C 2007 International Psychogeriatric Association

International Psychogeriatrics (2007), 19:3, 345354

doi:10.1017/S1041610207005145 Printed in the United Kingdom

CONSENSUS STATEMENT

International Psychogeriatric Association

consensus statement on dening and
measuring treatment benets in dementia
..............................................................................................................................................................................................................................................................................

Cornelius Katona,1 Gill Livingston,2 Claudia Cooper,3

David Ames,4 Henry Brodaty5 and Edmond Chiu6 on behalf
of the Consensus Group7
1

Chair, Old Age Taskforce, World Federation of Societies of Biological Psychiatry, Professor and Dean, Kent
Institute of Medicine and Health Sciences, University of Kent, Canterbury, U.K.
Secretary, Old Age Taskforce, World Federation of Societies of Biological Psychiatry, Professor of Psychiatry for
Older People, Centre for Ageing and Mental Health Science, University College London, London, U.K.
3
Medical Research Council Fellow, Centre for Ageing and Mental Health Science, University College, London, U.K.
4
Editor International Psychogeriatrics, University of Melbourne Academic Unit for Psychiatry of Old Age, Kew,
Australia
5
Immediate Past President Alzheimers Disease International, Professor of Age Care Mental Health, Primary
Dementia Collaborative Research Centre, University of New South Wales, Sydney, Australia
6
Chair Old Age Psychiatry Section World Psychiatric Association, Professorial Fellow in Psychiatry of Old Age,
University of Melbourne Academic Unit for Psychiatry of Old Age, Kew, Australia
7
Members of the consensus group are listed in Appendix 1.
2

ABSTRACT

Current symptomatic treatments for dementia have only modest efficacy.

Assessing meaningful benefits in this variably progressive syndrome is complex
and difficult. This consensus statement was generated by an international
group representing caregivers, organizations and professionals with expertise
in dementia.
We recommend the statement of clear, pre-defined diagnostic and severity
criteria and outcome measures, which include functional and executive capacity,
in treatment trials. Outcomes can include effects on people with dementia
(PWD) with regard to cognition, behavioral and psychological symptoms,
quality of life, global assessments, and activities of daily living, and must be
tailored to the education and culture of the participants. Outcomes can also
appropriately encompass effects on caregivers. New instruments may be needed,
as validation of many existing measures has been in relatively homogenous
populations. Treatment goals can be to prevent dementia emerging, or in those
with established dementia to slow deterioration, and to postpone, stabilize or
improve symptoms. Comparisons between treatment groups should be on the
Correspondence should be addressed to: Professor C. Katona, Kent Institute of Medicine and Health Sciences, University of
Kent, Canterbury CT2 7PD, U.K. Phone: +44 (0) 1227 824063; Fax: +44 (0) 1227 824054. Email: c.katona@kent.ac.uk.
First published online 27 March 2007.

345

346

C. Katona et al.
basis of clinically relevant measures with both risk and benefit reported for all
participants regardless of whether or not they continue to receive treatment
throughout the trial. Characterization of any groups that respond well to
treatment has been unsuccessful to date, but may be facilitated in the future
by measurement of putative biomarkers. Despite considerable recent progress
and several candidate biomarkers, none is yet satisfactory for determining
diagnosis, severity, progression or prediction of response.
To provide meaningful data, economic analyses should use up-to-date, countryspecific data. Health economic measures should be incorporated as secondary
outcomes in all Phase 3 trials since health systems are concerned with costeffectiveness as well as clinical outcome. Health utility measures are not,
however, validated satisfactorily in dementia, thus calling into question previous
health economic analyses. While current drugs appear to reduce the amount of
family caregiver time required by PWD, these costs fall in the main on older
individuals who often exert little political leverage, rather than on society at large.
Traditionally, elderly people have been marginalized in the political process. The
growth in the older population across the world, and their potential for increasing
political empowerment may lead to a radical re-evaluation of the economics of
treatment in dementia.
Key words: dementia, treatment, randomized controlled trials, outcomes, measurement, caregivers,
biomarkers, Alzheimers disease, health economics, culture

