3. Strengthening Health Systems

Assessment of Approaches to Strengthen Health System Capacity

Strengthening health system capacity to improve performance is a wide-ranging subject, likely to require action—often simultaneously or appropriately sequenced—on many fronts. In particular, it requires attention to the various functions of the health system, especially to the various dimensions of management, as well as to the relationships between the health system, its patients (clients), and their communities. Evidence on which approaches work best is limited. The coverage of this section is therefore selective, drawing on chapters in part III and focusing on stewardship and regulation, organizational structures and their financing, and general management functions—namely, human resources and quality assurance.

When possible, we identify general lessons and note instances of relevant country experiences. In interpreting them, readers will need to keep in mind the strengths and weaknesses of their own country's health system. For example, in South Africa, where basic hospital supplies are good, improved training of health staff members reduced case-fatality rates for severe malnutrition, whereas in settings that experience shortages of antibiotics, potassium, and milk powder and that lack a doctor, training alone is highly unlikely to reduce high case-fatality rates (Ashworth and others 2004).

Stewardship and Regulation

Saltman and Ferroussier-Davis (2000, 735) explain stewardship as a "function of governments responsible for the welfare of populations and concerned about the trust and legitimacy with which its activities are viewed by the general public." The importance of the stewardship role is indicated by analyses that suggest that, in countries with good governance, a relationship is apparent between increased health spending and reduced child mortality (chapter 9), but that such a relationship is not apparent in countries that scored less well on indicators of good governance.

Strengthening structures of accountability to communities and introducing mechanisms to ensure that users have a voice in the local health system and can influence priorities are likely to be important in encouraging good performance. Methods to increase the transparency of resource allocation to peripheral services are also needed. In Burkina Faso, participation by community representatives in public primary health care clinics has increased the coverage of immunization, the availability of essential drugs, and the percentage of women who get two or more prenatal visits. In Ceara, Brazil, strengthened community accountability mechanisms helped improve service delivery (chapter 9). Factors identified as important to the success of community-based health and nutrition programs in chapter 56 include the existence of an effective, respected, and socially inclusive organization at the community level that builds on established community procedures.

Because of the substantial role that private sources of care play in almost all low- and middle-income countries, regulating and developing creative ways to work with the private sector are important. This effort needs to be seen as part of the stewardship role. Even though most countries have a network of regulations controlling private providers and products such as drugs, the regulations are often outdated and poorly enforced and can even be counterproductive (box 3.1).

Evidence is growing that using a mix of measures to influence both consumers and providers can improve the quality of care obtained through private providers. Chapter 70 provides several examples, including introducing total quality management practices and training with peer review feedback. Providers in the informal sector are some of the hardest to reach because of their wide distribution, small scale, and minimal education; however, some evidence indicates that their dispensing practices can be improved (box 3.2).

Regulation can be used as an intervention in its own right, as well as a way to improve health service delivery. The list of interventions identified as success stories (chapter 8) includes these in which a change in regulation was at the root of success:

• regulations requiring all sex workers in brothels to use condoms in Thailand

• tobacco control legislation in Poland and South Africa

• provision of a legal and regulatory framework for adding fluoride to salt in Jamaica

• legislation banning the sale of noniodized salt in China.

Given that enforcement is the Achilles heel of regulation, a noteworthy point is that these countries are all middle-income countries with a reasonable level of enforcement capacity. In other countries, approaches such as that outlined in box 3.2, where the authorities work with the private sector rather than seeking to control it, may have a better chance of succeeding.

Organizational Structures and Financing

The appropriate configuration of health system structures can ensure a clear delineation of responsibilities and accountabilities inside organizations, linking performance with rewards. Governance and organizational structures can also help ensure organizations' accountability to the public.

In recent years, the approach known as new public management, explained further in chapter 73, has encouraged a rejection of traditional, hierarchical forms of public sector management, whereby a single organization both finances and provides health services. For example, the U.K. health service has introduced a clear separation between the entities purchasing services (deciding what services are required for a given population and allocating funds for them) and those providing services. One aim of such arrangements is to ensure that providers' interests—as opposed to users' interests—do not dominate decisions on what services are funded. In addition, separating purchasers and providers allows competition to be introduced in service provision. Although introducing competition is widely considered desirable to encourage efficiency, debate continues on the magnitude of potential adverse effects.

Examples of new organizational structures include removing national health services from civil service control, introducing executive agencies to manage health services, and using contracts to govern relationships, both within the public sector (between public purchasers and public providers) and between the public and private sectors (Preker and Harding 2003). Colombian reforms introducing competition in both insurance and provider markets are among the most comprehensive. Another reform example is Ghana's creation of the Ghana Health Service, which is separate from the Ministry of Health.

