A methodological review of randomised n-of-1 trials.

BACKGROUND: n-of-1 trials are a type of crossover trial designed to optimise the evaluation of health technologies in individual patients. This trial design may be considered for the evaluation of health technologies in rare conditions where fewer patients are available to take part in research. This review describes the characteristics of randomised n-of-1 trials conducted over the span of 12 years, including how the n-of-1 design has been employed to study both rare and non-rare conditions. METHODS: Databases and clinical trials registries were searched for articles including "n-of-1" in the title between 2011 and 2023. The reference lists of reviews identified by the searches were searched for any additional eligible articles. Randomised n-of-1 trials were selected for inclusion and data were extracted on a range of design, population, and analysis characteristics. Descriptive statistics were produced for all variables. RESULTS: We identified 74 studies meeting our eligibility criteria, 13 of which (17.6%) were conducted in rare conditions. They were conducted in a range of clinical areas with the most common being neurological conditions (n = 16, 21.6%). The median (Q1, Q3) number of participants randomised was 9 (4, 20) and 12 trials (16.2%) involved a single patient only. Forty-six (62.2%) trials evaluated pharmaceutical interventions and 49 (66.2%) trials were placebo controlled. Trials had a median (Q1, Q3) of six (4, 8) periods and 61 (82.4%) compared two health technologies. Fifty-seven (77.0%) trials incorporated blinding and 32 (43.2%) had a washout period. Forty-nine trials (66.2%) used patient-reported outcome measures (PROMs) to assess the primary outcome. Trials used a range of approaches to analysis and 48 (64.9%) combined data from multiple patients. The characteristics of the n-of-1 trials conducted in rare conditions were generally consistent with those in non-rare conditions. CONCLUSIONS: n-of-1 trials are still underused and the application of the n-of-1 design for the evaluation of health technologies for rare diseases has been particularly limited. We have summarised the characteristics of randomised n-of-1 trials in rare and non-rare conditions. We hope that it can inform researchers in the design of future n-of-1 studies. Further work is required to provide guidance on specific design considerations, implementation, and statistical analysis of these studies. TRIAL REGISTRATION: Not applicable.

Evaluation of the impact of letters to GP practices to promote asthma prescription uptake in school-age children during summer (TRAINS study): a pragmatic cluster-randomised controlled trial.

BACKGROUND: Asthma exacerbations peak in school-aged children after the return to school in September. Previous studies have shown a decline in collections of asthma prescriptions during August. The PLEASANT trial demonstrated that sending a reminder letter to parents increased prescription uptake; reduced unscheduled care, and was cost saving to the health service. We aimed to assess whether informing general practitioner (GP) practices about the PLEASANT trial and its results could lead to its implementation in routine practice. METHODS: The trial to assess implementation of new research in a primary care setting (TRAINS) was a pragmatic cluster-randomised (1:1) trial conducted in England involving GP practices contributing to the Clinical Practice Research Datalink (CPRD). The intervention was a letter informing the GP practice of the PLEASANT trial results with recommendations for implementation. GP practices in the control group continued with usual care without receiving any letters about PLEASANT trial. The intervention was distributed via CPRD by both mail and email in June 2021. The trial received both University of Sheffield Ethics approval and Independent Scientific Advisory Committee (ISAC) approval. The primary outcome was the proportion of children with asthma (aged 4-15 years) who had a prescription for a preventer between Aug 1 and Sept 30, 2021. This trial is registered with ClinicalTrials.gov, NCT05226091. FINDINGS: A total of 1326 GP practices, including 90 583 children with asthma, were included in the study. These practices were randomly allocated to the intervention group (664 practices, 44 708 children) or the control group (662 practices, 45 875 children). In assessing the impact of the intervention on the proportion of children collecting a preventer prescription, 15 716 (35·3%) of 44 708 children from the intervention group and 16 001 (35·1%) of 45 559 children from the control group picked up a prescription. There was no statistically significant difference observed (odds ratio [OR] 1·01, 95% CI 0·97-1·05), indicating that the intervention had no effect. INTERPRETATION: The study findings suggest that passive intervention of providing a letter to GPs did not achieve the intended outcomes. To bridge the gap between evidence and practice, alternative, more proactive strategies could be explored to address the identified issues. FUNDING: Jazan University.

A review of sample sizes for UK pilot and feasibility studies on the ISRCTN registry from 2013 to 2020.

