The objective of our study was to describe the prevalence and characteristics of falls in adults with intellectual disability living in a residential care setting and to define differences between fallers and non-fallers in younger and older resident groups. In contrast to the general population, falls are a problem for both aged and younger adults with intellectual disability living in a residential care setting. Falls of 147 residents, aged between 21-89 years with different grades of ID, were recorded prospectively over a 12 months period using a digital fall report form. For all participants, a total of 140 falls were reported and high fall rates per person-year were found in the younger (0.85) as well as in the older aged residents (1.06).
Aspects of Aging in People with ID
Age-related processes are marked by physiological, psychological and social decline, threatening health determinants, quality of life, functional status, frailty and mobility of individuals (Rapp, Becker, Cameron, König, & Büchele, 2011; Spirduso, Francis & MacRae, 2005). Due to these declines, research has demonstrated that the risk of falling increases with age. In the age group 65 years and older living in the community, 30% or more adults experienced one or more falls each year (Lamb, Jørstad-Stein, Hauer, & Becker, 2005; Rubenstein, 2006). This rate is even higher among those living in nursing homes or residential care (Becker & Rapp, 2010; Rubenstein, 2006). Furthermore, in the residential care setting, additional help and support necessitated by each fall will produce costs about $700-1,000US (Haines et al., 2013; Nurmi & Lüthje, 2002). In case of injuries leading to hospitalization, costs will increase to $5,500-42,000US (Heinrich, Rapp, Rissmann, Becker, & König, 2010) depending on the degree of injury.
However, in contrast to the existing expertise in community-dwelling and residential care settings, knowledge concerning the prevalence of falls, fall risk factors and prevention of falls in both younger and older adults with intellectual disability (ID) is scarce, and the topic is less frequently investigated (Willgoss, Yohannes, & Mitchell, 2010).
With advances in medical management and public health, the lifespan of people with ID has also increased (Haveman et al., 2010; World Health Organization, 2001). During the normal aging process this population is exposed to the same challenges as the general population (Bruckner & Herge, 2003; Evenhuis, Hermans, Hilgenkamp, Bastiaanse, & Echteld, 2012), but at an earlier age (Chiba et al., 2009; Cox, Clemson, Stancliffe, Durvasula, & Sherrington, 2010; Smulders, Enkelaar, Weerdesteyn, Geurts, & van Schrojenstein Lantman-de Valk, 2012), especially on the physiological side, which may cause declines in strength and balance and lead to higher fall risk at a younger age (Connolly, 2001; Smulders et al., 2012).
Falls and Fall Risk Factors in People With ID
The prevalence of falls in younger and older adults with ID (Chiba et al., 2009; Cox et al., 2010; Hsieh, Rimmer, & Heller, 2012, Rapp et al., 2011; Smulders et al. 2012; Spirduso et al., 2005) seems to be comparable to these in independent older people (O'Loughlin, Robitaille, Boivin, & Suissa, 1993; Rubenstein, 2006), although a higher fracture and injury rate is reported (Finlayson, Morrison, Jackson, Mantry, & Cooper, 2010; Lamb et al., 2005; Rubenstein, 2006; Sherrard, Tonge, & Ozanne-Smith, 2001). Several studies have observed that people with ID seem to have the same risk factors for falls as people in the older general population (Cox et al., 2010; World Health Organization, n.d.). For example, risk factors for falls in people with ID were found in the areas of physical function (gait, balance, strength), related impairments (epilepsy, incontinence, visual deficits, seizures), medication (anti-epileptic drugs, more than 4 drugs) and even in the environment (bad lighting, flooring) (Chiba et al., 2009; Cox et al., 2010; Enkelaar, Smulders, van-Schrojenstein Lantman-de Valk, Weerdeseteyn, & Geurts, 2013; Lamb et al., 2005; Wagemans & Cluitmans, 2006; Willgoss et al., 2010). Another study by Smulders et al. (2013) reported falls mainly occurring during walking; these falls were most often attributed to tripping or slipping, and most falls were reported in the afternoon or evening.
