When is prevention better than cure?

Benjamin Franklin once wisely observed, "An ounce of prevention is worth a pound of cure." On reflection, Michael Bixby Dudley noted in 1994: Zealous reformers have obviously discarded this proverb. Several billion pounds of cure are seemingly required to ensure "health care security" for all. Forget the ounce of prevention - it is no longer sufficient.

Zealous reformers have obviously discarded this proverb. Several billion pounds of cure are seemingly required to ensure "health care security" for all. Forget the ounce of prevention - it is no longer sufficient.

This paper asks a few questions and generates some answers that address the concept of HIV prevention across sectors in South Africa. From the outset, it is clear that we cannot separate the concept of prevention from the concept of spending and that the international funding context demands that we address the question: ?“How can HIV prevention interventions reduce the incidence of HIV infection and save financial resources in the process??”

Shared experience tells us that the HIV/AIDS funding made available falls short of the needs of the South African population. Marseilles et al (2002) estimate that international needs are under-resourced by approximately ninety-seven percent, while hiccups in the South African flow of committed Global Funds from national to provincial institutions and diversions from donor commitments have compounded the situation during 2003. In this context, there is increasing competition for material and financial resources, confounded by indications of society?’s declining confidence in the public health sector.

A current trend in addressing these issues is a move towards evaluation of interventions in terms of cost-effectiveness linked to outputs. In the HIV context, you may ask: ?“What are prevention interventions??” and: ?“How do we know they work??” Prevention interventions, as alternatives to more cost-heavy curative or treatment programmes, aim to reduce the incidence of HIV/AIDS, often with an overt focus on sustained behavioural change. UNAIDS recognises the following HIV prevention interventions: voluntary counseling and testing (VCT), prevention-of-mother-to-child-transmission interventions (PMTCT), treatment for sexually transmitted infections (STI), commercial sex worker peer education (CSW), screening blood for HIV infection, schools-based education (SBE), mass media interventions (IEC), prevention measures among injecting drug users (IDU), and social marketing of condoms (CSM). UNAIDS also sets standards for cost-efficiency and cost-effectiveness of these programmes.

On the basis of replicable international research, we assume that these interventions work. For example, researchers usually make assumptions about the efficacy of VCT or peer education when they model the cost-effectiveness of this intervention type using costing and other effectiveness data. That is to say, they assume that these interventions consistently lower risk-taking behaviour. Prior research to show the effects of a particular approach on behaviour is usually used to support the assumptions made.

In conducting cost-effectiveness analysis (CEA) in South Africa, we could ask: ?“What criteria have been used elsewhere to assess the effectiveness of particular intervention types??” Similarly, should we assume that certain behavioural interventions necessarily decrease risk-taking, and, if so, on what basis? Also, what scope exists for analyses that contribute qualitatively to an understanding of effectiveness?

Use: As stated, CEA answers the question: ?“How can we improve on existing intervention models, using available resources more effectively??” CEA answers this question by:

• identifying characteristics of the population under investigation and the epidemiology of HIV/AIDS in that population;
• describing broad socio-economic and other measures of effectiveness on a macro level; and
• addressing issues of performance and achievement in relation to the vision, mission and goals of interventions under investigation on a micro-level.


CEAs can be used to: evaluate existing programmes, sometimes comparatively, to assess pilot stage interventions, or to predict programme efficiency and effectiveness through modeling exercises that address the costs of scaling up interventions.

Benefits:When effectively implemented, CEAs can have macro-level impacts on (government) spending decisions. CEAs can also increase awareness of goals, stimulate creative thought on how to achieve goals, focus attention on behavioural change and measurement thereof, and effect qualitative improvements to business or project planning on a micro-level.

?“How to?…?” steps in the design of cost-effectiveness programmes:CEA measures inputs (financial, human and capital) and outputs (measures of performance as defined by organisations and, more expansively, by researchers). It is important to note that the results of CEAs are driven mainly by the assumptions that underlie them, based on micro- and macro-level measures of intervention effectiveness or baseline HIV incidence. Levin (1983) identified core steps in the CEA design that have been developed over time into complex processes that guide CEAs. Simply, the steps are:

(1) Identify and describe the macro and micro-level research problems:There may be many presenting problems in the HIV prevention field. CEAs identify themes and questions that attach to these problems and investigate these in the population. Questions that could be asked in the current context include: Are prevention interventions achieving qualitatively what they aim to achieve? and: What changes in the epidemiology of HIV/AIDS can inform best practice? These and other questions are important considerations in the planning of CEA as they inform basic assumptions.

(2 Identify the target population:It is critical to identify the characteristics of groups affected and targeted by HIV prevention interventions. In assessing the effects of a particular intervention type on risk-taking as a prerequisite to cost-efficiency estimates, use of randomised comparison and quasi-experimental (sometimes called study) groups (comprising members of the target population) are appropriate.

The full report can be downloaded on the righthand side of this page