A guide to useful research.

So that you don't have to, I spend a lot of time peering into management and research journals and e-mailing the authors, most of whom respond with grace. Every year, maybe 10 articles appear truly worthy of attention by the workers' comp professional. They cover management, insurance, law, medical care and safety. Because I've been in this field for 20 years, I could list 200. Within this space, however, I will mention but a few.

Research helps when it debunks, steering us away from conventional wisdoms that are really only overroasted anecdotes. It can speed change without chaos. Maybe my diagnosis of terminal optimism causes me to think that every senior executive should read at least the research summaries.

The most productive factory of research on work-injury risk is Liberty Mutual's Research Institute for Safety. You find it west of Boston tucked behind a large commercial nursery. Sometimes, visitors upon the long driveway enjoy a view of nursery workers knocking around trees and dirt with reckless abandon.

The safety and disability researchers there I have always found bright and hardworking, and it shows in the many papers they publish. You can scan their titles and order them online. Go to www.libertymutual.com, then "site map," and Research Institute for Safety is listed under "contacts." From aging to ergonomics to predictors of disability, this site is rich with studies you can use.

The most valuable published study in the past two decades is from the Upjohn Institute for Employment Research in Kalamazoo, Mich. In the early 1990s, researchers there asked if and how safety and workers' comp experience varied among employers within the same industry. They found that some employer safety and injury practices explained some huge variances. They also found, however, how some practices had very little effect. Their findings are valuable for anyone trying to predict, or explain after the fact, work-injury experience variances among similar employers.

The principal study is called "Disability Prevention Among Michigan Employers." To find the paper, you have to enter a back door of the Institute's Web site, at www.upjohninst.org/publications/wp/index. htm#9318. Look for the Michigan disability prevention study.

I've scanned or read hundreds of articles related to workers' comp since 2000. From a sample of 50, I asked professionals in the field to help me find the most useful. Almost all were interested in work by occupational doctors at Johns Hopkins School of Medicine to pilot medical provider networks in Maryland and Louisiana.

Ed Bernacki has led this multistate undertaking. Their reports explain why these networks performed well in reducing disability days and workers' compensation costs. Read the summary, or buy their articles in the January 2005 issue at www.joem.org, located in the archives section.

Failure of workers to file claims is a hot research topic. Two researchers at Michigan State University, Jeff Biddle and Karen Roberts, published an article based on a large survey. Many do not file, even for a disabling injury. Some of them could access other benefits. But that's not the whole complex and engaging picture. If you need to understand claiming behavior, or if you are looking out for the health and safety of large worker populations, get their analysis in the December 2003 issue of the Journal of Risk & Insurance. Google searches will bring up the publication and pertinent articles by other researchers, such as the ones at RAND.

Insurers, I think, should link up with researchers, but that's for another day or another column.

PETER ROUSMANIERE is a Vermont-based writer and columnist for Risk & Insurance[R]. He can be reached at [e-mail].