Article Text


What is the difference between comprehensive and selective primary health care? Evidence from a five-year longitudinal realist case study in South Australia
  1. Fran Baum1,
  2. Toby Freeman1,
  3. Angela Lawless1,
  4. Ronald Labonte2,
  5. David Sanders3
  1. 1 Southgate Institute for Health, Society and Equity, Flinders University, Adelaide, South Australia, Australia
  2. 2 Epidemiology and Community Medicine, University of Ottawa, Ottawa, Ontario, Canada
  3. 3 School of Public Health, University of the Western Cape, Cape Town, South Africa
  1. Correspondence to Professor Fran Baum; fran.baum{at}


Background Since the WHO’s Alma Ata Declaration on Primary Health Care (PHC) there has been debate about the advisability of adopting comprehensive or selective PHC. Proponents of the latter argue that a more selective approach will enable interim gains while proponents of a comprehensive approach argue that it is needed to address the underlying causes of ill health and improve health outcomes sustainably.

Methods This research is based on four case studies of government-funded and run PHC services in Adelaide, South Australia. Program logic models were constructed from interviews and workshops. The initial model represented relatively comprehensive service provision in 2010. Subsequent interviews in 2013 permitted the construction of a selective PHC program logic model following a series of restructuring service changes.

Results Comparison of the PHC service program logic models before and after restructuring illustrates the changes to the operating context, underlying mechanisms, service qualities, activities, activity outcomes and anticipated community health outcomes. The PHC services moved from focusing on a range of community, group and individual clinical activities to a focus on the management of people with chronic disease. Under the more comprehensive model, activities were along a continuum of promotive, preventive, rehabilitative and curative. Under the selective model, the focus moved to rehabilitative and curative with very little other activities.

Conclusion The study demonstrates the difference between selective and comprehensive approaches to PHC in a rich country setting and is useful in informing debates on PHC especially in the context of the Sustainable Development Goals.

Statistics from

Strengths and limitations of this study

  • This study provides a unique empirical examination of differences between comprehensive and selective primary healthcare in practice.

  • This difference is crucial and needs to be defined very clearly when health systems are being reoriented to Primary Health Care.

  • The comprehensive model as envisaged by these services was limited by the relative limited scope of action on social determinants of health.

  • This study is limited by the scope for generalisation from five case studies.


The initial WHO 1978 vision of Primary Health Care (PHC) was comprehensive, viewing health services as part of a new international economic order that would benefit all nations especially low income and groups living in disadvantage, that would encourage democratic participation in health, and help improve social and environmental contexts that create disease and risks for disease.1 Health services were to be multidisciplinary, attuned to local need, and emphasise disease prevention and health promotion. This comprehensive vision was overtaken by a pragmatic call for a more selective approach, although originally considered to be temporary until developing countries could afford a more comprehensive approach, which minimised the broader social change ambitions of the original vision, marginalised preventive and promotive actions, and emphasised responses to specific diseases or narrowly defined health outcomes.2 Although the WHO recommitted to PHC in 20083 and the Commission on the Social Determinants of Health4 endorsed PHC as the cornerstone of a health system and a strategy for taking action on social determinants of health at a local level, selective PHC has dominated health system reforms in most low-income and middle-income countries, abetted by growth in vertical (disease-specific) global health funds.5 Most empirical work on PHC implementation has come from low-income and middle-income countries, with few systematic studies of comprehensive PHC from high-income countries. This paper reports on an Australian study which tracked a shift from comprehensive to selective PHC and has enabled development of a program logic description of the two forms of PHC. We do not claim that our typification of selective PHC in this study necessarily captures all interpretations (past and present) of this form of PHC. Rather, it allows us to articulate the difference between the two models in a particular high-income country context when so much discourse about PHC (both within Australia, and more globally under the post-2015 Sustainable Development Goals of promoting Universal Health Coverage) does not make the distinction.

Primary healthcare in high-income countries

In high-income countries, the best examples of comprehensive PHC have been community health centres in Canada (, the USA6 and Australia.7 Community health centres are characterised by multidisciplinary teamwork, a social understanding of health, community participation in management, advocacy for policy changes to address the social determinants of health at higher government levels, and services that cover rehabilitation, treatment, prevention and promotion. These centres have remained marginal within their country health systems, faced opposition from mainstream medicine and struggled to maintain their comprehensiveness.

In Australia, community health centres were the legacy of a 1970s national programme and were maintained by state governments including the South Australian government, which is the focus of this study. There have been very few studies of whole PHC services. Labonte et al 8 found that most of the empirical PHC literature focused on ‘slices’ or particular programmes, rather than studying the overall service in a systematic way. Our research studied the totality of services in a way not previously reported in the literature.9 While we did not anticipate it at the outset, our 5-year study (2009–2014) witnessed a series of structural reorganisations and policy changes10 which undermined the comprehensive nature of our case study services. The aim of this paper is to describe the difference between a comprehensive and a selective model of PHC in a high-income country setting.


