Incidence of gestational diabetes in Australia

This section outlines the data source and statistical methods for analysing the data presented in the Incidence of gestational diabetes in Australia web report and accompanying fact sheet.

The National Hospital Morbidity Database (NHMD) is the main data source used for this web report. The NHMD is a collection of episode-level records from the Admitted Patient Care National Minimum Data Set. It contains information on episodes of care for admitted patients in hospital, and includes demographic, diagnostic, outcomes, interventions and procedural information.

Further information about the NHMD 2016–17 can be found in the Admitted patient care 2016–17: Australian hospital statistics report (Appendix A).

Counting births in the NHMD

The number of new cases of gestational diabetes (numerator) was calculated based on the number of hospitalisations of females with a birth event code (ICD-10-AM code Z37) and coexisting diagnosis of gestational diabetes (ICD-10-AM code O24.4) in the year of interest. A single birth event code is entered for each woman, regardless of the number of times she is hospitalised during the same pregnancy or the number of babies born.

As a result, this method counts unique pregnancies affected by gestational diabetes resulting in a hospital birth.

The population at risk of gestational diabetes (denominator population) was based on the number of hospitalisations (pregnancies) with a birth event code (Z37) in the year of interest.

All pregnancies, regardless of outcome (that is, stillbirth or live birth) are counted by this method.

The numerator, pregnancies affected by gestational diabetes, were divided by the denominator population, total hospitalisations (pregnancies) to give the proportion of pregnancies affected by gestational diabetes in the year of interest.

Diabetes in pregnancy status

Two sets of diabetes codes in ICD-10-AM have been taken into account: diabetes ‘E-codes’ and diabetes in pregnancy ‘O24-codes’. The matrix below shows the method used to assign diabetes in pregnancy status to records from the NHMD, where a record also includes an outcome of delivery code (Z37). The method uses a hierarchy, whereby a record with any E-code is assigned to pre-existing diabetes in pregnancy status first, and the remaining records are assigned a status based on the diabetes in pregnancy O24-codes. Gestational diabetes is only assigned where an O24.4 code exists in the absence of any E-code.

Table 1: Matrix of ICD-10-AM codes for assigning diabetes in pregnancy status

		Diabetes (E-codes)
		Type 1 (E10)	Type 2 (E11)	Other/unspecified (E13 and E14)	No E-code
	Pre-existing type 1 diabetes (O24.0)	Pre-existing type 1 diabetes	Pre-existing type 2 diabetes	Pre-existing other/unspecified diabetes	Pre-existing type 1 diabetes
	Pre-existing type 2 diabetes (O24.1)	Pre-existing type 1 diabetes	Pre-existing type 2 diabetes	Pre-existing other/unspecified diabetes	Pre-existing type 2 diabetes
Diabetes in pregnancy codes (O24-codes)	Pre-existing other/unspecified diabetes  (O24.2 and O24.3)	Pre-existing type 1 diabetes	Pre-existing type 2 diabetes	Pre-existing other/unspecified diabetes	Pre-existing other/unspecified diabetes
	Gestational diabetes (O24.4)	Pre-existing type 1 diabetes	Pre-existing type 2 diabetes	Pre-existing other/unspecified diabetes	Gestational diabetes
	Diabetes in pregnancy , unspecified onset (O24.9)	Pre-existing type 1 diabetes	Pre-existing type 2 diabetes	Pre-existing other/unspecified diabetes	Diabetes in pregnancy, unspecified onset
	No O24-code	Pre-existing type 1 diabetes	Pre-existing type 2 diabetes	Pre-existing other/unspecified diabetes	No diabetes in pregnancy

Source: ICD-10-AM/ACHI/ACS, Ninth edition.

ICD-10-AM codes

The following table presents the variables and ICD-10-AM codes used to count the population of women giving birth (outcome of delivery), their diabetes status, treatment status and the effects of pregnancy, labour and delivery.

Table 2: Analysis variables and ICD-10-AM code

Definition	ICD-10-AM code
Outcome of delivery (birth event)	Z37
Diabetes in pregnancy status
Type 1 diabetes	E10
Type 2 diabetes	E11
Other specified diabetes	E13
Unspecified diabetes	E14
Pre-existing diabetes, type 1, in pregnancy	O24.0
Pre-existing diabetes, type 2, in pregnancy	O24.1
Pre-existing diabetes, type 2, other specified type, in pregnancy	O24.2
Pre-existing diabetes, type 2, unspecified specified type, in pregnancy	O24.3
Diabetes arising in pregnancy (gestational diabetes)	O24.4
Diabetes in pregnancy , unspecified onset	O24.9
Treatment of gestational diabetes
Insulin-treated	O24.42
Oral hypoglycaemic therapy	O24.43
Other (diet, exercise and lifestyle management)	O24.44
Unspecified	O24.49
Effects: outcomes and complications
Pre-existing hypertension complicating pregnancy, childbirth and the puerperium	O10
Pre-eclampsia superimposed on chronic hypertension	011
Gestational (pregnancy-induced) hypertension	O13
Pre-eclampsia	O14
Maternal care for excessive fetal growth	O36.6
Pre-term labour and delivery	O60
Failed induction of labour	O61
Single delivery by forceps and vacuum extractor; Multiple delivery, all by forceps and vacuum extractor	O81 and O84.1
Single spontaneous delivery; Multiple delivery, all spontaneous	O80 and O84.0
Induced labour	ACHI Block number 1334
Caesarean section (includes emergency and elective)	ACHI Block number 1340

Source: ICD-10-AM/ACHI/ACS, Ninth edition.

