TRENDING RISK: Denial for Short Inpatient Stays + WHAT To DO

The last several months have seen the resurgence of audits by CMS and commercial payers. One prime area of interest is the hospitalized patient with a short inpatient stay (short stay). This new level of attention places hospitals at increased risk of payment denials for inpatient services. Now is a good time to do a spot check of how your hospital is managing this category of admissions, hopefully before the auditors come knocking. Let’s take a look at some strategies we employ on the frontend to monitor and manage short stays, and thereby decrease the risk of post discharge audit and denial.

WHO IS AUDITING?

Since CMS implemented the 2 midnight (2MN) rule in October 2013, there have been audits of its use by hospitals in determining the medical necessity (MN) of inpatient stays. Since October 2015, the quality improvement organizations (QIO), KEPRO and LIVANTA, have been major players in these audits.  However, this work was placed on hold in April 2019 after their contract expired. CMS awarded Livanta the sole national QIO claims review contract in early 2020, but this was contested by other players in the industry, leading to a pause in the work that has lasted into 2021. Finally, it was announced in April 2021 that audits by Livanta would resume later in the year, and in September, hospitals began to see additional document requests (ADR) from the QIO. Hospitals were told to expect a sample size consisting of 30 claims within a rolling 3-month period, with ADRs faxed (or mailed if no fax number is provided), to the point of contact listed in the facility's Memorandum Of Agreement on file with Livanta.

A BRIEF REVIEW OF THE "TWO MIDNIGHT RULE" (2MN)

Implemented in October 2013, CMS established a presumption that hospital stays spanning two or more midnights after the patient is formally admitted as inpatient are considered reasonable and necessary for Part A payment absent any evidence of systematic gaming, abuse, or delays in provision of care in an attempt to qualify for the 2MN presumption. Unlike the private payers, the Medicare stays exceeding a benchmark of 2MN are not the focus of most medical reviews for correct status.

The 2MN rule has been revised by CMS multiple times since it was implemented. The most impactful modifications have been the introduction of exceptions to the requirement of a LOS exceeding 2MN. The categories, and proper use of exceptions should be ingrained in the minds of all admitting physicians, and UM staff. Regardless of the exception used, it is critical that documentation reflects the physician’s expectation for an IPS,  and any unforeseen outcomes that qualify for an exception.

•  Procedure is on the Medicare Inpatient Only list

•  New Mechanical Ventilation Initiated During Visit

•  Unexpected Recovery

•  Patient leaving AMA (against medical advice)

•  Death

•  Transfer to another acute care facility

•  Election of Hospice Care

• “Rare and Unusual” exception  

There is one more situation recognized by CMS, described as the  "rare and unusual," exception. Under this modification, inpatient stays of <2MN may be deemed medically necessary, if the patient's clinical condition, and medical management could only be provided in the inpatient setting. I recommend against the routine use of this exception. Outside an CMS-deemed inpatient only procedure (IPO), there are vanishingly few situations where someone is so critically ill that only the inpatient setting is appropriate, yet the predicted LOS is <2MN. In cases with this severity of illness the hospitalization will exceed 2MN, and an earlier discharge suggests the patient has an earlier than expected recovery. CMS expects that stays under 24 hours rarely qualify for an exception to the benchmark of a 2MN stay for an inpatient stay. 

WHAT IS SHORT STAY?

When announcing the impending resumption of audits, CMS modified its long used terminology stating

“Formerly known as the “Two-Midnight Rule Review,” claim reviews for short hospital stays focus on the claims submitted by providers when a patient was admitted to the hospital as an inpatient but discharged less than two days later. Inpatient admissions are generally payable under Part A if the admitting practitioner expects the patient to require a hospital stay that crosses two midnights and the medical record supports that reasonable expectation.”

Under Traditional Medicare Insurance, any hospital episode of care with a LOS less than 2MN following the placement of an inpatient order, is considered a short stay. These episodes are at risk of audit under the reinvigorated QIO audits. With commercial payers, the LOS threshold of a short stay can vary by the diagnosis, payor standards, or hospital specific contract language. Hospitals when screening for short stay, under commercial payers, may decide to be more conservative in their cutoff, such as setting the bar at stays less than 72 hours. Regardless of the definition, there are several situations that lead to the appearance of a short stay, and thereby trigger an audit. More on this later. 

