br The paper in by Victoria Fan and

The paper in by Victoria Fan and colleagues is important because it describes what several people close to the Global Fund to fight AIDS, Tuberculosis and Malaria already know: there is a problem with its disbursement mechanisms. Fan and colleagues correctly state that the Global Fund should “redesign this system to explicitly link a portion of the funds to a simple performance measure in health coverage or outcomes, measured independently and robustly”. I applaud Fan and colleagues for having done the number-crunching, which involved 508 projects over a 10-year period. Yet they did not analyse the underlying reasons for this poor SB202190 between funding and performance and did not provide practical suggestions about what should be improved. This Comment aims to fill that gap. First, the Global Fund uses the term “performance-based financing” in a confusing manner because it refers to a contractual relationship at the macro level between the Global Fund and the recipients. Yet a rapidly growing number of countries (at least 30) use the term performance-based financing in a different manner. For them performance-based financing is a health-reform movement with a clear definition, best practices, and implementation instruments. Moreover, experts suggest that only by applying a minimum number of best practices and instruments can a programme be labelled as performance-based financing. The criteria include the existence of autonomous management for health facilities; a competitive environment for contracts with both public and private providers; good governance by separating the functions of provision, regulation, contract development, and payments; and moving away from input financing to performance payments. It is clear that the Global Fund\'s interpretation of performance-based financing does not meet these criteria, which explains the difficulty in finding a positive link between funding and performance. Fan and colleagues are kind to the Global Fund by only addressing the issue of linking predictors to fund disbursements. They could also have ventured into sensitive qualitative issues such as the corruption charges in Malawi, Mauritania, Mali, Zambia, and Djibouti. Yet, according to performance-based financing theory, the transparency problems were likely to happen because the Global Fund\'s performance-based financing approach stops at the national level and fails to work with recipient governments or organisations on how to ensure “performance” at the level where it matters: the health facilities, the community, and the patients. Good governance issues such as separation of functions are not addressed at the Global Fund, which, like several other aid agencies, supports corrupting input monopolies such as centrally buying essential drugs and equipment. Its projects require hundreds of millions of dollars\' worth of drugs, equipment, bednets, ambulances, etc, instead of injecting money into the local economy through competitive contracts with health facilities and thereby directly paying for performance in terms of outputs, quality of services, and equity.
A major failure of our global society in the 21st century is that many people in developing countries are not only born and live without any official record of their existence—a flagrant deprivation of an essential human right—but also die without having been seen by medically qualified personnel. The resultant uncertainty about the real burden of specific causes of death is being increasingly recognised by international health and funding agencies as a crucial limitation in the prioritisation of effective public health programmes and assessment of their effect. Recently published estimates of the main causes of global and cause-specific mortality have stirred a profound debate about the validity and adequacy of existing methods used to estimate cause of death. Complete diagnostic autopsies, indisputably the gold-standard method to estimate cause of death in developed countries, are undertaken infrequently in resource-poor settings. Reasons for this include the large proportion of deaths that occur outside the health system, insufficient facilities or trained human resources, cultural or religious apprehension about the practice of post-mortem procedures from the community perspective, and decreasing consent rates in such regions. To bypass such problems, WHO now recommends the use of non-invasive indirect methods such as the verbal autopsy, a protocolised procedure that allows the classification of causes of death through analysis of data derived from structured interviews with family, friends, and caregivers. However, the Achilles\' heel of the verbal autopsy is its accuracy, which depends largely on the quality of the diagnostic criteria, the type of diseases involved, the location of death, and the delay between death and verbal autopsy. Deaths associated with non-specific signs and symptoms are the most problematic, and are an especially common issue for perinatal and neonatal deaths. Despite these key limitations, verbal autopsies are the only source of data for cause of death in many settings, and their practice and improvement should therefore be encouraged. Assessment of the cause of in-hospital deaths is generally based on the clinician\'s diagnosis of the disease(s) that led to the fatal outcome. However, such estimations are also prone to frequent misclassification errors. Indeed, when clinical diagnoses have been contrasted with post-mortem findings, rates of major clinical–pathological discrepancies have ranged from 10% to above 30%, especially in the diagnosis of infectious diseases.