- Email: [email protected]
depressing reasearch
CORRESPONDENCE
COMMENTARY
CORRESPONDENCE e-mail submissions to [email protected]
Depressing research Sir—Suppression of the publication of negative trials of serotonin-selective reuptake inhibitors (SSRI) in children (Apr 24, p 1335)1 is more than just a matter of “confusion, manipulation, and institutional failure”. It is a crime. To blandly illustrate its severity, we can analogise the situation as follows: would you be concerned if a colleague prescribed penicillin to a child who had an uncomplicated upper respiratory tract infection of viral aetiology? Would you be more concerned if you heard that this child had a serious anaphylactic reaction to the penicillin at home? And how much more concerned would you be if you found out that the prescribing doctor was previously aware of the child’s severe penicillin allergy? The latter scenario could be deemed malpractice. How, then, is it acceptable for pharmaceutical companies to suppress publication of SSRI studies that showed a lack of efficacy and an increased risk of serious adverse events in the children and adolescents in experimental trial groups (other than for fluoxetine)?2 Intentional concealment of these data, an accusation for which there is already public evidence,3 must be considered a form of corporate violence. If our legal systems do not have adequate mechanisms to hold multinational corporations accountable for their harmful actions, then we should identify the individuals who are ultimately responsible for corporate decisions. Accountability must also be claimed in part by the physician investigators themselves who were silently complicit in the suppression of crucial information or the authorship of misleading articles. More important than any punitive or legal reaction is the need to ensure that all clinical trials are registered, conducted, analysed, and published in an appropriate and objective manner. Publication bias should be at least one easily avoided obstacle to the dissemination of the best medical evidence. Physicians must act quickly to firmly disentangle themselves from drug companies, since our interests can drastically diverge.4 Corporate manipulation of data from paediatric SSRI trials is an appalling revelation that should send shockwaves through the
international clinical research community. If evidence-based medicine is to continue to be an effective tool for improving care and ensuring patients’ safety, then we must now collectively condemn the companies and individuals whose actions undermine the integrity of medical science. *Daniel Roth, Erin Boyle, Darcy Beer, Anita Malik, Jen deBruyn Department of Pediatrics and Child Health, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta T6G 2R7, Canada (e-mail: [email protected]) 1 2
3
4
The Lancet. Depressing research. Lancet 2004; 363: 1335. Whittington CJ, Kendall T, Fonagy P, Cottrell D, Cotgrove A, Boddington E. Selective serotonin reuptake inhibitors in childhood depression: systematic review of published versus unpublished data. Lancet 2004; 363: 1341–45. Kondro W, Sibbald B. Drug company experts advised staff to withhold data about SSRI use in children. CMAJ 2004; 170: 783. Moynihan R. Who pays for the pizza? Redefining the relationships between doctors and drug companies, 2: disentanglement. BMJ 2003; 326: 1193–96.
