Shift handover: how to record a change of shift without losing information

Six to ten minutes. That is how long a nursing shift handover typically takes in Brazil: spoken, in the staff room, with nothing to rely on except whatever each professional manages to write down afterward.

A Brazilian study validating a shift handover instrument with the SBAR tool, published in the Revista Brasileira de Enfermagem, measured exactly that: the predominant modality is speech, in the nursing room, between 6 and 10 minutes. The same study shows what gets in the way: delays and early departures interfere with the quality of what is passed on.

There is a part of the handover that never reaches the medical record: what worries the team, what was left pending, what only makes sense to the people who were there. And that is precisely the part the incoming shift will need at 3 a.m.

This post is about closing that gap without writing anything twice. The structure of the handover was solved years ago. The record is what still lives in memory.

What a shift handover carries that the medical record does not

The medical record is a record of what has already happened. The shift handover is a heads-up about what can still happen.

That is where information with no field of its own in the system circulates: events that never turned into a test, changes in the patient’s behavior, what the team agreed among themselves, the “if it gets worse, call before waiting for the shift to end.” Current nursing guidance is that the handover happens verbally and that the records stay accessible to the incoming shift. The problem is that the accessible record usually contains less than the spoken word.

In practice, this creates a strange situation: the most important conversation of the day is also the only one that leaves no trace.

What gets lost when memory is the only thing holding the handover

Communication failure at shift change is not an operational detail. Miscommunication is described in the literature as one of the leading causes of serious medical errors.

The best known study on the subject, published in the New England Journal of Medicine in 2014, tested a standardization bundle called I-PASS across nine hospitals, with 10,740 admissions followed. The result:

In other words: a 23% drop in errors and a 30% drop in preventable adverse events, with no increase in handover time. Standardization did not make the handover longer. It made it complete.

A Brazilian study implementing I-PASS in a tertiary pediatric hospital found the same pattern in another metric, equally telling. After the training, explicit mention of pending items rose from 31% to 81% of handovers, and the presence of a contingency plan jumped from 16% to 73%. Time spent on interruptions and conversations that did not refer to the patient fell from 18% to 2.7%.

Notice what changed: what started getting said was what stayed pending and what to do if the worst happened. Those are exactly the two things that, when missing, force the next shift to rebuild the context in the middle of the hallway.

SBAR and I-PASS: which to use and when

Two structures solve the “what to say” problem. They do not compete with each other: they solve problems of different sizes.

SBAR has military origins: it was created for communication on nuclear submarines, a context in which a misunderstanding is not an inconvenience, it is catastrophic. It was adopted in healthcare for being fast and standardized, and it is recommended by the Joint Commission International.

I-PASS adds two elements SBAR does not have: situational awareness (what can go wrong and what to do if it does) and synthesis by the receiver, meaning whoever takes over repeats what they understood. That last item is the most underrated of all: it is a check that the information arrived, not just that it was sent.

The step almost nobody takes: recording the handover

If the structures already exist and the results are documented, why is the handover still nothing but spoken?

Because recording it was work. Two usual attempts:

1. Fill in a template after the shift. It works in the first week. On the day the shift ends 40 minutes late, the template stays blank, and it is the last task of a 12-hour turn.
2. Trust what is already in the system. What is in the system is what happened to the patient, not what the team agreed about the patient.

There is a third path, and it is the one transcription technology made viable: record the handover that is already happening and let the transcript become the record. No extra writing, no change to the format of the handover, and no third person taking notes while everyone else talks.

The flow looks like this inside Sintesy:

• Record the shift handover on your phone while it happens, or upload the audio afterward. If the audio is poor, the transcript gets it wrong, and the problem is almost always the capture, not the recognition.
• Transcription in Portuguese of what was said, bed by bed, in the order the handover happened.
• Automatic summary: the summary of the handover, with what happened to each patient and what stayed pending.
• Pending-items checklist when the handover contains concrete tasks. Worth saying clearly: Sintesy decides between a checklist and an outline based on the content, on the same principle that makes the meeting checklist exist only when there is a real agreement. A purely informative handover becomes a summary; it does not invent a task list where there is no task.
• Search later: when someone asks “when did that bed start getting worse?”, the answer shows up in the transcript of the handover where the subject was discussed, without listening to the whole audio again, the same logic as searchable transcripts.
• Export to PDF, DOCX, SVG or PNG to attach wherever the service needs it (export and sharing are part of the paid plan).

And the summary has a function no paper template has: it can be read by whoever is taking over the shift before they arrive at the station.

