@CardioMetaBoliK Good to see NICE asking the public. Human involvement is the point people keep raising, and it would help to spell out what that means in practice, including who reviews the case when an AI-supported decision goes wrong.
@HlthcareToday Faster letters are welcome. The safety question with ambient voice is what the clinician checks before signing, and how easily a draft error slips through when the letter is ready in minutes. How did the team handle review and sign-off?
@HSJnews An AI tool can only work with what the record holds, so gaps and workarounds in the EPR carry straight through to its output. When AI is part of an incident, the review needs to look at the record and the workflow around it as well as the tool.
@DrNeenaJha Waiting rooms do safety work that's easy to miss on a floor plan. Reception staff notice the patient who looks worse than when they booked. Take that space away and the risk moves somewhere less visible, as it has with corridor care in EDs.
@HSJnews Staff take their cue from how they are inspected. If a visit feels like fault-finding, people prepare a defence, and that habit carries into incident reviews. PSIRF asks trusts to look at systems and learning, so it helps when inspection asks the same questions.
@bmj_latest Learning here depends on someone owning the join when a patient moves between providers. Each organisation can run its own learning response and still miss the handover. A joint review, or an agreed route for sharing findings both ways, lets both sides see the whole pathway.
@HealthwatchE The accounts people gave are learning conversations i.e. what corridor care does to monitoring and to communication with families. If those conditions are recorded winter planning has evidence of how the system is running. Reviews can then treat corridor care as a standing risk.
@NIHRMaudsleyBRC@my_DHC@MaudsleyNHS@KingsMaudsley@KingsIoPPN A public adverse-incident database will only help if events are also reviewed in the trust where the tool was used. Regulation can set the duties and the monitoring. The learning still has to happen with the staff who used the system.
@healthtechworld An AI-related error usually sits between the tool supplier and Trust. If incidents involving AI are reviewed under PSIRF like any other safety event, trusts can see where risk lands. Trusts and suppliers sharing the learning save each org. finding the same problems on their own.
@FrontPubHealth Mapping what patients and staff wanted onto SEIPS is a useful step, as it's more often used to trace back from harm. Making interpersonal care visible matters, because in a crowded ED it's the first thing to slip and the hardest to see in a review afterwards.
@skittles_kitty Guidance says the CQC can check whether notification and apology happened, and it can prosecute a failure to notify. The unresolved issue is what triggers that scrutiny when the family has never been told there was an incident.
@GreigVile@NHSTayside@gmcuk@NHSScotland The missed chances to step in are often the most important finding in cases like this. Each organisation can review its own part and still miss how the warnings fell between them. Looking at the whole journey from the patient's side is how those gaps come to light.
@zoetidman Inappropriate access to records can show up in incident reports long before it reaches a solicitor. It would be good to know whether trusts review these cases with the same learning approach they use for clinical incidents, with the people affected involved from the start.
@SandipP61970438@skittles_kitty A shared record would help. Much of what we lose sits in how findings move between organisations. LFPSE and the HSSIB reports are meant to carry that learning across. What has worked best for spreading learning where you are?
@skittles_kitty When a patient has died, DoC still applies, the notification and apology go to the family. Trusts should record when that happened, and CQC can check it on inspection. A family who isn't sure they were told can ask the trust to show them its candour record
@ShaunLintern@YvetteCooperMP Families notice the warning signs before anyone else, a slow response can mean the same risk is missed again. When families are involved early in a learning response, their account shapes the questions of the review. The same principle should apply when they write to government.
@susiehewitt Long waits in a busy, noisy department can make a crisis worse, and that is hard for the staff trying to help too. When a learning response looks at an incident like this, the setting deserves as much attention as the decisions people made. I'll read HSSIB's findings closely.
@Damian_Roland Asking the caregiver at every set of observations makes family concern part of the routine. It also gives a learning response something concrete to look back on when a child's deterioration was missed.
@pcdosociety The finding that people with learning disabilities and mental health problems miss out on this technology is one local teams can act on. When they review insulin incidents, it's worth checking whether access to technology comes up at all.