Dataset Information

The use of percentage change from baseline as an outcome in a controlled trial is statistically inefficient: a simulation study.

ABSTRACT: BACKGROUND:Many randomized trials involve measuring a continuous outcome - such as pain, body weight or blood pressure - at baseline and after treatment. In this paper, I compare four possibilities for how such trials can be analyzed: post-treatment; change between baseline and post-treatment; percentage change between baseline and post-treatment and analysis of covariance (ANCOVA) with baseline score as a covariate. The statistical power of each method was determined for a hypothetical randomized trial under a range of correlations between baseline and post-treatment scores. RESULTS:ANCOVA has the highest statistical power. Change from baseline has acceptable power when correlation between baseline and post-treatment scores is high;when correlation is low, analyzing only post-treatment scores has reasonable power. Percentage change from baseline has the lowest statistical power and was highly sensitive to changes in variance. Theoretical considerations suggest that percentage change from baseline will also fail to protect from bias in the case of baseline imbalance and will lead to an excess of trials with non-normally distributed outcome data. CONCLUSIONS:Percentage change from baseline should not be used in statistical analysis. Trialists wishing to report this statistic should use another method, such as ANCOVA, and convert the results to a percentage change by using mean baseline scores.

SUBMITTER: Vickers AJ

PROVIDER: S-EPMC34605 | biostudies-literature | 2001

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

The use of percentage change from baseline as an outcome in a controlled trial is statistically inefficient: a simulation study.

Vickers A J AJ

BMC medical research methodology 20010628

<h4>Background</h4>Many randomized trials involve measuring a continuous outcome - such as pain, body weight or blood pressure - at baseline and after treatment. In this paper, I compare four possibilities for how such trials can be analyzed: post-treatment; change between baseline and post-treatment; percentage change between baseline and post-treatment and analysis of covariance (ANCOVA) with baseline score as a covariate. The statistical power of each method was determined for a hypothetical ...[more]

PMID: 11459516

Similar Datasets

Project description:Evidence concerning the importance of glucose lowering in the prevention of cardiovascular (CV) outcomes remains controversial. Given the multi-faceted pathogenesis of atherosclerosis in diabetes, it is likely that any intervention to mitigate this risk must address CV risk factors beyond glycemia alone. The SGLT-2 inhibitor empagliflozin improves glucose control, body weight and blood pressure when used as monotherapy or add-on to other antihyperglycemic agents in patients with type 2 diabetes. The aim of the ongoing EMPA-REG OUTCOME™ trial is to determine the long-term CV safety of empagliflozin, as well as investigating potential benefits on macro-/microvascular outcomes.Patients who were drug-naïve (HbA1c ?7.0% and ?9.0%), or on background glucose-lowering therapy (HbA1c ?7.0% and ?10.0%), and were at high risk of CV events, were randomized (1:1:1) and treated with empagliflozin 10 mg, empagliflozin 25 mg, or placebo (double blind, double dummy) superimposed upon the standard of care. The primary outcome is time to first occurrence of CV death, non-fatal myocardial infarction, or non-fatal stroke. CV events will be prospectively adjudicated by an independent Clinical Events Committee. The trial will continue until ?691 confirmed primary outcome events have occurred, providing a power of 90% to yield an upper limit of the adjusted 95% CI for a hazard ratio of <1.3 with a one-sided ? of 0.025, assuming equal risks between placebo and empagliflozin (both doses pooled). Hierarchical testing for superiority will follow for the primary outcome and key secondary outcomes (time to first occurrence of CV death, non-fatal myocardial infarction, non-fatal stroke or hospitalization for unstable angina pectoris) where non-inferiority is achieved.Between Sept 2010 and April 2013, 592 clinical sites randomized and treated 7034 patients (41% from Europe, 20% from North America, and 19% from Asia). At baseline, the mean age was 63?±?9 years, BMI 30.6?±?5.3 kg/m2, HbA1c 8.1?±?0.8%, and eGFR 74?±?21 ml/min/1.73 m2. The study is expected to report in 2015.EMPA-REG OUTCOME™ will determine the CV safety of empagliflozin in a cohort of patients with type 2 diabetes and high CV risk, with the potential to show cardioprotection.Clinicaltrials.gov NCT01131676.

Project description:IntroductionTeaching endotracheal intubation is uniquely challenging due to its technical, high-stakes, and highly time-sensitive nature. The GoPro is a small, lightweight, high-resolution action camera with a wide-angle field of view that can encompass both the airway as well as the procedurist's hands and positioning technique when worn with a head mount. We aimed to evaluate its effectiveness in improving intubation teaching for novice learners in a simulated setting, via a two-arm, parallel group, randomized controlled superiority trial with 1:1 allocation ratio.MethodsWe recruited Year 4 medical students at the start of their compulsory 2-week Anesthesia posting. Participants underwent a standardized intubation curriculum and a formative assessment, then randomized to receive GoPro or non-GoPro led feedback. After a span of three months, participants were re-assessed in a summative assessment by blinded accessors. Participants were also surveyed on their learning experience for a qualitative thematic perspective. The primary outcomes were successful intubation and successful first-pass intubation.ResultsSeventy-one participants were recruited with no dropouts, and all were included in the analysis. 36 participants received GoPro led feedback, and 35 participants received non-GoPro led feedback. All participants successfully intubated the manikin. No statistically significant differences were found between the GoPro group and the non-GoPro group at summative assessment (85.3% vs 90.0%, p = 0.572). Almost all participants surveyed found the GoPro effective for their learning (98.5%). Common themes in the qualitative analysis were: the ability for an improved assessment, greater identification of small details that would otherwise be missed, and usefulness of the unique point-of-view footage in improving understanding.ConclusionsThe GoPro is a promising tool for simulation-based intubation teaching. There are considerations in its implementation to maximize the learning experience and yield from GoPro led feedback and training.