Gaps in Fall Research in People With ID
Most of the previously mentioned risk factors for falls are based on retrospective data collection within periods of time that range from three months to one year (Bruckner & Herge, 2003; Chiba et al., 2009; Cox et al., 2010; Evenhuis et al., 2012; Hsieh et al., 2012). Research in the general population tells us that this method is known to produce a recall bias, and may lead to false conclusions (Chiba et al., 2009; Cox et al., 2010; Freiberger & de Vreede, 2011; Ganz, Higashi, & Rubenstein, 2005; Smulders et al., 2012). As suggested by the Prevention of Falls Network Europe (ProFaNE-Group), a prospective collection of data on falls over a 12 month period is the recommended way to collect and describe falls (Lamb et al., 2005). Only one other research group has described falls and fall risk factors in people with ID using this method (Smulders et al., 2012). In their longitudinal cohort study, they found a fall rate of 1.00 fall per person-year in older adults aged 51 and above, while 46% of the participants experienced one or more falls in a 12 month period. Other studies have observed that adults younger than 50 years experienced high fall rates as well (Cox et al., 2010; Hsieh et al., 2012), which is consistent with the previously noted early-onset declines in strength and balance in people with ID.
To acquire deeper information on prevalence and characteristics of falls even in younger adults with ID, more prospective research is required. In addition, further knowledge on risk factors for falls is necessary.
The objective of this article is two-fold: 1) to describe the prevalence and characteristics of falls in adults with intellectual disability aged 21 to 89 years living in a residential care setting, and 2) to define differences between fallers and non-fallers in younger and older adults.
To the best of our knowledge, this is the first prospective study on characteristics of falls in adults with ID over a broad age range that expands on the knowledge on older adults provided by Smulders et al. (2012). In our study a fall is defined as “an unexpected event in which the participants come to rest on the ground, floor, or lower level” in accordance with the ProFaNE recommendations (Lamb et al., 2005).
This article describes the results of a 12 months exploratory longitudinal cohort study with a prospective explorative design.
The study took place in a residential facility in Bavaria, Germany. The residential facility accommodates people with intellectual disability living in single or double private rooms. Regarding the definition of ID by the World Health Organisation (n.d.), people having limitations in the following seven factors live in the residential facility: (1) tasks of daily living (e.g., cooking, shopping); (2) individual supply (e.g., toileting, eating); (3) having social relations (e.g., friendship, relation to relatives); (4) participating at cultural and social life (e.g., meeting in groups, participation at residential program); (5) communication (e.g., use of communication aids, stereoscopic orientation); (6) emotional and psychological development (e.g., coping with fear and restlessness); (7) health promotion and health preservation (e.g., see the doctor). Therefore, our population had at least one of the above-mentioned limitations or a combination of limitations, which was also the basis for inclusion in our study.
Up to 20 residents live on the same floor or in the same building, supported by a number of staff members. At present there are more than 200 residents. Depending on their severity of disability, some residents work, some receive special support during the day, and some are already retired.
Recruitment and Ethical Approval
In accordance with the exploratory study design, all residents were allowed to take part. No exclusion criteria were defined and no power calculation was performed. The number of signed consent forms determined the sample size.
Participants were recruited by staff members; potential participants were asked to take part in the study after being provided with a short explanation of the study aims. In addition, a letter with further information was sent to the respective legal representatives of the residents. Two different versions of informed consent form were provided: one version in simple language was provided for the residents, while the other version, with more formal language, was provided for the legal representatives. The participants or their legal representatives signed written informed consent forms in accordance with the declaration of Helsinki. Ultimately 147 participants were enrolled. Prior to commencing the study, ethical clearance and clearance for data management was obtained by the ethical committee of the University of Erlangen-Nürnberg.
Data Source – Measurements
Collection of baseline data was carried out by exporting all relevant data from internal residential files. All staff members with personal access were able to update data on the internal network, which guaranteed an up-to-date database. Wherever possible, standardized procedures were used for obtaining information in the internal files (e.g., care-level, level of ID).
The following data were gathered from the internal residential files: gender, birthday, height and weight, care-level, use of wheelchair or walking aid, use of incontinence products, use of orthopedic shoes, use of glasses, diagnoses, and the name(s) and amount(s) of medication.
Data collection on falls
For collecting data on falls prospectively, a digital fall report form was implemented after a 3-month pilot test period, which was based on report forms previously used and validated by our study group (Blank et al., 2011; Freiberger, Haberle, Spirduso, & Zijlstra, 2012). After an educational training session, the form was made available to all staff members via the internal network of the institution. A copy of the report form was automatically sent via e-mail to the research group after the staff members saved the completed form following a fall. Staff members completed the fall report form if a fall was reported by the residents themselves or noticed by other staff members or residents.
During the 3-month pilot test period, the digital fall report form was adapted and modified based on the experience and feedback provided by staff members (via e-mail or telephone) to increase usability and to reduce missing data (e.g., identification number of participant, date of fall).
