Article Text


Dementia in prison: ethical and legal implications
  1. S Fazel1,
  2. J McMillan2,
  3. I O'Donnell3
  1. 1Section of Old Age Psychiatry, Department of Psychiatry, University of Oxford, Oxford, UK
  2. 2Centre for Ethics and Communication, University of Oxford, Departments of Public Health and Primary Care, Institute of Health Sciences, Oxford, UK
  3. 3Institute of Criminology, Faculty of Law, University College Dublin, Dublin, Ireland
  1. Correspondence to:
 Dr S Fazel, Department of Psychiatry, University of Oxford, Warneford Hospital, Oxford OX3 7JX, UK;


As the number of elderly prisoners increases in the UK and other Western countries, there will be individuals who develop dementia whilst in custody. We present two case vignettes of men with dementia in English prisons, and explore some of the ethical implications that their continuing detention raises. We find little to support their detention in the various purposes of prison put forward by legal philosophers and penologists, and conclude by raising some of the possible implications of The Human Rights Act 1998.

Statistics from

There are about nine million people held in penal institutions around the world. America now detains over two million persons in custody (a rate of 680 per 100 000 population), and there are about 65 000 people in prisons in England and Wales (a rate of 125 per 100 000).1 These numbers are projected to rise, and the British government expects there will be another 18 000 prisoners by 2007. The number of elderly prisoners is also rising. There are now more than 1 000 men aged 60 and over in English and Welsh prisons, more than three times the level of a decade ago.2 The proportion of elderly compared to younger adults in prisons has also been increasing, and a similar trend has been reported in the US and Canada.3 The implications of the changing demographic structures of prison populations have been discussed elsewhere but we are not aware of any exploration of the ethical issues raised by individuals developing dementia during their imprisonment.4

The large studies on the psychiatric morbidity of prisoners have excluded those over 65. They have found that the burden of treatable mental disorders is substantial, and that rates of psychotic illnesses and major depression are 2–4 times higher than community samples of similar age.5 We recently conducted a study on a group of 203 male prisoners in England and Wales aged over 59, representing a fifth of all elderly sentenced men in the prison system.6 Two cases of dementia were identified using a reliable and valid diagnostic instrument, a semi-structured standardised clinical interview called the geriatric mental state schedule (GMS). Geriatric-mental-state-generated cases of dementia have been shown to correlate well with DSM III diagnoses of dementia.7 A brief neuropsychological test—the mini-mental state examination (MMSE)—was administered to screen for cognitive impairment and a measure was taken of their ability to perform tasks associated with independent living (Barthel index). Both individuals developed dementia whilst serving their prison sentences. They remained in prison, probably due to their lack of any overwhelming physical disability. This paper will present brief details of these individuals, and then discuss some issues that their continuing custody raises.


Case 1

At interview, AB was 69 years old. He was convicted at the age of 66 for incest, and was given a six year sentence. In the third year of his sentence, he suffered a left posterior parietal infarct, which was identified on brain scanning by computed tomography (CT). He was interviewed four months after this stroke, and scored 12/30 on the MMSE (where a score below 24 indicates a high likelihood of dementia). After the GMS had been adminstered, he was given a diagnosis of vascular dementia. He remained physically independent for activities of daily living, scoring 20/20 on the Barthel. He had no recollection of his crime and no idea that he was in prison. He was so demented that he could not answer any of the questions put to him during interview. Information on previous convictions was not available because the local prison database only records the index offence.

Case 2

CD was 78 years old. He was convicted of the serious sexual assault and murder of a child, and received a life sentence. He had been continuously in custody for well over a decade. In 1997, his medical notes recorded mild cognitive impairment that had progressively worsened. On interview in 1999, he scored 9/30 on the MMSE. He was given a diagnosis of Alzheimer's dementia. His Barthel score was 17/20. He was housed on the ground floor as he needed help with mobility, and was unable to climb stairs. He was insightful about his crime, expressing remorse: “I am in for life, and I should continue to be in for life. I would never kill another child, but who could tell what you're likely to do”. He kept a photograph of his victim in his wallet, remembering his name, explaining that he “never got over it”. He explained that alcohol intoxication contributed to his crime: “If I hadn't been drinking, [I] wouldn't have strangled a little boy”. He read one of us [SF] part of his last annual sentence planning review that described him as having “a sexually deviant personality” with “sadomasochistic fantasies” and remarked that: “I don't have none of this now”. Information on previous convictions was not available.


The presence of men with dementia in prison raises important ethical issues. The various purposes of imprisonment, and punishment more generally, have been extensively discussed by penologists, philosophers, and those involved in public policy.8 There are legal purposes to imprisonment—prisons are necessary to hold those awaiting trial or sentence, to enforce court orders, and to incapacitate convicted offenders. Prison is also thought to have some deterrent value, both to the individual under punishment and to the general public. Finally, prisons may also serve symbolically as a means for the state and the public to express their collective denunciation of criminal behaviour.

