Comparative effectiveness of second line oral antidiabetic treatments among people with type 2 diabetes mellitus: emulation of a target trial using routinely collected health data.

OBJECTIVE: To compare the effectiveness of three commonly prescribed oral antidiabetic drugs added to metformin for people with type 2 diabetes mellitus requiring second line treatment in routine clinical practice. DESIGN: Cohort study emulating a comparative effectiveness trial (target trial). SETTING: Linked primary care, hospital, and death data in England, 2015-21. PARTICIPANTS: 75 739 adults with type 2 diabetes mellitus who initiated second line oral antidiabetic treatment with a sulfonylurea, DPP-4 inhibitor, or SGLT-2 inhibitor added to metformin. MAIN OUTCOME MEASURES: Primary outcome was absolute change in glycated haemoglobin A1c (HbA1c) between baseline and one year follow-up. Secondary outcomes were change in body mass index (BMI), systolic blood pressure, and estimated glomerular filtration rate (eGFR) at one year and two years, change in HbA1c at two years, and time to ≥40% decline in eGFR, major adverse kidney event, hospital admission for heart failure, major adverse cardiovascular event (MACE), and all cause mortality. Instrumental variable analysis was used to reduce the risk of confounding due to unobserved baseline measures. RESULTS: 75 739 people initiated second line oral antidiabetic treatment with sulfonylureas (n=25 693, 33.9%), DPP-4 inhibitors (n=34 464 ,45.5%), or SGLT-2 inhibitors (n=15 582, 20.6%). SGLT-2 inhibitors were more effective than DPP-4 inhibitors or sulfonylureas in reducing mean HbA1c values between baseline and one year. After the instrumental variable analysis, the mean differences in HbA1c change between baseline and one year were -2.5 mmol/mol (95% confidence interval (CI) -3.7 to -1.3) for SGLT-2 inhibitors versus sulfonylureas and -3.2 mmol/mol (-4.6 to -1.8) for SGLT-2 inhibitors versus DPP-4 inhibitors. SGLT-2 inhibitors were more effective than sulfonylureas or DPP-4 inhibitors in reducing BMI and systolic blood pressure. For some secondary endpoints, evidence for SGLT-2 inhibitors being more effective was lacking-the hazard ratio for MACE, for example, was 0.99 (95% CI 0.61 to 1.62) versus sulfonylureas and 0.91 (0.51 to 1.63) versus DPP-4 inhibitors. SGLT-2 inhibitors had reduced hazards of hospital admission for heart failure compared with DPP-4 inhibitors (0.32, 0.12 to 0.90) and sulfonylureas (0.46, 0.20 to 1.05). The hazard ratio for a ≥40% decline in eGFR indicated a protective effect versus sulfonylureas (0.42, 0.22 to 0.82), with high uncertainty in the estimated hazard ratio versus DPP-4 inhibitors (0.64, 0.29 to 1.43). CONCLUSIONS: This emulation study of a target trial found that SGLT-2 inhibitors were more effective than sulfonylureas or DPP-4 inhibitors in lowering mean HbA1c, BMI, and systolic blood pressure and in reducing the hazards of hospital admission for heart failure (v DPP-4 inhibitors) and kidney disease progression (v sulfonylureas), with no evidence of differences in other clinical endpoints.

The importance of investing in data, models, experiments, team science, and public trust to help policymakers prepare for the next pandemic.

