 |
 |
Toby Facinating stuff about curves! I wonder if there is some kind of a cut off point before which we feel confident that we know what we're doing and after which we freely admit we have to summon up some kind of help - and does it vary from GP to GP? Sets me thinking of other ways this kind of information could be amassed. Would be very interested to participate in this once it gets clearer what we're actually trying to do. John Williams > -----Original Message----- > From: [log in to unmask] > [mailto:[log in to unmask]]On Behalf Of Toby > Lipman 7, Collingwood Terrace, Jesmond, Newcastle upon Tyne. Tel > 0191-2811060 (home), 0191-2869178 (surgery) > Sent: 08 October 1998 15:18 > To: Ati Yates > Cc: [log in to unmask]; [log in to unmask] > Subject: Re: GP questions > > > In message <[log in to unmask]>, Ati Yates > <[log in to unmask]> writes > >Great idea, Martin! You can definitely sign me up for the effort. > > A belated me too! > > I > >think we might sort out early whether we want to look at presenting > >complaints (congestion, cough, stomach ache) which require diagnosis or > >diagnoses we tend to quickly make which require treatment (and then > >"retrospectively" information about the best diagnostic > methods), or both. > >I think I'd favor both since both show up in the clinics. > > Yes I like that - I've always thought that working around evidence on > therapy has predominated too much for primary care's needs. Of course > what we need is good evidence on the diagnostic power of common symptoms > and signs - much more than we need to be able to work out the benefits > of statins! > > > It would also be > >interesting to see them just how they come in--because some > symptoms might > >come in as frequently as other diagnoses, and people see a mix of both. > > > >I also think it will make sense to have people do this, say at > least 3 or 4 > >times, just to get closer to the average picture for each. Is > there a way > >to predict how many times will give you a reasonable picture if > there are, > >hmmm.....25 most common conditions as people sometimes say? > > > You *could* say there are 25 common conditions but it depends on your > definition of common (and common is not the same as important). When I > looked at 413 presentations of problems to GPs I found that the top ten > commonest problems presented in roughly a third of consultations, 26 > further problems presented in about another third, but there were no > fewer than 122 different problems in the remaining third. The difficulty > is that it's often the rarer problems which are more serious (or that we > are most ignorant about). Obviously EB guidelines would be useful for > patients in the top 10 (which included hypertension and depression but > also URTI). But what happens if someone asks for a PSA test because they > have a strong family history of prostate cancer, but no symptoms of > prostatism? Unless you are specifically interested in the PSA debate, > have some specialised knowledge (or just have a loads better memory than > me) you'll need to search for evidence to answer the question implied by > this patient's request (and it's no use just remembering reading about > it somewhere - was it in Bandolier? was it in EBM journal? really you > need access to all the usual databases, if not during the consultation, > at least before you leave the building) > > So yes I think this is a great project but I'd be fascinated to find out > how great the variety, number and frequency of problems is. Could I > suggest that we analyse them cumulatively and record not only the > frequency of problems, but also their number and variety. (how many > separate problems presented in the first 100, how many in the first 200, > the first 300 and so on). This would give us an idea of the true > information needs of primary care clinicians - and if plotted as a graph > (cumulative number of separate problems against numbers of patients > seen!) should show an initially steep rise followed by flattening out. > I'd guess, from my own work, that the curve would start to flatten at > about problem number 10 or 12 and would be really shallow after about > problem number 120 - but have no idea how long the shallow rise would > continue thereafter. > > Of course I'm aware that I've just suggested a different project from > what Martin originally intended...(!) - but wouldn't it be interesting > to find out also what proportion of our questions are about diagnosis cf > treatment? If we each record 40 consecutive questions we'd need at least > 25 clinicians to participate to get a sample of 1,000 > questions/problems. If the curve was still rising we could go back and > each collect a further 10 until it was flat. > > Cheers > > Toby > -- > Toby Lipman 7, Collingwood Terrace, Jesmond, Newcastle upon Tyne. Tel > 0191-2811060 (home), 0191-2437000 (surgery) > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%