As recommended by the ProFaNE- Group (Lamb et al., 2005), the following data were gathered using the fall report form: date, time, place, description of activity done prior to fall, worn shoes, injuries, used walking aid, glasses worn, and a detailed written description of the fall event. In addition, the fall report form asked for the most likely reason for the fall. These reasons were defined by Rubenstein (2006) and Todd & Skelton (2004).
Additionally, a questionnaire, reporting on the workload of staff members following a fall, was implemented (e.g., requesting information on additional care needed for residents, guidance to hospital or general practitioner, and the time required to carry out the fall report form). This data will be used later for analyses of costs associated with falls. In this paper, only the total number of minutes of additional workload is presented.
To reduce bias in collecting data on falls, all staff members were trained in workshops on how to define falls and reasons for falls, and how to use the fall report form. In addition, the definition and further examples of falls were presented at the beginning of each report form. All staff members were reminded to fill in the report form regularly with quarterly update letters. Based on the definition of falls used, both epileptic falls and falls that occurred during transfer actions by a staff member were included. Unlike the recommendations of using daily fall calendars (Lamb et al., 2005), falls were only reported when they happened, due to the requirements of the institution. All staff members in residential care were required by law to document each fall event in a digital residential file.
Two age groups were defined by splitting the sample at the median age so as to have two comparable sample sizes. In addition, a “mild-to-moderate” ID-level group (MM) and a “severe-to-profound” ID-level group (SP) were defined. Persons who reported no falls were classed as “non-fallers” (NF), persons with one or more falls as “fallers” (F). For medication, two groups were generated for persons taking ≤ 4 or > 4 pharmaceutical drugs (Todd & Skelton, 2004). Having a BMI > 30 was defined as being obese.
Level of care is integrated into the German health care system to categorize the amount of care needed in persons with ID. The assessment is carried out by a physician who categorizes the recipients into 1 of 3 levels (levels 1, 2, and 3 requiring basic care such as washing, feeding, or dressing for at least 0.75, 2.00 and 4.00 hours per day, respectively) (Becker, Leistner & Nikolaus, 1998). People with ID who need less than 0.75 hours per day are categorized as care-level “0”. Thus, the level of care can be seen as an indicator of functional status in persons with ID.
Overall, frequencies and percentages were used to describe categorical data, while means and standard deviations (SD) were used to describe continuous variables. In a first step, univariate statistics were used to analyze differences between fallers and non-fallers. In the case of categorical data (e.g., ID-level, age-group, being obese), χ2-analyses were performed. In the case of interval data (e.g., fall rate, injury fall rate), Mann-Whitney-U-Test was used, because the data were not normally distributed (tested with Shapiro-Wilk-Test).
In a second step, all significant variables from the univariate analyses were entered into multivariate analyses, using a binary logistic regression model, to determine the independent power of each variable in predicting a faller or not.
“Fall rate” was defined as the number of falls divided by person-years. This allowed for the inclusion of data collected from subjects who dropped out during the twelve months period. The alpha level was set to .05. Data management was carried out with IBM SPSS Statistics v20.
If missing data was found in baseline data, it was requested from the residential facility. The digital fall report form reported an error message if relevant data was missing, preventing the collection of missing data in fall reports.
During the 12 months period, 147 participants with a mean age of 55.2 (SD = 16.1) and a median age of 55 years were tracked. With a mean age of 41.5 years (SD = 8.4) in the younger and 68.3 years (SD = 9.2) in the older age group, the age ranged between 21 and 89 years. Two participants died during the 12 months and three moved away. Most of the participants did not use any kind of walking aid (61.2%), such as a cane or crutch, but 23.1% were fully dependent on a wheelchair. Further sample characteristics can be found in Table 1.
Prevalence and Fall Rates
All together, 140 falls were reported by 147 participants in 146.97 person-years. This results in a fall rate per person-year of 0.96 with a fall rate of 0.85 in the younger age group (reporting 61 falls) and a fall rate of 1.06 in the older age group (reporting 79 falls).
All falls were reported by 51 (34.7%) of the 147 participants, with 24 (33.3%) fallers in the younger and 27 (36.0%) fallers in the older age group. Out of the 51 fallers, 22 (15.0%) experienced two or more falls (recurrent-fallers) during the 12 months, with 10 (45.5%) recurrent-fallers in the younger group and 12 (54.5%) in the older group.
In 48 (34.3%) of the 140 falls, injuries like cuts, abrasions or fractures were reported, with a higher prevalence of injurious falls in the older group (28 injury-falls, or 35.4%).
