Caring for homeless people with mental illness and distress - challenging professional assumptions and recommendations for changing practice.

presented at Trinity College Dublin: 

The original study was funded by The Scottish Office Chief Scientist’s Health Service Research Committee, St Andrew’s House Edinburgh  Scotland UK and supervised by Professor MF Alexander and Professor J McIntosh, Department of Nursing and Community Health and Professor D Walsh, Department of Social Sciences, Glasgow Caledonian University, Scotland.  Acknowledgements to Mr J Dreghorn, project nurse.

Abstract

This paper describes evidence of mental illness and distress amongst a group of single homeless men in Glasgow. The evidence was gained during a wider study of district nursing intervention with residents of a private hostel for the homeless (Atkinson 1997).  Another hostel was utilised for a comparison sample.

A demographic profile of the hostel dwellers with comparison and contrast between hostels was undertaken.  The data was gained through answers to a structured questionnaire containing,  26 questions and the scores derived from questionnaires regarding physical function (Barthel Index), mental state Hospital Anxiety and Depression  (HAD) Scale, the effects on the individual’s life of alcohol use.

Using the structured questionnaire 106 men were interviewed in the Main Study Hostel and 100 men were interviewed in the Comparison Hostel.

 Throughout the study psychiatric morbidity and psychological distress featured highly.  Amongst many of the men, the use of the HAD Scale proved relevant.  In relation to seeking out residents with possible biogenic depression, a large group of men were found, particularly in the Main Study Hostel.  The evidence suggested that many of these men could be treated and that their depression could have been causal to their homeless state, rather than, as was the almost universal professional assumption, that the men were depressed as a response to their environment.

Introduction

Glasgow, in Central Scotland, is the country’s largest city with approximately one million residents.  This large industrial city used to have many transient workers and hostels to cater for them.  In more recent years the number of transient workers has reduced and now the hostels mainly cater for homeless single men. The hostels, containing about 1500 beds city wide are managed by local authority, church and private enterprise initiatives (GCSH 1991).

 The author worked from one of the health centres in the east end of the city as a District Nurse.  His generic duties comprised providing a home care service for the patients of two general practitioners’ (GP) practices.  As well as these general duties he undertook to provide a service for hostel dwellers in the same geographical area.  His main activities at that time were immediate treatment and enabling men to become registered with a GP (Author 1987).  These actions promoted  the chances of men gaining access to mainstream care and ameliorated immediate distress but they left the author and the men many unanswered questions and dilemmas.  To this end the author and his academic supervisors gained funding from The Scottish Office Department of Health to set up a study to examine nursing intervention with this group.

The study

The main objectives of the study were to present health profiles, including the measurement of levels of anxiety and depression and a nursing assessment, of the residents of a hostel (undertaken by a project nurse) and a comparison hostel (undertaken by the author), to make assessments and referrals and to evaluate their effect (Zigmond and Snaith 1983) (Roper et al 1990).

These objectives were successfully met along with the secondary objective to discover insights into the residents’ experiences and lifestyles and their interaction with health and nursing services.  The objectives were addressed by gathering and analysing quantitative and qualitative data and the use of theoretical perspectives: Roy’s nursing theory (1980) of adaptation (to study the men as individuals) and a sociological perspective, including Deviancy theory, to examine the men as a group.  Quantitative data was analysed using the Statistical Package for the Social Sciences 1990.  Qualitative tools included the project nurse keeping an incident diary and the formation of comments which were collated into themes and exemplified by paradigms which were contemporaneously collected (Benner 1984) (Norman and Parker 1990).

The purpose of this paper is to demonstrate  the main strands of the study and analysis by presenting a summary and discussion of the findings specific to mental health and psychological distress.  This will be followed by recommendations for practice and future research.

Important overall findings

It is important at this point to highlight points which were discovered during the study which, with the benefit of hindsight, would change the approach to a similar study in the future.

