Diagnostic misconceptions? A closer look at clinical research on Alzheimer's disease
- Correspondence to Lara Katharina Kutschenko, Institute for History, Philosophy and Ethics of Medicine, University Medical Center Mainz, Am Pulverturm 13, 55131 Mainz, Germany;
- Received 2 June 2011
- Revised 13 June 2011
- Accepted 15 June 2011
- Published Online First 25 August 2011
The current focus on early intervention trials in Alzheimer's disease research raises particular ethical issues. These arise out of problems of validating study results and translating them into general practice for one thing and out of unwanted effects of an uncertain diagnosis for diagnosed people for another. The first addresses the demands of scientific research compared to those of medical practice, questioning how the medical value of clinical trials is evaluated. The second relates the scientific and medical value of early intervention trials to the normative value of an uncertain diagnosis. Are people who are diagnosed with a potential early form of AD in clinical studies patients proper—although they would not have been diagnosed with the given “disease” in non-research-oriented medical settings? The very problem of framing this question in terms of diagnostic misconceptions connects conceptual with ethical issues of research into preclinical stages of neurodegenerative diseases.
- allocation of health care resources
- applied and professional ethics
- embryos and fetuses
- genetic screening/testing
- government/criminal justice
- medical ethics
- philosophical ethics
- philosophy of medicine
- research on special populations
Four years ago, Mr Lorenzo's mother was diagnosed with Alzheimer's disease (AD). After having learnt that only some moderately successful symptomatic treatments exist, Mr Lorenzo took his mother to a large memory clinic that conducted clinical trials. He hoped that she could be included in research but she was already classified as being severely demented, which was an exclusion criterion for the clinical trials. The principal investigator explained that at this stage of the disease, too much brain matter had already been irreversibly destroyed, thus limiting the chances of success for potentially disease-modifying treatments.
Two years later, Mr Lorenzo got the impression that he himself had become more and more forgetful. He went to his general physician, who told him that he was merely getting old, but Mr Lorenzo did not want to be ‘too late’ again and eventually obtained an appointment at the memory clinic. Mr Lorenzo underwent indepth neuropsychological testing and a MRI and positron emission tomography scanning were performed. Blood work was done and cerebrospinal fluid was collected for biomarker assessment. The resulting diagnosis sounded similar to the one made by the general practitioner: Mr Lorenzo had no dementia but was experiencing cognitive decline. He was viewed as having early ‘mild cognitive impairment’ (MCI), a possible prodromal stage of dementia. That meant that he was eligible to participate in a clinical study on a new anti-amyloid therapy. The physician informed him about the study design and possible risks and explained to him that his participation would not guarantee that he would receive the tested drug. Mr Lorenzo was glad that he could help promote research on AD and hoped he would profit in some way from his participation in the clinical trial. He signed the informed consent and was included as a patient–subject in the trial.
The presented case is fictitious but the situation itself is not by any means: AD has become one of the conditions most intensively researched, with at least 50 compounds currently being tested in clinical trials and volunteers being recruited by more than 250 studies.1 Online services, for example, the recently established ‘TrialMatch’ of the Alzheimer's Association (AA), foster this development by setting up databases of ongoing clinical trials as well as of possible research participants in order to ‘match’ them.2 Matching refers not only to the location but also to the stage of the disease. More and more clinical trials on AD therapeutics include non-demented persons as research subjects because they aim at the prevention of dementia and early intervention during the long asymptomatic phase of AD. Promotion of research on very early stages of the disease seems to solve several problems at once: first of all, timely intervention into AD-derived neurodegeneration is regarded as essential for evaluating whether a given antidementive drug may be disease modifying or not. Second, such research gets around involving demented patients, who are seen to have an impaired capacity for giving informed consent to participation in clinical trials.3
This paper calls for revisiting how these issues and the consequences of an uncertain early diagnosis should be balanced. While there is great talk of the potential advantages of research on dementia prevention, little attention is given to the fact that this development does not come without a cost: obviously, not involving people who already show symptoms of dementia requires the recruitment of other study populations. At this stage, the involved research subjects are usually considered to have (amnestic) MCI. MCI is understood to be a transitional stage between ‘normality’ and dementia and its diagnosis aims to identify people who are at high risk of developing dementia. However, even proponents of the concept agree that the predictive value of MCI (or any of its subtypes) is limited: the annual proportion of those initially diagnosed with MCI who progress to dementia is only 10–15% and, moreover, a significant number of them returns to being considered cognitively ‘normal’ within 5 years.4 The uncertainty of MCI diagnosis challenges the ethical evaluation of clinical trials on early intervention in AD in at least two respects, namely by complicating the validation of study results and their translation into general practice and by raising the question of whether diagnosed people should be regarded as patients.
