Article Text
PDF
Abstract
Background Fall-related injuries affect the lives of elderly to a substantial degree. This quasi-experimental study investigates the fall-injury reducing effect of impact absorbing flooring among female nursing home residents.
Methods The intervention site is a nursing home in Sweden where impact absorbing flooring was installed in parts of one of six wards (six out of 10 apartments (excluding bathrooms), the communal dining-room and parts of the corridor). The impact absorbing flooring is a 12 mm thick closed cell flexible polyurethane/polyurea composite tile (500×500 mm) with an exterior surface of polyurethane/polyurea. A generalised linear model (log-binomial) was used to calculate the RR of injury from falls on impact absorbing flooring compared to falls on regular flooring, adjusted for age, body mass index, visual and cognitive impairments.
Results During the study period (1 October 2011 to 31 March 2014), 254 falls occurred on regular flooring and 77 falls on impact absorbing flooring. The injury/fall rate was 30.3% for falls on regular flooring and 16.9% for falls on impact absorbing flooring. Adjusted for covariates, the impact absorbing flooring significantly reduced the RR of injury in the event of a fall by 59% (RR 0.41 (95% Cl 0.20 to 0.80)).
Conclusions This is, to our knowledge, the first study evaluating the injury-reducing effect of impact absorbing flooring in a nursing home showing statistically significant effect. The results from this study are promising, indicating the considerable potential of impact absorbing flooring as a fall-related injury intervention among frail elderly.
Statistics from Altmetric.com
Background
Fall-related injuries are a substantial public health problem among elderly, causing 1500 premature deaths per year in Sweden and costing upwards of Kr9.4 billion (US$1.5 billion as of October 2014).1 Old age, morbidity and level of care increase the risk of fall-related injuries,2 with injury rates considerably higher among elderly living in nursing homes compared to community dwelling elderly.3
Substantial efforts in preventing falls and fall-related injuries among elderly have been made during the last decades, both within the clinical field and in the field of research. Evidence-based interventions have been developed, for example, through improvements in elderly care, training programmes for community dwelling elderly and home safety assessments.4 For elderly living in care facilities or with mental health problems, however, limited interventions have proven to be effective.5 ,6
With few active, evidence-based, methods available to reduce falls and fall-related injuries in nursing homes, impact absorbing flooring has been suggested as a potential passive safety measure. Theoretically, the risk of injury following a fall is based on the relationship between the type of fall, the type of impact, energy absorption and bone strength.7–9 Impact absorbing flooring aims to affect the energy absorption in a fall, therefore, theoretically, reducing the risk of injury once a fall has occurred. The effectiveness of preventing injury by reducing force is well known10 and has been successfully applied in other fields of injury prevention, for example, within traffic safety.
There is evidence for these types of interventions from studies in laboratorial settings11–14 and a correlation between the risk of fracture and flooring stiffness has been established.15 Limited studies have evaluated the effect of impact reducing flooring in a clinical setting. In a randomised controlled pilot study of impact absorbing flooring (8.3 mm thick Tarkett Omnisports EXCEL), tested in a hospital setting, preliminary results show a non-significant fall-related injury reducing effect of 42%.16 In a retrospective study, evaluating the effect of the SoftCell flooring, a clinically significant reduction of hip fractures was seen with 2.4% of falls resulting in hip fractures on regular flooring and 0% on impact reducing flooring.17 Also, a prospective observational study, examining 6641 falls by differences in impact absorbing ability of the landing surface related to hip fracture outcome, showed the importance of adjusting flooring materials in housing for frail elderly, as the rate of hip fractures were significantly lower on wooden flooring with carpet compared to more rigid flooring.15
Previous research has shown that the risk of injury following a fall is highest in nursing homes.3 Therefore, in this quasi-experimental study, the aim is to evaluate the effect of impact absorbing flooring on fall-related injury risk in a nursing home setting.
Methods
The studied nursing home contains 60 apartments divided into six identical wards, accommodating approximately 60 individuals (of which approximately 70% are women). As with most Swedish nursing homes, a majority of the apartments are designed for one person, although in some situations couples can live together. Due to Swedish elderly care being regulated by the Social Services Act (SoL 2001) and private nursing homes being very rare, the population (elderly people living in nursing homes) is relatively homogenous.18 In general, older adults living in residential care facilities in Sweden are disabled with cognitive and/or physical impairments and require supervision or functional support and care. The residents in the studied nursing home had 24 h access to assistance with activities of daily living and medical care.
Impact absorbing flooring was installed in parts of one of the wards (six out of 10 apartments, the communal dining-room and parts of the corridor), a total of 350 m2. Flooring in the control areas was vinyl, linoleum or ceramic tiles, all with concrete underlay. Control areas were present on all six wards. These areas consisted of bedrooms, communal areas, living-rooms in similarity to the intervention areas. The impact absorbing flooring was not installed in bathrooms because the flooring was not approved for wet areas. Therefore, all bathrooms were included in the control areas.
