Document Sample
Letters Powered By Docstoc


Cancellation of debt must ensure maximum benefit to vulnerable                                                           There is no reason why a lecturer should
                                                                                                                     not charge high fees to a rich parent drug
                                                                                                                     company or claim expenses commensurate
                                                                                                                     with those paid to other professionals, but the
Editor—It should be difficult for any doctor          United Nations should support these local                      situation in Thailand, and probably in other
to ignore Abbasi’s recent article on “third           delegates. A committee formed between the                      developing countries, is different from that in
world” debt.1 The state of health care                UN partners and delegates from the country                     the West. Visiting speakers may not under-
available to families in between half and two         repaying the debt should thereby ensure                        stand that their expenses are not reimbursed
thirds of the world is unacceptable, and the          that all the money that has been repaid to                     by a parent company in the West but come
great differences in quality of health care           rich countries is used for health care and                     out of the local division’s budget. High fees
between rich and poor countries are unethi-           education.                                                     have thus had important repercussions
cal.2 However, the experiences of our aid                  Such a system would be more difficult to                  during Thailand’s recent economic recession.
agency in trying to develop hospital care for         organise than simple cancellation of debt.                     Also, visitors may not understand that the
children in disadvantaged countries leads us          Care would have to be taken to ensure that                     nature of the social hierarchy in Thailand is
to be cautious about advocating simply a              existing levels of finance for health care and                 such that people of lower seniority would not
cancellation of debts. We have witnessed that         education are maintained in addition to                        ask before a visit what the fees and expenses
an input of money alone to countries which            money resulting from debt repayment. The                       of a visitor are likely to be, nor can they ques-
are poorly governed, and where corruption             repayment of debts by the “third world” to                     tion large fees or expenses, even when they
is a way of life, may fail to reach the most          rich countries who have prospered from                         seem excessive.
vulnerable and needy within the community.            earlier exploitation clearly represents an                         The payment of high charges for speak-
     Our agency’s view is that all debt to poor       evil. However, our response to this should be                  ing fees uses funds that would otherwise be
and disadvantaged countries should be can-            “street wise” and should ensure that the abo-                  spent on education and training. The £1000
celled but in a more sustainable way than by          lition of debt results in maximum benefit to                   spent on a visiting speaker, for instance,
releasing states from their repayments.               the most vulnerable members of disadvan-                       could cover the costs of sponsoring a
Money earmarked for debt repayment                    taged countries.                                               resident for a three year training pro-
should be retrieved but ploughed back fully                                                                          gramme or could cover the total cost of a
                                                      David Southall honorary director, Child Advocacy
and immediately into the country’s health-            International                                                  Thai royal college’s annual meetings.
care and education systems. We suggest that           Academic Department of Paediatrics, North Staffs                   We suspect that these problems are not
this is implemented through the appoint-              Hospital (City General), Stoke on Trent,                       confined to Thailand, British lecturers, or
ment in each repaying country of carefully            Staffordshire ST4 6QG
                                                                                      any particular specialty, but we hope that by
chosen advocates for women and children.                                                                             understanding the implications of high
Ideally, partners appointed from within the                                                                          charges, speakers will be more considerate
                                                      1 Abbasi K. Free the slaves. BMJ 1999;318:1568-9. (12 June.)   in making their claims. In the meantime, it is
                                                      2 Sogan D, Bridel J, Arzomund M, Shepherd C, Southall D.
                                                        21st century health care for children in Afghanistan.        possible that colleagues in some poorer
Advice to authors                                       Pediatrics 1998;102:1193-8.                                  countries will no longer invite lecturers who
We prefer to receive all responses electronically,                                                                   ask for large fees and expenses. We
sent either directly to our website or to the                                                                        recommend that, however embarrassing it
editorial office as email or on a disk. Processing    Developing countries cannot                                    may be for them, these colleagues make it
your letter will be delayed unless it arrives in an                                                                  clear before the visit that only reasonable,
                                                      afford high fees for speakers                                  affordable expenses may be claimed.
electronic form.
     We are now posting all direct submissions to     Editor—Doctors visit Thailand to give                          Teerakiat Jargensettasin vice dean
our website within 24 hours of receipt and our        lectures at the invitation of Thai universities                Guy Edwards visiting professor
                                                                                                                     Faculty of Medicine, Khon Kaen University,
intention is to post all other electronic             or royal colleges or are invited by drug firms.                Khon Kaen 40002, Thailand
submissions there as well. All responses will be      Some pay all of their own expenses, or only
                                                                                                                     Competing interests: None declared.
eligible for publication in the paper journal.        claim minimal expenses, and give gener-
     Responses should be under 400 words and          ously of their time. These doctors contribute
relate to articles published in the preceding         much to the educational programme in
month. They should include <5 references, in the      Thailand and to good international rela-                       Antidepressants for old people
Vancouver style, including one to the BMJ article     tions. However, for the reasons illustrated by
to which they relate. We welcome illustrations.       the following examples, our Thai colleagues                    GPs should become familiar with one or
     Please supply each author’s current              have expressed concern about some of                           two antidepressants from each class
appointment and full address, and a phone or          those lecturers who are invited by Thai divi-                  Editor—Livingston and Livingston cast a
fax number or email address for the                   sions of international drug firms.                             sceptical eye over the newer antidepressants
corresponding author. We ask authors to declare            After lecturing in Bangkok one doctor                     in relation to the treatment needs of older
any competing interest. Please send a stamped         submitted a claim to a local division for two                  people with depression.1 Their editorial
addressed envelope if you would like to know          first class air tickets (one of which was for his              contained omissions and inaccuracies. They
whether your letter has been accepted or rejected.    wife) costing a total of £6000 ($9600).                        state that subjects aged over 64 with physical
     Letters will be edited and may be shortened.     Another speaker asked for an honorarium                        disease were excluded from “all the trials                                           of £2000 for his one hour talk, refused to be                  cited” in the editorial yet quote research                                       paid by cheque, and asked for a cash                           (their reference 5) that addressed precisely
                                                      payment.                                                       this group.

BMJ VOLUME 319      25 SEPTEMBER 1999                                                                                                            849

