Postoperative nausea and vomiting (PONV) is a relatively common side effect of general anaesthesia. It occurs in approximately one-third of untreated patients in a general surgical population [1–3]. In recent years, risk factors as well as prophylactic and therapeutic options for PONV have been identified and guidelines have been proposed [3–7]. Some studies [8–10] have shown that risk score-based strategies to prevent and treat PONV may potentially be effective and efficacious.
In a previous study , we demonstrated that decision support was effective in improving both identification and correct prescription of PONV prophylaxis in the preoperative screening clinic. However, that study was neither designed to test the execution of the decision to prescribe PONV prophylaxis nor was it set in the operating theatre. The current study investigates whether the prescription of PONV prophylaxis is correctly executed in the operating theatre. Further, a set of automated reminders was tested for their ability to improve the execution of a previously prescribed PONV prophylaxis order. As timing may be important in achieving the optimal effect of PONV prophylaxis, we also investigated the effect of a patient-specific reminder (decision support) on the timing of the PONV prophylaxis medication .
This study was conducted in the operating theatre complex of our regional teaching hospital in Amsterdam, The Netherlands. Patients at high risk for PONV should receive PONV prophylaxis consisting of dexamethasone 8 mg intravenously (i.v.) upon induction and granisetron 1 mg i.v. shortly before awakening. For a more comprehensive description of the hospital setting and the departmental guidelines on PONV in our hospital, we refer to a previous study  on decision support in our centre.
To study the effect of decision support on the actual administration rate of prescribed PONV prophylaxis, we developed a patient-specific decision support system (DSS). This system consisted of two different reminders that suggested administering PONV prophylaxis in case of impending nonadherence to the departmental guideline. The reminders were active both in the operating theatre (operating theatre reminder) and in the recovery room (recovery reminder). For a more comprehensive explanation of such reminders, we refer to a previous study  on decision support in our centre.
All consecutive adult patients presenting for noncardiac surgery who had been evaluated in the preoperative screening clinic were included in our study. The study was divided into three study periods, each lasting for 6–12 weeks, and was set up according to an off–on–off design. During the first 6-week study period (control period), PONV prophylaxis was managed according to the study version of the protocol as it had been in place before this study. In the second period, which lasted for 12 weeks (decision support period), the automated reminders as described above were activated. After deactivating the automated reminders, PONV prophylaxis in the postdecision support period was managed as it was in the control period. After all study periods were completed, the relevant data were extracted from the database by prewritten scripts.
Our primary outcome was guideline adherence. We selected all patients who were prescribed PONV prophylaxis at the preoperative screening clinic and calculated the percentage who actually received either dexamethasone or granisetron or both medications during anaesthesia in each study period. In addition, we evaluated the effect of decision support on the behaviour of each individual anaesthesiologist. Therefore, we determined the anaesthesiologist who had been responsible for each general anaesthetic administered to a high-risk patient and calculated the percentage of high-risk patients who actually received the prophylactic medication in the three study periods per anaesthesiologist.
As a secondary outcome, we analysed the timing of the prophylactic medication. For this analysis, we categorized the time at which dexamethasone and granisetron had been administered. Four categories were made for each parameter: before the first reminder, after the first and before the second reminder, after the second reminder and before the third reminder and after the third reminder. The change in timing of the prophylactic medication was analysed per study period. Further, we dichotomized the medication timing into correct and incorrect. We arbitrarily considered timing of a dexamethasone dose correct if it had been given in the operating theatre (before arrival in the recovery room) and within 30 min from induction of general anaesthesia. Timing of granisetron was arbitrarily considered correct when granisetron was given either before or within 15 min after arrival in the recovery room. The latter was mainly intended to distinguish granisetron as PONV prophylaxis from granisetron used as a PONV therapeutic.
For all percentages, confidence intervals were calculated, and the significance of the differences between the periods was established with a chi-squared test.
Between November 2005 and June 2006, a total of 5652 consecutive patients were included in the study: 1727, 2594 and 1331 patients in the control period, decision support and postdecision support periods, respectively.
Patients' characteristics and incidence of risk factors for the patients included in the three study periods are shown in Table 1. Generally, patients were equally distributed across the study groups. However, compared with the control period, PONV prophylaxis was prescribed more often in the decision support and the postdecision support periods.
