Vector Autoregressive Models and Granger Causality in Time Series Analysis in Nursing Research: Dynamic Changes Among Vital Signs Prior to Cardiorespiratory Instability Events as an Example.
Eliezer Bose, Marilyn Hravnak, Susan M Sereika
PMID 27977564WHAT IT FOUND
In 20 monitored patients, oxygen saturation was the first vital sign to cross instability thresholds in 60% of cases, but breathing-rate changes more often preceded heart-rate changes than the reverse.
This is a method example, not treatment proof.
Key findings
01In the 20 cases, oxygen saturation was the first vital sign to cross an instability threshold in 60% of cases, followed by respiratory rate in 30% and heart rate in 10%.
02Respiratory rate changes preceded heart rate changes more often than heart rate changes preceded respiratory rate changes (21% vs 15%), and respiratory rate changes preceded oxygen saturation changes more often than oxygen saturation changes preceded respiratory rate changes (15% vs 9%).
03The models were stable in 18 of 20 cases, while 3 cases had no causal structure and 2 cases had unstable models.
STILL TO COME
How it was doneWhat they foundWhat it means for RNs
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 analysis used a convenience sample of 20 patients from one 24-bed step-down unit, so it is a demonstration rather than a broad test in a general patient population. Not all patients had complete 6-hour vital-sign streams before the first instability event; some had only 2 or 4 hours, some had only two of the three vital signs, and some had at least 3 hours of intermittent monitoring. The authors used interpolation or imputation to make the data streams continuous. Two models could not be stabilized, and three cases had no causal structure, so the method did not produce a usable model for every case. The Granger causality method described here is based on improved prediction from lagged values, not proof that one vital sign physiologically causes another.
The easy way to misread this
Do not read the vital-sign ordering as a validated instability threshold. The study analyzed 20 patients from one 24-bed step-down unit, filled some missing data with interpolation or imputation, and used a forecasting method rather than a treatment test.