The installed impact absorbing flooring, marketed under the trademark Kradal, is a 12 mm thick closed cell flexible polyurethane/polyurea composite tile (500×500 mm) with an exterior surface of polyurethane/polyurea elastomers approximately 1.5 mm thick. The interior is closed cell flexible polyurethane. The tiles were laid onto the subfloor (concrete) with approximately 4 mm flexible grouts between tiles and then coated in order to seal the floor's top surface. The flooring, according to the product testing report, reduces the energy by 65%–80% compared to concrete.19
The Kradal flooring has previously been tested for its effect on balance, showing no difference in gait stability compared to vinyl and carpet with commercial underlay.20 It has also been tested for impact reduction during lateral falls on the pelvis, showing an overall protective effect, especially for frail elderly.21
Data collection
Data collection took place between 1 October 2011 and 31 March 2014 by the registered nurses at the nursing home. All residents were invited to participate in the study and a written consent was collected, with 94% agreeing to participate. Participants were continuously recruited, based on when they moved in, during the study period. In cases where the resident could not give consent, the next of kin was asked to determine whether the resident was to participate. Ethical approval was received from the regional ethical board at the University of Uppsala, registration number 2011/147.
During the data collection, all falls were registered in accordance with the organisation's existing injury surveillance system. A fall was defined as ‘an unexpected event in which the participants come to rest on the ground, floor or lower level’.22 Within the surveillance system, information concerning the fall was collected, for example, date, time, location, type of flooring, activity and whether a hip protector was worn. In order to obtain more detailed information on resulting injuries, the type of injury, diagnostic tools and treatment were noted. The outcomes of the fall-events were categorised as recommended;23 ‘no injury’, ‘minor’, ‘moderate’ or ‘severe’ with the addition of the category ‘death’. When it was clear that the injury occurred in contact with something other than the flooring, the event was excluded. Relevant individual information on age, sex, body height and weight was collected from the patient medical records by the nurses in charge of data collection. Downton fall risk index24 was used to collect data on the following covariates; medications (sedatives/tranquillizers/neuroleptics, antidepressives), sensory deficits, cognitive impairment (cognitive impairment or no cognitive impairment) and gait ability (stabile gait (with or without walking aid), non-stabile gait or no ability to walk). Bone mass density (BMD) (T-score) was measured at the heel with the DXL Calscan (Demetech).
Due to the fact that the occupants of the nursing home were not isolated to one type of flooring, two separate groups of individuals were not compared. None of the residents were exposed only to impact reducing flooring. Rather, two separate groups of fall-events (falls on regular flooring and falls on impact absorbing flooring) were compared.
Baseline
The study population consisted of 57 women, with a mean age of 85.42 (SD 6.41, range 69–98) and an average body mass index (BMI) of 24.7 (SD 4.53, range 15–41). No participant moved out but 20 women died and new residents moved in. Therefore, the length of stay varied between the participants. The data collection on fall-events lasted for a total of 31 748 bed days. Of the 57 women included in the study, 68.4% fell at least once. Six women fell on impact absorbing flooring, 36 on regular flooring and three on both surfaces. In total, 331 fall-events were recorded. Leading to 10.4 falls per 1000 (3.8/year) bed days in the total population.
Statistical analysis
The statistical analysis began with χ2 tests and t tests to investigate differences in the individual characteristics of the residents linked to fall-events on impact absorbing flooring versus fall-events on regular flooring. Factors that were statistically significant in the bivariate analyses were included in a generalised linear model (log-binomial), presenting results as RR. A generalised linear model (log-binomial), which estimates RR, is preferred over logistic regression analyses when the incidence of the outcome of interest is more than 10% and also has the advantage that results expressed as RR are interpreted more intuitively.25 SEs were clustered at the individual level to account for within-individual autocorrelation between fall-events. The predictor in the model was flooring surface (impact absorbing or regular) and the outcome was injury or no injury. The model describes how the risk of getting injured in a fall-event differs between impact absorbing and regular flooring adjusted for age, BMI, visual and cognitive impairments. Data was analysed using IBM SPSS Statistics V.20 and statistical significance was set at p<0.05.
