4.2.1 The child health check—background and context
4.2.2 Identifying the checked and non-checked groups—data sources and analyses
4.2.3 Coverage of the child health checks
This section focuses on the coverage of the child health checks, and examines the characteristics of children who received child health checks and those who did not. The purpose is to determine whether the children who got checked were a representative sample of the total children eligible for the check, or if they were a group of children who were otherwise not accessing health services.
Like MBS 708, the NTER child health checks were an assessment of the patient’s health and physical, psychological and social function. The checks also considered whether preventive health care and education should be offered to the patient to improve their health and physical, psychological and social function.
The MBS advises that a health assessment should generally be undertaken by the patient’s usual doctor and it should not take the form of a health screening service (MBS note A32, A33). As part of the CHCI, in communities where there was no resident doctor or the resident doctor declined to participate, this condition could not be met. In those cases, NTER Australian Government taskforce medical and nursing staff completed checks. These practitioners were mostly from interstate, often with little experience in Aboriginal and/or child health.
NTER taskforce doctors and nurses received a two-day orientation before travelling to communities, where they were expected to assess children’s health, make or arrange any necessary interventions and referrals, and document a ‘simple strategy for the good health of the patient’ (MBS note A33). They were required to do this while often having little knowledge and experience of cross cultural issues, the geographical isolation and general lack of health services, a child’s family and living circumstances or providing medical care to children with chronic infectious conditions and high rates of co-morbidity.
The extent to which the child health checks would have met the criteria of a quality child health assessment or a quality child health screening program (NHMRC 2002b) are likely to have varied by community, depending on factors such as whether the health checks were carried out by a resident doctor or delivered by visiting teams.
- Child health check dataset
Health services recorded information from each child’s health check on a paper form. Seven different versions of the child health check form were used, with earlier versions more likely to be missing data. Changes to the form addressed deficiencies of the earlier versions by adding questions and improving the structure of a number of existing questions. The final version of the form was 15 pages long. The changes did create difficulties when comparing information across versions (AIHW and DoHA 2009). The majority (72 per cent) of child health check forms processed were on version 5 or 6 (AIHW andDoHA 2008a).
Valid child health check forms were returned for 10,605 children. This figure is used to estimate overall child health check coverage rates. However, 1,232 of the forms returned were non-standard. Excluding these children leaves 9,373 children whose demographic data and other health check information could be analysed.
The AIHW was responsible for data management, analysis and reporting of information collected as part of the CHCI. PHC services recorded and transferred data on follow-up chart reviews, audiology and dental services to the AIHW. The AIHW, along with DoHA, reported on the progress of the CHCI and the results of the child health checks and follow-up data collections at three points (December 2008, May 2009 and December 2009).12 Top of page
- The DHF Client Master Index
DHF’s Client Master Index (CMI) is an NT-wide database that assigns a unique seven digit number (the Hospital Registration Number or HRN) to any client accessing public health services within the NT. As part of this evaluation, the CMI was used to identify all children who were eligible, as residents of prescribed areas who were under 16 years of age, to receive a child health check. A total of 14,647 children in the CMI were eligible to receive a child health check and 10,605 children received a check. Of those 10,605 children, 8,415 could be matched to the list of eligible children generated from the CMI and were tagged as having received a check. A total of 2,190 children received a check but could not be linked to the CMI.
Once the children within the CMI had been ‘tagged’ as either receiving a child health check or not, this data was linked to other NT datasets (Hospital Morbidity, Midwives and Growth Assessment and Action) to assess any differences in the characteristics of the two groups. The diagram in Appendix E shows how the populations of 14,647 and 8,415 were derived, and provides a breakdown of the status of the 10,605 children (child health check dataset) as a result of the data link with the CMI. Of the 2,190 children who received a child health check but were either not matched in the CMI or were identified as not being eligible for a child health check (the difference between 10,605 and 8,415), 1,093 (49.9 per cent) were recorded as residents of non-eligible communities. Given the high mobility of the remote Aboriginal population, (Condon et al 2001) it is quite likely that a number of children who received a child health check were not permanent residents in an eligible community but were visiting the community at the time the health checks were delivered. The remaining 50.1 per cent included records with missing HRNs, children whose status was recorded as non-Indigenous, and children with either an unknown residential status or a residential status within another state.
