Introduction

The high prevalence and burden of falls in older adults has been widely described [1, 2]. Meta-analyses suggest that interventions targeting adults aged 65+ could lower the risk of falls by up to 30% [3]. But despite this, population injury rates for falls and injuries from falls continue to rise independent of the ageing of the population [4, 5], suggesting a failure of intervention strategies and/or a change in risk factor profiles of those who fall. The rising rates in injuries from falls and the ageing of the population has led to calls for new approaches to falls prevention [6]. Current falls prevention guidelines predominantly focus on adults over the age of 65 with a high falls risk based on presence of risk factors [7, 8]. While this approach is sensible from the point of view of providing care to those with the highest need, it ignores the opportunity for early preventive interventions.

A past fall is the strongest predictor of a future fall [9], suggesting that primary prevention is important. The second strongest predictor of falls is abnormality of gait or balance [9]. At the population-level, onset of declines in balance ability and other measures of physical functioning are typically observed between ages 40 and 60 [10, 11]. The prevalence of other risk factors for falls, such as syncope, dizziness and chronic conditions, also increase after the age of 50 [12, 13], particularly in women after menopause. These findings suggest that middle-age may be a critical life stage for early interventions for falls prevention.

A first step in exploring opportunities for preventive strategies at younger ages is to establish the prevalence of falls in middle-aged adults. To date, few population-based studies have examined the prevalence of falls in this age group. However, those studies that have reported on this, report prevalences ranging from 11 to 30% [4, 14]. For example, in the 2008 U.S. National Health Interview Survey, the prevalence of falls in the past year among adults aged 45–64 years was 11.4% [4]. In the middle-aged cohort of the Australian Longitudinal Study on Women’s Health, the prevalence of falls varied between 21 and 31% between ages 53–58 and 62–67 [15]. Studies in older adults suggest that country differences exist in the prevalence of falls [2, 16–18], indicating that published prevalences cannot be automatically extrapolated to other countries. Moreover, these studies do not show how the prevalence changes during middle-age.

The aim of this study is to calculate the prevalence of falls in middle-aged adults (aged 40–64 years) from four countries, and to examine how the prevalence changes during middle-age. Data were used from population-based cohort studies from Australia, the Netherlands, Great Britain and Ireland. These studies were selected based on the availability of falls data measured using similar methods in the relevant age-range. Previous publications from these cohort studies reported on the prevalence of falls in older adults only (e.g. [19, 20]) or reported the prevalence for the 50–64 year old group as a whole [21].

Materials and methods

Results

Each cohort has unique features (Table 1). The ALSWH cohort includes a large sample (n = 10556) of women only. The TILDA cohort covers the full age-range of 40–64 years. The LASA cohort includes the most recent data (collected in 2012/13). The ALSWH and NSHD cohorts each have falls data collected at 3 and 2 time points spanning 6 and 10 years of follow-up, respectively.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Sample characteristics of the four cohorts.

https://doi.org/10.1371/journal.pone.0201989.t001

Across the cohorts, the crude prevalence of falls in the previous year varied from 12.8% in 53 year old men in Great Britain (NSHD) to 31.4% in 53–61 year old women in Australia (ALSWH) (Table 2). Weighting for age, sex and education had little influence on the prevalence estimates (i.e. less than 2 percentage points). Overall, the prevalence was higher in Australia (27.8%, women only) and the Netherlands (25.1%) than in Ireland (17.6%) and Great Britain (17.8%, p<0.001). Women (27.0%) had higher prevalences than men (15.2%, p<0.001). In women, the prevalence of falls increased with age from 8.7% in 40–44 year olds, 19.1% in 45–49 year olds, 20.9% in 50–54 year olds, 27.3% in 55–59 year olds, to 29.9% in 60–64 year olds (p<0.001). In men, the prevalence of falls increased with age from 14.7% in 45–49 year olds, 13.4% in 50–54 year olds, 19.3% in 55–59 year olds, to 15.7% in 60–64 year olds (p<0.001). However, there is substantial variation in prevalence between the cohorts within any specific age-band (Fig 1). For example, in 60–64 year old women, the prevalence ranged from 20% in TILDA to 31.4% in ALSWH. Sensitivity analyses estimating the prevalence in 4- and 6-year age-bands strengthen the observation that the prevalence of falls increased with age in women, but less so in men (S2 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1.

Prevalence of falls per 5-year age band in middle-aged women (top) and men (bottom). Presented are the prevalence of falls per 5-year age bands based on the harmonised data across the four cohorts, and the number of participants providing data in each 5-year age band. Note that ALSWH and NSHD participants could be included more than once if they provided data at multiple data collection waves while still falling within the defined age bands.

https://doi.org/10.1371/journal.pone.0201989.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Prevalence of falls in each of the cohorts at each survey.

https://doi.org/10.1371/journal.pone.0201989.t002

Discussion

Across the four cohorts, the prevalence of falls ranged from 8.7% to 31.1%, and varied by age-group, sex and country. On average, the prevalence was markedly higher than the 11.4% found in 45–64 year old adults in the 2008 U.S. National Health Interview Survey[4] and the 2-year prevalence of 21% in 45–64 your old adults in the Baltimore Longitudinal Study of Aging [14]. In addition to age and sex, factors that explain this variation in prevalence between countries include variations in prevalence of key fall risk factors and the measurement of falls. This will be discussed in the following paragraphs. Regardless of the differences in absolute prevalences between the countries, there is consistent evidence of an increase in prevalence of falls across mid-life, particularly in women.

