Project no. FP6 - 043345

CID

Culture and Innovation Dynamics: Explaining the uneven evolution of human knowledge

Instrument: STREP

Thematic Priority: NEST-2005-Path-CUL

D7.5 – Joint final report on AIDS vaccine development

Period covered: 01/07/2008-31/03/2010 Date of preparation: 13/05/2010

Start date of project: 01/01/2007 Duration: 39 months

Project coordinator name: Stefano BRUSONI

Project coordinator organisation name: KITeS-CESPRI, Bocconi University

1

Cultural and Innovation Dynamics

–

Explaining the uneven evolution of human knowledge

Deliverable D 7.5

WP 7 - The uneven development and distribution of vaccines across cultures: the case of AIDS

Final Report

Eugenia Cacciatori*

Rebecca Hanlin°

Stefano Breschi*

Stefano Brusoni*

Luigi Orsenigo*,^

* Università Bocconi

° Open University

^ Università di Brescia

Expert assistance by Gianluca Tarasconi and Riccardo Valboni in the preparation of data for analysis is most gratefully acknowledged. We are indebted to Nicoletta Corrocher for pointing us to the ‘doubles’ and ‘triples’ techniques for textual analysis.

1.  Introduction and objectives

This project has investigated how culture acts as a resource through which different groups address problems (c.f. Pickering, 1992), in this case the search for an AIDS vaccine. In this context, ‘culture’ refers to “the way people communicate and develop knowledge, beliefs and attitudes which shape the way organizational mechanisms develop and organizational behaviour is conducted.” (WP7-D2: 3-4). The starting point for our research was the assumption that dominant cultures contribute to shape choices, such as what alternative to explore and how to do it, that are critical in determining the features of innovation. We chose the example of the AIDS vaccine research because it provides an exemplary case of the difficulties of ensuring even distribution of innovation. In particular we have focused out attention on the International AIDS Vaccine Initiative (IAVI) the first product development partnership to focus on AIDS vaccine research. We have carried out the investigation of the role of culture at the field level and the level of our case study organization, IAVI.

At the field level, the evolution of the search for an AIDS vaccine has been depicted as a continuing struggle among conflicting paradigms. The struggle concerned both the institutional set up of the field and the best scientific approach to the discovery of the vaccine (cf. Cohen 2001), and is an essentially cultural conflict, as it centers around understandings that are broadly shared within specific communities of what scientific research is and how it should be organized (cf. Geertz 1973; Pettigrew 1979; Pickering 1992). With regard to the institutional set up, the debate has centered on the issue of the optimal level of central coordination in the decisions about which scientific routes and approaches to follow – is a centralized effort needed (a Manhattan Project for AIDS, similar to the centralized and directed effort that led to the successful development of the atomic bomb), or is it better to leave individual scientists to search autonomously in order to increase the chances of success? With regard to the issue of the best scientific approach, the debate has been among those favoring an “empiricist” approach, in which candidate vaccines are quickly developed and tested in clinical trials, and those wanting to base the design of the vaccine on an increased and detailed understanding of how the HIV virus works (the “rationalist” approach).[1] Both the available literature, in particular Cohen (2001), and our own interviews, converge on an account of the evolution of the field in which the rationalist approach in the scientific domain and a decentralized approach in the policy arena held sway until the mid-1990s. Research in this period, according to these accounts, applied the relative novel techniques of genetic engineering to produce candidate vaccines that worked by stimulating neutralizing antibodies. This approach, which has been termed “Follow the Hepatitis B road” (see also IAVI 2008, p.6) and Hilleman 2000, built on the mechanism through which the largest part of vaccines work and on the first success of genetic engineering in producing a vaccine. A natural candidate for these vaccines was the mushroom shaped protein covering the surface of the virus (see Figure 1), whose components are known as gp120 and gp41.