Introduction
There are over 24 million people with dementia worldwide (Ferri et al., 2005).
This number is expected to double every 20 years, and by 2040 there will be
81 million people affected by dementia, nearly two-thirds of whom will be in
the developing world. Symptomatic treatments offer modest benefits and there
is hope of disease-modifying therapies emerging within the next decade (Ritchie
et al., 2007). Assessing benefits in the context of a variably progressive syndrome
is complex, particularly in a climate that mandates prioritization in the use
of limited resources. A consensus as to what constitutes meaningful treatment
benefits is therefore necessary.
The Consensus Group (see Appendix 1 for details of its membership)
met in Canterbury, U.K. on 31 October and 1 November 2006 to develop
this statement on defining and measuring treatment benefits in dementia.
The meeting was organized by the International Psychogeriatric Association
(IPA) in association with the Old Age Taskforce of the World Federation of
Societies of Biological Psychiatry, Alzheimers Disease International and the
World Psychiatric Association Section of Old Age Psychiatry. Participants were
selected for their expertise and also included representatives of the World Health

International Psychogeriatric Association consensus statement

Organization and the European pharmaceutical regulatory authority. They came

from 16 countries in six continents, and included family caregivers, medical
specialists in psychiatry, neurology, geriatric medicine and primary care, as
well as allied health professionals (psychology, nursing, social work), health
economists and a regulator. Papers given at the meeting are published in this
(June 2007) issue of International Psychogeriatrics. The meeting was supported
financially by a number of pharmaceutical companies (see Appendix 2); their
representatives attended as observers (see Appendix 3) and were invited to
comment during the conference, but did not take part in the planning of the
meeting or the preparation of this report.
This consensus statement is intended to assist those with a personal, clinical,
academic or policy involvement in dementia, including researchers, clinicians,
regulators, funding bodies, health authorities, policy-makers and caregivers,
as well as those developing new treatments targeting the symptoms or course
of the dementias. Our working definition of the dementia syndrome is that
of the International Classification of Diseases, 10th edition (ICD-10) (WHO,
1992) though we wish to emphasize that assessment of functional and executive
impairments should be more prominent in future classificatory systems in view
of their importance to people with dementia (PWD) and their caregivers. This
statement is particularly informed by our knowledge of treatment effects in
Alzheimers disease (AD), as most evidence to date relates to this, the most
common form of dementia.

Selection of trial populations in dementia

The power of clinical trials in AD is significantly reduced by diagnostic
inaccuracy; this also reduces the interpretability of trial outcomes. There is
a tension between requirements for diagnostic accuracy, and for clinical trial
results to be generalizable to the clinical population that might subsequently
receive the intervention. The more complex the assessment involved in selecting
the population, the less likely it becomes that a trial will include people with
behavioral and psychological symptoms of dementia (BPSD or neuropsychiatric
symptoms). This is important as these symptoms are common and are associated
with greater patient and caregiver distress, decreased quality of life and increased
institutionalization and costs of care for PWD (Ryu et al., 2005). Similarly,
diagnostic accuracy and the detection of treatment effects may be easier to
attain in moderate than in mild dementia, but the potential benefits of disease
modification are likely to be greatest in the early or prodromal stages. At
the very least, pre-defined diagnostic and severity criteria should be made
explicit in future trials to ensure transparency with regard to comparability and
generalizability.

