Advancing Strategies to Optimise the PerIopeRativE Management of PostOperative Nausea and Vomiting (ASPIRE-PONV) study: rationale and design for a sequential, repeated crossover trial

Abstract

Introduction

This project, in adult surgical patients, will evaluate whether the creation of a customised checklist, driven by a clinical decision support tool, is able to improve anaesthesia providers’ adherence to consensus guidelines and standardised practice recommendations for the prevention of postoperative nausea and vomiting (PONV).

Methods and analysis

The intervention will be evaluated using a sequential, repeated crossover design at the institutional level, with designated washout, control and intervention periods. The surgical case will serve as the unit of analysis. The primary outcome is adherence to appropriate PONV prophylaxis administration guidelines. Secondary outcomes include the incidence of PONV and length of stay in the postanaesthesia care unit (PACU).

Ethics and dissemination

This protocol and statistical analysis plan provide an outline of the study design, primary and secondary end points and analytic approach. The Advancing Strategies to Optimise the PerIopeRativE Management of PostOperative Nausea and Vomiting trial has received approval from the Vanderbilt University Institutional Review Board (IRB: 250773). The results will be disseminated through peer-reviewed publications and presentations at national conferences. Findings from this trial will inform best practices for timely antiemetic prophylaxis, with the goal of reducing PONV incidence and shortening PACU stay.

Trial registration number

NCT07152249.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• Will assess the impact of an actionable clinical decision support tool that supports postoperative nausea and vomiting risk assessment and provides a real-time checklist for planning antiemetic prophylaxis strategies.

• Will test preprocedural and intraoperative delivery of an automated, workflow-integrated clinical decision support tool for personalised postoperative nausea and vomiting risk prevention through the electronic health record.

• Employs a sequential, repeated crossover design with defined washout, control and intervention periods.

• Blinding is not feasible due to the nature of the design and intervention.

• It will be conducted as a single-centre study at a quaternary academic centre, which may limit generalisability.

Introduction

Postoperative nausea and vomiting (PONV) is one of the most common and distressing complications of surgery, significantly impacting both clinical and economic outcomes. PONV is associated with prolonged postanaesthesia care unit (PACU) stays, increased healthcare costs and decreased patient satisfaction.1,3 Given its prevalence and implications, tailored interventions based on individualised patient risk assessments are essential to improve perioperative outcomes.2,7 Anaesthesia providers are uniquely positioned to lead PONV prevention and management efforts. However, existing clinical decision support (CDS) tools, such as single-time-point alerts or post hoc reporting mechanisms, have not achieved meaningful improvements in adherence to PONV prophylaxis guidelines.⁸ These tools are limited by their lack of integration into clinical workflows, resulting in poorly timed interventions and provider alert fatigue. A missed opportunity lies in designing CDS systems that are seamlessly integrated into the anaesthesia workflow and tied to critical moments in the perioperative care timeline. Instead of generic, static alerts, a robust CDS tool could deliver time-sensitive, patient-specific recommendations for PONV prophylaxis and reminders for medication administration during pivotal phases, such as during the preoperative evaluation, postinduction or pre-emergence. This would ensure that PONV prophylaxis is prompted precisely, and only, when it is most relevant. Preliminary evidence supports the efficacy of such strategically timed, workflow-integrated alerts in changing clinician behaviour.^{9 10}

At Vanderbilt University Medical Center (VUMC), the Multicentre Perioperative Outcomes Group Anaesthesia Performance Improvement and Reporting Exchange (ASPIRE) programme has pioneered systematic post hoc reporting on PONV prophylaxis adherence and outcomes.¹¹ However, current CDS tools at VUMC and other institutions demonstrate notable limitations in addressing PONV. For instance, at VUMC, the existing CDS tool provides alerts only once early in the anaesthesia workflow—typically during the opening of the electronic anaesthesia chart—and only for patients identified with three or more PONV risk factors¹² (figure 1). This conservative threshold for triggering alerts deviates from PONV consensus guidelines,³ which recommend prophylaxis for patients with at least one risk factor. Despite the institutional protocol’s intent to improve outcomes, the modest improvements in compliance and PONV rates observed at VUMC may partly result from this stricter threshold. Furthermore, the single early alert provided by the CDS tools lacks integration into other critical perioperative time points, such as the preoperative evaluation, postinduction or pre-emergence phases, further limiting its effectiveness.^{9 13 14} As a result, compliance with PONV prophylaxis guidelines and reductions in PONV rates have been modest. Common barriers include poor integration of CDS tools into workflows, lack of motivation among providers to prioritise PONV prevention and alert fatigue due to poorly timed or irrelevant notifications.¹⁴

Figure 1. Current intraoperative clinical decision support (CDS) reminder at our institution. This CDS is triggered in the ‘Preprocedure’ section under ‘Procedure Information’ of the electronic health record to suggest that the patient has three risk factors for postoperative nausea and vomiting, and that preoperative administration of an antiemetic medication should be considered. CSMD, Controlled Substance Monitoring Database.