     Likewise, they claim that fluoxetine is the                the ratio of the total number who stopped                 3 Martin RM, Hilton SR, Kerry SM, Richards NM. General
                                                                                                                            practitioners’ perceptions of the tolerability of antidepres-
“only newer antidepressant that has been                        taking antidepressants to the number who                    sant drugs: a comparison of selective serotonin reuptake
evaluated clinically in depressed patients                      started them was 22% for selective serotonin                inhibitors     and     tricyclic   antidepressants.      BMJ
with organic brain disease” but cite two                        reuptake inhibitors and 33% for tricyclic                 4 MacDonald TM, McMahon AD, Reid IC, Fenton GW,
papers, one concerning citalopram and the                       antidepressants.3 This suggests that in                     McDevitt DG. Antidepressant drug use in primary care: a
other moclobemide, that have evaluated                          clinical practice selective serotonin reuptake              record linkage study in Tayside, Scotland. BMJ
depression in old people with dementia.                         inhibitors offer a considerable advantage in              5 Spigset O, Martensson B. Clinical review: Drug treatment
They overlook important sources of                              terms of tolerability, an advantage that is                 of depression. BMJ 1999;318:1188-91. (1 May.)
information—for example, the comprehen-                         minimised in the highly standardised envi-
sive report of the National Institutes of                       ronment of clinical trials.                               Lack of evidence of efficacy is not
Health in 1991, recently updated2—and they                           Although antidepressants have similar                evidence of lack of efficacy
lump together all older people, when clearly                    efficacy in trials, this will occur in clinical           Editor—Livingston and Livingston’s edito-
there are important differences between a fit                   practice only if patients take a therapeutic              rial regarding the selection of antidepres-
65 year old and a frail 85 year old.                            dose. Studies in primary care consistently                sants for elderly people rightly points out
     The debate they are trying to promote is                   show that patients are more likely to be pre-             the dearth of reliable evidence on the
a tired one. Older people are prone to side                     scribed a recommended therapeutic dose of                 comparative efficacy and tolerability of vari-
effects and have contraindications to a large                   a selective serotonin reuptake inhibitor than             ous classes of antidepressants in this age
number of drugs (including some older                           of a tricyclic antidepressant. This difference            group.1 The authors conclude that it is hard
tricyclic antidepressants), so it makes sense                   remains when allowance is made for the fact               to recommend newer drugs on safety
to have a choice; that is precisely what the                    that elderly people often respond to doses of             grounds other than to those at risk of
newer antidepressants offer to older                            tricyclic antidepressants below those recom-              overdosing. They base this conclusion on a
patients. Having a choice improves out-                         mended in younger adults.                                 lack of evidence of better tolerability of
come. In one study of older people, sequen-                          The authors highlight the greater toxic-             newer compared with older antidepressants.
tial antidepressant regimens that used                          ity of tricyclic antidepressants in overdose.                 The authors cite a meta-analysis that
different classes of antidepressants resulted                   Another aspect of safety that deserves                    compared dropout rates with tricyclic
in a recovery rate of over 80%.3 The newer                      consideration is side effects. Elderly people             antidepressants and selective serotonin
antidepressants are here to stay. The impor-                    are particularly prone to develop these                   reuptake inhibitors in clinical trials, which
tant questions are not whether they should                      because of concurrent illness, drug interac-              suggested no significant difference between
be used at all but their place alongside                        tions, and the changes of normal ageing.                  these classes of drugs.2 Extrapolating drop-
psychological approaches in cases resistant                     Compared with newer agents the tricyclic                  out rates in clinical trials to either tolerability
to treatment with a first line antidepressant                   antidepressants have more potential to                    or compliance in clinical practice is, how-
and in the prevention of recurrence.                            cause serious side effects due to their                   ever, difficult, and a more recent meta-
     A more helpful message would be to                         non-selective action; 1 blockade can cause                analysis has disputed these findings.3
encourage practitioners who treat older                         postural hypotension, leading to falls and                    Perhaps of more relevance to clinical
patients to become familiar with one or two                     fractures; muscarinic blockade can cause                  practice is an examination of general practi-
antidepressants from each class, as recom-                      urinary retention, impaired cognitive func-               tioners’ prescribing habits for antidepres-
mended by the Royal Colleges of Psychia-                        tioning, and delirium; the membrane stabil-               sants; this showed that the ratio of dropping
trists and General Practitioners.4 It is                        ising effect can cause cardiac arrhythmias. A             out of treatment to starting treatment was
unnecessary to pit one class of antidepres-                     record linkage study showed a significantly               50% higher for tricyclic antidepressants
sants against another: there is room for                        increased risk of admission with acute                    than for selective serotonin reuptake inhibi-
both.                                                           urinary retention among patients pre-                     tors.4 How much of this difference is due to
R C Baldwin consultant old age psychiatrist                     scribed tricyclic antidepressants compared                tolerability problems in elderly people is
Manchester Royal Infirmary, Manchester M13 9BX                  with those prescribed selective serotonin                 unknown.                                  reuptake inhibitors.4 The number of contra-                   A lack of evidence of efficacy is not the
Competing interests: None declared.                             indications to and cautions against use is                same as evidence of a lack of efficacy. This is
                                                                greater for tricyclic antidepressants than for            also the case with regard to tolerability. Prac-
1 Livingston MG, Livingston HM. New antidepressants for         selective serotonin reuptake inhibitors.                  tice based randomised clinical trials are the
  old people? BMJ 1999;318:1640-1. (19 June.)
2 Lebowitz BD, Pearson JL, Schneider LS, Reynolds CF 3rd,
                                                                     A recent review concluded that newer                 gold standard. In the absence of a sound evi-
  Alexopoulos GS, Bruce ML, et al. Diagnosis and treatment      antidepressants were preferred in elderly                 dence base, however, it is surely best to act
  of depression in late life: consensus statement. JAMA         people, although previous response to anti-
                                                                                                                          on what evidence we have, coupled with our
3 Flint AJ, Rifat SL. The effect of sequential antidepressant   depressants may modify this recommen-                     clinical acumen. Clinical observations sug-
  treatment on geriatric depression. J Affect Disord            dation.5 Data are strongest for the selective             gest that patients of all ages generally
4 Katona C, Freeling P, Hinchcliffe K, Blanchard M, Wright      serotonin reuptake inhibitors. Tricyclic anti-            tolerate selective serotonin reuptake inhibi-
  A. Recognition and management of depression in late life      depressants are useful second line agents,                tors better than tricyclic antidepressants. Are
  in general practice: consensus statement. Primary Care
  Psychiatry 1995;1:107-13.
                                                                but those with high antimuscarinic activity               we to ignore these observations and deny
                                                                are best avoided. Irrespective of the anti-               elderly people drugs that are almost
Elderly people are particularly prone to                        depressant chosen, it is important when                   certainly easier for them to take while we
develop side effects                                            treating elderly patients to start with a low             wait for definitive studies to be conducted? If
                                                                dose, increase it gradually, and monitor for              so then Livingston and Livingston’s power
Editor—By discussing mainly data from
                                                                side effects.                                             calculations suggest that we may have a very
clinical trials Livingston and Livingston
                                                                Peter Haddad consultant psychiatrist                      long wait.
understate the advantages of the newer anti-
                                                                Moorside Unit, Trafford General Hospital,                 R H McAllister-Williams MRC clinical scientist
depressants in elderly people.1 They cite                       Manchester M41 5SL                                        Department of Psychiatry, University of Newcastle
Song et al’s meta-analysis of randomised                                                                                  upon Tyne, Royal Victoria Infirmary, Newcastle
controlled trials which found no significant                    Competing interests: Dr Haddad has received               upon Tyne NE1 4LP
                                                                lecture fees and conference expenses from the
difference in total dropout rates between tri-                                                                  
                                                                manufacturers of several antidepressants, including
cyclic antidepressants and selective sero-                      selective serotonin reuptake inhibitors.                  Competing interests: Dr McAllister-Williams has
tonin reuptake inhibitors.                                                                                                received lecture fees, conference expenses, and small
     A more recent meta-analysis found a                                                                                  sums of money to support research from the manu-
                                                                1 Livingston MG, Livingston HM. New antidepressants for   facturers of several antidepressants, including selec-
small but significant difference in total drop-                   old people? BMJ 1999;318:1640-1. (19 June.)             tive serotonin reuptake inhibitors.
out rates in favour of selective serotonin                      2 Anderson IM, Tomenson BM. Treatment discontinuation
                                                                  with selective serotonin reuptake inhibitors compared
reuptake inhibitors.2 An observational study                      with tricyclic antidepressants: a meta-analysis. BMJ    1 Livingston MG, Livingston HM. New antidepressants for
in primary care showed a larger difference:                       1995;310:1433-8.                                          old people? BMJ 1999;318:1640-1. (19 June.)

