Studies of the cost effectiveness of treatment versus prevention found little respect at this conference. The assumptions used in several models presented ranged from inadequate to laughable. One hapless economist was shouted down by a member of ACT UP Paris while showing an analysis that counted the value of the life of a South African who dies of AIDS as equivalent to $100 in lost tax revenue. (E. Masaki, TuOrG1248)

Although decision makers may consult cost/benefits analyses, it was clear that the decision to offer therapy encompasses many factors that may trump the bottom line.

Steven Forsythe of the Futures Group, a technical consultant to governments, non-governmental organizations (NGO) and businesses, reported the findings of a study contracted by a large multinational corporation to assess the feasibility of offering ART to its HIV-positive employees, spouses and children. The cost of treatment was compared to the impact on productivity to see if there would be a direct economic benefit for the company.

The assessment considered three regions, of low, moderate and high HIV prevalence. East Africa, a region with 22 percent prevalence, served as the model for an area where treatment costs were expected to be highest. For this company, it was estimated that eight new patients would come into care each year in its East African operations. The lifetime cost to the company under its medical plan for an East African worker with no access to treatment was estimated to be $4,500. The model predicted that the cost in lost productivity and direct costs of care from AIDS would be highest in the company's management strata, with 46% of overall costs coming from only 15% of the affected workforce. The distribution of costs was estimated as 48% among workers, 41% among spouses and 11% among their dependent children. The annual cost to the company for treating its employees in this high prevalence region was estimated at about $62,000. The model assumed that ART would add five years of life, with death delayed from year one to year six. With the provision of ART, the model predicted net savings during the first year, due primarily to the deferred costs of death and job retraining. The costs of death included end-of-life care and burial expenses. After the first year, however, the model predicted that costs would outweigh the savings benefit.

In this East African example, the model predicted that the cost of providing treatment would be approximately three times the savings benefit, a ratio of .29/1.00. For lower prevalence regions, the cost/benefit ratio was more favorable. The economists reported that for this multinational company's East African operations, the economic benefits of offering ART to its employees would not outweigh the costs. Nonetheless, the company determined that since providing ART was affordable and offered other, unmeasured benefits, it would make the decision to provide treatment to all of its employees. Forsythe stated that the role of the economist was simply to provide estimates of both the costs and benefits as well as the affordability of ART. Armed with this information, the company decided to treat. (S. Forsythe, TuOrG1245)

Although Forsythe didn't identify the client, it was assumed to be the multinational beverage maker and distributor, Heineken. In a subsequent session, Heineken official Stefaan Van der Borght discussed his company's decision to offer ART, their experiences with the program, and the challenges they still face. The company's management decided that one of the key criteria for moving ahead with providing ART was that any program should remain sustainable once begun. Heineken has been doing business in Africa for over 70 years, said Van der Borght, and in this context the decision to provide care was understood to mean a long-term commitment and therefore not to be taken lightly. He said that businesses are generally well prepared to undertake treatment programs because of policies and accountability systems that are already in place to administer programs with verifiable standards and quality controls.

Heineken began its program by establishing several principles: first, no discrimination would be allowed based on serostatus. Second, confidentiality would be strictly maintained. Next, all workers and their partners were to be treated with no distinction made between management and other workers. And once treatment had begun, a commitment was made to continue it, despite layoffs or illness. Finally, no suboptimal regimens would be used; treatment must be in accord with WHO guidelines.

To date, Heineken's operations in Rwanda and Burundi have provided treatment for 45 persons, with an estimated 10 deaths having been averted. The program will be expanding to new territories soon. According to Van der Borght, several enabling factors helped Heineken management make the decision to offer ART.

• A preexisting culture valued sustainability of the company in the region and placed stability over short-term profit.

• There was a willingness to provide leadership for the business community on this issue.

• The company determined that it had the capacity to absorb the costs involved and that adequate medical facilities were already in place to implement the program.

• Finally, the decision makers had known people with AIDS and most had been personally affected.

While the company was studying the issue, the price of drugs began to fall considerably, a development that enabled confidence the company could afford to mount a quality program. This last factor, Van der Borght noted, has been instrumental in stimulating other companies to start offering treatment to their employees. He also noted that the cost/benefit analysis was thought to be less meaningful than the affordability of the program.

The cost per treated employee is currently pegged at about $1600 per year with about half of the company's seropositive employees currently electing treatment.

Yet several obstacles remain, with the difficulty of managing treatment first among them. The cost and complexity of laboratory monitoring is also difficult, according to Van der Borght. The company has also recognized a conflict between the principle of confidentiality and proven methods to enhance treatment adherence such as directly observed therapy (DOT) provided in the workplace. And in very poor regions, where simply having a steady job sets one apart from most people, the added benefit of access to ART only increases that gap.

Relations with governments have also been less than ideal. Often government officials will agree in principle with the program but fail to participate meaningfully. In particular, Van der Borght said, it is difficult to obtain written commitments, perhaps because government officials are embarrassed by their own inability to offer their employees ART. Often when cooperation is offered, it is with strings attached that attempt to shift control away from the company. Other practical problems include assuring that drugs are registered and available for legal import -- the company has actually been faced with paying to expedite registration. For available drugs, stock supplies may be erratic due to changes in import mechanisms or, in one case, a drain on the expected drug supply when the army demanded first access.

The program has also experienced resistance from conservative medical providers who do not understand or believe that HIV can be treated. There is also the perception that if the company is providing this benefit then it must have unlimited financial resources; the program attracts much interest but little support from other actors. For example, NGOs seem reluctant to cooperate with the private sector and company officials are rarely invited to meetings about coordinating HIV care in the region. So far, the UNAIDS Accelerated Access Initiative has offered little assistance with supply lines.

Yet Heineken believes there are untapped benefits from public/private cooperation. Operational research involving the private sector could rapidly discover easier and more efficient ways to provide treatment. And, at the very least, governments could take a greater role in facilitating the success of such programs.

One cost-cutting avenue that Heineken will not pursue for the meantime is the use of non-branded medications. As a branded entity itself, the company has a policy to respect brands and not use generics. Although, Van der Borght says, this position could change if a government where they do business eventually offers an acceptable standard of care for its citizens that includes the use of non-branded ARVs. (S. Van der Borght, ThOr247)