COVID fueled debate over centralized vs. decentralized health systems

The COVID pandemic has added fuel to the debate over the benefits of centralized versus decentralized health systems.

Decentralization has long been advocated as a pathway for strengthening health systems and improving health outcomes. Many of the countries identified by Exemplars in Global Health as overperformers – countries that have rapidly accelerated their progress towards achieving their health goals – have done so, in part, through decentralization. These include Ethiopia, Nepal, Rwanda, and Senegal, among others. In each of these cases, decentralization allowed local decision makers to send resources where they were needed most, increasing equity and improving health outcomes.

However, the track record of decentralized health systems during the pandemic has been mixed. The United States, for example, which has a decentralized health system, marshaled a chaotic response to the pandemic and its infection and death rates have been nearly unrivaled around the world. The response to COVID by the United States and other countries with decentralized health systems underscores a few key weaknesses of such systems: that cross boundary coordination in such systems is purely voluntary; they tend to move more slowly; and they achieve scale mostly through painstaking collaboration.

To make sense of this conflicting evidence we spoke with health systems decentralization expert Thomas J. Bossert, Director of the International Health Systems Program of the Harvard T. H. Chan School of Public Health.

Have centralized health systems marshaled more effective responses to the COVID pandemic?

Bossert: There is no simple answer to this question. Consider Taiwan and Singapore – both countries that did very well initially. They are centralized and were poised from the outset to take action at scale quickly because of their previous experience with SARS and their centralized health systems. But if you look at those countries now they are struggling. COVID is evolving and what worked well initially may not be working well now. COVID is moving from a crisis phase to a chronic phase. As it does so, we can look at how health systems have addressed HIV/AIDS, another global crisis that has shifted to an ongoing chronic challenge, for a sense of how they might respond to COVID going forward. There are a lot of factors that are likely to affect health outcomes. We simply can’t say if centralization improved health outcomes during the pandemic.

Has the pandemic changed the sector’s thinking about decentralized health systems?

Bossert: COVID hasn't changed people's thinking about decentralization. But it has highlighted two lessons about decentralization. The first is that responding to a sudden crisis or shock effectively often requires making quick decisions and taking action at scale. Here speed can be key. In such instances centralized health systems can be more effective because they can move quickly at scale. But, of course, this only works if the central government moves quickly and at scale. Over the last two years we’ve seen centralized governments in a few countries fail to take effective action to prevent, prepare for, or respond to the crisis.

Which leads us to the second lesson, which is that decentralization has benefits in a crisis, especially if the decentralized units have the ability to collaborate and a tradition of collaborating. We can see this in Canada where provinces collaborate and talk to each other. Germany is another example of a country where there is a tradition of compromise and collaboration across decentralized health systems. Another example is how states in the Northeast of the United States worked together at the beginning when faced with the national government’s inaction to prepare for and respond to COVID.

How might decentralization have helped some countries respond to the pandemic?

Bossert: Even this question is not so simple. Because what we know about decentralization is that it isn’t one thing. Countries are decentralized in different ways and you have to look at the whole health system. Decentralized health systems are about three aspects of decision making: one, the amount of choice that is transferred from central institutions to institutions at the periphery of health systems; two, what choices local officials can make with their increased discretion; and three, what effect these choices have on the performance of the health system.

What this shows is that decentralization isn’t uniform. Local health leaders may have wide decision-making authority with regard to human resources, but narrow decision-making authority with regard to service delivery or governance.

Beyond this, we need to understand – do local officials have the capacity to make decisions? And how they are held accountable by national authorities and by the local population? These are key determinants of performance in decentralized health systems. At Harvard we consider these three factors – choice, capacity and accountability – the core of our analytical framework for decentralization called the “decision space” approach.

In addition, although the state administration may be decentralized, there may be other organizations, such as strong political parties that have ability to influence local officials. For example, Ethiopia’s government structure was officially decentralized. But the coalition of political parties in government was not decentralized. In fact, the parties were very strong and often set the national agenda, limiting the actual discretion for local leadership. Likewise in Vietnam, the national Communist party plays a strong role in setting the agenda.

We can look at how all of these factors play out in a country like Chile. Chile’s central government made an early and quick investment in a variety of COVID vaccines and participated in phase III clinical trials for vaccine candidates to help the country negotiate for more doses earlier. They bought enough vaccines to fully vaccinate against COVID almost everyone in their country multiple times. Moving that fast helped the country, no doubt. But it would not have been as effective if not for another factor: Chile has a well-established primary health care system that has a strong tradition of routine immunization. This is a decentralized system run by municipalities. That helped the country immensely vaccinate quickly. This decentralized vaccination system has been able to get shots in arms faster than any other country in Latin America.

The effectiveness of decentralization depends on a lot of factors. Therefore, it is important to consider both the character of decentralization and the country context in which it occurs before attributing success or failure of COVID response to decentralization.