| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
Rockville, MD
Correspondence to: Saul Malozowski, MD, PhD, 5600 Fisher Ln, Parklawn Bldg, Room 14B04, Rockville, MD 20857
To the Editor:
The article by Simmons et al (August 2000)1 and the accompanying editorial2 highlight how little we know about the lack of compliance with prescribed medication regimens despite the advances in clinical trial design. The articles also generate hypotheses in related fields, eg, the role of poor compliance and the rise in antibiotic resistance.
Nonetheless, the study design employed and the conclusions reached merit some discussion. The patients appear to have been randomized, but the details of the randomization, including whether block randomization occurred, were not presented. The sample size is relatively small (n = 236) and may not be representative of the larger population (n = 5,887). The study addresses only the issue of device manipulation and does not address other drug product handling variables that could alter the amount of the drug inhaled and absorbed, such as incorrect meter use or improper drug storage. Nor does it address less overt noncompliance and dumping, such as intentional proper and timely activation of the device without inhalation. The design of the study also cannot reveal whether compliance patterns change over time.
We agree with the authors that estimating noncompliance is problematic. Nonetheless, as suggested by the editor, the clinical data may reveal clues. Intent-to-treat (ITT) trial data include data from all enrolled patients. Per the protocol, trial data include information only on individuals who completed the study as designed. A divergence between the efficacy and safety results of those who complete a trial per protocol vs results from those who dropped out vs the results from an ITT analysis should raise questions about what will happen when a drug reaches the market. When exposed to a drug, the general public would behave in a manner that is more similar to the ITT population than to the per-protocol population. The general public will comprise a mixture of subjects with a wide range of concurrent diseases who will be receiving concomitant medications with varying degrees of compliance. If, however, there was a significant degree of noncompliance, the ITT results would also underestimate the potential for adverse drug reactions, as stated in the editorial. Events that require a threshold of exposure, ie, a minimum dose level or cumulative exposure, will be underestimated if patients are not taking the medication as prescribed. Similarly, drug reactions of an idiosyncratic nature may not occur if sufficient numbers of patients are not taking the medication.
References
University of California Los Angeles, CA
Correspondence to: Michael S. Simmons, MS, Department of Medicine, Division of Pulmonary and Critical Care Medicine, University of California, Los Angeles, CA 90095-1690; e-mail: msimm{at}ucla.edu
To the Editor:
Thank you for the opportunity to respond to the thoughtful comments of Drs. Malozowski and Koller. Their cautioning about the interpretation of the results of our study and the extrapolation of those results to other populations are well-taken and must be kept in mind with any clinical trial. I would point out, however, that the sample selected for our study was fairly large for a compliance study in which detailed information on medication use was collected with electronic monitors over a relatively long period, perhaps the largest study of its kind. It is hoped, of course, that even larger prospective compliance studies employing electronic medication monitors can be conducted in the future. Because the study design and randomization method employed in the Lung Health Study have been described previously, the readers are referred to the articles by Rand et al1 and Connett et al2 (references 5 and 15 in our article).
Perhaps future studies also can address additional factors that may affect the measurement of adherence to a study protocol, as mentioned by the correspondents, such as medication storage, metered-dose inhaler (MDI) technique, and the intentional actuation of MDIs without inhalation of the medication. Data of this type are difficult to collect at present (although devices that identify the latter are now available) and were outside the scope of this study. The assessment of changes in compliance patterns over time can be measured with the data available in the Lung Health Study; I refer the correspondents to reference 9 in an article,3 previously published in CHEST in 1996.
The correspondents also point out the limitations of the analysis of data obtained only from patients who return for follow-up visits, as opposed to an intent-to-treat (ITT) analysis. While dropouts should always be considered in the interpretation of data, it was not possible to conduct an ITT analysis in this assessment of deception in compliance with prescribed inhaler use. Furthermore, an assumption that the results of controlled clinical trials accurately represent the real world of clinical practice could be as much a mistake as assuming that all medication is taken as prescribed. Dr. Chapman’s excellent editorial4 in the same issue of CHEST also addresses these issues.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
