BackgroundBefore considering whether to use a multivariable (diagnostic or prognostic) prediction model, it is essential that its performance be evaluated in data that were not used to develop the model (referred to as external validation). We critically appraised the methodological conduct and reporting of external validation studies of multivariable prediction models.MethodsWe conducted a systematic review of articles describing some form of external validation of one or more multivariable prediction models indexed in PubMed core clinical journals published in 2010. Study data were extracted in duplicate on design, sample size, handling of missing data, reference to the original study developing the prediction models and predictive performance measures.Results11,826 articles were identified and 78 were included for full review, which described the evaluation of 120 prediction models. in participant data that were not used to develop the model. Thirty-three articles described both the development of a prediction model and an evaluation of its performance on a separate dataset, and 45 articles described only the evaluation of an existing published prediction model on another dataset. Fifty-seven percent of the prediction models were presented and evaluated as simplified scoring systems. Sixteen percent of articles failed to report the number of outcome events in the validation datasets. Fifty-four percent of studies made no explicit mention of missing data. Sixty-seven percent did not report evaluating model calibration whilst most studies evaluated model discrimination. It was often unclear whether the reported performance measures were for the full regression model or for the simplified models.ConclusionsThe vast majority of studies describing some form of external validation of a multivariable prediction model were poorly reported with key details frequently not presented. The validation studies were characterised by poor design, inappropriate handling and acknowledgement of missing data and one of the most key performance measures of prediction models i.e. calibration often omitted from the publication. It may therefore not be surprising that an overwhelming majority of developed prediction models are not used in practice, when there is a dearth of well-conducted and clearly reported (external validation) studies describing their performance on independent participant data.

H igh blood pressure is a major risk factor for global disease burden.1 Even modest reductions in blood pressure are important and would reduce the risk of associated morbidity and premature mortality. [2][3][4] In settings where health care and medicines are freely available, a substantial burden of cardiovascular disease may be attributable to suboptimal adherence to blood pressure-lowering treatments. 5 Missed appointments for collection of medicine and challenges with taking lifelong treatment are some of the major reasons for suboptimal Background-We assessed the effect of automated treatment adherence support delivered via mobile phone short message system (SMS) text messages on blood pressure. Methods and Results-In this pragmatic, single-blind, 3-arm, randomized trial (SMS-Text Adherence Support [StAR]) undertaken in South Africa, patients treated for high blood pressure were randomly allocated in a 1:1:1 ratio to information only, interactive SMS text messaging, or usual care. The primary outcome was change in systolic blood pressure at 12 months from baseline measured with a validated oscillometric device. All trial staff were masked to treatment allocation. Analyses were intention to treat. Between June 26, 2012, and November 23, 2012, 1372 participants were randomized to receive information-only SMS text messages (n=457), interactive SMS text messages (n=458), or usual care (n=457). Primary outcome data were available for 1256 participants (92%). At 12 months, the mean adjusted change in systolic blood pressure compared with usual care was −2.2 mm Hg (95% confidence interval, −4.4 to −0.04) with informationonly SMS and −1.6 mm Hg (95% confidence interval, −3.7 to 0.6) with interactive SMS. Odds ratios for the proportion of participants with a blood pressure <140/90 mm Hg were 1.42 (95% confidence interval, 1.03-1.95) for information-only messaging and 1.41 (95% confidence interval, 1.02-1.95) for interactive messaging compared with usual care. Conclusions-In this randomized trial of an automated adherence support program delivered by SMS text message in a general outpatient population of adults with high blood pressure, we found a small reduction in systolic blood pressure control compared with usual care at 12 months. There was no evidence that an interactive intervention increased this effect. Clinical Trial Registration-URL: http://www.clinicaltrials.gov.

Background The effectiveness of the COVID-19 vaccination programme depends on mass participation: the greater the number of people vaccinated, the less risk to the population. Concise, persuasive messaging is crucial, particularly given substantial levels of vaccine hesitancy in the UK. Our aim was to test which types of written information about COVID-19 vaccination, in addition to a statement of efficacy and safety, might increase vaccine acceptance. Methods For this single-blind, parallel-group, randomised controlled trial, we aimed to recruit 15 000 adults in the UK, who were quota sampled to be representative. Participants were randomly assigned equally across ten information conditions stratified by level of vaccine acceptance (willing, doubtful, or strongly hesitant). The control information condition comprised the safety and effectiveness statement taken from the UK National Health Service website; the remaining conditions addressed collective benefit, personal benefit, seriousness of the pandemic, and safety concerns. After online provision of vaccination information, participants completed the Oxford COVID-19 Vaccine Hesitancy Scale (outcome measure; score range 7–35) and the Oxford Vaccine Confidence and Complacency Scale (mediation measure). The primary outcome was willingness to be vaccinated. Participants were analysed in the groups they were allocated. p values were adjusted for multiple comparisons. The study was registered with ISRCTN, ISRCTN37254291. Findings From Jan 19 to Feb 5, 2021, 15 014 adults were recruited. Vaccine hesitancy had reduced from 26·9% the previous year to 16·9%, so recruitment was extended to Feb 18 to recruit 3841 additional vaccine-hesitant adults. 12 463 (66·1%) participants were classified as willing, 2932 (15·6%) as doubtful, and 3460 (18·4%) as strongly hesitant (ie, report that they will avoid being vaccinated for as long as possible or will never get vaccinated). Information conditions did not alter COVID-19 vaccine hesitancy in those willing or doubtful (adjusted p values >0·70). In those strongly hesitant, COVID-19 vaccine hesitancy was reduced, in comparison to the control condition, by personal benefit information (mean difference –1·49, 95% CI –2·16 to –0·82; adjusted p=0·0015), directly addressing safety concerns about speed of development (−0·91, –1·58 to –0·23; adjusted p=0·0261), and a combination of all information (−0·86, –1·53 to –0·18; adjusted p=0·0313). In those strongly hesitant, provision of personal benefit information reduced hesitancy to a greater extent than provision of information on the collective benefit of not personally getting ill (−0·97, 95% CI –1·64 to –0·30; adjusted p=0·0165) or the collective benefit of not transmitting the virus (−1·01, –1·68 to –0·35; adjusted p=0·0150). Ethnicity and gender were found to moderate information condition outcomes. Interpretation In the approximately 10% of the population who are strongly hesitant ...