Project description:BACKGROUND:In the field of health care, knowledge and clinical reasoning are key with regard to quality and confidence in decision making. The development of knowledge and clinical reasoning is influenced not only by students' intrinsic factors but also by extrinsic factors such as satisfaction with taught content, pedagogic resources and pedagogic methods, and the nature of the objectives and challenges proposed. Nowadays, professors play the role of learning facilitators rather than simple "lecturers" and face students as active learners who are capable of attributing individual meanings to their personal goals, challenges, and experiences to build their own knowledge over time. Innovations in health simulation technologies have led to clinical virtual simulation. Clinical virtual simulation is the recreation of reality depicted on a computer screen and involves real people operating simulated systems. It is a type of simulation that places people in a central role through their exercising of motor control skills, decision skills, and communication skills using virtual patients in a variety of clinical settings. Clinical virtual simulation can provide a pedagogical strategy and can act as a facilitator of knowledge retention, clinical reasoning, improved satisfaction with learning, and finally, improved self-efficacy. However, little is known about its effectiveness with regard to satisfaction, self-efficacy, knowledge retention, and clinical reasoning. OBJECTIVE:This study aimed to evaluate the effect of clinical virtual simulation with regard to knowledge retention, clinical reasoning, self-efficacy, and satisfaction with the learning experience among nursing students. METHODS:A randomized controlled trial with a pretest and 2 posttests was carried out with Portuguese nursing students (N=42). The participants, split into 2 groups, had a lesson with the same objectives and timing. The experimental group (n=21) used a case-based learning approach, with clinical virtual simulator as a resource, whereas the control group (n=21) used the same case-based learning approach, with recourse to a low-fidelity simulator and a realistic environment. The classes were conducted by the usual course lecturers. We assessed knowledge and clinical reasoning before the intervention, after the intervention, and 2 months later, with a true or false and multiple-choice knowledge test. The students' levels of learning satisfaction and self-efficacy were assessed with a Likert scale after the intervention. RESULTS:The experimental group made more significant improvements in knowledge after the intervention (P=.001; d=1.13) and 2 months later (P=.02; d=0.75), and it also showed higher levels of learning satisfaction (P<.001; d=1.33). We did not find statistical differences in self-efficacy perceptions (P=.9; d=0.054). CONCLUSIONS:The introduction of clinical virtual simulation in nursing education has the potential to improve knowledge retention and clinical reasoning in an initial stage and over time, and it increases the satisfaction with the learning experience among nursing students.

Project description:ObjectiveTo pilot test the proposed DiaPROM trial components and address uncertainties associated with conducting a full-scale randomised controlled trial (RCT) to evaluate whether such a trial is feasible.DesignTwo-arm pilot RCT.ParticipantsAdults aged ≥18-39 years, with minimum 1 year type 1 diabetes duration, attending outpatient follow-up. Exclusion criteria were pregnancy, severe cognitive, somatic or psychiatric conditions and impaired vision.Randomisation and interventionAll participants completed electronic Patient-Reported Outcome Measures (PROMs) prior to the annual diabetes consultation. Using computer-generated block-randomisation without blinding, we assigned participants in a 1:1 ratio stratified by sex to receive standard care or an intervention. Physicians reviewed diabetes distress scores (Problem Areas In Diabetes scale) and referred individuals with scores ≥30 or single item(s) ≥3 to minimum two diabetes nurse consultations where reported problems were reviewed and discussed.OutcomesRecruitment and retention rates; participants perceptions about intervention components. Variance and estimated between-group differences in follow-up scores (Diabetes Distress Scale (DDS), WHO 5-Well-being Index, Perceived Competence for Diabetes Scale and glycaemic control) and DDS correlation with baseline scores, to assist sample size calculations.ResultsWe randomised 80 participants to the control or intervention arm (one participant was later excluded). 23/39 intervention arm participants qualified for additional consultations and 17 attended. 67/79 attended the 12-month follow-up (15.2% attrition); 5/17 referred to additional consultations were lost to follow-up (29.4% attrition). Participants reported PROMs as relevant (84.6%) and acceptable (97.4%) but rated the usefulness of consultations as moderate to low. Baseline mean±SD DDS score was 2.1±0.69; DDS SD was 0.71 (95% CI: 0.60 to 0.86) at follow-up; correlation between baseline and follow-up DDS scores was 0.8 (95% CI: 0.7 to 0.9).ConclusionsThe pilot trial revealed need for intervention modifications ahead of a full-scale trial to evaluate use of PROMs in diabetes consultations. Specifically, participant acceptability and intervention implementation need further investigation.

Dataset Information

The use of percentage change from baseline as an outcome in a controlled trial is statistically inefficient: a simulation study.

Publications

The use of percentage change from baseline as an outcome in a controlled trial is statistically inefficient: a simulation study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets