Falls recall—limitations of the most used inclusion criteria

Firstly, several studies indicated that self-reported recalls of falls in older adults leads to underreporting and even minor injuries are often not reported [6, 12] . Not recognizing (the severity of) a fall or not remembering a fall may attribute to the low uptake by older persons who are offered fall prevention programs [41–43], as the recognition of falls by the older persons themselves is crucial for the uptake of a fall prevention program. Another problem is that different perceptions and definitions of a fall exist between older persons, researchers, and health professionals [15, 19, 28]. The different use of language between older persons and health care staff in assessing falls causes barriers in the screening procedure. Furthermore, with regard to even interchangeable use of different terms, e.g., slips, trips, and falls used differently by health care providers, and the older person himself, makes it even harder to obtain reliable data and information. Therefore, in the AGS guidelines [1] and the German DEGAM guideline [7], the question “did you have a fall during the last 12 months?” is extended by a follow-up question “..have you had any fall including a slip or trip on which you lost your balance and landed on the floor or ground or lower level?” [19, 20, p.1619]. Ganz et al. [11] even recommend in their algorithm for screening older persons for fall risks to ask the participant or patient “whether they had noticed any problems with gait, balance, or mobility” [11, p.84]. Chu et al. [4] found a positive likelihood ratio (LR) for future recurrent falls of 2.0 (95% CI 1,7–2.4) by self-perceived and self-reported problems in mobility.In a study comparing health care staff and seniors in their definition and understanding of falls Zecevic [44] found that both groups tended to focus most on the consequences of falls thus neglecting and disregarding the first warning sign of a light fall without any consequences and prevent early detection of an “at risk” person and effective fall prevention strategies [44]. This study confirms the aforementioned differences in fall awareness between practitioners, researchers, and participants underlining the necessarily of communication in depth a definition of falls being used and asked for.

Mackenzie and colleagues [21] raise the point that in a fall prevention study, there may be a difference of fall recall between control and intervention group due to the fact that fall definition and importance of falls are targeted probably by the intervention but not to the control group. Dickens [8] discussed another important point for identification of at risk persons by bringing up the question “how does an older person define a fall for himself?” Our own experience in an earlier research study for fall prevention [10] did exactly support this point. Our participants used daily falls calendars with a definition of different type of falls on it. When asked in a discussion session what they would call a fall, they came up with totally different definitions and assumptions (Freiberger personal communications). In addition, it became obvious that in the German language, there were different wordings for falls (ein Sturz or simple Fallen in German), and most persons would use those two words related to the outcomes for as fall, e.g., easy fall or very heavy fall, may be with injuries or even use the words trips and slips for a light fall, which would be in contrast to our aforementioned definition of falls. The statements of our participants corresponded well with Zecevic's statement that most persons have an idea about a fall, but having to define it, fail to be able to do so [44]. An older person's concept of falls varies and is related to their state of function on the broad continuum from high active to frailty.

Secondly, falls can be reported retrospectively or prospectively. In general, a prospective design is favored by researchers with “daily fall calendar” considered the golden standard. Mackenzie and colleagues [21] however found that many intervention studies aiming at reducing falls obtain the fall variable not always with the golden standard “daily fall calendar” but by self-reported falls retrospectively. This leads to an uncertainty of the outcome measure “falls rate and number of falls” because retrospective recall of falls by older persons are not always precise [6, 9, 21, 23]. Fleming et al. [9] investigated recall of falls in a cohort of oldest old and found a striking similarity between the proportion of persons who retrospectively recalled a fall (58%) and prospectively recorded a fall (60%). However, the authors discuss the point that the percentage of fallers obtained retrospectively and prospectively might not necessarily be the same person having experienced a fall.In line with the above findings are our own findings (presented at the World Congress of Gerontology and Geriatrics 2009 in Paris). We investigated the reliability of falls recall in a group of 66 community-dwelling, active older persons (age range 70–90 years). Falls were monitored with a daily falls calendar handed in monthly. Participants were being contacted by telephone if the monthly calendar was not returned in due time or to obtain data regarding the circumstances of a fall. After 12 months, participants were reassessed and questioned for fall history of the past 12 months. Twenty-nine participants (43.9%) reported no falls in the daily falls calendar over the 12-month period, whereas 37 participants (56.9%) reported no falls when asked retrospectively. The difference was even more distinct in the reporting of two or more falls. Participants 37.8% of the participants recorded two or more falls in the daily calendar, but only 17.2% recalled having two or more falls. This demonstrated that even in participants of fall prevention program and a raised awareness for falls the event of a fall is still underreported.

Recent developments showed that computerized interactive response technology may be of additional value to prospectively monitoring falls in older persons. Wijlhuizen and colleagues found that their Telephone Inquiry System (TIS) is feasible and reliable for registering accidents and falls among community-dwelling elderly individuals [46].

Different research has been conducted to find the most reliable time frame for self-recorded falls in the past [6, 9, 12, 23]. Cummings demonstrated in his early study [6] that older persons tended to forget falls often; thus, showing an underreporting of falls history in his cohort. Interestingly, the time frame of 12 months for falls recall was more reliable than a time frame of 3 months [6]. In addition, in the study by Cumming et al. [6], participants tended to forget falls with minor injuries as well, and persons with minor cognitive limitations showed even less reliability in the recall of their falls history. A systematic review of the available data for recall of fall history and obtaining falls data [12] concluded that recall of any fall in the last 12 months is relatively specific with 91–95% but demonstrate less sensitivity with 80–89% compared to collected prospective falls data [12]. The author did conclude that the information available was not enough to give recommendations based on the inhomogeneity of study populations and falls definitions [12].