Background

The definition of a fall as “an event whereby an individual comes to rest on the ground or another lower level with or without loss of consciousness” [1] was adopted by the National Institute for Health and Care Excellence (NICE) in its guideline on falls [2].

Approximately, one third of adults aged 65 and over fall each year [3]. A consequence of falls can be a fractured neck of femur, which has a monthly mortality rate approximately 10%, and a 1 year mortality rate of around 33% [4]. Increasing awareness of fall incidence and consequences, for both individuals and the UK NHS, is arguably a public health priority [5].

Falls account for over 600,000 emergency department (ED) attendances and lead to over 200,000 admissions annually in the UK NHS [6, 7]. However, some healthcare professionals are not aware of the role they can play in managing and preventing falls [7]. In particular, the identification of older adults at risk of falls (and repeat-falls) is required to prevent occurrence [7], and the ED may be able to facilitate this [8].

The NICE 2004 Falls guideline [2] was developed in order to reduce the impact of falls upon older adults and to limit NHS costs. The guideline was updated in 2013, although the recommendations on assessment after a fall not materially changed, and therefore, these recommendations have had ample time for them to be adopted into routine practice.

Falls are caused by various factors, all of which can be investigated through comprehensive patient assessments, which may reduce recurrent falls [9]. The guideline recommends a multifactorial falls risk assessment in older adults who have fallen (Table 1). The NICE Falls guideline applies to ED practice [2], and as EDs are usually the first, and often only, service consulted by older adults after a fall, they provide a setting in which assessments can be conducted or referrals made to a more appropriate setting [8, 10].

Table 1 Falls guideline recommendations [2] on multifactorial risk assessments in older adults
Full size table

Our focus was on multifactorial assessments, and the degree to which these are delivered according to the guideline (i.e. adhered to).

Our theoretical framework was developed from studies of determinants of adherence to the NICE Falls guideline in an ED. Our approach to studying adherence drew on the work of Cabana et al. [9] and Flottorp et al. [11]. It applies the idea of determinants of adherence that forms part of the model of tailored implementation [12]; the premise being that interventions tailored to the prevailing determinants are more likely to be effective than interventions chosen without tailoring. We have also drawn on research evaluating barriers to implementation, and the effectiveness of multi-faceted interventions, in implementing fall prevention guidelines [7, 13,14,15].

The study aimed to investigate (1) how falls are managed in EDs, (2) the barriers and enablers (collectively, referred to as the determinants of practice) influencing adherence to the guideline in EDs, and (3) ways barriers could be addressed.

Methods

Two sites were chosen to provide some variety of setting, individuals, and contexts. Recruitment from both hospitals allowed comparison of falls management in a busy city hospital ED (A) with that in a less busy town hospital ED (B). Further information is provided in Table 2, although limited in order to retain anonymity.

Table 2 Hospital sites - 2013
Full size table

Observation and semi-structured interviews were conducted at both sites, allowing adherence practices to be viewed from both the perspective of an outside observer, as well as the opinions of those working within the ED. Consent to participation in the study was sought from both ED professionals and patients presenting with a fall, participants being recruited through convenience sampling. Prior to seeking consent, potential participants were given information about the study and the characteristics of the researcher (e.g. non-clinical).

Care of patients aged 65 and over with a fall was observed from presentation at the ED, until discharge or admission to another department. The observer (HM) observed 30 professionals (20 doctors, 10 nursing staff) deliver 27 episodes of care. Of the 27 patients, 19 were male and 8 female, ranging in age from 67 to 98. Observation took place in major and minor injury departments, through watching, listening to, and recording interactions, but not participating in these [16]; Table 1 was referred to, to monitor adherence.

Contemporaneous notes were made for the duration of each patient’s ED stay; recording interactions between patients and professionals, questions asked, answers given, and the tests being conducted. Any potential barriers and enablers to guideline adherence were noted. After observing care, (HM) reviewed the patient’s clinical notes, using both sources to assess Falls guideline adherence. Observations were conducted until data saturation was reached-when the new information produced little or no change in the themes detected at the familiarisation stage of framework analysis.

