Hospital admissions from nursing homes: a qualitative study of GP decision making

Abstract

Background

Decisions regarding the hospitalisation of nursing home residents may present a difficult dilemma for GPs. There are pressures to admit very frail patients with exacerbations of illness even though such frailty may limit the possible health gains. As ‘gatekeepers’ to NHS, GPs are expected to make best use of resources and may be criticised for ‘inappropriate’ admissions. Little is understood about the influences on GPs as they make such decisions

Aim

To explore GPs views on factors influencing decisions on admitting frail nursing home residents to hospital.

Design and setting

A purposive sample of 21 GPs from two counties in the South of England.

Method

Data from semi-structured, one-to-one interviews with GPs were analysed using thematic analysis following principles of the constant comparative method.

Results

This study suggests that while clinical assessment, perceived benefits and risks of admission, and patients’ and relatives’ preferences are key factors in determining admissions, other important factors influencing decision making include medico-legal concerns, communications, capability of nursing homes and GP workload. These factors were also perceived by GPs as influencing the feasibility of keeping patients in the nursing home when this was clinically appropriate. Key areas suggested by GPs to improve practice were improving communication (particularly informational continuity), training and support for nursing staff, and peer support for GPs. Local initiatives to address these issues were very variable.

Conclusion

Developing a systematic palliative care approach to address poor documentation and communication, the capability of nursing homes, and medico-legal concerns has the potential to improve decision-making regarding hospital admissions.

INTRODUCTION

Older people living in nursing homes represent a population who are frail, in the last years of life, and whose health needs are often a challenge to those caring for them. The decision of whether or not to admit to hospital a patient who is frail and has a potentially life-threatening illness is complex and involves both the best interest of the patient and the effective use of resources.

A recent analysis of admissions of nursing home residents to a hospital in the UK showed that 34% did not survive the admission and that 52% had died within 6 weeks.¹ There has been a widespread concern that some hospital admissions from nursing homes are ‘inappropriate’ or ‘avoidable’^2–8 although there is little consensus about which admissions are avoidable.⁹

It is unclear how admission decisions are made and what the key influencing factors are. Improvements in nursing home staffing levels, hospital discharge procedures and clinical care may reduce hospital admissions from long-term care facilities in the US.¹⁰ A US postal survey sought physicians’ responses to hypothetical cases of patients who were seriously ill and in difficult clinical situations; despite the presence of advance directives, the doctors appeared to consider other factors such as prognosis, perceived quality of life, and the wishes of family or friends as more determinative than the advance directive.¹¹ An interview study of 27 Scottish GPs found that admission to a community hospital, rather than a district general hospital, depended on factors including the GP’s personal ‘comfort’ with the decision, particularly when medical complexity increased.¹²

A structured, palliative approach to nursing home care has been advocated^13–15 and this may reduce both admissions and deaths of residents in hospital.^16–18 Such a policy, which both reduces costs and ‘improves’ care, is compelling.¹⁰ In the UK — where a recent government-sponsored report suggested that 40% of people dying in hospital had no medical need to be there¹⁹ — this type of policy is implicit in recent NHS quality improvement initiatives regarding end-of-life care and the reduction of unscheduled admissions.²⁰ Advance care planning, which is an important element of this approach,²¹ is underpinned in the UK by the 2007 Mental Capacity Act. Although there is some evidence that structured approaches to care, advance care planning, and talking about dying have the potential to reduce unplanned hospital admissions, little is understood about the interplay between these factors and GP decision.

In this study the views of GPs in the UK were explored with regard to the factors that influence their decisions to admit to hospital nursing home residents who are frail.

How this fits in

Decisions regarding the hospitalisation of nursing home residents can present a difficult dilemma for GPs. As ‘gatekeepers’ to NHS services, GPs are expected to make best use of resources and may be criticised for ‘inappropriate’ admissions. This study identifies factors influencing GPs’ decision making and their suggestions for improving practice.