850                                                                                                                   BMJ VOLUME 319         25 SEPTEMBER 1999 

2 Song F, Freemantle N, Sheldon TA, House A, Watson P,            ine in elderly physically ill people who                        trials comparing low molecular weight
  Long A, et al. Selective serotonin reuptake inhibitors:
  meta-analysis of efficacy and acceptability. BMJ                are depressed1 and citalopram and                               heparins and unfractionated heparin in the
  1993;306:683-7.                                                 moclobemide in depressed people with                            treatment of venous thromboembolism.
3 Anderson IM. SSRIs versus tricyclic antidepressants in
  depressed inpatients: a meta-analysis of efficacy and toler-
                                                                  dementia.2 3                                                    Reduced mortality in patients with venous
  ability. Depression Anxiety 1998;7(suppl 1):11-7.                    The message of getting to know a few                       thromboembolism who were treated with
4 Martin RM, Hilton SR, Kerry SM, Richards NM. General            drugs well is sound, but for us the question is
  practitioners’ perceptions of the tolerability of antidepres-                                                                   low molecular weight heparins seemed to be
  sant drugs: a comparison of selective serotonin reuptake        which ones. Haddad would exhort us to                           confined to patients with cancer.2
  inhibitors      and     tricyclic   antidepressants.     BMJ    consider wider sources of information than                           In a meta-analysis of randomised clinical
                                                                  randomised controlled trials on prescribing                     trials (1723 subjects overall: 848 in the low
                                                                  in elderly people. We agree, but treatments                     molecular weight heparin group and 875 in
Research on antidepressants in the                                are available for depression that have
                                                                                                                                  the unfractionated heparin group), we
elderly population is scarce                                      already undergone clinical trial in this
                                                                                                                                  evaluated mortality in the first 15 days of
                                                                  population. We cite the very few studies of
Editor—Livingston and Livingston are                                                                                              treatment, when a true effect of low molecu-
                                                                  the newer drugs, and surely prescribers (and
right to draw our attention to the paucity of                                                                                     lar weight heparins would be expected, and
                                                                  consumers) would want new alternatives to
research on antidepressants in the elderly                                                                                        in the three months of subsequent oral anti-
                                                                  be submitted similarly to randomised con-
population.1 At least five studies comparing                                                                                      coagulation.3 After a follow up of three
                                                                  trolled trials.
tricyclic antidepressants and selective sero-                                                                                     months, overall mortality was 3.3% in the
                                                                       McAllister-Williams is correct to encour-
tonin reuptake inhibitors in elderly people                                                                                       low molecular weight heparin group and
                                                                  age doctors to do what they can in the clini-
have shown no difference. In all of these                                                                                         5.9% in the unfractionated heparin group
                                                                  cal context when faced with an elderly
comparison studies relatively low doses of
                                                                  person who is depressed, but when we are                        (relative risk 0.51, 95% confidence interval
tricyclic antidepressants were used, which
                                                                  driven to choose a treatment that has not                       0.2 to 0.9, P < 0.01). Although there was no
makes meaningful conclusions difficult to
                                                                  been subjected to a rigorous randomised                         significant difference in overall mortality in
draw, except perhaps that a low dose tricyclic
                                                                  controlled trial our treatment is founded on                    the first 15 days, the difference during oral
antidepressant may be equivalent to a stand-
                                                                  less firm ground. We do not assert that the                     anticoagulation was significant (2.5% v 4.5%;
ard dose serotonin reuptake inhibitor.
                                                                  newer antidepressants are less effective or                     0.48, 0.2 to 0.8, P < 0.03). In a separate
Another large study of selective serotonin
                                                                  less well tolerated than tricyclic antidepres-                  analysis in patients with a diagnosis of
reuptake inhibitors in elderly people
                                                                  sants in the over 65s, simply that claims for                   cancer at the time of initial diagnosis of
showed final mean Hamilton depression
                                                                  efficacy and tolerance have to be justified, as                 venous thromboembolism, 1.3% of patients
scores of 16 (not consistent with full
                                                                  with younger populations.                                       in the low molecular weight heparin group
                                                                       Gordon would seem to take our                              died during the first 15 days of treatment
     The risks versus the benefits of prescrib-
                                                                  position that, in the words of our editorial’s                  compared with 2.5% in the unfractionated
ing newer antidepressants in elderly people
                                                                  subtitle, the evidence of efficacy and                          heparin group (0.5, 0.01 to 7.7, P = 0.52).
depend on the person concerned. Although
                                                                  tolerability for many new antidepressants in
tricyclic antidepressants should be withheld                                                                                      The difference in mortality in patients with
                                                                  elderly people is thin. We do not see our
in people at high risk of suicide, death by                                                                                       cancer during the 16-90 day follow up
                                                                  editorial as promoting a tired debate (Bald-
overdose of tricyclics is relatively rare and                                                                                     period of oral anticoagulation was signifi-
                                                                  win) about tricyclics versus selective serot-
there is some evidence that these drugs may                                                                                       cant: 12% in the low molecular weight
                                                                  onin reuptake inhibitors and other new
be more efficacious at treating melancholia.3                                                                                     heparin group and 26% in the unfraction-
                                                                  antidepressant drugs. It is, in our view, a
     The major difficulty in using tricyclic                                                                                      ated heparin group (0.33, 0.1 to 0.8,
                                                                  reminder that depressed old people, just
antidepressants in elderly people is ortho-                                                                                       P < 0.01). During the combined period of
                                                                  like their younger counterparts, are entitled
static hypotension, which can occur at blood                                                                                      heparin and oral anticoagulation therapy,
                                                                  to receive treatment based on evidence and
concentrations well below therapeutic con-                                                                                        mortality was 14% in patients with cancer
                                                                  that more studies need to be carried out on
centrations. This effect, however, is probably
                                                                  antidepressant drug treatments in this                          treated with low molecular weight heparin
less a function of age than of increasing
                                                                  important population.                                           and 28% in those treated with unfraction-
illness and treatment with multiple drugs.
                                                                  Martin G Livingston consultant psychiatrist                     ated heparin (0.33, 0.1 to 0.8, P < 0.01). Mor-
On this evidence, tricyclic antidepressants
                                                                  Hilary M Livingston consultant psychiatrist                     tality in patients without cancer was not
remain the treatment of choice in elderly
                                                                  Southern General Hospital, Glasgow G51 4TF                      significantly different between the groups
patients with uncomplicated depression.                 
                                                                                                                                  (1.9% v 2.6%; 0.72, 0.3 to 0.6, P = 0.4).
Catherine Gordon specialist registrar in psychiatry               Competing interests: None declared.
Queens Medical Centre, Nottingham NG7 2UH                                                                                              Our observation does not support the                                                                                                               hypothesis that the reduction in death in
                                                                  1 Evans M, Hammond M, Wilson K, Lye M, Copeland J.
Competing interests: None declared.                                 Placebo controlled treatment trial of depression in elderly   patients with cancer is due to an effect of low
                                                                    physically ill patients. Int J Geriatr Psychiatry 1997;12:    molecular weight heparins in preventing
1 Livingston M, Livingston H. New antidepressants for old         2 Nyth AL, Gottfries CG, Lyby K, Smedegaard-Andersen L,         thromboembolism. No difference in mor-
  people? BMJ 1999;318:1640-1. (19 June.)                           Gylding-Sabroe J, Kristensen M, et al. A controlled           tality from all causes was observed in
2 Tollefson GD, Holman SL. Analysis of the Hamilton                 multicenter trial of citalopram and placebo in elderly
  depression rating scale factors from a double-blind,              depressed patients with and without concomitant demen-        patients without cancer. The cause of
  placebo-controlled trial of fluoxetine in geriatric major         tia. Acta Psychiatr Scand 1992;86:138-45.                     reduced mortality in cancer patients treated
  depression. Int Clin Psychopharmacol 1993;8:253-9.              3 Roth M, Mountjoy CQ, Amrein R. Moclobemide in elderly
3 Roose SP, Glassman AH, Attica E, Woodring S. Compara-             patients with cognitive decline and depression; an
                                                                                                                                  with low molecular weight heparins is there-
  tive efficacy of selective serotonin reuptake inhibitors and      international double blind placebo controlled trial. Br J     fore difficult to explain.
  tricyclics in the treatment of melancholia. Am J Psychiatry       Psychiatry 1996;168:149-57.
  1994;151:1735-9.                                                                                                                Sergio Siragusa haematologist and angiologist
                                                                                                                                  Istituto di Ricovero e Cura a Carattere Scientifico,
                                                                                                                                  Policlinico S Matteo, Pavia, Italy 27100
Authors’ reply                                                    Low molecular weight heparins                         
Editor—In our editorial we were concerned                         could be important in cancer
with the primary evidence of efficacy and                                                                                         1 Kakkar AK, Williamson RCN. Antithrombotic therapy in
safety of newer antidepressants in elderly                        Editor—Kakkar and Williamson clearly                              cancer. BMJ 1999;318:1571-2. (12 June.)
people, and therefore we focused mainly on                        state the need for well conducted prospec-                      2 Green D, Hull RD, Brant R, Pineo GF. Lower mortality in
                                                                                                                                    cancer patients treated with low-molecular weight heparin
randomised controlled trials rather than on                       tive clinical trials to evaluate the potential                    versus standard heparin. Lancet 1992;339:1476.
meta-analyses or reviews even from august                         role of low molecular weight heparins in                        3 Siragusa S, Cosmi B, Piovella F, Hirsh J, Ginsberg JS. Low-
bodies such as the National Institutes of                         improving survival in cancer patients.1                           molecular weight heparins and unfractionated heparin in
                                                                                                                                    the treatment of patients with acute venous thrombo-
Health. Baldwin is correct to point out that                           In the absence of properly designed                          embolism: results of a meta-analysis. Am J Med 1996;100:
we cite trials that support the use of fluoxet-                   trials, indirect data can be extrapolated from                    269-77.