Time-sensitive, workflow-driven CDS alerts have the potential to overcome these barriers by seamlessly integrating into perioperative care and delivering actionable recommendations at the most critical moments.^{15 16} However, no large-scale, pragmatic study has evaluated the impact of linking CDS alerts to specific points in the perioperative timeline, such as preprocedure evaluation, postinduction or pre-emergence. This approach could enhance compliance with PONV prophylaxis with the aim of reducing PONV incidence and shortening PACU length of stay. Furthermore, PONV represents a widespread clinical problem, with approximately 2.5–10 million adult surgical patients in the USA affected annually. Addressing this issue has the potential to significantly improve patient outcomes and satisfaction on a national scale.

In this proposed study, we aim to improve PONV prophylaxis and reduce PONV rates by developing and implementing an anaesthesia workflow-driven CDS tool for personalised PONV prevention (AW-D CDS Tool). We hypothesise that by delivering time-sensitive, patient-specific checklists for antiemetic prophylaxis at critical perioperative stages, this tool will increase compliance with PONV prophylaxis guidelines, reduce the incidence of PONV and shorten PACU length of stay. Furthermore, this project has the potential to establish a scalable, generalisable model that could be adopted by other institutions to improve PONV management.

Methods and analysis

Our manuscript was prepared in accordance with Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) guidelines (online supplemental table 1), and the SPIRIT checklist is provided in online supplemental table 2.¹⁷ Our manuscript describes key elements of the Advancing Strategies to Optimise the PerIopeRativE Management of PostOperative Nausea and Vomiting (ASPIRE-PONV) trial protocol and statistical analysis plan.

Study design and intervention

The ASPIRE-PONV trial is a prospective, unblinded, pragmatic, sequential, repeated crossover trial conducted in the operating rooms and PACU, with washout, control and intervention periods, at VUMC in Nashville, Tennessee, USA. This multiple crossover study is intended to limit the errant detection of secular institutional trends. The ASPIRE-PONV trial protocol was approved by the VUMC Institutional Review Board (IRB: 250773). The trial protocol was also registered with ClinicalTrials.gov on 25 August 2025, prior to initiation of study participants’ enrolment on 3 September 2025 (ClinicalTrials.gov identifier: NCT07152249).

The ASPIRE-PONV trial will examine whether and how specifically linking CDS alerts directly to each patient’s care timeline—alerting the anaesthesia provider at critical points when prophylaxis should be administered—will impact compliance with PONV societal consensus and standardised practice recommendations.^{2 3} Provider compliance and patient outcomes will be monitored to help determine the intervention’s effectiveness on adequate PONV prophylaxis administration and the incidence of PONV compared with our current clinical care (ie, a CDS tool not linked to critical time points along the perioperative care timeline).

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Study site and population

The ASPIRE-PONV trial is being conducted in the Vanderbilt University Adult Hospital Main Operating Rooms (ORs), Medical Center East, Jim Ayers Tower and Gynaecological (4S) ORs, Radiology, Gastrointestinal-Endoscopy Suites, Vanderbilt Health Belle Meade and Vanderbilt Surgery Center Franklin at VUMC.

The inclusion criteria are as follows:

1. Patients:

   1. Age ≥18 years.

   2. Requiring general anaesthesia with endotracheal intubation or laryngeal mask airway.

   3. Have zero or more risk factors for PONV, including:

      1. Sex: female.

      2. History of PONV or motion sickness.

      3. Non-smoker.

      4. Duration of inhalational anaesthesia >1 hour.

      5. Undergoing selected surgical procedures:

         1. Cholecystectomy.

         2. Laparoscopy.

         3. Gynaecologic.

      6. Perioperative opioid use.

2. Anaesthesia providers:

   1. Any OR anaesthesia provider of eligible patients. The expected number of anaesthesia providers involved in the trial is approximately 570 providers.