Background A previous efficacy trial found benefit from inhaled budesonide for COVID-19 in patients not admitted to hospital, but effectiveness in high-risk individuals is unknown. We aimed to establish whether inhaled budesonide reduces time to recovery and COVID-19-related hospital admissions or deaths among people at high risk of complications in the community. Methods PRINCIPLE is a multicentre, open-label, multi-arm, randomised, controlled, adaptive platform trial done remotely from a central trial site and at primary care centres in the UK. Eligible participants were aged 65 years or older or 50 years or older with comorbidities, and unwell for up to 14 days with suspected COVID-19 but not admitted to hospital. Participants were randomly assigned to usual care, usual care plus inhaled budesonide (800 μg twice daily for 14 days), or usual care plus other interventions, and followed up for 28 days. Participants were aware of group assignment. The coprimary endpoints are time to first self-reported recovery and hospital admission or death related to COVID-19, within 28 days, analysed using Bayesian models. The primary analysis population included all eligible SARS-CoV-2-positive participants randomly assigned to budesonide, usual care, and other interventions, from the start of the platform trial until the budesonide group was closed. This trial is registered at the ISRCTN registry (ISRCTN86534580) and is ongoing. Findings The trial began enrolment on April 2, 2020, with randomisation to budesonide from Nov 27, 2020, until March 31, 2021, when the prespecified time to recovery superiority criterion was met. 4700 participants were randomly assigned to budesonide (n=1073), usual care alone (n=1988), or other treatments (n=1639). The primary analysis model includes 2530 SARS-CoV-2-positive participants, with 787 in the budesonide group, 1069 in the usual care group, and 974 receiving other treatments. There was a benefit in time to first self-reported recovery of an estimated 2·94 days (95% Bayesian credible interval [BCI] 1·19 to 5·12) in the budesonide group versus the usual care group (11·8 days [95% BCI 10·0 to 14·1] vs 14·7 days [12·3 to 18·0]; hazard ratio 1·21 [95% BCI 1·08 to 1·36]), with a probability of superiority greater than 0·999, meeting the prespecified superiority threshold of 0·99. For the hospital admission or death outcome, the estimated rate was 6·8% (95% BCI 4·1 to 10·2) in the budesonide group versus 8·8% (5·5 to 12·7) in the usual care group (estimated absolute difference 2·0% [95% BCI –0·2 to 4·5]; odds ratio 0·75 [95% BCI 0·55 to 1·03]), with a probability of superiority 0·963, below the prespecified superiority threshold of 0·975. Two participants in the budesonide group and four in the usual care group had serious adverse events (hospital admissions unrelated to COVID-19). Interpretation Inhaled budesonide improves time to recovery, with a chance of also r...

BackgroundIn the last decade several authors have reviewed the features of pilot and feasibility studies and advised on the issues that should be addressed within them. We extend this literature by examining published pilot/feasibility trials that incorporate random allocation, examining their stated objectives, results presented and conclusions drawn, and comparing drug and non-drug trials.MethodsA search of EMBASE and MEDLINE databases for 2000 to 2009 revealed 3652 papers that met our search criteria. A random sample of 50 was selected for detailed review.ResultsMost of the papers focused on efficacy: those reporting drug trials additionally addressed safety/toxicity; while those reporting non-drug trials additionally addressed methodological issues. In only 56% (95% confidence intervals 41% to 70%) were methodological issues discussed in substantial depth, 18% (95% confidence interval 9% to 30%) discussed future trials and only 12% (95% confidence interval 5% to 24%) of authors were actually conducting one.ConclusionsDespite recent advice on topics that can appropriately be described as pilot or feasibility studies the large majority of recently published papers where authors have described their trial as a pilot or addressing feasibility do not primarily address methodological issues preparatory to planning a subsequent study, and this is particularly so for papers reporting drug trials. Many journals remain willing to accept the pilot/feasibility designation for a trial, possibly as an indication of inconclusive results or lack of adequate sample size.