In all, 598 out of 644 reports were included in the analysis; 46 incidents were excluded due to reports not describing an actual patient safety incident (n = 18), insufficient narrative detail (n = 24), or no free-text description available (n = 4). Cohen’s κ statistics were calculated for the primary incident type (chronologically closest to the outcome experienced by the patient) at κ = 0.79 (95% confidence intervals = 0.78 to 0.81). Reports were submitted from 116 different locations (for example, health boards in Wales, and formerly primary care trusts in England). Some organisations reported a discharge-related incident once, while others reported over 40 reports, although most organisations reported <10 reports.
The most frequent contributory factors were inefficient or poorly followed protocols by staff (n = 308, 52%); lack of, or insufficient, organisational protocols (n = 184, 31%); and environmental issues such as a lack of equipment availability (n = 97, 16%). The degree of harm to the patient was identifiable in 554 (93%) of the 598 included reports, with 44 reports not describing patient outcome at all, and a further 91 describing incidents where the patient was not harmed. Most reports (n = 463, 77%) described harm occurring to patients, with 381 (64%) experiencing low harm and 82 (14%) moderate harm or worse (see examples 1 and 2 in Box 1).
Box 1. Edited extracts of incident reports (salient points highlighted by the authors for illustration)
Example 1. Severe harm
Contacted by Ward **** on
Friday afternoon to inform us that patient was coming home with a catheter and would need a visit on Monday 19.10.09. Over the weekend patient had a visit from on call GP and district nurse. He was virtually
immobile, confined to bed, had two grade 2 pressure sores, one on each buttock. He couldn’t eat due to extreme oral thrush. None of those problems were addressed on the
discharge letter. Patient was also sent home with
no analgesia despite being on morphine in hospital, and was vomiting virtually all weekend. Has been struggling to tolerate any diet and fluids and developed UTI. Patient had been told to return to hospital on Tuesday 20 October for trial without catheter, but
did not know where to go or what time. Condition deteriorated and
readmitted on 21.10.09.
Example 2. Moderate harm
91-year-old patient was admitted to ********** in June 2009. The patient was then discharged to ******* Nursing Home, who then performed a home assessment and the patient was discharged on 28 September. The district nurse rang on 29 September advising that she felt
discharge was inappropriate. The room was too small for
equipment (hospital bed, hoist, commode) and care staff were unable to care for him properly. Apparently the patient was hardly eating or drinking (GP spoke to the patient’s daughter who confirmed this). GP discussed with *** Ward at ******** and intermediate care team who felt that the patient should be
readmitted to nursing home bed via social services.
Example 3. Low harm
The health visitor carried out a primary birth visit following the twins’ discharge from the
Special Care Baby Unit (SCBU). There was
no discharge letter with information for the service, or medications required.
No discharge plan. No
resuscitation training given to the parent. The mother stated that she was told it would be given before discharge, but that it was not received. Twin discharged on oxygen therapy.
No apnoea monitor.
No risk assessment surrounding this twin’s care.
No official referral to the paediatric community nurse and no involvement pre-discharge. The paediatric community nurse was
not informed of the discharge. The twins had been cared for over the past 7 weeks in the SCBU.
No liaison had been made with the community staff.
Example 4. Low harm
This patient was discharged from ******** on 24/12/08 having had a
laparotomy and subtotal d2 gastrectomy for gastric adenocarcinoma on 17/12/08. This man was discharged home
without a referral to district nurses. His wife is very poorly and expected to cope with his care, medication, and heparin injections. This was a very poor discharge
that could have resulted in readmission and has been
very stressful for this couple.
Example 5. Low harm
Message received from GP 10.4.10 — patient was discharged from … ward 9/4/10
late pm —
no referral sent to child district nurse. Urinary
catheter (long-term) in situ.
No advice given to family re changing bags/care of catheter and
no bags supplied on discharge.
No information as to whether district nurse can change catheter.
Example 6. Low harm
A 92-year-old man was discharged from hospital after being recommenced on warfarin therapy. It
was assumed by the medical staff on the ward that the GP surgery would take over the monitoring of the patient’s INR. The only correspondence the surgery received was an anticoagulation form with the patient target INR and recent INR (International Normalised Ratio) recordings and warfarin dosages. No
indication for the warfarin was documented or date of discharge. The surgery only became aware of the patient discharge when a receptionist was contacted to request warfarin. The ward sister was contacted and she felt that as a district nurse had been arranged to take an INR nothing else needed to be done. When questioned about the patient suitability she felt that as the patient was taking a lot of medication in hospital there wasn’t an issue. When asked if the patient suitability to change his dose of warfarin was checked she felt that if there was a problem perhaps his daughter could administer the warfarin.