-CoV-2 test, COVID-19 hospitalisation, COVID-19 death and non-COVID-9 death. Booster vaccine effectiveness was defined as 1−HR. Results: Among 4,352,417 BNT162b2 booster recipients matched with unboosted controls, estimated effectiveness of a booster dose compared with two doses only was 50.7% (95% CI 50.1-51.3) for positive SARS-CoV-2 test, 80.1% (78.3-81.8) for COVID-19 hospitalisation, 88.5% (85.0-91.1) for COVID-19 death, and 80.3% (79.0-81.5) for non-COVID-19 death. Estimated effectiveness was similar among those who had received a BNT162b2 or ChAdOx1-S two-dose primary vaccination course, but effectiveness against severe COVID-19 was slightly lower in those classified as clinically extremely vulnerable (76.3% (73.1-79.1) for COVID-19 hospitalisation, and 85.1% (79.6-89.1) for COVID-19 death). Estimated effectiveness against each outcome was lower in those aged 18-65 years than in those aged 65 and over. Conclusion: Our findings are consistent with strong protection of BNT162b2 boosting against positive SARS-CoV-2 test, COVID-19 hospitalisation, and COVID-19 death.