These efforts proved to be unsuccessful. By mid-1990s, it was clear that the relatively simple strategy that had worked for Hepatitis B would not work for the HIV. The failure of the search up to that point led also to a rethinking of both the scientific avenues to be explored and the institutional set up of the search. From a scientific point a view, researchers begun to investigate vaccines based on cell mediated immunity. The shift in the research trajectory was seen as dramatic, as one of our interviewee put it “nobody was doing antibodies any longer!”. On the institutional side, during this period many criticism were raised against the traditional organization of scientific research in US.[2] These included the fact that there was too little central coordination, and this lack of direction meant that promising avenues were not followed through with adequate resources. More fundamentally, it was argued that lack of coordination caused heterogeneity in the scientific infrastructure (assays, procedures, cell cultures, challenge stocks and similar), which in turn strongly limited the ability to compare alternatives in order to identify those more promising. Critics also claimed that the peer review system, which dominated the funding decision of the largest funder of AIDS vaccine research (the NIH), is an inherently conservative mechanism and, coupled with a relatively small community, prone to bandwagon effects. The result was fragmentiation in the scientific infrastructure coupled with much less diversity in the lines of research than theoretically possible and practically desirable. Finally, the lack of interest in the pursuit by the private sector and in particular by the ‘big pharma’ was also a cause of concern. This concern was further deepened when, in 1994, NIH controversially decided not to move ahead with the clinical trial of a candidate vaccine developed in collaboration with industry. While there seem to be agreement in the field that, in hindsight, this decision was scientifically correct, it had the effect to further dampen the already weak interest of industry in the search.

All of this led to a major shift and reorganization in the field around the mid-1990s. WHO took a more leading role, and a not-for-profit organization totally devoted to the search for an AIDS vaccine was founded (The Internation AIDS Vaccine Initiative – IAVI). NIH also restructured its activities, with the objective to increase coordination. A main observable consequence of these changes was a shift towards a more empiricist approach, so that in a few years several vaccines were tested in a variety of clinical trials. This trend towards more explicit coordination among research groups was further reinforced from the mid-2000, as funders begun to require large consortia with very tight data sharing agreements, as opposed to individual investigators, to bid for their funding (Butler 2006).

The issue of cultural conflict, however, was not only present in the scientific arena. The devastating consequences of AIDS in developing countries had also became clear. An increasing amount of resources were devoted to vaccines that would work on HIV types and clades prevalent in developing countries, and several trial conducted in developing countries. In turn, this led to an incidental, but all important, building up of scientific capacity in developing countries (Chataway and Smith 2006; Chataway and Hanlin 2007). On this background, organizations that aspired to ‘lead the field’ such as IAVI, attempted to bring together not only different strands of basic scientific research, but also basic with more applied clinical research. In doing so, they were confronted with further cultural heterogeneity among developed and developing countries, and for-profit and not-for-profit organizations.

The aim of this report is twofold. First, we provide a (partial) validation of the prevalent account of the evolution and characteristics of the search for an AIDS vaccine. Using a database of the scientific publications related to the search for an AIDS vaccine, we map the areas of research by a textual analysis of the abstracts of the publications. In our deliverable D.7.4 we provided a preliminary analysis of the literature based on the number of papers that contained specific keywords in their title, abstracts or keywords for the period up to 2007. Here we report a more detailed analysis based on the method employed by Corrocher, Malerba and Montobbio (2007), in which we analyze the frequencies of occurrences of couple or triples of words that form technical expressions (e.g., neutralizing antibodies, cell mediated immunity). We also expended the initial database to include papers up to April 2009. We map the relationships between different institutions by analyzing co-authorship patterns. Finally, we explore the evolution of the topics of research of key organizations.

The second aim of our report is to provide a more micro level analysis of the role of cultural heterogeneity in the activities of key actors, and in particular our case study organization IAVI, exploring if and how the different cultures of basic and clinical science, and for profit and not-for-profit organizations manage to collaborate. The report is therefore organized in two main sections, one dealing with the evolution of the field as seen through the analysis of the scientific literature and the other dealing with the cultural dynamics of IAVI’s activities in India. The final section draws the conclusions of the work carried out in our project and described here as well as in our previous reports.

