Level-of-care guidelines--prompting clinicians in making admisision and referral decisions up and down the healthcare continuum--are not new. The American Society of Addiction Medicine has been publishing them for years. What's more, a private firm--McKesson Corporation, a large pharmaceutical management and information technology company--has issued practice guidelines under the InterQual brand for a quarter century. It has only been in the last two years, though, that McKesson has made such algorithms available for behavioral healthcare. It introduced in 2000 criteria marketed under the InterQual brand for guiding adult mental health and substance abuse management decisions. Since then, McKesson Health Solutions has added instruments for managing adolescent and pediatric care. Geriatric and residential adolescent offerings are in development, with the former set to be released this fall. Recently Lollie Dubiel, director of behavioral health for the company's medical management group, discussed with Behavioral Health Management Editorial Director Richard L. Peck the concept of automated decision support and how it works.
Peck: Would you describe the InterQual Behavioral Health Criteria tool?
Dubiel: It is a decision-support tool using evidence-based, level-of-care criteria derived from the literature, nationwide clinical expertise and feedback from clients.
Peck: What does 'evidence-based' mean in this context?
Dubiel: Every year our staff submits to a national panel of experts-psychiatrists, psychiatric nurses, psychologists, pediatricians, teachers and social workers--proposed criteria revisions based on a literature review. Based on the panel's feedback and recommendations, the basic decision-support algorithms for the particular behavioral health criteria set are revised.
Peck: How do these algorithms work?
Dubiel: They are designed to guide the clinician in decision making from the initial assessment, to referral to appropriate level of care, to review of the appropriateness of continuing at that level, to referring the client up or down the continuum, to discharge. They do this with a series of linked questions and recommendations branching from the answers. For example, upon initial contact with a patient complaining of mental health symptoms, the decision tree raises the question of whether this patient is at immediate or potential safety risk. Based on the user's input, the algorithm might identify the patient as being at immediate risk and recommend hospitalization. If the algorithm identifies the patient as being potentially at risk, a series of prompts guides the clinician toward appropriate placement. In a case in which a patient presents with chemical dependency, an initial question would be whether the user is at severe, moderate or no risk of withdrawal. Again, the algorithm would recommend the most appropriate level of care.
Peck: For which caregivers and in what format are these programs available?
Dubiel: The hands-on users of these guidelines are typically care managers in a managed care organization (MCO). However, hospitals have used them for team clinical review meetings or in their clinical assessment centers. The criteria are available in a browser-based version under the name CareEnhance Review Manager, which is HIPAA-enabled to protect privacy, or in book form.
Peck: What kind of outcomes can you cite from the use of these instruments?
Dubiel: One example is of an MCO with 2.2 million covered lives that began using the Adult Psychiatric guidelines when they were introduced in 2000. The MCO reported about a 5% reduction in overall utilization in the first six months.
Peck: What kind of return on investment (ROI) might the user expect from this approach to level-of-care decision making?
Dubiel: In assessing ROI, there are three issues that the typical MCO looks at: First, there is usually significant variation among its clinicians with regard to clinical decision making. Many have different patient care philosophies and worldviews. Particularly on the mental health side, various caregiving professionals are involved with decision making-physicians, nurses, social workers, psychologists, each group with its own particular take on the situation. Use of decision-support criteria can help standardize the decision making while still providing a patient-specific approach based on the individual's presentation.
Second, if variations continue between clinicians' decision making and the established criteria, this might show that network expansion or other quality initiatives are in order. Or, it could demonstrate the cost savings that could be achieved by using the criteria appropriately.
Third, many plans experience a savings of clinical reviewer resources because of the increased efficiency of the review process itself.
In general, level-of-care criteria offer a way to break away from lockstep reauthorization of fixed amounts of days of care, such as renewals of seven-day inpatient stays. Clinicians now have guidelines for deciding whether the more appropriate referral might be partial hospitalization, or an intensive outpatient or outpatient setting, according to evidence-based clinical criteria.
Peck: Are there specific advantages to an automated format?
Dubiel: Having the criteria in an interactive, browser-based application enables organizations to gather data and produce useful management reports to identify any areas where they can improve care management processes. Well-designed software allows for flexible use of data and daily, monthly, quarterly and yearend reporting. Another major advantage of automation is its impact on workflow. Our program, for example, prompts staff on days when patient follow-up should occur and finds the necessary clinical review data for that patient to expedite the evaluation. Also, well-designed software should have a security function built in to help with HIPAA compliance.
For further information, phone (800) 582-1738, ext. 1506, or e-mail bill.barton@mckesson.com.