347

348

C. Katona et al.

The measurement of treatment effects

The natural history of deterioration in dementia is variable; further naturalistic
data in community (rather than clinic) populations are needed to inform analyses
addressing slowing of deterioration.
Historically, licensing requirements for the anti-dementia drugs have centered
on the achievement of statistically significant effects on cognitive, functional and
global domains. More recently, some regulators have indicated that improvement
in quality of life could be an acceptable alternative to global improvement.
The consensus group identified a number of other meaningful and measurable
domains which may constitute appropriate targets for interventions. These
include caregiver outcomes (including psychological consequences and time
spent in delivering care), BPSD, quality of life, activities of daily living, global
assessment, health utility and healthcare costs. Treatments may also enhance
personal autonomy and capacity.
Intervention trials should pre-specify outcome measures and effect sizes that
are accepted as clinically meaningful. Within a given clinical trial, there may
be dissonance between the need for measures that test very specific hypotheses
and for outcomes that are comprehensive and relevant. Measures should be both
relevant to the trials hypotheses and appropriate for the cultural and educational
background of the population under study. The validity of outcome measures
is limited to the populations in which they have been tested and cannot be
assumed to extend to others. Cognitive and behavioral measures are particularly
vulnerable to the effects of culture, language and education. The relevance of
some domains may vary between cultures; functional disability, for example,
may be less important in cultural contexts where independence and autonomy
in activities of daily living are not part of the older persons role (Gureje et al.,
2006). Similarly, memory impairment is not regarded as important in all cultures
(Chiu and Zhang, 2000).
Adjustment of cut-points may affect a tests performance characteristics.
In addition, the psychometric properties of scales need to be taken into
account in analyzing outcomes; for example, scale linearity should not be
assumed. Heterogeneity of items may mean that scales are not linear, while
analyzing each of many items individually may lead to spurious statistical
significance.
Where possible, primary comparisons must be on the basis of clinically
relevant ordinal scales and with all trial participants followed up regardless of
whether or not they continue to receive the trial treatment. It is also essential to
assess adverse effects of treatment in order to determine the risk-benefit ratio.
Staging instruments may allow the whole spectrum of dementia to be considered,
but since such instruments often lack the precision to detect clinically significant

International Psychogeriatric Association consensus statement

changes within an individual stage while also being vulnerable to cusp effects,
they are not sufficient as a sole outcome measure.
Health-Related Quality of Life instruments can capture how people feel about
their health, its impact and personal interactions. They have, however, only
been validated in specific populations and their sensitivity to clinically relevant
change in dementia requires further study. While AD-specific measures have
been developed, they do not allow comparisons with other conditions. More
generic health-related quality of life instruments validated across a range of
diseases could also be included in trials. Existing quality of life instruments may
not be appropriate for all cultures. They are not, for example, valid for use in
Asia, as they have been developed on the basis of priorities and values expressed
by non-Asian populations (Chiu and Chiu, 2005).
Information about health-related quality of life can be given by either PWD
or by family/professional caregivers, or both, and can also be observed directly.
Responses of caregivers, PWD and health professionals may diverge as there
are inherent differences in judgment and memory in an illness characterized by
cognitive impairment.

The need for new measures

Despite progress in the measurement of individual domains in dementia over
the last decade, we identified a number of areas for further development. In
particular, as the number of people with dementia in developing countries
soars, culturally appropriate instruments are needed. This may require new
instruments rather than merely adaptation of existing ones. Priorities vary
between individuals and at different stages of dementia, regardless of culture, and
measures have also been developed which take into account baseline functioning
and individual priorities. These allow the identification of individually relevant
goals and the monitoring of their attainment or loss (Rockwood et al., 2006).
Clinicians are now aware that neuropsychiatric symptoms (also known as
BPSD) are very common and diverse, and scores on specific symptom clusters,
such as measures of affect, apathy and psychosis (Lyketsos et al., 2001) may be
more meaningful than an overall score. These need to be developed further to facilitate valid measurement of treatment effects on specific symptoms or domains.

The goals of treatment

In any trial the goal may be to prevent dementia emerging or, in those
with established dementia, to slow deterioration, and postpone, stabilize or
improve symptoms. Other aims may be to improve quality of life, prevent

349

350

C. Katona et al.

institutionalization or improve function. Trial aims should be clearly stated and

relate to the primary hypothesis, and the tools selected for measurement should
be in line with this.
An outcome formulated in terms of comparison of means may be more
difficult to understand outside of a specialist research context than one expressed
as a change in health state. Slowing of deterioration may be amenable to slope
analysis. Numbers Needed to Treat (NNT) analyses are also relatively easy to
interpret and make comparisons with treatments inside and outside the field
possible, but must be on predetermined and clinically meaningful outcomes.
The effect of treatment on caregivers is of great importance. In future trials of
disease modifying agents, participants may be independent in activities of daily
living and so not have a caregiver. In addition, the priorities of the caregiver will
vary according to stage of the dementia, personal circumstances and culture of
origin. Qualitative interviews with small numbers of caregivers and PWD may
give valuable additional information to complement and inform quantitative
data.

