Intent-to-treat: A curse of clinical trials

1998 ◽  
Vol 4 (1) ◽  
pp. 1-2 ◽  
Author(s):  
Jay N. Cohn
1998 ◽  
Vol 26 (2) ◽  
pp. 57-65 ◽  
Author(s):  
R Kay

If a trial is to be well designed, and the conclusions drawn from it valid, a thorough understanding of the benefits and pitfalls of basic statistical principles is required. When setting up a trial, appropriate sample-size calculation is vital. If initial calculations are inaccurate, trial results will be unreliable. The principle of intent-to-treat in comparative trials is examined. Randomization as a method of selecting patients to treatment is essential to ensure that the treatment groups are equalized in terms of avoiding biased allocation in the mix of patients within groups. Once trial results are available the correct calculation and interpretation of the P-value is important. Its limitations are examined, and the use of the confidence interval to help draw valid conclusions regarding the clinical value of treatments is explored.


2020 ◽  
Vol 4 (Supplement_1) ◽  
Author(s):  
George J Kahaly ◽  
Terry J Smith ◽  
Robert Holt ◽  
Saba Sile ◽  
Raymond S Douglas

Abstract Introduction: Teprotumumab, an insulin-like growth factor 1 receptor inhibitory monoclonal antibody, was recently shown to significantly reduce proptosis in patients with active, moderate-to-severe thyroid eye disease (TED) in phase 2 and phase 3 clinical trials.1,2 Prior analyses have demonstrated a combined trial proptosis response (≥2 mm reduction) rate of 77.4% in the teprotumumab group and 14.9% in the placebo group after 24 weeks of therapy (p < 0.001).3 The current analysis was performed to investigate whether or not patient demographic characteristics influence the teprotumumab proptosis response. Methods: Data from two 24-week randomized, double-masked, placebo-controlled, parallel-group, multicenter studies (Phase 2 [NCT01868997], Phase 3 [NCT03298867[) were combined. All patients had active TED associated with Graves’ disease. The study eye designated at baseline manifested more severe TED and a clinical activity score of > 4. Subjects were divided into subgroups based on gender, smoking status, and age at baseline (younger: <65, older: ≥65). The percentage of proptosis (≥2 mm) responders and proptosis change from baseline were examined in each of these subgroups. Because most of both teprotumumab (85%) and placebo (87%) subjects were white, there were insufficient numbers of subjects to examine the effect of race on the teprotumumab proptosis response. All analyses were performed on the intent-to-treat (ITT) population using data from the study eye. Results: A total of 171 patients comprised the population from the two studies. Eighty-four and 87 patients were randomized to the teprotumumab and placebo groups, respectively, and the treatment groups had balanced baseline characteristics. At week 24, significantly more teprotumumab than placebo patients were proptosis responders in all examined subgroups (male: 73.1% vs. 5.0%, female: 79.3% vs. 17.9%, smokers: 70.0% vs. 23.1%, non-smokers 79.7% vs. 11.5%, younger: 76.1% vs. 16.2%, older: 84.6% vs. 7.7%; all p < 0.001). In continuous variable analyses, the mean proptosis reduction from baseline was also significantly greater at week 24 in teprotumumab-treated patients than placebo patients (male: -3.34 vs. -0.07 mm, female: -3.10 vs. -0.42 mm, smokers: -2.99 vs. -0.72 mm, non-smokers: -3.20 vs. -0.31 mm, younger: -3.10 vs. -0.39 mm, older: -3.55 vs. -0.22 mm; all p < 0.001). Conclusion: Teprotumumab was effective across subgroups of age, gender, and smoking status in the pooled 24-week clinical trials. Reference: (1) Smith TJ, et al. N Engl J Med 2017;376:1748-1761. (2) Douglas RS, et al. AACE 2019 late-breaking abstract. (3) Kahaly GJ, et al. Thyroid 2019;29(Suppl1):A-1 [abstract].