The COVID-19 pandemic has brought about valuable insights regarding models, data, and experiments. In this narrative review, we summarised the existing literature on these three themes, exploring the challenges of providing forecasts, the requirement for real-time linkage of health-related datasets, and the role of 'experimentation' in evaluating interventions. This literature review encourages us to broaden our perspective for the future, acknowledging the significance of investing in models, data, and experimentation, but also to invest in areas that are conceptually more abstract: the value of 'team science', the need for public trust in science, and in establishing processes for using science in policy. Policy-makers rely on model forecasts early in a pandemic when there is little data, and it is vital to communicate the assumptions, limitations, and uncertainties (theme 1). Linked routine data can provide critical information, for example, in establishing risk factors for adverse outcomes but are often not available quickly enough to make a real-time impact. The interoperability of data resources internationally is required to facilitate sharing across jurisdictions (theme 2). Randomised controlled trials (RCTs) provided timely evidence on the efficacy and safety of vaccinations and pharmaceuticals but were largely conducted in higher income countries, restricting generalisability to low- and middle-income countries (LMIC). Trials for non-pharmaceutical interventions (NPIs) were almost non-existent which was a missed opportunity (theme 3). Building on these themes from the narrative review, we underscore the importance of three other areas that need investment for effective evidence-driven policy-making. The COVID-19 response relied on strong multidisciplinary research infrastructures, but funders and academic institutions need to do more to incentivise team science (4). To enhance public trust in the use of scientific evidence for policy, researchers and policy-makers must work together to clearly communicate uncertainties in current evidence and any need to change policy as evidence evolves (5). Timely policy decisions require an established two-way process between scientists and policy makers to make the best use of evidence (6). For effective preparedness against future pandemics, it is essential to establish models, data, and experiments as fundamental pillars, complemented by efforts in planning and investment towards team science, public trust, and evidence-based policy-making across international communities. The paper concludes with a 'call to actions' for both policy-makers and researchers.

Improving transportability of randomized controlled trial inference using robust prediction methods.

Randomized trials have been the gold standard for assessing causal effects since their introduction by Fisher in the 1920s, since they can eliminate both observed and unobserved confounding. Estimates of causal effects at the population level from randomized controlled trials can still be biased if there are both effect modification and systematic differences between the trial sample and the ultimate population of inference with respect to these modifiers. Recent advances in the survey statistics literature to improve inference in nonprobability samples by using information from probability samples can provide an avenue for improving population causal inference in randomized controlled trials when relevant probability samples of the patient population are available. We review some recent work in "transporting" causal effect estimates from trials to populations, focusing on the setting where there is a "benchmark" or population-representative sample along with the RCT sample. We then propose estimators using either inverse probability weighting (IPWT) or prediction that can accommodate unequal probability of selection in the "benchmark" or population, and use Bayesian additive regression trees for both inverse probability of treatment weighting and prediction estimation that do not require specification of functional form or interaction. We also consider how the assumption of ignorability may be assessed from observed data and propose a sensitivity analysis under the failure of this assumption. We compare our proposed approach with existing methods in simulation and apply these alternative approaches to a study of pulmonary artery catheterization in critically ill patients. We also suggest next steps for future work.

The Evolving Nature of Health Technology Assessment: A Critical Appraisal of NICE's New Methods Manual.

OBJECTIVES: The National Institute for Health and Care Excellence (NICE) recently completed a review of its methods for health technology assessment, involving a 2-stage public consultation. We appraise proposed methodological changes and analyze key decisions. METHODS: We categorize all changes proposed in the first consultation as "critical," "moderate" or "limited" updates, considering the importance of the topic and the degree of change or the level of reinforcement. Proposals were followed through the review process, for their inclusion, exclusion, or amendment in the second consultation and the new manual. RESULTS: The end-of-life value modifier was replaced with a new "disease severity" modifier and other potential modifiers were rejected. The usefulness of a comprehensive evidence base was emphasized, clarifying when nonrandomized studies can be used, with further guidance on "real-world" evidence developed separately. A greater degree of uncertainty was accepted in circumstances when evidence generation raised challenges, in particular for children, rare diseases, and innovative technologies. For some topics, such as health inequality, discounting, unrelated healthcare costs, and value of information, significant changes were possibly warranted, but NICE decided not to make any revisions at present. CONCLUSION: Most of the changes to NICE's health technology assessment methods are appropriate and modest in impact. Nevertheless, some decisions were not well justified and further research is needed on several topics, including investigation of societal preferences. Ultimately, NICE's role of protecting National Health Services resources for valuable interventions that can contribute toward improving overall population health must be safeguarded, without accepting weaker evidence.

How does a local instrumental variable method perform across settings with instruments of differing strengths? A simulation study and an evaluation of emergency surgery.

