> -----Original Message-----
> From: GP-UK [mailto:[log in to unmask]] On Behalf Of Declan Fox
> One of my worries, from seeing RAT in use, is that it perhaps reduces the
> tendency to question the patient as to WHY control might have changed. I
> know what the prompts say but.....
> I don't think I have worked anywhere using INRStar--yet. INRStar may be
> better?

When you use INRstar and you plan to add anew INR result you first have to
get past a splash screen which requires you to ask the patient important
questions:

1. Check that this is the right patient, right name, right date of birth.
2. Check whether or not there have been any adverse reactions since last
visit
3. Check the dose currently being taken is dose agreed at last visit
4. Check for missed doses during preceding seven days
5. Check for changes in medication - prescribed or OTC - since last visit.

The answers to all these questions will provide food for thought if the
patient's control starts to waver. The clinician can also then move on to
dietary factors such as alcohol and brassica vegetables which cause all the
problems with control at Christmas.

> <<It seems reasonable to assume that good INR control equates to less
> thrombo-embolic events but I'm not sure you would get ethical approval for
a
> study to confirm this hypothesis! >>
> 
> I was curious so I checked Clinical Evidence which said no studies done.
> Reason I am a wee bit worried is that all the studies showing the benefits
> of anticoag (in those situations where benefit has indeed been shown) were
> done with humans. Who do seem to get it wrong a bit more often than the
> software. And IF---and I agree it may be a big IF---the humans tend to
> underdose more often than they overdose, then the software may well cause
> patients to be more anti-coagulated and such additional effect may tilt
the
> balance towards more bleeding. I don't know---but I am sure that
information
> is somewhere in the studies comparing humans to software.

I agree that humans tend to underdose - when I ran a manual system my brain
defaulted to an INR of between 2-3 for everyone. When you enter patients
onto INRstar you have to choose the indication for warfarin and the program
then pops up the default target value for that condition. If you don't like
that national recommendation you can alter it, either when you start the
patient on warfarin or at any later stage, if they develop bruising at the
high end of their target range, for example. You can even skew the target
range around the target value so that the algorithm is allowed greater
tolerance below target than above.
 
> But I would have to agree that decent software must be a hell of a lot
> better than some of the docs I have seen attempting to control Warfarin.

True - the average point prevalence figure for practices transferring their
statistics onto INRstar is 55-60%. The average for those practices using
INRstar for a few months is currently 80% and during the last quarter we had
three user practices achieve 100%.

Robert