Thanks. I did consider creating a new variable for every combination, but as I'm sure you can imagine, there are many possible combinations, and the n values for some of the sub-groups risk being very small.
I think the most useful compromise is to have the binary variable pharmacological & non-pharmacological, and then have a different variable for the most common multiple uses.
Thanks for your suggestions
Andrew
Dr. Andrew Symon
Senior Lecturer
School of Nursing & Midwifery
University of Dundee
>>> [log in to unmask] 22/09/2005 17:36 >>>
Hi
in regards to multiple use of analgesia (e.g. TENS and
pethidine; pethidine and epidural) you can creat a new variables that combined every two methods together for example pethidin and epidural, pethidin alone, epidural alone, TENS and pethidine then you can get percentage and frequencies for those women who had combined method. and again you can put them on an ordinal scale. i think no harm to have two type of measurment for the same question as in your case. The first time you put on nominal scale pharmacological coded (1)and nonpharmacological coded (2) as binary respose and next to this variable (coding) you insert another coding on an ordinal scale that includes every method alone and two method together.
hopwe i did not missunderstand.
thanks
Andrew Symon <[log in to unmask]> wrote:
Thanks to all who've responded to this thread.
I think the pharmacological * non-pharmacological distinction is helpful, and I liked the idea of calculating a simple percentage rate of different types of analgesia by unit. While that gives some indication, there will still be occasions of multiple use of analgesia (e.g. TENS and pethidine; pethidine and epidural) and that's what I was trying to get around.
Thanks again,
Andrew Symon
>>> [log in to unmask] 21/09/2005 16:59 >>>
Hi Andrew,
When I did a comparison of the level of intervention in maternity units, I
compared similar women, i.e. healthy nulliparous and parous women at term
of a singleton pregnancy, in spontaneous labour, cephalic
presentation. That enabled me to distinguish the level of intervention,
including the use of low level pain relief (nil, Entonox, water), pethidine
and epidural as this was recorded on the computer programme used by 11
units. This suggested that there were major differences in the care given
to similar women in different units.
I did not allocate particular points for the particular use of whatever
intervention method, but rather ranked units according to their percentage
rate of use for each intervention. That prevented a quality judgment but
maintained the ranked order of use of intervention.
I wonder if this might be useful for what you want to do. If so, the
reference is
Mead MM Kornbrot D (2004) An intrapartum intervention scoring system for
the comparison of maternity units' intrapartum care of nulliparous women
suitable for midwifery-led care. Midwifery, 20(1): 15-26.
Best wishes
Marianne
>This request relates to comparing units, not individuals.
>Is anyone aware of a method of categorising the analgesia used in labour
>as a quantifiable outcome measure? I know that pain is a subjective
>experience, and I'm not trying to infer anything about the women who
>receive certain forms of pain relief.
>It's easy to describe the amount and types of analgesia used within a
>certain unit, but what can we infer from that? If we were to be told that
>a certain unit had an 80% opiate analgesia rate, or an 80% epidural rate,
>we might make certain assumptions about that unit, especially if we are
>told that another similar unit has much lower rates.
>Can this be scored? E.g. TENS is worth one 'point', Entonox worth two,
>opiates / epidurals worth three / four / whatever?
>I appreciate that this would be a very crude measure, but is it feasible?
>(I'm fully expecting some howls of anguish)
>Many thanks.
>Andrew Symon
>
>Dr. Andrew Symon
>Senior Lecturer
>School of Nursing & Midwifery
>University of Dundee
---------------------------------
Yahoo! for Good
Click here to donate to the Hurricane Katrina relief effort.
|