Sir—Your Editorial1 deplores the use of antidepressants for childhood depression on the basis of the selective reporting of favourable research.2 This state of affairs seems to be much more widespread than is generally recognised, and also applies to SSRIs for depression in adults.3 Drug regulatory agencies worldwide routinely rely on selective data presented by companies. For example, the pivotal requirement for drug approval by the US Food and Drug Administration (FDA) is evidence of superiority of active drug over placebo from two studies of sufficient size to establish statistical (if not clinical) significance. However, so long as companies provide evidence from two such studies, the FDA discounts the results of studies in which evidence of drug efficacy is lacking. Thus, drugs may be licensed even when evidence of efficacy is weak. Only meta-analysis would reveal this fault, but regulators rarely attempt it and secrecy prevents others from doing so. A leading academic psychiatrist and industry consultant recently noted that the manufacturers of SSRI anti-
THE LANCET • Vol 363 • June 19, 2004 • www.thelancet.com
depressants have now established that they need to plan for eight placebocontrolled efficacy studies in order to get positive results from two. Over half of all such studies on SSRIs have reportedly failed to show efficacy, and the proportion could be higher. The same source commented that Pharmacia (Pfizer) had failed to get US marketing approval for reboxetine because “they had the bad luck of only 1 of the 8 studies definitively showing that it works”.4 Reboxetine was licensed in the EU in 1997: it may be inferred that European regulatory standards for efficacy are no higher than in the USA. This is a serious and global problem. WHO estimates that 80% of countries have limited drug regulatory capacity, or none.5 They are obliged to adopt standards agreed between the US, EU, and Japan. Drug efficacy determined in this way constrains rational prescribing and is ethically objectionable. *Charles Medawar, Anita Hardon, Andrew Herxheimer *Social Audit Ltd, Box 111, London NW1 8XG, UK (CM); University of Amsterdam, Amsterdam, Netherlands (AH); and 9 Park Crescent, London N3 2NL, UK (AH) (e-mail: [email protected]) 1 2
3
4
5
The Lancet. Depressing research. Lancet 2004; 363: 1335. Whittington C, Kendall T, Fonagy P, et al. Selective serotonin reuptake inhibitors in childhood depression: systematic review of published versus unpublished data. Lancet 2004, 363: 1341–45. Medawar C, Hardon A. Medicines out of control. Amsterdam: Aksant Academic Publishers, 2004. Thase M. Assessing efficacy of antidepressants: clinician and FDA perspectives. http://www.medscape.com/ viewprogram/2546 (accessed June 1, 2004). World Health Organization. Effective medicines regulation: ensuring safety, efficacy and quality, WHO Policy Perspectives on Medicines, WHO/EDM/2003.2. Geneva: WHO, 2003.
Sir—Your Editorial1 highlights the importance of ensuring that medicine’s evidence base is unbiased and that industry-sponsored studies are reported responsibly. Although I am encouraged that The Lancet is highlighting such important issues, it would have been helpful to mention that several initiatives that aim to address them are already underway. A set of guidelines entitled Good Publication Practice for
2087
For personal use. Only reproduce with permission from The Lancet Publishing Group.
CORRESPONDENCE
Pharmaceutical Companies has been published2 which calls on drug companies to endeavour to publish results of all clinical trials of marketed products, and to produce publications in a responsible and ethical manner. These guidelines have been endorsed by some drug companies and are gaining support from individuals, companies, and journals involved in publishing industrysponsored research. The European Medical Writers Association is also in the process of developing guidelines for professional medical writers working on industry-sponsored publications. Doctors and patients might be encouraged to know that there are people within the industry who take these issues seriously and are working to improve practice. You might also have mentioned that, in Spain, a new law came into force in May as part of the implementation of the EU Clinical Trials Directive. This law requires drug companies to publish the results of their research.3 The need to publish both positive and negative findings has also been included in the revised version of the Declaration of Helsinki.4 Clinician-investigators, and others involved in industry-sponsored research such as writers and editors, can therefore now refer to these guidelines and join these initiatives to bring pressure on companies to prevent the suppression of unfavourable findings and the irresponsible presentation of results. Elizabeth Wager Sideview, 19 Station Road, Princes Risborough HP27 9DE, UK (e-mail: [email protected]) 1 2
3
4
The Lancet. Depressing research. Lancet 2004; 363: 1335. Wager E, Field EA, Grossman L. Good publication practice for pharmaceutical companies. Curr Med Res Opinion 2003; 19: 149–54. Ministerio de Sanidad y Consumo. Real Decreto 223/2004, de 6 de febrero, por el que se regulan los ensayos clínicos con medicamentos. Boletín Oficial de Estado 2004; 33: 5429–43. Hirsch LJ, Guess HA. What are the effects of the fifth revision of the Declaration of Helsinki? BMJ 2001; 323: 1417–23.
