Comment

What can be learned from data for nancing of global health?

In The Lancet today, Nirmala Ravishankar and colleagues1 valuably quantify and characterise development assistance for health (DAH) since 1990. This endeavour was dicult, and the investigators have had to grapple with fragmented data for aid ows for health from various bilateral and multilateral donor agencies, private philanthropic institutions, pharmaceutical companies, and civil-society initiatives. But there are limitations to what the data, so extensively massaged, can communicate. First, there is a US-centric quality to this analysis. Some of the non-governmental organisations (NGOs) that are treated as receiving money from the USA are, in fact, international NGOs, even though some of their constituent organisations are US-incorporated. Equally, some US-based NGOs receive money from non-US sources, such that the contributions of these other sources are eectively understated. Some nonUS NGOs and foundations as well as bilateral aid agencies in countries outside the Organisation for Economic Co-operation and Development (OECD) are not included because of scarcity of data. Over time, the exclusion of this group would impute some bias to the results because these sources of funding have become increasingly important to DAH. Second, although the focus of Ravishankar and coworkers analysis is on DAH, I am concerned that assistance for health-related areas such as water and sanitation are excluded from the database. The inclusion of aid for these areas would probably not substantially change the messages of the study in terms of the orientation of DAH. However, that nding itself would yield an interesting insight, since the provision of clean water and sanitation would probably do as much to facilitate good health as does the assistance provided to direct medical care. Third, todays study underscores that much health assistance is provided not in usable budgetary resources, but rather as in-kind assistance (technical assistance and commodity aid). Although Ravishankar and colleagues call attention to the questionable valuation of drugs provided by pharmaceutical companies, implicitly raising the question of value for money, one could pose the same question with respect to the high opportunity cost of technical assistance in the form of high-priced external experts.
www.thelancet.com Vol 373 June 20, 2009

Fourth, the investigators allude to the high fraction of the total assistance provided by US-based institutions in the private sector, and the possibility that there might be a tax benet associated with these contributions. I would be interested to know the amount of tax expenditure associated with these contributions since, eectively, a share of private contributions is actually nanced by US taxpayers. A rough imputation of these tax expenditures at, say, the corporate tax rate would change the eective shares derived from the public and private sectors. Such an imputation would be no more arbitrary than that associated with many of Ravishankar and colleagues calculations. Fifth, I believe a far stronger qualication is needed with respect to the inherent articiality of some of the numbers derived from various imputation approaches used in the study. This need for qualication relates to: pre-2002 disbursement data from the OECD creditor reporting system database; data on health spending from NGOs; 2007 data for overseas spending from NGOs, which is estimated on the basis of the growth rate from 2001 to 2006; and grants that were for more than one area (where the grant was divided equally across the matched areas). It is hard to judge the fraction of dierent categories of results derived from such imputation methods as opposed to when hard data are provided by the agencies.

See Articles page 2113

2087

USAID

Comment

Finally, although Ravishankar and colleagues report a distinct reduction in the role of the UN agencies, I think this seems simply to reect the decision of the international community to rechannel previously donated resources and provide new resources through new entitiesparticularly the Global Fund to Fight AIDS, Tuberculosis and Malaria and the GAVI Alliancethat were established to bypass UN agencies. But these new entities are linked indirectly to the UN system, with heavy international participation. I would be wary of overstating this reduction as an important trend, except to the extent that it can be seen as a no condence vote in the capacity of the UN system. This study does, however, make an important and helpful contribution to our understanding of the

magnitude and sources of DAH. If nothing else, the complexities entailed in producing this report show the inadequacies and absence of compatibility in the way in which data for aid ows are provided by the private and government donor community. Peter S Heller
Paul H Nitze School of Advanced International Studies, The Johns Hopkins University, Washington, DC 20016-1905, USA phellerdc@gmail.com
I declare that I have no conicts of interest. 1 Ravishankar N, Gubbins P, Cooley RJ, et al. Financing of global health: tracking development assistance for health from 1990 to 2007. Lancet 2009; 373: 211324.

Thiazolidinediones and clinical outcomes in type 2 diabetes

Published Online June 5, 2009 DOI:10.1016/S01406736(09)61029-1 See Articles page 2125

Type 2 diabetes is characterised by two main metabolic defects: target-cell resistance to the action of insulin (insulin resistance) and insucient secretion of insulin by pancreatic cells (-cell dysfunction).1 Thiazolidinedione medications (rosiglitazone and pioglitazone), through modulation of the transcription factor peroxisome proliferator-activated receptor , have remarkable benecial eects on both insulin action and -cell function.13 The clinical signicance of these eects is underscored by the fact that treatment with thiazolidinediones results in more durable glycaemic

Atheromatous artery

control compared with other antidiabetic agents (especially sulfonylureas).4 Furthermore, the thiazolidinediones have several other benecial cardiometabolic eects, including reduction of visceral fat mass, decreased systemic inammation, and improvement in biomarkers and surrogate outcomes associated with atherosclerosis.1,2,5,6 In clinical practice, the benecial eects of a therapy should be considered in relation to its potential risks. Adverse eects that have been associated with both rosiglitazone and pioglitazone include weight gain, increased incidence of fractures, uid retention, and a two-fold increased risk of heart failure.7,8 Most importantly, meta-analyses911 have reported a 3040% increase in the risk of myocardial infarction in patients treated with rosiglitazone. These ndings have raised considerable uncertainty about the eects of thiazolidinediones on cardiovascular disease. In The Lancet today, the rosiglitazone evaluated for cardiovascular outcomes in oral agent combination therapy for type 2 diabetes (RECORD) study was specically designed to assess the eect of rosiglitazone on cardiovascular outcomes.12 In this open-label noninferiority study, 4447 patients with type 2 diabetes on metformin or sulfonylurea monotherapy were randomly assigned to either add-on rosiglitazone or metformin and sulfonylurea combination therapy. Over 55 years mean follow-up, non-inferiority
www.thelancet.com Vol 373 June 20, 2009

Science Photo Library

2088