Preliminary Simple Screening Tool for Identifying Patients at Risk for severe Postcesarean Section Pain - Actual versus Prediction: A Quality Assurance Study
Abstract Number: 59
Abstract Type: Original Research
Acute pain after delivery has been shown to vary significantly despite similar analgesic regimen. Besides having acute pain itself, severity of acute pain can associate with persistent pain & depression after delivery. If those at risk for severe pain can be identified, early aggressive management may prevent the short & long term negative consequences. Our institution has recently developed a preliminary predictive model from a 3-question screening test to identify those at risk for having severe post C/S evoked pain (pain with movement). We have recently incorporated the 3-question screening in our preop clinic. As part of quality assurance, this study prospectively evaluated the effectiveness of this preliminary predictive model in identifying patients at risk of having severe evoked pain (top 20%) after C/S.
After IRB approval, parturients scheduled for elective C/S with spinal anesthesia & spinal morphine were routinely asked a 3-question screening test in the preop clinic. The 3 questions asked the patient to rate her anxiety to surgery(VAS 0-100), level of anticipated postop pain(VAS 0-100) & estimated postop pain medication needed. The scores were entered into a handheld palm pilot to generate a predicted postop evoked pain score & whether the score was within top 20%. The ongoing development of the predictive software is being reported elsewhere. The software utilized a general linear model with 2 & 3 way interactions of the answers to the questions and evoked pain from previous data, through which a regression equation was developed to calculate the predictive evoked pain scores and classification of the top 20th percentile. Participants were interviewed 24 hours after surgery to assess with visual analog pain scale (0-100) their resting pain (in bed rest) & evoked pain (patients transferring themselves from bed rest to sitting up right). Effectiveness of the preliminary predictive model was assessed prospectively for correlation, sensitivity & specificity on predicting evoked pain scores to be within top 20th percentile.
64 evaluable parturients of the 200 planned completed the assessments & underwent an elective C/S with spinal anesthesia & spinal morphine. Average actual evoked pain score at 24 hours reported by patients was 42 (SD=26, range 0-85), while the average score predicted by the model was 59 (SD 28, range 0-100). Pearson correlation was 0.36 (P<0.008) between predicted and actual scores with sensitivity (80%) and specificity (58%) in identifying those within top 20th percentile.
These results suggest that the simple 3-question screening was promising at identifying patient at risk of severe evoked pain at 24hr after C/S. However, further refinement of the predictive regression equation is needed from a larger dataset than the preliminary one we used in order to improve the specificity in identifying patient at risk for severe acute evoked pain after C/S.