Demand Forecasting Takes Off

This is a joint post with Prashant Yadav.

In early 2006, CGD convened a working group–led by Ruth Levine- to address a pervasive problem in global health: poor forecasting of expected demand for key products. Long-term strategic demand forecasts are needed in order for manufacturers to make capacity investments, make more accurate long term plans for manufacturing and distribution, and for donors to conduct better multi-year program planning. Medium-term demand forecasts are equally essential. When such forecasts are off, manufacturers have to dispose of unsold drugs; donors and ministries of health may face uncertain prices and availability of essential products; and –most importantly- communities and individuals can face the terrible prospect of shortages, incomplete treatments and the emergence of drug resistance.  The Wall Street Journal’s recent coverage of shortages of pediatric tuberculosis (TB) treatments demonstrates the weaknesses –or perhaps absence- of adequate demand forecasting in India’s anti-TB programs.

Six years later, global health funders are clearly headed towards better forecasting, as the major players make progress in areas that are closely aligned with the CGD forecasting working group’s recommendations

First, the working group report recommended that funders take forecasting more seriously, and global health funders have delivered. UNITAID has been instrumental in supporting global forecasting for artemisinin-combination therapies and artemisinin. Since early 2010, it has been funding a consortium consisting of Boston Consulting Group, CHAI and MIT-Zaragoza* to generate global demand forecasts for ACTs with quarterly updates that inform policymakers for budgeting and program planning, and market participants for sourcing and supply planning. The forecasts are overseen by a steering committee consisting of UNITAID, Roll Back Malaria (RBM) Secretariat, WHO and the Global Fund to vet their validity and robustness but also to ensure there is coordination across these agencies around a single global forecast.

For vaccines, the GAVI Alliance has set up its in-house demand forecasting group; and recently put out a RFP for an evaluation of the accuracy of their forecasts. The US Government’s PEPFAR mega-project – the Supply Chain Management System (SCMS), whose $7 billion pre-bidding solicitation was recently released – includes many activities to improve effective forecasting and supply planning. The Global Fund has been slower to act, but there too recent actions bode well –under a newly reorganized Procurement Department, clearer responsibilities for forecasting have been put in place.

Second, the working group report recommended the creation of an “infomediary” that would gather and analyze data to forecast demand across a variety of diseases and products and make information available to all stakeholders, thereby enabling better budgeting from funders and clearer incentives for manufacturers to respond with best prices and adequate supply. Here again, UNITAID is leading on behalf of all funders, awarding a grant to IMS Health to develop a single repository of market information for the various markets UNITAID operates in.  While the data tables generated by the project are mainly for the use of UNITAID, the architecture will be designed to facilitate access by other donors and stakeholders. If other funders buy in quickly, the infomediary can be a reality.

Finally, the working group recommended that global funders better align incentives and share risks through restructuring contractual agreements. Here, there has been some experimentation around the core idea. A group of agencies including UNFPA, BMGF, CHAI, CIFF, NORAD, DFID and USAID, recently set up a “volume guarantee” for the purchase of a new contraceptive implant. As was argued in the CGD working group report, a volume guarantee can help reduce prices and increase supply availability, and indeed the price of one of the implants was cut by almost half as a result of the volume guarantee. However, as was the case with the Advanced Market Commitment – such innovative risk sharing schemes always come with criticisms from some actors. For example some argue they lock in a single manufacturer for two years, potentially affecting long term price.

We revisit the issue of demand forecasting at the Global Fund in our recent report, ‘More Health for the Money’, now up as a consultation draft for comments. In the report, our colleague Victoria Fan finds that short lead-times in ordering are correlated with higher prices of a common first-line antiretroviral combination, suggesting poor demand forecasting and resulting “emergency” purchases. Of course, many other factors may also play a role via availability of funding, market structure, and contract and payment arrangements. However, although it’s completeness and consistency can be improved, the mere availability of the Global Fund’s Price and Quality Reporting (PQR) data enables a better and more rigorous understanding of what drives prices of products purchased with Global Fund monies. Other global donors should also consider investing in such transaction data recording systems.

Better demand forecasting is a low-hanging fruit in the quest for more health for the money and global health funders are well on their way; we’ll be watching for further progress.

*Prashant Yadav was previously employed at the MIT-Zaragoza International Logistics Program.


CGD blog posts reflect the views of the authors, drawing on prior research and experience in their areas of expertise. CGD is a nonpartisan, independent organization and does not take institutional positions.