Blood ◽  
2011 ◽  
Vol 118 (21) ◽  
pp. 995-995 ◽  
Author(s):  
Ashutosh Wechalekar ◽  
Stefan O Schonland ◽  
Efstathios Kastritis ◽  
Philip N Hawkins ◽  
Meletios A. Dimopoulos ◽  
...  

Abstract Abstract 995 N-terminal fragment of BNP (NT-proBNP) and cardiac troponin –T (TnT) or I (TnI) form a useful staging system in AL amyloidosis and poor outcomes have been reported in stage III patients treated before routine use of novel agents. These patients are routinely excluded from clinical trials and prospective outcome data is limited but recent studies suggest that some such patients may have better outcomes. We report the outcomes of 347 patients with Mayo stage III AL amyloidosis seen at the amyloid centres in London (UK), Pavia (Italy), Heidelberg (Germany) and Athens (Greece). Organ involvement and responses are defined according to 2005 amyloidosis consensus criteria. Presenting features were [n (%)/median (range)]: cardiac, renal and liver involvement in 338 (97%), 216 (62%) and 77 (22%) respectively, NT-proBNP 9106 ng/L (379–216187); TnI – 0.18 ng/ml (0.1–12); TnT −0.09 ng/ml (0.04–8.2) and IVS 15 mm (7–24). Treatments given were: Bortezomib combinations - 23 (7%), MDex - 150 (43%), thalidomide combinations - 96 (28%), lenalidomide combinations - 13 (4%). 30 (8%) were deemed too ill for treatment or died prior to treatment initiation. Only 37% completed the planned treatment course. The haematological responses on an intention to treat basis were seen in (Overall response rate/complete response (CR)/partial response (PR))(n(%)): MDex – 63(42%)/24 (16%)/39 (26%); Thalidomide combinations 31(32%)/11(12%)/20(21%), bortezomib combinations 10(43%)/6 (23%)/4 (17%), lenalidomide combination 5(38%)/0(0%)/5(38%). The median overall survival (OS) was 7.1 mos. The overall survival at 12 months from response evaluation was 74% for CR, 52% for PR and 18% for NR and from diagnosis was (median): CR – 59 mo, PR 28 mo, NR 10 mo and not assessable for response 2.9 mos. Stage III patients without echocardiographic evidence of cardiac involvement had excellent outcomes with 80% estimated 2 year OS. Using best fit cut-off, in multivariate model (including NT-proBNP., systolic blood pressure (SBP), ejection fraction, NYHA, ECOG, dFLC, LV wall thickness), NT-proBNP > 8000 ng/L (HR 2.3; p <0.0001) and SBP < 100 mm of Hg (HR 1.6; p<0.0001) were the only independent predictors of poor outcome. Using NT-proBNP >8000 ng/L and SPB <100 mm of Hg as high risk criteria, stage III patients can be further subdivided based on presence of none, one or two criteria with OS of 25 mo, 6 mo and 3 mo respectively. Using these criteria, on intent to treat basis, OS by CR/PR/NR was: no high risk factors – median not reached/69 mo/7 mo and one high risk factor - 59 mo/23 mo/4 mos respectively and too few patients for patients with two high risk factors making comparison unreliable. In conclusion, outcomes amyloidosis patients with stage III disease remain poor. However, stage III patients are heterogeneous and combination of NT-proBNP and SBP can usefully sub-classify these patients. Patients with abnormal biomarkers just due to renal failure in absence of cardiac involvement should be excluded from the current Mayo staging system. Although, treatment responses of stage III patients, on intent to treat basis, are poor with all regimes, it is encouraging that haematological responses improve outcomes and patients who achieve a CR have best outcomes. Clinical trials are urgently needed in patients with stage III disease to confirm these findings and define optimal treatment options. Disclosures: No relevant conflicts of interest to declare.


