The National Institute for Health Care Excellence (NICE, 2013) defines a fall as an unintentional or unexpected loss of balance resulting in coming to rest on the floor, on the ground or on an object below knee level. Falls are frequent in older adults, are the primary cause of injury in those aged 65 years and over, and are associated with high mortality, morbidity and immobility (Rubenstein, 2006; Darnell et al, 2012). Evidence suggests that 35% of those aged >65 years of age fall annually (Tinetti et al, 1988).
Ambulance services across the UK respond to about 300 000–400 000 emergency (999) calls for falls in older people each year (Close et al, 2002), many of which result in conveyance to the emergency department (ED). It has been estimated that the cost of falls accounts for 3% of total NHS expenditure (Scuffham and Chaplin, 2003).
Ambulance services are under continued, sustained, demand-led pressure and have seen a 6% year-on-year increase in calls between 2009–2010 and 2016–2017 (NHS England and NHS Improvement, 2019).
Historically, it has been common practice for ambulance services to convey older adults who fall to the ED. The Ambulance Improvement Programme identified falls as one of six pillars of improvement (NHS England and NHS Improvement, 2019), suggesting older adults who fall could be safely managed and discharged on scene.
In response, most ambulance services in the UK use strategies to prevent falls in older adults, promote healthy ageing and reduce conveyance to the ED. In conjunction with Gateshead Health NHS Foundation Trust, North East Ambulance Service (NEAS) NHS Foundation Trust launched a pilot falls rapid response service (FRRS) covering Newcastle upon Tyne and Gateshead in September 2018. Although evidence supports associations between several risk factors and falls (Rubenstein, 2006), predictors of conveyance to the ED are less clear.
The population is undoubtedly ageing and the number of older adults who contact the ambulance service after a fall will continue to rise.
The authors hypothesised that: some of these patients will be unsuitable for on-scene discharge and will require conveyance to the ED; and predictive characteristics of conveyance could be identified.
Consequently, despatching the FRRS to all older adults who fall is not practical and may not be desirable, and predicting patients likely to need conveyance to hospital could result in more effective use of the FRRS.
The purpose of this study was to identify the predictive characteristics of conveyance to the ED in older adults using the FRRS.
Ethics
North West–Greater Manchester North West Research Ethics Committee (18/NW/0763) and the Health Research Authority provided ethics approval for this study.
Setting
NEAS is one of 10 ambulance services in England, covering 3230 square miles and a population of 2.71 million. NEAS serves metropolitan, urban and rural areas. The FRRS responded to calls within the Newcastle and Gateshead geographical area serving a population of 200 153. The FRRS comprised a paramedic and occupational therapist (OT) who could access alternative and rapid care pathways to promote falls prevention and to facilitate safe non-conveyance of older adults. The FRRS is available from 7:00–17:00, 7 days per week.
Methods
This cross-sectional study ran from December 2018 until September 2020 and the authors set out to enrol consecutive patients who were attended by the FRRS in this period. Eligible for inclusion were all patients aged ≥60 years who had mental capacity and who had had a fall within the relevant geographical area.
The FRRS paramedic or OT screened patients for eligibility and approached appropriate patients regarding the study following clinical assessment if they had mental capacity. Patients were provided with an information sheet, had an opportunity to ask questions and provided written informed consent before enrolment.
The FRRS paramedic or OT prospectively collected baseline sociodemographic characteristics and health status data, which were attached to the ambulance electronic patient care record. In addition, patients were asked to self-report a fear of falling (FoF) score using an 11-point Likert scale (0=no fear of falling; 10=worst fear of falling). A frailty score for the day of contact with the FRRS was calculated using the Clinical Frailty Scale (CFS) (Rockwood et al, 2005). A long lie, where a patient remains on the floor or ground for >1 hour following a fall (Lord et al, 2007), was defined as the time from the emergency call to the arrival of the FRRS >1 hour. Polypharmacy was defined as the use of ≥4 medications (Akyol, 2007).
Statistics
Descriptive statistics are presented using counts (n) and frequencies (%). Age, CFS and self-reported FoF Likert scales were analysed on a continuous scale from high to low. Comparisons of age and sex between different groups were analysed using a t-test. A generalised linear model and odds ratio (OR) were used to analyse predictive characteristics. A P value of <0.05 was accepted as statistically significant. Data are presented correct to 2 decimal places. Data were analysed using R 3.6.2 (R Project).
