There were 566 patients admitted to the SIU of which 135 (24%) were females. Over the 5-year period, 132 people (23% of the total number of admissions) experienced at least one fall incident. Table 1 reports the characteristics of fallers.

Forty-one (31%) participants experienced more than one fall (i.e., referred to as recurrent fallers). Most recurrent fallers fell twice (n = 23); this was followed by three falls (n = 11); four falls (n = 4); five falls (n = 2); six falls (n = 1); and seven falls (n = 1). Seventy-eight (59%) participants had at least one fall with a consequence (i.e., consequential fall) including subjective (pain and psychological) and objective (laceration, soft tissue injury, fracture, and head injury) consequences and 11 (8%) participants experienced two or more consequential falls. Table 1 shows characteristics of people who fell once (single fallers) compared to recurrent fallers and people who had no consequences from their fall compared to those with consequences. Univariate analyses found no significant differences in characteristics of single versus recurrent fallers, therefore a multivariate regression model was not generated to predict recurrent fallers. Univariate analyses found a significant relationship between consequential fallers and older age (B = 0.024, S.E = 0 .011, Wald = 5.018, df = 1, p = 0.025) and female gender (χ2 (1) = 4.961, p = 0.026). In the multiple logistic regression model (Table 2), females were approximately 3.5 times more likely to have a consequential fall (Wald χ 2 (1) = 5.809, p = 0.016, OR = 3.581,95% CI, 1.269 to 10.103) and the odds of having a consequential fall increased by approximately 3.8% for each increased year in age (Wald χ 2 (1) = 8.371, p = 0.004, OR = 1.038,95% CI, 1.012 to 1.064).

There were a total of 207 falls during the 5-year period. Table 3 reports the circumstances of falls. Falls occurred throughout participants’ admissions, from as early as two days after admission to as late as one day before discharge. Some falls occurred while accessing the community (for leisure or in preparation for discharge home) with the earliest fall in the community recorded 16 days after admission. Falls occurred mostly during the day (68%) and most were unwitnessed (72%). Sunday was the most common day in which falls occurred.

Figure 1 shows the activities at the time of the falls and the locations where falls occurred. The activities most engaged in at the time of falling were transferring and sitting. The most common location of falls was the bedspace with falls occurring during transfers between wheelchair and bed. Falls while sitting occurred doing activities such as picking items up from the floor, completing bowel therapy, reaching for items (e.g. phone charger), putting clothes away in cupboards and reaching down to wash self. Falls while wheeling occurred mostly off the ward in the community or on hospital grounds and were most commonly attributed to environmental factors such as inclines and gutters.

SIU Spinal injuries unit.

Of the 207 total falls, 116 (56%) were experienced by the 41 recurrent fallers. Of the recurrent fallers, 22 (54%) fell more than once in the same location and 26 (63%) fell more than once doing the same activity. Twelve of the recurrent fallers (29%) fell more than once in the same location, while completing the same activity. Recurrent falls occurred mostly in sitting (n = 37 falls, 32%), transferring (n = 33, 28%) and wheeling (n = 26, 22%). Locations of recurrent falls were mostly in the bedspace (n = 54 falls, 47%), bathroom (n = 20, 17%), community (n = 15, 13 %) and hospital grounds (n = 14, 12%).

Nearly half of the total falls, 98 (47%) were consequential. The majority (n = 93, 95%) had minor consequences including pain (n = 45, 44%), lacerations (n = 43, 44%), soft tissue injuries (n = 22, 22%), psychological consequences (n = 6, 6%), and other which included headache, change to taste, difficulty sleeping, haematuria and drowsiness (n = 6, 6%). Serious consequences occurred from five falls (2%) in five different locations and included four fractures (lower limb and nose) and one head injury. These falls occurred mainly with male participants (n = 4, 80%) who only fell once (n = 4, 80%). Their falls were mostly unwitnessed by staff (n = 4, 80%), occurred mostly from sitting (n = 4, 80%). Figure 2 shows the comparison of consequential falls versus non–consequential falls for each activity.

W/C wheelchair, MSCC mobile shower commode chair.

Table 3 shows the comparison between characteristics of non-consequential and consequential falls. Consequential falls had a significantly greater proportion of falls unwitnessed by staff (χ2 = 6.097, df = 1, p = 0.014) and falls occurring on a Sunday (χ2 = 10.820, df = 1, p = 0.001). Fewer consequential falls occurred in the bedspace (χ2 = 11.595, df = 1, p = 0.001) and when completing a transfer (χ2 = 13.861, df = 1, p = 0.000). Table 4 shows the post hoc analysis of non-binary variables for non-consequential versus consequential falls. Other variables that appeared different between non-consequential and consequential falls but were not statistically significant included falls occurring in March and April, in 2020, in autumn, and in the first third of admission.

In the multiple logistic regression model (See Table 2), falls on a Sunday were approximately 5 times more consequential compared to falls on any other day of the week (Wald χ2 = 7.472, df = 1, p = 0.006, OR = 0.196, 95% CI, 0.061 to 0.630). Falls while transferring were approximately 4 times less likely to be consequential compared to falls while completing other activities (Wald χ2 (1) = 9.945, p = 0.002, OR = 4.100, 95% CI, 1.706 to 9.856). Non-significant variables have also been reported in Table 2 for interest.