BMJ VOLUME 319           25 SEPTEMBER 1999                                                                                                                          851

Hypoalbuminaemia and                                          prolong the action potential in vitro; and                    seemed to concentrate on assimilating both
                                                              that mibefradil (again unlike those drugs                     sides of the case with a view to proceeding to
transcapillary pressures have                                 mentioned above) does not induce malig-                       a service hearing.
role in nephrotic syndrome                                    nant arrhythmias in animal experiments.                           In another case, not related to my
                                                              Notwithstanding the doubts expressed dur-                     practice, the same health authority pro-
Editor—In their article on understanding
                                                              ing the discussion, the committee recom-                      ceeded with a late complaint against a fund-
oedema Diskin et al have overlooked an
                                                              mended that mibefradil should be approved                     holding practice in which there had been an
important body of work that challenges the
                                                              for marketing.                                                allegation of clinical mismanagement due to
traditional pathophysiological explanation
                                                                   During the time that mibefradil was                      a delayed referral. The general practitioner
for the development of oedema in the
                                                              marketed, there were no reported cases of                     was accused of putting the financial fund-
nephrotic syndrome secondary to renal dis-
                                                              ventricular arrhythmia associated with                        holding interests of the practice above clini-
ease.1 Animal as well as several human stud-
                                                              mibefradil in patients who did not have                       cal need. The health authority invoked a
ies have shown that sodium retention seems
                                                              other risk factors and who were not                           fundholding        complaints       procedure,
to occur as a primary phenomenon in the
                                                              receiving concomitant treatment with pro-                     although it had no remit under the NHS
nephrotic syndrome as a result of increased
                                                              arrhythmic drugs. Here again, mibefradfil                     regulations to proceed along these lines.
sodium reabsorption in the collecting
                                                              differs from most drugs that are recognised                   Over several months it persistently
duct.2–4 This in turn may be due to resistance
                                                              to be proarrhythmic.                                          attempted to facilitate an informal meeting
to atrial natriuretic peptide.5
                                                                   Mibefradil was ultimately removed from                   between the parties and refused to accept
    Hypoalbuminaemia and changes in
                                                              the market because of its ability to derange                  that there was no procedure or regulation
transcapillary pressures, as well as volume
                                                              the metabolism of numerous more essential                     laid down for such cases. Finally, only the
regulating hormones and sodium retention,
                                                              drugs. This problem was anticipated by one                    threat of legal action by the practice
are mechanisms that do have a role in neph-
                                                              member of the Cardio-Renal Advisory                           persuaded the authority to drop the case.
rotic, oedematous patients. Whether sodium
                                                              Committee and by at least one regulatory                          Facilitation by the health authority tends
retention is the dominant mechanism for
                                                              agency (in Sweden), but no similar finding                    to delay proceedings and allow the parties to
the development of oedema remains uncer-
                                                              had ever been the basis for not approving or                  reach unchallenged conclusions that quickly
tain. For clinicians it is not the oedema but
                                                              for withdrawing a drug in the United States,                  become entrenched and lead to a service
the assessment of intravascular volume
                                                              and the potential impact of the interaction                   hearing. In-house mediation with an early
status (using clinical and laboratory data) in
                                                              of these drugs was not adequately appreci-                    exchange of views has helped to limit the
children with the nephrotic syndrome that
                                                              ated by other members of the committee,                       number of hearings mediated by the health
remains an inexact science.
                                                              the sponsor, the Food and Drug Administra-                    authority and has greatly reduced the
Nadeem E Moghal consultant paediatric nephrologist            tion, or (to my knowledge) other regulatory                   accompanying stress to general practition-
Department of Paediatric Nephrology, Royal
Victoria Infirmary, Newcastle upon Tyne NE1 4LP               agencies. All concerned regret what hap-                      ers brought on by well meaning facilitators.                                    pened, but we do not believe that our hind-                   Peter Thomson general practitioner
                                                              sight is less acute than that of Landow.                      Greyswood Practice, London SW16 6NT
Competing interest: None declared.
                                                              Robert R Fenichel deputy division director
                                                              Division of Cardio-Renal Drug Products, Food and              1 Jain A, Ogden J. General practitioners’ experiences of
1 Diskin CJ, Stokes TJ, Dansby LM, Carter TB, Radcliff L,     Drug Administration, HFD-110, 5600 Fishers Lane,                patients’ complaints: qualitative study. BMJ 1999;318:
  Thomas SG. Towards an understanding of oedema. BMJ          Rockville, MD 20857, USA                                        1596-9. (12 June.)
  1999;318:1610-3. (12 June.)
2 Ichikawa I, Rennke HG, Hoyer JR, Badr KF, Schor N, Troy
  JL, et al. Role for intrarenal mechanisms in the impaired                                                                 Mentoring should be more widespread
  salt excretion of experimental nephrotic syndrome. J Clin
  Invest 1983;71:91-103.
                                                              1 Landow L. FDA approves drugs even when experts on its       Editor—Jain and Ogden’s study of general
                                                                advisory panels raise safety questions. BMJ 1999;318:944.
3 Geers AB, Koomans HA, Roos JC, Boer P, Mees EJD.              (3 April.)                                                  practitioners’ experiences of patients’ com-
  Functional relationships in the nephrotic syndrome.                                                                       plaints and Baker’s editorial on learning
  Kidney Int 1984;26:324-30.
4 Bohlin AB, Berg U. Renal sodium handling in minimal                                                                       from complaints both mention the need for
  change nephrotic syndrome. Arch Dis Child 1984;59:825-                                                                    support for doctors about whom complaints
  30.                                                         General practitioners’                                        are made.1 2 It seems unfortunate that
5 Perico N, Remuzzi G. Edema of the nephrotic syndrome:
  the role of the atrial peptide system. Am J Kidney Dis      experiences of patients’                                      support is not universally available—a
                                                              complaints                                                    feeling heightened when we read quotes