- NT Hospital Morbidity dataset
This dataset contains NT public hospital information on episodes of hospitalisation. We used data on hospital admissions that occurred from 1 July 1991 to 30 June 2007 to capture all pre-child health check admissions among the eligible population, and matched 12,438 children to the 14,647 children identified as eligible for a check in the CMI (84.9 per cent match).13 Compared to the unmatched population (those for whom we could find no hospital record), the matched population was less likely to be aged under one year and more likely to live in remote regions (Barkly, East Arnhem, Alice Springs Rural and Darwin Rural).
The 12,438 children whose records were matched had a total of 46,042 hospitalisation episodes. As well as rate of hospitalisation, we looked at the demographic profile of the child health check eligible population who had been hospitalised and data on the reason for hospital admission.
- NT Midwives dataset
This dataset is a census of all births in the NT and includes information about antenatal care received, maternal health, pregnancy, labour and childbirth and perinatal health. We used data on births of Aboriginal children from 2000–01 to December 2008. We used 2000–01 as a cut-off point because of the low coverage of HRNs within the dataset for earlier years. We matched 5,743 children to the 8,133 children identified as eligible for a check in the CMI who had been born since the cut-off date (70.6 per cent match). We used the resulting data to compare the birth weights of these checked and nonchecked children.
- Growth Assessment and Action (GAA) dataset
The ongoing GAA program, delivered by the NT Government, aims to improve the growth and nutritional status of children 0–4 years of age who live in remote areas.14 Data collected as part of the GAA includes measurements of children’s weight, height/length and haemoglobin.
We used data from DHF service providers for the period 1992–93 to 2008 and matched 7,058 children to the 14,647 children identified as eligible for a child health check in the CMI. These children had a total of 41,054 GAA attendances. We also used data from all service providers (DHF and ACCHOs) for the period 2003–2006 and identified a total of 10,852 GAA attendances from an unknown number of individual children.
We were unable to get satisfactory access to the GAA dataset to undertake analysis of the specific items in the dataset.
Other sourcesThe evaluation also uses data on community population estimates to assess rates of child health check coverage in the five case study communities. We used this information to validate the coverage estimates in the CHCI progress reports.
Finally, the evaluation draws on qualitative information collected from the five evaluation case study communities to help understand why some children did not get a child health check.
Number of children who had a CHC(a)
Proportion of eligible population who had a CHC
|Analysis of CMI linked data(b)||8,415||14,647||57.5%|
|AIHW analysis of CHCI data(c)||10,605||16,259||65.2%|
(a) - Figures exclude 4,000 checks provided under the MBS Item No. 708. These checks were not specifically funded through the NTER.
(b) - Estimated eligible population based on the Australian Bureau of Statistics definition of excluding urban areas and on the population aged 0–15 years at the child health check finishing time in each region.
(c) - Estimated eligible population based on Australian Bureau of Statistics Indigenous resident population figures for 2006 for children aged 0–15 years who live in communities and town camps covered by the CHCI.
Source: DHF analysis of CMI data; AIHW and DoHA (2009).
Child health check coverage varied markedly by region (Table 10) from 28.9 per cent in Darwin Urban to 63.4 per cent in Barkly and East Arnhem. AIHW and DoHA (2009) also report a variance in coverage by region of between 52.2 per cent in Darwin Rural to 77.4 per cent in Barkly/Katherine.
Table 10: Number and proportion of children who had a child health check and coverage by region (July 2007 to 30 June 2009)
Children who had a CHC
Coverage of CHCs
Proportion of eligible population who had a CHC
|Alice Springs Rural||1,926||22.9||3,202||60.1|
|Alice Springs Urban||96||1.1||256||37.5|
|East Arnhem District||2,045||24.3||3,227||63.4|
Source: DHF analysis of CMI data.
Child health check coverage rates for the five case study communities are shown in Table 11. Two estimates are provided. One is based on the child health check dataset and associated estimates from the Australian Bureau of Statistics/DoHA. The other is based on community-level data including community-level population estimates. This further highlights the challenge in estimating coverage rates.