Figure 1 – The HIV virus

2.  The evolution of the field through an analysis of the scientific literature

Method

We used the ISI Web of Science database to select all items (conference proceedings, journal articles and other material) containing in title, abstract, or keywords the words AIDS/HIV vaccine and immunization (with various spellings), for the period 1989-April 2009.[3] The search yielded 13,179 papers. Several of these were about other topic, and only cited AIDS vaccines incidentally. As the relevance of papers for our aims was difficult to assess in an automated way (e.g., papers containing “BCG” could be irrelevant, or could be about using BCG as a vector), we manually cleaned the database, bringing it to 10,250 papers, of which 8,336 have abstracts. Random checks of the database show that the number of irrelevant records is now between 10% and 15%. The database contains very few papers for 1989. Therefore, we excluded 1989 from the analysis, bringing the number of paper to 10,240.

Figure 2 and Table 2, Table 3, Table 4,Table 5, and Table 6 and provide an overview of the type of items contained in the database by year. Overall, these tables show that an analysis of all items containing abstracts allows a good coverage of the literature, while enabling us to consider the frontier of the scientific debate thanks to the inclusion of conferences.

Figure 2 – AIDS vaccine papers by year

Table 1 – Material and material with abstract per year

All material / Material w AB / % material with abstract
1989 / 6
1990 / 93 / 14 / 15.1
1991 / 285 / 211 / 74.0
1992 / 376 / 265 / 70.5
1993 / 428 / 263 / 61.4
1994 / 543 / 368 / 67.8
1995 / 482 / 383 / 79.5
1996 / 427 / 352 / 82.4
1997 / 488 / 414 / 84.8
1998 / 526 / 430 / 81.7
1999 / 504 / 424 / 84.1
2000 / 561 / 478 / 85.2
2001 / 536 / 465 / 86.8
2002 / 552 / 478 / 86.6
2003 / 607 / 516 / 85.0
2004 / 677 / 586 / 86.6
2005 / 735 / 647 / 88.0
2006 / 693 / 615 / 88.7
2007 / 694 / 617 / 88.9
2008 / 876 / 665 / 75.9
2009 / 161 / 145 / 90.1
Total / 10,250 / 8,336 / 81.3

Table 2 – Type of material in the database

Type of material / %
Article / 6,813 / 66.5
Bibliography / 2 / 0.0
Biographical-Item / 3 / 0.0
Book Review / 16 / 0.2
Correction / 31 / 0.3
Correction, Addition / 11 / 0.1
Discussion / 1 / 0.0
Editorial Material / 443 / 4.3
Letter / 179 / 1.7
Meeting Abstract / 549 / 5.4
News Item / 272 / 2.7
Note / 110 / 1.1
Proceedings Paper / 767 / 7.5
Reprint / 4 / 0.0
Review / 1,049 / 10.2
10,250 / 100.0

Table 3 – Type of material with abstract

Type of material with abstracts / %
Article / 6,549 / 78.6
Bibliography / 1 / 0.0
Correction / 3 / 0.0
Discussion / 1 / 0.0
Editorial Material / 93 / 1.1
Letter / 3 / 0.0
News Item / 15 / 0.2
Note / 82 / 1.0
Proceedings Paper / 659 / 7.9
Reprint / 2 / 0.0
Review / 928 / 11.1
8,336 / 100.0

Table 4 – Type of material without abstract

Type of material without abstracts
Biographical-Item
Book Review
Correction, Addition
Meeting Abstract

1

Table 5 – Type of material per year

Table 6 – Material with abstract per year

56

In order to proceed with the content analysis of the abstracts frequencies of single words are insufficient to investigate research topics that are described by combinations of two or more words, as it is often the case in the AIDS vaccine literature. For this reason, we have adopted the technique based on the analysis of doubles and triples (i.e., two or three words that occur one after the other in the text, e.g., neutralizing antibodies or cell-mediated immunity) used in Corrocher, Malerba and Montobbio (2007). However, we first conducted a visual exploration of frequencies and co-occurences of words, with distances up to 10 words, using the software package developed by Prof. Elmar Wolff for WP2. Further, we interviewed five scientists involved in the search asking them which words and expressions would enable us to identify the type of research carried out. This gave us a first idea of what words were most prominent.