Trial design and evaluation are complicated because MCI is associated with a heterogeneity that seriously challenges the definition of a good study population. Given the annual reversion rates, it is difficult to assess whether an observed positive effect in a trial is due to the treatment tested or to therapy-independent reversion. This problem concerns the scientific value of MCI trials regarding the efficacy or internal validity, that is to say, reproducibility of results in a particular setting. Furthermore, it impacts on the purported medical value of the obtained study results in terms of their effectiveness or external validity.5 This is a particularly important issue for the application of research results in general practice, as I will lay out in the following.
One strategy for dealing with the loss of power is to include more participants for a longer period of time. This alone challenges the internal validity of study results, however, for example, with regard to declining compliance and higher drop-out rates of study participants over time.6 Another strategy increasingly employed is to change the outcome measure, aiming for biomarker evaluation rather than clinical diagnosis of dementia. This tendency has become evident in the recently published recommendations for updating the diagnostic criteria of the influential AA (National Institutes of Health (NIH)–AA criteria). The NIH–AA criteria workgroup presents a framework of subclassifications that refer to the model of AD progression along three states (‘preclinical AD’, ‘MCI due to AD’, ‘dementia due to AD’), the diagnostics applied (in the case of MCI ‘core clinical criteria’ for all healthcare settings, and ‘research criteria’ that include biomarker assessment), and the degree of certainty of the diagnosis based on the abundance of certain biomarkers.7 To date, several genetic, imaging and chemical biomarkers are under discussion as possibly providing partial indication of AD progression. However, none is considered to qualify as a true surrogate marker. This situation reflects a typical epistemological paradox: one enormous limitation on and at the same time main motivation for the search for surrogate markers or so-called biosignatures is the unclarified connection between (perhaps ‘non-normal’ but therefore not necessarily pathological)8 9 cognitive ageing and the development of dementia.10
Classifying degrees of diagnostic certainty by biomarker profiling may indeed be an important step for increasing the comparability of MCI study results; historians of medicine have stressed the importance of the standardisation of diagnostic practices as preconditions for the standardisation of drug trials in general and the evaluation of drug efficacy and effectiveness in particular.11–13 However, the NIH–AA recommendations also show that the ability to translate study results into clinical practice not only depends upon their internal validity but also on the extent to which the study conditions need to be transferred. This refers to practical issues, such as the availability and applicability of costly and invasive, thus risk-associated, diagnostic techniques (eg, positron emission tomography scanning and the collection of cerebrospinal fluid) and elaborate classifying procedures (eg, by means of sophisticated, time-consuming neuropsychological test batteries) in non-research-oriented medical settings. Therefore, already at the stage of clinical trial design—and especially against the background of the purported prevalence of (pre-)MCI conditions—not only risk–benefit but also allocation issues come to view. Some research ethicists, most prominently perhaps Joffe and Miller, have argued that issues that are not immediately associated with the scientific value of a clinical study should not guide ethical evaluations of clinical trials.14 This implies a strong focus on the internal validity of the study design that will, however, not necessarily give us the relevant evidence.5 15 Would Mr Lorenzo receive a treatment plan for a condition that does not (yet) give rise to serious problems in daily life activities (one criterion of the definition of MCI) if he were not included in a clinical trial? Given that more and more people who are afraid of having early-stage AD are looking for medical assistance, this question mirrors a real concern leading to new demands of patients and physicians alike.