Sir—There are many reasons to be concerned by the events that have led to the Medicines and Healthcare products Regulatory Agency’s advice on SSRIs for children and adolescents, particularly with respect to what these events say about the evidence base for child and adolescent psychiatry and practice implications in the field. However, I cannot feel that your Editorial1 properly reflected these or even the results of the paper by Craig Whittington and colleagues.2
2088
You start the Editorial with what can only be described as tabloid-type sensationalism—appealing to parents’ feelings after an iatrogenic medicationrelated death of their child. Why? No such death has been reported.2,3 To watch how the response to the SSRI debate has become polarised, with much moral high ground being assumed, has been disconcerting. Yes, there are serious concerns about the way that the drug companies handled the trials and their internal and external validity, and Whittington and colleagues point to evidence that there has been a publication bias in favour of positive results. But this kind of publication bias against negative findings has been a widespread concern and is not just confined to drug-company-sponsored research. An equal concern in my mind has been the failure over many years of funding bodies to invest properly in psychopharmacological research in childhood and adolescence which could have replicated some of these findings or at least designed studies robust enough to test the issues properly. Things as they stand are not clear-cut. Whittington and colleagues properly conclude that the extant studies were not designed well enough to be able to make any proper conclusion on adverse effects from SSRIs in this age-group. Although the evidence on apparent lack of effectiveness seems compelling, it is potentially undermined by the lack of external validity of the studies undertaken: there is no stratification for age and severity nor identification of subgroups that might benefit from these drugs. The clinical consensus before the recent debates was that, judicially used, SSRIs are an important part of the child psychiatry pharmacopoeia, especially in serious problems of teenage depression. No alternative treatments have been shown effective in this severe symptom group. The precautionary principle has been invoked, but in the absence of well designed studies in high-risks groups, we cannot be sure that this judgment sufficiently balances the risks of nontreatment. It is to be hoped that the lessons taken from this episode extend primarily to a change in funding culture to allow properly designed trials to be undertaken by independent academic institutions alone or in collaboration with pharmaceutical companies. Jonathan Green Academic Department of Child and Adolescent Psychiatry, Booth Hall Childrens Hospital, Blackley, Manchester M9 7AA, UK (e-mail: [email protected]) 1 2
The Lancet. Depressing research. Lancet 2004; 363: 1335. Whittington CJ, Kendall T, Fonagy P, et al.
3
Selective serotonin reuptake inhibitors in childhood depression: systematic review of published versus unpublished data. Lancet 2004; 363: 1341–45. Brent DA, Birmaher B. British warnings on SSRIs questioned. J Am Acad Child Adolesc Psychiatry 2004; 43: 379–80.
Improving quality of care for severe malnutrition Sir—Ann Ashworth and colleagues (Apr 3, p 1110)1 blame their findings of high case fatality from severe malnutrition in rural South Africa on doctor and nurse errors. Only one of the 46 deaths was attributed to “systems fault”. This flies in the face of the quality assurance experience in health, and even seems to contradict much of the discussion in the paper about weak health-system infrastructure and low morale, which are surely system failures. Although the clinical errors mentioned seem serious, it is somewhat tenuous to allege that without them all 36 deaths, representing 43% of deaths in this study, would have been avoided, since even appropriately treated sepsis in severe malnutrition has a significant unavoidable mortality. Ashworth has stated that poor case management is the main determinant of case fatality and that implementing WHO guidelines alone will reduce it to below 5%.2 It is instructive to compare this study from a wealthy African country (gross national income [GNI] per capita US$2900) to our study in nearby Malawi (GNI per capita $170). We implemented paediatric supervisory visits to promote compliance with the WHO guidelines over a 12-month period, studying prospectively 1625 consecutive admissions with severe malnutrition to two central hospitals, two district hospitals (similar to those in the south African study), and three rural clinics.3 Case-fatality rates were 30·5% (275/901) at central hospitals, 25·8% (90/351) at the district level, and 7·5% (28/373) at the rural level. In a further study, we expended substantial research resources to improve nursing and paediatric supervision of an unselected cohort at a central hospital, which only reduced the case-fatality rate non-significantly (p=0·34) to 25% (19/75).4 These and other studies show that severity of illness is the main determinant of mortality, and reducing case-fatality to below 5% is not possible in some settings without reducing the severity of illness (eg, admitting milder cases). In Ashworth and colleagues’ study, the use of historical controls, the
THE LANCET • Vol 363 • June 19, 2004 • www.thelancet.com
For personal use. Only reproduce with permission from The Lancet Publishing Group.
