As researchers uncover more about the genetic and biological markers of the common causes of dementia, such as Alzheimer’s disease⁷, clinicians will have to address the ethical challenges of the presymptomatic stage of dementia. Can genes, biomarkers or imaging⁸ – or some combination of these – reliably predict a person’s risk of developing dementia? Communicating such prediction models to patients presents clinicians with real chal-lenges. For example, the use of the apolipoprotein E4 allele in predictive testing – which initially indicated a threefold increased chance of developing Alzheimer’s disease if heterozygous for the allele but with quite poor specificity⁹ – already suggests that significant challenges face both the clinician in interpreting the meaning of predictive tests for patients and patients in applying probabilistic knowledge to future decisions. Similar questions and challenges apply to so-called biomarkers of common causes of neurodegenerative dementia such as tau protein in spinal fluid.

An approach to presympotomatic testing that addresses these issues is as follows. Patients are entitled to good quality predictive testing, if available, given the value that this information may have for some. In accordance with the Genetic Nondiscrimination Act of 2008 (GINA), this information has at least the same legal protections as other med-ical information¹⁰. The clinician (and the medical system) are only obligated to provide tests, however, that significantly improve upon current predictive information (e.g. family history, neuropsychiatric testing) and which the patient can understand and reasonably make use of (e.g. not make dramatic changes to future plans based on small differences in probability). The decision to provide a predictive test is an important clinical decision and, like the provision of other tests or treatments, requires clinical judgement. An essential clinical skill in exercise of this judgement is how to assess a patient’s understanding and appreciation of such information. Below, we discuss techniques to do this.

The first ethical issue that many clinicians face with regard to the detection of dementia is when, how and if to deliver a diagnosis of dementia¹¹. The question of ‘if’ is largely settled as a matter of policy. Patients are morally and legally entitled to their diagnosis. Although this approach has been widely endorsed by the Alzheimer’s Association¹²,the AMA¹³ and a consensus statement¹⁴, studies sug-gest that this policy has not fully filtered down to practice. Surveys of caregivers suggest that some physicians often do not disclose the diagnosis to the patient¹⁵. Instead, they disclose the diagnosis to the family¹⁶. And many generalists¹⁷ and dementia specialists^18–20 choose not to disclose a diagnosis of dementia, particularly at a more advanced stage when comprehension is seriously compromised. This practice persists despite evidence that most cognitively intact adults say they would want to know if they were diagnosed with dementia²¹ (though the strength of this desire seems to wane with the increasing age at which this diagnosis would be made²²). This obviously pater-nalistic practice reflects the complexity of a disease with progressive cognitive decline that impairs a person’s ability to adequately under-stand relevant information. The most compelling reason to disclose even at later stages of disease is that the patient with dementia is owed the chance, however small in some cases, of understanding (or gaining a glimpse of) why the things that are happening are happen-ing and what this portends for the future. To deny the patient this, whether to make things easier for a caregiver or to prevent purported psychological harm to the patient (‘therapeutic privilege’), is to deny an opportunity, perhaps the most important and last opportunity, for self-knowledge left to the patient with dementia.

A reasonable approach to guide whether to disclose a diagnosis to a patient is to converse with the patient about his or her awareness of cognitive problems, and, if the patient is aware of these problems, about the patient’s desire to know what is their cause²³. Useful ques-tions to elicit these symptoms include Are you having problems with your memory? and Do your memory problems bother you? If a patient endorses these symptoms, a follow-up question is Would you like to know what is the cause of these problems? In the case of a patient who has awareness of problems with memory and a desire to know the cause of these problems, a clinician has a clear warrant to dis-close the diagnosis. Two useful techniques include asking the patient what he or she knows about Alzheimer’s disease and clarifying any misunderstandings about the disease (One of the most common causes of memory problems is Alzheimer’s disease. Have you ever heard of that? Tell me what you know about it?). This needs to be done in a way that balances truth with kindness. For example, after disclosing a diagnosis to a patient, it is sensible to ask how the patient feels about this diagnosis and whether any questions or concerns remain.

The central issue that is at the heart of many of the ethical concerns during the symptomatic stage of dementia care is the loss of decision-making capacity. Dementia progressively undermines the abilities necessary for competent medical decision making (e.g. understand-ing, appreciation, reasoning and expression of choice²⁴) and con-sequently undermines the capacity for giving informed consent or refusal for treatment or research. While the development of some clinical instruments has shown some promise for aiding the assess-ment of capacity^25–27, the determination of whether someone has sufficient capacity is still ultimately a clinical judgement.

When a clinician does judge that a patient has lost the capacity to make a decision, the appointment of a surrogate provides a way to protect and promote the patient’s interests. In clinical practice, a surrogate tends to appear naturally as part of that person’s other roles as knowledgeable informant and caregiver. Often this person is a spouse, partner, adult child or sibling. A surrogate can be designated preemptively by the patient (via an advance directive) or appointed through established (often state) standards for next of kin. The raison d’être of the surrogate is to make decisions on the patient’s behalf.

Acceptable standards for how the surrogate makes decisions vary²⁸. Traditionally, ethics has posited a hierarchy of standards. The surrogate can try to decide as the patient would have decided given what is known of her previous beliefs, values, character and so on (called the substituted judgement standard). Absent the ability to fulfil this standard, the surrogate should decide based on what the surrogate believes is in the best interest of the patient (best interest standard). Studies suggest that most older adults do not want a surrogate to practice this hierarchy of standards, but instead to exercise a leeway that balances the two standards. In a study of 150 dialysis patients, two thirds preferred that their surrogate exercise at least ‘a little leeway’ in interpreting their advance directives to withdraw dialysis were they to develop advanced Alzheimer’s disease²⁹. This has substantial implications for clinicians when they talk with surrogates. Specifically, it suggests a useful approach to guiding surrogates in decision making; consideration should be given both to what preferences, if any, their relative told them, and what they think will maximize their relative’s dignity and quality of life. A surrogate decision should be challenged, in court if necessary, if clear evidence suggests that the surrogate is not acting with the patient’s interests in mind. In the absence of such, the presumption lies in favour of the surrogate knowing how best to act on the patient’s behalf.

Patients with dementia can formulate instructional