 1.  Constraints

In the preparation for the study certain elements of the proposed study practice were not obvious.  This constrained, to a certain extent, the  scope of the study.  These constraints were time, access and lack of previous similar work.  the author did not know how stable the population of single homeless men was or whether a large proportion would be willing to  engage with the study and how much time they would be prepared to spend in the interviews.  These factors combined meant that the study and approach were pioneering in nature.  As it happened the population, particularly in the Main Study Hostel, was very stable, the majority were willing to engage in the study and, once recruited, were extremely obliging in terms of time  (with a few notable exceptions). 

Had this been known before the study commenced it may have been possible to use more complex validated tools, particularly in relation to physical and social function.  It may also have been possible to introduce into the design follow up interviews which may have given more precise insight into  clinical and personal outcomes.  Previous work in the area had suggested a more transient population (Featherstone and Ashmore 1988) (Williams and Allen 1989) (Atkinson 1987).

2.  Critique of the validated tools

Barthel Index

Taking into account the constraints it was decided to use a simple tool to define physical function and one which, as far as possible, was not environment or context specific (to ensure that the men’s homeless state did not distort the findings).  The Barthel Index (Mahoney and Barthel 1965) only measures physical function and therefore appeared to have the simplicity required for this study.  However as the findings demonstrate the outcome of its use was not as predicted.

Firstly it is more successfully used on a recognisable physically disabled population to discern levels of relative normal and abnormal function (it was originally used with patients who had suffered from stroke).  Most people in the general population score very highly and it is limited in detecting disability or functional problems if the individual is able to achieve the tasks by himself or if he tolerates a certain amount of dysfunction (taking 20 minutes to get to the toilet or urinary dribbling for example).

Secondly the tool and disability itself are entirely environment/context dependent.  One is disabled because one cannot achieve certain tasks in particular circumstances which therefore dictates where one can live, work and exist.  Because the Comparison Hostel had better facilities it was possible for disabled people to live there.  The population scores on the Barthel Index therefore described the context better than they described the individuals.

Taking the previous points together and the fact that the men usually demonstrated a willingness to spend time with the study, it may be possible in future studies to use more complex tools examining function and quality of life taking, particular care to exclude or adapt questions which would obviously provide ambivalent answers given the homeless context (eg questions regarding holidays, possessions, activities).

HAD Scale

Conversely the use of the HAD scale proved immensely valuable even though it too had been chosen, in part, for its simplicity, ease and speed of use.  The possible use of the Present State Examination (PSE) (Win, Cooper and Surtonus 1974) which was considered, would not have been appropriate (even in hindsight) as, even with the men’s cooperation, it would have taken too long (1.5 hours) and the PSE concentrates on psychotic illness which would not have discovered this study’s major findings regarding anxiety and possible biogenic depression.  There is also some question as to whether the PSE could be administered by any community nurse as a primary assessment tool.

3.  Limitations of the study

Taking the constraints and the critique of the Barthel Index into account it is essential to stress that this study was only of one hostel in one city.  Comparative elements were included, in particular the Comparison Hostel and by attempting to provide background context in the literature review.  The study does not therefore claim that the men in the study are representative of homeless men in Britain for example, although other workers in the area may draw resonance from the study with their own experience.

4.  Use of the District Nurse:  impact on the findings

 As described in Chapter 1 I came from a District Nursing background  and was motivated to initiate a rigorous study arising from exploratory work undertaken in the area of homelessness as a practising District Nurse.  For the purposes of an academic study it was also necessary to be as precise as possible when describing and setting up the study.

 However as an individual, and supported by the findings, I do not categorise nurses into tight professional bands and class myself only as a community nurse.  In terms of practice, what is more important is the clinical competency of the individual nurse and the ability/autonomy to proactively case find, assess and monitor the physical and psychological wellbeing of individuals in a defined group.

