The effect of publication on Internet-based solicitation of personal-injury litigants

Methods

On Mar. 1, 2006, the New England Journal of Medicine released ahead-of-print a pharmacoepidemiologic study demonstrating that the antibiotic gatifloxacin (Tequin, Bristol-Myers Squibb) was strongly associated with the development of hypoglycemia and hyperglycemia among outpatients aged 66 years and older.4 This study was released 14 days after the manufacturer and the US Food and Drug Administration jointly strengthened warnings about the drug and 29 days before the article's full publication on Mar. 30, 2006. On May 1, 2006, the manufacturer ceased marketing the drug worldwide.

Beginning on the date of acceptance of the manuscript (Feb. 9, 2006), we conducted daily Internet searches for websites offering legal representation to people who had possibly been injured by Tequin. Using the Internet search engine Google, each of the authors determined the number of “hits” each day using a structured search strategy that was piloted and refined in the weeks before the manuscript's acceptance. For the purposes of this study, we defined a “hit” as any website that contributed to the total website count generated using the search strategy, regardless of whether duplicate or highly similar websites were included. The goal of our search was specificity rather than sensitivity.

Results

Figure 1 shows the mean number of websites that we identified during the study period. About 20 websites were identified each day during the week following the manuscript's acceptance. This number doubled in the week before the article's early release, presumably as a result of the warning issued by the manufacturer and the US Food and Drug Administraion.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1: Average number of medicolegal Internet hits before and after publication of an article4 in the New England Journal of Medicine about the adverse events associated with gatifloxacin.

Within a week of the article's early release, about 100 Internet hits were identified each day. This number increased to more than 300 hits within 1 week of full publication in print. Following a transient decline, the manufacturer's withdrawal of the drug was associated with an additional increase in the number of hits to about 400 per day. Within-rater agreement for observations throughout the study interval was excellent (intraclass correlation coefficient –0.814).

To investigate the durability of this phenomenon, we repeated the search 1 year following the article's early release (Mar. 1, 2007). At that time, we identified 522 websites that were soliciting personal-injury claimants.

Interpretation

We found that the publication of a study concerning the adverse drug events associated with gatifloxacin led to a rapid, dramatic and sustained increase in Internet-based solicitation for litigants for personal-injury claims. Because we had a baseline measure of Internet activity (i.e., the period before the article's early release), we were able to document the magnitude of the change in Internet-based medicolegal solicitation.

A detailed review of the contents of each identified website every day would have been prohibitive; however, we examined many of the websites on an informal and periodic basis. Most websites appeared to be for law firms in North America (mainly in the United States), possibly because we restricted our search to English-language websites. Many sites offered representation for a wide array of drug-and device-related lawsuits (e.g., Vioxx, Accutane, Paxil, Ortho-Evra), and the websites invited prospective claimants to submit details of their cases online.

Some limitations of this study deserve mention. Although our search strategy was chosen for its specificity, it is possible that some hits did not reflect solicitation of personal-injury claimants. In addition, some hits may have represented duplicate links to the same group of attorneys, and the number of websites that were not identified by this strategy is unknown. The period of baseline Internet activity before the article's online release was relatively short (20 days) as a result of the expedited publication process. However, this is of little consequence given the magnitude of the change in Internet activity and because we did not use the baseline period to generate statistical inferences.

In conclusion, we found that publication of an article in a general medical journal can dramatically influence Internet-based solicitation of litigants for personal-injury claims.