Recommendations for responder analyses

There are some people within trial cohorts who respond much better to
interventions than others, and their characterization would clearly be very helpful
in order to identify subgroups for which a particular treatment is more likely to be
beneficial. Efforts to do this, based on clinical and demographic characteristics,
or by using patients changes on rating scales, have so far been unsuccessful.
Possible strategies to facilitate work in this area in the future include
stratification or enrichment of study populations to identify predictors such
as putative biomarkers of response (e.g. APOE 4). Statistical pooling or metaanalysis of all studies may enhance post-hoc responder analyses; selective pooling
of studies is not appropriate.

The role of biomarkers

Despite considerable recent progress, there is as yet no satisfactory biomarker
for diagnosis, severity, progression or prediction of response (Galasko, 2005).
Of several candidate biomarkers, volumetric neuroimaging shows particular
promise as a measure of progression; markers of oxidative stress may also become
useful. Biomarkers may be informative indicators of the mechanism of action
of novel treatments or may help identify potential responders but would be
supplementary to and not a proxy for clinically significant measures of treatment
response.

International Psychogeriatric Association consensus statement

The role of health economics

Health utility measures such as Quality Adjusted Life Years (QALY) and
Disability Adjusted Life Years (DALY) are based on the values people place on
different health states but are not satisfactorily validated in dementia (Wimo and
Winblad, 2006). To provide meaningful data, economic analyses should aim
to use up-to-date data applicable to the country being studied. Incorporating
standard clinical measures (cognitive, staging, etc.) in population studies and
trials is also important in informing the assumptions underpinning economic
modeling.
It is unclear to what extent PWD can complete relevant questionnaires validly
and reliably. Despite these limitations, health economic measures should be
incorporated as secondary outcome variables in all Phase 3 trials since these are
crucial to understanding effectiveness and to the information base required by
third-party funders of dementia treatment and care, and because, more generally,
health systems are concerned to achieve cost-effectiveness as well as to improve
outcomes for PWD and their carers.
It is currently unclear to what extent current drugs alter the patterns of use
of health or social care services, and the associated direct cost savings may not
be large unless admission to full-time care is significantly delayed. While current
drugs may substantially reduce family caregiver time required by PWD, there
remains disagreement as to how such savings should be valued and in particular
whether these should be based on replacement costs (which would be very high)
or on opportunity costs (generally extremely low). There are also difficulties
in defining which activities constitute caring. These caregiver-related costs fall
mainly on individuals rather than on society at large. At present this produces
little political leverage when decisions are made about whether the cost of a
drug is worth paying. Traditionally, older people have been marginalized in the
political process. The growth in the older population across the world, and their
increasing political empowerment, may lead to a re-evaluation of the economic
contribution of dementia caregivers and therefore the economics of treatment in
dementia.

Conict of interest declaration

Cornelius Katona has received research support, assistance to attend conferences
and/or speakers/consultants honoraria from Eisai, Eli Lilly, GlaxoSmithKline,
Janssen-Cilag, Novartis, Pfizer, and Lundbeck.
Gill Livingston has received research support and/or assistance to attend
conferences and/or speakers/consultants honoraria from Eisai, Eli Lilly,
GlaxoSmithKline, Janssen-Cilag, Novartis, Pfizer, Quintiles and Lundbeck.

351

352

C. Katona et al.

Claudia Cooper has worked on a project supported by Lundbeck.