 A Community Psychiatric Nurse attached to a psychiatrist’s caseload in a hospital who had little experience of treating physical illness would have difficulty in undertaking the work described in this study.  However many community nurses now work across professional boundaries and it is suggested that the findings of this study are not specific to the District Nurse but rather identify a model of community nursing practice.

 The Study - results

Results from the questionnaire and validated tools were analysed using the Statistical Package for the Social Sciences (SPSS 1990).

 Age - Length of stay in hostel

 

The largest group of men in each hostel were aged between 51 and 70 years (50% in the Main Study Hostel, 53% in the Comparison).  There were more young men (< 25 years) in the Comparison Hostel which was used as an emergency admission unit by the Council. 

 

Two broad  groups of men emerged in both hostels, although not every resident was typical of either group.

 

(i)         Group One  tended to be older and had made a choice to live in hostel type           accommodation with other residents.  Many of them considered the hostel their             home.  This group had a lower prison record and often lower anxiety levels.

 

(ii)               Group Two  tended to be younger and had come to live in a hostel because of some break down in their family relationships, social circumstances or health.  Alcohol use, high anxiety levels and higher prison admittance were seen in this group. This group tended not to think of the hostel as their home but as a temporary alternative.

 

Hospital admissions

 

Within the previous two years, half of the men in both hostels had been hospital inpatients - predominantly as a result of diagnosed physical illness.  Only 10% of men in both hostels had never been in hospital.

 

Twenty one percent of the Main Study Hostel sample and 14% of the Comparison Hostel sample had been admitted to a mental hospital within the last five years.  There was little evidence that men had been discharged from long stay care directly into hostel accommodation.  There was evidence that some had been in long term care, discharged into supervised accommodation and at a later date become hostel residents following a breakdown in their social circumstances.

 

Residents’ perception of their state of health

 

The question “How do you feel ?” proved one of the most useful questions in the study.  Just over half of the sample in the Main Study Hostel and two thirds of the Comparison Hostel sample felt well.  Significant statistical relationships were found between those men who stated that they felt unwell or ill with several hospital admissions, high levels of anxiety and depression, high use of Accident and Emergency and with high impact of alcohol use on lifestyle.

 

Preferences for companionship and type of accommodation

 

Similarly asking residents “Where would you like to live?”, “With whom would you like to live?” and “Where is home?” provided great insight into residents’ present wellbeing, aspirations and adaptive processes.

Two thirds of the sample in the Main Study Hostel and of the Comparison sample wanted to live in a house or flat but a third of the men wanted to live in hostel accommodation.

 

 

Although the largest group in the Main Study Hostel wanted to live alone, a third wanted to live with their wife and/or family.  This evidence, borne out by the project nurse’s experience, showed that residents’ personal aspirations had an effect on how they felt at the present, and on their level of contentment with their circumstances and adaptation.  The project nurse found men who had lived in hostels for many years but who wanted to live with their families and tended to see their lifestyle as aberrant or dysfunctional because of this.  Conversely other men, who had lived in hostels only a short time, would be reasonably happy with their lot and considered themselves to be living a “normal” lifestyle.

 

The Hospital Anxiety and Depression Scale

Anxiety

 

The Hospital Anxiety and Depression Scale had been used in both hospital and community settings but it had not been used with homeless men.  In discussion, some health professionals were doubtful about the use of the tool with this group.  Some suspected that the men might give answers in order to gain attention.  This notion was disproved in the study.  The scores of 46% of the Main Study Hostel sample fell within normal score range, as did 65% in the Comparison Hostel.  Twenty per cent were in the mid range at the Main Study Hostel and 14% at the Comparison Hostel.  Thirty five per cent were in the high range at the Main Study Hostel and 20% at the Comparison Hostel.

 

The important feature about a high anxiety level is that, unlike a high depression level, it is susceptible to environmental factors.  There is also a well documented connection between anxiety and alcohol use.  These factors were borne out in the study.  Significant relationships were seen between high anxiety levels and environmental factors such as length of residency, whether individuals wished to live in their hostel and individuals’ feelings of well being.

