The Myth of Length of Stay
The introduction of Medicare' prospective payment system in 1983 prompted some deep strategic thinking by hospital executives about future revenue and how to best preserve margins. The thinking went something like this: "If Medicare is giving the hospital a flat fee to cover its costs based on an expected length of stay, then if the patient stays in the hospital for less than the expected length of stay, the hospital will make a profit."
It sounds reasonable based on the reimbursement methodology, but it was flawed logic simply because it was assumed that if the patient stayed in the hospital fewer days than expected, then fewer resources would be consumed. With fewer resources consumed, the difference between Medicare's DRG reimbursement and the hospital's actual cost would result in an attractive margin that every hospital needs to financially thrive. The reality turned out to be quite different. Resource consumption did not decline and physicians continued to practice exactly the way they did pre-DRG—but they did it in fewer days. It is this legacy assumption that still pervades much of the thinking about hospital costs, margins, and length of stay, and hospitals are still feeling the pinch.
A recipe for bankruptcy?
Since 1983, approximately 2,500 hospitals have gone bankrupt, closed, or have been acquired. Though decreasing margins continue to grab headlines in healthcare magazines, hospital executives remain slavishly devoted to the LOS measure. It appears on their desk every morning, every month, and every year according to patient units, physicians, service lines, and any other drill-down meant to communicate the LOS impact. And when the hospital's LOS is higher than that of the state, regional, or national indicators—or if it's not in line with budgeted expectations-the occupants of executive suites across the country issue directives to beef up the hospital's control and command tactics to 'get the patients out quicker.'
That's not to say that LOS is not a valuable metric. Indeed, it is important to gauge hospital efficiency, to maximize the volume of patients in profitable service lines, to attract managed care contracts, to maintain competitive edge and to stay aligned with regional and national benchmarks. It is also a very easy metric to obtain. Though hospitals are replete with data, it is often difficult to access information that would prove helpful to hospital leaders in their quest to improve efficiency, reduce costs, and decrease length of stay.
There are many internal operational situations that every utilization review coordinator or, more recently, case managers routinely encounter in their efforts to advance a patient's treatment plan and effect a timely transition to a lower level of care or discharge. Singularly or in combination, these situations add up to a complexity of variables that needlessly prolong LOS. Situations ranging from inaccurate registration information to delays in delivery of care, consulting specialist responsiveness, ancillary services availability, timely diagnostic testing and reporting, scheduling limitations, and communication mishaps are among the numerous controllable situations that are often cited as cause for excessive length of stay.
Since 1983, length of stay has been the organizational mantra, and since 1983 doctors have been begged, harassed, coaxed, and sweet-talked into discharging their patients as soon as they no longer require an acute level of care. As part of this effort, physicians are asked to consider some other acceptable alternative that can be provided to the patient at a lower level of care. In addition, utilization review coordinators or case managers have pleaded, encouraged, and negotiated with department heads to schedule the diagnostics the doctors have ordered and expedite the results so the patient can be discharged promptly. These ‘discussions' have been going on since 1983 and nothing has changed. Physician practice decisions and delivery of care processes still drive the patient's navigation through the hospital system with the end result of a length of stay that keeps the chief financial officer awake every night. This scenario is a classic example of that overused example about doing the same thing over and over and expecting different results.
Physician practice behaviors
In the past, and according to Medicare's Conditions of Participation, it was left to the hospital's utilization review committee to monitor individual cases with excessive LOS. That directive, while still on the books, is beyond the reality of most hospitals and is rarely practiced. Instead, hospitals generally use LOS trends to identify physicians whose practice decisions routinely put patients over the accepted LOS threshold or whose use of resources far exceed those used by their peers caring for similar patients. But even with objective outlier evidence, offending physicians face little or no consequences. Hospital boards of directors and executive leadership are either unaware of the problem or incapable of implementing solutions. Research from Dartmouth University and the RAND Corporation have consistently confirmed that physicians aren't prescribing more resources because their patients are sicker, but rather because of practice patterns, resource availability, and a perverse reimbursement system that pays physicians for doing more-not doing better.
Back in the 1980s, hospital information systems and technical support staff were ill-prepared for the kind of monitoring it takes to examine physician practices beyond length of stay. Today however, every practice decision made by the physician—how many tests are ordered, how many consultants are brought in, how many days the patient remains warehoused before elective surgery is done—are known. The data exist at the hospital, state, and federal level, and are often used by a variety of public and proprietary oversight agencies and insurance companies to compare a physician's practice against those of his peers caring for similar patients. While the use of the claims data is an inexact science and subject to the vagaries of hospital-specific financial policies, they can be used as a consistent measure of resource consumption. Even if the data are unreliable—and often they are, given the inconsistencies in the hospital's CDM or revenue codes—they will be consistently unreliable. Statistical consistency rather than reliability makes the process a suitable technique to objectively confirm physician practices. This data provides practice information about medical interventions that are not related to the patient's reason for admission; wasteful discretionary treatments; disregard of evidence-based medical protocols; excessive LOS; and compliance with quality measures. Armed with this information, especially when presented in a comparative format, physicians tend to self-modify practice patterns to stay under the bell curve and avoid scrutiny as long as patient quality is never compromised.
Hospitals are complex organizations and suffer from the silo-centric nature of their structure. Administrators are put to the test daily to improve delivery-of-care processes to meet staggering demands and overcome the challenge of communication between multiple caregivers. Tracking the timeliness of scheduling a treatment or service, providing that treatment or service, and reporting the results of that treatment or service are tests of patience. The number of steps in every delivery-of-care process, the sheer number of people who have a hand in the care process, and the process variables that have to be considered are overwhelming, especially in large academic medical centers. Hospitals have tried. From Deming to Six Sigma, administrators have introduced methodologies to help hospital departments streamline processes to make them more patient-centric. But the structural and operational complexities of the hospital defy every effort, and obstacles continue to emanate from every service delivery department. Just think of the number of people and tasks involved in a simple order for a blood test. The complexity and cost of delivering a treatment or service create barriers.