Rehabilitation is an important goal for prisons. There are essentially two ways that prisons can rehabilitate individuals—first, by positive action such as access to work, training and education, or through exposure to various types of treatment programmes. Second, by addressing the factors that contribute to offending (unemployment, drug abuse, temperament), the prison regime may reduce the risk of re-offending. The rehabilitative ideal waned in the early 1970s but there has recently been a resurgence of interest in “what works” for offenders.9,10 All of these ideas about prison focus on its consequences—deterrence, rehabilitation—and are embraced as primary rationales for punishment because they are believed to have the potential to reduce the crime rate. Prison is also seen as a deserved punishment for certain offences. In this penological tradition, the primary rationale is desert and the key criteria are the seriousness of the crime committed and the offender's culpability. The emphasis is on retribution rather than crime reduction.

How relevant are these purposes of prison for individuals with dementia? For those who developed dementia before sentencing, it is unlikely that they will be fit to stand trial and hospital diversion is the most likely outcome. For those who develop dementia whilst in prison, as in the two cases presented above, we will explore the relevance of various purposes of prison.


At first there might appear to be some reasons for thinking that keeping people like AB and CD in prison might add to the general deterrent effect of imprisonment. Those contemplating similar crimes would know that if caught they would be punished even if they were unwell. This knowledge would strengthen the disincentive to perform illegal actions of this kind.

The problem with this argument is that prisons already place people with other severe illnesses into health care facilities. This means that, in order to keep people like AB and CD in prison to add to the general deterrent effect of imprisonment, people contemplating crime would need to consider the possibility that they would not be released from prison even if demented. Studies of offender decision making show, however, that the likelihood of arrest has greater deterrent value than the severity of punishment, and that increasing penal severity has a modest effect on the decision making calculus of individual offenders.11

What then about the effect of individual deterrence? Might the prospect of continued incarceration reduce the incentive to commit crimes for those who already have dementia? This argument does not work as people who are as demented as AB and CD are not able to think through the consequences of criminal actions. It might be objected that those with early dementia might think that even if caught they would be released when they became sufficiently unwell. We do not have the same concern, however, that transferring out of prison those with other illnesses creates an incentive for those at the early stages of illnesses to commit crimes. It would be unethical not to release into care those suffering from advanced cancer because it might make others who have just been given a diagnosis of cancer more likely to commit crimes. Once diagnosed, their primary role is patient rather than prisoner.


There is evidence from studies carried out in the courts that individuals with mild dementia commit crimes,12 and it is therefore possible to argue that by detaining them in prison, society is protected from any further offences. It may be that some would feel that prison serves this purpose in the case of CD, who himself admitted that he was unsure whether he would reoffend. The parole board, who regularly review the case of CD and other life-sentenced prisoners, may have felt that he continued to pose a risk to the public, and this may be the reason for his continued detention. But there will be a point at which the level of dementia of such prisoners as AB and CD renders them incapable of committing further crimes, particularly if they are also physically disabled. When this point is reached, the security of a prison is not required to prevent further offending.


This leaves the question of whether prisons serve a symbolic role in detaining individuals with dementia, reflecting the desire of the state and the public to express their collective disapproval of crime. There is little doubt that some elements of the general public support the incarceration of individuals with mental illness; this was reflected in the trial of Peter Sutcliffe, the so-called “Yorkshire Ripper”, where the jury refused to accept the views of psychiatrists and sent Sutcliffe to prison (from where he was moved to a secure hospital). Public opinion should not determine public policy on the disposal of mentally disordered offenders just as it does not determine policy in other health care areas. The horror felt by most people at CD's crime may mean that his continued detention in prison serves a symbolic purpose for society. As a mandatory lifer, his release decision rests with the home secretary, and thus its timing may be influenced by political considerations.


Rehabilitation through offender treatment programmes, education, and work are of little relevance as they require cognitive abilities incommensurate with dementia. Addressing factors that may have contributed to offending will make little difference, either because they are no longer present, such as drug abuse, or cannot be addressed, such as personality problems. Treating co-existing severe psychiatric illnesses, such as psychoses, may, however, reduce the likelihood of further offending. Research has shown that aggressive behaviour in elderly persons in a nursing home was associated with the presence of delusions,13 and that effective treatment of psychosis is possible in the elderly. Overall, however, it is likely that the risk of recidivism on release to the community, hospital, or nursing and residential homes is low in individuals with dementia, although they may show some aggression towards staff and other patients. The risk of future aggression is not sufficient reason to incarcerate these individuals, and specialist nursing homes would be better equipped to manage these behaviours than prisons.


We have left our discussion of retributive punishment until last because it is the most controversial and difficult issue. It is controversial because many people think that once the deterrent and rehabilitative functions of punishment have been performed, there is no additional reason to punish, so long as the individual concerned no longer poses a threat to public safety. People who hold this view would argue for the release of AB and CD. It is a difficult issue because there are so many supposed defences of retributive punishment. As well as having a basis in many religions, it is also an idea with a long philosophical history. Thinkers such as Kant, Hegel, and Mill have written about it. It has also been given attention in more recent years.14–16 Philosophical opinion about which theory of retributive punishment is most relevant is so divided that Nigel Walker has distinguished between 15 different schools of thought.17 Clearly we will not able to survey all of the theories and compare their claims with the situations of AB and CD. What we will do is outline some of the basic views of one of the more prominent (albeit controversial) accounts of retributive punishment and show that even on this account there are good reasons for ceasing the punishment of AB and CD.