                                                                                                           

Depression

 

In the Main Study Hostel only 38% of the sample had scores which fell within the normal range, 35% had scores which fell within mid range and 27% had scores which fell within the high range.  In contrast the majority of the men in the Comparison Hostel, 71%, had scores which fell within the normal range, 18% within mid range and 11% fell within the high range.  It is important to highlight that the express purpose of using the HAD depression tool was to isolate treatable clinical, biogenic depression by measuring “anhedonia”, the loss of the ability to experience pleasure.  This condition has been found not to be susceptible to environmental factors or alcohol use.  Comparison with these variables discovered that  residents with high depression scores were found to be equally distributed throughout the impact of alcohol and length of stay comparisons.

 

These findings were presented to Dr R P Snaith at the University of Leeds who developed the HAD tool.  Dr Snaith estimated that the occurrence of this form of non reactive depression in the general population was 5% or less.  The levels in both hostels, therefore, are high especially in the Main Study Hostel.  This discovery became one of the most important findings of the study.

 

Impact of Alcohol on Lifestyle

 

The purpose of the Impact of Alcohol on Lifestyle Questionnaire was to measure the effects of alcohol consumption, rather than the amount of consumption.  This tool proved useful when results were compared with other variables, particularly anxiety and prison record.  Over half the men in the Main Study Hostel sample scored within the two higher score levels compared to just under half in the Comparison Hostel sample.  These show a very high impact of alcohol on both populations.  However it is important to point out that not all the men drank alcohol or drank alcohol to excess.

 

Discussion

 

Mental health:  Use of the HAD Scale

 

Throughout the study psychiatric morbidity and psychological distress featured highly.  Amongst many of the men, the use of the HAD Scale proved very relevant.  In relation to seeking out residents with possible biogenic depression, a large group of men were found, particularly in the Main Study Hostel.  The evidence of the expert witness who devised the scale, (see Chapter 7) suggested that many of these men could be treated and that their depression could have been causal to their homeless state, rather than, as was the almost universal professional assumption, that the men were depressed as a response to their environment.

 

In the case of the Anxiety scale, a strongly positive relationship was found between high scores on the HAD scale and alcohol use problems.  Also the group of mainly older men, who considered the hostels their home and preferred to have the companionship of other residents, tended to have lower HAD anxiety scores, whereas the group of mainly younger men, leading a more chaotic life, who did not feel at home, tended to score highly on the HAD anxiety scale.  The HAD scale thus assisted in the definition of these two groups.

 

It may be the case that if individuals suffering from depression, in particular, and some of those suffering from anxiety, were treated systematically, then a number of these people may be able to motivate themselves to take positive action.  This may take the form of increased exercise of personal responsibility and/or uptake of mainstream services.  It is difficult to see how an individual, living in homeless accommodation, and suffering from depression, can move forward in his life without help

 

 

The author and project nurse did not feel that they possessed the expertise to agree with or repudiate these reactions, although they did know from the literature that the HAD Scale had been used in a variety of settings.  As the responses were a common reaction of health professionals who received referrals it was important to present the HAD scores to an expert witness, who could add other dimensions to the discussion.

 

It was decided, therefore, to present the findings to Dr RP Snaith, Consultant Psychiatrist at St James’s University Hospital, Leeds, and originator of the HAD Scale.  Following this decision the author and project nurse visited Dr Snaith and interviewed him, presenting him with the reactions of the health professionals.  The results of this interview are presented in this chapter. 

 

The author sent Dr RP Snaith anonymised results from the statistical analysis of the scores from the HAD Scale gained during the interviews with residents from the Main Study Hostel and the Comparison Hostel.  These comprised  frequencies - how many men had scored in each of the three categories (normal, medium and high) with the concomitant percentage relationships.