                                                                                                                            from some doctors who participated in the
                                                                                                                            study, who clearly had not come to any satis-
Mibefradil was not rushed to                                  In-house mediation can help in reducing                       factory resolution after the complaint. Reso-
                                                              general practitioners’ stress                                 lution, the word used by the authors for the
market in United States
                                                              Editor—I work as a principal in general                       third phase of response to a complaint,
Editor—Landow gave a misleading impres-                       practice, with part of my list in Lambeth, and                implies a good outcome, but for some
sion of the discussion about mibefradil at                    was summoned to a service hearing by                          general practitioners considerable distress
the US Food and Drug Administration’s                         Lambeth, Southwark and Lewisham Health                        or defensive working patterns persisted.
Cardio-Renal Advisory Committee meeting                       Authority before 1996. I was therefore                            Neither article mentions what form sup-
of 28 February 1997 and the subsequent                        particularly interested in Jain and Ogden’s                   port might take. Support for general practi-
withdrawal of the drug from the market in                     article about general practitioners’ experi-                  tioners is developing in various forms
the United States.1                                           ences of complaints.1                                         around the United Kingdom, mainly as
    The question most extensively discussed                       Although I appreciate the impartiality                    counselling services available in times of dif-
at the meeting was whether mibefradil                         shown by the health authority, I would ques-                  ficulty or crisis. But why wait for a crisis? We
prolongs the QT interval and is therefore                     tion whether its facilitator function is                      exhort patients to live health promoting life-
likely to induce malignant ventricular                        helpful. In my experience the authority’s                     styles and regularly have our cars serviced to
arrhythmia. Mibefradil does induce electro-                   role largely entailed gathering non-critical                  lessen the likelihood of malfunction, but we
cardiographic changes that look disturbing,                   information; it failed to analyse the quality                 don’t extend the philosophy of maintenance
but the committee (and later the Food and                     of the information collected. My summons                      and prevention to ourselves.
Drug Administration) was impressed by the                     to a service hearing seemed to be deter-                          Mentoring and co-tutoring are ways of
fact that similar changes can be induced by                   mined more by the acrimony of the                             providing health promotion for doctors3 4;
verapamil and diltiazem; that mibefradil                      complainant than by the case itself. The                      meeting regularly away from work with a
(unlike quinidine, terfenadine, cisapride,                    health authority seemed unable or unwilling                   trained colleague, one has the focused and
astemizole, and other well known drugs that                   to come to a decision itself about the merits                 sympathetic attention of a skilled person,
induce torsades de pointes) does not                          of the allegations, and from an early stage it                with confidentiality, to allow reflection on

852                                                                                                                     BMJ VOLUME 319       25 SEPTEMBER 1999

any aspect of work, difficulties experienced                        I approached every prisoner identified                     acute care—that is, structured long term
and opportunities arising, educational                         to me who was available (15 in total) and                       care.
issues, and professional development. This                     interviewed each in turn in private for 10                          Campbell et al showed that nurse run
may not only lead to better management of                      minutes. I was explicit about the purpose of                    secondary prevention clinics in primary care
difficulties such as complaints but even help                  the interview and the limits of confiden-                       can reduce risk factors for the recurrence of
to prevent them.                                               tiality; all were willing to speak to me. Two                   coronary events.2 The widespread adoption
     Doctors who take time to reflect are                      thirds showed clear signs of mental illness,                    of such clinics will need the same amount of
more aware of what they find difficult, of                     most being floridly psychotic. Subsequent                       funding as that currently provided for the
their thoughts and feelings about interac-                     inspection of prison medical records                            management of asthma and diabetes.
tions with patients; this awareness may lead                   showed that prison healthcare staff were not                        Carrick Primary Care Group are to be
to improvements in quality of work and                         aware of any important mental health prob-                      congratulated for doing what the govern-
healthier responses to upsets. There is inter-                 lems in seven cases, two of the prisoners                       ment should have done when it initiated
est in mentoring from general practitioners;                   were thought to have schizophrenia in                           payment for some management of chronic
“regular review of work with a mentor” was                     remission, and one of the prisoners was                         diseases in general practice. All primary care
thought to be appropriate by a quarter of                      deemed to have a personality disorder. One                      groups should be persuaded to follow this
general practitioners surveyed about their                     man posed such a risk to his cellmate by vir-                   example, if necessary using innovative
support needs.5                                                tue of his delusional beliefs that I breached                   methods of joint funding.
     Mentoring and co-tutoring may also                        confidentiality and informed one of the                         H M Dalal general practitioner
help doctors to feel valued; being listened to                 prison medical officers.                                        Truro TR1 2LZ
does this to people, as we know from                                This was only a snapshot survey of two           
patients. Baker says that in the new culture,                  prisons. These findings suggest, however,                       Hugh Bethell chairman
of which clinical governance is part, “a high                  that prison officers can readily identify at                    Secondary Prevention and Rehabilitation
                                                                                                                               Committee, Coronary Prevention Group, London
value is placed on practitioners,” but he                      least some of the prisoners with severe                         WC1H 0BT
doesn’t say how this might be communi-                         hidden psychiatric morbidity. If individuals
cated. Mentoring supported by employers                        identified as having such morbidity under-                      1 Bradley F, Cupples ME. Reducing the risk of recurrent
might be one way.                                              went a more formal screen for mental health                       coronary heart disease. BMJ 1999;318:1499-500. (5 June.)
                                                                                                                               2 Campbell NC, Thain J, Deans HG, Ritchie LD, Rawles JM,
     Doctors who feel valued and have time                     problems a considerable number of men-                            Squair J. Secondary prevention clinics for coronary heart
to reflect and develop probably provide                        tally ill prisoners might get access to the                       disease: randomised trial of effect on health. BMJ
better quality services to patients; mentoring                 treatment they need.                                              1998;316:1434-7.

