Relationship between test frequency and outcomes of anticoagulation: A literature review and commentary with implications for the design of randomized trials of patient self-management
Gp. Samsa et Db. Matchar, Relationship between test frequency and outcomes of anticoagulation: A literature review and commentary with implications for the design of randomized trials of patient self-management, J THROMB TH, 9(3), 2000, pp. 283-292
Background: Patient self-management (PSM) of anticoagulation, which is prim
arily based upon the premise that more frequent testing will lead to tighte
r anticoagulation control and thus to improved clinical outcomes, is a prom
ising model of care. The goals of this paper are (1) to describe the streng
th of evidence correlating more frequent testing with improved outcomes; an
d (2) to discuss implications of these findings for the design of randomize
d controlled trials (RCTs) assessing the effectiveness and cost-effectivene
ss of PSM.
Methods: We performed two literature reviews: one examining the strength of
the relationship between time in target range (TTR) and the clinical outco
mes of major bleeding and thromboembolism; and the second examining the str
ength of the relationship between frequency of testing and TTR.
Results: We found that (1) the relationship between TTR and clinical outcom
es is strong, thus supporting use of TTR as a primary outcome variable; and
(2) more frequent testing seems to increase TTR, although the studies supp
orting this latter conclusion were relatively few and not definitive. Stati
stical analysis suggested that a study which uses clinical event rates as i
ts primary outcome would need to be much larger than a comparable study whi
ch is based upon TTR.
Conclusions: When designing randomized trials of PSM, the design should (1)
use as its control group high quality anticoagulation management rather th
an usual care; (2) include the maximum possible amount of self-management i
n the intervention group; (3) include different testing intervals in the in
tervention group; (4) use TTR as the primary outcome variable and event rat
es as a secondary outcome; and (5) base the sample size calculations upon a
5-10% absolute improvement in TTR. Additional RCTs are needed in order to
determine how the promise of PSM can best be fulfilled.