BACKGROUND: Pilot and feasibility studies provide information to be used when planning a full trial. A sufficient sample size within the pilot/feasibility study is required so this information can be extracted with suitable precision. This work builds upon previous reviews of pilot and feasibility studies to evaluate whether the target sample size aligns with recent recommendations and whether these targets are being reached. METHODS: A review of the ISRCTN registry was completed using the keywords "pilot" and "feasibility". The inclusion criteria were UK-based randomised interventional trials that started between 2013 (end of the previous review) and 2020. Target sample size, actual sample size and key design characteristics were extracted. Descriptive statistics were used to present sample sizes overall and by key characteristics. RESULTS: In total, 761 studies were included in the review of which 448 (59%) were labelled feasibility studies, 244 (32%) pilot studies and 69 (9%) described as both pilot and feasibility studies. Over all included pilot and feasibility studies (n = 761), the median target sample size was 30 (IQR 20-50). This was consistent when split by those labelled as a pilot or feasibility study. Slightly larger sample sizes (median = 33, IQR 20-50) were shown for those labelled both pilot and feasibility (n = 69). Studies with a continuous outcome (n = 592) had a median target sample size of 30 (IQR 20-43) whereas, in line with recommendations, this was larger for those with binary outcomes (median = 50, IQR 25-81, n = 97). There was no descriptive difference in the target sample size based on funder type. In studies where the achieved sample size was available (n = 301), 173 (57%) did not reach their sample size target; however, the median difference between the target and actual sample sizes was small at just minus four participants (IQR -25-0). CONCLUSIONS: Target sample sizes for pilot and feasibility studies have remained constant since the last review in 2013. Most studies in the review satisfy the earlier and more lenient recommendations however do not satisfy the most recent largest recommendation. Additionally, most studies did not reach their target sample size meaning the information collected may not be sufficient to estimate the required parameters for future definitive randomised controlled trials.

Appropriate design and reporting of superiority, equivalence and non-inferiority clinical trials incorporating a benefit-risk assessment: the BRAINS study including expert workshop.

Background: Randomised controlled trials are designed to assess the superiority, equivalence or non-inferiority of a new health technology, but which trial design should be used is not always obvious in practice. In particular, when using equivalence or non-inferiority designs, multiple outcomes of interest may be important for the success of a trial, despite the fact that usually only a single primary outcome is used to design the trial. Benefit-risk methods are used in the regulatory clinical trial setting to assess multiple outcomes and consider the trade-off of the benefits against the risks, but are not regularly implemented in publicly funded trials. Objectives: The aim of the project is to aid the design of clinical trials with multiple outcomes of interest by defining when each trial design is appropriate to use and identifying when to use benefit-risk methods to assess outcome trade-offs (qualitatively or quantitatively) in a publicly funded trial setting. Methods: A range of methods was used to elicit expert opinion to answer the project objectives, including a web-based survey of relevant researchers, a rapid review of current literature and a 2-day consensus workshop of experts (in 2019). Results: We created a list of 19 factors to aid researchers in selecting the most appropriate trial design, containing the following overarching sections: population, intervention, comparator, outcomes, feasibility and perspectives. Six key reasons that indicate a benefit-risk method should be considered within a trial were identified: (1) when the success of the trial depends on more than one outcome; (2) when important outcomes within the trial are in competing directions (i.e. a health technology is better for one outcome, but worse for another); (3) to allow patient preferences to be included and directly influence trial results; (4) to provide transparency on subjective recommendations from a trial; (5) to provide consistency in the approach to presenting results from a trial; and (6) to synthesise multiple outcomes into a single metric. Further information was provided to support the use of benefit-risk methods in appropriate circumstances, including the following: methods identified from the review were collated into different groupings and described to aid the selection of a method; potential implementation of methods throughout the trial process were provided and discussed (with examples); and general considerations were described for those using benefit-risk methods. Finally, a checklist of five pieces of information that should be present when reporting benefit-risk methods was defined, with two additional items specifically for reporting the results. Conclusions: These recommendations will assist research teams in selecting which trial design to use and deciding whether or not a benefit-risk method could be included to ensure research questions are answered appropriately. Additional information is provided to support consistent use and clear reporting of benefit-risk methods in the future. The recommendations can also be used by funding committees to confirm that appropriate considerations of the trial design have been made. Limitations: This research was limited in scope and should be considered in conjunction with other trial design methodologies to assess appropriateness. In addition, further research is needed to provide concrete information about which benefit-risk methods are best to use in publicly funded trials, along with recommendations that are specific to each method. Study registration: The rapid review is registered as PROSPERO CRD42019144882. Funding: Funded by the Medical Research Council UK and the National Institute for Health and Care Research as part of the Medical Research Council-National Institute for Health and Care Research Methodology Research programme.

Randomised controlled trials are considered the best way to gather evidence about potential NHS treatments. They can be designed from different perspectives depending whether the aim is to show that a new treatment is better than, equal to or no worse than the current best available treatment. The selection of this design relates to the single most important outcome; however, often multiple outcomes can be affected by a treatment. For example, a new treatment may improve disease management but increase side effects. Patients want a treatment to work but not at the price of poor quality of life; therefore, a trade-off must be made, and the recommended treatment depends on this trade-off. Benefit­risk methods can assess the trade-off between multiple outcomes and can include patient preference. These methods could improve the way that decisions are made about treatments in the NHS, but there is currently limited research about the use of these methods in publicly funded trials. The aim of this report is to improve the design of clinical trials by helping researchers to select the most appropriate trial design and to decide when to include a benefit­risk method. The recommendations were created using the opinions of experts within the field and consisted of a survey, review of the literature and a workshop. The project created a list of 19 factors that can assist researchers to select the most appropriate trial design. Furthermore, six key areas were identified in which researchers may consider including a benefit­risk method within a trial. Finally, if a benefit­risk assessment is being used, a checklist of items has been created that identifies the information important to include in reports. This report is, however, limited in its applicability and further research should extend this work, as well as provide more detail on individual methods that are available.

