Table 2. Measurements to Assess the Completeness of the Documented Clinical Observations Among Randomly Selected Patientsa.
| Mean No. of Measurements (SD)b | Between-Group Mean Difference (95% CI)c |
P Value | Adjusted P Valued |
||||
|---|---|---|---|---|---|---|---|
| BedsidePEWS | Usual Care | ||||||
| Baseline Period |
Intervention Period |
Baseline Period |
Intervention Period |
||||
| Total No. of patient assessments | 1270 | 2588 | 1419 | 2832 | |||
| Heart rate | 6.97 (5.1) | 7.40 (4.8) | 6.48 (4.9) | 6.45 (4.7) | 0.58 (−0.11 to 1.26) | .10 | .30 |
| Respiratory rate | 6.99 (5.5) | 7.38 (4.9) | 5.59 (4.2) | 5.53 (4.0) | 0.85 (0.02 to 1.68) | .05 | .18 |
| Systolic blood pressure | 4.09 (4.2) | 5.05 (3.5) | 3.58 (4.5) | 3.59 (3.6) | 1.12 (0.59 to 1.65) | <.001 | <.001 |
| Transcutaneous oxygen saturation | 6.64 (5.3) | 7.30 (5.0) | 5.29 (5.4) | 5.21 (5.3) | 1.06 (0.27 to 1.85) | .009 | .04 |
| Respiratory effort | 1.99 (3.6) | 7.16 (4.7) | 2.80 (3.3) | 3.00 (3.6) | 4.67 (3.23 to 6.12) | <.001 | <.001 |
| Capillary refill | 1.96 (3.5) | 6.66 (4.4) | 1.48 (3.0) | 1.66 (3.1) | 4.65 (3.49 to 5.80) | <.001 | <.001 |
| Oxygen therapy | 6.80 (5.2) | 7.30 (4.8) | 5.83 (5.6) | 6.09 (5.7) | 0.37 (−0.71 to 1.46) | .50 | >.99 |
| Observation sets with ≥5 vital signs, No. (%) | 960 (75.6) | 2563 (99.0) | 883 (62.2) | 1725 (60.9) | 38.1 (20.8 to 55.4)e | <.001 | <.001 |
Abbreviation: BedsidePEWS, Bedside Paediatric Early Warning System.
At each hospital during each study week, 5 patients admitted to a ward for at least 24 hours were randomly selected for documentation review. Study coordinators abstracted the number of documented measurements during the 24 hours before assessment for each of the 7 clinical observations in this Table and abstracted the number of clinical observation types included in the last set of clinical observations. A set of clinical observations was regarded as those that were documented as being from the same time. There were 8190 (21 sites × 5 patients per week × [26 + 52] weeks) case report forms anticipated and 8109 (99%) were obtained. The missing data may reflect the challenges of finding 5 randomly selected patients admitted for more than 24 hours at a smaller hospital, and in some cases, the fifth patient may have been unintentionally missed.
Unless otherwise indicated.
Calculated from a generalized estimating equation linear regression model that adjusted for clustering by hospital using an exchangeable correlation structure. The BedsidePEWS group and the hospital baseline mean value were used as predictors.
Adjustment for multiple comparisons was performed using the method of Holm.
Calculated using a generalized estimating equation regression model with a binomial variance and a linear link function. The BedsidePEWS group and the hospital baseline proportion were the only covariates.