METHOD

A semi-structured interview schedule was developed and piloted (Box 1). Key areas included participants’ perceptions of:

• factors that influenced their decision making with regard to admitting hospital nursing home residents who are frail;

• factors affecting the level of confidence/comfort in making those decisions;

• influence of health policies or research on those decisions; and

• suggestions on strategies for improving practice.

Box 1. Semi-structured interview schedule

• What recent experiences can you recall of considering admission to hospital nursing home residents who were frail?

• What are the most important influences on you in that situation?

• What do you see as the potential benefits and disadvantages of hospital care for this type of patient?

• In your experience, how do these decisions to admit (or not) frail older patients impact on their quality of life?

• What are the factors that affect your personal comfort as a GP and as a person when making these decisions?

• How is your clinical practice, in terms of deciding whether or not to admit frail nursing home residents influenced by your awareness of research and policy (for example, 2007 Mental Capacity Act).

• What might be done to improve practice and to help patients and carers or GPs in the sorts of situations we have been discussing?

Purposive sampling was conducted between August 2009 and November 2010 by inviting 41 GPs from within the study area (two counties in the south of England) to participate and ensuring that each successive participant provided a different combination of personal and practice characteristics; these were represented on a sampling grid to maximise heterogeneity. Participants gave written informed consent. With the exception of two pilot interviews, all interviews were conducted by the same researcher at the participant’s surgery or in another agreed location. The average interview duration was 46 minutes (range 25–75 minutes). Interviews were audiotaped and transcribed, with identifying details removed. The data were analysed using NVivo 8 and thematic analysis.22 Once the data had been collected, five transcripts were independently analysed by one researcher to draw up a preliminary list of codes. Two of the remaining researchers independently coded the same data to enhance dependability of results. As one of the researchers was a GP working in the same geographical area as the study participants, initial interpretation of the data was carried out by two of the other researchers.

Based on initial codes, the full data corpus was coded and a final thematic framework agreed by all authors. All authors checked the validity and consistency of coding in data sessions in accordance with the constant comparative method of analysis.²³ Negative case analyses (seeking out disconfirming examples) safeguarded against reaching a final thematic scheme prematurely and ensured that all views were represented.

RESULTS

In total, 41 GPs were invited by email or in person; 21 of these were interviewed, four were unable to participate within the study period, 12 did not reply, and four declined to participate. Table 1 summarises participants’ professional, personal, and practice characteristics. The role of the GP in making decisions regarding hospital admissions for residents of nursing homes was perceived by participants in terms of acting as the patient’s advocate to make a decision that would not only be medically appropriate, but also fit the overall needs of the patient at a crucial point in their life journey. Two main themes were identified and captured in GP reports; these comprised GPs’ views on:

• factors influencing decision making; and

• strategies for improving practice.

Table 1.

Participant and practice characteristics, n = 21

GP characteristics	n
Sex
Male	11
Female	10
Age, years
20–30	1
31–40	8
41–50	6
51–60	6
Experience in care of older people
Substantial	7
Moderate	13
Minimal	1
Out-of-hours worka
Substantial	7
Moderate	7
Minimal	3
Nil	4
Practice demographics
Ruralityb
Urban	5
Suburban	5
Rural	1
Mixed (rural and town)	10
Deprivationb
Substantial	4
Moderate	8
Minimal	9
GP role in practice
GP partner	17
Salaried	2
Locum	1
GP registrar	1

^aGPs’ current level of work out of hours.

^bGPs’ assessment of the rurality of their practice area. c GPs’ assessment of the level of deprivation within their practice area.

Each is discussed in turn; quotations are used to illustrate each theme and its sub-themes.

Factors influencing decision making

All factors — anticipated and emergent — that influenced GPs’ decisions on whether or not to admit a nursing home resident to hospital are summarised in Box 2. Anticipated factors included familiarity with the patient’s clinical picture, risks versus benefits of admission, and patient and relative/carer preferences.