Table 11 indicates that nearly all the children living in the smaller case study communities (A, C and E) received a child health check. Health service staff working in these communities at the time the checks were completed have confirmed this impression. In contrast, between 50–81 per cent of eligible children received checks in the larger case study communities (B and D).Top of page
Table 11: The proportion of children in the five case study communities who received child health checks according to different eligible child population estimates (July 2007 to 30 June 2009)
Population aged 0–15 years
Number of children who had a child health check
Australian Bureau of Statistics/DoHA estimates(a)
Alternative estimate based on community data
(a) AIHW analysis of CHCI data.
(b) DHF data as at 3 September 2009.
(c) Health service estimate dated 31 August 2008. Population estimate as at the commencement of the CHCI. The number is likely to include records for children who also use the services of a nearby Homelands health service (an Australian Government taskforce team is said to have travelled to one Homeland community only otherwise children living in Homelands communities were checked at the health service in Community B); children who periodically come and live in the community and use the service; visiting children; and children who have since moved to live in other communities.
(d) HIC MBS Item No. 708 Child Health Check.
(e) Health service staff in this community believe that an approximately 89 per cent coverage rate accurately reflects the proportion of children who received checks.
Age group and sexThe remainder of this section uses data from the CMI on child health check coverage, as it enables a comparison between those who received a child health check and those who did not. As was shown in Table 9, of the 14,647 children who were eligible to receive a child health check, 8,415 received a check and 6,232 did not.
Of the 14,647 children eligible to receive a child health check, the proportion of male to female children was similar—51.4 per cent male and 48.6 per cent female. The proportion of the 8,415 eligible male and female children who received a check was also similar—51.1 per cent male and 48.9 per cent female children.
Table 12 shows child health check coverage by age. Coverage was greatest amongst 2–9 year olds and lowest for children aged 14–15 years old and children who were under one year old. Among 15-year-olds, only 32.8 per cent received a child health check. These results support the experience of health service staff working in one of the case study communities—eligible children in the older age ranges, for a number of different reasons, did not wish to have a child health check. It may indicate that this type of annual check does not meet their needs and that perhaps older children are more likely to access health care when they see a reason to.
The low coverage in the under one-year-old age group may be due to the carers of these children choosing not to attend an additional check-up on top of other checks recently attended at their health clinic. We are unable to confirm this assumption.
Child health check
Reasons why children did not get checkedQualitative data collected at the community level provided information on reasons why some eligible children did not receive a child health check. Staff in the three smaller case study communities were confident that the only reason children might not have been checked in their communities was because they were not present in the community over the period the child health checks were completed. Therefore, Table 13 relates mainly to the two larger case study communities.Top of page
Table 13: Reasons why children in the case study communities did not get checked (February–April 2010)
Number of times reason given by informant groups
Guardian of child who did not get a CHC
Community member/ guardian of child who got a CHC
Health service staff(a)
|Child living at another community||3||-||5||1||9|
|Older children chose not to be checked (e.g. for fear of getting an injection, or did not consider themselves children)||-||2||5||-||7|
|Some people frightened||1||1||3||-||5|
|Guardian thought not needed as child checked recently||-||-||3||-||3|
|Guardian shy or shamed||1||1||-||-||2|
|Guardian sleeping or ‘lazy’ (drinking kava)||-||2||-||-||2|
|Did not want outsiders doing check||1||-||-||-||1|
|Mother away and child in the care of a relative at time of CHC||1||-||-||-||1|
Some guardians were frightened as they thought their children might be taken away by the visiting teams, and some informants spoke about parents
taking their children out bush or to outstations until the ‘threat’ was over. A number of staff and guardians related these feelings to the way the CHCI was first publicised, with the focus on sexual abuse and a lack of local level communication on what the initiative was about. It was not possible to quantify how many eligible children may have missed their NTER child health check for these reasons.
In one community staff doing the checks noted that fewer children in the 12–15 year age group received a check. The Australian Government taskforce teams working there made considerable efforts to recruit more children in this age range by, for example, having lucky draws.
This strategy had little or no success. Key informants offered a number of likely reasons why children in this
age range did not present for checks, including:
- adolescents were wary of the checks following the extensive media coverage and controversy over the initial intent to do sexual abuse checks
- the voluntary nature of the checks meant guardians and other responsible adults did not encourage or force children to have checks—many older primary school and adolescent children may have made their own decision whether or not to have a check
- some children had already participated in an annual screening program and did not want to be subjected to another check
- many older children would not wish to place themselves in a position where there was any chance of feeling self-conscious or embarrassed
- adolescent boys initiated as men into their customary group would be reluctant to present for what they may consider a program designed for children.
