Dear Chris,
Your unhappy patients have a familiar ring, and its great your researching
this. Thoroughly investigated series of patients with a particular
complaint often turn up no identifiable cause (my estimate from a few
series is about 30%). What is going on? Our ignorance is probably much
larger than most of us realize, and there is much to be discovered. The
method of inquiry will depend on the question. My current classification of
questions and appropriate study types is below (I'd be interested in
variants of this). I currently put your unhappy patients in category 1 - we
just don't know much about the phenomena. But when you've found out more
several of the other categories of question will become important (what
happens to these people? Can we alter the problem? etc) and then you'll
need larger numbers.
1. What are the phenomena/problems?
Observation (e.g., qualitative research)
2. What is frequency of the problem? (FREQUENCY)
Random (or consecutive) sample
3. Who has the problem? (DIAGNOSIS)
Random (or consecutive) sample with Gold Standard
4. Who will get the problem? (PREDICTION)
Follow-up of inception cohort
5. How can we alleviate the problem? (INTERVENTION)
Randomised controlled trial
Best wishes,
Paul Glasziou
At 03:47 PM 9/12/99 , Chris Ellis wrote:
>Mark Gabbay's question arises so often that I think we should perhaps
>readdress it as case histories and their significance, power, validation,
>place in the hierarchy of research and their extrapolatability (if I may
>use such a word) where discussed on this list about 18 months ago.
> This is also covered in some new textbooks which other members will be
>able to give to Mark.
>
>My interest is a parallel one of using small numbers in research (not the N
>of 1 thing) which has also be discussed in this forum before.
>Several years ago I completed a doctorate in family medicine\general
>practice. My thesis was that a few patients were coming in to see me who
>we as doctors were packaging in to boxes such as anxiety neurosis,
>depression, personality disorder, somatisation disorder etc etc. I felt
>they were unhappy so we researched the phenomenon of Unhappiness (ref.Can
>Fam Physician 1996:42:645-651).
> Only 4 patients in my practice fitted our selection criteria reasonably
>uniformly. The only appropriate research methodology to investigate this
>condition was the purely qualitative method of existential phenomenology.
>My promoters who included a professor of research psychology and a senior
>research philosopher spent over 200 hours over 3 years discussing and
>teaching me the method.
>Whilst we were doing the verification\validation\trustworthiness via the
>standard qualitative processes we were continually questioned by those we
>submitted it to for triangulation\explosion\free imaginative variation etc
>with the question, How can you do research with only 4 patients? (And what
>the dickens is existential phenomenology?).
>
>Which brings me, rather long windedly, to my point about small numbers in
>medical research and case histories (or clinical biographies or patient
>stories or several other similar texts that are used as research
>information).
>Parkinson described only 6 cases of his eponymous disease from his practice
>in Shoreditch in London in his original description (Parkinson J. An Essay
>on the Shaking Palsy. Sherwood, Neely and Jones, 1817).
>The Argyll Robertson Pupil was described initially in one case (Argyll
>Robertson D. On an interesting series of eye symptoms in a case of spinal
>disease. Edinburgh Medical Journal 1869; 14:696-708).
>(By the way if anyone else has examples of similar small series from
>practice could they let me know, as I am building up my case)
>And this brings me to my real and final point (at last) which is that I am
>keen on trying to keep general practice research within my practice. As
>soon as I leave it and go into the general practice unit at the medical
>school I seem to have the living reality of myself and my practice removed.
> My patient's relatives, the receptionist's background whispered
>information, my partner's comments in the tea room (the so called rich or
>thick texts) are removed and I am told I must have a trillion cases and
>something called statistical significance (I am numerically challenged).
>I am trying to write a post doctoral monograph which I have tentatively
>called :
>In Search of an Intrinsic General Practice Research Methodology.
>Could anyone help me with advice? This may have already been covered
>before.
>Yours
>Dr Chris Ellis
>29 Oriel Road, Pietermaritzburg, KwaZulu\Natal, South Africa.
>Fax 033-3868400. Tel : 033-3869208. E mail : [log in to unmask]
Paul Glasziou
Harvard Centre for Risk Analysis
718 Huntington Avenue, E-221
Boston, MA 02115
Fax: 1-617-432-0190 ph: 1-617-4320095
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|