Collaborate with us to create the 28-day readmission data visualisation

One of the key takeaways from our stakeholder engagement sessions in 2020 was the importance of context: that a standalone number isn’t as valuable as knowing the full story.

In preparing the data for the full acute inpatient suite of KPIs, we have observed some interesting patterns, but we would like to check what they mean. For example, do they represent data entry errors, conflicting interpretations of codes, variable business practices, genuine differences in service – or some combination?

Here is what we’ve noticed:

28-day readmission (discharges since 1 January 2019)

• High proportion of readmissions occur on the first day after inpatient discharge
  • Nationally, 10 per cent of 28-day readmissions occur on Day 1 (and 17 per cent occur on Days 1 and 2 combined)
  • Ten DHBs have 20+ per cent of their 28-day readmissions occur on Days 1 and 2
  • Six DHBs have 10+ per cent of all readmissions occur on Days 1 and 2
• Across the board, 75 per cent of Day 1 readmissions are into the same team as the original inpatient discharge
• Excluding Day 1 and Day 2 readmissions changes the 28-day readmission rate by over 4 percent at five DHBs.

We’re wondering…

• Could some of these be data entry errors, for example an accidental discharge by someone who doesn’t have access/permission to undo the discharge, and so creates a new referral in the IT system to re-activate the client record?
• Should some of the 25 per cent of inpatient discharges almost immediately readmitted into another team or organisation have had DT or DW end codes to indicate a transfer?

Length of stay (referrals started and ended since 1 January 2020)

• Frequency and duration of leave (T37) in acute inpatient referrals are quite variable across DHBs
  • In four DHBs, 1 per cent (or fewer) of nights in an acute inpatient referral are spent on leave; in other DHBs, up to 29 per cent of nights in acute inpatient referrals are reported as leave
  • In five DHBs, fewer than 10 per cent of acute inpatient referrals include any leave nights at all; in other DHBs, up to 71 per cent of acute inpatient referrals include at least one night of leave.
• Variable patterns of sub-acute inpatient (T04) bednights delivered in acute inpatient referrals
  • 16 DHBs report no sub-acute bednights against acute inpatient referrals, whereas 3 DHBs do deliver sub-acute bednights against acute inpatient referrals
  • At one DHB, 21 per cent of all nights in acute inpatient referrals are sub-acute inpatient bednights (and 29 per cent of all acute inpatient referrals included at least one sub-acute bednight).

We’re wondering…

• Are leave patterns just quite different between DHBs, or do these patterns represent different data entry practices for when tāngata whai ora are not present in a bed?
• Do some DHBs regularly deliver sub-acute inpatient support as part of the acute inpatient referral while others transfer to a different sub-acute team to provide that step-down support if required?

Next steps

Before we progress the development of the data visualisation for 28-day readmission we would love to collect more perspectives about the above patterns and why they might exist.

So, we have set up a virtual Zoom meeting on Friday 12 March 2021 at 1pm for anyone from our sector who may be interested in collaborating with us to investigate and draw out the right context for this KPI. To register, please see the event details below.