Results are reported in accordance with the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies (von Elm et al, 2007).
Results
During the study period, the FRRS attended 1202 patients. After ineligible patients (n=86) and repeat calls (n=25) to patients who had already been enrolled or excluded were removed, 1091 patients were eligible for enrolment. Sixty-one per cent (664/1091) of patients were not enrolled for various reasons: 29.6% (n=324) lacked mental capacity; 16% (n=171) were too unwell to approach regarding the study; 7% (n=77) had other reasons why they could not be approached, such as complex on-scene situations; 4.3% (n=48) were missed recruitment opportunities; 3% (n=34) refused to participate; 0.5% (n=6) were intoxicated; and 0.3% (n=4) were precluded from participating because of language barriers. One participant was enrolled but later withdrew, resulting in 426 patients eligible for analysis. The participation recruitment flowchart is shown in Figure 1. Those not enrolled in the study or not conveyed to the ED did not significantly differ from those who were, in terms of age and sex (Table 1).
| Enrolled (n=426) | Not enrolled (n=665) | P | |
|---|---|---|---|
| Mean age (years) | 82.61 SD 8.4 (range 60–99) | 82.7 SD 8.91 (range 60–103) | 0.85 |
| Female sex | 259 (60.7%) | 392 (58.9%) | 0.54 |
| Not conveyed (n=325) | Conveyed (n=101) | ||
| Mean age (years) | 82.8 SD 8.36 (range 60–99) | 81.9 SD 8.52 (range 60–96) | 0.34 |
| Female sex | 194 (59.6) | 65 (64.3%) | 0.4 |
| Home status (n=426) | |||
| Community | 380 (89.2%) | ||
| Sheltered accommodation | 7 (1.6%) | ||
| Residential care | 39 (9.1%) |
SD: standard deviation; P: probability value of age and sex of patients enrolled versus not enrolled in the study, and not conveyed versus conveyed to the emergency department
Of the 426 patients enrolled in the study, the mean age was 82.61 years (SD 8.4; range 60–99) and 60.7% (259/426) were women. In brief, study participants were more likely to live alone (62.4%; 266/426) and had a history of fall events before contact with the FRRS (65.4%; 279/426). Patients had comorbidities and were impaired by known geriatric syndromes, the most common being mobility dysfunction (82.8%; 353/426), balance disorders (81.9%; 349/426), polypharmacy (87%; 371/426), visual impairment (49.2%; 210/426), osteoarthritis (42.9%; 183/426) and incontinence (35.6%; 152/426). Most patients lived in the community (89.2%; 380/426).
The predictive characteristics of non-conveyance and conveyance to the ED were derived from the generalised linear model and are shown in Table 2. Predictive characteristics of non-conveyance reaching significance were a witnessed fall (OR 0.46; 95% CI (0.23–0.85); P=0.01), cognitive impairment (OR 0.3; 95% CI (0.10–0.71); P=0.01), dementia (OR 0.13; 95% CI (0.01–0.64); P=0.04), mobility dysfunction (OR 0.45; 95% CI (0.26–0.78); P≤0.01), balance disorders (OR 0.46; 95% CI (0.27–0.79); P≤0.01) and walking aid use (OR 0.38; 95% CI (0.23–0.64); P≤0.01).