 

The author and project nurse visited Dr Snaith in Leeds and asked him in a semi structured interview, to comment on the findings and to answer some of the questions which health professionals had presented to the author and project nurse.  What follows is a summary of that meeting.

 

Dr Snaith began his response in the interview session by describing the development of the HAD Scale. It was developed as a response to the high number of apparently depressed people passing through a hospital Out Patient Department and to the need to find some way by which to filter out those who were in need of intervention / treatment (with biogenic or clinical depression) (Snaith 1991). 

 

 

The scale identifies loss of the ability to experience pleasure,  ‘Anhedonia’, as the indicator of depression (Snaith 1992).  When asked by the author about the relatively large numbers of residents found to have high HAD scores, Dr Snaith said he was not surprised and described a situation where someone suffering from biogenic depression  becomes a “flounderer”, unable to cope with life, going from one crisis to the next and drifting into homelessness. He commented that these may well be people who have slipped through the health care net, never having been diagnosed as suffering from depressive illness.

 

He went on to talk about the relationship between alcohol and anxiety, stating that although there was a relationship, it was not always obvious which came first.

In relation to context or environmental influences on the HAD scores, he stated that environment should have a minimal effect on HAD (Depression) but a greater effect on HAD (Anxiety).

 

Dr Snaith was very interested in the HAD (Depression) results from the study. He thought that the relatively high numbers (approximately 14%) of residents who scored 10 or more was significant, and this compared to an expected frequency of

around 5% in the general population.  He did not think that this could be explained by environmental factors and asked whether any of these men had been prescribed anti depressant therapy as a result of the study.  When given the example of someone with a HAD (Depression) score of 14 he stated that he would expect some form of intervention to take place for someone scoring at this level.       

 

The ‘Hospital’ prefix, Dr Snaith stated, was used because the tool had originally been developed for use in hospital, but because, in his view, it was non contextual, the tool could be used in most areas of practice.  Dr Snaith was aware that the tool had been used in several studies in community or non institutional settings and with different groups, for example in

1.         a study of Swedish mothers of retarded children,

2.         an Asian Clinic study,

3.         a study into elderly dementing patients attending a geriatrician, and

4.         a psychiatric Outpatient Clinic.

 

He considered that the HAD Scale was an appropriate tool to use with the particular sample in the study i.e. homeless men.  He also stated that it had been used in a wide variety of environments and with widely differing groups.

 

Asked if there were other validated tools which could have been used to achieve similar results, Dr Snaith  cited the Present State Examination (PSE) (Win,Cooper and Surtonus 1974) as a possible tool to screen for psychotic illness.  When asked how long it would take to administer this tool he stated this would be an hour and a half approximately. 

 

To return to the three reactions presented to the author and project nurse by the health professionals receiving referrals....

 

“Of course he’s depressed, he lives in that terrible place.”

“ The HAD Scale was made for hospitals and is not relevant here.”

“We are looking for treatable psychotic illness. These other illnesses are not treatable in these conditions”......

 

the interview with Dr Snaith certainly presented another view.  Firstly, at least some of those with high depression scores, if Dr Snaith was correct, would not have been depressed because of their environment;  indeed they may have become homeless as a result of their depression.  Secondly as Dr Snaith had predicted during his interview there was a statistically significant relationship between alcohol use and anxiety scores in that high impact of alcohol on lifestyle scores correlated to high anxiety scores.

 

Thirdly it would appear that the prefix “Hospital” is rather an indication of where the tool originated, not an exclusion factor from the evidence that the HAD Scale has been used in a variety of institutional and community settings, with success.

Finally, Dr Snaith stated that biogenic depression is a treatable condition which tends not to be influenced by environmental factors.  The health professionals who received referrals, based on the men’s HAD score, were pessimistic about the success of treating residents with depression.

