Back to Evidence

Appraising SBAR Handover for Therapeutic Endoscopy with an AI Medical Search Engine | QSevidence 2026

Evidence-Based Medicine23 min read

Completing therapeutic endoscopy does not establish that risk information has reached the next team. During transfer to recovery or a ward, staff need to know what was done, the patient's current condition, what requires observation and whom to contact about deterioration.

Appraising SBAR Handover for Therapeutic Endoscopy with an AI Medical Search Engine | QSevidence 2026

Structured handover should answer these questions explicitly.

QSevidence can assist nursing and quality teams with retrieving, comparing and tracing evidence. This discussion of an AI medical search engine examines SBAR study design and local workflow adaptation. The available research narrative still requires verification against original records and consistent statistical definitions; its effect figures are not presented as established safety improvements.

Define the research question around an actual transfer

SBAR organizes communication into Situation, Background, Assessment and Recommendation. For endoscopic mucosal resection, submucosal dissection, ERCP and endoscopic hemostasis, the same structure must accommodate different procedural priorities. Transfers involve endoscopists, anesthesia personnel, endoscopy nurses and receiving nurses. Evaluating whether a form was completed is insufficient.

The proposed question is whether a specialist SBAR checklist plus training improves completeness and accuracy of applicable information, and reduces communication-attributable errors compared with local usual care. Completeness can be a primary process outcome. Patient harm needs separate definitions, observation windows and adjudication. Staff satisfaction and perceived patient safety describe experience rather than substituting for clinical safety outcomes.

Turn SBAR into a checkable specialist framework

Information domains for therapeutic endoscopy handover
DomainInformationReceiver's check
SituationConfirmed identity, actual procedure, completion time, current observations and consciousnessCorrect patient and observation time; recent changes not concealed by old values
BackgroundRelevant conditions, allergies, antithrombotic history, key tests, procedural treatment and sedation recordCompleted and pending actions distinguished; doses and times checked against actual orders and records
AssessmentRecovery, breathing and circulation, pain, concerns about bleeding or perforation, ERCP risks and applicable devicesEvidence, trends and uncertainty rather than an unsupported label of stable
RecommendationMonitoring, pending investigations, treatment, diet and activity instructions, escalation routeNamed responsibility, timing, triggers and confirmed acceptance

This is a communication framework for local review, not a universal order set. A device that is absent should be marked not applicable; a test not performed should not be entered as normal. Anticoagulant or antiplatelet interruption and restart plans should reflect the responsible clinician's confirmed decision, not a decision made by the form. Differences between verbal, paper and electronic records require clarification and a correction trail.

Recovery cannot be reduced to one score

Handover should distinguish baseline consciousness from the current state and identify sedation or anesthesia exposure, airway and oxygen support, problems during recovery and outstanding observations. Recovery scores such as Aldrete may be used according to institutional policy. A score does not replace assessment of respiratory, circulatory and procedural concerns, or automatically establish readiness for discharge.

The British Society of Gastroenterology sedation guideline supports recovery monitoring, documented discharge assessment, and clear follow-up advice and contact information for patients and carers. Its setting must be considered when adapting local practice. Recovery from sedation does not exclude every delayed procedural risk.

Procedure-specific concerns need explicit transfer

After mucosal resection or dissection, communication can focus on the treatment site, hemostasis or closure performed, and observations relevant to bleeding or perforation. ERCP handover should add the actual intervention, stent or drainage details, and the observation plan for concerns such as pancreatitis or biliary infection. Identical laboratory tests, imaging and monitoring intervals are not appropriate for every patient.

Persistent or worsening abdominal pain, hematemesis or substantial rectal bleeding, altered breathing or consciousness, or circulatory deterioration warrant prompt clinical assessment and institutional escalation. A checklist should not make staff wait for an unvalidated numerical threshold before seeking help. Monitoring schedules, investigations, diet and medicines follow actual risk assessment and orders; fixed intervals in an example are not universal protocols.

Close the loop rather than assuming information was heard

The sender identifies the most pressing concern and pending actions. The receiver can ask questions and repeat the key actions, which the sender confirms or corrects. Important tasks have an owner and an escalation route if they cannot be completed. Acute deterioration requires care without waiting to finish documentation.

The Agency for Healthcare Research and Quality describes structured information sharing in its SBAR tool and explains receiver repetition with sender confirmation in its Check-Back tool. These resources inform process design; they do not themselves establish a reduction in harm at a particular endoscopy center.