Local instrumental variable (LIV) approaches use continuous/multi-valued instrumental variables (IV) to generate consistent estimates of average treatment effects (ATEs) and Conditional Average Treatment Effects (CATEs). There is little evidence on how LIV approaches perform according to the strength of the IV or with different sample sizes. Our simulation study examined the performance of an LIV method, and a two-stage least squares (2SLS) approach across different sample sizes and IV strengths. We considered four 'heterogeneity' scenarios: homogeneity, overt heterogeneity (over measured covariates), essential heterogeneity (unmeasured), and overt and essential heterogeneity combined. In all scenarios, LIV reported estimates with low bias even with the smallest sample size, provided that the instrument was strong. Compared to 2SLS, LIV provided estimates for ATE and CATE with lower levels of bias and Root Mean Squared Error. With smaller sample sizes, both approaches required stronger IVs to ensure low bias. We considered both methods in evaluating emergency surgery (ES) for three acute gastrointestinal conditions. Whereas 2SLS found no differences in the effectiveness of ES according to subgroup, LIV reported that frailer patients had worse outcomes following ES. In settings with continuous IVs of moderate strength, LIV approaches are better suited than 2SLS to estimate policy-relevant treatment effect parameters.

Emulating Target Trials With Real-World Data to Inform Health Technology Assessment: Findings and Lessons From an Application to Emergency Surgery.

OBJECTIVES: International health technology assessment (HTA) agencies recommend that real-world data (RWD) are used in some circumstances to add to the evidence base about the effectiveness and cost-effectiveness of health interventions. The target trial framework applies the design principles of randomized-controlled trials to RWD and can help alleviate inevitable concerns about bias and design flaws with nonrandomized studies. This article aimed to tackle the lack of guidance and exemplar applications on how this methodology can be applied to RWD to inform HTA decision making. METHODS: We use Hospital Episode Statistics data from England on emergency hospital admissions from 2010 to 2019 to evaluate the cost-effectiveness of emergency surgery for 2 acute gastrointestinal conditions. We draw on the case study to describe the main challenges in applying the target trial framework alongside RWD and provide recommendations for how these can be addressed in practice. RESULTS: The 4 main challenges when applying the target trial framework to RWD are (1) defining the study population, (2) defining the treatment strategies, (3) establishing time zero (baseline), and (4) adjusting for unmeasured confounding. The recommendations for how to address these challenges, mainly around the incorporation of expert judgment and use of appropriate methods for handling unmeasured confounding, are illustrated within the case study. CONCLUSIONS: The recommendations outlined in this study could help future studies seeking to inform HTA decision processes. These recommendations can complement checklists for economic evaluations and design tools for estimating treatment effectiveness in nonrandomized studies.

Adapting Patient and Public Involvement processes in response to the Covid-19 pandemic.

BACKGROUND: The COVID-19 pandemic brought rapid and major changes to research, and those wishing to carry out Patient and Public Involvement (PPI) activities faced challenges, such as restrictions on movement and contact, illness, bereavement and risks to potential participants. Some researchers moved PPI to online settings during this time but remote consultations raise, as well as address, a number of challenges. It is important to learn from PPI undertaken in this period as face-to-face consultation may no longer be the dominant method for PPI. METHODS: UK stay-at-home measures announced in March 2020 necessitated immediate revisions to the intended face-to-face methods of PPI consultation for the ESORT Study, which evaluated emergency surgery for patients with common acute conditions. PPI plans and methods were modified to all components being online. We describe and reflect on: initial plans and adaptation; recruitment; training and preparation; implementation, contextualisation and interpretation. Through first-hand accounts we show how the PPI processes were developed, experienced and viewed by different partners in the process. DISCUSSION AND CONCLUSIONS: While concerns have been expressed about the possible limiting effects of forgoing face-to-face contact with PPI partners, we found important benefits from the altered dynamic of the online PPI environment. There were increased opportunities for participation which might encourage the involvement of a broader demographic, and unexpected benefits in that the online platform seemed to have a 'democratising' effect on the meetings, to the benefit of the PPI processes and outcomes. Other studies may however find that their particular research context raises particular challenges for the use of online methods, especially in relation to representation and inclusion, as new barriers to participation may be raised. It is important that methodological challenges are addressed, and researchers provide detailed examples of novel methods for discussion and empirical study. PATIENT AND PUBLIC CONTRIBUTION: We report a process which involved people with lived experience of emergency conditions and members of the public. A patient member was involved in the design and implementation, and two patients with lived experience contributed to the manuscript.

