Mean arterial pressures bracketing prolonged monitoring interruptions have negligible systematic differences from...

Comparing intraoperative hemodynamic data from anesthesia information management systems (AIMS) among hospitals involves handling missing or edited values. There routinely are periods >5 minutes ("gaps") in recorded blood pressure in AIMS records. Previous studies showed the importance of monitoring the incidences of unexplained gaps, because providers interpolate when charting vital signs in gaps. We studied whether ignoring missing vital signs during gaps systematically biases monitoring results.

Mean arterial pressures (MAPs) from noninvasive blood pressure cuffs or arterial lines were studied from 14,860 noncardiac cases at a hospital. The hospital's AIMS does not permit manual editing or deleting of blood pressure data. Matched intervals from the same case were found for 15,338 of the 22,046 MAP gaps >5 minutes. The difference was taken between the first MAP at the end of a gap and the MAP at the start of the gap. The difference was also taken between the MAP at the end minus the beginning of the matched interval.

The mean difference between a gap's MAP difference and matched interval's MAP difference was significantly larger than zero (P = 0.0003), but hemodynamically negligibly (0.9 mm Hg, 95% confidence interval [CI] 0.4-1.4 mm Hg). There was no association between the differences of the MAP differences and durations of gaps (P = 0.47, Spearman r = 0.00, 95% CI -0.01 to 0.02). The pairwise difference was taken of the absolute MAP difference for the gap minus the absolute MAP difference for the matched pair. This mean difference was significantly larger than zero (4.4 mm Hg, 95% CI 4.0-4.7 mm Hg, P < 0.00001). Gaps' MAP absolute differences more often exceeded 20 mm Hg than did matched intervals' MAP differences (odds ratio 1.85, 95% CI 1.75-1.96, P < 0.00001).

Our results show that when comparing hospitals using mean MAP data from hundreds of AIMS cases, statistical issues related to gaps are of minor importance. The more important issues when comparing hospitals are the incidences of gaps themselves and/or the manual editing of automatically recorded vital signs. Nevertheless, when quantifying hemodynamic variability (e.g., brief periods with rapid changes in MAP), gaps cannot be ignored. Furthermore, none of our results apply to individual patients (i.e., it is best not to have gaps in blood pressure during anesthesia).