1995 ◽  
Vol 6 (6_suppl) ◽  
pp. 28-33 ◽  
Author(s):  
R. Kay

The correct application of appropriate statistical methods in clinical trials is essential for deriving valid conclusions. This paper covers a number of the key principles, with particular relevance to studies in herpes zoster. The distinct roles of the P-value and the confidence interval are investigated and their use in the assessment of the equivalence of two treatments evaluated. The principle of intent-to-treat is discussed and is seen to be vitally important for obtaining correct conclusions. The definition of end-points in the assessment of pain in herpes zoster is controversial; it is argued here that the only valid approach is to consider the time from start of treatment to complete cessation of all pain. Finally, methods for the evaluation of time to event data are explained in general terms, and particular drawbacks relevant to herpes zoster are discussed.


2018 ◽  
Vol 16 (1) ◽  
pp. 63-70 ◽  
Author(s):  
Fred Yang ◽  
Janet Wittes ◽  
Bertram Pitt

Introduction: Assessing safety is important to evaluating new medications. In many randomized clinical trials, assessment of safety relies on so-called on-treatment analysis, where data on adverse events are collected only while the participant is taking study medication and perhaps for a few (7, 14, or 30) days after stopping. This article discusses the consequence of such failure to use intent-to-treat analyses in assessing safety. Methods: This article discusses two approaches to analysis of safety data: intention-to-treat and on-treatment analysis with reference to principles of the design of randomized clinical trial. Results: On-treatment analysis violates randomization and is often not well defined. Moreover, because the typical on-treatment analysis ignores the reason participants in clinical trials stop treatment, on-treatment analyses can lead to biased estimates of risk. Examples show biases that can result from failure to count all adverse events. An example from a study of rofecoxib shows an on-treatment analysis that led to likely underestimation of harm; an example from a study of saxagliptin shows an on-treatment analysis that led to a likely overestimate of harms. Conclusion: For major safety outcomes in long-term clinical trials, intention-to-treat analysis should be performed in the framework of benefit–risk evaluation. More generally, analyses of safety should be tailored to the specific question being asked with the specific study design under consideration. On-treatment analyses are subject to bias; however, the direction of that bias is not necessarily clear.


1999 ◽  
Vol 33 (2) ◽  
pp. 87-95 ◽  
Author(s):  
Sati Mazumdar ◽  
Kenneth S. Liu ◽  
Patricia R. Houck ◽  
Charles F.Reynolds Iii

2021 ◽  
pp. 1-15
Author(s):  
Steven D. Targum ◽  
Lisa Fosdick ◽  
Kristen E. Drake ◽  
Paul B. Rosenberg ◽  
Anna D. Burke ◽  
...  

Background: Age may affect treatment outcome in trials of mild probable Alzheimer’s disease (AD). Objective: We examined age as a moderator of outcome in an exploratory study of deep brain stimulation targeting the fornix (DBS-f) region in participants with AD. Methods: Forty-two participants were implanted with DBS electrodes and randomized to double-blind DBS-f stimulation (“on”) or sham DBS-f (“off”) for 12 months. Results: The intervention was safe and well tolerated. However, the selected clinical measures did not differentiate between the “on” and “off” groups in the intent to treat (ITT) population. There was a significant age by time interaction with the Alzheimer’s Disease Assessment Scale; ADAS-cog-13 (p = 0.028). Six of the 12 enrolled participants <  65 years old (50%) markedly declined on the ADAS-cog-13 versus only 6.7%of the 30 participants≥65 years old regardless of treatment assignment (p = 0.005). While not significant, post-hoc analyses favored DBS-f “off” versus “on” over 12 months in the <  65 age group but favored DBS-f “on” versus “off” in the≥65 age group on all clinical metrics. On the integrated Alzheimer’s Disease rating scale (iADRS), the effect size contrasting DBS-f “on” versus “off” changed from +0.2 (favoring “off”) in the <  65 group to –0.52 (favoring “on”) in the≥65 age group. Conclusion: The findings highlight issues with subject selection in clinical trials for AD. Faster disease progression in younger AD participants with different AD sub-types may influence the results. Biomarker confirmation and genotyping to differentiate AD subtypes is important for future clinical trials.


1996 ◽  
Vol 17 (2) ◽  
pp. S87
Author(s):  
Kathryn Hirst ◽  
Barbara Tilley ◽  
Mei Lu

Sign in / Sign up

Export Citation Format

Share Document