Effect of Angiotensin-Converting Enzyme Inhibitor and Angiotensin Receptor Blocker Initiation on Organ Support-Free Days in Patients Hospitalized With COVID-19: A Randomized Clinical Trial.

IMPORTANCE: Overactivation of the renin-angiotensin system (RAS) may contribute to poor clinical outcomes in patients with COVID-19. Objective: To determine whether angiotensin-converting enzyme (ACE) inhibitor or angiotensin receptor blocker (ARB) initiation improves outcomes in patients hospitalized for COVID-19. DESIGN, SETTING, AND PARTICIPANTS: In an ongoing, adaptive platform randomized clinical trial, 721 critically ill and 58 non-critically ill hospitalized adults were randomized to receive an RAS inhibitor or control between March 16, 2021, and February 25, 2022, at 69 sites in 7 countries (final follow-up on June 1, 2022). INTERVENTIONS: Patients were randomized to receive open-label initiation of an ACE inhibitor (n = 257), ARB (n = 248), ARB in combination with DMX-200 (a chemokine receptor-2 inhibitor; n = 10), or no RAS inhibitor (control; n = 264) for up to 10 days. MAIN OUTCOMES AND MEASURES: The primary outcome was organ support-free days, a composite of hospital survival and days alive without cardiovascular or respiratory organ support through 21 days. The primary analysis was a bayesian cumulative logistic model. Odds ratios (ORs) greater than 1 represent improved outcomes. RESULTS: On February 25, 2022, enrollment was discontinued due to safety concerns. Among 679 critically ill patients with available primary outcome data, the median age was 56 years and 239 participants (35.2%) were women. Median (IQR) organ support-free days among critically ill patients was 10 (-1 to 16) in the ACE inhibitor group (n = 231), 8 (-1 to 17) in the ARB group (n = 217), and 12 (0 to 17) in the control group (n = 231) (median adjusted odds ratios of 0.77 [95% bayesian credible interval, 0.58-1.06] for improvement for ACE inhibitor and 0.76 [95% credible interval, 0.56-1.05] for ARB compared with control). The posterior probabilities that ACE inhibitors and ARBs worsened organ support-free days compared with control were 94.9% and 95.4%, respectively. Hospital survival occurred in 166 of 231 critically ill participants (71.9%) in the ACE inhibitor group, 152 of 217 (70.0%) in the ARB group, and 182 of 231 (78.8%) in the control group (posterior probabilities that ACE inhibitor and ARB worsened hospital survival compared with control were 95.3% and 98.1%, respectively). CONCLUSIONS AND RELEVANCE: In this trial, among critically ill adults with COVID-19, initiation of an ACE inhibitor or ARB did not improve, and likely worsened, clinical outcomes. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT02735707.

A Comparison of Ordered Categorical versus Discrete Choices within a Stated Preference Survey of Whole-Blood Donors.

HIGHLIGHTS: This article compares the relative preferences from stated preference (SP) questions requiring ordered categorical versus discrete choice responses. The approaches were contrasted for blood donation service characteristics that offer opportunities to donate blood.The estimates of relative preferences for alternative blood donation service characteristics were similar between the 2 forms of SP approach.This study illustrates how SP survey questions can be formulated to provide responses on an ordered categorical scale and to estimate marginal rates of substitution between different attributes, which can be compared with those derived from discrete choice experiment (DCE) choices.The article highlights the potential value of considering alternative choice framings rather than relying solely on DCEs.

Ethnic and socioeconomic disparities in initiation of second-line antidiabetic treatment for people with type 2 diabetes in England: A cross-sectional study.