Differences between the younger and older age groups regarding the fall rate and injury fall rate per person-year were found to be not significant (fall rate: U = 2620, z = -37, ns.; injury fall rate: U = 2537, z = .87, ns.).
Comparison Between Fallers and Non-Fallers
Univariate analyses were employed, with the following factors being significant at p < .05 for the full sample (Table 1): care-level, level of ID, autism, use of incontinence products and orthopedic shoes.
For the younger group, only care-level was found to be significant, with those having care-level 2-3 (19; 79.2%) being more likely to be a faller. For the older group, the use of incontinence products was found to be significant, with those using incontinence products (16; 59.3%) being more likely to be a faller.
Fall Risk Identification
After entering all significant variables from univariate analyses in Table 1 into a binary logistic regression model, only “orthopedic shoes (yes)” remained as an independent risk factor for being a faller or not (Table 2) in the full sample. With an odds ratio of 3.32, this variable indicated a much higher fall risk for those residents.
Additional Fall Characteristics and Workload
Table 3 presents further characteristics related to the falls reported in falls per person-year. For example, it can be seen that 42.1% of all falls reported happened in the winter, resulting in high fall rates for both the younger group (0.38) and the older group (0.43). Looking at the time of day, younger residents had the highest fall rate in the afternoon (0.24), while older residents had the highest fall rates in the morning or evening (both 0.22). The most mentioned reasons for falls presumed by the staff members were gait and balance problems, followed by inattention or accidental falls.
Looking at the additional workload for staff members after a fall event, we found 29.7 minutes (SD = 64.54) as a mean of additional workload for one fall, with a minimum of 15 seconds for just helping the fallen person to stand up, and a maximum of nearly 10 hours for bringing the faller to the hospital and giving additional support. The mean time for filling in the fall report form itself was 6.48 minutes (SD = 5.23).
This study was primarily developed to observe and describe the prevalence and characteristics of falls in younger and older adults with ID living in a residential facility, using a prospective study design. With 140 falls reported over one year by all participants, and a fall rate of 0.95 falls per person-year and 34.7% of the participants experiencing one or more falls, this study demonstrated that falls are highly relevant to both younger and older adults with ID. Conducting this study has allowed us to expand on existing knowledge already reported by Smulders et al. (2012) and other researchers (e.g., Chiba et al., 2009; Cox et al., 2010).
In addition, our data confirm the conclusions reached by the previously cited researchers whose studies reported falls retrospectively, and our time period of 12 months allows easier comparison with other data on falls. Finally, the use of a standard fall definition, a standardized fall record form, and training of staff members on the topic of falls strengthened our findings.
A limitation of our study is the fact that falls were only reported by staff members. Therefore, some falls may not have been reported due to incorrect interpretation of a fall or the possibility of forgotten and unreported falls due to the high workload of the staff. Furthermore, falls that were not seen by staff members may have been missed if the residents themselves did not report them.
Another limitation is the use of a convenient sample that was chosen in designing the study. The sample was determined by the residents living in the facility and was not limited by the use of specific exclusion or inclusion criteria. Therefore, generalizing these findings to others (e.g., adults with ID living in community setting) has to be carried out very carefully.
Prevalence and Rates of Injuries
Our results support previous findings (Chiba et al., 2009) that fall prevalence in people with ID is approximately 30%. This is close to fall prevalence rates seen in the general older population (30-60%) (Rubenstein, 2006). The high fall prevalence and fall rates per person-year in the older population with ID found in one study, that is, 45%; 1.00 (Smulders et al., 2013), are now supported by our findings. In addition, we found a prevalence rate of 33.3% and a fall rate per person-year of 0.85 in our younger age group, which is lower than those reported for older people from the general population living in residential care (falls per bed annually, mean = 1.7), but higher than those from community-dwelling persons (mean = 0.65) (Rubenstein, 2006). These results demonstrate the urgent need to include younger adults with ID in research on falls in residential care settings.
Looking at injurious falls, we found much lower prevalence rates of injury falls (>30%) compared to previous findings from older residential people with ID (>70% (Smulders et al., 2013) or to those living in the community (>80% (Cox et al., 2010); 50% (Hsieh, Keller, & Miller, 2001)). One explanation for our low prevalence rate of injurious falls could be differences in the definition of an “injurious fall” (Schwenk et al., 2012). It is of utmost importance to further investigate injurious falls in persons with ID in order to target the right ID population with future fall prevention strategies. In addition, circumstances of injurious falls have to be investigated in this specific population in order to develop the right preventive strategy.