In Table 2, prophylaxis administration is presented for patients who did have PONV prophylaxis prescribed and received general anaesthesia. Dexamethasone was given in 46% of these patients in the control period. In the decision support period, this rate increased significantly to 95%, and, after deactivating the automated reminders, it decreased to 47% in the postdecision support period (P < 0.001). For granisetron, these percentages were 53, 81 and 51%, respectively (P < 0.001).
In the control period, only 39% of patients eligible for PONV prophylaxis received both medications within the time frame we considered acceptable: dexamethasone in the operating room within 30 min after induction and granisetron before – or within – 15 min after arrival in the recovery room. This increased to 79% in the decision support group and decreased again to 41% in the postdecision support group (P < 0.001).
When analysing the data for each individual anaesthesiologist, we found a similar pattern for all users. As shown in Fig. 1a, all anaesthesiologists were more compliant with administering dexamethasone during the decision support period and less compliant during the control and postdecision support periods. As shown in Fig. 1b, this pattern is the same for granisetron. The single anaesthesiologist scoring 0% adherence in the postdecision support period attended to only a single case of general anaesthesia during that period; therefore, we consider this line an artefact.
Figure 2a demonstrates that, apart from an increase in dexamethasone administration, the timing changed as well. Significantly, dexamethasone was given more often within the first 30 min after induction. A similar pattern is shown in Fig. 2b for granisetron: a significant increase in administration before – and within 15 min – after arrival in the recovery room.
As demonstrated in Table 3, prophylactic medication was in some cases given to patients who had no prescription for PONV prophylaxis. This occurred in only 1–2% of patients. This rate remained constant throughout the study periods. Moreover, 39% of patients receiving prophylaxis without prescription were actually at high risk for PONV and should have had a prescription.
The present study showed that guideline adherence in administering PONV prophylaxis in our hospital was very poor, and that automated decision support was able to improve guideline adherence significantly. Moreover, we demonstrated a significant improvement in timing of the PONV prophylaxis administration; both dexamethasone and granisetron were more frequently given within a timeframe considered acceptable and even before the first reminder.
Previous studies [11,13,14] on decision support have focused on either the prescription or the execution of an intervention. In the current study, we focused on the execution of a previously prescribed order by co-workers. The results showed that compliance regarding the actual administration of previously prescribed medication was poor. This suggests that the use of decision support only for the prescription of an intervention, such as the one we studied in a previous report, may not be an effective strategy as, even if prescription is good, the execution lags behind . Instead, decision support designed to improve the actual administration of PONV prophylaxis seems to be more effective than using decision support only to improve the correct prescription of PONV prophylaxis.
As in our previous study, we found that, after discontinuation of the decision support guideline, adherence returned to control levels. The absence of any form of learning from the system is in accordance not only with our previous study but also with the only other study that reported on a control group after the intervention; both describe a marked decrease in adherence after discontinuing decision support [11,15].
Future research could include a study on the optimal (combination of) reminders, so as to attain maximum effect with a minimum number of reminders. Applied to PONV prophylaxis, it may be most effective to use decision support for the identification of high-risk patients in the operating room instead of at the preoperative screening clinic and support the actual administration of PONV prophylaxis medication at the same time. A reminder at the preoperative screening clinic could then be omitted, saving a potentially redundant pop-up message.
The importance of timing and form of decision support may be illustrated by the residual nonadherence regarding granisetron. On average, 95% of patients who were eligible for dexamethasone received it within an acceptable timeframe during the decision support period. For granisetron, the adherence was lower; approximately 81% of eligible patients received it within the accepted timeframe. Although this is a marked improvement over the control period, guideline adherence of approximately 80% is clearly suboptimal. This may be related to the timing of the reminder. Upon arrival in the recovery room, many different tasks have to be performed simultaneously. This may not be the optimal time to remind users of yet another action. The importance of timing was also demonstrated in a different setting by Vigoda et al. . These authors showed that an electronic mail reminder the night before elective, noncardiac surgery had no effect on anaesthesiologist behaviour, as guideline adherence regarding perioperative β-blockade in eligible patients remained unchanged. Both studies demonstrate that, for decision support to work, the user needs to be able to process the information presented to him or her. Therefore, this information needs to be presented to the user as close as feasible to the time a decision is made.