COMMENTARY
CORRESPONDENCE e-mail submissions to [email protected]
Depressing research Sir—Suppression of the publication of negative trials of serotonin-selective reuptake inhibitors (SSRI) in children (Apr 24, p 1335)1 is more than just a matter of “confusion, manipulation, and institutional failure”. It is a crime. To blandly illustrate its severity, we can analogise the situation as follows: would you be concerned if a colleague prescribed penicillin to a child who had an uncomplicated upper respiratory tract infection of viral aetiology? Would you be more concerned if you heard that this child had a serious anaphylactic reaction to the penicillin at home? And how much more concerned would you be if you found out that the prescribing doctor was previously aware of the child’s severe penicillin allergy? The latter scenario could be deemed malpractice. How, then, is it acceptable for pharmaceutical companies to suppress publication of SSRI studies that showed a lack of efficacy and an increased risk of serious adverse events in the children and adolescents in experimental trial groups (other than for fluoxetine)?2 Intentional concealment of these data, an accusation for which there is already public evidence,3 must be considered a form of corporate violence. If our legal systems do not have adequate mechanisms to hold multinational corporations accountable for their harmful actions, then we should identify the individuals who are ultimately responsible for corporate decisions. Accountability must also be claimed in part by the physician investigators themselves who were silently complicit in the suppression of crucial information or the authorship of misleading articles. More important than any punitive or legal reaction is the need to ensure that all clinical trials are registered, conducted, analysed, and published in an appropriate and objective manner. Publication bias should be at least one easily avoided obstacle to the dissemination of the best medical evidence. Physicians must act quickly to firmly disentangle themselves from drug companies, since our interests can drastically diverge.4 Corporate manipulation of data from paediatric SSRI trials is an appalling revelation that should send shockwaves through the
international clinical research community. If evidence-based medicine is to continue to be an effective tool for improving care and ensuring patients’ safety, then we must now collectively condemn the companies and individuals whose actions undermine the integrity of medical science. *Daniel Roth, Erin Boyle, Darcy Beer, Anita Malik, Jen deBruyn Department of Pediatrics and Child Health, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta T6G 2R7, Canada (e-mail: [email protected]) 1 2
3
4
The Lancet. Depressing research. Lancet 2004; 363: 1335. Whittington CJ, Kendall T, Fonagy P, Cottrell D, Cotgrove A, Boddington E. Selective serotonin reuptake inhibitors in childhood depression: systematic review of published versus unpublished data. Lancet 2004; 363: 1341–45. Kondro W, Sibbald B. Drug company experts advised staff to withhold data about SSRI use in children. CMAJ 2004; 170: 783. Moynihan R. Who pays for the pizza? Redefining the relationships between doctors and drug companies, 2: disentanglement. BMJ 2003; 326: 1193–96.
Sir—Your Editorial1 deplores the use of antidepressants for childhood depression on the basis of the selective reporting of favourable research.2 This state of affairs seems to be much more widespread than is generally recognised, and also applies to SSRIs for depression in adults.3 Drug regulatory agencies worldwide routinely rely on selective data presented by companies. For example, the pivotal requirement for drug approval by the US Food and Drug Administration (FDA) is evidence of superiority of active drug over placebo from two studies of sufficient size to establish statistical (if not clinical) significance. However, so long as companies provide evidence from two such studies, the FDA discounts the results of studies in which evidence of drug efficacy is lacking. Thus, drugs may be licensed even when evidence of efficacy is weak. Only meta-analysis would reveal this fault, but regulators rarely attempt it and secrecy prevents others from doing so. A leading academic psychiatrist and industry consultant recently noted that the manufacturers of SSRI anti-