The exclusion criteria:

1. Any patient or anaesthesia provider that does not meet the above specified inclusion criteria.

2. American Society of Anesthesiologists Physical Status 6, including organ procurement.

3. Patients anticipated to be transferred directly to the intensive care unit intubated.

4. Procedures:

   1. Electroconvulsive therapy cases.

   2. Intubation only cases.

   3. Labour epidurals.

   4. Transoesophageal echocardiography and cardioversion.

5. Surgery duration ≤30 min.

The time of enrollment for the trial (ie, ‘time zero’) is the time of completing and signing the plan section of the ‘anaesthesia evaluation’ electronic note, after the patient is met by a member of the anaesthesia provider team who will assume anaesthesia care of the patient.

Crossover design and intervention allocation

The 12-month period of enrolment in the ASPIRE-PONV trial is divided into four 12-week intervention and control blocks. During intervention blocks, the predicted risk score for PONV will be calculated in the preprocedural evaluation note, according to the patient’s specific number of PONV risk factors.¹² An automated PONV prophylaxis plan will then be generated based on their risk score, suggesting an appropriate number of prophylaxis medications according to our departmental guideline (online supplemental figure 1). An intraoperative, real-time checklist will then be available to the in-room anaesthesia provider, showing the recommended PONV medications to be administered throughout the workflow of the anaesthetic case. This real-time checklist will serve to optimise the timing of PONV medication administration (eg, preprocedure evaluation, postinduction, pre-emergence, figures2 3) by providing a visual reminder of the PONV prophylaxis plan, as well as real-time feedback on which medications have been administered, and which medications remain. During control blocks, the time-sensitive AW-D CDS Tool will be turned off, and care will revert to the routine use of an untimed CDS tool and institution-specific standards of care (figure 1). The anaesthesia providers will have access to departmental PONV guidelines either via a link embedded in the preprocedure evaluation electronic note or in the ‘I-care’ application from the electronic health record (EHR; online supplemental figure 1). The trial will begin with a 12-week control block followed by a 12-week intervention period. The first 2 weeks of the intervention period will serve as a washout, followed by 10 weeks of intervention during which the time-sensitive AW-D CDS Tool will be activated. After the first 12-week intervention period, the time-sensitive AW-D CDS Tool will be switched off for 12 weeks (2-week washout followed by a 10-week control period). This will be followed by the final 12-week intervention block of the trial (2-week washout followed by a 10-week intervention period) (online supplemental table 3). The trial began on 3 September 2025 and is ongoing.

Figure 2. Proposed clinical decision support reminder tested in the study. If the selected plan includes ‘general’ as the anaesthesia type, and any type of endotracheal tube or laryngeal mask airway is selected, the anaesthesia workflow-driven clinical decision support tool will appear below. The patient’s specific postoperative nausea and vomiting (PONV) risk factors will be automatically generated, along with a suggested plan for PONV prophylaxis based on their risk score. MAC, monitored anesthesia care, DLT, double lumen tube; LMA, laryngeal mask airway; ETT, endotracheal tube; NC, nasal cannula; CVC, central venous catheter; PAC, pulmonary artery catheter; TEE, transesophageal echocardiogram; H/O, history of; GYN, gynaecology; ICU, intensive care unit; DHCA, deep hypothermic circulatory arrest; IV, intravenous; Med Hx, medical history; PACU, postanaesthesia care unit; ROS, review of systems; TIVA, total intravenous anesthesia; CEBA, Center for Evidence-Based Anesthesia.

Figure 3. Demonstrating the ‘Intraoperative Sidebar’ showing planned prophylactic antiemetics, showing those administered and those remaining to be administered. H/O PONV, History of Post Operative Nausea and Vomiting; GYN, gynaecology; IV, intravenous; PONV, postoperative nausea and vomiting; H/O, history of; TIVA, total intravenous anesthesial CEBA, Center for Evidence-Based Anesthesia.

Randomisation was not feasible at either the provider or patient level. In our anaesthesia care team model, attending anaesthesiologists supervise Certified Registered Nurse Anaesthetists (CRNAs), anaesthesiology residents or Fellows and Student Registered Nurse Anaesthetists (SRNAs). Randomisation at the provider level was impractical, as the staffing structure inevitably creates combinations of attending anaesthesiologists with CRNAs, SRNAs and residents who might otherwise belong to conflicting study groups. Patient-level randomisation was also not feasible due to concerns about repeatedly withholding the reminder tool on a case-by-case basis. Because providers rely on reminders to reduce cognitive burden, we limited withdrawal of the tool to a single transition—from the initial experimental period to the subsequent control period.