One might argue that we did not include PONV itself as an outcome measure. However, our study was neither powered nor designed to provide evidence for or against the hypothesis that PONV prophylaxis works. Instead, our study was designed to investigate the effect of decision support on the execution of previously prescribed orders. Moreover, the effectiveness of PONV prophylaxis medication used in this study has already been demonstrated by several excellent studies [1,2,17].
Decision support significantly improved administration of PONV prophylaxis medication. A complete disappearance of this effect was shown after discontinuing the reminders. In addition, timeliness of administration was also significantly improved by decision support. The effect on adherence was stronger regarding dexamethasone stressing the importance of timing of decision support.
1 Kovac AL. Prevention and treatment of postoperative nausea and vomiting. Drugs 2000; 59:213–243.
2 Apfel CC, Korttila K, Abdalla M, et al
. A factorial trial of six interventions for the prevention of postoperative nausea and vomiting. N Engl J Med 2004; 350:2441–2451.
3 Gan TJ, Meyer TA, Apfel CC, et al
. Society for ambulatory anesthesia guidelines for the management of postoperative nausea and vomiting. Anesth Analg 2007; 105:1615–1628.
4 Apfel CC, Laara E, Koivuranta M, et al
. A simplified risk score for predicting postoperative nausea and vomiting: conclusions from cross-validations between two centers. Anesthesiology 1999; 91:693–700.
5 Sinclair DR, Chung F, Mezei G. Can postoperative nausea and vomiting be predicted? Anesthesiology 1999; 91:109–118.
6 Koivuranta M, Laara E, Snare L, Alahuhta S. A survey of postoperative nausea and vomiting. Anaesthesia 1997; 52:443–449.
7 HaR AS, El-Moalem HE, Gan TJ. The efficacy of the 5-HT3 receptor antagonists combined with droperidol for PONV prophylaxis is similar to their combination with dexamethasone. A meta-analysis of randomized controlled trials. Can J Anaesth 2004; 51:311–319.
8 Pierre S, Corno G, Benais H, Apfel CC. A risk score-dependent antiemetic approach effectively reduces postoperative nausea and vomiting: a continuous quality improvement
initiative. Can J Anaesth 2004; 51:320–325.
9 Biedler A, Wermelt J, Kunitz O, et al
. A risk adapted approach reduces the overall institutional incidence of postoperative nausea and vomiting. Can J Anaesth 2004; 51:13–19.
10 Rusch D, Eberhart L, Biedler A, et al
. Prospective application of a simplified risk score to prevent postoperative nausea and vomiting. Can J Anaesth 2005; 52:478–484.
11 Kooij FO, Klok T, Hollmann MW, Kal JE. Decision support increases guideline adherence for prescribing PONV prophylaxis. Anesth Analg 2008; 106:893–898.
12 Wang JJ, Ho ST, Tzeng JI, Tang CS. The effect of timing of dexamethasone administration on its efficacy as a prophylactic antiemetic for postoperative nausea and vomiting. Anesth Analg 2000; 91:136–139.
13 Dexter PR, Perkins S, Overhage JM, et al
. A computerized reminder system to increase the use of preventive care for hospitalized patients. N Engl J Med 2001; 345:965–970.
14 O'Reilly M, Talsma A, VanRiper S, et al
. An anesthesia information system designed to provide physician-specific feedback improves timely administration of prophylactic antibiotics. Anesth Analg 2006; 103:908–912.
15 Bosman RJ, Rood E, Oudemans-van Straaten HM, et al
. Intensive care information system reduces documentation time of the nurses after cardiothoracic surgery. Intensive Care Med 2003; 29:83–90.
16 Vigoda MM, Elf RB, Wu EY, et al
. Intraoperative beta blocker use doesn't change with electronic reminder of patients' cardiac status. Anesthesiology 2006; 105:A-1377.
17 Carlisle JB, Stevenson CA. Drugs for preventing postoperative nausea and vomiting. Cochrane Database Syst Rev 2006; 3:CD004125.