Box 2. Factors influencing GPs’ decisions on admitting nursing home residents to hospital

Clinical factors

• clinical picture — nature of current health episode

• patient’s level of health prior to this health episode

• patient’s quality of life prior to this health episode

Potential benefits of admission

• medical investigations to clarify diagnosis

• specific interventions available in hospital

Potential risks of admission

• hospital-acquired infection

• trauma of relocation, leading to distress or increased confusion

• increased stress or communication problems for patient due to unfamiliarity of hospital nursing staff and/or pace of hospital ward

Patient preference

• patient’s current wishes (if patient is able to communicate)

• patient’s documented wishes

Relatives’ preferences

Medico-legal issues

Informational continuity and communication issues

• tendency to admit to hospital is increased when key information is missing (either clinical information, or that about patient’s or relatives’ preferences)

Nursing home capabilities and attitudes

• staff qualifications

• staffing levels

• care capability

• staff confidence

• views of staff on admission

Resources and GP workload

• awareness of cost implication of admissions

• work pressures — tendency to admit patients to hospital is perceived as increasing when workload is high

GP personal factors

• age and level of experience — perception that experience in general practice tends to make GPs more willing to avoid admissions that may be of doubtful benefit

• previous experiences — past experiences of either positive or negative outcomes of decision making

• personality of GP — GPs who are cautious or risk adverse are perceived as tending to admit to hospital more readily

Clinical picture

Assessing the patient’s current medical condition was perceived as the starting point when making admission decisions. Participants reported that medical conditions perceived to have a strong probability of successful treatment in hospital, and/or which would cause acute pain or distress if left untreated, tended to be an automatic reason for admission:

‘There may be occasions where just correcting a fluid balance or correcting electrolytes or sugar would, would be helpful or where treating an infection would make all the difference.’ (Participant 10)

Examples given included fractures and acute conditions that were surgically operable. Other medical conditions — examples given included chest infections and left ventricular failure — could present greater dilemma, unless a specific hospital-based intervention offered clear benefits.

Risks versus benefits of hospital admission. Overall, hospital admission tended to be regarded as a potentially distressing experience for patients and one that could involve adverse effects on overall wellbeing (see Box 1).

‘I think when people have been ill in bed, any movements, turning them is painful and uncomfortable, they may end up on a trolley in A&E[accident and emergency], it’s certainly more likely they’ll get pressure sores, from what I’ve seen. And sleep deprivation with noisy wards, so I think it would be hard to say that going to hospital would be a pleasant experience, though it might be a lifesaving one.’(Participant 14)

‘Apart from the mortality, which could increase following hospitalisation, their morbidity could increase as well and[the patient] could pick up also germs or problems, and the inconvenience for them and when they are in the nursing home in their own bed … that’s all disadvantages for them.’(Participant 11).

In consequence, GPs tended to perceive admission as appropriate only when there were clear medical benefits that outweighed the potential adverse effects of admission.

Patient and relative preferences

When the patient’s own wishes were expressed, or clearly documented, all participants regarded this as a key factor in decision making:

‘Knowing the patient’s wishes … particularly if they’ve got capacity to make decisions, but even if they haven’t got full capacity, somebody who perhaps is very confused but is adamant they don’t want to go to hospital; it’s very upsetting to have to make somebody go or try and encourage them to go, against their will. So patients’ wishes would be the first thing.’(Participant 16)

Emergent factors included medico-legal considerations, communication, nursing home capabilities, NHS resources and GP workload, and GPs’ personal factors.

