Falls Screening Mobile Application (FALLSA©): its discriminative ability, predictive validity and effectiveness on falls prevention behaviour and physical function among older adults

Abstract

Early fall risk assessment is recommended as one of the falls prevention strategies in older adults. While FallSA© has been shown to be accepted and reliable, its discriminative ability, predictive validity and effects of use is not known. Therefore, the objective of this study was to examine the discriminative ability and predictive validity of FallSA© and determine its effects of use on falls prevention knowledge, behavior, balance confidence, falls risk and physical performance besides upgrading FallSA© to FallSA 2.0. Phase 1 was a prospective study, involving a six months follow-up of 610 participants from four states of Malaysia, for incidence of falls using falls diaries, FallSA© and monthly phone calls. In phase 2, a six months randomized control study, 59 participants were involved. Participants were randomly allocated into experimental (n=30, 66.2±5.3 years) and control (n=29, 69.2±5.0 years) groups. The experimental group received falls prevention education and FallSA© installation for its use, while the control group received only falls prevention education. Falls prevention awareness and practice of falls prevention behavior (FaB), fall risk knowledge (FRAQ), balance confidence (ABC-16), physical function (SPPB) and physical activity (MBPAQ) of the participants were measured. In phase 3, the limitations of FallSA© was identified through observation, field work and research team discussions, inclusion of secondary data and FallSA© was upgraded to FallSA 2.0. The study results showed that FallSA© had a moderate to strong discriminative ability in classifying fallers and non-fallers (n=610, mean age=71.78±4.7 years). FallSA© demonstrated a predictive validity with pooled sensitivity of 82% and specificity of 64%, positive likehood ratio of 2.27 and area under the curve of 0.80 at cut-off score of >5 among 454 participants (mean age=71.79±4.9 years) who were successfully followed-up. The outcomes of the empowerment of using FallSA© showed no significant time x group interaction effects on all outcome measures (p>0.05). However, there was significant time effect in improving FaB score (p=0.001, η2p= 0.17, power= 0.92), reducing FallSA© risk (p=0.03, η2p= 0.08, power= 0.59), ABC-16 (p=0.009, η2p= 0.11, power= 0.75) and SPPB (p=0.008, η2p= 0.12, power= 0.77) scores. A larger mean change was found in FaB, ABC-16, FallSA© risk score and MBPAQ in the experimental compared to the control group. The limitations of FallSA© were addressed in FallSA 2.0 via additional and upgraded features and design that included mainly falls prevention educational package, sit to stand test with normative values and improved back end application for monitoring by health professionals. In conclusion, FallSA© is a valid and reliable self screening falls risk assessment software application. Its’ use has the potential to improve awareness and prevention behavior in regard to falls prevention strategies among older adults. Further studies are required to explore the wider use of FallSA 2.0, and test its cost effectiveness for falls prevention in older persons