Lawsuits have been filed against tobacco companies for more than 50 years. Three distinct waves of tobacco litigation have been identified. The first wave began when scientific evidence linking smoking to cancer was first published, and continued to the early 1970s (1954 to 1973). The second wave occurred from 1983 to 1992 and included the landmark Cipollone v. Liggett Group, Inc case. Neither of these waves of litigation resulted in payments to plaintiffs. The third wave began in 1994 when the class action Castano v. American Tobacco Company was filed.¹ It continues to the present, and has included thousands of lawsuits of many different types, including several that have yielded large payments to plaintiffs from settlements or jury verdicts.²

Law professor Richard Daynard, chair and founder of the Tobacco Products Liability Project (which has supported tobacco litigation since 1984), has identified seven achievements from this legal activity: (1) it has resulted in the disclosure of millions of tobacco industry documents, providing great fodder for research and advocacy; (2) large verdicts in several cases “have added to the industry’s confusion and loss of legitimacy”; (3) the cases have stimulated “the first stirrings of responsible behaviour” among the tobacco manufacturers; (4) tobacco companies’ huge legal costs have forced them to increase the price of their cigarettes, which has in turn contributed to lower smoking rates; (5) the state attorney general lawsuits and their settlements (including the Master Settlement Agreement) resulted in large payments from tobacco companies to the states, which in a few cases have been used for tobacco control; (6) tobacco litigation has served as a model for litigation against the manufacturers of other harmful products, such as handguns and “fast foods”; and (7) tobacco lawsuits in other countries besides the United States have resulted in judicial rulings that have curbed smoking in public places and illegal tobacco company marketing practices.³ One additional benefit of tobacco litigation is that it has produced a new and valuable “database” for research—millions of pages of transcripts of testimony in tobacco lawsuits, many of which now comprise the Tobacco Deposition and Trial Testimony Archive (DATTA).

ORIGINS OF THE TOBACCO DATTA PROJECT

On 17 July 1998, President Bill Clinton issued an executive memorandum directing the Secretary of Health and Human Services “to coordinate a public health review of tobacco industry documents and develop a plan to make the documents more accessible to researchers and the public at large”.⁴ Further, the Secretary was asked to:


 Provide a strategy for coordinating a broad public and private review and analysis of the documents to gain critical public health information. Issues to be considered as part of this analysis include: nicotine addiction and pharmacology; biomedical research, including ingredient safety; product design; and youth marketing strategies.⁵

As one component of the Department of Health and Human Services’ response to those directives, the National Cancer Institute (NCI) launched a programme in June 1999 to fund tobacco documents research. According to the NCI’s first “program announcement” soliciting grant proposals, the purpose of the programme was:


 … to stimulate research on a wide variety of scientific, technical, marketing and tactical undertakings by the tobacco industry, which were documented in papers, memos, and other records. The systematic, comprehensive analysis and evaluation of these documents will enhance understanding of the tobacco industry’s knowledge, strategies and tactics, provide a greater understanding of the determinants of tobacco use and addiction and help researchers and public health practitioners identify effective strategies to prevent and reduce tobacco use.⁶

In March 2001 the NCI issued a second programme announcement for its tobacco documents research programme.⁷

Since the programme’s inception, several funding cycles have been used to review grant proposals, and many grants have been awarded. Most of the research projects funded by the NCI have analysed the previously secret tobacco industry documents that are now available in various depositories and on many websites (such as http://tobaccodocuments.org and http://legacy.library.ucsf.edu). The Tobacco DATTA project, funded in the NCI’s first cycle of grants for tobacco documents research, is unique among these NCI-funded studies in that it analyses a different (but complementary) set of documents—transcripts of testimony in tobacco-related lawsuits.

Two of us (RD and CD) have been involved in tobacco litigation—one (RD) as an expert witness and the other (CD) as a consultant to, or co-counsel with, law firms involved in tobacco-related lawsuits. We were thus aware of the valuable, but often inaccessible, information in depositions and trial testimony within tobacco litigation. To make that information easily accessible, and to begin to analyse it, we crafted a grant proposal responsive to the NCI’s 1999 programme announcement. In September 2000, the NCI awarded a grant to the Michigan Public Health Institute (MPHI), initiating the Tobacco DATTA project. The American Legacy Foundation provided continuation funding in 2004.

WHY STUDY LITIGATION TESTIMONY?