This paper draws on a 5-year longitudinal realist11 case study of PHC services which used program logic modelling to describe the services and their expected outcomes (for details see reference12). This paper draws on a synthesis of our findings to examine the difference between comprehensive and selective PHC. Our study was conducted with seven PHC services and this paper draws on data from five state-managed PHC services (the other two are non-government services and did not experience the changes reported in this paper). The services are anonymised as A, B, C, D (an Aboriginal health team) and E. Service B withdrew from further participation in the study in 2012 due to high staff workloads and significant organisational change. Service E agreed to join as a replacement. Further details of the services in 2010 and 2103 are provided in table 1. Each case study service adopted a reasonably comprehensive PHC approach at the onset of the study although A, C and E did not provide medical services reflecting the historical opposition of the organised medical profession to these centres.12 In 2009, all services had organisational statements which demonstrated strong commitment to the Alma Ata Declaration principles including an explicit commitment to social determinants of health and health promotion. These documents were analysed as part of this study. This paper also draws on previous work in our 5-year study which reports on a detail analysis of Federal and State government policy documents which demonstrate the changing context that drove the change from comprehensive to selective PHC detailed in this paper.13

Table 1

Characteristics of case study state-managed PHC services: 2010 and 2013

Staff interviews

We interviewed staff in 2009 and 2013. The details of the interviews have been reported elsewhere.14 In 2013, a total of 63 interviews were conducted with service practitioners and managers in the seven PHC sites, and regional and central health executives.

Interview questions were developed by the research team based on the attributes of PHC and data collected on changes in PHC during 2009–2013, and piloted on two practitioners and one manager from non-participating PHC services. Interviews were audio recorded, transcribed and deidentified. Ethics approval was received from the Southern Adelaide Clinical Human and Aboriginal Health Research Ethics Committees. Written consent was obtained from all participants.

A team approach was taken to thematic analysis, aided by NVivo software. Codes were discussed and revised in team meetings, and four interviews were double coded or triple coded, ensuring rigour through constant monitoring of analysis and interpretation.15

Program logic models

An overarching model of comprehensive PHC in Australia was constructed in 2010 using a collaborative process16 and drawing on the models constructed for each service. Following the interviews conducted in 2013, a new program logic model was constructed by the research team reflecting the changes and revealing the more selective nature of the state-managed services. The program logic models we used are not akin to practice audits although we note that the dimension specified in the Australian Quality and Safety Commission’s17 PHC practice level indicators of quality overlaps significantly with the mechanism and activities in the selective program logic.


Figures 1 and 2 present the before and after pictures of PHC. In figure 1 the comprehensive nature of the services in 2009–2010 is shown according to the operating context, underlying mechanisms, service qualities, activities, activity outcomes and community health outcomes. By 2014, these had changed significantly in the services and these changes are shown in figure 2 and elaborated below.

Figure 1

The Southgate model for CPHC in Australia. CPHC, comprehensive primary healthcare; FTE, full time equivalent; PHC, Primary Health Care.

Figure 2

The selective primary healthcare model evident in the South Australian state government-managed services in 2013. CPHC, comprehensive primary healthcare; FTE, full time equivalent; HP, health promotion; NGO, non-governmental organisation; PHC, Primary Health Care.


In 2009–2010, the context of the services was reasonably supportive of comprehensive service delivery. By 2014, the context for the services had changed so that their work had little political or bureaucratic support,  and their mandate changed from being responsive local services to one in which their agenda was centrally driven with a focus on chronic disease management. This changing context partly reflected an ongoing dispute between the Australian Federal and State governments regarding which authority was responsible for PHC and health promotion, largely in terms of who was to pay for the activities. While the Federal government had introduced regional PHC authorities (first Medicare Locals and then Primary Health Networks), their mandate and their practice were not comprehensive18 and they did not work with the state-funded PHC services.19


The main difference between the selective and comprehensive models was that the service components had contracted considerably by 2014. Rather than offering services that responded to a wide range of community health issues, the service model was reduced to a focus on chronic disease management and some limited early childhood services. Previously, the services had responded to a much broader range of health issues including domestic violence, injury prevention and food quality. The new selective model was also inwardly focused whereas the comprehensive model had relied on health workers linking with other sectors, reaching out to the community and, although in a limited way, paying some attention to social determinants of health. Most significantly the selective model was based on a biomedical understanding of health with little or no attention to social factors.