CONSORT Harms 2022 statement, explanation, and elaboration: updated guideline for the reporting of harms in randomized trials.

Randomized controlled trials remain the reference standard for healthcare research on effects of interventions, and the need to report both benefits and harms is essential. The Consolidated Standards of Reporting Trials (the main CONSORT) statement includes one item on reporting harms (i.e., all important harms or unintended effects in each group). In 2004, the CONSORT group developed the CONSORT Harms extension; however, it has not been consistently applied and needs to be updated. Here, we describe CONSORT Harms 2022, which replaces the CONSORT Harms 2004 checklist, and shows how CONSORT Harms 2022 items could be incorporated into the main CONSORT checklist. Thirteen items from the main CONSORT were modified to improve harms reporting. Three new items were added. In this article, we describe CONSORT Harms 2022 and how it was integrated into the main CONSORT checklist and elaborate on each item relevant to complete reporting of harms in randomized controlled trials. Until future work from the CONSORT group produces an updated checklist, authors, journal reviewers, and editors of randomized controlled trials should use the integrated checklist presented in this paper.

A retrospective analysis of conditional power assumptions in clinical trials with continuous or binary endpoints.

BACKGROUND: Adaptive clinical trials may use conditional power (CP) to make decisions at interim analyses, requiring assumptions about the treatment effect for remaining patients. It is critical that these assumptions are understood by those using CP in decision-making, as well as timings of these decisions. METHODS: Data for 21 outcomes from 14 published clinical trials were made available for re-analysis. CP curves for accruing outcome information were calculated using and compared with a pre-specified objective criteria for original and transformed versions of the trial data using four future treatment effect assumptions: (i) observed current trend, (ii) hypothesised effect, (iii) 80% optimistic confidence limit, (iv) 90% optimistic confidence limit. RESULTS: The hypothesised effect assumption met objective criteria when the true effect was close to that planned, but not when smaller than planned. The opposite was seen using the current trend assumption. Optimistic confidence limit assumptions appeared to offer a compromise between the two, performing well against objective criteria when the end observed effect was as planned or smaller. CONCLUSION: The current trend assumption could be the preferable assumption when there is a wish to stop early for futility. Interim analyses could be undertaken as early as 30% of patients have data available. Optimistic confidence limit assumptions should be considered when using CP to make trial decisions, although later interim timings should be considered where logistically feasible.

Developing guidance for a risk-proportionate approach to blinding statisticians within clinical trials: a mixed methods study.

BACKGROUND: Existing guidelines recommend statisticians remain blinded to treatment allocation prior to the final analysis and that any interim analyses should be conducted by a separate team from the one undertaking the final analysis. However, there remains substantial variation in practice between UK Clinical Trials Units (CTUs) when it comes to blinding statisticians. Therefore, the aim of this study was to develop guidance to advise CTUs on a risk-proportionate approach to blinding statisticians within clinical trials. METHODS: This study employed a mixed methods approach involving three stages: (I) a quantitative study using a cohort of 200 studies (from a major UK funder published between 2016 and 2020) to assess the impact of blinding statisticians on the proportion of trials reporting a statistically significant finding for the primary outcome(s); (II) a qualitative study using focus groups to determine the perspectives of key stakeholders on the practice of blinding trial statisticians; and (III) combining the results of stages I and II, along with a stakeholder meeting, to develop guidance for UK CTUs. RESULTS: After screening abstracts, 179 trials were included for review. The results of the primary analysis showed no evidence that involvement of an unblinded trial statistician was associated with the likelihood of statistically significant findings being reported, odds ratio (OR) 1.02 (95% confidence interval (CI) 0.49 to 2.13). Six focus groups were conducted, with 37 participants. The triangulation between stages I and II resulted in developing 40 provisional statements. These were rated independently by the stakeholder group prior to the meeting. Ten statements reached agreement with no agreement on 30 statements. At the meeting, various factors were identified that could influence the decision of blinding the statistician, including timing, study design, types of intervention and practicalities. Guidance including 21 recommendations/considerations was developed alongside a Risk Assessment Tool to provide CTUs with a framework for assessing the risks associated with blinding/not blinding statisticians and for identifying appropriate mitigation strategies. CONCLUSIONS: This is the first study to develop a guidance document to enhance the understanding of blinding statisticians and to provide a framework for the decision-making process. The key finding was that the decision to blind statisticians should be based on the benefits and risks associated with a particular trial.