AIMS: To assess any disparities in the initiation of second-line antidiabetic treatments prescribed among people with type 2 diabetes mellitus (T2DM) in England according to ethnicity and social deprivation level. MATERIALS AND METHODS: This cross-sectional study used linked primary (Clinical Practice Research Datalink) and secondary care data (Hospital Episode Statistics), and the Index of Multiple Deprivation (IMD). We included people aged 18 years or older with T2DM who intensified to second-line oral antidiabetic medication between 2014 and 2020 to investigate disparities in second-line antidiabetic treatment prescribing (one of sulphonylureas [SUs], dipeptidyl peptidase-4 [DPP-4] inhibitors, or sodium-glucose cotransporter-2 [SGLT2] inhibitors, in combination with metformin) by ethnicity (White, South Asian, Black, mixed/other) and deprivation level (IMD quintiles). We report prescriptions of the alternative treatments by ethnicity and deprivation level according to predicted percentages derived from multivariable, multinomial logistic regression. RESULTS: Among 36 023 people, 85% were White, 10% South Asian, 4% Black and 1% mixed/other. After adjustment, the predicted percentages for SGLT2 inhibitor prescribing by ethnicity were 21% (95% confidence interval [CI] 19-23%), 20% (95% CI 18-22%), 19% (95% CI 16-22%) and 17% (95% CI 14-21%) among people with White, South Asian, Black, and mixed/other ethnicity, respectively. After adjustment, the predicted percentages for SGLT2 inhibitor prescribing by deprivation were 22% (95% CI 20-25%) and 19% (95% CI 17-21%) for the least deprived and the most deprived quintile, respectively. When stratifying by prevalent cardiovascular disease (CVD) status, we found lower predicted percentages of people with prevalent CVD prescribed SGLT2 inhibitors compared with people without prevalent CVD across all ethnicity groups and all levels of social deprivation. CONCLUSIONS: Among people with T2DM, there were no substantial differences by ethnicity or deprivation level in the percentage prescribed either SGLT2 inhibitors, DPP-4 inhibitors or SUs as second-line antidiabetic treatment.

Cost-effectiveness of alternative minimum recall intervals between whole blood donations.

BACKGROUND: The INTERVAL trial showed shorter inter-donation intervals could safely increase the frequency of whole-blood donation. We extended the INTERVAL trial to consider the relative cost-effectiveness of reduced inter-donation intervals. METHODS: Our within-trial cost-effectiveness analysis (CEA) used data from 44,863 whole-blood donors randomly assigned to 12, 10 or 8 week (males), and 16, 14 or 12 week inter-donation intervals (females). The CEA analysed the number of whole-blood donations, deferrals including low- haemoglobin deferrals, and donors' health-related quality of life (QoL) to report costs and cost-effectiveness over two years. FINDINGS: The mean number of blood donation visits over two years was higher for the reduced interval strategies, for males (7.76, 6.60 and 5.68 average donations in the 8-, 10- and 12- week arms) and for females (5.10, 4.60 and 4.01 donations in the 12-, 14- and 16- week arms). For males, the average rate of deferral for low haemoglobin per session attended, was 5.71% (8- week arm), 3.73% (10- week), and 2.55% (12- week), and for females the rates were: 7.92% (12-week), 6.63% (14- week), and 5.05% (16- week). Donors' QoL was similar across strategies, although self-reported symptoms were increased with shorter donation intervals. The shorter interval strategies increased average cost, with incremental cost-effectiveness ratios of £9.51 (95% CI 9.33 to 9.69) per additional whole-blood donation for the 8- versus 12- week interval for males, and £10.17 (95% CI 9.80 to 10.54) for the 12- versus 16- week interval arm for females. CONCLUSIONS: Over two years, reducing the minimum donation interval could provide additional units of whole-blood at a small additional cost, including for those donor subgroups whose blood type is in relatively high demand. However, the significance of self-reported symptoms needs to be investigated further before these policies are expanded.

Impact of the first wave of COVID-19 on outcomes following emergency admissions for common acute surgical conditions: analysis of a national database in England.

