Activities of Daily Living Assessment

gardening

Usually, the assessment of frailty is often observational and self-reported. As subjective measures are prone to bias, an objective frailty screening is needed. In particular, lower physical function represents a key point in frailty assessment and is generally recorded by the use of questionnaires. But using questionnaires may result in imprecise data, as people may have difficulties to recognize low-intensity activities of daily living (ADLs).

In contrast, wearable sensors are an objective way to distinguish between different activity levels and may detect important frailty related changes in physical activity patterns. In addition, wearable sensors are portable and cost-effective, enabling frailty screening at home as well as in the community in a continuous way.

Frail individuals seem to score less in ADLs, compared to pre-frail elderly. Therefore, ADL monitoring appears to be an important way to detect the onset of frailty. Accelerometers have been shown to effectively recognize characteristic movement patterns (e.g., running, walking, lying). However, little is known about the analysis of upper extremity-based ADLs like tea making and gardening.

Experiment

As part of the FRAIL project, we did the following experiment. We asked 17 participants to execute the tea making and the gardening task. The tasks were executed one after another and in random order. Among our participants, 35 % were frail, 47% were prefrail, and 18% were robust.

The subjects received the following introduction for the tea making:

tea making

‘Can you prepare a cup of tea with one sugar cube, standing behind the table? Please do the task in a natural way, as you would do that at home and in the speed which is appropriate to you.’

And the following introduction for the gardening task:

‘Can you replant this plant into the pot and water it, standing behind the table? Please do the task in a natural way, as you would do that at home and in the speed and a way which is appropriate to you.’

gardening

 

Participants wore the smartwatch (HUAWEI ll) on their dominant hand. We installed our app’ SensorRecorder’ on the smartwatch before the experiment. The set-up was kept constant. The participant was standing behind a table and all equipment was placed in a similar location.

Results

figure 1

Figure 1

Figure 1 shows the parameters duration, median activity, variance of acceleration, and smoothness of the ADL “Tea Making” in relation to the frailty status (0 = robust, 1-2 prefrail, 4-5 frail). The median activity and variance show the tendency of a negative correlation (r = – 0.49 and r = – 0.44), whereas duration shows a positive trend with r = 0.43. Smoothness demonstrates a small negative coefficient. None of the parameters reached a significance level of p < 0.05.

Figure 2

Figure 2

Figure 2 shows the parameter duration, median activity, variance of acceleration, and smoothness of the ADL’ Gardening’ in relation to the frailty status (0 = robust, 1-2 prefrail, 4-5 frail). The median activity, smoothness and variance show the tendency of a negative correlation (r = – 0.47, r = – 0.49 and r = – 0.37). Duration shows a small positive correlation. None of the parameters reached a significance level of p < 0.05.

In this experiment, we used activity tracking via the smartwatch to collect ADL data. The results show that frail people may need more time to perform instrumental ADLs. The lack of correlation between duration and frailty status in the ‘Gardening,’ however, can be traced back to possible differences in performance demands.

In addition, frail people have a tendency towards reduced activity and variance of acceleration compared to healthy older people during task execution. Differences between the ADLs can be induced by different dynamics of the tasks. Hence, more complex tasks seem to be more prone to deficits in dynamic performance.

Leave a Comment