TRial to Assess Implementation of New research in a primary care Setting (TRAINS): study protocol for a pragmatic cluster randomised controlled trial of an educational intervention to promote asthma prescription uptake in general practitioner practices.

BACKGROUND: There is a marked increase in unscheduled care visits in school-aged children with asthma after returning to school in September. This is potentially associated with children not taking their asthma preventer medication during the school summer holidays. A cluster randomised controlled trial (PLEASANT) was undertaken with 1279 school-age children in 141 general practices (71 on intervention and 70 on control) in England and Wales. It found that a simple letter sent from the family doctor during the school holidays to a parent with a child with asthma, informing them of the importance of taking asthma preventer medication during the summer relatively increased prescriptions by 30% in August and reduced medical contacts in the period September to December. Also, it is estimated there was a cost-saving of £36.07 per patient over the year. We aim to conduct a randomised trial to assess if informing GP practices of an evidence-based intervention improves the implementation of that intervention. METHODS/DESIGN: The TRAINS study-TRial to Assess Implementation of New research in a primary care Setting-is a pragmatic cluster randomised implementation trial using routine data. A total of 1389 general practitioner (GP) practices in England will be included into the trial; 694 GP practices will be randomised to the intervention group and 695 control group of usual care. The Clinical Practice Research Datalink (CPRD) will send the intervention and obtain all data for the study, including prescription and primary care contacts data. The intervention will be sent in June 2021 by postal and email to the asthma lead and/or practice manager. The intervention is a letter to GPs informing them of the PLEASANT study findings with recommendations. It will come with an information leaflet about PLEASANT and a suggested reminder letter and SMS text template. DISCUSSION: The trial will assess if informing GP practices of the PLEASANT trial results will increase prescription uptake before the start of the school year. The hope is that the intervention will increase the implementation of PLEASANT work and then increase prescription uptake during the summer holiday prior to the start of school. TRIAL REGISTRATION: ClinicalTrials.gov ID: NCT05226091.

Optimal pharmacotherapy pathway in adults with diabetic peripheral neuropathic pain: the OPTION-DM RCT.

BACKGROUND: The mainstay of treatment for diabetic peripheral neuropathic pain is pharmacotherapy, but the current National Institute for Health and Care Excellence guideline is not based on robust evidence, as the treatments and their combinations have not been directly compared. OBJECTIVES: To determine the most clinically beneficial, cost-effective and tolerated treatment pathway for diabetic peripheral neuropathic pain. DESIGN: A randomised crossover trial with health economic analysis. SETTING: Twenty-one secondary care centres in the UK. PARTICIPANTS: Adults with diabetic peripheral neuropathic pain with a 7-day average self-rated pain score of ≥ 4 points (Numeric Rating Scale 0-10). INTERVENTIONS: Participants were randomised to three commonly used treatment pathways: (1) amitriptyline supplemented with pregabalin, (2) duloxetine supplemented with pregabalin and (3) pregabalin supplemented with amitriptyline. Participants and research teams were blinded to treatment allocation, using over-encapsulated capsules and matching placebos. Site pharmacists were unblinded. OUTCOMES: The primary outcome was the difference in 7-day average 24-hour Numeric Rating Scale score between pathways, measured during the final week of each pathway. Secondary end points included 7-day average daily Numeric Rating Scale pain score at week 6 between monotherapies, quality of life (Short Form questionnaire-36 items), Hospital Anxiety and Depression Scale score, the proportion of patients achieving 30% and 50% pain reduction, Brief Pain Inventory - Modified Short Form items scores, Insomnia Severity Index score, Neuropathic Pain Symptom Inventory score, tolerability (scale 0-10), Patient Global Impression of Change score at week 16 and patients' preferred treatment pathway at week 50. Adverse events and serious adverse events were recorded. A within-trial cost-utility analysis was carried out to compare treatment pathways using incremental costs per quality-adjusted life-years from an NHS and social care perspective. RESULTS: A total of 140 participants were randomised from 13 UK centres, 130 of whom were included in the analyses. Pain score at week 16 was similar between the arms, with a mean difference of -0.1 points (98.3% confidence interval -0.5 to 0.3 points) for duloxetine supplemented with pregabalin compared with amitriptyline supplemented with pregabalin, a mean difference of -0.1 points (98.3% confidence interval -0.5 to 0.3 points) for pregabalin supplemented with amitriptyline compared with amitriptyline supplemented with pregabalin and a mean difference of 0.0 points (98.3% confidence interval -0.4 to 0.4 points) for pregabalin supplemented with amitriptyline compared with duloxetine supplemented with pregabalin. Results for tolerability, discontinuation and quality of life were similar. The adverse events were predictable for each drug. Combination therapy (weeks 6-16) was associated with a further reduction in Numeric Rating Scale pain score (mean 1.0 points, 98.3% confidence interval 0.6 to 1.3 points) compared with those who remained on monotherapy (mean 0.2 points, 98.3% confidence interval -0.1 to 0.5 points). The pregabalin supplemented with amitriptyline pathway had the fewest monotherapy discontinuations due to treatment-emergent adverse events and was most commonly preferred (most commonly preferred by participants: amitriptyline supplemented with pregabalin, 24%; duloxetine supplemented with pregabalin, 33%; pregabalin supplemented with amitriptyline, 43%; p = 0.26). No single pathway was superior in cost-effectiveness. The incremental gains in quality-adjusted life-years were small for each pathway comparison [amitriptyline supplemented with pregabalin compared with duloxetine supplemented with pregabalin -0.002 (95% confidence interval -0.011 to 0.007) quality-adjusted life-years, amitriptyline supplemented with pregabalin compared with pregabalin supplemented with amitriptyline -0.006 (95% confidence interval -0.002 to 0.014) quality-adjusted life-years and duloxetine supplemented with pregabalin compared with pregabalin supplemented with amitriptyline 0.007 (95% confidence interval 0.0002 to 0.015) quality-adjusted life-years] and incremental costs over 16 weeks were similar [amitriptyline supplemented with pregabalin compared with duloxetine supplemented with pregabalin -£113 (95% confidence interval -£381 to £90), amitriptyline supplemented with pregabalin compared with pregabalin supplemented with amitriptyline £155 (95% confidence interval -£37 to £625) and duloxetine supplemented with pregabalin compared with pregabalin supplemented with amitriptyline £141 (95% confidence interval -£13 to £398)]. LIMITATIONS: Although there was no placebo arm, there is strong evidence for the use of each study medication from randomised placebo-controlled trials. The addition of a placebo arm would have increased the duration of this already long and demanding trial and it was not felt to be ethically justifiable. FUTURE WORK: Future research should explore (1) variations in diabetic peripheral neuropathic pain management at the practice level, (2) how OPTION-DM (Optimal Pathway for TreatIng neurOpathic paiN in Diabetes Mellitus) trial findings can be best implemented, (3) why some patients respond to a particular drug and others do not and (4) what options there are for further treatments for those patients on combination treatment with inadequate pain relief. CONCLUSIONS: The three treatment pathways appear to give comparable patient outcomes at similar costs, suggesting that the optimal treatment may depend on patients' preference in terms of side effects. TRIAL REGISTRATION: The trial is registered as ISRCTN17545443 and EudraCT 2016-003146-89. FUNDING: This project was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme, and will be published in full in Health Technology Assessment; Vol. 26, No. 39. See the NIHR Journals Library website for further project information.

