Tools for assessing fall risk in the elderly: a systematic review and meta-analysis

Seong-Hi Park

Methods: conducted according to the Cochrane guidelines

 

Participants: age > 60, total of 9,743. Mean age in the 70 was most common (32 studies, followed by mean age in the 80s (7 studies). 3 studies had a mean age in the 60s.

Selection: 33 articles of the initial 2571 were included in the study

• 31/33 (93.9%) had a low risk of bias in all domains. And items, all studies had a low risk of bias in the patient selection
• Seven studies, which accounted for 4,067 patients, were done in acute care setting and most patients were admitted to a geriatric or rehabilitation ward. 3 studies (364 patients) were conducted in long term care facilities and 23 studies (5312 participants) were conducted in the community. It included elderly people living at home and in retirement communities.

Methodology: In most studies >/= to one wall was considered a faller, in 3 studies, they needed greater or equal to 2 falls.

• Typical follow up was 6 and 12 months. 2-18 months was typical of the studies conducted in the community.
• 26 different fall risk assessment tools were used throughout the articles.
  • Acute care setting: Hendrich II Fall Risk Model, STRATIFY, and Timed Up and Go (TUG) test
  • Long term care setting: mobility interaction fall chart, Downton fall risk index
  • Community dwelling: 23 different assessment tools were used. Assessment tools used in 2 or more studies: BBS, TUG test, Downton Fall Risk Index, and Tinetti Balance scale, while the remaining 20 assessment tools were each used in a single study.

 

Different types of tests assess different things, it is important to pick a test based on the setting and what you’re trying to assess.

• The Hendrich II fall risk model and STRATIFY is designed to be done by medical staff and involves matters such as conscious state, urinary function, drug taking.
• The BBS and TUG tests assess balance ability in everyday activities, such as walking up and down stairs and speed.

 

Conclusions:

BBS showed both pooled sensitivity and pooled specificity>0.7, but inter-study heterogeneity in sensitivity was high. The Downton Fall Risk Index was a relatively stable tool with a moderate level of inter-study heterogeneity in both sensitivity and specificity and had a pooled sensitivity of 0.84. However, the pooled specificity was too low (0.26). The pooled sensitivity and specificity of the Hendrich II Fall Risk Model were acceptable, at 0.7 and 0.6, respectively, but inter-study heterogeneity was high, particularly in specificity. The pooled sensitivity of the MIF chart was low (0.53) and inter-study heterogeneity was high. STRAT- IFY had the highest pooled sensitivity (0.89), and a pooled specificity at a level similar to that of Hendrich II Fall Risk Model. In both the TUG test and Tinetti Balance scale, the pooled sensitivity was approximately 0.7 with no inter- study heterogeneity, whereas the pooled specificity was low (0.5) and inter-study heterogeneity was high. Regarding the assessment tools used each in a single study, the sensitivity was >0.7 for the majority of the assessment tools, while it was <0.5 for quite a few assessment tools. Overall, both pooled sensitivity and pooled specificity were <0.6 and inter-study heterogeneity was high in both measures.

Predictive validity if fall assessment tools currently used for elderly is not sufficient. Evidence shows that the use of a large variety of tools does not predict elderly fallers with sufficient accuracy. With a focus on the primary goal of each selective tool, they recommend the Downton fall risk index, Hendrich II fall risk model, STRATIFY and TUG test which all show a pooled sensitivity of >0.7 and low inter-study heterogeneity.

Final conclusions: two assessment tools in combination will maximize the characteristics and predictability of each test. STRATIFY or the Hendrich II Fall Risk Model, which has a high sensitivity, and BBS or MIF chart which has a stable specificity. This approach will increase the predictability of the risk of fall regardless of the setting.

Limitations: they used prospective studies, and did not cover all tools used in practice