David Ames and Henry Brodaty have received research support, and/or
assistance to attend conferences and/or speakers/consultants honoraria from:
Astra Zeneca, Bristol Myers Squibb, Eisai, Eli Lilly, GlaxoSmithKline, JanssenCilag, Novartis, Parke Davis, Pfizer, Quintiles, Sanofi-Aventis, Servier, and
Voyager.
Edmond Chiu has received conference sponsorship and speaker honoraria
from Pfizer, Esai and Janssen, and has been an investigator in trials sponsored
by Pfizer, Janssen, Novartis, Servier and Sanofi-Aventis.

Description of authors roles

The consensus conference was conceived by Cornelius Katona and Gill
Livingston, and was organized by IPA via a committee chaired by Cornelius
Katona which included Cornelius Katona, Gill Livingston, David Ames,
Edmond Chiu and Henry Brodaty. Claudia Cooper acted as rapporteur for
the conference. The consensus statement was written by Claudia Cooper,
Cornelius Katona, Gill Livingston, David Ames, Henry Brodaty and Edmond
Chiu and was then circulated to all consensus participants listed in Appendix 1
for comment prior to final approval.

References
Chiu, H. and Chiu, E. (2005). Dementia care in Asia. International Psychogeriatrics, 17, 12.
Chiu, H. and Zhang, H. (2000). Dementia research in China. International Journal of Geriatric
Psychiatry, 15, 947953.
Ferri, C. P. et al. (2005). Global prevalence of dementia: a Delphi consensus study. Lancet,
366, 21122117.
Galasko, D. (2005). Biomarkers for Alzheimers disease clinical needs and application. Journal
of Alzheimers Disease, 8, 339346.
Gureje, O., Oguyinni, A. and Kola, L. (2006). The profile and impact of probable dementia in
a sub-Saharan African community: results from the Ibadan Study of Aging. Journal of
Psychosomatic Research, 61, 327333.
Lyketsos, C. G., Breitner, J. C. S. and Rabins, P. V. (2001). An evidence-based proposal for
the classification of neuropsychiatric disturbance in Alzheimers disease. International Journal
of Geriatric Psychiatry, 16, 10371042.
Ritchie, C., Ames, D., Masters, C. and Cummings, J. (2007). Treatment Strategies in
Dementia, Oxford: Clinical Publishing.
Rockwood, K. et al. (2006). Attainment of treatment goals by people with Alzheimers disease
receiving galantamine: a randomized controlled trial. Canadian Medical Association Journal,
174, 10991105.
Ryu, S. H., Katona, C., Rive, B. and Livingston, G. (2005). Persistence of and changes in
neuropsychiatric symptoms in Alzheimers disease over 6 months the LASER-AD Study.
American Journal of Geriatric Psychiatry, 13, 976983.

International Psychogeriatric Association consensus statement

Wimo, A. and Winblad, B. (2006). Pharmacoeconomic outcomes. In K. Rockwood and S.
Gauthier (eds.), Trial Designs and Outcomes in Dementia Therapeutic Research (pp. 250258).
Abingdon: Taylor and Francis.
World Health Organization (1992). International Classification of Diseases, 10th edn. Geneva:
World Health Organization.