therefore has a contribution to make in the                    Luke Birmingham lecturer in forensic psychiatry
development of clinical governance.                            University of Birmingham, Reaside Clinic,
                                                               Birmingham B45 9BE                                              Is recruitment more difficult
Gail Young general practitioner
Newcastle upon Tyne NE4 9BB                                    1 Singleton N, Meltzer H, Gatward R, Coid J, Deasy D. Psychi-
                                                                                                                               with a placebo arm in RCTs?
                                                                 atric morbidity among prisoners in England and Wales.
                                                                 London: Stationery Office, 1998.
1 Jain A, Ogden O. General practitioners’ experiences of       2 Brooke D, Taylor C, Gunn J, Maden A. Point prevalence of      Methodological issues will have affected
  patients’ complaints: qualitative study. BMJ 1999;318:         mental disorder in unconvicted male prisoners in England
  1596-9. (12 June.)                                             and Wales. BMJ 1996;313:1524-7.                               results
2 Baker R. Learning from complaints about general              3 Gunn J, Maden A, Swinton M. Mentally disordered prisoners.
                                                                 London: Home Office, 1991.                                    Editor—Welton et al found that the
  practitioners. BMJ 1999;318:1567-8. (12 June.)
3 Freeman R. Towards effective mentoring in general            4 Birmingham L, Mason D, Grubin D. Prevalence of mental         proportion of women who were willing to
                                                                 disorder in remand prisoners: consecutive case study. BMJ
  practice. Br J Gen Pract 1997;47:457-60.
                                                                                                                               participate in a randomised controlled trial
4 Sackin P, Barnett M, Easthaugh A, Paxton P. Peer-
  supported learning. Br J Gen Pract 1997;47:67-8.             5 Birmingham L, Mason D, Grubin D. A follow-up study of         of hormone replacement therapy was
5 Young G, Spencer J. General practitioners’ views about the     mentally disordered men remanded to prison. Criminal          higher if participation in a trial without a
  need for a stress support service. Fam Pract 1996;13:          Behaviour and Mental Health 1998;8:202-13.
                                                                                                                               placebo arm was offered rather than partici-
                                                                                                                               pation in one with a placebo arm, although
                                                                                                                               the difference was of borderline signifi-
                                                               Reducing risk of recurrent                                      cance.1 Their conclusion was that inclusion
Prison officers can recognise                                  coronary heart disease in                                       of a placebo arm may reduce patients’
hidden psychiatric morbidity                                   Cornwall                                                        willingness to participate in a trial. We do
in prisoners                                                                                                                   not believe that this conclusion is justified.
                                                               Editor—Bradley and Cupples end their                                 Allocation of participants to either of
Editor—Prisons in England and Wales are                        editorial on reducing recurrent coronary                        the comparison groups was by the woman’s
known to house a considerable number of                        heart disease by asking whether anyone has                      choice of appointment time: the nurses
mentally ill people.1–3 Many of these people                   a better idea of how to improve perform-                        offered participation in trials with and with-
probably pass through health screening                         ance in this area1 Earlier they say that                        out placebo groups in alternate weeks. This
undetected when they are received into                         adequate resources would be needed to                           is not random allocation, in which only
prison and remain on ordinary location (on                     ensure a systematic approach to the second-                     chance can influence which group a partici-
prison wings) without ever coming to the                       ary prevention of coronary heart disease in                     pant is allocated to.2 In this study the nurses
attention of a doctor.4 5                                      primary care and that primary care groups                       could predict in which weeks they would be
     During two visits to large local male                     would be well placed to implement this.                         offering participation in trials with or
remand prisons as a guest medical inspector                    Such a scheme has begun in the Carrick                          without placebo groups and so could have
with the Inspectorate of Prisons I decided to                  Primary Care Group in Cornwall.                                 influenced the choice of appointment time.
see how difficult it was to identify hitherto                      For the next two years the British Heart                    This kind of bias would have been avoided
undetected mental illness in prisoners on                      Foundation will fund two liaison nurses to                      by randomising the weekly blocks, the inter-
ordinary prison location. I asked prison offic-                ease the discharge of coronary patients from                    vention for that week being revealed to the
ers to identify inmates on their landing whom                  the Royal Cornwall Hospital into the                            nurses only after all appointments for that
they considered to be odd, strange, or behav-                  community, as well as the training of a lead                    week were made.
iourally disturbed. The officers had no                        nurse from each practice. The primary care                           The nurses’ explanation of the reason
difficulty identifying between two and five (out               group will give extra monies to each of their                   for using a placebo could have influenced
of 35-40) inmates per landing on each of the                   13 practices for a scheme that includes                         willingness to participate. We would have
six landings I surveyed. The description of the                creating a register of patients with coronary                   been reassured that nurses’ explanations
behaviour of some of these prisoners strongly                  heart disease, starting a call and recall                       were consistent and neutral by knowing that
suggested that they were psychotic.                            system, and developing improved links with                      rates of willingness to participate across

BMJ VOLUME 319          25 SEPTEMBER 1999                                                                                                                      853

