Classic risk assessment is usually based on population, using basic epidemiological studies and a managed care organization's demographic data to create estimates of disease and subsequent cost risks. For example, the incidence of heart disease in the general population may be x percent, which would be modified (higher, in this case) if your MCO has an older membership with a higher proportion of men.
While risk assessment is well established as an underwriting tool, it is less established for medical interventions and is in its infancy as a pharmacy intervention tool.
With pharmacy cost trends approaching vertical and hundreds of new drugs in the pipeline, many MCOs are struggling to ensure appropriate pharmaceutical utilization while containing costs. Pharmacy-based risk assessment will help HMOs and other health plans to manage future utilization and costs.
To date, MCO responses to increasing drug costs have been reactive, not proactive. Multiple-tier formulary designs, rediscovery of coinsurance as a benefit design, and prior authorization are emerging as strategies for controlling drug costs. The assumption is that the increased utilization of pharmaceuticals is always bad, but clearly, evidence-based medical studies, such as the Beta Blocker Heart Attack Trial (BHAT) and Metoprolol In Acute Myocardial Infarction (MIAMI) studies suggest otherwise for a significant number of diseases.
A typical example of risk assessment from the pharmacy perspective is the recent introduction of Cox-II specific inhibitors. These new nonsteroidal anti-inflammatory agents are used for various acute and chronic conditions. While NSAIDs as a class are relatively inexpensive — often averaging 16 cents a day — high volume and near-ubiquitous use makes them a significant part of a pharmacy budget. Even before the release of the Cox-II inhibitors, NSAIDs were often in the top 10 of an MCO's pharmacy-budget expense.
At nearly $2.40 a day, replacement of the older NSAIDs with the Cox-II inhibitors could cost an MCO millions more annually with no difference in efficacy.
Nevertheless, the fewer adverse drug reactions associated with these new NSAIDs mean that Cox-II inhibitors may be truly advantageous for a segment of the MCO population. Avoidance of standard NSAIDs in members at risk for their associated adverse reactions could result not only in positive clinical effects, but cost savings as well. Indeed, in that segment, a rearrangement of benefit design or formulary tiers may be structured to encourage the use of the new technology.
The challenge is risk identification.
A large Midwest health plan decided to conduct a risk assessment. It would identify members at risk for an adverse reaction to standard NSAIDs, stratify the magnitude of that risk, and develop an intervention for increasing the number of the members on an appropriate NSAID or Cox-II inhibitor.
Analysts created member-centered data sets from administrative claims to achieve full database integration. Selection criteria included continuous enrollment from Jan. 1, 1996, to Dec. 31, 1997, and complete benefit coverage. All age groups and genders were included, but the greater-than-65 age group received special attention.
The analysis included medical data to determine the presence of targeted diseases such as rheumatoid arthritis, osteoarthritis, peptic ulcer disease, and gastrointestinal hemorrhage (from ICD-9 codes). Additionally, E codes, which identify causes of injuries or poisonings, were used to identify prior or current adverse reactions to standard NSAIDs.
From pharmacy data, the researchers identified the exposure to standard NSAIDs, dosing regimen, number of prescriptions, and multiple NSAIDs prescriptions. These data also indicated exposure to corticosteroids and concomitant use of gastrointestinal protective agents. From the pharmacy data, the researchers established a quality-of-life indicator using the total number of daily medications.
Overall, 496,664 members met the selection criteria. The slight majority of the members (51.4 percent) were female, with 12.9 percent being 65 or older. Not surprisingly, the elderly had a higher prevalence of the targeted disease states. Compared with the under-65 group, the elder group had a nearly fourfold prevalence of rheumatoid arthritis, a sevenfold prevalence of osteoarthritis, a nearly threefold prevalence of peptic ulcer disease, double the rate of known NSAID adverse reactions, and a nearly fourfold rate of gastrointestinal hemorrhage. Exposure to prescription NSAIDs was also higher in the elderly group.
Within that group, 28.7 percent had at least one prescription NSAID during the time frame. For those under age 65, 17.9 percent had at least one prescription NSAID. Based on dose regimen criteria, both groups had similar dose trends. Moderate doses were used in 80 percent of both groups. The vast majority of both patient groups used only one NSAID during the study.
Corticosteroid exposure was similar in both groups and dependent on the presence of an arthritis diagnosis. In those members with a diagnosis of either type of arthritis, 20 percent had a prescription for corticosteroids. Gastrointestinal protective use followed a similar pattern, but a slightly greater number of elderly patients received either a histamine-2 blocker or a proton pump inhibitor.
The greatest difference between the age groups came in the quality-of-life indicator. The elderly are much more likely to be taking multiple prescription medications daily than the under-65 group. Of the over-65 group, 55.8 percent took at least two prescription medications daily; 12.8 percent took more than five daily. In contrast, 18.1 percent of the younger group took more than two daily and only 2.7 percent took more than five prescription medications daily. Overall, approximately 25 percent of all members consumed at least one prescription medication daily during the study period.
With the above findings aggregated and analyzed for risk of an adverse reaction to a standard NSAID, it becomes evident that age played the greater role in increasing risk. Of the younger group, 95.1 percent had low risk, versus 52.1 percent of the elderly. Nearly 10 percent of the elderly population appeared at high risk for an adverse event with standard NSAIDs. When the high-risk group was stratified, the elderly were the "highest of the high." Overall, 4.3 percent of the members were in the high-risk category.
As a result of this research, but before the Cox-II inhibitor launches, the MCO implemented a proactive strategy designed to notify prescribers about patients who were in the high-risk group for adverse reactions to standard NSAIDs. The MCO published the criteria used to identify that group and encouraged prescribers to put patients who met those criteria on a Cox-II inhibitor rather than another NSAID. The health plan published prior authorization and appeals procedures for those patients who were not identified as high-risk but were identified by the prescriber as Cox-II inhibitor candidates.
One-year follow-up data have revealed that 4.06 percent of this MCO's population was exposed to a Cox-II inhibitor. A comparison was drawn between the percentage of persons at high risk for adverse reactions to standard NSAIDs and those who actually received prescriptions to the Cox-II inhibitors due to their having experienced adverse reactions to other NSAIDs.
While the comparison showed 4.3 percent of patients were at risk for adverse reactions and 4.06 percent of patients were actually taking Cox-II inhibitors, the cost difference between NSAIDs and Cox-II inhibitors is great enough to have generated a measurable savings. It was determined that a $1.5 million annual savings was realized by the MCO — a successful demonstration that risk assessment can be successfully applied in a large-scale managed care environment.