BACKGROUND: This study assessed the impact of the first COVID-19 wave in England on outcomes for acute appendicitis, gallstone disease, intestinal obstruction, diverticular disease, and abdominal wall hernia. METHODS: Emergency surgical admissions for patients aged 18 years and older to 124 NHS Trust hospitals between January and June in 2019 and 2020 were extracted from Hospital Episode Statistics. The risk of 90-day mortality after admission during weeks 11-19 in 2020 (national lockdown) and 2019 (pre-COVID-19) was estimated using multilevel logistic regression with case-mix adjustment. The primary outcome was all-cause mortality at 90 days. RESULTS: There were 12 231 emergency admissions and 564 deaths within 90 days during weeks 11-19 in 2020, compared with 18 428 admissions and 542 deaths in the same interval in 2019. Overall, 90-day mortality was higher in 2020 versus 2019, with an adjusted OR of 1.95 (95 per cent c.i. 0.78 to 4.89) for appendicitis, 2.66 (1.81 to 3.92) for gallstone disease, 1.99 (1.44 to 2.74) for diverticular disease, 1.70 (1.13 to 2.55) for hernia, and 1.22 (1.01 to 1.47) for intestinal obstruction. After emergency surgery, 90-day mortality was higher in 2020 versus 2019 for gallstone disease (OR 3.37, 1.26 to 9.02), diverticular disease (OR 2.35, 1.16 to 4.73), and hernia (OR 2.34, 1.23 to 4.45). For intestinal obstruction, the corresponding OR was 0.91 (0.59 to 1.41). For admissions not leading to emergency surgery, mortality was higher in 2020 versus 2019 for gallstone disease (OR 2.55, 1.67 to 3.88), diverticular disease (1.90, 1.32 to 2.73), and intestinal obstruction (OR 1.30, 1.06 to 1.60). CONCLUSION: Emergency admission was reduced during the first lockdown in England and this was associated with higher 90-day mortality.

Unmeasured confounding in nonrandomized studies: quantitative bias analysis in health technology assessment.

Evidence generated from nonrandomized studies (NRS) is increasingly submitted to health technology assessment (HTA) agencies. Unmeasured confounding is a primary concern with this type of evidence, as it may result in biased treatment effect estimates, which has led to much criticism of NRS by HTA agencies. Quantitative bias analyses are a group of methods that have been developed in the epidemiological literature to quantify the impact of unmeasured confounding and adjust effect estimates from NRS. Key considerations for application in HTA proposed in this article reflect the need to balance methodological complexity with ease of application and interpretation, and the need to ensure the methods fit within the existing frameworks used to assess nonrandomized evidence by HTA bodies.

Application of quantitative bias analysis for unmeasured confounding in cost-effectiveness modelling.

Due to uncertainty regarding the potential impact of unmeasured confounding, health technology assessment (HTA) agencies often disregard evidence from nonrandomized studies when considering new technologies. Quantitative bias analysis (QBA) methods provide a means to quantify this uncertainty but have not been widely used in the HTA setting, particularly in the context of cost-effectiveness modelling (CEM). This study demonstrated the application of an aggregate and patient-level QBA approach to quantify and adjust for unmeasured confounding in a simulated nonrandomized comparison of survival outcomes. Application of the QBA output within a CEM through deterministic and probabilistic sensitivity analyses and under different scenarios of knowledge of an unmeasured confounder demonstrates the potential value of QBA in HTA.

Effect of High-Flow Nasal Cannula Therapy vs Continuous Positive Airway Pressure Therapy on Liberation From Respiratory Support in Acutely Ill Children Admitted to Pediatric Critical Care Units: A Randomized Clinical Trial.