Medico-legal issues

Medico-legal issues were widely perceived as a source of pressure in decision making. Admitting a patient to hospital was generally viewed as less likely to expose the GP to the risk of litigation; in contrast, keeping the patient in the nursing home tended to be perceived as preferable for the patient, unless admission could offer clear benefits. Tension between a desire to prioritise the patient’s best interests and concern over medico-legal vulnerability emerged as a major issue for some participants:

‘I think the medico-legal thing is always present, probably much more so now than it was 20 years ago.’ (Participant 14)

‘You’re going to, in the back of your mind, follow the least line of resistance and send them in, rather than not send them in, which may be clinically justified. Yet you’re then embroiled in answering letters of complaint. So why would you do that? Why would you expose yourself?’ (Participant 7)

Communication and informational continuity. Communication issues were perceived as playing a major role in influencing the decision-making process:

‘Avoiding hospitalisation should be the rule, but that can only be achieved if there is a good, well-balanced communication between the primary care team, which includes us, the nurses and the home, whether it is a nursing or residential or family or independent or care home. If we achieve that good communication and we deliver the service that is required, hospitalisation will be very minimal and will be only reserved for those who are in real need for it.’ (Participant 11)

A key aspect of communication was the documentation, transfer, and availability of clinical information, as well as previously recorded information on patient wishes, if the patient was currently too unwell to communicate.

Access to adequate information was an issue that recurred throughout the interviews and appeared to mediate other factors involved in decision making. In cases where crucial clinical or background information was missing or inadequate, uncertainty resulted in leaving admission to hospital as the ‘default’ option:

‘If I don’t have any of this information to be able to decide, I will admit the patient, even if I think that they’re not going to … there’s less chance of survival, then we have to try our best to treat them if we don’t know their plan.’ (Participant 12)

Capability of nursing home staff and attitudes

The capability of nursing home staff and positive attitudes towards palliative care were regarded as strong factors in supporting a decision to keep a patient in the nursing home. Although participants reported excellent skills in some nursing homes, overall, a need for more training of nursing home staff was highlighted:

‘In my experience the nursing homes are very, very different in the level of expertise of the staff and their feelings about admission to hospital. The better ones tend to prefer not to admit patients unless it’s absolutely necessary. Some of the less confident staff are, well, do seem to be asking for admission quite a lot for things that we probably feel like would look after themselves.’ (Participant 9)

Resources and GP workload

The cost of admissions to NHS resources emerged from the analysis as a minor factor; all interviewed GPs were aware of it, but none felt it had much influence on their decision making:

‘I suppose we’ve always been aware that admissions are expensive and[require] a lot of resources so we’ve always been a little bit wary about admitting people anyway, and I suppose the factors that needed to admit them probably wouldn’t be swayed too much by just thinking about cost and resources.’(Participant 14)

In contrast, participants noted that not admitting the patient generally necessitated additional GP workload, which could contribute to the dilemma surrounding decision making, especially during busy periods:

‘I think, actually in hospital, it’s someone else’s problem and so I think definitely it would be a saintly GP who would say that they’ve never been influenced by their own personal pressures in the day.’ (Participant 9)

Personal factors

Personal factors included GP age, personality, and experience. Younger or less-experienced GPs were perceived as tending to admit more readily than experienced GPs. This was linked to the view that keeping a patient in the home could induce concerns about clinical and medico-legal vulnerability. Some participants reflected on how their own confidence in decision making had changed over time:

‘If you are far less experienced, you’re probably more likely to admit to hospital than maybe if you have more experience; you would tend to only refer to hospital where you feel it would add quality of life for the patient.’ (Participant 18)

GPs’ views on strategies for improving practice

Participants were asked how decision making could be improved and how ‘inappropriate’ admissions avoided. Four key areas were identified:

• improving communication;

• increased nursing home training;

• use of specialist nurses; and

• peer support for GPs.

Improving communication

Good communication was seen as crucial to reducing inappropriate admissions:

‘Communication and support, I think are the two most important things. So for us to communicate with nursing homes and relatives, for them to communicate with us, and for, actually, everybody to feel supported that they’re making a decision, that runs right from relatives right up to, you know, doctors and hospital consultants.’ (Participant 10)

A particular area in which improvement in communicating clinical information was seen as crucial was between GPs and out-of-hours service (OOH) doctors. In some geographical areas, this issue had been addressed through strategies to convey information directly from GPs to the OOH database:

‘Our out of hours is perfect, at [name of OOH service] here, it is very good. It works, because it’s run by us. We are the GPs running the out of hours more or less and we have a system where we write specific notes for specific patients and we fax it to [name of OOH service]. So whoever deals with them, as soon we go on the computer, we see these special notes and we deal according to the usual GP, patient’s GP advice, so, for example, he’s said it’s for palliative care, we know it’s for palliative care, avoid hospitalisation, we know to avoid hospitalisation and so on, and that has helped a lot.’ (Participant 11)

Some participants highlighted the importance of nursing homes making and maintaining detailed notes and care plans in their locality.