Design a prospective nonrandomized evaluation carefully

A local center and receiving wards could first observe usual handover, then introduce the specialist checklist, standardized training, simulation and feedback. Actual usual practice should be documented rather than presumed entirely unstructured. Sequential periods constitute a before-and-after comparison, not random allocation. Case mix, rosters, equipment and concurrent quality initiatives may also change outcomes.

The population can include adults receiving therapeutic endoscopy who require transfer, together with participating staff in different roles. Inclusion of emergency, high-risk or direct-to-intensive-care transfers should be specified prospectively. Temporary inability to answer a questionnaire after sedation does not remove the need for safe handover. Clinical care and questionnaire eligibility must remain separate.

A patient may undergo several transfers, while one staff member may participate in many cases. Specify whether the analytical unit is the patient, handover event or staff member, and link these levels without conflating them. Event-level analyses should address correlation within patients, personnel and periods. Repeated staff satisfaction responses should not be treated as independent people.

Measure completeness, accuracy and interaction separately

Freeze the checklist version and define core information, interaction items and handling of nonapplicable fields. Scores from 20 items, 23 items or a percentage scale cannot be mixed without explaining conversion. A study could report the proportion of applicable information transmitted both completely and correctly, plus critical omission rates, while separately recording questions, confirmation and transfer of responsibility. A total score can otherwise conceal a high-risk omission.

Mentioning an item does not establish its accuracy. Observers need to compare the handover with orders, procedural records and valid observation times. Expert content review, cognitive interviews and piloting should precede assessment of inter-rater agreement. Internal consistency alone does not establish comprehensive checklist validity. Recordings, direct observation and chart review have different biases and require appropriate privacy authorization and de-identification.

Adverse events need independent attribution and denominators

An omission is a process defect. Medication errors, investigation delays, near misses and actual harm require separate definitions; a composite rate should not be described simply as fewer complications. Postprocedural bleeding is not attributable to communication merely because it occurred after handover. Independent reviewers can apply prespecified criteria for attribution, severity and preventability within an explicit inpatient or postdischarge window.

Sample size should follow a plausible primary-outcome difference, variability, clustering and missingness. Adding one arbitrary percentage does not replace a design-effect calculation. Analysis can address procedural risk, procedure type, sedation and time trends, and should report effect estimates with confidence intervals. Nonsignificant baseline tests do not establish equivalent groups, and simple group comparisons cannot remove nonrandomized bias.

Result status and implementation outcomes

What can currently be presented clearly is the workflow, evaluation domains and testable research questions. Sample records, score maxima and event proportions in the research narrative have not been reconciled with original data and consistent definitions. Consequently, this article reports no confirmed effect size, P value or risk reduction. In particular, absence of an observed serious event in a short period must not be interpreted as elimination of risk.

Implementation evaluation should also track actual checklist use, training completion, handover duration, duplicate documentation, workload and acceptability. Patient or family perceptions should be collected when respondents can understand and answer, with reasons for nonresponse recorded. Sustainability, electronic downtime procedures, refresher training and version control also matter. Future work could use multicenter cluster randomization or a controlled time series with sufficient observations.

Use evidence retrieval for local quality improvement

QSevidence can assist searches around therapeutic endoscopy, handover, SBAR, recovery and communication errors, then compare populations, procedures, designs and outcomes for nursing, endoscopy, anesthesia and statistical review. Communication resources, specialist guidance and local research data answer different questions. Effect sizes from other departments should not be transferred directly. Local adoption still requires appropriate ethics and quality oversight; the platform does not replace bedside assessment, order confirmation or emergency decisions.

Frequently asked questions

How can QSevidence support preparation of an endoscopy SBAR checklist?

QSevidence can help retrieve communication and recovery information and organize shared fields and procedure-specific details. The local endoscopy, anesthesia and nursing team reviews the final checklist against practice and refines it through simulation.

How can QSevidence support comparison of handover studies?

QSevidence can assist with comparing score versions, analytical units, event definitions and follow-up windows. Researchers interpret effects using original data, clustered analysis and bias assessment rather than satisfaction or P values alone.

How can QSevidence assist sustained handover improvement?

QSevidence can help organize audit outcomes, training questions and source updates. Authorized professional teams monitor actual implementation, clinical safety and responsibility transfer, refining practice from real feedback.