Reducing The Hospital Length Of Stay

One of the best ways to improve a hospital’s financial margin is to reduce the average patient length of stay. Since hospitals are paid by the “DRG” (diagnosis related group), the hospital is going to get paid the same amount for a patient with say, pneumonia, if that patient spends 4 days in the hospital or spends 8 days in the hospital. Therefore, the quicker the hospital can get a patient discharged from a hospital bed, the sooner the hospital can put a new patient in that bed. If the hospital can get all of their pneumonia patients discharged in 4 days, as opposed to 8 days, then the hospital can admit twice as many patients into that bed over the course of a year and consequently can make almost twice as much money.

Hospitals report their length of stay (in days) as a standard administrative metric. But I would argue that the length of stay in days is a meaningless number. The more important number is the length of stay index.

The case mix index is a way of adjusting for how sick a patient is, as defined by co-morbid medical conditions. For example, an otherwise healthy patient with pneumonia will have a lower case mix index than a patient with pneumonia who also has diabetes, COPD, heart failure, and cancer. Medicare and insurance companies pay hospitals more if a patient with a given diagnosis has a lot of these co-morbidities. Thus, hospitals are actually paid by the Medicare Severity-Diagnosis Related Groups (MS-DRGs) rather than by the plain DRG.

There are more statistics available for length of stay than for length of stay index. For example, there are geographic differences in length of stay: hospitals in the Northeastern U.S. have an average length of stay fo 4.9 days, the South 4.5 days, the Midwest 4.3 days, and the West 4.2 days (2012 AHRQ data). There are also state-specific differences in length of stay with South Dakota having the longest length of stay (9.2 days) and Utah having the lowest length of stay (4.3 days) in 2014. There are economic differences with hospitals in low income areas having an average length of stay of 4.6 days and hospitals in high income areas the average is 4.4 days. There are also differences by payers: Medicare = 5.2 days, Medicaid = 4.3 days, commercial insurance = 3.8 days, and uninsured = 4.0 days (although this may be a reflection of different ages of patients served by different payers since older [Medicare] patients have a longer length of stay than younger patients).

Here are some practical steps that the hospital can take to reduce the length of stay index:

1. Increase the case mix index. This does not mean that the hospital has to seek out sicker patients to admit. It does mean that the hospital has to have a robust process for ensuring that all of the various co-morbidities that each patient has are captured at the time of admission to the hospital. This generally requires a physician document that these various conditions were present on admission in their H&Ps and consults. There are two steps in doing this. First, educate all of the physicians (and especially the hospitalists) in the various co-morbid conditions that they need to be on the lookout for when admitting a patient. Some of these conditions may seem minor to the physicians (for example, hyponatremia and hypernatremia) but can significantly increase the case mix index. Second, documentation specialists should do real-time audits of the patients’ charts in the first days of their hospital stay to identify any co-morbidities that the physicians may have missed and request that the physicians addend their notes accordingly.
2. Adequately staff the hospital’s case management department. If a case manager or social worker is so overworked that he/she cannot get to a patient on a given day, then you have likely added an extra day to the length of stay for that patient. Those “opportunity days” add up and their cost can quickly exceed the expense of hiring an additional case manager or social worker.
3. Develop arrangements with high-performing skilled nursing facilities. Patients being discharged to SNFs often have the highest length of stays, not because they take longer to get well in the hospital but because there is a delay in getting a bed in the SNF. The best SNFs are those that can promptly obtain prior authorization from the insurance company and accommodate the patient as soon as the patient is medically stable for transfer. Ultimately, the hospital may need to even lease beds from SNFs.
4. Schedule the operating room strategically. Ideally, patients who will require an inpatient stay post-operatively should be scheduled as early in the morning as possible. This allows them to get a head start on their recuperation. For example, the patient with a hip replacement surgery in the morning can get their first physical therapy session the same day as their surgery and potentially be ready for discharge a day earlier than if they have surgery in the afternoon.
5. Whenever possible, be a 7-day a week hospital. For procedures and tests that can delay discharge, try to offer these tests every day of the week. Common examples include cardiac stress tests, echocardiograms, endoscopy, bronchoscopy, and venous duplex ultrasounds. It may be impractical to run a full-day schedule for these tests on Saturdays and Sundays but at least be able to do half-day schedules on either a Saturday or Sunday shorten hospital stays.
6. Consultants should be co-managers. In academic medical centers, consultation services typically consist of a resident or fellow who will initially see a patient in the morning and then round with the attending consultant physician later in the afternoon with the consult getting finalized at the end of the day. Furthermore, in order to preserve the residents’ autonomy and ownership of patients on the admitting teaching services, the consulting physicians generally leave recommendations in the chart but do not order tests or medication changes. Frequently, this can result in delays in care when, for example, the consulting attending leaves recommendations for blood tests or x-rays in the chart at the end of the day and then the resident on the admitting service does not see that recommendation until the following day with the result that the test gets delayed. Ideally, inpatient medicine should be a team effort and the consulting physicians should be empowered to order the tests that they want and medications that they recommend. Furthermore, consulting physicians should continue to follow patients after the initial consultation to assess the patients’ response to treatment and to help interpret the tests that they have recommended/ordered. Although this culture change can result in the admitting physicians feeling like they have lost control over their patients, the benefit to the patients’ care and to the patients’ length of stay is worth it. I believe that this strategy is one of the most over-looked strategies in length of stay reduction programs.
7. Consult liberally. Many hospitalists are loath to consult specialists, a mindset borne out of the tenet that “a consult is a sign of weakness”. Although it is true that consultation will add to overall healthcare costs because of the additional professional service that will be billed to the patients’ insurance company, consultation may also reduce overall healthcare costs by expediting inpatient throughput, reducing unnecessary medication use, and facilitating outpatient follow-up arrangements.  Many experts feel that United States healthcare should reduce the use of subspecialists; I take a contrarian view that when it comes to inpatients, more liberal consultation has a net effect of reducing costs while improving quality of care.
8. Discharge planning starts on the day of admission. If the case management staff do not do an initial evaluation of the patient until the patient is close to discharge, discharge delays ensue. Sometimes, you just know that a patient is going to need to go to a SNF or an LTACH or inpatient rehab from the minute they arrive on the hospital floor. By initiating discussions with the patient and their families early, discharge choices can be made early in the hospital stay and the staff can initiate prior-authorization process with insurance companies as soon as possible, thus avoiding delays in discharge.
9. Institute multidisciplinary rounds. These are typically done mid to late morning, after the hospitalist has made “work rounds”. The hospitalist, nurse, case manager, and often pharmacist then round on each patient together so that all parties are on the same page regarding tests that need to be done, discharge planning barriers, etc. Multidisciplinary rounds are the most effect way to reliably communicate patient care among all of the care providers and optimize patient throughput in the hospital.