‘As I said, it does make a big difference the more information you have about the patient, the much easier it is for you to make a decision about the … future care. So having a document about the care plan in the notes will make a big difference.’ (Participant 12)

‘Medico-legally as well, because you can’t make decision not to treat the patient, but if everything was documented, well then, then you can, because if it’s already been agreed that the patient doesn’t need any intervention when they go really downhill, then it makes it easier for us. So GPs writing down special notes will help.’ (Participant 14)

Training and support for nursing home nurses

For some participants, the role of district nurses in visiting nursing homes was perceived as a means of reducing admissions. In one locality, GPs reported an initiative in which the primary care trust funded a specialist nurse to liaise between nursing homes and GPs, and to provide nursing home staff with training. This was perceived to be a valuable resource in reducing avoidable admissions.

Overall, increased training for nursing home staff was seen as an avenue towards reducing admissions:

‘Some nursing home staff are really well trained, they give confidence for us, ‘cause we know that they know when the patient is unwell and when to call us back, but sometimes, you think “Oh, they’re not very sure about what’s going on …” So you feel uncomfortable as well. So staff training and increasing the facilities of treatment in the nursing home will make us more comfortable and put us in a better situation to make decisions about treatment in the nursing home, without sending the patient to hospital.’ (Participant 12)

Peer support and education for GPs

Learning from the skills and experiences of other medical colleagues was highly regarded as a crucial aspect of developing good decision making. Examples of this included: formal postgraduate education sessions, led by consultants, on care of older people; GP peer supervision groups; and seeking advice from experienced colleagues:

‘If I go and see a patient who’s not well known to me, I usually speak to one of my colleagues … particularly, my two older partners here, who are in their late 50s, have been in the situation where they’ve been looking after either these patients or have many more years of looking after elderly people and I’ll very often ask them for advice.’ (Participant 16)

Societal trends towards an ageing population and the need for GPs to provide good end-of-life care within often-complex situations was reiterated throughout the interviews:

‘It’s a minefield for doctors because it’s not really … it’s not a big part of our training and I think it’s something we need … we need more knowledge in because this is increasingly more prevalent, especially with the ageing population and people are much more aware of their rights and people do tend to think of it more — about end-of-life care, than they used to. So I think, as a group, we need to get better at this.’ (Participant 17)

DISCUSSION

Summary

These findings suggest that the issues surrounding nursing home residents’ admissions to hospital are complex and cause GPs concern. The main factors considered in such decisions were not only the clinical assessment of the patient, the perceived benefits and risks of admission, and patients’ and relatives’ preferences, but also perceptions of the capability of the nursing home, medico-legal concerns, and GP training and workload. Lack of adequate clinical information and poor documentation of patients’ and relatives’ preferences were seen as hindrances to good decision making.

Strengths and limitations

Although decisions to admit nursing home residents to hospital often involve a broad care team, the study chose to focus on GPs’ views because they carry the ultimate responsibility for most admissions.

A routine limitation of interview studies is that they capture the participant’s report on events, rather than the actual observation of events; the study also presented GPs’ views on decision making rather than provided a window through which decision-making moments could be observed.

The study recruited GPs from two counties in the south of England and it is possible that GPs in other areas might hold different views. However, the data suggest the dilemmas experienced in such decision making accord with those described both elsewhere in the UK and in other countries.

