Editorial: Dealing with the Missing Data Challenge in Clinical Trials
- PDF / 5,724,708 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 69 Downloads / 215 Views
Thomas Permutt CDER/FDA
DATA
403
Editorial: Dealing With the Missing Data Challenge in Clinical Trials
Jose Pinheiro Novarfis Pharmaceuticals
Correspondence Address Thomas Permuff. 10903 New Hampshire Ave.. Silver Spring, MD 20993 (email: Thomas.Permutt @ fda.hhs.gov),
Missing data remains a pervasive a n d important problem in clinical studies. How to properly include information from a patient who did not complete t h e study will often depend o n unverifiable assumptions about t h e underlying missing data mechanism. Different assumptions a n d associated analysis methods may lead to conflicting conclusions about the statistical significance a n d magnitude of treatment benefit. Understandably, there is considerable interest from key stakeholders in d r u g development (including regulatory agencies, pharmaceutical companies, a n d academic institutions) to reach a n agreement o n acceptable a n d recommended approaches for dealing with missing data. In recognition of t h e relevance of t h e topic, o n e of t h e two tracks in t h e Second FDA-DIA Statistics Forum (held in Bethesda, Maryland, o n April 14-16,2008) was dedicated to missing data, being cochaired by t h e two of us. Renowned experts o n missing data methods presented at t h e conference a n d led lively discussions with t h e participants o n t h e pros a n d cons of different approaches. This special issue of the Drug Information Journal intends to document t h e most relevant topics presented a n d discussed a t the meeting, serving as a reference for future research a n d debate in t h e field. We start, in this editorial, with o u r thoughts o n framing a n d addressing t h e problem of missing data in clinical trials.
THE PROBLEM OR PROBLEMS Consider t h e following five possible outcomes for a patient in a clinical trial. 1. The patient is enrolled in a cancer trial. Most pa-
tients in the trial will die of cancer over the course of the trial. The goal of therapy is to delay their deaths without affecting quality of life too adversely. The primary measured outcome is the time to death. Our patient, however, is somehow lost to follow-up. 2. The patient is enrolled in a trial of a drug to control premature ventricular contractions (PVC), which are associated with sudden cardiac death. The goal is to prevent sudden cardiac death, but the number of PVC is considered an acceptable surrogate. At the end of the trial, our patient, fortunately, has no PVC. Unfortunately, he is dead, of sudden cardiac arrest. 3. The patient is enrolled in a trial of an antibiotic for otitis media. The primary measured outcome is to be a clinical assessment of cure based on signs and symptoms after seven days of treatment. Our patient reports that her symptoms were gone after three days and declines to be seen. 4. The patient is hypertensive. The goal of therapy is to mitigate the long-term consequences of hypertension, but diastolic blood pressure at the end of six months is the primary, surrogate outcome. The patient discontinues therapy after two weeks beca
Data Loading...