The number of people with diabetes is growing rapidly in the UK and is predicted to rise to over 5 million by 2025. Diabetes causes nerve damage that can lead to severe painful symptoms in the feet, legs and hands. One-quarter of all people with diabetes experience these symptoms, known as 'painful diabetic neuropathy'. Current individual medications provide only partial benefit, and in only around half of patients. The individual drugs, and their combinations, have not been compared directly against each other to see which is best. We conducted a study to see which treatment pathway would be best for patients with painful diabetic neuropathy. The study included three treatment pathways using combinations of amitriptyline, duloxetine and pregabalin. Patients received all three treatment pathways (i.e. amitriptyline treatment for 6 weeks and pregabalin added if needed for a further 10 weeks, duloxetine treatment for 6 weeks and pregabalin added if needed for a further 10 weeks and pregabalin treatment for 6 weeks and amitriptyline added if needed for a further 10 weeks); however, the order of the treatment pathways was decided at random. We compared the level of pain that participants experienced in each treatment pathway to see which worked best. On average, people said that their pain was similar after each of the three treatments and their combinations. However, two treatments in combination helped some patients with additional pain relief if they only partially responded to one. People also reported improved quality of life and sleep with the treatments, but these were similar for all the treatments. In the health economic analysis, the value for money and quality of life were similar for each pathway, and this resulted in uncertainty in the cost-effectiveness conclusions, with no one pathway being more cost-effective than the others. The treatments had different side effects, however; pregabalin appeared to make more people feel dizzy, duloxetine made more people nauseous and amitriptyline resulted in more people having a dry mouth. The pregabalin supplemented by amitriptyline pathway had the smallest number of treatment discontinuations due to side effects and may be the safest for patients.

A descriptive study of samples sizes used in agreement studies published in the PubMed repository.