| Characteristic | n (%) | OR | 95% CI | P |
|---|---|---|---|---|
| Lives alone | 266 (62.4) | 0.97 | 0.61–1.54 | 0.88 |
| Witnessed fall | 86 (20.1) | 0.46 | 0.23–0.85 | 0.01 |
| History of suffering falls | 279 (65.4) | 0.67 | 0.42–1.06 | 0.08 |
| Known to falls services | 77 (18.0) | 1.48 | 0.84–2.54 | 0.16 |
| Cognitive impairment | 53 (12.4) | 0.3 | 0.10–0.71 | 0.01 |
| Dementia | 24 (5.6) | 0.13 | 0.01–0.64 | 0.04 |
| Parkinson's disease | 20 (4.6) | 0.45 | 0.05–1.22 | 0.15 |
| Stroke | 82 (19.2) | 0.88 | 0.48–1.55 | 0.67 |
| Diabetes | 92 (21.5) | 1.01 | 0.58–1.72 | 0.95 |
| Mobility dysfunction | 353 (82.8) | 0.45 | 0.26–0.78 | <0.01 |
| Balance disorders | 349 (81.9) | 0.46 | 0.27–0.79 | <0.01 |
| Osteoarthritis | 183 (42.9) | 0.93 | 0.59–1.46 | 0.75 |
| Osteoporosis | 113 (26.5) | 0.95 | 0.56–1.56 | 0.83 |
| Use of mobility devices | 336 (78.8) | 0.38 | 0.23–0.64 | <0.01 |
| Blackouts or fainting | 34 (7.9) | 1.38 | 0.61–2.91 | 0.41 |
| Visual impairment | 210 (49.2) | 0.96 | 0.61–1.50 | 0.85 |
| Incontinence | 152 (35.6) | 0.66 | 0.40–1.07 | 0.09 |
| Polypharmacy | 371 (87.0) | 0.9 | 0.48–1.78 | 0.74 |
| Injurious fall/pain | 188 (44.1) | 8.25 | 4.89–14.50 | <0.01 |
| Long lie | 130 (30.5) | 1.6 | 1.00–2.56 | 0.04 |
| Clinical frailty score (1–9) | 426 (100.0) | 1.15 | 0.72–1.88 | 0.55 |
| 11-point fear of falling score (0–10) | 391* (92.0) | 1.19 | 0.73–1.91 | 0.48 |
OR: odds ratio; P: probability value of conveyance to hospital for each predictive characteristic; CI: confidence interval
Predictive characteristics of conveyance reaching significance were an injurious fall or pain as a result of the current fall (OR 8.25; 95% CI (4.89–14.50); P≤0.01) and a long lie (OR 1.6; 95% CI (1.00–2.56); P=0.04).
Discussion
This cross-sectional study investigated the predictive characteristics of conveyance and non-conveyance to the ED of older adults using the FRRS. Patients enrolled into the study were more likely to be older women who had comorbidities and were clinically complex with a variety of geriatric syndromes.
The authors identified distinct differences between those who were not conveyed and those who were. Conveyance was predicted by the consequence of a fall, such as an injury or pain and a long lie. Predictors of non-conveyance were characteristics traditionally associated with falls; these were having a witnessed fall, cognitive impairment, dementia, mobility dysfunction or balance disorders and using mobility devices.
Mobility dysfunction and balance disorders have consistently been identified as strong risk factors for falls (Tinetti et al, 1988; Rubenstein et al, 1994). Impaired mobility results in a requirement for mobility solutions (Grimmer et al, 2019) such as walking aids, and includes frames, crutches, canes and wheelchairs; older adults who use walking aids have lost their functional independence and the use of a walking aid indicates a decline in the person's posture, gait and step length (Pirker and Katzenschlager, 2017). These syndromes are synonymous with ageing and, although the symptoms can often be managed, the syndromes themselves are unmodifiable. This study indicates that, when the FRRS attends older adults with these characteristics, the patient can frequently be managed in the community with the additional interventions available to the clinical team.
Almost half (44.1%; n=188) of falls in this study resulted in injury or pain, suggesting that negative consequences from them are commonplace. These findings reflect previous research suggesting the prevalence of pain or injury subsequent to a fall is in a range of 40–60% (Kenny et al, 2017). Minor injuries include contusions, grazes or scratches, but many falls result in more serious injuries such as head injuries and fractures (Ash et al, 1998; Kallin at el, 2002).
In this study, the most common reasons for conveyance to the ED (Table 3) were minor trauma, including limb injuries (53.4%; n=54), presenting as medically unwell (24.7%; n=25) and head injuries, including complications origination from anticoagulation therapy (9.9%; n=10).
| Reason | n |
|---|---|
| Acute pain | 1 (0.99%) |
| Head injury | 10 (9.9%) |
| Long lie | 2 (1.9%) |
| Loss of mobility | 1 (0.99%) |
| Major trauma | 1 (0.99%) |
| Medically unwell | 25 (24.7%) |
| Minor trauma | 54 (53.4%) |
| Not recorded | 6 (5.9%) |
| Patient safety | 1 (0.99%) |
| Total | 101 |
Older adults are more likely to have musculoskeletal conditions such as osteoporosis, which are known to cause pain. In addition, they are more likely to develop reduced bone density than younger adults, which increases their propensity to experience a fracture or bone injury.
