Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

8. Health service provision and perceptions of service adequacy

S.R. Walker and I.H. Burnley

It has been proposed that of all services access to medical facilities is of the greatest concern to residents in remote areas (Mosely 1979). If certain physical or environmental conditions are causally related to health, then remote towns in arid lands might be expected to be faced with an even greater need for services, although care must be taken in assuming that climatic factors are necessarily deleterious. Within Australia, however, there is further cause for concern, in that many of the remote settlements have a responsibility to provide services to help redress the health disadvantage with which the Aboriginal people are faced (see also Burnley, chapter 9). Morbidity rates and infant mortality rates (though falling) are higher among the Aboriginal than non-Aboriginal population (tables 8.1 and 8.2), and these people comprise approximately 25 per cent of the population of the Northern Territory. The hospitals in central Australia have large catchment areas with difficult environmental conditions in relation to transport especially and "face many problems in the delivery of health care not normally confronted in Victoria or New South Wales" (Northern Territory Department of Health Submission 1979).

TABLE 8.1. Patients Treated, Northern Territory Hospitals 1977/78: Rates Per Capita and Length of Stay

Hospital Rate per 1,000 Population Total
Aboriginal Non Aboriginal
Darwin and East Arm 243 192 199
Katherine 229 265 250
Alice Springs 318 211 248
Tennant Creek 333 173 213
Gove 169 163 165
Total 255 197 211
Patient's average stay:      
Total (days) 11.8 6.2  

Source: Northern Territory Department of Health Submission, 1979. p. 26.

The first question to be examined is whether, in absolute terms, the remote settlements in the arid zone are in fact more disadvantaged in the availability of hospital and medical care than their coastal, humid zone counterparts. On a per capita basis, at least, if these towns are not adequately provided with facilities, then one could state at the outset that they are indeed disadvantaged. To investigate the relationship between distance from metropolitan centres and assumed service advantage, correlation coefficients were calculated. A number of indicators of health service provision were selected: the ratios to total population of doctors, nurses, and hospital beds, and the ratio of maternity beds to the number of births. These were related to three measures of remoteness: distance from the nearest state capital, distance from the nearer of Sydney or Melbourne, and distance from the nearest town in the study, in a similar manner to the discussion about education services in the previous chapter.

The analysis (tables 8.3 and 8.4) shows that remote areas are not necessarily disadvantaged on the measures of health service provision. On average, service provision does not differ significantly from that for all cities, although there are marked variations betwen certain towns (table 8.4). In the arid zone, Alice Springs, for example, tends to be relatively well provided with health facilities compared to the national average, whereas Mt. Isa is notably under-provided. These variations and indeed the overall level of provision may be partly explained by the size and nature of the catchment areas of these centres. Alice Springs' services treat a relatively large number of patients, especially Aboriginal people from fringe camps and surrounding areas. The average length of stay in hospital for these people is also longer compared to that required for urban non-Aboriginal residents (as indicated in table 8.1).

On a regional basis, certain arid zone areas may even seem to be advantaged, rather than disadvantaged, in terms of available health services. Figure 8.1 illustrates this point: "acute" hospital beds are relatively better provided in central Australia areas (as well as in the very inner city districts of metropolitan areas). Central-west Queensland, Wyndham region, and Kalgoorlie region are all arid or semi-arid zone regions relatively well supplied with hospital beds, and the Northern Territory is also above average on this count. Reference back to the Royal Flying Doctor Service map also indicates the extensive coverage of the arid regions by medical services, which may take less time for delivery to a point 200 kilometres away than may occur in some outer metropolitan locations.

TABLE 8.2. Northern Territory Infant Mortality Rates, 1965-1979

Year Aboriginal Non-Aboriginal Total N.T. Alice Springs
and Barkly region
(Aboriginal only)
Total Australia
1965 142.7 30.1 80.0 196.9 ....
1966 147.3 20.8 78.0 204.3 18.7
1967 100.0 28.4 59.9 92.3 18.3
1968 80.9 29.3 51.2 90.1 17.8
1969 94.8 17.7 46.8 88.8 17.9
1970 115.1 26.4 58.2 181.8 17.9
1971 142.9 19.8 60.2 219.3 17.3
1972 87.0 17.0 39.1 117.0 16.7
1973 79.7 14.3 34.8 107.1 16.5
1974 55.6 18.8 29.9 59.7 16.1
1975 50.1 22.1 33.5 54.2 14.3
1976 52.8 16.1 27.7 42.7 13.8
1977 74.6 9.9 30.5 73.8 12.5
1978 48.1 9.7 22.4 38.9 12.2
1979 45.0 9.8 21.1 54.6  

Source: Northern Territory Department of Health, Supplement to Bulletin Issue No. 26 (1979).