Results
During the studied time period, 254 falls were registered on regular flooring and 77 falls on impact absorbing flooring. Most falls occurred during the afternoon and evening (14:00 to 22:00 (42%)), followed by the night (22:00 to 06:00 (32%)) and during the day (06:00 to 14:00 (25%)). The rate of injuries/falls varied with the highest rate in the morning (06:00 to 10:00 (42%)) and lowest during the evening (18:00 to 22:00 (20%)). The falls mainly occurred in the residents’ apartments (72%), followed by communal areas (17%) and bathrooms (12%). The most common activity before falling was walking or transferring (eg, from bed to chair). Hip protectors were used in 24% of the events (28% missing). In terms of injury per fall, 30.3% of the falls on regular flooring and 16.9% of the falls on impact absorbing flooring led to injury (table 1).
Number of falls with no injuries, number of falls with injuries, total number of falls and percentage of injuries/falls, presented by flooring surfaces
A majority (80%) of the injuries were minor (distinct pain, bruising, swelling), 1%–2% were hip fractures and the remainder were other fractures and abrasions/cuts. The distributions of injury types, when comparing falls on different surfaces, were similar for severe injuries. A decrease in injury rate/fall was seen for minor and moderate injuries (table 2).
Distribution of injury severity divided by no injury, minor, moderate major and death, presented by flooring surfaces. Type of injury within major injury presented by fracture type
Differences in covariates between fall-events on different surfaces were tested, and the results showed significant differences in age, BMI, BMD, visual and cognitive impairments (table 3). These covariates were thus considered for inclusion in the log-binomial regression model to adjust the effect estimate for potential confounding (table 4).
Differences for covariates, divided by flooring surface, significance in differences calculated by χ2 test or t tests
Generalised linear model (log-binomial) calculating fall-injury reduction effect of impact absorbing flooring, adjusted for age, BMI, visual and cognitive impairment s
The unadjusted effect estimate of impact absorbing flooring showed a risk reduction of 47% per fall-event compared to regular flooring (RR 0.53 (95% CI 0.29 to 0.96)). Adjusted for covariates, the effect size was increased to 59% (RR 0.41 (95% CI 0.20 to 0.80)) (table 4). BMI was the only significant covariate in the final model. BMD was excluded from the final model due to a high level of missing values. The addition of the BMD covariate did not change the effect size to any substantial degree (RR 0.38 (95% CI 0.22 to 0.67)). The estimated risk difference per fall-event was −20.03 percentage points (95% CI −35.36 to −4.69).
A sensitivity analysis, including hip protector use as a covariate along with the other independent variables used in the final model, showed an increase in effect size of the impact absorbing flooring on injuries (RR 0.31 (95% CI 0.21 to 0.47)).
Discussion
The results indicate that after controlling for observed confounders, the tested impact absorbing flooring can reduce the risk of injury per fall-event by 59% compared to regular flooring among women in a nursing home setting. Comparable research in clinical settings is limited, but the results from this study are similar to those of previous research.
Previous intervention studies in fall-related research have rarely used injury as an outcome measure. Instead, common outcomes are reductions in fall rate or the number of people falling.26 By applying a biomechanical injury prevention perspective on fall-related injury risk, other approaches to fall-related injury prevention are possible. Previous biomechanical research in laboratorial settings has demonstrated the potential of energy-reducing interventions. Focusing primarily on hip fractures, several studies have shown the importance of reducing the force applied to the hip by using materials with energy-reducing capabilities, concluding therefore that impact absorbing flooring is a promising intervention for reducing fall-related injuries among frail elderly.11 ,27 ,28
The present study suggests that the results from the biomechanical studies can be translated to a clinical setting. The results are also promising in comparison to other fall-related injury interventions with regards to effect size. Although few have used injury reduction as an outcome variable, the studies have shown an effectiveness of interventions (regarding community dwelling elderly) between 6% and 33%.29
An interesting incidental finding was the statistically significant effect of BMI. In the final model, BMI was the only other statistically significant variable showing a decrease in risk of injury by 8% for every unit increase in BMI. The inverse correlation between BMI and risk of hip fracture is well established30 ,31 and our results indicate that this could apply to other fall-related injuries as well. How the effect of impact absorbing flooring interacts with BMI is yet to be established; however, results from laboratory experiments indicate that the benefit from such flooring is greater for frail elderly with low BMI.32 Given that the individuals living in nursing homes in Sweden are generally frail, this further supports the use of impact absorbing flooring in these settings.
Strengths and limitations
The main strength of this study is that it adds evidence from a clinical perspective to the body of knowledge involving impact absorbing flooring that to date has been mainly theoretical, combined with studies on younger participants in laboratorial settings. There are inherent difficulties with randomised controlled trials within injury research due to aspects related to ethics and study costs. Therefore, the quasi-experimental study design, which has been used in this study, has an important role to play when evaluating the effect of interventions in injury prevention.33 ,34
There are, however, also a number of limitations in this study. First, due to limited data it is at present not possible to differentiate by injury severity in the statistical model due to a lack of statistical power. This is especially relevant with regards to analysing the effect of impact absorbing flooring on hip fractures, the most common serious fall-related injury.35 In order to achieve sufficient statistical power regarding severe injuries considerably more data would be required.