THE LANCET • Vol 363 • June 19, 2004 • www.thelancet.com
depressants have now established that they need to plan for eight placebocontrolled efficacy studies in order to get positive results from two. Over half of all such studies on SSRIs have reportedly failed to show efficacy, and the proportion could be higher. The same source commented that Pharmacia (Pfizer) had failed to get US marketing approval for reboxetine because “they had the bad luck of only 1 of the 8 studies definitively showing that it works”.4 Reboxetine was licensed in the EU in 1997: it may be inferred that European regulatory standards for efficacy are no higher than in the USA. This is a serious and global problem. WHO estimates that 80% of countries have limited drug regulatory capacity, or none.5 They are obliged to adopt standards agreed between the US, EU, and Japan. Drug efficacy determined in this way constrains rational prescribing and is ethically objectionable. *Charles Medawar, Anita Hardon, Andrew Herxheimer *Social Audit Ltd, Box 111, London NW1 8XG, UK (CM); University of Amsterdam, Amsterdam, Netherlands (AH); and 9 Park Crescent, London N3 2NL, UK (AH) (e-mail: [email protected]) 1 2
3
4
5
The Lancet. Depressing research. Lancet 2004; 363: 1335. Whittington C, Kendall T, Fonagy P, et al. Selective serotonin reuptake inhibitors in childhood depression: systematic review of published versus unpublished data. Lancet 2004, 363: 1341–45. Medawar C, Hardon A. Medicines out of control. Amsterdam: Aksant Academic Publishers, 2004. Thase M. Assessing efficacy of antidepressants: clinician and FDA perspectives. http://www.medscape.com/ viewprogram/2546 (accessed June 1, 2004). World Health Organization. Effective medicines regulation: ensuring safety, efficacy and quality, WHO Policy Perspectives on Medicines, WHO/EDM/2003.2. Geneva: WHO, 2003.
Sir—Your Editorial1 highlights the importance of ensuring that medicine’s evidence base is unbiased and that industry-sponsored studies are reported responsibly. Although I am encouraged that The Lancet is highlighting such important issues, it would have been helpful to mention that several initiatives that aim to address them are already underway. A set of guidelines entitled Good Publication Practice for
2087
For personal use. Only reproduce with permission from The Lancet Publishing Group.
CORRESPONDENCE
Pharmaceutical Companies has been published2 which calls on drug companies to endeavour to publish results of all clinical trials of marketed products, and to produce publications in a responsible and ethical manner. These guidelines have been endorsed by some drug companies and are gaining support from individuals, companies, and journals involved in publishing industrysponsored research. The European Medical Writers Association is also in the process of developing guidelines for professional medical writers working on industry-sponsored publications. Doctors and patients might be encouraged to know that there are people within the industry who take these issues seriously and are working to improve practice. You might also have mentioned that, in Spain, a new law came into force in May as part of the implementation of the EU Clinical Trials Directive. This law requires drug companies to publish the results of their research.3 The need to publish both positive and negative findings has also been included in the revised version of the Declaration of Helsinki.4 Clinician-investigators, and others involved in industry-sponsored research such as writers and editors, can therefore now refer to these guidelines and join these initiatives to bring pressure on companies to prevent the suppression of unfavourable findings and the irresponsible presentation of results. Elizabeth Wager Sideview, 19 Station Road, Princes Risborough HP27 9DE, UK (e-mail: [email protected]) 1 2
3
4
The Lancet. Depressing research. Lancet 2004; 363: 1335. Wager E, Field EA, Grossman L. Good publication practice for pharmaceutical companies. Curr Med Res Opinion 2003; 19: 149–54. Ministerio de Sanidad y Consumo. Real Decreto 223/2004, de 6 de febrero, por el que se regulan los ensayos clínicos con medicamentos. Boletín Oficial de Estado 2004; 33: 5429–43. Hirsch LJ, Guess HA. What are the effects of the fifth revision of the Declaration of Helsinki? BMJ 2001; 323: 1417–23.