practices were similar. A lower rate of                             There is nothing in our paper to justify            aggregated by electoral ward of residence
willingness to participate in some practices                    not including a placebo when it is scientifi-           rather than by hospital of admission. The
would imply differences between nurses in                       cally necessary to do so. Rather, we hope that          aggregate discharge data would be the same
the way they described the reasons for inclu-                   our results will help researchers to set realis-        (except that true standardised admission
sion of a placebo—that is, that it is not the                   tic recruitment targets. We also drew                   rates could be used). The community data
placebo itself but the way in which it is                       attention to the importance to participants             could then be used directly (with numbers of
explained that influences the decision to                       of trial information indicating the potential           general practitioners per head calculated
participate. This would make the study’s                        benefits for others besides themselves.                 from family health services registers), but
conclusion invalid.                                             A J Welton health psychologist                          the hospital data would need to be cross
     The study shows only that the use of pla-                  M R Vickers senior scientist                            attributed. It would also then be possible to
cebos may influence rates of willingness to                     J A Cooper statistician                                 match hospital mortality data with commu-
                                                                T W Meade professor
participate, not actual participation rates, in                                                                         nity data on general mortality and life
                                                                Medical Research Council Epidemiology and
a trial of hormone replacement therapy. We                      Medical Care Unit, Wolfson Institute of Preventive      expectancy.
are concerned that this paper may be used                       Medicine, St Bartholomew’s and Royal London             Tom Hennell strategic analyst
to justify not including a placebo in                           Hospital School of Medicine and Dentistry,              NHS Executive North West, Warrington WA3 7QN
                                                                London ECIM 6BQ                               
randomised controlled trials. It is important
to include a placebo when the only                              T M Marteau professor
                                                                Psychology and Genetics Research Group, King’s          1 Jarman B, Gault S, Alves B, Hider A, Dolan S, Cook A, et al.
treatments available are of unknown efficacy                    College London, Guy’s Campus, London SE1 9RT              Explaining differences in English hospital death rates
or safety.3 The real challenge is to develop                                                                              using routinely collected data. BMJ 1999;318:1515-20. (5
ways of effectively explaining the reasons for                                                                            June.)
                                                                                                                        2 Welch WP, Miller ME, Welch HG, Fisher ES, Wennberg JE.
using a placebo to potential participants                                                                                 Geographic variation in expenditures for physicians’ serv-
when a placebo is ethically justified.                          Differences in death rates in                             ices in the United States. N Engl J Med 1993;328:621-7.
                                                                                                                        3 Health Care Financing Administration. Hospitalization and
Lindsay Forbes lecturer in public health medicine               English hospitals                                         mortality data for states. Washington, DC: Department of
Sue Chinn reader in medical statistics                                                                                    Health and Human Services, 1992. (Medicare hospital
Jose Figueroa-Munoz lecturer in public health                                                                             information report; 1992 technical supplement, section D.)
                                                                Effects of admission rates may have been                4 Manheim LM, Feinglass J, Shortell SM, Hughes EFX.
Paul Seed lecturer in medical statistics                        understated                                               Regional variation in Medicare hospital mortality. Inquiry
Department of Public Health Sciences, Guy’s, King’s             Editor—Unlike Jarman et al, I do not yet
and St Thomas’s School of Medicine, London
SE1 3QD                                                         think that we can state with confidence that                                        “more doctors means fewer deaths.”1 As the              Data are inadequate basis for drawing
                                                                authors make clear, there has been a lively             conclusion of paper
1 Welton AJ, Vickers MR, Cooper JA, Meade TW, Marteau           discussion on comparative hospital death                Editor—Efforts to compare the quality of
  TM. Is recruitment more difficult with a placebo arm in       rates in the United States. From this debate            care among hospitals have defeated many
  randomised controlled trials? A quasirandomised, inter-
  view based study. BMJ 1999;318:1114-7. (24 April.)            two points emerge clearly: for any given                investigators in the past. In the latest attempt
2 Altman DG, Bland JM. Treatment allocation in controlled       population the standardised admissions rate             Jarman et al try to explain differences in
  trial: why randomise? BMJ 1999;318:1209. (1 May.)
3 Pocock SJ. Clinical trials: a practical approach. New York:   is positively correlated with the standardised          English hospital death rates using routinely
  John Wiley, 1983.                                             death rate but is inversely correlated with the         collected data.1 Their main finding is an
                                                                standardised hospital death rate (defined as            inverse association between hospital death
Authors’ reply                                                  any death within 30 days of a hospital admis-           rates and the number of hospital doctors
Editor—Our conclusion that including a                          sion).2 3 Where a population is admitted to             and general practitioners, from which they
placebo arm may reduce willingness to take                      hospital fairly frequently a higher proportion          conclude that ratios of doctors to popula-
part in a trial was tentative and qualified. We                 of admissions will not be associated with sub-          tion “seem to be critical determinants of
drew explicit attention to the difference                       sequent death; hence there will be a lower              standardised hospital death rates; the higher
between expressed willingness and actual                        apparent hospital mortality. A study in Ohio            these ratios, the lower the death rates.” By
participation. So on these two points Forbes                    found African-Americans to have consist-                implication, and as widely interpreted by the
et al have simply repeated what we ourselves                    ently higher admissions and lower severity-             press, hospitals with fewer doctors have, as a
said.                                                           adjusted hospital mortality than white                  result, higher death rates. This may be true,
     Random allocation as suggested by                          patients treated at the same hospitals.4                but the data are an inadequate basis for
Forbes et al would have been ideal but was                            It was therefore informative that the             drawing that conclusion.
not practicable. Appointments were                              authors included with their explanatory                     The strength or weakness of such inves-
arranged and rearranged entirely according                      variables a calculation of the standardised             tigations hinges on the accuracy of the
to patients’ preferences and their own time-                    admission ratio (though not separated into              measurement of prognostic or risk factors in
tables and often less than a week in advance.                   “all cases” and “emergencies” for the two               patients treated by the hospitals being com-
To have insisted that all appointments were                     versions of their model). Unfortunately, the            pared. In this study the severity of the
made for each week before we allocated the                      figures used were the aggregates for the                primary illness could not be estimated and
trial group would have reduced flexibility to                   health authority of hospital location rather            comorbidity was limited to a count of sub-
make and reschedule appointments, thus                          than individual figures for each hospital’s             diagnoses for each patient in hospital
increasing the length and costs of the study.                   emergency catchment area (often very                    episode statistics. The inadequacy of such
There was no evidence to suggest that the                       different). Even in this form, standardised             routine data has been well documented by
method we used resulted in the biased selec-                    admissions entered significantly into both              the NHS Executive,2 Iezzoni et al3 (Iezzoni
tion of women to weeks during which one or                      the all cases and the emergency multiple                being one of the authors), and Jencks et al.4
other of the trials was being explained.                        regression models. It is not impossible that a          In addition, comorbidity, when measured
     Nurses were centrally trained before-                      fully specified version of this variable might          carefully, has an impact on mortality.5 The
hand, and their explanations of the two trials                  displace some of the doctor variables; the              dangers of using inadequate information
were according to standardised scripts and                      observed relation could simply be “More                 has been shown by Iezonni et al, who
monitored by the use of tape recordings of a                    doctors means more admissions.”                         reported the counterintuitive finding that
patient’s visit with each nurse. Any differ-                          One way to resolve this issue would be to         many subdiagnoses—for example, adult
ences between practices in expressed will-                      exploit a generally observed feature of the             onset diabetes mellitus, essential hyper-
ingness to take part in either trial will almost                NHS in England—that there is relatively                 tension, previous myocardial infarction,
certainly have been due to a variety of influ-                  little overlap of acute hospitals’ geographical         angina, and ventricular premature beats—
ences, such as the local awareness of issues                    catchment areas for adult emergency admis-              were associated with lower death rates.3
to do with hormone replacement therapy                          sions. Hence it should be possible to repeat            Much the same finding had previously been
and the prevalence of its use.                                  the authors’ analysis exactly but with data             reported by Jencks et al.4

854                                                                                                                  BMJ VOLUME 319       25 SEPTEMBER 1999 

