| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters and Responses |
We agree with Hart that differential item functioning (DIF) across patients with various impairments is clinically relevant and an issue deserving of additional study. Although this issue was not the focus of our paper, we did examine the potential presence of DIF in the AM-PAC items most frequently administered in the AM-PAC-CAT across outpatients in our sample with different primary impairments. Only 1 item out of 36 displayed a significant DIF. We suspect that more focused constructs such as basic mobility and daily activity function may have less potential for significant DIF than broader health-related concepts. We do agree with Dr Hart that DIF is an issue that should be examined both during item bank development and in CAT applications in various patient populations.1
With respect to Hart's suggestion that the daily activity and basic mobility domains of the AM-PAC might not be distinct in an outpatient population, we wish to clarify that outpatients were represented in our calibration samples. The AM-PAC's combined calibration samples of 1,041 patients in postacute care included patients from 4 different care settings: outpatient therapy (n=237), home health care (n=246), skilled nursing or transitional care (n=138), and inpatient rehabilitation (n=420).2 The AMPAC was intentionally developed and tested in samples drawn from several post–acute care settings to provide users with one instrument that had the ability to track functional recovery across settings throughout an entire episode of post–acute care.
In separate analyses done on the outpatient sample used in this pilot study, we confirmed a distinction between the daily activities and basic mobility domains of the AM-PAC. We saw only a moderate positive correlation between the basic mobility and daily activity scales (0.40 on admission and 0.55 at discharge), suggesting the psychometric and clinical merits of keeping these 2 domains of activity function separate and distinct.
Given the dynamic nature of CAT outcome instruments, a feature that allows for periodic refinements and updating, Dr Hart raises an interesting concern about a potential challenge of keeping users (as well as journal editors and reviewers) current with pertinent changes in various CAT instruments being used with increasing frequency in health care. We agree that this is an important issue that must be taken seriously. Our current thinking is that CAT instrument developers might look to the broader software development field for guidance on how this might be efficiently accomplished by adopting a policy of labeling different versions of CAT software. For instance, although "version 1" (AM-PAC-CATv1) was examined in this pilot study, "version 2" (AM-PAC-CATv2) is soon to be released and will be the subject of future study. Accurate labeling of software and instruments may help readers and various users keep track of the evolution of CAT software.
Again, we thank Hart for his letter and look forward to further discussions of these and related issues relevant to the introduction and use of CAT instruments in health care.
AM Jette, PT, PhD, FAPTA, is Director, Health & Disability Institute, Boston University of Public Health
 | Footnotes |
|---|
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
