So Suppose You Want to Fix African Health Systems

My colleague April recently blogged about the LA Times Article which suggests that the problems of African health systems are primarily due to efforts to expand single disease vertical programs. Clearly the problems have been there for a long time. So suppose you are the Gates Foundation and you want to fix the African health systems in which you are operating your single-disease programs. What would you do? A surprising place to start for inspiration is to read the fascinating article by Atul Gawande in the December 10 issue of The New Yorker entitled "The Checklist." At first glance this would not seem to be a promising source for information about how to improve service delivery in the rural clinics of Africa. After all, Gawande's article is about the effort to improve the delivery of a vastly more technical health care service, that of intensive care units in American hospitals, some of which are at the pinnacle of American medical care. As Gawande eloquently explains, these ICUs suffer from extreme complexity. In the typical American ICU "you have a desperately sick patient, and in order to have a chance of saving him you have to make sure that a hundred and seventy-eight daily tasks are done right." While a few ICUs exist in Africa that attempt care of equal complexity, the daily routine of peripheral African health care clinics is far more mundane. So how could Gawande's story hold lessons for helping the poor of Africa?For the American ICUs, the "fix" is what Peter Pronovost, the protagonist in Gawande's story, calls a "checklist." Pronovost has been able to show that the risk of an infection caused by the routine administration of intravenous fluids can be cut from over ten percent to close to zero by enforcing the use of a checklist. Over 15 months the use of this checklist in one hospital "prevented forty-three infections and eight deaths, and saved two million dollars." The ICU achieved even greater savings of lives and costs when the doctors and nurses worked with Pronovost to develop their own checklists. "Within a few weeks, the average length of patient stay in intensive care dropped by half." As the tagline of Gawande's article asks, "If something so simple can transform intensive care, what else can it do?" The idea of a checklist, which Peter Pronovost has shown to be so powerful in improving the most complex health care, has a long history of improving basic care under a different name. Since the 1970's when US physicians began delegating more and more tasks to physicians' assistants, nurse practitioners and other "intermediate health care practitioners," health service researchers have been attempting to systematize the delivery of medical services by developing and testing so-called "medical decision algorithms" or "medical decision protocols" or "diagnosis and treatment guidelines." Around the same time, Professors Mercenier and Van Balen from the Institute of Tropical Medicine in Antwerp were developing a system for delivering "district based health care" in the Kasongo District of the Democratic Republic of the Congo, which at that time was called Zaire. Their approach to maximizing the quality of the health care subject to severe resource constraints was comprehensive, involving district-level health insurance and a holistic concept of the entire health care system from primary care through referral networks to a hospital at the apex. An essential element of their system was that a set of standardised diagnosis and treatment guidelines should be implemented, laying down criteria for the referral of patients to the next level of care (e.g. a child with a serious infection or severe malnutrition)," which was subsequently applied by their students to small projects throughout Africa. I personally visited one of these called the Pikine Project which was located in a suburb of Dakar, Senegal in 1979. Like the earlier Kasongo Project, which had served as the prototype, the Pikine Project trained its health care providers in the application of simple diagnostic and treatment protocols which improved the quality of care. (Also see "Disease control in primary health care: a historical perspective" and the classic primary health care manual: Where There Is No Doctor). The World Health Organization picked up the idea and began to collaborate with teams of experts to prepare "standard case management" protocols for a variety of diseases, including pneumonia, hypertension, and sexually transmitted infections. For example, the most recent protocols for the management of sexually transmitted infections include the following detailed decision tree for use when a newborn baby has a suspicious discharge from the eye:Of course, reforming African health care systems is not as simple as distributing diagnostic and treatment protocols. Gawande recounts substantial resistance to the use of checklists by ICU personnel in the US. Health professionals in Africa also resist the introduction and application of decision rules. According to a former colleague who is a veteran of the Kasongo Project, the Congolese nurses who were taught the decision protocols resisted at first because the rules seemed to suggest that they would not know on their own what to do. But my colleague said that the nurses eventually came to accept the protocols because they saw with their own eyes how the rules saved time, prevented repeat visits and improved the health outcomes. Gawande points out two other impediments to protocols that were overcome in the US ICUs, but would prove more refractory in Africa: stock-outs of key supplies and inadequate supervision. When the ICU staff in the Indiana hospital complained of insufficient drapes to protect patients from infection, which were called for by their checklist, the hospital executives were quick to provide the missing materials. To make protocols work in Africa, the supply chain for all the required supplies would have to be greatly improved. A common finding in the US and poor countries is the need for tight supervision to assure the systematic use of checklists and algorithms.