Semi-structured interviews were conducted after the observation phase of data collection was completed. A range of professionals were recruited, including both doctors (20) and nursing staff (10). Sampling was opportunistic and professionals who had been observed delivering care could also be interviewed (3) were both observed and interviewed. Professionals were interviewed once, and interviews were conducted until data saturation was reached, that is, when there was repetition of interview responses.

Interviews (conducted by HM) were guided by an interview schedule (Table 3) developed from the Falls guideline recommendations and studies of adherence and falls management [1, 17,18,19], but flexibility was allowed based upon interviewees’ responses. The interview schedule was piloted with two research clinicians. The interviews were audio-recorded and transcribed anonymously. On average, interviews lasted 20 min.

Table 3 Interview schedule
Full size table

The interviews included five consultants, four registrars, five junior doctors, four Advanced Nurse Practitioners, four senior nurses, two staff nurses and six healthcare assistants. They sought to develop understanding of the patterns of care that had been observed and to gain views on determinants of guideline adherence.

Observation and interview data were combined to enable development of a single set of determinants, including those identified by both approaches and only one. The process included listing events that occurred during observations, characteristics of the environments, and people’s views, experiences and behaviours (HM) and (RB) collectively developed the themes from the data. Framework analysis was used as this suited our aim of producing practice-orientated findings to solve practical problems and generate policy [20,21,22].

Data were drawn together to define concepts, the nature of the phenomena under investigation, and categorise behaviours, motivations, and attitudes. We drew on emergent issues, including those raised by participants and what participants reported as, or were observed to be, determinants of adherence. The analysis consisted of coding transcripts and notes and then organising these into themes. Framework analysis’ five stages were adopted:

(i) Familarising ourselves with the data; (ii) identifying a thematic framework through drawing on a priori themes; topic aims, existing models of adherence [9, 11], and the interview schedule; (iii) indexing (linking data with themes); (iv) charting (providing evidence for each theme in a table); and (v) mapping and interpretation (drawing themes together, marrying up the findings from both phases of data collection, and mapping into super-ordinate and sub-ordinate themes-determinants of practice). Associations were analysed, looking for explanations, and developing ideas and strategies. Analysis of observation notes and transcribed data was supported by QSR NVivo7 software.

To help identify determinants that would be most appropriate to target in interventions to improve adherence, each determinant of practice was assessed on two criteria: (1) its apparent size of effect on adherence and (2) whether it was likely to be amenable to change. This assessment relied on our own judgement, but drew on previous research on the development of tailored implementation models [12, 23].

Results

Levels of adherence

Overall adherence was 62% in hospital A and 64% in hospital B.Footnote 1 For none of the 27 observed care episodes were all 11 multifactorial assessment guideline recommendations completed. In both EDs, completion ranged from 4 to 9 of the 11 recommendations. Adherence to each recommendation varied and is summarised in Table 4.

Table 4 The frequency in which guideline recommendations were adhered to
Full size table

At both sites, the recommendation most frequently adhered to was the completion of a cardiovascular examination (27/27 occasions). Assessments of osteoporosis risk and urinary incontinence were both completed less than a quarter of the time. A medication review was completed least frequently (1/27 occasions).

Determinants of practice

Several determinants of adherence were identified. At initial indexing, there were 17 themes, which were mapped into 7 super-ordinate and 6 sub-ordinate themes (Fig. 1).

Fig. 1
figure 1

A map of the themes (barriers and enablers to adherence)

Full size image

Example quotes from observation notes and, in italics, interviews are provided, in which A/B refers to the hospital, and numbers refer to the observation/interviewee number.

Communication

Staff-patient

Professionals initially tried to talk directly to the patient, but in some circumstances, this was not possible, preventing the identification of some possible causes and consequences of falls, and ways in which to prevent re-occurrence.

Staff-staff

When communication within the ED and/or between departments was poor, breakdowns in care were prone to occur. Team-working could be influenced by the staff rota and the effectiveness of communication between staff.

‘Unbeknown to the professional treating the patient, they had been taken straight to an observation unit. Staff-staff member communication was a concern’.

(27- A)

“Blood tests might be omitted. You just have to hope that doctors in other area(s) will pick up on these.”