Research has demonstrated that musculoskeletal pain leads to mobility limitations (Mottram et al, 2008) and functional decline (Dunlop et al, 2005), which in turn results in an increase in the risk of falls and concerns regarding the consequences of falling (Stubbs et al, 2016). Therefore, pain before or after a fall has a significant impact on older adults and is detrimental to quality of life.
Diagnosing and managing the cause of pain or injuries in older adults can be complicated and often impossible outside an inpatient setting, so it is not surprising that pain and/or injuries result in conveyance to the ED.
Of all fall events in this study, 30.5% (n=188) resulted in the person lying on the floor for >1 hour. Long lie events are evidence of a delayed initial recovery and may occur for several reasons. Injuries or pain may preclude the patient being able get themselves up from the floor and can be both a cause and result of lying on the floor for a long time (Fleming and Brayne, 2008). Long lies can result if a fall occurs when there is no one around to help. In this study, 62.4% (n=266) patients lived alone but 79.8% (n=340) had an unwitnessed fall, where they were found on the floor after the event; this indicates that even when older adults do live with someone, unwitnessed falls are frequent.
The findings here echo previous research, which indicates the prevalence of unwitnessed falls to be as high as 82% (Segev-Jacubovski et al, 2011) and that older adults living alone in the community or in sheltered housing are unlikely to be able to get up unaided (Fleming and Brayne, 2008). Poor balance and gait and the use of mobility devices (which were highly prevalent in this study's sample) are known to further inhibit older adults from being able to get up from the floor themselves (Tinetti et al, 1993).
Long lie events are associated with further sequelae such as rhabdomyolysis, pressure ulcers, dehydration, hypothermia and pneumonia, and they frequently result in hospital admission (Tinetti et al, 1993) and a move into long-term care (Fleming and Brayne, 2008).
In some situations, long lie events are modifiable and the ambulance service can directly reduce occurrence of long lies by improving the identification of patients who have experienced a period of recumbency on the floor of <1 hour and responding accordingly. Given the deleterious consequences associated with long lie events, more emphasis needs to be placed upon timely identification of modifiable long lies at the time of first contact with the ambulance service, although high, sustained pressures on ambulance services may preclude this.
Limitations
The authors were unable to collect data for all older adults eligible to be attended by the FRRS as some patients lacked capacity, for example, and could not be approached regarding the study. This is a limitation as, for instance, older adults who lack mental capacity may fall for different reasons than those with capacity.
The FRRS was unavailable at night so patients who had a fall between 19:00 and 07:00 were not enrolled. It is plausible that older adults who fall at night differ from those who fall during the day.
This study examined a sample of the population of North East England and, arguably, older adults elsewhere in the UK may present differently and may fall for different reasons. The authors could not account for this variability in their study, which serves to reduce the generalisability of the findings.
The study was not statistically powered to detect a difference between the predictive characteristics of those conveyed and those who were not.
Implications for practice
This study has demonstrated that despite the FRRS having access to alternative and rapid care pathways, conveyance to the ED is unavoidable when the patient has experienced an injurious fall, pain or a long lie event.
In many instances, it is possible to confirm this at the first point of contact with the ambulance service, traditionally during telephone triage. To use the FRRS more efficiently and prioritise the intervention for patients who can receive the most benefit, ambulance services need to promote the identification of the predictors of conveyance identified in this study at the time of the call.
Conclusion
This study has demonstrated it is possible to predict when conveyance or non-conveyance is more likely to occur. The predictive characteristics of conveyance to the ED are different from the predictive characteristics of experiencing a fall. Conveyance to the ED is a consequence of a fall event rather than being linked to an increased risk of experiencing a fall. An undifferentiated approach towards dispatching the FRRS to all older adults who fall is unwarranted.
Further research is required to see how patients can be stratified at the first point of contact and the impact this has on practice and patient satisfaction.