Comparison with existing literature

The study found a general acceptance of the concept of ‘avoidable’ or ‘unnecessary’ admissions — a finding that accords with that of Wyman and Hazzard (2010),⁸ although all participants said they would admit a patient if they felt it was in the patient’s best interest. Echoing the findings of Patel et al in the literature,²⁴ there appeared to be widespread recognition of the complexity of factors involved in making a ‘good’ decision.

The finding that medico-legal issues are a significant concern regarding admission decisions seems to have received little attention in the literature to date and implies that, in making such difficult decisions, GPs feel somewhat vulnerable to the risk of criticism or even legal sanction. The impact on GP workload of avoiding hospital admissions has also been little reported; however, one study exploring GPs’ views on decisions to admit patients from their home to community hospitals rather than to a district general hospital, found that job satisfaction appeared to offset the additional workload and responsibility.¹² Concerns about inadequacies in GP training in end-of-life care have also been acknowledged in a recent UK evaluation by The King’s Fund²⁵ and other policy reports.^{19, 26}

Problems with communication, including the availability of adequate clinical documentation and nurses’ knowledge of the patient’s clinical history, were commonly mentioned. Lack of understanding of patients’ prior state of health has been found to contribute to inappropriate hospital admissions from the community in a UK focus group study;²⁷ adequate healthcare records and advance planning documentation in anticipation of care management decisions are a common feature of palliative care initiatives.^16–18,28

Participants viewed the capability of nursing home care staff as a crucial factor in decision making and many suggested further training and support for nursing home staff. These concerns are highlighted in the international literature, but solutions are elusive.^9,10

Participants placed considerable weight on the views of patients and relatives, but there was relatively little mention of detailed advance care planning despite its advocacy in the UK,^17,21 the US,¹⁸ and Australia.¹⁶ The concept of a palliative care approach in nursing homes, as previously highlighted in other published research,^13–15 was implicit in many of the responses; a few responders described experiences of admission avoidance initiatives suggesting that, as outlined by Partridge,²⁹ a more proactive approach to avoid residents being sent to hospital to die may be gaining ground.

Implications for practice and research

These findings highlight a number of problems, such as inadequate documentation, concerns about the capability of some nursing home staff, and a sense of vulnerability in making difficult decisions; they also support the advocacy of a systematic palliative approach to nursing home care.²⁸ Adopting such strategies, including better training and more general use of advance planning and associated documentation, could both improve care and avoid possibly inappropriate admissions. In the UK, these ideas are now central to the NHS strategy for end-of-life care³⁰ and the related Quality Improvement Productivity and Prevention workstream.³¹ Consideration might also be given to providing more professional support for nursing home staff and specific training in how to deal with residents who are acutely ill.

In addressing GPs’ medico-legal concerns, greater transparency may be needed in terms of justifying and supporting decisions about whether or not to admit to hospital nursing home residents. Given the perceived importance of experience, also noted by The King’s Fund report,²⁵ local decision support might be helpful — perhaps as telephone advice from a more experienced GP or specialist.

The research base on which many of the current initiatives are based is limited. Understanding the impact of the many different factors involved requires robust ongoing research. There is also a need to identify the most effective advance planning tools and training programmes, and to address the paucity of robust outcome data, including quality-of-life measures, comparing the risks and benefits associated with hospital admission or continuing nursing home care. It is also unclear what resources, particularly in terms of professional support, will best assist nursing homes to fulfil their enhanced role.

In addition to the clinical assessment of the patient, perceived benefits and risks of admission, and patients’ and relatives’ preferences, important factors in determining admissions were cited as poor documentation, perceptions of the capability of the nursing home, medico-legal concerns, and GP workload. Developing a systematic palliative care approach may address many of these concerns.

Funding

Wessex Primary Care Network.

Ethical approval

The study received ethics approval from Berkshire Research Ethics Committee (reference 08/H0505/48) and research and development approval from the relevant NHS trusts.

Provenance

Freely submitted; externally peer reviewed.

Competing interests

The authors have declared no competing interests.

Discuss this article

Contribute and read comments about this article on the Discussion Forum: http://www.rcgp.org.uk/bjgp-discuss