One of the more recent attempts to develop a coherent notion of retributive punishment was made by Robert Nozick.18 He thinks that in addition to its deterrent and rehabilitative functions, punishment is important for what he calls “connecting with correct values”.19 This is based on the presumption of a rational offender. The vocabulary of punishment does not apply with the same force to the actor whose capacity to reason has become clouded. If what is required in order for people to be reconnected with correct values is that they come to see the actions for which they are being punished as wrong and ones that they should not have performed, then there would seem to be little point in continuing to punish AB and CD. AB does not know that he is in prison and cannot remember what it is that he has done so there is little prospect of him re-evaluating his past preferences. CD is already remorseful for what he did, so for different reasons continued punishment holds little purpose for him. The problem is that Nozick has something different in mind when he discusses connecting with correct values. He thinks it's not just about people appropriating the correct values but has more to do with them being compelled to live in accordance with correct values.

Correct values are themselves without causal power, and the wrongdoer chooses not to give them effect in his life. So others must give them some effect in his life, in a secondary way. When he undergoes punishment these correct values are not totally without effect in his life (even though he does not follow them), because we hit him over the head with them. Through punishment, we give the correct values, qua correct values, some significant effect in his life, willy-nilly linking him up to them.20

This seems to be highly applicable to AB and CD. Both of them have been found guilty of sexual offences with minors. Offences of this sort are ones that exhibit values and sexual preferences that most people find utterly unacceptable. Punishing them for these crimes serves to reconnect them with appropriate values even if they do not come to see that what they did was wrong. If we think of a thief who believes, and will continue to believe, that breaking into houses and cars is justifiable, we may want this person to be punished or reconnected with correct values even if he is not going to change his personal ethical system. What is still problematic is the level of dementia that is consistent with reconnection of this sort. For Nozick it is sufficient that people know they are being punished.

Punishment does not wipe out the wrong, the past is not changed, but the disconnection with value is repaired (though in a second best way); non-linkage is eradicated. Also, the penalty wipes out or attenuates the wrongdoer's link with incorrect values, so that he now regrets having followed them or at least is less pleased that he did.21

The suggestion here is that if we cannot get people to reform then a second best is for them to regret having followed the values they did. This means that there is a lower threshold for those capable of meeting the point of retributive punishment. If people are incapable by reason of dementia of re-evaluating inappropriate values then it might be sufficient that they know others think what they did was bad, and that they are being punished for this. He suggests the following condition:

It is not so very difficult to get someone to understand that they are being punished because others view what they did as wrong, and intend for them to realize that this is happening.21

This may apply to the ordinary case where somebody is punished and is not prepared to adjust their values. Merely being able to know that others disagree and that this is why they are being punished is sufficient. What is less clear is whether AB and CD satisfy this condition. AB does not know that he is being punished or what he has done. This is a good reason for thinking that it does not make sense on Nozick's account of retribution to continue to punish him. CD, on the other hand, does know what he's done and why he's being punished so meets Nozick's condition.

It does not follow from this that we ought to relocate AB and not CD to a non-punitive institution. Even if we think that Nozick is right, it does not imply that imprisonment ought exclusively to be about retribution. There are powerful arguments that its function is limited to rehabilitation and prevention or that a variety of penal purposes compete for priority.22 Even within a rigid retributive framework we should look more carefully at the continued detention of prisoners with dementia.

Ethically, the same question can be posed towards the incarceration of those with severe physical chronic illness. Is it appropriate to keep someone with a severe progressive cancer in prison? What purpose does it serve? The introduction of the Human Rights Act 1998 (adopted from the European Convention on Human Rights), which came into full force in the UK on 2 October 2000, may assist in the clarification of how justifications for punishment may change in accordance with the deteriorating health of a prisoner. Article 3 forbids inhuman or degrading treatment or punishment of those in detention. The act also stipulates a new statutory duty for all public authorities to act compatibly with human rights (section 6).23 Good practice in the psychiatric care of the elderly would incorporate a community orientation, a multidisciplinary approach, an emphasis on abilities as well as deficits, and the aim to improve quality of life rather than simply to alleviate symptoms.24 It requires the development of skills and facilities that are specialised, and unlikely to be met in the prison setting. The lack of appropriate health care for sick prisoners has been declared as a form of inhuman or degrading treatment. In the case of a prisoner who developed a severe depression during an unusually long and stressful remand period, a violation of the convention was confirmed by the European Court.25 On two occasions, governments have had to demonstrate that high levels of medical care were available for chronically sick prisoners in order to counter charges under article 3 of the convention.25 In many ways, then, the groundwork has been laid for a test case involving someone detained in custody with dementia. As a ruling has been made in favour of prisoners with chronic physical illness, it would seem unlikely that a radically different approach would be taken to those experiencing similar difficulties as a result of dementia.


We are grateful to Professor Tony Hope for his comments on earlier drafts of this paper. Seena Fazel was supported by a grant from the Wellcome Trust.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.