The diagnosis of a prodromal state, even if it only occurs within a clinical trial, has a normative value9—in particular if it is connected with the threat of developing dementia. As a result, patients, their relatives and ultimately also the wider public are increasingly aware of early-stage AD and search for medical assistance. This is, for example, reflected within the recently published draft of the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders. It envisages the inclusion of diagnostic criteria for MCI and other ‘minor neurocognitive disorders’ because ‘as these conditions are increasingly seen in clinical practice, clinicians have a pressing need for reliable and valid diagnostic criteria in order to assess them and provide services including (…) appropriate early interventions’.16
In this framework, the question arises concerning the extent to which AD trials that include people who have been diagnosed with MCI or preclinical AD do actually recruit patient–subjects in the strict sense of the word. Are patient–subjects facing ‘diagnostic misconceptions’? This notion signifies the gap between identifying patient–subjects within clinical trials and diagnosing diseases or disorders of patients in the clinics: People who undergo extensive diagnostic procedures for research purposes are likely to be regarded as and/or to see themselves as patients proper although they would not have been diagnosed (in this case with preclinical AD or MCI) in non-research-oriented medical settings. It has long been noted that diagnostic technologies and risk assessment influence what is considered as belonging to the medical domain as well as patients' self-conceptualisations.17–22 The diagnostic misconceptions of patients have been discussed in the context of therapeutic misconception.23 24 The latter refers to patient–subjects who fail to recognise that participation in clinical studies may bring with them certain individual disadvantages, for example, placebo treatment, which arises from scientific rather than therapeutic constraints. Notably, not only patient–subjects but also physician–researchers can conflate research and patient care: diagnostic strategies of clinical research, which aim to obtain well-defined study populations (eg, invasive screening techniques), may be scientifically useful but not necessarily serve the needs of clinical practice.20
Clinical trials are one possibility to alter diagnostic conceptions of lay-people and physicians alike. This paper has sketched out how these diagnostic conceptions are shaping official diagnostic criteria and classifications. Standardised classification, again, stabilises diagnostic conceptions and drives the need for more research into early conditions. These interconnections shine some light on the very difficulty of conceptually and empirically identifying misconceptions—a fact that should, however, not be taken as an excuse not to question current conceptions and their effects on medical practice and on individual patients. Rather, questioning diagnostic, or for that matter, therapeutic conceptions, should raise awareness of the ethical obligations of physician–researchers and institutional review boards.25 Further research is needed to assess the implications for the toolkit of research ethics: how should the ‘side effect’ of becoming an AD patient be modelled in risk–benefit analysis and how could informed consent be correspondingly amended?
AD that is associated with a deep fear of losing one's self, that plays a pivotal role in advanced directives and, in The Netherlands, for euthanasia, requires not only much research but also particular attention and psychological care for the research subjects, namely people being diagnosed as potential dementia patients. These people are already in the spotlight of the pharmaceutical industry and of physician researchers. The time has come to let them also gain centre stage of research ethics.
I would like to thank two anonymous referees as well as Norbert W. Paul and Lara Huber for helpful comments on earlier drafts of this paper. The article draws on work of my dissertation “Towards an Epistemology of Medical Classification Systems” (in preparation) that is embedded in the German-French Doctoral Programme (funded by the German–"French University) of the Philosophy Department of the École normale supérieure in Paris and the Institute for History, Philosophy and Ethics of Medicine at the University Medical Center in Mainz.
Funding Bundesministerium für Bildung und Forschung, 01GP0807. The German Federal Ministry of Education and Research funded a project on ‘Normality, Normalisation and Cognitive Enhancement’ from 2008 to 2011. The author of this paper was supported by this grant.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.