Hospital admissionsChildren who received a child health check were significantly more likely (P < 0.01) to have been hospitalised than children who did not. Age-adjusted hospitalisation rates (per 1,000 population) were 253 for the group of children who had a check and 207 for the group of children who did not. Actual (non-age-adjusted) rates for the two groups show that, over the period 1 July 1991 to 30 June 2007, children who had a check had an average of 3.4 hospitalisations, compared to 2.8 hospitalisations per child for children who did not have a check. Put another way, children who received a check comprised 57.5 per cent of the eligible population (Table 12) but accounted for 62.7 per cent of hospital admissions for the eligible population. The hospitalisation rate was significantly higher (P < 0.01) for all age groups among eligible children who received child health checks than those who did not receive checks (Figure 2).
Figure 2: Hospitalisation rates by age (at child health check) for eligible children who did and did not receive a child health check (1 July 1991-30 June 2007)
The number and per cent of hospitalisations by the International Classification of Diseases (ICD) chapter for eligible children who did and did not receive a child health check showed little difference between the two groups. Table 14 shows the results for the top four ranked conditions. Appendix F includes a full list of the ICD classifications.
Table 15 shows the number and percentage of avoidable hospitalisations for the top four ranked conditions. Children who received a check were
significantly more likely (P < 0.01) to have been hospitalised for dehydration and gastroenteritis and less likely to have been hospitalised for influenza and pneumonia than children who did not get checked.Top of page
Table 14: Top four hospitalisations by ICD Chapter for eligible children who did and did not receive a child health check (1 July 1991–30 June 2007)
Child health check
|Factors influencing health(a)||1,374||14.3||2,473||14.7||3,847||14.6|
|Total (top four)||9,640||100.0||16,785||100.0||26,425||100.0|
(a) - I CD Chapter XXI: Factors influencing health status and contact with health services. This provides for circumstances other than a disease, injury or external cause. For example, when a person encounters the health services for some specific purpose, such as to donate an organ, to receive a prophylactic vaccination or to discuss a problem which is in itself not a disease or injury. This chapter also provides for occasions when some circumstance or problem is present which influences the person’s health status but is not in itself a current illness or injury (WHO 2007, International Statistical Classification of Diseases and Related Health Problems, 10th Revision).
Source: DHF analysis of NT Hospital Morbidity dataset.
Table 15: Most common reasons for avoidable hospitalisations for eligible children who did and not receive a child health check (1 July 1991–30 June 2007)
Child health check
|Number||Per cent||Rank||Number||Per cent||Rank||Number||Per cent||Rank|
|Dehydration and gastroenteritis||496||24.1||3||1,170||33.7||1||1,666||30.1||2|
|Iron deficiency anaemia||587||28.5||2||926||26.7||3||1,513||27.3||3|
|Influenza and pneumonia||306||14.8||4||349||10.1||4||655||11.8||4|
|Total (top four)||2,061||100.0||3,471||100.0||5,532||100.0|
Source: DHF analysis of NT Hospital Morbidity dataset.We compared the four most common reasons for avoidable hospitalisation and the four most common ICD chapters for child health check eligible children between those living in Central Australia and Top End regions. Only small differences were observed.Top of page
The top four health conditions identified among child who received a child health check were (AIHW and DoHA 2009):
- untreated caries (40.2 per cent of children who had a child health check)
- history of recurrent chest infection (37.2 per cent)
- any skin problem (30.4 per cent)
- ear disease (30.0 per cent of children).
These conditions were usually chronic in nature and would alone, or in combination with other conditions, contribute to high hospitalisation and high avoidable hospitalisation rates among Aboriginal children in the NT.
Hospitalisation rates are difficult to interpret as reliable indicators of presence or severity of a condition in a population. Rates can be skewed by the incidence or prevalence of a condition in a community or defined population. For example, a population with a higher proportion of very young or very elderly people will generally have higher hospitalisation rates. Communities with easier access to hospital services will tend to be hospitalised more often, whereas a person living in a remote location with a comparable condition may be less likely to be hospitalised. Conversely, children living in remote communities may be more likely to be hospitalised for acute conditions because of concern about sudden deterioration in a remote community, whereas a child living locally may be observed and then discharged.