Washout periods

The initial 12-week control block will be followed by a 12-week intervention block during which anaesthesia providers will receive the time-sensitive AW-D CDS Tool. The first 2 weeks of the intervention block will serve as a washout period, intended to mitigate potential carryover effects as providers adapt the tool to their clinical practice. Similarly, after the first 12-week intervention block, a 12-week control block will follow, during which providers will stop receiving the AW-D CDS Tool. The first 2 weeks of this control block will serve as a washout period, designed to mitigate residual effects of the tool in supporting the time-sensitive administration of antiemetic medications. The same structure will be repeated between the remaining 12-week intervention block, as illustrated in online supplemental table 3. Patient data collected during all washout periods will be excluded from the primary analysis to ensure consistency and eliminate potential carryover effects.

Blinding

Because of the anaesthesia care team model—where multiple providers are often involved in the intraoperative management of a single patient—blinding of both patients and anaesthesia providers will not be performed.

Data collection

The ASPIRE-PONV trial will minimise observer bias through two complementary methods of data collection. First, structured perioperative data routinely documented in clinical care will be automatically captured from the EHR and exported daily into Epic Clarity, the VUMC Data Warehouse and the Vanderbilt Department of Anesthesiology Perioperative Data Warehouse. Second, trained study personnel will perform manual data collection to capture any missing information and to validate the accuracy of the automated electronic data capture process (online supplemental table 4).

Outcomes

Primary outcome

The primary outcome of the study is adherence to appropriate PONV prophylaxis administration measured as a binary variable (yes/no).

Secondary outcomes

The ASPIRE-PONV trial will evaluate several prespecified secondary outcomes: (1) the occurrence of PONV within the immediate postoperative period, defined as the time from arrival in the PACU up to 6 hours after anaesthesia ends; (2) the occurrence of PONV or the administration of rescue antiemetic medication within this same immediate postoperative period and (3) the duration of PACU length of stay.

Safety outcomes

The ASPIRE-PONV trial will not include specific safety outcomes, as the use of single or multiple antiemetic medications for PONV prevention is standard, approved and routinely practised in clinical care.

Statistical analysis and reporting

Sample size estimation and power calculation

Based on data from the most recent 12 months of adult patients undergoing surgery at VUMC, the rate of the primary outcome—appropriate PONV prophylaxis administration—was 48%. If the true PONV prophylaxis rate in the adult intervention groups is 52.8% (a clinically significant 10% increase from 48%), we will need to study 1947 intervention and 1947 control patients to be able to reject the null hypothesis that the failure rates for intervention and control groups are equal with power 85%. The type I error probability associated with the tests of these null hypotheses is 0.05. Our accrual rates on average were around 487 in adult patients per week. Assuming a similar average weekly accrual rate, we expect to recruit 9740 patients per treatment group (2 periods of 10 weeks for each treatment group), respectively. Although smaller sample sizes are enough to detect a 10% change in the primary outcome, we would likely recruit a larger number of patients to mitigate potential seasonal and random differences between the groups, and to have more power to study secondary outcomes and the intervention effect heterogeneity in the primary outcome.

Data and Safety Monitoring Board and interim analysis

An independent Data and Safety Monitoring Board (DSMB) will oversee the trial, with a single planned interim safety analysis conducted after the first experimental period. This analysis will include reporting adherence rates to appropriate PONV prophylaxis administration in both the control and intervention groups. The DSMB is composed of two external physicians with expertise in anaesthesiology clinical practice and research, one bioethicist and one biostatistician.

Statistical analysis

The reporting of statistical results will follow the guidelines outlined in the SPIRIT statement.¹⁷ Continuous variables will be summarised as mean±SD or median (IQR), while categorical variables will be summarised as counts and percentages. Unadjusted comparisons between patients in the control and intervention groups will be conducted using either the Wilcoxon rank-sum test or Pearson’s χ² test, depending on the type of data. Differences between the control and exposure groups will be quantified using standardised mean differences.¹⁸

To test the primary hypothesis of no association of intervention and outcome in either time period, we will fit logistic regression with adequate PONV prophylaxis (primary outcome, yes/no) as an outcome and intervention (control/intervention), time period (first/second round of control and intervention periods) and interaction of intervention and time period. The model will also adjust for demographics, clinical characteristics and surgery characteristics. To avoid overfitting, the model’s df will not exceed one per 15 events or non-events.¹⁹ Inclusion of covariates will be prioritised by clinical relevance to PONV risk, starting with sex, smoking status, history of motion sickness, age, American Society of Anesthesiologists Physical Status classification, body mass index, followed by surgical and anaesthetic factors.