Appendix 1
Consensus conference participants
Ricardo Allegri (Neurologist, Argentina), David Ames (Psychiatrist and
Editor-in-Chief of International Psychogeriatrics, Australia), Olusegun Baiyewu
(Psychiatrist, Nigeria), Karl Broich (Physician and Pharmaceutical Regulator,
Germany), Henry Brodaty (representing Alzheimers Disease International,
Psychiatrist, Australia), Roger Bullock (Psychiatrist, U.K.), Alistair Burns
(Psychiatrist and former President of IPA, U.K.), Edmond Chiu (representing
Section of Old Age Psychiatry of World Psychiatric Association and former
President of IPA, Psychiatrist, Australia), Helen Chiu (President Elect of IPA
representing IPA, Psychiatrist, Hong Kong SAR, China), Claudia Cooper
(Psychiatrist, U.K.), Marshal Folstein (Psychiatrist, U.S.A.), Serge Gauthier
(Neurologist, Canada), Guk-Hee Suh (Psychiatrist, South Korea), Juanita Hoe
(Nurse, U.K.), Cornelius Katona (representing the Old Age Taskforce of the
World Federation of Societies of Biological Psychiatry, Psychiatrist, U.K.), Steve
Iliffe (General Practitioner and member of the NICE/SCIE dementia guidelines
development group 20046, U.K.), Sirrka-Liisa Kivela (representing the Old
Age Taskforce of the World Federation of Societies of Biological Psychiatry,
Primary Care Physician, Finland), Martin Knapp (Health Economist, U.K.),
Alexander Kurz (Psychiatrist, Germany), Carlos Augusto de Mendonca
Lima (representing the World Health Organization, Psychiatrist, Brazil), Gill
Livingston (representing the Old Age Taskforce of the World Federation of
Societies of Biological Psychiatry, Psychiatrist, U.K.), Constantine Lyketsos
(representing the Old Age Taskforce of the World Federation of Societies of
Biological Psychiatry, Psychiatrist, U.S.A.), Jacobo Mintzer (Treasurer-elect
IPA representing IPA, Psychiatrist, U.S.A.), Shirley Nurock (Caregiver, U.K.),
Anne OLoughlin (Social Worker, Ireland), Peter Rabins (Psychiatrist, U.S.A.),
Jill Rasmussen (General Practitioner, U.K.), Barry Reisberg (former President
of IPA, Psychiatrist, U.S.A.), Karen Ritchie (Psychologist, France), Kenneth
Rockwood (Geriatrician, Canada), Joel Sadavoy (President of IPA representing
IPA, Psychiatrist, Canada); Barbara Sahakian (Clinical Neuropsychologist,
U.K.), Lon Schneider (Psychiatrist, U.S.A.), Masatoshi Takeda (representing
IPA, Psychiatrist, Japan), Anders Wimo (Health Economist, Sweden), Mirka
Wojciechowska (Caregiver, Poland).

353

354

C. Katona et al.

Appendix 2
IPA would like to thank the following sponsors whose financial support enabled
the consensus conference to take place:
Eisai Japan
Forest Laboratories
Lundbeck
Merz Pharmaceuticals
Myriad Pharmaceuticals (IPA Corporate Partner)
Novartis Pharmaceuticals, Corp.
Pfizer, Inc. U.S. Medical Team
Pfizer, Inc. Worldwide
Merck & Co., Inc.
Shire Pharmaceuticals
Wyeth Pharmaceuticals

Appendix 3
The following observers were present at the consensus conference:
Andrew Abbott (Wyeth Pharmaceuticals, U.S.A.), Kirsty Barttelot (Shire
Pharmaceuticals, U.K.), Megan del Valle (Pfizer Inc. North American Shared
Services, U.S.A.), Gina Eagle (Novartis Pharmaceuticals Corp., U.S.A.), Gail
Farfel (Novartis Pharmaceuticals Corp., U.S.A.), Jose Fojas (IPA Corporate
Partner, Myriad Pharmaceuticals, U.S.A.), Nanco Hefting (Lundbeck,
Denmark), Timothy Hsu (Eisai, U.S.A.), Wayne Laslie (IPA Corporate
Partner, Myriad Pharmaceuticals, U.S.A.), Tom McRae (Pfizer Inc. U.S.
Medical Team, U.S.A.), Galin Michailov (Merz Pharmaceuticals, Germany),
James Perhach (Forest Laboratories, U.S.A.), Larry Radican (Merck &
Co. Inc., U.S.A.), Benot Rive (Lundbeck, France), Elliott Schwam (Pfizer Inc.
North American Shared Services, U.S.A.), Amber Selwood (University College
London, U.K.), Albrecht Stoffler (Merz Pharmaceuticals, Germany), Ed Swabb
(IPA Corporate Partner, Myriad Pharmaceuticals, U.S.A.), Jina Swartz (Eisai,
U.S.A.), Lara Verdian (Eisai, U.S.A.), Jonathon Woods (University College
London, U.K.).