INTRODUCTION: A sample size justification is required for all studies and should give the minimum number of subjects to be recruited for the study to achieve its primary objective. The aim of this review is to describe sample sizes from agreement studies with continuous or categorical endpoints and different methods of assessing agreement, and to determine whether sample size justification was provided. METHODS: Data were gathered from the PubMed repository with a time interval of 28th September 2018 to 28th September 2020. The search returned 5257 studies of which 82 studies were eligible for final assessment after duplicates and ineligible studies were excluded. RESULTS: We observed a wide range of sample sizes. Forty-six studies (56%) used a continuous outcome measure, 28 (34%) used categorical and eight (10%) used both. Median sample sizes were 50 (IQR 25 to 100) for continuous endpoints and 119 (IQR 50 to 271) for categorical endpoints. Bland-Altman limits of agreement (median sample size 65; IQR 35 to 124) were the most common method of statistical analysis for continuous variables and Kappa coefficients for categorical variables (median sample size 71; IQR 50 to 233). Of the 82 studies assessed, only 27 (33%) gave justification for their sample size. CONCLUSIONS: Despite the importance of a sample size justification, we found that two-thirds of agreement studies did not provide one. We recommend that all agreement studies provide rationale for their sample size even if they do not include a formal sample size calculation.

An investigation of the constancy of effect in Cochrane systematic reviews in context with the assumptions for noninferiority trials.

When designing a noninferiority (NI) study one of the most important steps is to set the noninferiority (NI) limit. The NI limit is an acceptable loss of efficacy for a new investigative treatment compared to an active control treatment - often standard care. The limit should be a value so small that the loss efficacy is clinically zero. An approach to the setting of a noninferiority limit such that an effect over placebo can be shown through an indirect comparison to placebo-controlled trials where the active control treatment was compared to placebo. In this context, the setting of the NI limit depends on three assumptions: assay sensitivity, bias minimisation, and the constancy assumption. The last assumption of constancy assumes the effect of the active control over placebo is constant. This paper aims to assess the constancy assumption in placebo-controlled trials. METHODS: 236 Cochrane reviews of placebo-controlled trials published in 2015-2016 were collected and used to assess the relation between the placebo, active treatment, and the standardised treatment different (SMD) with the time (year of publication). RESULTS: The analysis showed that both the size of the study and the treatment effect were associated with year of publication. The three main variables that affect the estimate of any future trial are the estimate from the meta-analysis of previous trials prior to the trial, the year difference in the meta-analysis, and the year of the trial conduction. The regression analysis showed that an increase of one unit in the point estimate of the historical meta-analysis would lead to an increase in the predicted estimate of future trial on the SMD scale by 0.88. This result suggests the final trial results are 12% smaller than that from the meta-analysis of trials until that point. CONCLUSION: The result of this study indicates that assuming constancy of the treatment difference between the active control and placebo can be questioned. It is therefore important to consider the effect of time in estimating the treatment response if indirect comparisons are being used as the basis of a NI limit.

Blinding of study statisticians in clinical trials: a qualitative study in UK clinical trials units.

BACKGROUND: Blinding is an established approach in clinical trials which aims to minimise the risk of performance and detection bias. There is little empirical evidence to guide UK clinical trials units (CTUs) about the practice of blinding statisticians. Guidelines recommend that statisticians remain blinded to allocation prior to the final analysis. As these guidelines are not based on empirical evidence, this study undertook a qualitative investigation relating to when and how statisticians should be blinded in clinical trials. METHODS: Data were collected through online focus groups with various stakeholders who work in the delivery and oversight of clinical trials. Recordings of the focus groups were transcribed verbatim and thematic analysis was used to analyse the transcripts. RESULTS: Thirty-seven participants from 19 CTUs participated in one of six focus groups. Four main themes were identified, namely statistical models of work, factors affecting the decision to blind statisticians, benefits of blinding/not blinding statisticians and practicalities. Factors influencing the decision to blind the statistician included available resources, study design and types of intervention and outcomes and analysis. Although blinding of the statistician is perceived as a desirable mitigation against bias, there was uncertainty about the extent to which an unblinded statistician might impart bias. Instead, in most cases, the insight that the statistician offers was deemed more important to delivery of a trial than the risk of bias they may introduce if unblinded. Blinding of statisticians was only considered achievable with the appropriate resource and staffing, which were not always available. In many cases, a standard approach to blinding was therefore considered unrealistic and impractical; hence the need for a proportionate risk assessment approach identifying possible mitigations. CONCLUSIONS: There was wide variation in practice between UK CTUs regarding the blinding of trial statisticians. A risk assessment approach would enable CTUs to identify risks associated with unblinded statisticians conducting the final analysis and alternative mitigation strategies. The findings of this study will be used to design guidance and a tool to support this risk assessment process.

A practical guide to sample size calculations: Installation of the app SampSize.

In 2016 we published three articles in Pharmaceutical Statistics that gave a practical guide to sample size calculations. In each of the articles there were instructions on how to obtain the App SampSize. This short communication updates these instructions and highlights the updates and added functionality to the App.

Expected Value of Sample Information to Guide the Design of Group Sequential Clinical Trials.