TABLE 8.3. The Relationship between Isolation and Health Service Provision

Health service   DIST1 DIST2 DIST3
HEALTH1 (Doctors:population) .024 -.026 -.047
HEALTH2 (Nurses:population) -.004 -.073 -.068
HEALTH3 (Hospital beds:population) -.188 -.077 -.062
HEALTH4 (Maternity beds:births) -.133 .065 .064

For a key to DIST1-DIST3 variables see table 8.5.

TABLE 8.4. Comparisons in the Provision of Services and Infant Mortality

  Overallč ± s National average Alice Springs Katherine Tennant Creek Kalgoorlie Mt. Isa Broken Hill
HEALTH1 1.1 (0.7) 3 2 1 1 1 1
HEALTH2 9.1 (3.3) 14 10 8 15 5 10  
HEALTH3 15.3 (9.1) 16 19 13 26 6 13
HEALTH4 2.0 (1.2) 2 2 2 4 1 2
Infant mortality rateČ 159.8 (73.6) 286 528 441 205 183 112
POPCH 5.5 (11.0) 27 24 25 -8 .. -7

1 Mean for 118 urban areas, except for IMR (data not obtained for Victoria). Key to variables: see tables 8.3 and 8.5.
2 Infant mortality rate here = deaths per ten thousand live births.

FIG. 8.1. Regional Acute Bed to Population Ratios, Australia 1972-1973

However, such apparent "oversupply" in arid zone areas must be reviewed in the context of the very remoteness of these areas. It is important to locate facilities within a reasonable distance of users, and this "necessitates the location of hospitals in places where they may not be utilised continuously by the local population at a reasonable level of occupance" (Northern Territory Department of Health Submission). Hospital bed occupancy rates tend to be lower in remote areas (under 60 per cent) than in metropolitan centres (over 70 per cent, Hospitals Commission 1974), but to suggest that these areas had a surplus of hospital beds would be erroneous. Basic services have to be provided to care not only for a constant demand but also seasonal variations and peak admission levels, which could adversely affect hospitals in isolated areas that do not have nearby back-up facilities. A simple bed-population ratio cannot be used as a standard for measuring the adequacy of need in different areas. This is especially so in Alice Springs, where a population of 150,000 visitors inevitably makes casualty and other demands on hospital services.

Moreover, it cannot be assumed that the same proportions of different types of services are required in arid zone areas as in other regions. First, selective migration has meant that certain services need to be provided at relatively higher rates, while others can be reduced. Figure 8.2 demonstrates the imbalance in the age structure of the Northern Territory: with the large number of young mothers, more obstetric and gynaecology facilities have to be provided, and conversely, fewer geriatric facilities are required. Such differences in facility provision do in fact occur (Hospitals and Health Services Commission 1974, p. 61), and special units are established. The Alice Springs Child Care Unit assists in convalescence and treatment of malnourishment of children from the whole region (Northern Territory Department of Health, Southern Region Annual Report, 1978-1979).

FIG. 8.2. Population Comparison by Grouped Ages

The special medical needs of the local population also require different emphasis on specialist services (see Burnley, chapter 9). Among Aboriginal people, for example, infective and parasitic diseases and diseases of the respiratory system are the most common reasons for hospital admittance, apart from accidents, pregnancy, and childbirth (Parliamentary Report, p. 177) Hospitals in arid regions of Australia, therefore, are required, to a certain extent, to concentrate specialist resources on such difficulties, leading to relatively large expenses on a per capita basis (Submission, p. 29). A further effect on high morbidity levels is also alcoholism, the rate of alcohol consumption in the Northern Territory being twice as high as that for the rest of the country (Submission, p. 30, and Burnley, chapter 9).

Given these more difficult environmental conditions, it might therefore be expected that residents in remote areas are in need of more health facilities per capita than those in less arid areas. Physical environmental conditions have been proposed as one of the most significant factors in Aboriginal morbidity rates (Parliamentary Report, p. 15): infections, diseases of the skin, respiratory diseases, and trachoma are environmentrelated (see also Burnley, chapter 9). It could therefore be proposed that in areas experiencing poorest environmental conditions, health services would be required in excess of an "average" level of provision. The next part of the analysis, therefore, was to ascertain whether there was any significant relationship (in a positive direction) between levels of service and poorer environmental conditions. Our meaning of poor quality environmental conditions will be explained below.