The core characteristic of a short stay is a LOS so short that payers may flag the case for audit to determine if care may have been provided at a lower level, e.g. observation services (OBS). Any LOS which is below the average for patients with the same or similar diagnosis, is a target. In addition, certain diagnoses may be targeted when there is a large discrepancy between the actual and expected LOS. Consider a patient with septic shock, who is discharged (alive) less than three days after admission. In general, the diagnosis of sepsis is a common target of the payors due to a (actual or perceived) high error rate. A short stay with recovery and discharge of a critically ill patient with septic shock would be uncommon, and may suggest over coding. 

OCCURRENCE SPAN CODE 72 (OC72). This code is commonly used, but not required, to indicate that the patient has passed two necessary midnights in the hospital, but less than two as an inpatient. Occurrence Span Code 72 on the face of the claim informs providers and auditors of the total (including observation and inpatient) number of nights of hospital care. There are debates as to whether it is better for hospitals to include or exclude OC72 from the claim. A deeper discussion is outside the scope of this article, but for most hospitals, I recommend its use.

Hospital encounters with an inpatient LOS <2MN should be screened as a short stay, even if there is a preceding period of observation services and total LOS exceeding 2MN, or 48 hours. Make sure to note the specific date and time of care initiation (especially transfers), the total LOS from initiation of care, and the presence of an OC72 if applicable.

REVIEW YOUR SHORT STAY GAME PLAN

Once you’ve established the SS parameters for your facility, you can now perform a gap analysis and identify your vulnerabilities. Areas to assess include:

• ADMISSION ORDERS. What wording is used in the physician statement regarding the LOS?
• MEDICAL NECESSITY STATEMENTS. How are physicians documenting the MN for inpatient admissions?
• MONITORING SHORT STAY ADMISSIONS. What is the internal process for identifying and reviewing short stay encounters? Is your screening process capturing too many false short stays?
• DENIALS and APPEALS. What is your strategy for appealing denials of short stay encounters?

ADMISSION ORDERS and the PHYSICIAN STATEMENT 

If an auditor were to read the medical record, is the physician’s intent for inpatient status immediately evident or easily found? Or, is this information buried in the chart and likely to be overlooked by the auditor? Trusting that the payor will be thorough, is not a good strategy, when experience suggests they are more likely to skim through the record. Is the order for admission clear or ambiguous, and do they support  or hurt the defense of a short stay? 

MEDICAL NECESSITY STATEMENTS

How well does the physician convey the MN of the admission? Examples of physician documentation identified during chart audit include; 

• NONE. The defense of a short stay is weakened when the physician’s reasoning for inpatient status is missing.
• WEAK SUPPORT. “The patient requires inpatient treatment.”
• MODERATE SUPPORT. “His worsening urinary tract infection despite oral antibiotics, requires inpatient management with a LOS exceeding three midnights.”
• STRONG SUPPORT. “He is at high risk of progression to severe sepsis, due to his renal transplant on immunosuppressive medications, and having failed > 72 hours of appropriate outpatient antibiotics.”

While stronger arguments tend to be longer, it is the statement's ability to convey a picture of the medical necessity, rather than the length, which is key.

MONITORING SHORT STAY ADMISSIONS

The utilization management protocol at many hospitals includes flagging any Medicare cases that have a LOS <2MN. Fewer, will also identify commercial payor inpatient status with LOS <48 hours. I recommend a process encompassing all payers.

• Cases should be identified as early as possible, and preferably before the submission of charges to the payer. Earlier review, and correction of chart deficiencies, is likely to result in better outcomes such as decreased denials. Through manual processes, or utilizing artificial intelligence (AI), these cases may even be flagged before, or immediately after discharge. Manual processes tend to be inconsistent and labor intensive. For these reasons, we are likely to see increasing use of AI processes.
• Cases may be screened by senior UM RNs, or automatically forwarded for secondary review by the physician advisor (PA). Involvement of a PA experienced in denial and appeals will allow crafting of a preemptive statement in defense of the SS. 
• Consider screening specific diagnoses at high risk for audit and denial following a SS. This will vary by hospital, but common targets are sepsis, acute respiratory failure, and complex pneumonia.
• Make it a practice to screen all traditional Medicare with a time from the inpatient status order until discharge of <2MN.
• Discuss your hospital's target cutoff for commercial payers.
• Review your escalation policy. Which cases are sent for secondary review by the PA?
• Decide whether you’ll include or exclude the time spent under observation when determining whether an episode of care should be treated as a short stay.