INTRODUCTION: Adaptive designs allow changes to an ongoing trial based on prespecified early examinations of accrued data. Opportunities are potentially being missed to incorporate health economic considerations into the design of these studies. METHODS: We describe how to estimate the expected value of sample information for group sequential design adaptive trials. We operationalize this approach in a hypothetical case study using data from a pilot trial. We report the expected value of sample information and expected net benefit of sampling results for 5 design options for the future full-scale trial including the fixed-sample-size design and the group sequential design using either the Pocock stopping rule or the O'Brien-Fleming stopping rule with 2 or 5 analyses. We considered 2 scenarios relating to 1) using the cost-effectiveness model with a traditional approach to the health economic analysis and 2) adjusting the cost-effectiveness analysis to incorporate the bias-adjusted maximum likelihood estimates of trial outcomes to account for the bias that can be generated in adaptive trials. RESULTS: The case study demonstrated that the methods developed could be successfully applied in practice. The results showed that the O'Brien-Fleming stopping rule with 2 analyses was the most efficient design with the highest expected net benefit of sampling in the case study. CONCLUSIONS: Cost-effectiveness considerations are unavoidable in budget-constrained, publicly funded health care systems, and adaptive designs can provide an alternative to costly fixed-sample-size designs. We recommend that when planning a clinical trial, expected value of sample information methods be used to compare possible adaptive and nonadaptive trial designs, with appropriate adjustment, to help justify the choice of design characteristics and ensure the cost-effective use of research funding. HIGHLIGHTS: Opportunities are potentially being missed to incorporate health economic considerations into the design of adaptive clinical trials.Existing expected value of sample information analysis methods can be extended to compare possible group sequential and nonadaptive trial designs when planning a clinical trial.We recommend that adjusted analyses be presented to control for the potential impact of the adaptive designs and to maintain the accuracy of the calculations.This approach can help to justify the choice of design characteristics and ensure the cost-effective use of limited research funding.

Adjusting for bias in the mean for primary and secondary outcomes when trials are in sequence.

When designing a clinical trial, one key aspect of the design is the sample size calculation. The sample size calculation tends to rely on a target or expected difference. The expected difference can be based on the observed data from previous studies, which results in bias. It has been reported that large treatment effects observed in trials are often not replicated in subsequent trials. If these values are used to design subsequent studies, the sample sizes may be biased which results in an unethical study. Regression to the mean (RTM) is one explanation for this. If only health technologies which meet a particular continuation criterion (such as p<0.05 in the first study) are progressed to a second confirmatory trial, it is highly likely that the observed effect in the second trial will be lower than that observed in the first trial. It will be shown how when moving from one trial to the next, a truncated normal distribution is inherently imposed on the first study. This results in a lower observed effect size in the second trial. A simple adjustment method is proposed based on the mathematical properties of the truncated normal distribution. This adjustment method was confirmed using simulations in R and compared with other previous adjustments. The method can be applied to the observed effect in a trial, which is being used in the design of a second confirmatory trial, resulting in a more stable estimate for the 'true' treatment effect. The adjustment accounts for the bias in the primary and secondary endpoints in the first trial with the bias being affected by the power of that study. Tables of results have been provided to aid implementation, along with a worked example. In summary, there is a bias introduced when the point estimate from one trial is used to assist the design of a second trial. It is recommended that any observed point estimates be used with caution and the adjustment method developed in this article be implemented to significantly reduce this bias.

Characteristics of patients in platform C19, a COVID-19 research database combining primary care electronic health record and patient reported information.

BACKGROUND: Data to better understand and manage the COVID-19 pandemic is urgently needed. However, there are gaps in information stored within even the best routinely-collected electronic health records (EHR) including test results, remote consultations for suspected COVID-19, shielding, physical activity, mental health, and undiagnosed or untested COVID-19 patients. Observational and Pragmatic Research Institute (OPRI) Singapore and Optimum Patient Care (OPC) UK established Platform C19, a research database combining EHR data and bespoke patient questionnaire. We describe the demographics, clinical characteristics, patient behavior, and impact of the COVID-19 pandemic using data within Platform C19. METHODS: EHR data from Platform C19 were extracted from 14 practices across UK participating in the OPC COVID-19 Quality Improvement program on a continuous, monthly basis. Starting 7th August 2020, consenting patients aged 18-85 years were invited in waves to fill an online questionnaire. Descriptive statistics were summarized using all data available up to 22nd January 2021. FINDINGS: From 129,978 invitees, 31,033 responded. Respondents were predominantly female (59.6%), white (93.5%), and current or ex-smokers (52.6%). Testing for COVID-19 was received by 23.8% of respondents, of which 7.9% received positive results. COVID-19 symptoms lasted ≥4 weeks in 19.5% of COVID-19 positive respondents. Up to 39% respondents reported a negative impact on questions regarding their mental health. Most (67%-76%) respondents with asthma, Chronic Obstructive Pulmonary Disease (COPD), diabetes, heart, or kidney disease reported no change in the condition of their diseases. INTERPRETATION: Platform C19 will enable research on key questions relating to COVID-19 pandemic not possible using EHR data alone.