Second, a possible limitation in this study is a potential change in the distribution of injury severity as a result of the effectiveness of the intervention. If, for example, the floor was effective in reducing severe injuries, an increase in minor injuries might have been expected. However, this was not the case. Instead, decreases were found within the categories minor and moderate injuries, while the percentages of major injuries per fall were the same for regular and impact absorbing flooring. The reason for this is most likely due to the limited amount of data. When not differentiating between fractures and other injuries in the model, the effect of BMD becomes less relevant. In this study, BMD was not included in the final model, partly due to the minimal potential effect and also the number of missing cases. In future studies, however, when differentiating between injury severities, BMD should be included. Future studies, based on an increased data collection or larger sample size, will also generate the possibility of studying the effect of impact absorbing flooring on men; an aspect that was not possible in this study due to considerably fewer men living at the nursing home. Further, a number of different flooring types (vinyl, linoleum or ceramic tiles on concrete) were classified in this study as regular flooring. Although a difference in energy reduction does exist between these flooring types, the differences are minimal compared to the intervention flooring. Therefore, all could be considered rigid and were regarded to be comparable as control areas. A risk of bias is that the type of falls and fall injuries can vary between rooms, especially with regards to bathrooms. However, a sensitivity analysis, excluding falls that occurred in bathrooms (n=41), did not affect the effect estimate to any considerable degree (RR 0.43 (95% CI 0.22 to 0.84)).
Third, although previous studies show that softer floors do not affect balance20 ,36 and that the intervention flooring has previously been tested for a potential effect on balance,37 the risk of the impact absorbing flooring affecting the risk of falling cannot be eliminated. In order to measure and compare fall-rates on different surfaces, it would be necessary to control for exposure. This has not been possible in this study, though should be included in future studies. As the flooring was installed in parts of the ward, the same individual could be exposed at certain times of the day and not exposed at other times. The problem with controlling for exposure has been noted before15 and needs to be addressed in future studies.
Lastly, in all intervention studies, the risk of a Hawthorne effect is present.38 In this study, the risk is low due to a majority of the participants having cognitive impairments and that the intervention became a part of their living environment. However, there is a risk that the staff at the nursing home changed or adapted their behaviour with the knowledge, or perceived knowledge, of a reduced risk of injury among the elderly; something that might lead to misclassification for minor injuries when someone falls on protected surfaces. This aspect is important to investigate in forthcoming studies.
Conclusions
Limited evidence-based methods are available to reduce the risk of falls in a nursing home setting.5 Therefore, a focus on fall-related injury prevention can be complementary to fall prevention measures. In populations such as those living in nursing homes, passive interventions offer promising possibilities,39 and impact absorbing flooring is an alternative. This is, to our knowledge, the first study evaluating the injury-reducing effect of impact absorbing flooring in a nursing home showing statistically significant effects. The results from this study are promising, indicating a considerable potential for impact absorbing flooring as a fall-related injury intervention among frail elderly. However, further studies are required to reach conclusive evidence for this type of intervention. A blinded randomised controlled study is presently ongoing, evaluating the SoftCell flooring, with results due in September 2017.40 In the future, it is also important to include effects on the work environment in evaluations of flooring interventions.
What is already known on the subject
Fall-related injuries are a significant public health problem.
Elderly living in nursing homes are especially prone to falls and fall-related injuries.
There are limited evidence-based methods in preventing falls and fall-related injuries among frail elderly.
What this study adds
The results indicate that impact absorbing flooring reduces the risk of fall-related injuries among nursing home residents.
Fall-related injury prevention is complementary to fall prevention, especially among frail elderly.
In populations such as those living in nursing homes, passive interventions should be examined further.
Acknowledgments
We thank the staff at the nursing home, especially Birgit Sandberg, Katrin Oskarsson, Britt-Inger Olsson and Ann-Katrin Nilsson, for their participation in the project and Jari Appelgren for statistical guidance.
References
Footnotes
Contributors JG was responsible for obtaining approvals, the design and implementation of the study and compiling the initial draft of the manuscript. RA, FN and CB contributed significantly to the study design, data interpretation and manuscript drafting. All authors contributed to revisions and approved the final manuscript. The flooring company did not contribute to the study by any kind of founding nor had any role in data collection, analysis or preparation of the manuscript.
Funding This project was supported by Karlstad Municipality, Lansforsakringar (grant number P1/13) and the Swedish Civil Contingencies Agency (grant number 2013-4176).
Competing interests None.
Ethics approval Central Ethical Review Board, Uppsala.
Provenance and peer review Not commissioned; externally peer reviewed.