Age-standardised proportions

Direct age-standardisation was used in this report to remove the differences in age structure from the analysis, and highlight the contributions of diabetes to differences in the occurrence of effects of pregnancy, labour, and delivery between diabetes in pregnancy status groups.

Of all women who gave birth, those with gestational diabetes tend to be older than women with no diabetes in pregnancy. Aboriginal and Torres Strait Islander women who gave birth were younger than non-Indigenous or other Australian women who gave birth. So, age-standardising these groups removed the influence of older or younger maternal age, making comparisons of outcomes more valid.

Proportions have been calculated by dividing the number of cases of a particular outcome in a single diabetes in pregnancy status group (for example, gestational diabetes) by the total number of women (aged 15–49) who gave birth in hospital.

Age-standardised proportions have been calculated using the direct method, using the 30 June 2001 Australian female estimated resident population, based on the 2001 Census as the standard population.

Confidence intervals

In the Incidence of gestational diabetes in Australia web report and fact sheet, 95% confidence intervals were calculated around age-standardised proportions, to determine whether differences in pregnancy outcomes between diabetes pregnancy status groups were significant. A difference was deemed statistically significant if the 95% confidence intervals of the age-standardised proportions did not overlap. Confidence intervals have not been presented in the report.

Incidence

Incidence is the number of new cases (of an illness or event) occurring in a population during a given period. Incidence can be described as either a whole number or rate relative to the total number of people at risk.

Incidence should not be confused with prevalence, which refers to the total number or proportion of cases (of an illness or event) in a population at a given point in time.

In this report, incidence of gestational diabetes is reported by financial year—that is, the number of new cases from 1 July to 30 June in the year being reported.

Throughout this report, incidence is calculated and presented as a proportion of the age-standardised population.

For example, the incidence of gestational diabetes is calculated as a proportion of women aged 15–49 who gave birth in an Australian hospital.

Estimated resident populations

Population data were used to derive incidence of gestational diabetes. Population data that the AIHW holds are sourced from the Australian Bureau of Statistics (ABS), and updated as revised or new estimates become available.

All population estimates that the ABS currently produces are based on area of usual residence. These estimated resident populations are derived from the ABS Census of Population and Housing, and adjusted for deaths, births and net migration. The estimated resident populations used in this report are based on the population estimates for 30 June 2017.

Remoteness

Comparisons of regions in this report use the ABS Australian Statistical Geography Standard (ASGS) 2011 Remoteness Structure, which groups Australian regions into 6 remoteness areas. The 6 remoteness areas are Major cities, Inner regional, Outer regional, Remote, Very remote and Migratory. These areas are defined using the Accessibility/Remoteness Index for Australia (ARIA), which is a measure of the remoteness of a location from the services that large towns or cities provide.

Accessibility is based on distance to a metropolitan centre. A higher ARIA score denotes a more remote location. The category Major cities includes Australia’s capital cities, with the exceptions of Hobart and Darwin, which are classified as Inner regional. Note that Remote and Very remote areas have been combined in this publication, and the sixth remoteness area, Migratory, is excluded.

Further information on the ASGS, see Australian Statistical Geography Standard is available on the ABS website.

The coverage of the NDSS may be lower in Remote and very remote areas or across states and territories with large remote communities, which may influence estimates on the number of people with insulin-treated diabetes in these areas on the NDR. This may in part be due to the distribution of NDSS Access Points which assist in delivering support services and products to people with diabetes in all states and territories. These Access Points may be limited in rural Australia and unavailable in remote communities, with other programs sometimes being available in these areas to assist with the purchase of diabetes-related products.

Socioeconomic group

Socioeconomic classifications in this report are based on the ABS Index of Relative Socio-economic Disadvantage (IRSD). Geographic areas are assigned a score based on social and economic characteristics of that area, such as income, educational attainment, public sector housing, unemployment and jobs in low skill occupations.

A low score means an area has, on average, more low-income families, people with less training, and higher unemployment, and may be considered disadvantaged relative to other areas with higher scores. High scores reflect a relative lack of disadvantage, rather than advantage, and the IRSD relates to the average disadvantage of all people living in a geographical area. It cannot be presumed to apply to all individuals living in the area.

For the analysis in this report, the population is divided into 5 socioeconomic groups, with roughly equal populations (each around 20% of the total), based on the level of disadvantage of the statistical local area of their usual residence. The first group includes the 20% of the population living in areas with the highest levels of relative disadvantage (referred to as Group 1, most disadvantaged), while the last group includes the 20% of the population living in areas with the lowest levels of relative disadvantage (referred to as Group 5, least disadvantaged).

Further information about the IRSD values used in this report are based on the 2011 Census. Further information is available on the ABS website.