The association of intervention and adherence (yes/no) to appropriate PONV prophylaxis administration will be considered significant if the chunk test p value of the intervention and interaction of intervention and time period coefficients is <0.05. The treatment effect will be reported for each time period as ORs with 95% CIs for having the outcome in the intervention versus the control group. We will perform two secondary analyses to examine heterogeneity of treatment effect: (1) by surgery type and (2) by PONV risk score.¹² These two secondary models will be fit with a three-way interaction term of intervention, time period and surgery type or PONV risk score, accordingly. The treatment effect will be reported for each time period and for each category of the variable of interest (ie, surgery type or PONV risk score) as ORs with 95% CIs for having the outcome in intervention versus control patients.

The analyses for the secondary outcomes—the occurrence of PONV within the immediate postoperative period (yes/no, defined by nursing documentation in the PACU up to 6 hours after the end of anaesthesia indicating nausea or vomiting) and the occurrence of PONV or the administration of rescue antiemetic medication within the same period (yes/no, defined as either nausea/vomiting OR administration of a rescue antiemetic)—will be performed similarly to the primary analysis.

The analysis for the secondary outcome, PACU length of stay defined in minutes, will be performed like the primary analysis but using a proportional odds logistic regression model. In the analyses for the exploratory aim, we will summarise the frequencies of antiemetic medication use, stratified by intervention. We will also perform exploratory analysis using logistic regression models, with the occurrence of PONV within the immediate postoperative period and the occurrence of PONV or the administration of rescue antiemetic medication within the same period as outcomes, and with the type of antiemetic medications along with demographic, clinical and surgical characteristics as covariates.

We expect the washout period to prevent carryover effects, but if they persist, the treatment effect may be biased towards the null without inflating type I error. We will also perform a sensitivity analysis by ignoring the data during the first 2 weeks after the washout periods, essentially extending the washout period from 2 to 4 weeks. This analysis may help us understand whether the 2-week washout periods were adequate to eliminate the carry-over effect.

Missing demographic, clinical and surgical characteristics will be imputed using multiple imputation with chained equations and predictive mean matching,^{20 21} generating 10 datasets. Model results for these 10 datasets will be combined using Rubin’s rules.²² Analyses will be conducted at a 2-sided type I error rate of 0.05 and performed using R software.²³

Study status

The ASPIRE-PONV trial has already been initiated. The first patient was enrolled on 3 September 2025. The estimated completion date for our study is 3 September 2026.

Ethics and dissemination

IRB approval and waivers of informed consent

This study was approved by the Vanderbilt University Institutional Review Board (IRB: 250773) with a waiver of informed consent for patient participation, given the minimal risk and the impracticality of obtaining informed consent from the target study population. The PONV prevention and management guideline used in this study has departmental approval and is already in clinical use. We do not anticipate any additional risks to patients.

A waiver of consent for provider participation was also granted by the Vanderbilt University IRB. There are no provider-specific eligibility criteria, as the CDS information delivered through the AW-D CDS Tool is already available to providers. The tool links CDS checklists directly to each patient’s care timeline, alerting anaesthesia providers at critical points when prophylaxis should be administered based on the approved departmental guideline. Providers will not be directly identified, and no identifiable data will be collected. Interactions with providers will occur only electronically through the AW-D CDS Tool in Epic, along with an explanation of the tool at trial initiation. Provider behaviour will be evaluated only as a secondary outcome, based on adherence to appropriate PONV prophylaxis administration and its relationship to the secondary outcomes.

Protocol changes

Any modifications to the trial protocol will be documented on ClinicalTrials.gov in accordance with SPIRIT guidelines.

Data handling and sharing

All study data will be de-identified prior to analysis by trained study personnel. Only vetted and approved research staff will have access to the data.

Dissemination plan

Findings from this trial will be submitted to peer-reviewed journals for publication and presented at national and international scientific conferences.

The funder did not influence the design or conduct of the study, despite authors’ affiliations with the funder.