To approach this analysis, a number of variables were selected to measure different aspects of the physical and social environment: housing conditions (2 variables), availability, utilities /2), remoteness (3), population (3), socioeconomic status (3), and a measure of infant mortality. A regression analysis was undertaken for 69 urban areas for which this information was obtainable to provide an indication of which variables accounted for most of the variation in (a) ratios of doctors to population and (b) ratios of hospital beds to population. Tables 8.5 and 8.6, correlation coefficients (r), also provide a measure of the strength of the relationship between each variable separately and the health-provision variables.

The data relations imply that health provision based on these two indices is more related to demographic and socioeconomic factors than to environmental factors, such as housing conditions or provision of utilities. In both regression models, four such variables together account for at least 33 per cent of the variance, with all other variables contributing at most an extra 16 per cent. The remoteness variables notably play a minor role-once the effect of other variables has been taken into account, the distance factors add very little to the explanation. This suggests that remoteness itself is not the critical problem but the economic and demographic conditions that may be associated with it.

TABLE 8.5. Doctors Per Capita Ratios: Multiple Regression

Variableč Multiple R r Overall F SignificanceČ
POPCH .378 .143 .378 11.2 .001
INC1 .458 .210 .335 8.8 .001
INC2 .533 .284 -.101 8.6 .001
POP76 .577 .333 .342 8.0 .001
UTIL1 .600 .360 .309 7.1 .001
HOUSE2 .651 .424 .215 7.6 .001
UTIL2 .662 .437 -.075 6.8 .001
D IST1 .671 .451 .159 6.2 .001
IMR .686 .470 .024 5.8 .001
DIST3 .693 .481 .066 5.4 .001
HOUSE 1 .695 .484 .174 4.9 .001
DIST2 .697 .486 .056 4.4 .001
UNEMP .698 .487 .032 4.0 .001
ABOR .698 .487 .150 3.7 .001

1 Variable list (1976data):

ABOR Percentage of the population Aboriginal or Islanders
POP76 Total population, as recorded 1976 census
POPCH Population change 1971-1976
DIST1 Distance by road to the nearest state capital
DIST2 Distance by road to the nearer of Sydney or Melbourne
DIST3 Distance by road to nearest town in the study ( 1 18 towns)
HOUSE1 Percentage of total occupied dwellings with three or less rooms
HOUSE2 Percentage of total occupied dwellings improvised or mobile
INC1 Percentage of households with incomes of A$15,000 or more

INC2 Percentage of households with incomes of A$6,000 or less
UNEMP Percentage of the labour force unemployed
UTIL1 Percentage of total occupied private dwellings with no piped water
UTIL2 Percentage of occupied private dwellings with no sanitation
IMR Infant mortality rates; infant deaths: live births ratio (1974-1978)

2 69 cases (New South Wales, Queensland, Northern Territory)

TABLE 8.6. Hospital Beds Per Capita Ratios: Multiple Regression

Variable Multiple R r Overall F Significance
POP76 .450 .202 -.450 17.0 < .001
POPCH .538 .290 -.430 13.5 < .001
UNEMP .571 .326 -.302 10.5 <.001
INC1 .586 .344 -.155 8.4 <.001
UTIL1 .602 .362 .035 7.2 < .001
UTIL2 .616 .379 -.165 6.3 < .001
IMR .622 .386 -.008 5.5 < .001
HOUSE2 .627 .394 -.061 4.9 <.001
INC2 .630 .396 .055 4.3 <.001
ABOR .631 .397 .021 3.8 <.001
DIST1 .631 .398 -.146 3.4 <.001
DIST2 .632 .399 .029 3.1 <.001
DIST3 .632 .399 .026 2.8 < ,001

1 See table 8.5.

However, the precise relationship between the medical and other variables differs in each of the models. For example, the analysis suggests that doctors are relatively well provided in urban areas of rapid growth, large size, and high-income households; whereas hospital beds per capita are better provided in smaller centres, of slower growth, but with relatively low levels of unemployment. This implies a difference between the private and public provision of services: the location of government hospitals is not subject to the attractions of financial gain or attractive living environments one presumes. Given this, it is perhaps surprising to find that the provision of doctors is relatively high in Alice Springs, which is an arid zone urban centre, and also in the Northern Territory in general. In a sense, this can also be explained in the context of public service provision. In Alice Springs, there is now only one private doctor, yet substantial medical assistance must be provided to the area. Hence, the remaining medical staff are employed in public facilities. This at least means a favourable relationship between certain environmental indicators of need and the doctors-per-capita ratio.

