Automated detection of near falls: algorithm development and preliminary results

BMC Research Notes

Table 2 Sensitivity and specificity for detecting near falls using multiple parameters.

Parameters*	State	Sensitivity (%)	Specificity (%)	Detection (%)**
Max-V, Maxp2pdiff-V	and	85.71	90.12	17.37
[Max-V, Maxp2p-V, Maxp2pdiff-V	and	85.71	90.12	17.37
Maxp2p-V, Maxp2pdiff-V	and	85.71	89.37	17.81
[Max-V, Maxp2p-V	and	90.48	84.13	18.51
Max-M-L, Maxp2pdiff-V	or	90.48	81.74	20.59
Maxp2p-M-L, Maxp2pdiff-V	or	95.24	79.49	21.06
Max-V, Maxdiff-V, Maxp2pdiff-V	and	80.95	90.42	21.32
Maxdiff-V, Maxp2p-V, Maxp2pdiff-V	and	80.95	89.82	21.60
Maxp2pdiff-V, Maxp2pDiff-A-P	or	95.24	78.74	21.79
Maxdiff-A-P, Maxp2pdiff-V	or	95.24	78.59	21.93

The sensitivity and specificity values obtained for detecting near falls (n = 21) as compared to non-near falls (n = 668) using multi-parameter combinations Non-near falls included regular gait intervals combined with the irregular, non-near falls intervals (e.g., kicks, stepovers, stops).
*The parameter combinations for the algorithm's detection criterion included single parameters, and multiple parameters. For the multiple parameter combinations, we checked the case of passing the detection criterion for all parameters (state "and"), versus passing the detection criterion for at least one parameter (state "or").
Max = maximum acceleration amplitude; Maxdiff = maximum acceleration derivative; Maxp2p = maximum peak-to-peak acceleration amplitude; Maxp2pdiff = maximum peak-to-peak acceleration derivative; Std = standard deviation; V: vertical; M-L: medio-lateral; A-P: anterior-posterior.
**The detection% is the distance from the ideal roc curve (the lower, the better). For brevity, the results are shown for only the best 10 combinations.

ISSN: 1756-0500