Service qualities

The comprehensive model encouraged individual and community empowerment and responded to community needs. The health professionals also saw that the comprehensive model was holistic, used by those most in need and placed high emphasis on being culturally respectful. By contrast, the selective model paid very little attention to these attributes. The Aboriginal health workers at Service D felt less able to work in ways that suited the community, and some staff at the other services felt their service may be less welcoming to Aboriginal and Torres Strait Islander clients than in the past:

“I don’t think the centre is particularly safe or friendly for Aboriginal clients. It’s just a little bit more clinical … we don’t have the Aboriginal flags and we don’t have the things that would make Aboriginal people feel especially welcome to this service, unfortunately.”

Some services also had less capacity to flexibly respond to incorporate the needs of client groups, such as Aboriginal and Torres Strait Islander peoples, people from supported housing, and migrants, as local initiatives, health promotion activities, outreach and community development work were curtailed. Limits to comprehensiveness reported were that resource limitations meant services had to be targeted. The Aboriginal health service was open to all Aboriginal people, however. Under the comprehensive model, the aim was a broad response to community health needs which were identified in consultation with the community. Thus, a practitioner spoke of this engagement:

“Community health was very much around the Ottawa Charter and things like that, about being very accountable to your local community, and a lot of local community involvement and a lot of local ownership of how the centre operated and what services the centre provided, and a lot of local initiatives.” (practitioner, Service B)

Other comments demonstrated that community advice was no longer valued:

“No community involvement whatsoever. The only thing we do have is a client feedback form.” (practitioner, Service C)

“… you can’t go out and work with the community or plan with the community or other agencies because it’s become that siloed work.” (practitioner, Service A)

The selective model had a narrow focus on reducing hospital admissions:

“We really are now refocusing [Service A] to the high end chronic conditions that we feel we can create a service continuum interfacing with the acute sector and really focusing on hospital avoidance for clients with those conditions.” (regional health executive)

Service activities

The main difference in service activities was that the selective model focused solely on the treatment and secondary prevention activities for individuals. Nearly all the focus was on chronic disease management and the only other services remaining were for children but their scope had been reduced. The comprehensive model had a wider gaze and saw its mandate as working with individuals and the community as a whole in a variety of ways as this comment indicates:

“In the past we’ve run a wider scope of programs and groups, so it wasn’t uncommon to team up with a nurse and do some more preventative lifestyle programs, which we can’t do anymore.” (practitioner, Service E)

Many of the activities lost were of benefit in relation to many diseases. For example, activities that promote social connection are good for mental health and physical health,20 and exercise is a key component in management of mental health issues such as depression21 and diabetes. The comprehensive model included a wider range of activities, shown in Box 1.

Box 1

Activities present in comprehensive but not selective Primary Health Care

Community advocacy campaigns including domestic violence, Aboriginal and Torres Strait Islander rights and cultural pride

Individual advocacy: supporting individuals in quest to gain housing, welfare benefits

Support groups: for domestic violence, men’s groups, women’s groups, and mothers and babies groups

Community activities: lunches, school and childcare centre engagement

Community development: engagement with community members on health issues they are concerned about

Intersectoral actions: membership of regional roundtables, engagement with other sectors on range of health issues

In 2009, the health professionals reported working with the community in many different ways, often going out to community sites, but by 2013 institutional support for this activity had gone:

“It’s a lot more client-coming-into-the-service-based, rather than going out into the community … we’re not working with the [adult education school] or the local childcare centre, whereas probably in 2009 we were stepping outside our doors a little bit.” (practitioner, Service A)

The move from a comprehensive to a selective set of activities was summed up by this nutritionist’s comment:

“We would visit community groups regularly and be a guest speaker for example. We would run group programs that were really around increasing personal knowledge and skills, very hands on practical—like cooking programs. That work has slowly been whittled out of the role. We would do like a split of time, like 30% of the time would be client direct, 30% on groups and then 30% would be health promotion and other activities. So it might be networking with a local childcare centre, for example, helping do menu reviews, supporting community initiatives and really being responsive to the local community needs. That has turned into now just offering one on one nutrition work.” (practitioner, Service E)

In 2009, the services worked with other sectors, including, for example, at Service E a series of roundtables on issues including early childhood development, domestic violence and injury prevention. By 2014, all that work had ceased. Thus, a narrowing down of service activities typified the changes over the study period.

Activity outcomes

The impact of the changes to the mechanisms and service activities is evident in terms of the expected activity outcomes. Under the comprehensive model, outcomes were expected in individuals and also for communities (eg, more supportive environments, increased social capital). The selective PHC outcomes were limited to improved chronic disease management and aimed for more planned, managed care and decreased acute, episodic care for chronic disease, and a reduction in hospital admissions. Thus, the activity outcomes are less ambitious.

