Why Analytics by Themselves Do Not Change Processes

Decision-making
Research suggests 40% of major business decisions are based not on facts, but on the manager’s gut.1 Hopefully, most of those gut-based decisions are not life endangering. Yet, many of them directly impact organizational viability.

In Workers’ Compensation, critical decisions are made daily by front-line workers. Claims adjusters and medical case managers make course-swerving decisions every day. On what basis do they make these decisions? What kind of decision support is available to them? How timely are the decisions they make? How are outcomes traced back to the decisions made? Are any of the decision based on analytics?

Only analytics fused into operations can respond to these questions. Only analytics that are linked to operations can consistently provide good decision support and positively affect outcomes.

Distributed knowledge
“If you really want to put analytics to work in an enterprise, you need to make them an integral part of everyday business decisions and business processes—the methods by which work gets done and value gets created.”2

Both garden variety and highly sophisticated analytics are common now in many organizations. A few apply the analytics to their operational process effectively, thereby making significant impacts on outcomes and profitability. Unfortunately, those exceptions are not often found in the Workers’ Comp industry. Applied analytics in Workers’ Comp, and particularly those relating to the medical aspect of claims, is rare.

In Workers’ Comp, analytics are most often sequestered on the executive floor. Analytic results are displayed at board meetings and are lavishly portrayed at marketing shindigs. They are represented in colorful graphics while decision-makers ponder them. Nevertheless, just executing and reviewing analytics has little impact on outcome. Analytics must be linked to operations seamlessly to the end that processes are changed and changes are documented.

Dashboards—an attempt that falls short
Dashboards have become a fashionable way to display analytics, but they don’t necessarily link the analytics to operations. They are designed to present conditions in the organization across a broad swath of indicators in one view.

An example is a hospital where a dashboard displays vital operational statistics including admissions and discharges for the period, average lengths of stay, acuity rates, and mortality rates. Dashboards are interesting and informative of activity and organizational performance, but what will or should be done operationally to influence the indicators going forward is not always clear. Results may vary depending upon the leader present that day.

Basically, dashboards are for viewing, and unless the organization has designed response procedures and accountable persons, the impact is negligible. Dashboards have no direct relationship with operations and usually there is no mechanism for tracking action responses to the information.

Changing processes in this top-down manner is difficult, time-consuming and often inaccurate and costly. Corporate communications, regardless of how sophisticated, do not effectively translate analytic knowledge into actions on the front line.

Actionable analytics
For analytics to be actionable they must be linked to, and fused into operations electronically. The best way to do this is to continuously monitor current and historic data, execute the analytics in real time, and initiate the desired actions among workers by means of an electronic message. This approach hurdles the communication log jam found in most organizations with an immediate, specific directive. It requires a computer system designed to monitor and analyze all transactions and to automatically send the action message, thereby communicating the results of current analytics to the appropriate persons.

Computer-aided analytics
Design and build intelligent systems designed to monitor the data continuously and identify data combinations that portend risk. Set up rules-based profiles defining data combinations that can be captured by the computer. IT will build the system or outsource to a vendor that specializes in this service. The latter is usually a much less expensive and quicker option. Outsource to a turn-key solution provider.

Infused accountability
When the computer finds a claim containing the data elements in a profile, the appropriate person is automatically notified electronically. The system should also keep an audit trail noting all claims tagged, the reason it was tagges, and to whom the alert was sent. The end-to-end process will infuse analytics into the process, render the process more efficient, and keep everyone accountable.

It’s true, analytics by themselves will not change processes. But analytics linked to operations with built-in tracking systems will.

MedMetrics is an Internet-based Workers’ Compensation analytics company that provides the services described here. You are invited to read other articles dealing with issues of Workers' Comp medical cost management. Click MedMetrics Blogs.

1. Davenport, T. Harris, J., and Morison, R. Analytics at Work, Smarter Decisions, Better Results. Harvard Business School Publishing Corporation. 2010.
2. Ibid.