Age is the factor most likely to explain the variation in prevalence between the countries. As shown in the current and previous studies, falls risk increases with age [1, 28]. One would therefore expect the prevalence to be higher in cohorts with older average age. Indeed, the prevalence was higher in cohorts that were older at the time of assessment (Table 2). However, even within 5-year age-bands, substantial variation in prevalence was found between the cohorts. For example, within the 60–64 year old women, the prevalences were 20.1% in Ireland, 22.7% in Great Britain, 28.7% in The Netherlands and 31.1% in Australia (Fig 1). This suggests that other factors also play a role, such as socio-demographic, health and lifestyle factors.

In adults over the age of 65, women typically have a higher prevalence of falls than men [2, 28]. For example, a meta-analyses of observational studies showed that women have a 1.3 times higher odds of falls than men [28]. The current study also confirms this for 40–64 year olds. The sex-differences in prevalence were more pronounced in the Netherlands and Great Britain than in Ireland (Table 2). The Australian cohort included women only and also reported the highest prevalence. It therefore seems likely that differences in sex distribution contribute to the variation in prevalence between countries. A potential explanation for the stronger increase in prevalence of falls during midlife in women than men may be explained by the concurrent stronger increase in prevalence of risk factors for falls, such as arthritis and cardiovascular diseases, post menopause in women.

The four cohorts were designed to be nationally representative and included mainly Western-Europeans. The Australian cohort included mostly women born in Australia of British or Irish descendency (77.8%). Few studies have examined ethnic differences in falls risk of older adults and the findings are inconsistent, but there is some evidence that falls risk may be lower in African Americans than in White Americans [16, 18] and lower in Asians than in Caucasians globally [16, 17]. Hence the current results may not be generalizable to other populations with different ethnic backgrounds.

Many risk factors for falls have been identified in older adults. It is yet unclear whether the same risk factors are important in middle-aged adults. In older adults, some of the strongest risk factors include mobility limitations and chronic conditions such as diabetes and arthritis [28]. Indeed, the prevalence of diabetes and arthritis were higher in the Dutch cohort than in the British and Irish cohorts (Table 1) and could partly explain the higher prevalence of falls in this cohort. In contrast, the prevalence of arthritis was lower in the Australian cohort than in the Dutch cohort, while the prevalence of falls was higher in the Australian cohort. A previous study by our group found that, diabetes was not identified as a risk factor of falls in Australian middle-aged women, but presence of joint symptoms was [15]. That study also found that different factors were associated with falls at early, middle and later midlife, highlighting the complexity of falls risk at this life stage [15]. For example, high levels of alcohol intake and hearing problems were significant predictors of falls at ages 59–67, but not at earlier ages [15]. The differences found by age, sex and country could be explained by variations by age, sex and country in risk factors for falls. Further research is required to identify key risk factors of falls in middle-age.

Finally, the variation in prevalence between the cohorts may also be partially explained by differences in measurement of falls. While there was little variation in the wording used to ask participants about falls in the past year, there were variations in how the questions were presented. The Irish, British and Dutch cohorts use a conditional approach whereby follow-up questions about the consequences of a fall are only asked if the participant responds positively to the first question about a fall in the past year. In the Australian cohort, the question specifically asked about ‘falls to the ground’ and the follow-up questions about the consequences of a fall are visible in the survey to all participants. It may be that the specific wording and these follow-up questions facilitate the recall of falls. Moreover, in the Australian cohort, some participants responded ‘yes’ to either ‘having had an injury from a fall’ or ‘seeking medical attention after a fall’, but not to the actual falls question. As one cannot have an injury from a fall without actually having a fall, these participants were classified as fallers. This resulted in a 2–2.6 percentage point increase in the prevalence of fallers at each survey. Follow-up questions about the consequences of falls may help improve recall and reduce misclassification.

Falls are currently perceived as a major problem in older adults, but receive little attention in middle-age. While the prevalence of falls is lower in middle-aged adults than in older adults, the current findings show that the prevalence is not low. Previous studies have suggested a U-shaped association between age and fall risk, in which the fall risk is highest in children and seniors [29]. This may result in the misperception that falls are not of concern in middle-aged adults. While falls that result in injuries have the greatest impact on health services use, any fall, with or without injury, can result in fear of falling and avoidance of physical or social activities [20, 30]. Therefore all falls have the potential to influence wellbeing [20]. Repeating the analyses for ‘falls requiring medical attention’ (S2 Fig) showed similar trends as for all falls (Fig 1). However, data were available only for the Irish and Australian cohorts, and for the 2006–2010 assessment in the British cohort. On average, a third of all reported falls (32%) required medical attention,which is similar to the proportion reported for older adults [14, 30]. Moreover, the current data show a marked increase in prevalence from 8.7% at age 40–45 to 19.1% at age 45–50 in Irish women (Fig 1). The timing of this increase in falls coincides with the onset of menopause [31], decline in balance performance [11] and increase in prevalence of syncope and vertigo [12, 13]. These collective findings warrant further research to inform preventive strategies in middle-aged adults. A better understanding of the factors that drive this increase in fall risk in middle-age may be the key to effective preventive interventions earlier in life with potential sustained benefits into older age.

Supporting information

Acknowledgments

We are grateful to NSHD study members for their continuing support. We are grateful to all of the LASA and TILDA respondents for participating in the study. Some of the research on which this article is based was conducted as part of the Australian Longitudinal Study on Women’s Health by The University of Newcastle and The University of Queensland. We are grateful to the women who provided the survey data.