 Primary versus secondary care patterns of assessment

Further to the discussion on the specialist and generalist nursing roles of community nurses, the CPN’s responses (see Chapter 6) demonstrated how the main focus of the psychiatric team’s approach was to seek out residents suffering from psychotic illness, considering residents with other mental illness as being more able to seek help independently elsewhere. 

 

 

 

 

This arrangement of priorities demonstrated a highly specialised, secondary (or hospital) care model of assessment as opposed to a primary care approach which would, perhaps, concentrate more on the effect that an individual’s mental health was having on his ability to function normally.  The primary care approach would, possibly, identify and be of more benefit to the individuals suffering from depression.

 

Theoretical and practical links

The care of the marginalised

 

The pattern of recruitment employed by the project nurse, a mixture of opportunistic encounters with the hostel residents and targetted letters, proved effective at reaching the sample population.  It is suggested that this method, used alongside the assessment pack, which proved easy to administer, would be an effective way of monitoring other vulnerable and marginalised groups in society.

 

In Chapter 2 it was shown how the health and social profiles of the homeless, and particularly their difficulties in accessing mainstream care, were also seen in other groups, for example, prisoners, “travelling people” and the single elderly.  In Glasgow there are several areas which have a concentration of elderly people, and many “travelling people” in the city live on recognised sites.  Using the same nursing assessment and intervention techniques, as were used in this study, may be of benefit to these individuals.

 

In a research study such as this one it is important to specify a particular sample group, in the case of this study, single homeless men.  However I have come to the conclusion, from the study of the literature and the experience of the study, that the methods and approach would be relevant and appropriate for the study and care of other marginalised groups, including those mentioned above.

 

Recommendations to Service  Providers

 

The study found that many service providers had deeply held preconceptions about homeless people, such as, that residents were not interested in their health.  The study found that many of the men were interested in their health, had treatable conditions, for example biogenic depression, and were prepared to attend appointments and undergo treatment.

 

The study also found that some service providers were prepared only to address the presenting complaint for each client without instituting ongoing monitoring.  Most of the men were found to come into contact with primary health care teams on a regular basis.  It recommends that this provides health services with an ideal opportunity to set up assessment and monitoring programmes.

 

It should be recognised that many individuals and groups will not take an assertive role in expressing their health and personal needs.  Some people will prefer to hand over this responsibility to others while others are not capable of expressing themselves. So whereas “empowering” patients to express their specific needs as consumers, thus enabling nurses and others to react appropriately, may be a successful strategy in general, it may not be successful with particular individuals.

References

Atkinson J (1997)  A Descriptive and Evaluative Study of District nursing Intervention with Single Homeless Men from a Private Hostel in Glasgow.  Unpublished PhD thesis.  Glasgow Caledoniain University.

Ross F  (1993)  Editorial. Journal of Interprofessional Care Vol 7 No1 p6

Smith S J, McGuckin A, Knill Jones R eds  (1991)  Housing for Health.  Longman Harlow.

Snaith R P, Taylor C M  (1985)  Rating Scales for Depression and Anxiety:  A Current Perspective.  British Journal of Clinical Pharmacology  19  pp 17S-20S.

Snaith RP(1987) The Concepts of Mild Depression. British Journal of Psychiatry Vol 150  pp 387-393.

Snaith R P. (1991): Measurement in Psychiatry, British Journal of Psychiatry   Vol 157P. 78-82.

Snaith RP  (1992)  Anhedonia: exclusion from the pleasure dome - A useful marker of biological depression.  British Medical Journal Vol 305 p 134.

Zigmond, A S and Snaith, R P (1983)  The Hospital Anxiety and Depression Scale.  Acta. Psychiatr. Scand. Vol 67 : pp 361-370.

 

(GCSH 1991).

(Zigmond and Snaith 1983) (Roper et al 1990).

(Benner 1984) (Norman and Parker 1990).

(Featherstone and Ashmore 1988) (Williams and Allen 1989) (Atkinson 1987).