Cost-effectiveness analysis is used to compare the costs and outcomes of alternative policy options. Each resulting cost-effectiveness ratio represents the magnitude of additional health gained per additional unit of resources spent. Cost-effectiveness thresholds allow cost-effectiveness ratios that represent good or very good value for money to be identified. In 2001, the World Health Organization's Commission on Macroeconomics in Health suggested cost-effectiveness thresholds based on multiples of a country's per-capita gross domestic product (GDP). In some contexts, in choosing which health interventions to fund and which not to fund, these thresholds have been used as decision rules. However, experience with the use of such GDP-based thresholds in decision-making processes at country level shows them to lack country specificity and this - in addition to uncertainty in the modelled cost-effectiveness ratios - can lead to the wrong decision on how to spend health-care resources. Cost-effectiveness information should be used alongside other considerations - e.g. budget impact and feasibility considerations - in a transparent decision-making process, rather than in isolation based on a single threshold value. Although cost-effectiveness ratios are undoubtedly informative in assessing value for money, countries should be encouraged to develop a context-specific process for decision-making that is supported by legislation, has stakeholder buy-in, for example the involvement of civil society organizations and patient groups, and is transparent, consistent and fair.

Within health economic studies, it is often necessary to adjust costs obtained from different time periods for inflation. Nevertheless, many studies do not report the methods used for this in sufficient detail. In this article, we outline the principal methods used to adjust for inflation, with a focus on studies relating to healthcare interventions in low- and middle-income countries. We also discuss issues relating to converting local currencies to international dollars and US$ and adjusting cost data collected from other countries or previous studies. We outlined the 3 main methods used to adjust for inflation for studies in these settings: exchanging the local currency to US$ or international dollars and then inflating using US inflation rates (method 1); inflating the local currency using local inflation rates and then exchanging to US$ or international dollars (method 2); splitting the costs into tradable and nontradable resources and using method 1 on the tradable resources and method 2 on the nontradable resources (method 3). In a hypothetical example of adjusting a cost of US$100 incurred in Vietnam from 2006 to 2016 prices, the adjusted cost from the 3 methods were US$116.84, US$172.09, and US$161.04, respectively. The different methods for adjusting for inflation can yield substantially different results. We make recommendations regarding the most appropriate method for various scenarios. Moving forward, it is vital that studies report the methodology they use to adjust for inflation more transparently.

Background: Tuberculosis (TB) remains a major cause of mortality in developing countries, and in these countries diabetes prevalence is increasing rapidly. Diabetes increases the risk of TB. Our aim was to assess the potential impact of diabetes as a risk factor for incident pulmonary tuberculosis, using India as an example. Methods: We constructed an epidemiological model using data on tuberculosis incidence, diabetes prevalence, population structure, and relative risk of tuberculosis associated with diabetes. We evaluated the contribution made by diabetes to both tuberculosis incidence, and to the difference between tuberculosis incidence in urban and rural areas. Results: In India in 2000 there were an estimated 20.7 million adults with diabetes, and 900,000 incident adult cases of pulmonary tuberculosis. Our calculations suggest that diabetes accounts for 14.8% (uncertainty range 7.1% to 23.8%) of pulmonary tuberculosis and 20.2% (8.3% to 41.9%) of smear-positive (i.e. infectious) tuberculosis. We estimate that the increased diabetes prevalence in urban areas is associated with a 15.2% greater smear-positive tuberculosis incidence in urban than rural areas – over a fifth of the estimated total difference. Conclusion: Diabetes makes a substantial contribution to the burden of incident tuberculosis in India, and the association is particularly strong for the infectious form of tuberculosis. The current diabetes epidemic may lead to a resurgence of tuberculosis in endemic regions, especially in urban areas. This potentially carries a risk of global spread with serious implications for tuberculosis control and the achievement of the United Nations Millennium Development Goals.