Using data management for CM populations

Define the population and establish a baseline

Many case management departments are taking a new approach in terms of how they utilize case managers for outcome evaluation and process metrics. According to Vicky Mahn-DiNicola, RN, MS, vice president of ACS MIDAS+ in Tucson, AZ, many case managers now are merging various roles into what are called "outcome managers" or "outcome specialists."

Mahn-DiNicola says many case managers actually are having their titles changed to "outcome specialist" and are being organized around key clinical populations. In that capacity, they frequently are responsible for the information about that population and act as the clinical specialist. She says these case managers not only monitor their clinical populations but also examine the data to more accurately determine what is taking place within that clinical population. In addition, they are more involved in finding "leverage points" in the data, she says.

As an example, Mahn-DiNicola points to one case manager with a behavioral health specialty who reviews financial outcomes, clinical practice patterns, and performance improvement and utilization patterns, and performs physician profiling, as these functions relate to a specific population. Because the case manager knows the physicians and best practice standards surrounding the specialty, she is able to recognize opportunities to improve clinical outcomes or financial performance.

Mahn-DiNicola cites four trends that case management departments should be aware of in this area. First, she says hospital-based case managers increasingly are being organized around a group of patients who have similar clinical conditions, such as orthopedic, cardiac, or pediatric.

Even in the community environment, some case management programs specialize in oncology or another specific disease state. For example, she notes one case manager who is focused on multiple sclerosis patients and performs community clinics, workshops, and swimming therapy for that patient population. "Even though her focus is community-based, she is a clinical specialist in that population," Mahn-DiNicola explains.

Second, Mahn-DiNicola says, case managers more frequently are being asked to demonstrate the value of their roles and their interventions to their sponsoring organizations.

Third, case managers often are responsible for outcomes and process performance data within their organizations. "By virtue of that fact, case managers are becoming natural stewards to the business of metric design, performance improvement, and data mining," she says.

Finally, Mahn-DiNicola says case managers are being forced to become more "data-savvy" and more comfortable with data mining in order to identify opportunities for improvement and facilitate changes in clinical practice and care management processes. "It is a good fit," she says, "but to do that and to do it well, case managers must sharpen their skills."

For example, a length of stay of only a day or two for a group of pneumonia patients might look good on paper, but it actually may reflect the fact that these patients were not sick enough to be in this hospital in the first place. Perhaps they could have been managed in an alternative environment at a reduced cost, she argues.

To remedy this situation, Mahn-DiNicola says it sometimes is necessary for case managers to redirect their energy to helping the emergency department establish some form of risk protocol and then to manage the patients identified as low-risk patients using home health. If a large percentage of these patients are coming from nursing homes, it might make sense to have an advanced-practice nurse start making rounds in the nursing homes.

According to Mahn-DiNicola, different strategies are required for different situations. "You are not going to figure out what to do to better manage a population of patients until you have someone with clinical eyes’ review the data and the patterns and the trends," she says. "That is where we are seeing case managers more frequently utilized in terms of outcome management."

Mahn-DiNicola says the first step in this process is for auditors to be very clear about the population they are examining. "It is not enough to say, I follow stroke patients,’" she explains. Rather, they have to communicate why that population was selected for case management in the first place. She says it could be that it was high volume or high risk or possibly experimental in nature and therefore controversial.

Alternatively, Mahn-DiNicola says it might be an area subject to a regulatory focus by the Joint Commission on Accreditation of Healthcare Organizations or the Centers for Medicare & Medicaid Services (CMS).

Mahn-DiNicola says case managers also must be well-versed in best practice literature relating to the care of a specific population. She says that means, among other things, keeping abreast of national guidelines, standards of care, and medical literature. Then she says they must extract from that literature all of the best practice strategies for managing a patient population. For example, if the population in question is carotid endarterectomy, they must know when it is appropriate to do an angiography vs. a duplex ultrasound.

Mahn-DiNicola says case managers also must implement best practice strategies in their organization, reduce variation, and implement standards. It might be a matter of the timeliness of an intervention, she says. For example, the goal might be to get acute myocardial infarction patients in the emergency room and treated with a thrombolytic or an angioplasty in less than two hours. In that case, important measures will include the timeliness of antibiotics, timeliness of certain assessments, timeliness of certain drugs, and code arrest situations.

"Certainly, all the surgical procedures and complications studies that are being undertaken now by CMS are looking at the antibiotic selection and timeliness of when it stopped," she reports.

Mahn-DiNicola says case managers are natural stewards for these kinds of strategies, as well as for patient education. "There are key ingredients that we have to educate patients and families about," she explains. "Case managers are in a unique position to do that."

Another key factor in defining the population is to understand the data available in the hospital system. That means understanding what data elements are included in administrative claims data for UB-92 data and whether there is a pharmaceutical database or payer database to draw on. "Understanding where these data live and where you can tap into it is important," she says.

According to Mahn-DiNicola, it is very important in defining the population to give ample consideration to whether to use a DRG or an ICD-9 code. Many times when you pull patients by DRG, you get very different patients from when you pull them by ICD-9, she says. "It helps to look at the patients, pull them both ways, and see what you have," she explains. "You might have two very different clinical populations."

Mahn-DiNicola says she generally recommends pulling by ICD-9 code. Then case managers have to think about whether they want to select by primary or by secondary ICD-9 codes. "That makes a difference," she adds. "They need to educate themselves about coding to some degree by spending some time with the coders in their medical records departments." That means sitting down with them for a day and learning how these experts actually read medical records and come up with codes. "It is ultimately part of the billing process," she notes. "Case managers simply need to understand how they can capitalize on that and where the problems are."

Identification of population is crucial

Finally, Mahn-DiNicola says it is necessary to identify patients in the database. She says this is critical in order to follow a population of patients over time because case managers must identify the population the same way consistently to trend it over time or to compare it with other populations.

According to Mahn-DiNicola, if case managers do not pull exactly the same population with the same exact inclusion and exclusion criteria, they will wind up with conflicting data. She says it is very important for case managers to clearly identify the population based on an electronic data source such as an ICD-9 code, a CPT-4 code, some type of DRG, or another identifier and then consider the different ways to segment that population into electronic data that can be defined in a homogenous clinical population.

In some cases, Mahn-DiNicola says, case managers might want to look at Medicare patients independently, sort by payer, or use a particular age range. She says it is a matter of becoming familiar with all the data elements in the database where those patients reside.

Because case managers typically are not trained in informatics, they must put a lot of effort into learning what information exists. "That means they have to knock on a few doors and start asking questions and requesting access to raw data," she says.