“Such groups tend to be technically stronger, have more effective access to decision-makers, and despite higher costs, can often be very cost-efficient because of the scale of their operations…but very few such large groups can actually convince a young man to use condoms if he has not tried them before.” (O’Malley et al 1996).

There are a number of positive features that represent a comparative advantage for NGOs in working on HIV/AIDS, as compared to other types of organisation, although these characteristics are by no means shared by all NGOs. The commitment of the truly constituency-based organisations to being rooted within their communities, which they often know intimately, for example, enables them to work in a personalised manner and ethically with the sensitive and stigmatising issues raised by HIV/AIDS. In turn, they are able to come to understand the specific local risk factors for HIV transmission. Without the impediments faced by larger and more visible organisations, they can work with marginalised groups such as sex workers, drug users and illegal immigrants. Their small size enables them to respond flexibly to the rapidly evolving nature of the epidemic, and they can change directions quickly as they adapt to lessons learned. Less weighed down by bureaucracy than their larger or governmental counterparts, they can more easily integrate work across functions, rather than being restricted to single types of activity. Finally, their commitment to their communities means that they provide often the only sustained response to the epidemic, and are more likely to offer care and support as well as engage in preventive activities.

Therefore the challenge of scaling up activities in HIV/AIDS must address the question of whether such values of characteristics should and can be preserved as the programme or institution expands. This means not only extending the technical aspects of HIV/AIDS work, but also maintaining the integrity of the processes and values that have been the hallmark of NGO approaches to the epidemic.

At the Horizons/Alliance Seminar, it was not the consensus among organisations present that such aspects of quality are necessarily jeopardised with expansion. Indeed, several cited cases where the aspects of the original model were not lost or diluted in the process of scaling up (e.g. Foster re FACT/FOCUS; ASI). Some argued that a slight drop off in quality may be an inevitable stage of the process of scaling up (Geoff Foster, Family AIDS Caring Trust, Horizons/Alliance Seminar) – that is, that in the interest of reaching a greater number of people, some sacrifice of quality is acceptable. This may be particularly the case if many elements or programmes are scaled up simultaneously (International HIV/AIDS Alliance 2000b), although ensuring proper planning of the scaling up may limit this risk. The question then becomes whether the quality drops below an “acceptable” level and indeed whether such a level can be specified (International HIV/AIDS Alliance 2000b). Yet, as noted above, many NGOs lack sufficient monitoring capacity or data to assess changes in quality as the scaling up proceeds.

One example of an organisation that did pay careful attention to monitoring quality during the process of expansion is that of the Asociation de Salud Integral (ASI) in Guatemala. By enlisting volunteer medical students to collect data, they were able to track changes in quality as a rapid expansion of the counselling and clinical service coverage took place. Analysis of that data indicated that while an increasing number of people were coming to their services for voluntary testing, they were often not returning for their results. Further investigation into the reasons for this revealed that the time between test and results had increased due to the increase in numbers. Thus data was helpful in guiding a redirection of strategy.

There have, however, been documented cases in which a rapid increase of coverage has led to a decline in quality. Anti-AIDS Clubs in Zambia began in the 1980s as a result of a presentation by a doctor on sex education in high schools. Students sold the idea to their friends during the holidays, and the clubs continued to multiply across the nation. Over time, the activities of the Anti-AIDS Clubs in Zambia, for example, had proliferated rapidly and for a number of reasons they were no longer particularly effective and the clubs starting dying out. Thus the scaling up efforts of the Family Health Trust aimed initially to rethink and boost the quality of an initiative that was already operating at national level. The experience reinforces the view that during a scaling up process, one may need to concentrate on one dimension – such as quality or coverage – possibly at the expense of the other (Horizons/Alliance Seminar).

The struggle to maintain quality while the organisation expands may pose a strain on staff. As Naomi Gonahasa of TASO stated at the Horizons/Alliance Seminar: “It began so well but things got tough when it turned to scaling up.” In some cases, NGOs have realised that a key challenge in scaling up is knowing when to stop. As Dr. Suniti Solomon of YRG Care in India has stated “One of the most important things for YRG Care is the quality of services we provide. As a result, we chose to stop expanding our in-patient care when we reached 24 beds so that the quality we provided remains at a high level.” (International HIV/AIDS Alliance 2001)

This understanding of dynamic relationships between quality and coverage has also been examined within the broader development arena by David Korten, who attempted to draw commonalities among five community-based projects in Asia which successfully expanded the scale of their operations.15 He concludes that there is no single blueprint for scaling up, but organisations go through a number of stages during expansion (figure 5) each of which entails a learning process. The first is ‘learning to be effective’, when the organisation learns what strategies are effective among particular populations. At this stage, errors are made, but effectiveness steadily increases; efficiency has not yet been achieved and coverage is likely to be relatively low. At the next stage, the organisation is ‘learning to be efficient’, and unit costs tend to fall. Only in the third and final stage does the organisation move into expansion, and there is likely to be some loss of effectiveness and efficiency as the expansion occurs. As the process of expansion continues, it is likely that average costs will tend to rise, and so raising the issue of when it is optimal to continue to expand rather than replicating the project at a smaller scale. The extent to which this “trial and error” process is permitted and learning is encouraged depends on both the external environment – in terms of finance and political commitment – and the internal characteristics of the organisation. This is addressed in the final section of this publication.

15 This discussion draws on Korten 1980 and Myers 1992

Figure 5 - Programme Learning Curves (as defined by Korten) Source - Korten, 1980.