Exploratory analysis of real personal emergency response call conversations: considerations for personal emergency response spoken dialogue systems.
Victoria Young, Elizabeth Rochon, Alex Mihailidis
PMID 27842598WHAT IT FOUND
In real emergency calls, older adults spoke more slowly, used shorter turns and more one-word answers than care providers.
High-risk calls had fewer words, fewer turns and shorter response times. This describes speech patterns, not tested interventions.
Key findings
01Of the 84 transcribed calls, 72 were analysed, and 50 calls were made by older adult callers while 22 were made by care providers.
02Older adult callers had shorter average turns (5.81 words) than care provider callers (9.13 words), and they used more one-word utterances (3.06) than care provider callers (1.64).
03High-risk calls had fewer words (49.26 mean) and shorter response times (40.64 seconds) than medium-risk calls (82.14 mean words and 69.59 seconds).
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study analysed existing call recordings, not people receiving a therapy or a tested emergency system, so it cannot show that any intervention improves emergency response. The sample was small and unbalanced, with 72 calls analysed and 22 calls made by care providers. All calls came from one private PERS provider in Toronto, so other providers with different protocols may not match these results. Caller details were limited and partly inferred from the audio; age was determined for 53 of 84 calls, and gender was inferred. The study excluded false alarms and follow-up calls, and it did not analyse the meaning of the words, speech intelligibility, prosody, or separate call taker effects. Transcription may have varied because human error and poor audio made some words hard to hear. Classification accuracy came from this dataset, and checking each call against the other calls lowered the results, especially for care provider callers.
The easy way to misread this
Do not conclude that a speech-based emergency system has been shown to improve help-seeking. The work analysed 72 existing calls from one provider and did not test a spoken dialogue system with older adult users in real emergencies.