The IPRU regularly uses New Zealand publicly funded hospital discharge (National Minimum Data Set – NMDS) and Mortality Collection data from the Ministry of Health in producing NIQS, customised queries, and for projects on particular injury-related topics. There are a number of important issues to consider when using these data. This page highlights some of these issues and explains the restrictions that IPRU commonly apply to the data.

Publicly Funded Hospital Discharge Data Specific Notes

IPRU determines a count of injury incidence considering the following criteria:

Principal diagnosis of injury

Hospital discharge data is restricted to patients with a principal diagnosis of injury, within the range of ICD10 diagnosis codes S00-T98 (Injury, poisoning and certain other consequences of external causes). Within the NMDS, a given hospital event may have several contributing diagnoses, but the principal diagnosis is given to be the primary reason for admission.

Day patients

Day patients are excluded to make the data comparable across time, hospital, and cause of injury. Length of admission for less serious cases can vary over time as treatment practice changes. Such an exclusion also serves as a proxy for severity and a factor minimising reporting bias from short stay emergency department discharges (see below).

Readmissions

In considering the incidence of injury events, only first admissions for an individual event are included. A re-admission is defined as an event that shares the same date of injury or is admitted within one day of a previous discharge, and is excluded from the count.

Discharge Type

To avoid double counting with the Mortality Collection, hospital discharge data includes only non-fatal data. i.e.it does not include patients who were discharged dead from hospital.

Short stay emergency department reporting

Short stay emergency department (ED) discharges are health events in which the patient is seen in the ED and discharged without admission as an inpatient. Historically these events were not included in the National Minimum Dataset (NMDS). In 1999 one District Health Board (DHB) began reporting these events to the NMDS when a patient received a minimum of three hours of treatment. Other DHBs followed, and in 2008 approximately half of all DHBs were reporting short stay ED events to the NMDS. From July 2009 reporting short stay ED events to the NMDS has been compulsory. This change in reporting practice may affect trends by showing an increase in the frequency and rate that may not be reflective of a real change in injury.

All Injury vs All External Causes

Events due to medical procedures have traditionally not been considered injuries. These include:

• “misadventures to patients during surgical and medical care” ( ICD10-AM: Y60-Y69)
• “surgical and medical procedures as the cause of abnormal reaction of patients or later complication, without mention of misadventure at the time of procedure” (ICD10-AM: Y83-Y84)
• “drugs, medicaments and biological substances causing adverse effects in therapeutic use” (ICD10-AM: Y40-Y49)
• “medical devices associated with misadventures in diagnostic and therapeutic use” (ICD10-AM: Y70-Y82)

These types of events are included in NIQS but can be excluded by selecting “All Injury” rather than “All External Causes”.

General Notes

Change from ICD-9 to ICD-10

In mid1999 New Zealand hospitals switched from using the ICD-9 coding scheme to using the ICD10-AM coding scheme to summarise the injury(s)/disease(s) of patients. Among other changes, this meant the codes used to describe the external causes of injury (e.g. fall, motor vehicle traffic accident) changed significantly. Non-fatal injury data in NIQS is based on ICD-9 prior to July 1999, and ICD10-AM from then onwards. Fatal injury data is based on ICD-10 from January 2000.

Trends

Interpreting trends over time should be done cautiously as changes may reflect coding practices, changes in hospital admission or treatment policies (service delivery issues), or funding, and not necessarily a change in the population incidence of a particular cause of injury.

Population and injury/fatality rates

Injury/fatality rates from 1991 onwards are calculated based on Statistics New Zealand estimated resident population counts as of 30 June each year, for each region, age group, and gender. “The estimated resident population is based on the census usually resident population count, updated for residents missed or counted more than once by the census (net census undercount); residents temporarily overseas on census night; and births, deaths, and net migration between census night and the date of the estimate” (Statistics New Zealand).

Population estimates are adjusted after each census to bring the preceding intercensal estimates back in line with the latest census population counts. Injury/fatality rates are not calculated for fewer than 5 cases, regardless of population size, or where the specific sub-population is effectively zero.

This information sheet is intended to be used as a guide only. If you would like further information or assistance in interpreting the statistics on our website, our factsheets, NIQS, or if you have any other questions, please contact gabrielle.davie@otago.ac.nz.