Importance: The optimal first-line mode of noninvasive respiratory support for acutely ill children is not known. Objective: To evaluate the noninferiority of high-flow nasal cannula therapy (HFNC) as the first-line mode of noninvasive respiratory support for acute illness, compared with continuous positive airway pressure (CPAP), for time to liberation from all forms of respiratory support. Design, Setting, and Participants: Pragmatic, multicenter, randomized noninferiority clinical trial conducted in 24 pediatric critical care units in the United Kingdom among 600 acutely ill children aged 0 to 15 years who were clinically assessed to require noninvasive respiratory support, recruited between August 2019 and November 2021, with last follow-up completed in March 2022. Interventions: Patients were randomized 1:1 to commence either HFNC at a flow rate based on patient weight (n = 301) or CPAP of 7 to 8 cm H2O (n = 299). Main Outcomes and Measures: The primary outcome was time from randomization to liberation from respiratory support, defined as the start of a 48-hour period during which a participant was free from all forms of respiratory support (invasive or noninvasive), assessed against a noninferiority margin of an adjusted hazard ratio of 0.75. Seven secondary outcomes were assessed, including mortality at critical care unit discharge, intubation within 48 hours, and use of sedation. Results: Of the 600 randomized children, consent was not obtained for 5 (HFNC: 1; CPAP: 4) and respiratory support was not started in 22 (HFNC: 5; CPAP: 17); 573 children (HFNC: 295; CPAP: 278) were included in the primary analysis (median age, 9 months; 226 girls [39%]). The median time to liberation in the HFNC group was 52.9 hours (95% CI, 46.0-60.9 hours) vs 47.9 hours (95% CI, 40.5-55.7 hours) in the CPAP group (absolute difference, 5.0 hours [95% CI -10.1 to 17.4 hours]; adjusted hazard ratio 1.03 [1-sided 97.5% CI, 0.86-∞]). This met the criterion for noninferiority. Of the 7 prespecified secondary outcomes, 3 were significantly lower in the HFNC group: use of sedation (27.7% vs 37%; adjusted odds ratio, 0.59 [95% CI, 0.39-0.88]); mean duration of critical care stay (5 days vs 7.4 days; adjusted mean difference, -3 days [95% CI, -5.1 to -1 days]); and mean duration of acute hospital stay (13.8 days vs 19.5 days; adjusted mean difference, -7.6 days [95% CI, -13.2 to -1.9 days]). The most common adverse event was nasal trauma (HFNC: 6/295 [2.0%]; CPAP: 18/278 [6.5%]). Conclusions and Relevance: Among acutely ill children clinically assessed to require noninvasive respiratory support in a pediatric critical care unit, HFNC compared with CPAP met the criterion for noninferiority for time to liberation from respiratory support. Trial Registration: ISRCTN.org Identifier: ISRCTN60048867.

A Machine-Learning Approach for Estimating Subgroup- and Individual-Level Treatment Effects: An Illustration Using the 65 Trial.

HIGHLIGHTS: This article examines a causal machine-learning approach, causal forests (CF), for exploring the heterogeneity of treatment effects, without prespecifying a specific functional form.The CF approach is considered in the reanalysis of the 65 Trial and was found to provide similar estimates of subgroup effects to using a fixed parametric model.The CF approach also provides estimates of individual-level treatment effects that suggest that for most patients in the 65 Trial, the intervention is expected to reduce 90-d mortality but with wide levels of statistical uncertainty.The study illustrates how individual-level treatment effect estimates can be analyzed to generate hypotheses for further research about those patients who are likely to benefit most from an intervention.

Local Instrumental Variable Methods to Address Confounding and Heterogeneity when Using Electronic Health Records: An Application to Emergency Surgery.