(2, Doctor, B)

Patient complexity

Often older adults who presented with a fall had other health problems, which sometimes took priority, making it difficult to determine causes of falls and prevent re-occurrence. There were various factors to be taken into consideration during assessment and developing treatment plans.

A patient with Alzheimer’s was found on the floor.

‘Q - (To relative) if you had the option would you rather she was kept in hospital or went home?

A - Relative stated that they needed to weigh-up the decision. With regards to Alzheimer’s they thought the patient would be better in their own environment, but with regards to their unsteadiness they would be better in hospital.’

(19 - A)

Education and training

Lack of education about the Falls guideline was described as influencing adherence. Some interviewees argued staff could not be familiar with all guidelines, and the department selected which to prioritise. This was due to the large number of guidelines, the large number of staff to train, and the busyness of the department.

“I don’t necessarily think some of the staff within the department understand why they are doing it. Hopefully through education you’ll get people to do it meaningfully.”

(3, Nurse, A)

Patient presented with abdominal pains after slipping.

Healthcare professional noted to (HM) that they did not think it was a fall and that it was a surgical problem; their understanding contrasted with the NICE definition.

(20 - A)

Variation in care pathways

Variation between professionals’ perspectives on care could affect adherence. Care pathways were viewed as a key factor.

“There are a lot of operating procedures, sometimes it varies with who’s in charge.”

(4, Doctor, A)

Access to resources

While there were aids to prompt falls assessments being conducted, problems in accessing the guideline were both observed and described.

“There are lots of guidance but not easily available, you have to do lots of searching.”

(6, Doctor, B)

Influence of seniors

Peer influence may be important. If senior members of staff were to enforce and follow guidelines, juniors may be more likely to follow suit.

“You need somebody with clinical credibility to champion work.”

(3, Doctor, A)

“My guidelines would be whatever [seniors] asked me to do.”

(7, Healthcare Assistant, B)

ED care processes

Level of activity

The busyness of the EDs and the kind of patient care provided had varied causes and implications.

“So it’s down to the ‘busyness’ of it, it’s not built for purpose, you don’t have the facilities; there’s not necessarily appropriate handovers.”

(5, Doctor, A)

‘The department was busy, consequently the professional ran out of time when conducting falls assessments and focused on the patient’s injury.’

(4 - B)

Targets

Professionals had competing priorities, for example, in-depth assessments, or meeting target treatment times. The 4-h target for patients to been seen and discharged from an ED [22] was a particular concern and professionals argued it could have a negative impact by not allowing time for comprehensive assessments.

“I think the four-hour target is often not conducive in providing a full assessment to somebody old.”

(3, Doctor, A)

Variation in ED staff and attitudes towards guidelines

Variation in staff on duty was observed to influence care, and may be associated with communication gaps. Professionals may have to adapt to working with various colleagues with different opinions or ways of working.

“If you work in a teaching hospital then guidelines are quite prominent.”

(8, Doctor, B)

The rotation of junior doctors was reported as a factor influencing care.

“Perhaps the junior doctors don’t see the value of it [guideline] because they’re not here to deal with when people get brought back in.”

(9, Nurse, A)

Cross-boundary care

Cross-boundary care refers to the care a patient received for their fall before or after ED discharge, and it was thought to influence care. The ED may more easily care for a patient with a fall if they have received adequate support before ED presentation and the patient is willing to accept advice. Likewise, they may more easily care for a patient if resources are in place to meet their requirements post-discharge or ward transfer.

Patient with multiple health problems, presented after falling in the bathroom.

‘The relative stated that they had had an assessment before which had not helped as they were given a Zimmer frame, which is not practical in the house.

The patient had also had a bad experience within the hospital previously, they were reluctant to accept recommended care.’

(6 - B)

“I think identifying somebody that is a risk of falls or is a ‘frequent faller’ is a key part, and you can utilise a community nurse.”

(10, Healthcare Assistant, B)

Assessment of determinants

The determinants of practice were reviewed to identify those most likely to be suitable for addressing in a potential intervention, by assessing firstly, whether the determinant was likely to have a large impact on care (if the impact was thought to be small, little would be gained by addressing it), and secondly, whether it was possible, through intervention, to alter the effect of the determinant.