The high transaction costs involved in creating and managing these types of arrangements and the lack of evidence that competition improves the quality of care have moderated initial enthusiasm for new forms of public management. In addition, critics argue that such arrangements are more demanding on management capacity than is direct service provision (Mills and others 2001). Moreover, implementation has proved challenging. For example, in Trinidad and Tobago and in Zambia, reforms to create new health service agencies have run into major opposition from public sector workers, who oppose changes in their terms and conditions of service.

Some of the more successful elements of new public management reforms are those that involve contracting out services, especially to nongovernmental organizations (NGOs). Early evaluation of contracting experiences indicated that, even though contracting had been perceived as a way to avoid the inefficiencies inherent in public sector provision, it nonetheless required public sector capacity to manage the contracting process (Mills 1998). This situation was particularly a problem if the contractor was a commercial firm or individual provider with incentives to maximize profits (box 3.3). Contracting with individuals and firms that are strongly influenced by a profit motive requires a certain level of state capacity to ensure that the arrangements work in the interests of the state and the general public. In some countries, therefore, NGOs may be more appropriate service providers (Palmer and Mills 2003). A number of quite positive results from contracting with NGOs are now available (World Bank 2004), and the example of Cambodia is one of the most frequently quoted (chapter 13). Nevertheless, most evidence comes from programs with substantial external financial and technical resources, and long-term experience of sustainability is lacking.

Management decentralization has been another continuing theme in recent years. One variant is its application to hospital management, which involves giving hospitals autonomous or corporate status along with much greater responsibilities for raising income and managing their own affairs. A second variant is the creation of autonomous government agencies, and a third is decentralization to general management structures at lower levels, such as a health authority or local government.

Some pushing down of the locus of control over decision making is a prerequisite for effective management at the local and facility levels. However, without the necessary resources and management expertise at these levels and the right incentives, adverse consequences may arise for both efficiency and equity. For example, experience with hospital autonomy in low-capacity settings suggests that making the hospital partially dependent on fees for its income will restrict access by the poor to the hospital and also worsen the care they receive when admitted (Castano, Bitran, and Giedion 2004). However, for nonpatient care services, whose functions are easier to specify and monitor, autonomous agencies may have some advantages. For example, the Tamil Nadu Medical Supply Corporation has greatly improved the efficiency and effectiveness of drug purchasing and distribution (Mills and others 2001).

For decentralization of general health service management to succeed, attention must be paid to the entire management system, including management skills, information, analytical tools, and accountability mechanisms both to the community and to higher levels of management. Because decentralization is a complex process, takes a variety of forms, and is affected by the local context, research on its merits and demerits has been inconclusive (Alliance for Health Policy and Systems Research 2004). Some evidence indicates that decentralization to local governments can lead to neglect of broader public health functions and disease control, because these types of care are less visible to the public than curative care, as Khaleghian and Das Gupta (2004) indicate occurred in the Philippines.

Reviews of the merits of integrating services and of the effect of vertical programs on health systems have also been inconclusive. Some positive examples are available, such as the strengthening of health infrastructure and surveillance systems by the polio elimination campaign in Latin America and the Caribbean (Levine 2004). Nonetheless Briggs, Capdegelle, and Garner's (2001) review of the effects of strategies for integrating primary health care services on performance, costs, and patient outcomes finds too few studies of good enough quality to draw firm conclusions.

Human Resources

Achieving health policy goals depends on being able to train, recruit, and retain a staff with the necessary bundles of skills. In planning for human resource needs, countries must relate the numbers and levels of each category of staff members to health policy goals and the priorities that are set, given the overall availability of resources and local labor market constraints.

In recent years, concerns about the international brain drain have increased greatly, with evidence indicating that migration by doctors and nurses is severely affecting health services in some Sub-Saharan African countries (Physicians for Human Rights 2004). Actions by developing countries to improve recruitment and retention should either raise the rewards, both financial and nonfinancial, of local employment or reduce the attractiveness of alternative employment—for example, by making qualifications less portable across countries (chapter 71). Raising the remuneration of health workers may be difficult because it is likely to lead to demands for increased pay from other public sector employees. There is a long history of making use of local cadres, which can also allow training that is more specific to the needs of the local health system and its priorities.

Examples include nurses with extended training and roles and people working at subnurse levels with training of a few weeks to three years. For example, Bangladesh employs family welfare visitors, health assistants, and medical assistants; Uganda provides three years of training to clinical officers, who function as subdoctors, and three months of training to nursing aides; and Malawi trains clinical officers, who carry out surgical procedures and administer anesthetics in addition to providing medical care. Despite widespread use of such workers, evidence on how they perform relative to more qualified staff members is limited, though a study of clinical officers in Malawi suggests that well-trained clinical officers can safely substitute for doctors in performing cesarean deliveries (Fenton, Whitty, and Reynolds 2003).