Sample sizes for cluster-randomised trials with continuous outcomes: Accounting for uncertainty in a single intra-cluster correlation estimate.

Sample size calculations for cluster-randomised trials require inclusion of an inflation factor taking into account the intra-cluster correlation coefficient. Often, estimates of the intra-cluster correlation coefficient are taken from pilot trials, which are known to have uncertainty about their estimation. Given that the value of the intra-cluster correlation coefficient has a considerable influence on the calculated sample size for a main trial, the uncertainty in the estimate can have a large impact on the ultimate sample size and consequently, the power of a main trial. As such, it is important to account for the uncertainty in the estimate of the intra-cluster correlation coefficient. While a commonly adopted approach is to utilise the upper confidence limit in the sample size calculation, this is a largely inefficient method which can result in overpowered main trials. In this paper, we present a method of estimating the sample size for a main cluster-randomised trial with a continuous outcome, using numerical methods to account for the uncertainty in the intra-cluster correlation coefficient estimate. Despite limitations with this initial study, the findings and recommendations in this paper can help to improve sample size estimations for cluster randomised controlled trials by accounting for uncertainty in the estimate of the intra-cluster correlation coefficient. We recommend this approach be applied to all trials where there is uncertainty in the intra-cluster correlation coefficient estimate, in conjunction with additional sources of information to guide the estimation of the intra-cluster correlation coefficient.

Risk Predictors and Symptom Features of Long COVID Within a Broad Primary Care Patient Population Including Both Tested and Untested Patients.

INTRODUCTION: Symptoms may persist after the initial phases of COVID-19 infection, a phenomenon termed long COVID. Current knowledge on long COVID has been mostly derived from test-confirmed and hospitalized COVID-19 patients. Data are required on the burden and predictors of long COVID in a broader patient group, which includes both tested and untested COVID-19 patients in primary care. METHODS: This is an observational study using data from Platform C19, a quality improvement program-derived research database linking primary care electronic health record data (EHR) with patient-reported questionnaire information. Participating general practices invited consenting patients aged 18-85 to complete an online questionnaire since 7th August 2020. COVID-19 self-diagnosis, clinician-diagnosis, testing, and the presence and duration of symptoms were assessed via the questionnaire. Patients were considered present with long COVID if they reported symptoms lasting ≥4 weeks. EHR and questionnaire data up till 22nd January 2021 were extracted for analysis. Multivariable regression analyses were conducted comparing demographics, clinical characteristics, and presence of symptoms between patients with long COVID and patients with shorter symptom duration. RESULTS: Long COVID was present in 310/3151 (9.8%) patients with self-diagnosed, clinician-diagnosed, or test-confirmed COVID-19. Only 106/310 (34.2%) long COVID patients had test-confirmed COVID-19. Risk predictors of long COVID were age ≥40 years (adjusted Odds Ratio [AdjOR]=1.49 [1.05-2.17]), female sex (adjOR=1.37 [1.02-1.85]), frailty (adjOR=2.39 [1.29-4.27]), visit to A&E (adjOR=4.28 [2.31-7.78]), and hospital admission for COVID-19 symptoms (adjOR=3.22 [1.77-5.79]). Aches and pain (adjOR=1.70 [1.21-2.39]), appetite loss (adjOR=3.15 [1.78-5.92]), confusion and disorientation (adjOR=2.17 [1.57-2.99]), diarrhea (adjOR=1.4 [1.03-1.89]), and persistent dry cough (adjOR=2.77 [1.94-3.98]) were symptom features statistically more common in long COVID. CONCLUSION: This study reports the factors and symptom features predicting long COVID in a broad primary care population, including both test-confirmed and the previously missed group of COVID-19 patients.

Utilising benefit-risk assessments within clinical trials-a protocol for the BRAINS project.

BACKGROUND: Depending on the treatment to be investigated, a clinical trial could be designed to assess objectives of superiority, equivalence or non-inferiority. The design of the study is affected by many different elements including the control treatment, the primary outcome and associated relationships. In some studies, there could be more than one outcome of interest. In these situations, benefit-risk methodologies could be used to assess the outcomes simultaneously and consider the trade-off between the benefits against the risks of a treatment. Benefit-risk is used within the regulatory industry but seldom included within publicly funded clinical trials within the UK. This project aims to gain an expert consensus on how to select the appropriate trial design (e.g. superiority) and when to consider including benefit-risk methods. METHODS: The project will consist of four work packages: 1. A web-based survey to elicit current experiences and opinions, 2. A rapid literature review to assess any current recommendations, 3. A two-day consensus workshop to gain agreement on the recommendations, and 4. Production of a guidance document. DISCUSSION: The aim of the project is to provide a guideline for clinical researchers, grant funding bodies and reviewers for grant bodies for how to select the most appropriate trial design and when it is appropriate to consider using benefit-risk methods. The focus of the guideline will be on publicly funded trials however, the vision is that the work will be applicable across research settings and we will connect with other organisations and committees as appropriate.