The study of previously secret tobacco industry documents has become a sizable field of inquiry. A bibliography of reports and articles based on these documents includes more than 450 items.⁸ Research on these documents can be enriched by also assessing related testimony in DATTA. Litigation transcripts included in DATTA relate to the industry documents in at least three ways: (1) many of the authors of the latter documents have been asked to elaborate on them in the course of depositions and trial testimony, while under oath, creating a richer context in which the documents can be considered; (2) tobacco company executives, scientists, and consultants have been asked to comment on industry documents in depositions and trial testimony, providing a contemporaneous record of their awareness of and opinions about specific content in the documents; and (3) plaintiffs’ witnesses have been asked about industry documents in depositions and trial testimony, providing expert opinion about the documents and an indication of which documents are considered by the expert witnesses to be most important.

Besides the examination of industry documents, depositions and trial testimony provide three other types of useful information: (1) tobacco industry witnesses have been asked about industry research, providing details beyond those found in the company documents; (2) questions posed to tobacco company deponents and trial witnesses help reveal their knowledge and opinions (past and current) about nicotine pharmacology and addiction, the health consequences of tobacco use, advertising and promotion, and other areas of interest to researchers, clinicians, policymakers, and health advocates; and (3) questions posed by tobacco company lawyers to plaintiffs and plaintiffs’ witnesses help reveal tobacco company opinions, arguments, and strategies.

The industry documents might be compared to the frames of a Hollywood film. The individual documents—and the individual frames of a film—are revealing and intriguing. But the deposition and trial testimony—just like the movie camera—allows us to connect the pieces so that the big picture begins to take shape. Then we can understand more fully how tobacco products are designed, how they are marketed, how smoking behaviour is controlled, and how tobacco companies have disrupted efforts to reduce tobacco use.

AIMS OF THE TOBACCO DATTA PROJECT

The intended outcomes of the Tobacco DATTA project were: (1) the creation of a new, rich, accessible, user-friendly database pertaining to tobacco industry documents, research, knowledge, conduct, and strategies; and (2) research articles and reports, based on the study of that database, concerning nicotine addiction and pharmacology, the health consequences of tobacco use, tobacco-product design and manufacture, advertising and promotion, and other topics relevant to tobacco control.

After the project began, the architecture for DATTA was built and incorporated into Tobacco Documents Online (TDO) (http://tobaccodocuments.org), where many tobacco industry document depositories are housed. TDO offers several tools for tobacco documents researchers, including powerful search capabilities across multiple document depositories, high-quality images, conversion of document images to searchable text through optical character recognition (OCR) software, and the ability to collect and annotate documents. The DATTA site (http://tobaccodocuments.org/datta) is integrated with other collections on TDO, and allows searching by keywords and browsing by witness, case (lawsuit), topic (for example, ammonia, heart disease, second-hand smoke), persons and organisations mentioned in the transcripts, and tobacco-product brand names mentioned in the transcripts.

Since 2004, transcripts in DATTA have been added to the Legacy Tobacco Documents Library (http://legacy.library.ucsf.edu), a digital library of internal tobacco industry documents maintained by the University of California, San Francisco. Inclusion of DATTA transcripts in this library provides additional access and security to this unique collection.

After the first key goal of the DATTA project was achieved, project resources focused on the research agenda. Articles published in this journal supplement represent the product of that effort. The processes and methods used to organise and carry out that research are described below.

CONTENT OF DATTA

Types of lawsuits

In the initial formulation of the DATTA project, we decided to include tobacco lawsuits filed in US courts since 1990, including the following: (1) state attorney general lawsuits (and similar cases filed by counties and cities); (2) class actions (for example, the Castano series of class actions; Broin v. Philip Morris Companies, Inc.; Engle v. R.J. Reynolds Tobacco Co); (3) personal injury cases filed on behalf of individual plaintiffs (including the Kent Micronite filter cases and smokeless tobacco cases); (4) health care cost recovery cases patterned after the attorney general cases (for example, those filed by union-managed health and welfare funds and Blue Cross Blue Shield health insurance plans); and (5) other, miscellaneous cases, such as secondhand-smoke cases filed by individual plaintiffs, the Mangini v. R.J. Reynolds Tobacco Co, et al case on tobacco advertising, cigarette fire damage suits, litigation involving federal agencies (for example, US Department of Justice, US Food and Drug Administration, and US Environmental Protection Agency), and international/foreign claims filed in US courts. We have excluded business-related lawsuits against tobacco companies that do not have public health implications.

The 1990 cut-off was somewhat arbitrary, but was used to include all of the lawsuits filed in the third wave of tobacco litigation (which began in 1994, as noted above), and to help keep the scope of the project from expanding excessively. In addition, we assumed that the transcripts from the older cases would be more difficult to obtain. Nevertheless, we recognised that the older cases have substantial historical value (especially the Cipollone case), and that unusual pre-1990 cases such as Kyte v. Store 24 (regarding youth access to cigarettes) might be worth including. We adopted two other exclusions initially: (1) administrative hearings, such as those pertaining to the worksite smoking rules proposed in 1994 (but never finalised) by the US Occupational Safety and Health Administration (OSHA); and (2) tobacco lawsuits filed in the courts of foreign countries.

These exclusions were followed mainly to limit the scope of this project, which was already quite large. However, the exclusions were never considered to be absolute and unalterable. If, during the course of the project, important materials were made available to the project from pre-1990 lawsuits, administrative hearings, or lawsuits filed in foreign courts, and if resources were available, then we reserved the option to include some or all of these materials. Indeed, we have made several exceptions to these exclusions:

• More than 230 transcripts from the Cipollone case have been added to DATTA because of their historical interest and the importance of some of the testimony given in that case by tobacco industry witnesses.

• More than 90 transcripts from JTI-MacDonald Corp. [RJR-MacDonald Inc.] v. Canada, a lawsuit that challenged Canadian legislation banning tobacco advertising and promotion, have been added to DATTA because of their relevance to scientific and political issues concerning the marketing of tobacco products.

• More than 300 transcripts from OSHA’s administrative hearings for its proposed rules on workplace smoking have been added to DATTA. This material is a valuable addition to DATTA because of the paucity of lawsuits and DATTA transcripts focusing on second-hand smoke.

Sources of documents

We obtained transcripts and other documents for DATTA from five main sources: (1) law firms involved in tobacco litigation; (2) court reporter firms that transcribe testimony at depositions and at trial; (3) individual lawyers and witnesses who have been involved in these cases; (4) the depository of discovery items created by the settlement of the Minnesota Attorney General’s lawsuit; and (5) tobacco company websites (as available in TDO document collections). The Motley Rice law firm (http://motleyrice.com), which assisted 26 state attorneys general in their lawsuits against tobacco companies, contributed to DATTA several hundred transcripts from the firm’s own collection.

When possible, transcripts were obtained in electronic form, which facilitated the uploading of properly formatted and accurate versions of the documents to the DATTA website. Documents that were available only in hard copy were scanned and then uploaded to the DATTA website, where they were processed with OCR software. Transcripts appear on the DATTA website as original text files or PNG (portable network graphic) files accompanied by OCR text. Transcripts may also be viewed as PDF (portable document format) or TIFF (tagged image file format) images. Transcripts may be downloaded as text files (either original or OCR text) or PDF or TIFF images. About 45% of transcripts in DATTA are in original text format, with the balance in an image format accompanied by OCR text.

The OCR text is relatively accurate when the original hard copies contained full-sized and clear type. However, the OCR text contains errors at varying degrees when the clarity of the type was poor, when type size was very small, and when stray marks or tobacco company-imposed stamps or watermarks appeared on the hard copies. An example of a company-imposed stamp or watermark can be found at http://tobaccodocuments.org/datta/ABERLED111000.html.

Types and numbers of documents

Most of the documents in DATTA are transcripts of depositions and trial testimony for individual witnesses. In some cases, transcripts of opening and closing statements have been obtained and added to DATTA. When available, we have included in DATTA deposition and trial exhibits and “expert reports” prepared by expert witnesses. Exhibits are items that the witness is asked about, such as journal articles, scientific reports, letters and memoranda, medical records, internal industry documents, the witness’s curriculum vitae, and so on. In some legal cases, expert witnesses are asked to produce an expert report covering the scope of their testimony before they are deposed or testify at trial; attorneys for the plaintiffs and defendants usually focus on these reports during the direct examination and cross-examination of expert witnesses.

As of 3 March 2006, the publicly available portion of DATTA contained 4850 transcripts, which included a total of about 820 000 transcript pages. The transcripts spanned the period from 1957 to 2005, but 85% of the transcripts were for testimony from 1990 to 2005. Transcripts covered testimony presented by more than 1500 witnesses in a total of 232 cases. The number of lawsuits filed against the tobacco companies greatly exceeds 232,² but only cases that have progressed to the point where depositions were taken, or that have gone to trial, have been targeted for inclusion in DATTA.

Transcripts for testimony by witnesses were evenly divided between those for expert witnesses (49.6%) and those for fact witnesses (50.4%). (An “expert witness” is a person the judge allows to express opinions that the jury or judge may consider. A “fact witness” is allowed to report only what he or she knows from observation of some kind and is not (or should not be) permitted to offer opinions.) About 55% of all transcripts—and 48% of expert witnesses’ transcripts—were for witnesses appearing for the defence.

A restricted collection on the DATTA website includes transcripts containing sensitive personal information that was never presented in open court. These transcripts are made available to DATTA project researchers provided that they are not used in a manner in which the sensitive information could be linked publicly to a specific individual. In some cases, DATTA project staff have redacted sensitive personal information on a transcript, thereby allowing the transcript to be placed in the publicly available collection.

METHODS FOR CONDUCTING DATTA RESEARCH

The large number of transcripts and transcript pages in DATTA presented a rather daunting challenge for applying research to this dataset. The first step in carrying out our research programme was to identify topics for analysis, and to recruit small teams of researchers to lead studies in those topical areas. A few of the topics on the original list were dropped early on, because the DATTA transcripts provided minimal coverage of those areas. Ultimately research teams covering the following topics conducted research on DATTA transcripts:

• Health consequences of tobacco use

• Health consequences of exposure to second-hand smoke

• Addiction and pharmacology

• Tobacco advertising and promotion

• Tobacco-product design and manufacture

• Economic impact of tobacco use

• Youth initiation of tobacco use

• Tobacco control policies and policy research

• Public understanding of the risks of tobacco use and exposure to second-hand smoke

• Tobacco industry attacks on (or harassment of) people—plaintiffs and their families and friends, and scientists

• Tobacco industry attacks on research methods and the science of epidemiology

• Other tobacco industry activities (for example, lobbying, public relations activity, political action, research, philanthropy).

Research assistants working in the DATTA project office processed each transcript by preparing an abstract summarising the testimony, indexing standardised information about the testimony (“metadata”), and coding any portions of text that corresponded to the research topics listed above. Coding of text according to research topic was guided by a detailed document describing the scope of each topic. The metadata included the following: title of the transcript; witness’s name; document type (deposition, trial testimony, or opening or closing statement); document length and page range for the relevant testimony in the transcript; date of testimony; name of case (lawsuit); type of witness (fact or expert witness); witness’s affiliation (testifying for the plaintiff or defendant); witness’s employer and title; subjects covered; persons, organisations, and tobacco-product brands mentioned; and the research topics covered (from those listed above), along with corresponding transcript page numbers. Where possible, indexing fields and terms were consistent with the fields endorsed by the UCSF (University of California, San Francisco) Tobacco Control Digital Archive⁹ and the subject terms (keywords) used in the UCSF/ANRF (Americans for Nonsmokers’ Rights Foundation) Tobacco Documents Thesaurus.¹⁰

Each research team member received one or more CD-ROMs containing all DATTA transcripts that had text coded for the topic assigned to his or her team. The CDs also contained a file for each transcript with the indexing data for that transcript. The project gave each researcher a copy of winMAX 98 Pro software for qualitative data analysis (Scolari Sage Publications, Inc). Although some research teams used this software and others did not, the expectation was that all of the teams would apply qualitative research methods to their studies.

The researchers were asked to attend one of three planning meetings held in June and October of 2003. For the most part, members of a given research team attended the same meeting. These planning meetings were used to discuss common approaches to the study of DATTA transcripts across research teams, and to explore ideas for the analyses to be conducted by each research team.

In order to foster a common approach among the research teams, project leaders encouraged the teams to address one or more of the following research questions:

• What are the major themes or arguments emanating from tobacco industry-affiliated witnesses or lawyers in this topical area during the course of litigation testimony?

• How do these themes and arguments compare to scientific facts (for example, conclusions of the US Surgeon General and other authorities)?

• How do these themes and arguments compare to what the industry says to the public (via tobacco company websites, press releases, informational advertising, etc)?

• How do these themes and arguments compare to what the industry says in its internal (previously secret) documents?

• Have these themes and arguments changed over time—for example, before and after new admissions on company websites, or before and after adoption of the Master Settlement Agreement (MSA)? (The MSA between the tobacco companies and 46 attorneys general, which was signed in November 1998, prohibits any “material misrepresentation of the facts regarding the health consequences of using any tobacco product”.)

• Do these themes and arguments vary according to geography (where the case has been filed)?

• Are different themes, arguments, and strategies used in depositions versus at trial?

• Do these themes and arguments vary in short versus long trials? Which ones rise to the surface when time is limited?

The amount of text in DATTA transcripts that was coded as pertaining to a particular topic (and research team) was, in many cases, enormous. As a result, the research teams—in discussions during and after the planning meetings—explored strategies for narrowing the scope of their investigations. Different teams adopted different strategies for making their analyses feasible in light of time and resource constraints. One team, for example, focused on a specific type of litigation (for example, personal injury cases brought on behalf of individual plaintiffs). Other teams restricted their analyses to a small number of important witnesses. To aid those teams, members of the DATTA Project Advisory Committee who have had experience in tobacco litigation identified key tobacco industry witnesses whose testimony might warrant special study.

Some research teams decided to confine their research to trial transcripts, thereby excluding depositions. One reason for that preference is that defendants’ arguments in trial testimony are clearly meant for juries and judges, whereas industry arguments in depositions may serve other purposes (for example, to explore the opinions of the plaintiff’s expert witnesses—for possible use within, or even outside, the current case). Other research teams constricted their “dataset” even further by studying only the opening and/or closing statements presented at trial. Another practical approach was to define and pursue a research question that was much more narrow than (but still within) the overall topic assigned to the research team. The specific methodologic approaches used by the various research teams are described in other articles in this journal supplement.

On 1 June 2004, the DATTA project and the Tobacco Products Liability Project at Northeastern University School of Law co-hosted a symposium at which DATTA research teams presented their preliminary findings. The symposium was held in San Diego just before an NCI Tobacco Control Investigators Meeting. This event gave the DATTA researchers the opportunity to exchange information about the challenges, methods, and results of their research. A project listserve allowed for continued communication across research teams and between project leaders and researchers.

During succeeding months, the research teams completed their studies and submitted manuscripts for publication in this journal supplement. A peer-review process was used to evaluate those papers, and decisions on acceptance of the papers were made by the guest editor of the supplement, Stella Aguinaga Bailous.

NEXT STEPS

The research conducted thus far on DATTA transcripts, as reflected in the papers published in this journal supplement, demonstrates the value of studying this material. Nevertheless, much more analysis of this dataset is needed to tap its wealth fully. We encourage tobacco researchers to seek grant funds to conduct further investigations using the documents in DATTA.

To support continued research, we will add transcripts from current and future tobacco litigation to DATTA to maintain its currency, and we will fill gaps in the archive to the extent possible. Contingent on the availability of resources and partnerships with colleagues in other countries, we would like to add material to DATTA from tobacco litigation in other regions of the world.

We are working to improve the functionality of the archive. For example, we are adding web links to transcripts for easy retrieval of key documents mentioned in testimony. A link to “A Frank Statement to Cigarette Smokers” (http://tobaccodocuments.org/misc_trial/EXHIB2_70005662-5662.html)—perhaps the most frequently cited document in tobacco litigation—has been added to more than 750 DATTA transcripts. In addition, we have cross-linked 86 expert reports with the trial testimony or depositions of the authors of those reports. As of 12 March 2006, we had created more than 1650 links to 175 key documents and reports within DATTA transcripts.

CONCLUSION

Easily retrievable information on tobacco is growing in amount and variety. Research data are available through online journals and reports and bibliographic databases such as PubMed/Medline. Tobacco industry documents are available at Tobacco Documents Online (TDO), the Legacy Tobacco Documents Library, and other websites. Collections of tobacco advertisements appear on TDO (http://tobaccodocuments.org/advertising/) and the Campaign for Tobacco-Free Kids’ website (http://tobaccofreekids.org/adgallery/). Now, through the Tobacco DATTA project, testimony from tobacco litigation is becoming increasingly accessible.

Another potentially rich database, which would add one more piece to the tobacco control puzzle, remains to be constructed and analysed—testimony by tobacco industry representatives and allies at hearings held by Congress and other legislative bodies. As the puzzle is completed in the coming decades, we will have a comprehensive chronicle of the rise and fall of tobacco use, and the factors that promoted and impeded the avoidable epidemic of tobacco-attributable disease.