10. Ensure that guardians are appointed in a timely fashion. Many states have long waits for guardianship determination. For patients who are incompetent, the wait for guardianship can have a huge effect on the length of stay. Hospitals can work with state authorities to develop processes to expedite guardianship in order to move patients through the system in a more timely fashion.
11. Eliminate emergency department boarding. Patients who wait for hours in the emergency department for a bed to become available in a hospital that is at full capacity do not get the same care in the ER as they do on a nursing unit. Tests do not get done, medications often are not started, and consultants often do not see the patient. The faster the hospital can get a patient out of the ED and into a nursing unit, the sooner definitive evaluation and treatment of the patient’s medical problems can occur.
12. Right-size your hospitalist staffing ratios. A hospitalist who is managing 25 inpatients is not going to be as efficient as one who is managing 15 inpatients and consequently, the overburdened hospitalist is going to have a longer length of stay. When your hospitalists are spending most of their work days getting admission H&Ps done and attending to unstable patients, the lowest priority will be expediting discharges. That hospitalist needs to have sufficient extra time in the day to meet with family members, help the case managers with discharge plans, and critically evaluate new test results. When the hospitalist has sufficient time to round on their patients twice a day, as opposed to once a day, care is expedited.
13. Right-size your nurse staffing ratios. Similar to hospitalists, if a nurse is caring for too many patients, he/she will not have the time to help expedite the patients’ throughput and length of stay can suffer.
14. Measure it. In order to know where the length of stay problems are, the hospital has to be able to identify which services and nursing units are strong performers and which have excessive length of stays. Also, there needs to be a mechanism for real-time assessment of the effect of any interventions you make on the length of stay. Ideally, the length of stay index should be measured and reported monthly by service and by nursing unit.
15. Know which tests and procedures can be done as an outpatient. Not everything needs to be done when a patient is in the hospital. For example, a patient who has a lung mass incidentally identified on an abdominal CT done to evaluate their cholecystitis can just as easily have their PET scan and lung biopsy done as an outpatient. In order to do this, there must be reliable outpatient follow-up available. Transition clinics are a great way of ensuring that these tests get done timely after discharge.
16. Use palliative medicine strategically. Palliative medicine can reduce the ICU length of stay but has considerably less effect on non-ICU length of stay. However, palliative medicine is costly, requiring heavy subsidy by the hospital since palliative medicine rarely, if ever, pays for itself by regular billing for physician services. By prioritizing palliative resources on the ICU, the most cost-efficient use of palliative medicine can be achieved.
17. Draw morning labs early. Patients do not like to be woken up at 2:00 AM to get their morning labs drawn. However, by getting those blood samples sent to the lab early, results can be available at the time of the physicians’ morning rounds. This permits clinical decision making early in the day.
18. Save some time slots for inpatients on the procedure schedules. Most hospital procedure areas do testing for both inpatients and outpatients. There is a tendency to schedule on a first-come, first-served basis and as a consequence, the earliest time slots get assigned to outpatients who schedule their procedures days or weeks in advance. This often leaves the only times available for inpatients at the end of the day as “add-on” cases. This can add an extra day to the patient’s hospital stay if procedure results are not available to the hospitalist or consultants when they are doing their regular daytime rounds. By reserving some time slots for inpatients (particularly on Mondays and days after holidays), throughput can be expedited for tests such as venous duplex ultrasounds, cardiac stress tests, echocardiograms, cardiac catheterizations, and endoscopies.
19. Never admit a patient at 11:59 PM. Most hospitals calculate the length of stay by patients in an inpatient bed at midnight (the so-called “midnight census”). A patient admitted 10 minutes before midnight will have a length of stay that is 1-day longer than a patient who is admitted at 10 minutes after midnight, even if they are discharged at exactly the same time and day. The reality is that the true cost of caring for that patient will be the same, but your LOS index numbers will look worse.
20. Be aware of unintended consequences of admission practices. Patients tend to come to the ER in the late afternoon and evening and consequently, these times are when most patients get admitted to the hospital. If there is a service that only admits during the morning and early afternoon (for example, a resident teaching service), then those services can get a jump on patient testing that can be done on the same day of admission. On the other hand, if there is a service that admits disproportionately more patients in the evening when routine testing is not available, that service will have a 1-day delay in getting initial testing done and as a consequence will have a longer length of stay than the service that admits earlier in the day. Knowing the typical time of day of admissions can help you avoid penalizing services that primarily admit in the evenings when their length of stay looks higher than other services.
21. Do not put over-utilize observation status. Many times, it is not entirely clear if a patient will be in the hospital for less than 2 midnights (the CMS definition of observation status). By placing a patient who could legitimately be an inpatient into observation status, the length of stay index will increase for the inpatients since these patients generally have a lower length of stay.

The challenge of reducing length of stay is to do it without reducing quality. In this regard, there is a limit to how low the hospital length of stay index can safely go. In my opinion, for most hospitals, a length of stay index of 0.80 – 0.95 is optimal. Below that, quality of care can suffer and above that, costs are excessive.

February 11, 2019