Skip navigation

'Stops walking when talking'

One of Bandolier's colleagues bemoaned the difficulty of deciding for which patient to prescribe hip protectors to try and prevent fractures - identified as useful for injury prevention [1]. The key prescribing point turned out to be a judgement of whether the patient was likely to use them or not.

Selecting patients most likely to fall would be another important factor. A simple test comes from Sweden [2].

The observation

This was that some frail elderly people stopped walking when starting a conversation. Presumably this is because the attention needed to hold a conversation made a demand, and that the 'resources' available were insufficient to do two things at once.

The study

Residents (mean age 80 years; 72% women) in sheltered accommodation in Umeå who were able to walk with or without aids and able to follow simple instructions were included. Some had dementia, others a previous stroke, and some were depressed. They were observed by physiotherapists who noted whether they stopped walking when a conversation started. Falls during a six-month follow up were noted.

The results

There were 58 patients, of whom 12 stopped walking when talking. In the next six months 10 of these had at least one fall. Of the 46 patients who kept walking when talking, 11 had a fall in the next six months.

The overall prevalence of falls in this population was 21 of 58 patients, or 36%. The likelihood ratio of a positive test (stopped walking when talking) was 10, so that the post-test probability of a fall over the next six months with a positive test was 90%. With a negative test (carried on walking when talking), the post-test probability of a fall over the next six months was 24%.


Of course this is only one small observational study in a group of frail elderly people at high risk of falls. But it is an exemplar of a simple, zero-cost observation that can identify people of high risk of fall, for whom special action could be taken to reduce that risk. Hip fracture is devastating ( Bandolier 49 ): avoiding it is a good thing.


  1. R Lyons et al. Injury Prevention. Health Evidence Bulletins Wales, September 1998.
  2. L Lundin-Olsson, L Nyberg, Y Gustafson. 'Stops walking when talking' as a predictor of falls in elderly people. Lancet 1997 349: 617.

previous or next story in this issue