Three determinants were judged to be most important in influencing adherence and potentially most amenable to change.

  1. a)

    Support from seniors. Agreement amongst senior staff could lay the foundation for care throughout the ED.

  2. b)

    Education. Professionals need to be familiar with guidelines in order to adhere to them.

  3. c)

    Cross-boundary care. Adherence may be improved through considering care pathways and services to be used in conjunction with ED management.

An overview of all of the determinants and the reasons why they were considered more or less amenable to change is provided in Table 5.

Table 5 Determinants of practice and their amenability for change
Full size table

Discussion

Various determinants were found to influence adherence practices, relating to knowledge, attitudes and behaviours; this supports previous research findings [9, 11]. Building on the influence of senior staff is a feasible approach to improve guideline adherence. If the importance of Falls guideline adherence is agreed by seniors, and a process to achieve consensus amongst them is effectively managed, juniors may receive more support and adherence may improve.

Education is often considered an appropriate intervention in implementing guidelines. It is feasible to deliver and may improve awareness of the Falls guideline [18, 19, 24]. However, its effects on adherence may be limited by other factors; in the absence of senior staff support, for example, education may have only limited effect on care by junior staff. Education may be a necessary preliminary, but it should be supplemented by other interventions [25].

Cross-boundary care has the potential to be addressed through professionals and commissioners considering care pathways and services to be used in conjunction with ED treatment of falls patients, and through ensuring collaborative working. EDs can be busy, and consistent adherence to all of the Falls guideline within the ED may not be practical. Some less time-sensitive guideline elements, such assessments of home hazards, may be more appropriately completed later on. The ED could fulfill the role of gatekeeper, in recognising the need for falls assessments and referring patients to services with more time for managing people with falls in accordance with the guideline. Service re-design with a collaborative approach to care may lead to streamlined referrals and more consistent falls assessments [17]. This approach reflects the issue of cross-boundary care, but rather than using a new care pathway to improve adherence in the ED, it could be used to appropriately direct aspects of care away from the department.

Limitations

The study was undertaken in only two EDs, and therefore, the findings cannot be assumed to apply more widely within or outside the UK. However, most countries are likely to have some forms of screenings or guidelines in their ED and the results of our study might help inform the response to, or changes in, guidelines proposed or in use, for example, in Australia [13], Austria [14] and Singapore [15].

Convenience sampling in both observations and interviews allowed recruitment of a variety of individuals within the ED, but it was necessary to exclude a small number of patients from care homes who attended EDs unaccompanied. Excluding some individuals who could not give consent reduced variation in the patient sample. In addition, when it was busy, it was difficult to approach professionals to gain the necessary written consent to observe them attending a particular patient; this inevitably led to some data collection opportunities being missed.

Observation research has its strengths in avoiding reliance on self-report and is appropriate for use in various contexts [26]. However, observation does not collect information on the internal processes of decision-making. For example, it may have appeared that a professional did not complete an assessment of a patient’s balance when they had done so, but not recorded it, as they deemed them not to be at risk. Some gaps in decision-making were filled by reviewing clinical notes for records of tests that had been conducted and professionals’ interpretation of results.

Observation might have prompted some professionals to change their behaviour (the Hawthorne effect); however, in view of our findings of poor adherence to the guidelines, we think it unlikely that our presence encouraged healthcare professionals to adhere to the guidelines. In addition, the interview phase of research helped to shed light on what professionals knew about guidelines, and this could be compared to observed behaviour.

Implications for practice

Through identifying determinants of practice, this study provides a basis for developing ways to address barriers to Falls guideline adherence. In contrast with other studies, we used both observation and interview research, providing two sources of evidence on the key determinants to be addressed by interventions.

Conclusions

It appears that there is no ‘quick’ fix solution to overcoming barriers to Falls guideline adherence within the ED, but a planned approach to improving adherence has potential if it involves agreement and leadership amongst seniors, supported by staff education. Cross-boundary collaborative service designs also offer a potential approach.