The salaries necessary to recruit and retain staff members will depend on the opportunities they have for other employment both within the country and in other countries. Salary levels will also depend on health workers' preferences between financial and nonfinancial incentives. Evidence suggests that influences on motivation, though reflecting universal principles, will vary considerably from place to place (Brown 2002). Therefore, compensation and incentive structures need to be adapted to countries' circumstances; however, evidence is scanty on how countries have attempted to adapt such structures and whether they have been successful in improving recruitment and retention.

One approach to improving health workers' performance is to link performance and remuneration. The Chinese national tuberculosis (TB) program, identified as a success story (chapter 8), provided village doctors with incentives to treat TB patients. However, performance-related pay requires a good regulatory framework, skilled managerial resources, and careful monitoring to counter adverse effects—all features that are unlikely to be available in countries with limited capacity. Even in China, other experiences are much less positive because managers were not required to take likely adverse health consequences into account (box 3.4). Similar comments apply to the widespread practice of allowing doctors to work in both the public and the private sectors to increase their incomes. Doctors may exploit their private practice rights by encouraging patients to attend privately if they want better quality care—or even by diverting government resources, such as drugs, to private patients. Thus, the effects of private practice on incentives in public practice tend to be negative unless carefully monitored and controlled.

Nonmonetary rewards to encourage staff retention can be useful in such settings, as well as easier to manage. They include the availability of facilities and materials; of opportunities for learning and career progression; of subsidized housing and education for dependents; and of a culture that values the contribution of health workers to the achievement of organizational and system goals. In addition, the methods and levels of funding, the extent of organizational autonomy, the nature of support and supervisory systems, the role of the organization and of providers in the health system, and the regulation and accountability structures all influence how organizations and individuals function. Thailand provides an example in which the provision of both monetary and nonmonetary rewards has improved the recruitment, retention, and status of rural doctors (box 3.5).

The introduction of well-funded disease control programs runs the risk of attracting the most able staff members away from other positions. Past programs have successfully used combinations of financial and other incentives to encourage good worker productivity and program performance (chapter 71). Incentives have included better salaries; field and transportation allowances; and nonfinancial incentives such as streamlined management, specialized training, availability of facilities and material resources, and results-oriented management that provides effective administrative and technical support. Governments need to find ways of benefiting and learning from these experiences. For example, governments might allow periods of secondment to externally funded programs, after which staff members return to the government with enhanced skills. The success of such an approach will depend on remuneration not differing too greatly and on government bureaucracies providing the scope for staff members to use their new skills.

However, the history of civil service reform is not encouraging (Nunberg 1999). Reforms have sought to reduce the size of the civil service and to improve productivity using incentive schemes such as performance-based pay and promotion structures. Such reforms have been largely unsuccessful because of the political difficulties of reducing the size of the civil service. Structural and organizational changes are typically unpopular with labor unions, especially if they perceive such changes as threatening workers' well-being. Experience demonstrates the difficulties of aligning system and organizational objectives with individual workers' objectives (Martineau and Buchan 2000) and suggests that solutions need to be sought that do not involve radical reform of employment patterns unless the country setting is particularly propitious.

Where contracting with NGOs or other private providers is an option, doing so may permit changed employment patterns and improved performance without the widespread disruption that can result from attempting to change government workers' terms and conditions of service. In Cambodia, a project contracted to an NGO (HealthNet) obtained some improvements in staff performance by establishing clear agreements with staff members concerning issues such as the working hours that would be expected and the informal charges that staff members were not to demand from patients. In return, staff members received substantial incentive payments (box 3.6).

Quality Assessment and Assurance

The quality of health services has a number of important implications. It affects the outcomes that a health system can achieve—both directly, through patient treatment, and indirectly, by encouraging or discouraging use of the services. It also affects staff morale, because working in an environment where employees know the treatment quality is poor is not motivating.

Substantial evidence, reviewed in chapter 70, indicates that the quality of care is often suboptimal and varies widely within countries. In part this suboptimal quality is attributable to resource constraints, but providing good-quality care is possible even in resource-poor settings.

Evidence on how providers' practices can be improved can be grouped into two categories: policies that indirectly affect providers' practices by changing structural conditions, including the practice environment, and policies that directly affect individual and group practices.

In the first category, legal mandates and administrative regulations can be used to bar unqualified workers from practicing; professional oversight and clinical guidelines can encourage good practices; contracts can specify and monitor quality standards, such as immunization coverage targets; and accreditation can stimulate quality improvements. Among policies that directly affect providers' behavior, training with peer review feedback has been shown to improve quality, as have total quality management approaches; remuneration can be made dependent on performance subject to the caveats raised earlier. Measures that improve quality can increase use, strengthen the public sector's capabilities, and be highly cost-effective—even cost saving.