With the latter exceptions, it is evident that the supply of health services does not necessarily match the "potential need."* Provision is dependent upon a number of social and development factors. In newly established areas there may be some time lag in establishing hospitals (regardless of the centre's location), but individual doctors may be mobile, prepared to move to growth areas but less inclined to become established in low-income or small centres. Hence, whether a town in an arid area can attract personnel may depend much more on social, financial, and political factors than on the physical environment. (As indicated below, the Northern Territory is attempting to attract staff by offering a number of financial benefits.)

Tables 8.5 and 8.6 also indicate that not only is service provision not strongly related to many of the environment indicators that may affect health, but also that supply is not significantly associated with the index of health that was included in the model, infant mortality rates (IMR), and this definitely was not closely associated with service provision. Thus, while some arid zone towns were characterized both by high infant mortality rates and the services to meet such potential demand (e.g., Alice Springs), other towns with relatively high rates of infant mortality were not so fortunate (Tennant Creek). In general, however, it was the more remote and arid zone towns that experienced higher infant mortality rates.

TABLE 8.7. Determinants of Infant Mortality Rates

Variableč Multiple R r Overall F Significance
HOUSE 1 .574 .329 .574 32.9 < .001
POP76 .636 .405 -.158 22.5 <.001
HEALTH4 .659 .435 -.022 16.7 <.001
ABOR .672 .452 .486 13.2 <.001
INC1 .688 .473 .158 11.3 < .001
DIST2 .696 .484 .348 9.7 <.001
DIST1 .712 .507 .277 8.9 < .001
UTIL1 .718 .516 .559 8.0 < .001
HOUSE2 .729 .530 .457 7.4 <.001
HEALTH2 .737 .543 -.028 6.9 <.001
HEALTH3 .739 .546 -.008 6.2 <.001
HEALTH1 .740 .548 .024 5.7 <.001
DIST3 .741 .550 .342 5.2 <.001
UTIL2 .743 .552 .079 4.8 < .001
UNEMP .744 .554 -.087 4.8 <.001
INC2 .744 .554 -.222 4.0 <.001

1 See table 8.5.

TABLE 8.8. Northern Territory Interstate Specialist Referrals, 1979/80

Reason for referral (principal reason) Aboriginals Non Aboriginals
Orthopaedic 11 100
Cardiology 4 61
Ophthalmology 4 55
Plastic surgery 1 32
Oncology 1 32
Neurology 3 47
Radiology 0 30
Urology 4 28
Gynaecology 0 28
Total interstate (all reasons) 39 578
Total interstate escorts 23 166
Total intrastate referrals 327 532
Total intrastate escorts 60 87

Source: Northern Territory Department of Health, Darwin.

The importance of environmental effects (including remoteness) rather than the provision of services in influencing health measures can be gleaned from table 8.7, which shows the dependence of infant mortality on a number of variables. Variables that relate to environment (poor utilities provision, distance, housing) are all significantly correlated with infant mortality rates, whereas those concerning services (HEALTH 1-4) are not. Again, higher infant mortality rates in areas with large percentages of Aboriginal population show the poor standard of health experienced in outback settlements with inadequate housing, water supply, and sanitation (see also Moodie 1969; Parliamentary Report No. 60, 1979; and Burnley, chapter 9). It is important to point out that improvements to Aboriginal health are dependent on many factors-not only on the remoteness of many of the communities (as is evidenced by relatively high rates even in large urban areas). Indeed, infant mortality rates of non-Aboriginal people in remote towns are frequently even lower than the national average (table 8.2).

The preceding discussion has shown that remoteness per se is not necessarily a disadvantage in terms of the number and absolute level of health services provided. However, such quantitative analysis omits an important item: the quality of the services provided. For example, it has already been mentioned that in the case of Alice Springs, limitations on finance and the relatively large size of the town and region mean that only a limited range of specialist services can be provided. What this means to residents is a long journey to Adelaide or elsewhere for certain services (table 8.8). Furthermore, not only is the patient disadvantaged but this adds to hospital costs since travel expenses (both for referrals and escorts) have to be paid. Table 8.8 shows only those numbers who are officially referred elsewhere-the rest have to pay their own expenses. it is also interesting to note in that table the relatively small numbers of Aboriginal people who are referred. Part of the explanation may lie in the types of specialist requirements sought-cardiology and plastic surgery, for example; other reasons relate to cost of transport, personal prejudice, as well as cultural factors.

However, the difficulty of providing more specialist services in arid zone areas is not easily resolved. There are two major obstacles: the first relates to economies of scale.

