Richard F. Denning - RMIS Columns

Columns from one of the Leading RMIS Industry Experts

Richard F. Denning

Richard F. Denning is President of Shelter Island Risk Services, a leading RMIS consulting firm based in Shelter Island, New York. You may contact Rick at (631) 749-1535.

"It's the Data, Stupid!: Effective limitations on risk management information systems", 1993

When then candidate Clinton was advised, "it's the economy, stupid", the advice focused him on the truly critical issue in the voters' minds. There are always distractions to confuse us from recognizing such pivotal issues.

As a sixth-grader, I resisted spelling lessons on the argument that "when I go to work, a machine will do that". Unfortunately, spell-checkers did not come on the screen in time to save me from my weekly spelling quiz. But, I was right.

Obviously the information highway is not even completely paved and traffic jams--even bridge collapses--are common place. Nevertheless, these technologies are moving in the right, user-friendly direction. Five years ago, you needed a computer sciences degree to install software. Today, the only requirement is the ability to read and faith that the computer knows what it is doing.

Walk into a modern computer superstore and you are bombarded by the 'big three'--hardware, software and communications. In risk management, we have experienced and explosion of information services in each of these areas. There are literally hundreds of firms offering various information products.

As technology has expanded, the RMIS systems used in risk management have become lower priced, more consistent in their features and even similar in the way they operate. They are becoming commodities. Using the annual RIMS convention as a guide, in 1980 a computer was a draw ("wow, you can do that!"); but, by 1990, a yawn ("ah, when did you say lunch was...").

Yet it is not the technology that will differentiate the service providers of the future. New bits and bytes from Silicon Valley of Bothel Washington will dominate headlines, but the critical issues are the scope and quality of the data.

The insurers and tpa's have accepted the once radical notion that the information they provide is as valuable as their operations, such as claims handling. The content, accuracy and accessibility of the data are the critical battlefronts for those who are using information as a component of their services.

Given this importance of data quality, how do you measure it? The insert highlights some criteria to monitor your program. Although bench making is much in vogue, I believe the most reliable bench making source for any organization is its own experience. Competing with yourself (continuos quality improvement) was strongly endorsed by Dr. W.E. Deming as the most reliable and objective for monitoring performance. The trends shoeing improvements in year to year are the best measure available.

The flow of data from the source to the computer database can have a significant impact on quality. The problems of coding usually start with the original documents from the field. Unclear, unfamiliar and bureaucratic forms and procedures can start a workers' compensation claim down the expensive path to lawyer involvement. An alternative is 1-800 telephonic services. A phone call replaces a form and all the frustrating procedures surrounding paper systems. But the key benefit of telephonic systems is their ability to get the right people in contact with the claimant quickly.

Kemper and Travelers have published their results on the savings in average claim settlements when even a few days are cut in contacting claimants. Today 50% of workers compensation claims are initiated through such telephonic services.

Data quality must be considered in light of the value of the resulting information. Data collected, but never used, is irrelevant. Indeed there is a positive correlation between use of data and its accuracy. The more data is relied upon, the more its errors become obvious. In the ideal, the data is collected as an auxiliary process to that person's routine processing responsibilities. For and example, consider an automobile adjuster who can directly transfer photo images of damage to the computer-based claim file. Authorized users, even thousands of miles away, have near immediate access to these photo images.

Not collecting data at all can improve quality! How, you ask. Simply by building the link to the needed data when such data (for example, employee personnel information is available more accurately elsewhere).

In summary, you can improve data quality in three ways:

1. Use it.
2. Collect in the normal course of someone's responsibilities.
3. Find the best source available; avoid redundancy.

The Holy Grail of Risk Management in the 1990's is saving money. Risk Management is no more exempt from cost containment pressures than any other corporate department. The attention on cutting cost of risk via lower expense and smaller risk financing charges has squeezed these areas to a point of diminishing returns. The paragraphs that follow illustrate dollar & cents payoffs from improved data quality.

Financial data in claim records (reserves, payments, allocated expenses, fees, etc.) are the most scrutinized information. Yet, from a Risk Manager's Perspective, there are many questions to examine. Retrospectively-rated premiums often are manually calculated even by some of the largest insurers. Having an accurate and flexible database of claims gives the Risk Manager the opportunity to verify all the calculations used to compute a retro-premium. The effort may prove so cost effective, that its savings pays for the cost of the RMIS.

Defining coverage is a simple enough task. Now one would confuse an automobile case with a workers' compensation claim. Or is it really so easy? There are practical pricing issues involved as, for example, when a workers' compensation medical-only claim is mislabeled as an indemnity case. Another example: are the number of closed no payment cases reasonable and in line with historical trends. Since fees are often based on the coverage type, consider how you can test your data to confirm that the claim characteristics match the indicated coverage.

Fundamentally, it is an exercise in expert systems to test this logical connection between coverage and claim characteristics. This type of checking has broader applications--for example, in testing the reasonableness of outstanding reserves or the appropriateness of risk control codes.

Reducing the cost of risk means cutting losses. Effective risk control must be in the plan for every Risk Manager. Unfortunately, risk control represents the most severe data challenges of the areas above. The claim administrator often plays the foot soldier's role in collecting this data. Yet from their perspective it is not used, not edited, and rarely appreciated in terms of potential down-stream usefulness.

The problem is compounded by the lack of communication between the risk control/ergonomic specialists and the claims administrators. The challenge is also reflected in the fact that claims data alone is not enough to study the opportunity for reduced claims. One needs access to the demographics of entire work force to monitor characteristics of workers with claims and those characteristics associated with claim-free workers. We have the statistical analysis methods; we just need the data to run them.

From a distance, objects have a look of perfection that is lost as one approaches closer and closer. Data has the same character. It's apparent accuracy and completeness break down under the scrutiny of use. We have no need for data we do not use. Use of inaccurate or incomplete data can lead to embarrassment, errors and sometimes disaster (read any good spy novel). The starting point for improvement is to determine the quality of what is already being provided. When you have this baseline, then you have an objective means to measure the rate and nature of improvements.

Copyright© 1996 R. F. Denning & Associates

| RMIS-Web |
| Business Insurance Review |
| Articles |
| Columns |
| Press Release |
| Directory of Software Vendors |
| Directory of RMIS Consultants |