How to do this on your next shift (in 5 steps)

1. Agree first. A recorded handover without the awareness of the team and of management becomes a trust problem. Alignment comes first, always.
2. Start recording before the handover starts. The first seconds carry context that usually gets lost at the end, when fatigue sets in.
3. Speak bed by bed, in SBAR or I-PASS order. One patient at a time, no going back. A pending item always with a name: “the tech is to check oxygen saturation at 22:00” is a record; “someone was going to take a look” is noise.
4. Finish with the read-back. Whoever takes over repeats what they understood. It is the synthesis element of I-PASS and it takes less than 30 seconds.
5. Generate the transcript and name the file with your unit’s standard. Something like 2026-09-18_medical-ward_night. A handover with no name becomes a lost audio file on someone’s phone.

LGPD, confidentiality and health data: what to decide before you record

Here it is worth being direct: health data is sensitive personal data, with processing restricted by the LGPD, Brazil’s data protection law. The decision to record a handover is not the individual professional’s: it belongs to the service, which needs a policy, an adequate legal basis and a designated officer.

What usually works in practice:

• Align with management and with the data protection officer (DPO) before recording. If the service does not authorize it, do not record: the structured written record still holds.
• Minimize identifiers. If the policy allows, record by bed and initials instead of full name. The less identification, the lower the exposure.
• Restrict access to the people involved in care and define how long the files are kept.
• Discard the audio once the transcript has served its purpose as a record. Audio is the input; the text is the record.
• Keep professional confidentiality as set out in your profession’s code of ethics. Recording does not change the duty of confidentiality.

None of this is legal advice: it is the list of questions the service needs to answer before the first recording. And it is worth separating the two records: this post is about the shift handover, the conversation between shifts. The record of the care itself, the patient’s progress note, is another flow, which we cover in clinical notes with AI: SOAP notes and medical records.

The same logic outside nursing

Shift change is not exclusive to healthcare, and the pattern repeats in any operation that runs 24 hours:

• Service handover in manufacturing, with what is running, what is stopped and what needs attention.
• Support shifts (NOC, SOC, 24-hour support), where the overnight incident has to be explained to whoever arrives at 8 a.m.
• Front desk, hospitality and security, where an occurrence reported out loud rarely becomes a record.
• Project continuity between teams and vendors, in the baton pass from one owner to another.

In every case, speech is where context lives and the record is what makes context survive. The difference is that recording no longer requires writing.

Five mistakes that turn a handover into a lost meeting

• Handing over “all quiet” without mentioning pending items: if nobody mentions what is still open, the next shift finds out from the patient.
• Skipping the read-back: information sent is not information received.
• Turning the handover into a team meeting: interruptions and topics that are not about the patient are the item that consumes the most time in the process.
• Recording without agreement: technically easy, institutionally expensive.
• Leaving the audio with no name and no summary: it exists, but nobody can find it. A 10-minute file saved as audio_1234 is the same as nothing.

Quick questions

How long should a shift handover take? In studies with standardized handover, the average sits around 2 to 3 minutes per patient, per professional taking over. The oral team handover measured in Brazil, across all the beds in a unit, came in between 6 and 10 minutes.

What is the difference between SBAR and I-PASS? SBAR organizes information into four blocks and works to communicate a single case, including on a phone call. I-PASS adds the action list, the contingency plan and the synthesis by the receiver, and it is built for a full shift handover with several patients.

Does this apply to day shifts too, or only to nights? It applies to every shift change where someone takes on responsibility for patients, systems or operations. Nights tend to concentrate more risk because the team is smaller and support is slower.

I don’t have time to take notes during the handover. How do I record it? That is exactly the point: you don’t take notes. You record the handover that is already happening and review the transcript afterward, or you use the generated summary as the shift record.

Do I need a paid plan to do this? Not for the basics. The free plan includes 60 minutes of audio per day and up to 10 summaries per day, more than one shift handover per turn. Export features (PDF, DOCX, SVG, PNG) and sharing are part of the paid plan, which costs R$ 39,90 per month or R$ 149,90 per year.

The next shift starts with the record of the previous one

The shift handover already has structure, already has acronyms and already has evidence that it works. What is still missing on most teams is the record, because recording meant typing it up afterward, and nobody does that at the end of a shift.

Recording the handover and letting the transcript do the work solves the problem in the right order: the team keeps talking the way it has always talked, and the next shift receives something better than the memory of someone who was tired.

If you want to start small, start with one unit and one shift. Compare a week of recorded handovers with what the team remembered before.