In order to begin an effective population health management initiative, providers must first identify and stratify patients by the costs they incur, says a new white paper by the Health Care Transformation Task Force. By proactively flagging patients in the highest brackets of spending and service utilization, providers can target their population health and chronic disease management programs more effectively.
“Identifying the high cost population is important for practical reasons. Health care costs are highly concentrated in a very small patient subpopulation,” says the paper, compiled by the High Cost Patient Work Group.
“For example, the top 5 percent of patients, ranked by individual health care dollars spent, are responsible for almost half of the nation’s total personal health care dollars spent. Finding and managing care for this group of patients can be an efficient and effective way to increase quality and reduce total costs for the entire population.”
The Health Care Transformation Task Force, a coalition of providers and industry groups attempting to rapidly accelerate the adoption of value-based reimbursement and accountable care, suggests that healthcare organizations divide their patients into two major subgroups based on their spending, disease patterns, and likely outcomes.
Patients with advanced or terminal illnesses
While data from the National Institute for Health Care (NIHC) indicates that the top five percent of patients incur more than 49 percent of spending each year, not all patients are chronic users of services. Many of these patients are in the final stages of expensive illnesses and will not reappear in the spending bracket the following year.
Nearly a third of patients ranked as top decile spenders died within two years, Medicare data shows. Population health management strategies for these patients typically center on end-of-life care planning, including hospice use and the collection of advanced directive information.
Increased hospice use can save significant revenue for healthcare organizations, the white paper says. “Aetna reports average savings of nearly $13,000 per enrollee associated with an 82 percent hospice election rate, 82 percent reduction in acute hospital days and 86 percent reduction in ICU days,” the report states. “Similarly, Sutter Health’s Advanced Illness Management (AIM) program found savings of more than $4000 per enrollee per month associated with significant reductions in hospital utilization.”
“Focusing on this patient population has the potential to greatly reduce costs, while also providing care that is appropriate and valued by individuals with advanced illness and their families.”
While clinical analytics based on claims data can help to identify patients who fall into this category, current analytics technology is not always sensitive or comprehensive enough to rely upon fully, the Task Force says. Claims data does not always contain the necessary elements to distinguish between persistent high-users and end-of-life users, and EHR data is not consistently collected and applied to big data analytics projects in order to make large-scale population health analytics effective just yet.
Persistent super-users and high spenders
The second group of patients, those with complex chronic disease management needs that are projected to last for a longer term of care, have been the subject of most population health management projects thus far. Hypertension, diabetes, and high cholesterol are extraordinarily common diagnoses for patients who comprise the top five percent of spenders: more than 65 percent of elderly patients and 35 percent of younger patients who rank in the top tier of spending have been diagnosed with hypertension, the report states.
“In these patients it is useful to distinguish common diagnoses from diagnoses that drive spending,” the paper points out. “For example, hypertension and hyperlipidemia are widely prevalent conditions, but do not necessarily directly result in high costs, as renal failure, congestive heart failure (CHF), and COPD do. On the other hand, complex patients with low patient activation—as measured by the Patient Activation Measure (PAM)—are at particular risk due to the inability to perform adequate self-care to manage their condition.”
Big data analytics can be an important tool for making these distinctions. Prospective risk analytics that integrate claims and EHR data into risk scores can help providers target their services appropriately. However, the time delays involved in using claims data as a primary source for actionable analytics can leave some patients at risk of falling through gaps in care coordination.
As providers work to move from descriptive to predictive analytics, the increased use of real-time EHR and patient-generated health data will help ensure that chronic disease management for high-spending patients can be enacted more immediately.
Nonetheless, claims data is an important jumping off point for providers who hope to improve their population health management techniques by starting with spending data. This information is critical for achieving the cost savings and quality outcomes that support emerging value-based reimbursement strategies.
“Changing overall clinical, utilization, and cost outcomes for the entire population may best be accomplished by intervening with the small number of patients with highest need and highest cost,” the paper concludes. “Care management programs can be developed and brought to scale fairly easily to manage this cohort. Strategies to improve care for patients with advanced illness or persistently high costs are challenging to implement, but several viable innovative models exist.”
Future efforts by the Health Care Transformation Task Force and the High Cost Patient Work Group will detail successful use cases for cost-based population health management programs that achieve sustainability and widespread acceptance in the provider community.
View the original content and more from this author here: http://ift.tt/1en2iV7
from health IT caucus http://ift.tt/1JCB5c0
via IFTTT
No comments:
Post a Comment