Community health outcomes

The differences between the models become starkest in the likely outcomes. The selective model is expected to lead to improved chronic disease management for some individuals and so have negligible population health impact. By contrast, the comprehensive model anticipates improving health and well-being in individuals (including those with chronic disease) and the community and also to reduce health inequities. Selective PHC leads to a chronic disease treatment focused health system with little capacity to prevent disease or promote health. The comprehensive model provides a health system that would make some contribution to reducing the burden of disease and also promote well-being more generally, although the model depends on being supported by broader government action on the upstream social determinants.

Study limitations

This study is limited by the scope for generalisation from five case studies. Inevitably, case studies are context dependent and so care has to be taken in extrapolating from this study to other settings. The changing context of the study sites meant that the collaborative processes we used to develop the program logic models in 2009 were not possible for the 2014 model. We are, however, confident that the model does reflect the reality in the services concerned because our analysis draws on in-depth interviews offering detailed insights to the changes since the original model was developed.


Our findings have shown that while there are similarities between the two models of PHC in that they are both community based, involve multidisciplinary staff and respond to individuals in need of care, beyond that there are significant differences that mean the capacity for community health improvement is reduced significantly. This difference is crucial and needs to be defined very clearly when health systems are being reoriented to PHC. Previous delineations between comprehensive and selective PHC have been limited to short theoretical accounts22 23 whereas this study provides a unique empirical examination of differences in the two visions on ground practice.

Chronic disease management is vital given the increasing burden of chronic disease. But it is short sighted to design a PHC system solely for this purpose. A more comprehensive model offers many benefits to a community. Community involvement in management and planning of a health service helps ensure they respond to community need.24 A focus on prevention and the promotion of well-being in PHC is an important component of a health system’s capacity to prevent disease. As Rose25 has demonstrated, prevention requires more than a focus on those already ill, rather making smaller changes across the whole population and reducing the risk by clinically insignificant amounts adds up to a far greater contribution to prevention. Thus, while selective PHC appears to have an inherent logic in that it focuses on people with disease making high demands on the health service, it sets the health system up to run endlessly, like a rat on a wheel, because there is no prospect that it can stem the flow into this disease category. There appears to be nothing in the logic of the selective approach that suggests it can prevent new cases emerging.

We acknowledge that the comprehensive model as envisaged by these services did not include extensive advocacy on upstream determinants such as income inequity, unemployment or housing. Thus, its claim to be comprehensive was limited by the relative limited scope of action on social determinants of health. Elsewhere, we have detailed the management and funding pressures that led to a retreat from a more comprehensive model in South Australia.10 This retreat was despite the fact that Australian reviews of the health system have reinforced the importance of PHC and health promotion in particular.26 27 Our models show that the broader sociopolitical context is crucial in shaping the implementation of PHC. Because comprehensive PHC challenges the powerful dominant biomedical model of health, a particularly supportive political context is required for its implementation. In Australia, there has been declining investment in prevention—the spending has dropped from 2.2% to 1.4% (Ref. 28, p. 255). In this context, comprehensive PHC is unlikely to flourish.

Unlike the selective model, comprehensive PHC reaches out to people for whom health services are hard to reach through a range of community development activities.29 Actions to address local social determinants of health also seek to create supportive community environments for health and so promote health for the whole population. The importance of this continuum of action has long been recognised,30 yet its acceptance and integration into health systems is proving very difficult. Baum and Fisher31 have argued that there are structural pressures against a social approach to prevention including the inherent individualism driving political and social thought in many industrialised countries, and the considerable corporate pressures that exert influences on policy dialogues to keep the focus on individual behaviours rather than structural factors that drive poor health and health inequities. These pressures make it even more important to be clear on the different styles of PHC and to specify what constitutes a comprehensive and a selective model. The two models presented in this paper enable others to assess the extent to which their PHC services are comprehensive and operating in an environment which is supportive of such approaches.


We would like to acknowledge the staff of each participating service and thank them for their time and trust in allowing us to conduct research in partnership with them.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
View Abstract


  • Contributors FB oversaw the research, led the design of the research, contributed to data collection and analysis, and led the writing of this journal article. TF and AL contributed to the design of the research, and to data collection and analysis, and the writing of this journal article. RL and DS contributed to the design of the research, analysis of findings and the writing of this journal article.

  • Funding This research was supported by the National Health and Medical Research Council (grant number 535041).

  • Competing interests None declared.

  • Ethics approval Southern Adelaide Clinical Human and Aboriginal Health Research Ethics Committees.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement No additional data available.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.