BACKGROUND: Electronic health records (EHRs) offer opportunities for comparative effectiveness research to inform decision making. However, to provide useful evidence, these studies must address confounding and treatment effect heterogeneity according to unmeasured prognostic factors. Local instrumental variable (LIV) methods can help studies address these challenges, but have yet to be applied to EHR data. This article critically examines a LIV approach to evaluate the cost-effectiveness of emergency surgery (ES) for common acute conditions from EHRs. METHODS: This article uses hospital episodes statistics (HES) data for emergency hospital admissions with acute appendicitis, diverticular disease, and abdominal wall hernia to 175 acute hospitals in England from 2010 to 2019. For each emergency admission, the instrumental variable for ES receipt was each hospital's ES rate in the year preceding the emergency admission. The LIV approach provided individual-level estimates of the incremental quality-adjusted life-years, costs and net monetary benefit of ES, which were aggregated to the overall population and subpopulations of interest, and contrasted with those from traditional IV and risk-adjustment approaches. RESULTS: The study included 268,144 (appendicitis), 138,869 (diverticular disease), and 106,432 (hernia) patients. The instrument was found to be strong and to minimize covariate imbalance. For diverticular disease, the results differed by method; although the traditional approaches reported that, overall, ES was not cost-effective, the LIV approach reported that ES was cost-effective but with wide statistical uncertainty. For all 3 conditions, the LIV approach found heterogeneity in the cost-effectiveness estimates across population subgroups: in particular, ES was not cost-effective for patients with severe levels of frailty. CONCLUSIONS: EHRs can be combined with LIV methods to provide evidence on the cost-effectiveness of routinely provided interventions, while fully recognizing heterogeneity. HIGHLIGHTS: This article addresses the confounding and heterogeneity that arise when assessing the comparative effectiveness from electronic health records (EHR) data, by applying a local instrumental variable (LIV) approach to evaluate the cost-effectiveness of emergency surgery (ES) versus alternative strategies, for patients with common acute conditions (appendicitis, diverticular disease, and abdominal wall hernia).The instrumental variable, the hospital's tendency to operate, was found to be strongly associated with ES receipt and to minimize imbalances in baseline characteristics between the comparison groups.The LIV approach found that, for each condition, there was heterogeneity in the estimates of cost-effectiveness according to baseline characteristics.The study illustrates how an LIV approach can be applied to EHR data to provide cost-effectiveness estimates that recognize heterogeneity and can be used to inform decision making as well as to generate hypotheses for further research.

Effect of High-Flow Nasal Cannula Therapy vs Continuous Positive Airway Pressure Following Extubation on Liberation From Respiratory Support in Critically Ill Children: A Randomized Clinical Trial.

Importance: The optimal first-line mode of noninvasive respiratory support following extubation of critically ill children is not known. Objective: To evaluate the noninferiority of high-flow nasal cannula (HFNC) therapy as the first-line mode of noninvasive respiratory support following extubation, compared with continuous positive airway pressure (CPAP), on time to liberation from respiratory support. Design, Setting, and Participants: This was a pragmatic, multicenter, randomized, noninferiority trial conducted at 22 pediatric intensive care units in the United Kingdom. Six hundred children aged 0 to 15 years clinically assessed to require noninvasive respiratory support within 72 hours of extubation were recruited between August 8, 2019, and May 18, 2020, with last follow-up completed on November 22, 2020. Interventions: Patients were randomized 1:1 to start either HFNC at a flow rate based on patient weight (n = 299) or CPAP of 7 to 8 cm H2O (n = 301). Main Outcomes and Measures: The primary outcome was time from randomization to liberation from respiratory support, defined as the start of a 48-hour period during which the child was free from all forms of respiratory support (invasive or noninvasive), assessed against a noninferiority margin of an adjusted hazard ratio (HR) of 0.75. There were 6 secondary outcomes, including mortality at day 180 and reintubation within 48 hours. Results: Of the 600 children who were randomized, 553 children (HFNC, 281; CPAP, 272) were included in the primary analysis (median age, 3 months; 241 girls [44%]). HFNC failed to meet noninferiority, with a median time to liberation of 50.5 hours (95% CI, 43.0-67.9) vs 42.9 hours (95% CI, 30.5-48.2) for CPAP (adjusted HR, 0.83; 1-sided 97.5% CI, 0.70-∞). Similar results were seen across prespecified subgroups. Of the 6 prespecified secondary outcomes, 5 showed no significant difference, including the rate of reintubation within 48 hours (13.3% for HFNC vs 11.5 % for CPAP). Mortality at day 180 was significantly higher for HFNC (5.6% vs 2.4% for CPAP; adjusted odds ratio, 3.07 [95% CI, 1.1-8.8]). The most common adverse events were abdominal distension (HFNC: 8/281 [2.8%] vs CPAP: 7/272 [2.6%]) and nasal/facial trauma (HFNC: 14/281 [5.0%] vs CPAP: 15/272 [5.5%]). Conclusions and Relevance: Among critically ill children requiring noninvasive respiratory support following extubation, HFNC compared with CPAP following extubation failed to meet the criterion for noninferiority for time to liberation from respiratory support. Trial Registration: isrctn.org Identifier: ISRCTN60048867.