Birth weightData from the NT Midwives dataset was analysed to compare birth weight for children who had a child health check and children who did not. Birth weight is an important indicator of both the immediate health of the newborn and of the long-term risk of adult chronic disease. Low birth weight infants (less than 2,500 grams) have a much higher risk of poor health as adults (Li et al 2007).
Figure 3 shows the proportion of children with low birth weights who had a child health check and children who did not. Since 2002 the proportion of low birth weight children has been consistently lower among the population who got a check. With a total of 6,246 records over the eight-year period there are less than 100 children in each ‘underweight’ cohort which may explain why this result was not statistically significant (P = 0.08).
Figure 3: The proportion of eligible children who had a birth weight <2500g who did or did not have a child health check (from 2000–01 to 2007–08)
Source: DHF analysis of NT Midwives dataset.[D]
Participation in the GAA programAs discussed in Section 4.2.2 we used two sets of data from the GAA program—one set covering attendance on GAA programs delivered by DHF service providers from 1992–93 to 2008 and one set on attendance at GAA programs delivered by both DHF and ACCHOs for the period 2003–06.
Data on the GAA program delivered by DHF service providers shows that children who received a child health check were significantly more likely (P < 0.01) to have participated in the GAA program than children who did not get a check. Table 16 shows that eligible children who received a child health check participated in the GAA program approximately twice as often as eligible children who did not receive a child health check (average number of attendances of 3.5 compared to 1.9).Top of page
Child health check
|Number of GAA attendances||11,998||29,056||41,054|
|Number of children||6,232||8,415||14,647|
|Average number of attendances||1.9||3.5||2.8|
Source: DHF analysis of GAA dataset.This is an expected outcome as parents/guardians or older children willing to attend one type of check, may be pre-disposed to attending another.
We were unable to obtain the same level of data on GAA program attendance from all NT service providers. For the four years before the CHCI that data was available, Table 17 shows that children who received a child health check had an estimated average number of attendances of 3.9 compared to 1.7 for children who did not receive a health check.
Child health check
|Number of GAA attendances||2,684||8,168||10,852|
|Number of children(a)||1,558||2,104||3,662|
|Average number of attendances||1.7||3.9||3.0|
(a) T he number of children is calculated as a quarter of the total child health check eligible children, given the GAA data period is four years, while the child health check data covers 16 years.
Source: DHF analysis of the GAA dataset.
Children who received a check were significantly more likely to have had a previous hospitalisation than children who did not have a check. Data on low birth weight and on reasons for hospitalisation suggests that the population who received a child health check was no better (or worse) off than the 35 per cent who did not receive a check. This could mean that the population who received a child health check had a poorer health status than the population who did not. It may also mean, however, that the population who received a child health check had better access to hospital services, or had other more general ‘health seeking’ characteristics.
There is some evidence to suggest that the population who received a health check was more likely to attend checks. In other words, this was a population already accessing the health system through existing health screening programs available in the NT. This does not mean that these children had no unmet health needs, only that they were already accessing screening programs. The lower level of average GAA attendance for children who were not checked suggests that the CHCI was not an effective mechanism for reaching the unscreened population.
These analyses are influenced by the contextual variations and uncertainties and are limited by complexities inherent in identifying and comparing the checked and non-checked populations. The results from analyses done tend to indicate that apart from age and hospital admissions, the samples that could be compared are largely similar. As a result, these comparisons should be treated with a high degree of caution.
12 - Reports available at: http://www.health.gov.au/internet/main/publishing.nsf/Content/health-oatsih-nt
13 - The match rate is calculated as the proportion of the CMI eligible population whose record could be found.
14 - The GAA program is the previous name for the NT Government’s Healthy Under 5 Kids program. We use the name GAA in this report because it is commonly known and it remains the name for the data collection associated with this program.
15 - This percentage is based on the total number of resident children measured as part of the GAA program, as a proportion of the total number of resident children.
16 - Eligibility is based on residence in DHF’s CMI and the ABS definition of identifying the population in prescribed areas.
17 - Estimate is based on child health check dataset and pro-rata of the ABS population estimate of Indigenous children aged 0–5 years living in prescribed areas.