Because many towns are small, it is uneconomic to provide all the services residents may require. At the extreme, the Hospitals and Health Services Commission (1974) quote the "usual recommendation" for organ transplant and renal units to be one unit per 2 million people, radiotherapy, one unit to 2 million, and neurosurgery, one unit to 0.5-1.0 million (p. 232). Even costs for specialist services that are provided in arid zone towns at the moment are "more expensive on a per capita basis" (Northern Territory Department of Health Submission).

TABLE 8.9. Annual Employment Turnover of Hospital Staff in the Northern Territory

Hospital staff Northern Territory Alice Springs Tennant Creek
Turnoverč (NoČ) Turnover (No) Turnover (No)
Medical 90 (84) 113 (23) .. (2)
Nursing 85 (678) 102 (195) 140 (15)
Orderlies 67 (57) 105 (21) 200 (1)
Paramedical & technical 64 (59) 71 (17) 200 (1)
Total (including clerical & industrial) 75 (1,472) 101 (395) 95 (43)

1 Turnover = Number of cessations (including transfers) x 100/average number employed.
2 Number employed. Source: Northern Territory Medical Service Bulletin 18,1977.

The second major obstacle to improvements in specialist services is the problem of attracting staff. As table 8.9 illustrates, staff turnover is exceptionally rapid in the Northern Territory, and especially in the arid land centres. A turnover of over 100 per cent per annum in most clinical employment categories (including doctors) in Alice Springs is astonishing. In 1978/79, 74 per cent of new nursing staff failed to complete their 12-month contract. Apart from all the difficulties such staff changes cause within the hospital, there is the continuing problem of finding staff replacements. Positions are frequently left vacant until a replacement can be found who is willing to live in the town, and various inducements have to be offered to attract personnel. The higher salaries and auxiliary benefits increase total hospital expenditure. In 1978/79, the Alice Springs Hospital reported that "the positions of Radiologist and Paediatric Registrar are vacant .... The revised housing policy has made recruitment of married R.M.O.'s difficult .... The retention of nursing staff improved although the turnover rate, 82%, is still very high" (Northern Territory Department of Health, Southern Region Report, p. 16). Staff are offered higher commencing salaries, a district allowance, additional annual leave, recreation leave fares, and recruitment travel and removal allowances, yet still it is difficult to attract permanent staff.

It is possible, however, that these staffing difficulties do not prevent the continued efficiency of the hospital. In Alice Springs new initiatives in many areas of health care delivery have been taken in the last two years and new or additional appointments in ophthalmology, psychiatry, gynaecology, and medical records have increased the capacity of the hospital in these areas.

Field studies in Alice Springs in 1980 found that the lack of quality of medical services in the town (rather than the absolute lack of facilities) was identified strongly by residents as a major drawback of living in Alice Springs (Burnley, chapter 9). Specific problems residents mentioned are listed in table 8.11. That so many residents perceive such problems with the health service is of importance. One improvement which would help to alleviate some of the criticism would be the addition of another private doctor in the town to offer those who dislike the present dominantly hospital-based system a viable alternative. At present the single private doctor cannot meet apparent market demand alone.

Even to maintain the standard of medical service provision at its present level requires greater financial resources, on a percapita basis, than other areas of Australia. For example, it has been estimated for the Northern Territory as a whole that expenditure on hospitals is twice as much per capita as that for New South Wales or Victoria, and expenditure on other health services five times as much. This situation is attributed to higher Aboriginal bed usage and need for services; lack of private hospitals; higher cost structure, both in providing labour and goods; diseconomies of scale and dispersion of facilities; lower proportion of private medical and dental practitioners compared to the other states (Northern Territory Department of Health Submission). Many of these factors relate to such conditions as remoteness of the Territory, and even with additional finance it is apparent that some service disadvantage will remain.

TABLE 8.10. Adequacy of Treatment by Nature of Illness, 1980

Persons with illnesses Adequate treatment Inadequate treatment Total persons
Heart disease 1 - 1
Respiratory illness 12 5 17
Arthritis and back pain 5 8 13
Intestinal illness 4 2 6
Psychological illness 3 1 4
Cerebral illness 1 4 5
Hay fever 2 2 4
Accident 1 - 1
Eye disorders 3 - 3
Others 5 4 9
Short-term illnesses 13 7 20
Total 50 33 83

Source: Sample Survey, August 1980.

TABLE 8.11. Adequacy of Medical Help in Alice Springs by Problems in Getting Care

Problems Frequency of responses
Long waiting time at hospital 57
Lack of specialist 14
Lack of private doctor 49
Lack of transport to hospital 4
Poor quality of medical practitioners 27
Refusal of doctors to visit home 6
Unavailability of medical help after hours 11
Total 168

Source: Sample Survey, August 1980.


Contents - Previous - Next