     Severity of illness and comorbidity are                     Pawan Randev general practitioner                           investigators, any research that entails the
best judged prospectively, preferably in                         Measham Medical Unit, Measham, Near                         participation of many centres (as in any Euro-
                                                                 Swadlincote, Derbyshire DE12 7HR
“consultation with physicians.”3 Second best                                                 pean study by definition) and the collection of
is retrospective review of individual patients’                                                                              information on human subjects, whether by
                                                                 Competing interests: None declared.
records comparing condition specific diag-                                                                                   questionnaires, direct examination (treated or
noses and procedures case by case. The least                                                                                 observed), or consultation of already existing
satisfactory approach is to rely on routinely                                                                                records, usually demands a minimum of five
collected undifferentiated data, as in the                       Policy must separate need for                               years. This includes a minimum of one year to
study of Jarman et al. The authors are aware                     hospital beds from demand                                   activate an already formed international
of possible shortcomings of their approach,                                                                                  group of investigators, two to three years to
writing that “a matched pair study of                            Editor—Whether or not the private finance                   collect data, and two years to check, edit, and
patients admitted to hospitals with high and                     initiative has any real advantages (and I have              analyse data.
low standardised mortality ratios could help                     yet to meet a defender of it) is one question;                  What will be the consequences of grant-
to elucidate [their] findings.” Without such                     the appropriate number of hospital beds is                  ing three year contracts?
an investigation, the public, as well as the                     another.1 There is of course no right                           Firstly, funds may fall short of achieving
profession, is left to draw conclusions that                     number of hospital beds: if beds are there                  the aim of starting new research rather than
may be incorrect.                                                they will tend to be filled.                                just supporting the ongoing programme.
John P Bunker visiting professor
                                                                      Bed occupancy rates are relatively                     Secondly, projects will produce only prelimi-
Department of Epidemiology and Public Health,                    consistent (at about 80%) throughout                        nary results, often not publishable other
University College London School of Medicine,                    western Europe, although the numbers of                     than in abstract form. Thirdly, time consum-
London WC1                                                       hospital beds per capita vary fourfold. If                  ing quality control procedures essential to
Nick Black professor                                             anything, countries with fewer beds tend to                 ensure comparability of results between and
Department of Public Health and Policy, London                   have lower occupancy rates. It is almost
School of Hygiene and Tropical Medicine, London                                                                              within countries may be sacrificed to the
WC1E 7HT                                                         impossible to separate need for hospital                    time constraint. Fourthly, centres with less
                                                                 beds from demand. Icelanders are almost                     experience, which would benefit from
1 Jarman B, Gault S, Alves B, Hider A, Dolan S, Cook A, et al.
  Explaining differences in English hospital death rates         certainly not four times more likely to                     participation in multinational studies with
  using routinely collected data. BMJ 1999;318:1515-20. (5       become ill than Danish or Spanish people,                   leading centres, will tend not to be included
2 NHS Executive. Quality and performance in the NHS: clinical
                                                                 but they do spend almost four times as many                 because they usually slow down the pace of
  indicators. Leeds: NHSE, 1999.                                 days in hospital.                                           research. Finally, requests to extend the con-
3 Iezonni LI, Foley SM, Daley J, Hughes J, Fisher ES, Heeren          The problem with hospital beds is that
  T. Comorbidities, complications, and coding bias: does the
                                                                                                                             tract will occur at the last minute, generating
  number of diagnosis codes matter in predicting in-hospital     as long as they are there they tend to inhibit              administrative overload to gain extensions
  mortality? JAMA 1992;267:2197-203.                             the development of alternatives to admis-                   that will still be inadequate for the time that
4 Jencks SF, Williams DK, Kay TL. Assessing hospital-
  associated deaths from discharge data: the role of length of   sion to hospital. The real policy decision is               the research needs.
  stay and comorbidities. JAMA 1988;260:2240-6.                  therefore whether we should devote more                         To overcome this serious shortcoming
5 Imamura K, Black N. Does comorbidity affect the outcome
  of surgery? Total hip replacement in the UK and Japan. Int
                                                                 health services resources to hospital care or               we offer a simple proposal. In the negotia-
  J Qual Health Care 1998;10:113-23.                             more resources to care outside hospitals.                   tion phase (starting soon) between the
                                                                      Pollock has nailed her colours to the                  European Commission and the investiga-
                                                                 mast as a defender of hospitals; this implies               tors of the projects that have successfully
                                                                 that she believes fewer resources should be                 gone through the peer review process, a
Competing interests are                                          devoted to care outside hospitals. My own                   given amount of the funds assigned to an
relevant to lectures approved                                    instinct is towards demedicalisation, but cur-              epidemiological project should be spread (if
for PGEA                                                         rently there is little to say which approach                required and well argued for by the
                                                                 has more merits.                                            principal investigator) over five years rather
Editor—In journals authors are expected to
                                                                 Tom Marshall honorary lecturer in public health             than be concentrated in an illusory three
declare any competing interests. In the                          medicine                                                    years. Should the administrative manage-
context of lectures or workshops that are                        University of Birmingham, Birmingham B15 2TT                ment for the European Commission be
approved for postgraduate education allow-             
                                                                                                                             more onerous over five years than over
ance, however, when a group of general prac-                     1 Pollock AM, Dunnigan MG, Gaffney D, Price D, Shaoul J.    three, a small percentage of the assigned
titioners gather specifically to learn, there is                   The private finance initiative: Planning the “new” NHS:
                                                                   downsizing for the 21st century. BMJ 1999;319:179-84.     sum could be retained for the purpose—this
no such onus on the lecturer. Thus a talk to                       (17 July.)                                                will always be much less costly than the
30 primary care prescribers can be delivered
                                                                                                                             destructive wastage of resources caused by
by a person who has a direct stake in a
                                                                                                                             highly incongruous time schedules.
particular treatment. Would it not be sensible
to ask all providers of talks approved for post-                 New epidemiological research                                Rodolfo Saracci director of research in epidemiology
                                                                                                                             International Epidemiological Association,
graduate education allowance or continuing                       in Europe may be thwarted by                                National Research Council, 56100 Pisa, Italy
medical education to offer this information at                   short term funding                                          Josep Anto’ head of respiratory and environmental
the time they give their talks?                                                                                              health research unit
     If I was aware that a lecture on                            Editor—The European Union’s fifth frame-                    Institut Municipal d’Investigacio’ Medica, 08003
impotence was being given by a urologist                         work programme of research is making                        Barcelona, Spain
whose department was funded by Rhino                             available in 1999 some 850 million euros for                Jorn Olsen professor of epidemiology
Horn International my own interpretation                         projects in biomedicine, with the main                      Danish Epidemiological Science Centre, 8000
                                                                                                                             Aarhus C, Denmark
of his recommendations would take that                           objective of prompting new research on key
                                                                                                                             Anthony McMichael professor of epidemiology
into account. Of course this would never                         themes for the health of European citizens.                 London School of Hygiene and Tropical Medicine,
happen—our ethical standards are too high                        Whether this objective is achieved will                     London WC1E 7HT
for a colleague’s impartiality to be ques-                       depend critically on the length of funding.                 Dimitrios Trichopoulos professor of epidemiology
tioned. If it is important enough for journals                        The standard maximum time for a                        Harvard University and University of Athens,
to take a stand on competing interests, how-                     contract issued by the European Commission                  Boston, MA 02115, USA
ever, then why not clinicians generally? It                      is still three years. Clearly, this is totally un-
would perhaps be appropriate for those who                       realistic for most research in epidemiology, as
have undeclared competing interests to be                        it is for research in environmental and
fined—in          postgraduate       education                   occupational health and in evaluating treat-
allowance/continuing medical education                           ments or health services through randomised
units for the professional transgressors.                        trials. In our extensive experience as principal

BMJ VOLUME 319           25 SEPTEMBER 1999                                                                                                               855

Shared By: