Lead author Dr. Pascal A. Diethelm, at OxyRomandie in Geneva and colleagues conducted a search of these documents, which were posted on public Websites, as well as other information they found.

Based on their findings, they say that "those involved in reviewing evidence on smoking and health should be aware of what appears to be the selective nature of what is eventually published by some scientists with links to the industry, and the evidence that sometimes mechanisms appear to have been used to disguise these links."

Internal memos that Diethelm's group cites showed that executives of Philip Morris first identified a need for the company to conduct its own biological research in 1968.

The company then bought a research facility in Germany, the Institut fur Industrielle und Biologische Forschung GmbH (INBIFO), which came on the market in 1970. Though 100-percent owned by Philip Morris, the investigators note, the company developed a complex mechanism to ensure that work done by INBIFO could not be linked back to the company.

According to the Lancet article, direct contact with INBIFO was avoided by routing information through another subsidiary, Fabriques de Tabac Reunies (FTR), and a coordinator whose main employment was at the University of Gothenburg in Sweden.

Evidence showed that the company sought to maintain confidentiality over results of research conducted at INBIFO. In one memo from 1977 that the Lancet authors found, a senior Philip Morris executive stated, "we have gone to great pains to eliminate any written contact with INBIFO."

Information was often communicated verbally, on a "strict need to know basis," routed through FTR to avoid any direct contact with Philip Morris, or sent to home addresses where documents could be destroyed, Diethelm and his associates report.

In the 1980s, animal experiments conducted by INBIFO demonstrated high levels of toxicity from so-called sidestream smoke. Diethelm's team traced "more than 800 scientific reports dealing with sidestream smoke undertaken by INBIFO between 1981 and 1989."

However, they add, it was not until 1994 that researchers at INBIFO published research concerning sidestream smoke.

Studies published in scientific journals during this time "appear to be of considerable value to the industry," the investigators write, "casting doubt upon the value of markers of passive smoking and suggesting alternative explanations for the observed epidemiological association between passive smoking and lung cancer."