As already found by Willgoss et al. (2010), risk factors for falls in people with ID seem to be multifactorial and can include mobility, behavioral issues, environment or other personal factors. For example, Hsieh et al. (2012) found differences between fallers and non-fallers based on the use of incontinence products and the level of ID, respectively, which is supported by our findings. In contrast to Cox et al. (2010) and Hsieh et al. (2012), we found that age does not play an important role. In congruence with Chiba et al. (2009), we also found that gender does not appear to be a risk factor in people with ID. These heterogeneous findings should be further investigated.
People with ID are known to have a high rate of medication use (Doan, Lennox, Taylor-Gomez, & Ware, 2013) and the intake of antiepileptic drugs (Wagemans & Cluitmans, 2006) or drugs with a psychiatric basis (Hale, Bray, & Littmann, 2007) are known to potentially increase the fall risk in people with ID. In contrast to the findings of Todd & Skelton (2004) in the general older population, we found no significant differences to fallers taking more or less than 4 medications. Therefore, the risk factor of “taking more than 4 medications” should be reviewed and adapted for people with ID due to their high rate of medication. Further, more research is needed on the influence of different medication groups and their potential influence in increasing fall risk. Due to the high rate of medication over their lifespan, people with ID may be more familiar with some side effects of certain medications (e.g., dizziness, tiredness), and may have consequently adapted their daily life to those side effects.
Overall, in our final model, only the use of orthopedic shoes remained as an independent significant risk factor. This is in line with the findings by Smulders et al. (2013), who found high fall prevalence in those residents wearing orthopedic shoes. Again, more research is needed in comparable ID populations to define the most important risk factors for falls. Furthermore, our results demonstrated that fall risk factors may be different in younger and older persons with ID. No information is currently available on this issue.
Smulders et al. (2013) found the winter season to have the highest prevalence of falls (34.2%) and also observed that most falls occurred in the afternoon (45%) and while walking (63.3%). Above 50% of all falls were caused by gait and balance problems in their sample (aged 51 and above). We found similar trends in our older population, and in our younger population, we observed high fall rates associated with the above-mentioned factors. As most falls were caused by gait and balance problems, physical decline seems to be one of the most important determining factors for falls in people with ID; this is congruent with falls in the general population (Rubenstein, 2006).
Finally, taking into account the multifactorial approach of a fall—regarding time, place, circumstances and other reasons—complex interventions may provide a good approach in addressing the special needs of younger and older adults in residential facilities.
Nearly 30 minutes of additional help were required for helping a fallen person, filling in the fall report form and carrying out other tasks involved. The additional workload on the staff members caused by a single fall event is high, and other studies have already reported the economic consequences of a fall in the general older population (Findorff, Wyman, Nyman, & Croghan, 2007; Nurmi & Lüthje, 2002; Stevens, Corso, Finkelstein, & Miller, 2006). Nevertheless, further economic investigation due to the special needs of people with ID is needed in this field if effective fall prevention strategies and interventions are to be installed.
Falls in adults with ID living in a residential care facility are not only a problem for older persons: because both younger and older age groups demonstrate nearly the same prevalence rates and comparable causes of falls, we can conclude falls occur nearly as often in younger persons with ID. With such high fall rates, it is important to develop appropriate fall prevention strategies that target both older and younger adults with ID, as both populations are at risk. There is need for further research to define fall risk factors in adults with ID from young adulthood through older ages.
Oral presentation of parts of the manuscript at the 8th World Congress of Active Ageing in Glasgow.
The authors would like to thank Mr. Scharfenberg and Mr. Renner from Regens-Wagner-Stiftung Lauterhofen for their intensive support during the study period.
The PreFallID-study is funded by Regens-Wagner-Stiftung Lauterhofen in Lauterhofen, Bavaria, Germany. Materials for intervention are funded by AOK Bayern in Munich, Bavaria, Germany.
Johannes Salb, University of Erlangen-Nürnberg, Institute for Biomedicine of Aging; Carol Woodward, University of Erlangen-Nürnberg, Institute of Sportscience and Sport; Jens Offenhäußer, University of Hohenheim, Institute of Health Care & Public Management; Clemens Becker, Robert-Bosch Hospital; Cornel Sieber, University of Erlangen-Nürnberg, Institute for Biomedicine of Aging; and Ellen Freiberger, University of Erlangen-Nürnberg, Institute for Biomedicine of Aging.