Sir—There are many reasons to be concerned by the events that have led to the Medicines and Healthcare products Regulatory Agency’s advice on SSRIs for children and adolescents, particularly with respect to what these events say about the evidence base for child and adolescent psychiatry and practice implications in the field. However, I cannot feel that your Editorial1 properly reflected these or even the results of the paper by Craig Whittington and colleagues.2
2088
You start the Editorial with what can only be described as tabloid-type sensationalism—appealing to parents’ feelings after an iatrogenic medicationrelated death of their child. Why? No such death has been reported.2,3 To watch how the response to the SSRI debate has become polarised, with much moral high ground being assumed, has been disconcerting. Yes, there are serious concerns about the way that the drug companies handled the trials and their internal and external validity, and Whittington and colleagues point to evidence that there has been a publication bias in favour of positive results. But this kind of publication bias against negative findings has been a widespread concern and is not just confined to drug-company-sponsored research. An equal concern in my mind has been the failure over many years of funding bodies to invest properly in psychopharmacological research in childhood and adolescence which could have replicated some of these findings or at least designed studies robust enough to test the issues properly. Things as they stand are not clear-cut. Whittington and colleagues properly conclude that the extant studies were not designed well enough to be able to make any proper conclusion on adverse effects from SSRIs in this age-group. Although the evidence on apparent lack of effectiveness seems compelling, it is potentially undermined by the lack of external validity of the studies undertaken: there is no stratification for age and severity nor identification of subgroups that might benefit from these drugs. The clinical consensus before the recent debates was that, judicially used, SSRIs are an important part of the child psychiatry pharmacopoeia, especially in serious problems of teenage depression. No alternative treatments have been shown effective in this severe symptom group. The precautionary principle has been invoked, but in the absence of well designed studies in high-risks groups, we cannot be sure that this judgment sufficiently balances the risks of nontreatment. It is to be hoped that the lessons taken from this episode extend primarily to a change in funding culture to allow properly designed trials to be undertaken by independent academic institutions alone or in collaboration with pharmaceutical companies. Jonathan Green Academic Department of Child and Adolescent Psychiatry, Booth Hall Childrens Hospital, Blackley, Manchester M9 7AA, UK (e-mail: [email protected]) 1 2
The Lancet. Depressing research. Lancet 2004; 363: 1335. Whittington CJ, Kendall T, Fonagy P, et al.
3
Selective serotonin reuptake inhibitors in childhood depression: systematic review of published versus unpublished data. Lancet 2004; 363: 1341–45. Brent DA, Birmaher B. British warnings on SSRIs questioned. J Am Acad Child Adolesc Psychiatry 2004; 43: 379–80.
Improving quality of care for severe malnutrition Sir—Ann Ashworth and colleagues (Apr 3, p 1110)1 blame their findings of high case fatality from severe malnutrition in rural South Africa on doctor and nurse errors. Only one of the 46 deaths was attributed to “systems fault”. This flies in the face of the quality assurance experience in health, and even seems to contradict much of the discussion in the paper about weak health-system infrastructure and low morale, which are surely system failures. Although the clinical errors mentioned seem serious, it is somewhat tenuous to allege that without them all 36 deaths, representing 43% of deaths in this study, would have been avoided, since even appropriately treated sepsis in severe malnutrition has a significant unavoidable mortality. Ashworth has stated that poor case management is the main determinant of case fatality and that implementing WHO guidelines alone will reduce it to below 5%.2 It is instructive to compare this study from a wealthy African country (gross national income [GNI] per capita US$2900) to our study in nearby Malawi (GNI per capita $170). We implemented paediatric supervisory visits to promote compliance with the WHO guidelines over a 12-month period, studying prospectively 1625 consecutive admissions with severe malnutrition to two central hospitals, two district hospitals (similar to those in the south African study), and three rural clinics.3 Case-fatality rates were 30·5% (275/901) at central hospitals, 25·8% (90/351) at the district level, and 7·5% (28/373) at the rural level. In a further study, we expended substantial research resources to improve nursing and paediatric supervision of an unselected cohort at a central hospital, which only reduced the case-fatality rate non-significantly (p=0·34) to 25% (19/75).4 These and other studies show that severity of illness is the main determinant of mortality, and reducing case-fatality to below 5% is not possible in some settings without reducing the severity of illness (eg, admitting milder cases). In Ashworth and colleagues’ study, the use of historical controls, the
THE LANCET • Vol 363 • June 19, 2004 • www.thelancet.com
For personal use. Only reproduce with permission from The Lancet Publishing Group.