In workers’ compensation, the medical provider network philosophy has been in place for years. Most networks were developed using the logic that all doctors are essentially the same. Rather than evaluate performance, the focus was on obtaining discounts on bills, thereby saving money.
Physician selection by adjusters and others has frequently been based on subjective criteria. Those include familiarity, repetition, proximity and sometimes just assumption or habit. Often the criteria is something as flimsy as, “We always use this doctor,” or “The staff returns my calls.” The question is, which doctors really are best, and why?
The first assumption that must be debunked is that discounts save money. Doctors are smart—no argument there. So to make up the lost revenue for discounted bills, they increase the number of visits or services to the injured worker or extend the duration of claims by prolonging treatment. To uncover these behaviors, examine the data.
Amazingly, even doctors do not always make the best choices about other doctors. They may recommend doctors they know socially, professionally or by informal reputation, but they may not know how the doctors actually practice. They may not know a physician upcodes bills, dispenses medications or over-prescribes Schedule II drugs. The data will reveal that information.
Doctors may be unaware they are adding to claim complexity by referring to certain specialists. Again, familiarity and habit are often the drivers. On the other hand, duplicity among providers is fraudulent behavior, and it can be uncovered by examining the data.
Analysis of data can expose clustering of poorly performing, abusive or fraudulent providers referring to one another. The analysis may also divulge patterns of some providers associated with certain plaintiff attorneys.
Treating doctors influence claims and their outcomes in other ways. Management indicators unique to workers’ compensation such as return to work, indemnity costs and disability ratings can be analyzed in the data to spotlight both good and poor medical performance. These outcome indicators are either directed by or influenced by the physician, and they can be uncovered through data analysis.
Claims adjusters and other non-medical persons simply cannot evaluate the clinical capability of medical providers, especially doctors. Performance analysis must take place at a higher level. Evaluations for specific ICD-9 diagnoses and clinical procedures such as surgery must be made. Frequency, timing and outcome can be examined in the data in context with diagnoses and procedural codes, thereby disclosing the excellence or incompetency of physicians.
Negative clinical outcomes that can be analyzed include hospital readmissions, repeated surgery or infection. Physicians associated with negative medical outcomes should be avoided.
When analyzing clinical indicators for performance, care should be taken to compare only similar conditions and procedures. Without such discrimination, the results are dubious. Specificity is critical.
When using data analysis to find the best doctors and other medical providers, fairness is also important. Provider performance should be compared only with similar specialty providers for similar diagnoses and procedures. Results will not be accurate or reliable if performance analysis is not apples-to-apples.
Medical providers may question data analysis to evaluate performance, claiming they treat the more difficult cases. The data can be analyzed to determine diagnostic severity, as well. Diagnostic codes in claims can be measured and scored, thereby disclosing medical severity.
Now is the time to step up to a much more dignified and sophisticated approach to selecting medical providers. Decisions about treating physicians must be based on fact, not assumption or habit. Fortunately, the data can be analyzed to locate the best-in-class and expose the others.