THE APPEARANCE of a SHORT STAY, AKA FALSE STAY 

What do we mean “appearance” of a short stay? There are several circumstances where a cursory review may suggest an encounter is a short stay, but with a more thorough search, it becomes evident that the total LOS exceeded 2MN or 48 hours. The most common mistake leading to a false short stay, is not paying attention to the clock, and the time the patient began receiving care.

A hospital LOS can be calculated based upon the time between the admission order and discharge order. Or, it can be measured as the time between start of care and end of care. Hospital data sets most often report the former, when SS admissions should be judged by the latter.  Be aware of how your facility captures the LOS. To illustrate this further (for the following, assume the payor is traditional medicare):

SCENARIO 1. Care begins in the ED at 2000 on day one, an IPS order is placed at 0200 on day two, and the discharge is 1100 on day three. The LOS from the admission order to discharge was <2MN. However, calculating LOS from the start of care, until the end of care, the total time was >2MN, and meets the 2MN benchmark.

SCENARIO 2. After presenting at 1400, a patient is hospitalized with observation services for evaluation of chest pain. Testing shows an acute pulmonary embolism. The second day, there is instability for discharge with mild hypoxia and continued pain. An IPS order is placed. He discharges on the morning of day three. This scenario qualifies for the use of OC72.  If the hospital reports OC72, the payors' filters should alert them to a false SS. The benefits v. risks of reporting OC72, is a separate discussion, but for most hospitals, we recommend its use.

SCENARIO 3. Transfers. At 2000, a patient begins care at hospital A., is transferred to Hospital B, arriving at 0200 on day two, and an IPS order is placed. He is discharged on day three at 1100.. His LOS at hospital B is <2MN, but the total LOS since initiation of care is >2MN. This appears to be a SS, but actually meets the benchmark of >2MN.

SCENARIO 4. Outpatient procedures. Again, the clock starts with the initiation of care. An outpatient elective total hip arthroplasty is performed. On day two, she is not meeting discharge criteria, with poor performance during physical therapy, and uncontrolled nausea. An order for an inpatient is placed, and on day three she is discharged home. THe LOS since placement of the IPS order is < 2MN, but the total LOS is >2MN. 

PEARL: Many EMRs will display the LOS based on the most recent order. If your physicians like to place several orders, then the care team may lose track of the total LOS. I’ve sat through many multidisciplinary rounds, where the team reports the LOS is 1 day, yet the true LOS is 3+ days. To counter this, I set my EMR screen to display the arrival date, not the admission day. Many EMRs allow this option.

PEARL: CMS expects that stays under 24 hours rarely qualify for an exception to the IPS benchmark of a 2MN stay. To avoid triggering denials, care should be taken to ensure that IPS orders are placed as early as possible, the EMR captures the correct time of admission, and that physician practices such as placing multiple admission orders, are not contributing to the appearance of IPS falling below 24 hours in length.

DENIALS and APPEALS - HOW TO DEFEND the SHORT STAY

As a physician advisor there are several approaches and arguments I employ in defense of the short stay. Medical and surgical admissions may differ in some respects. Assume for the purposes of discussion, that we’re talking primarily about medical admissions, unless otherwise stated. The exact approach will vary depending upon the scenario, and payer. Consider the following:

• EXPECTED, OR PREDICTED LOS. Hopefully, your physician has documented an expected LOS >2MN, or 48h. But what does “expected” mean? The expected LOS can be based on
• DIAGNOSTIC RELATED GROUP (DRG). If the patient’s DRG, has a typical LOS of 3.5 days, it is reasonable for the attending (and reviewer) to predict the inpatient status thresholds as defined above will be met.
  PEARL: If the ALOS is >2MN, and the patient has other significant comorbidities, I may argue "Based on the clinical presentation and comorbidities, it is unreasonable to expect the patient will/would have a faster than expected recovery," or similar words.
• PREVIOUS EPISODES of CARE. A prediction of the current LOS can be based on the patient’s previous admissions. Consider a person hospitalized for the fourth time this year, with COPD exacerbation. All the previous stays resulted in a LOS of 5+ days. It is reasonable for the attending to predict this hospitalization will be similar to the others and therefore place an inpatient order.
  PEARL: My argument in this case will include "Similar admissions for this patient have far exceeded the maximum appropriate duration of observation services," or similar wording.
• COMORBIDITIES. LOS predictions may be based on the presence of other issues that confound the hospital course and contribute to prolonging the medically necessary LOS, Some examples include patients requiring contrast studies, who also have acute kidney injury, or patients requiring anticoagulation for acute pulmonary embolism, but also with worsening anemia in the setting of subacute Gastrointestinal bleeding. One of the reasons we do not adhere solely to clinical guidelines such as InterQual or MCG, is that an experienced clinician can assess the entirety of the medial record, and make predictions regarding the most likely duration of the hospitalization, or at least, be reasonably certain the LOS will meet IPS thresholds.

• THE PATIENT's CONDITION at the TIME the INPATIENT STATUS ORDER is PLACED. Payors often forget the MN of an inpatient hospitalization is judged on the information available to the attending physician at the time the decision was made for inpatient status. Any subsequent information, including the overall clinical course, is not to be included, unless that information further supports the original impression that inpatient status was medically necessary. In other words, the fact that everything turned out o.k. Is not a basis for denying an inpatient status. If you disagree, then consider whether hospitals should be paid only for the abnormal MRIs during workups for possible stroke, or only the abnormal stress tests when ruling out coronary artery disease. When defending an inpatient denial the standard remains; at the time the attending placed the inpatient status order, there was medical necessity for the status, and an expectation of >2MN.

• THE CLINICAL COURSE. Many of the short stay admissions begin with observation services, with later conversion to inpatient status.  By their very nature, this group tends to have a lower overall acuity, and intensity of service, making them a prime target for denials.  With most appeals, we should include clear statements that the patient "failed observation services," or "failed to meet discharge criteria, despite appropriate observation services."
• THE DURATION of CARE. For inpatient status with the appearance of a short stay, I will include specific wording, describing at a 6th grade level of reading, why the total length of care was really in excess of 2MN or 48 hours. For example "this patient presented to an outside hospital on 11/2/2021, with initiation of care at 2300, transferred to Good Community Hospital on 11/3/2021. The inpatient order was placed at 0730.  With the discharge date of 11/4/2021, his total length of stay exceeded 2MN."
• CMS EXCEPTIONS. The medical record reflects a CMS-recognized exception to the 2MN benchmark. It is critical that there is clear and easily discoverable documentation of the CMS exception for any inpatient stay lasting <2MN between the start and end of care.  Early screening at or near the time of discharge, and correction of any documentation deficiencies, will save the work doing downstream appeals.

PEARL: A common mistake is to assume a formal order of leaving Against Medical Advice is required for this short stay exception. In reality, it is a much lower threshold. The chart simply needs to show that it was the patient’s preference to leave, and the physician recommended continuing the hospitalization.

PEARL: There are times when the attending physician ordered inpatient status, the medical necessity review shows it is more appropriate for observation, and the case is escalated to the physician advisor on day 3. If the same case were sent on the first day of hospitalization, the determination would be observation. In these situations I argue " the hospital course and actual length of stay for medically necessary care, exceeds the 2 midnight benchmark, and supports the initial decision by the attending physician for inpatient status."

CONTINUE TO IMPROVE

Learn from your internal audits, and from the short stay encounters denied by payers. In the end, any chart audit should reveal four patient categories: 

• False short stays. Inpatient stays with a total LOS >2MN, including time since care started as an outpatient, or while under observation services. 
• True inpatients with a CMS recognized exception
• True observations, incorrectly hospitalized as inpatient.
• True outpatients, incorrectly hospitalized as inpatient

The ratio of these groups at your facility will provide great insights into how the hospital is performing. Armed with this knowledge, system improvements can be developed. Since each category has unique solutions which are often hospital and market specific, I recommend engaging your Physician Advisors to help craft your next steps

SUMMARY

With the QIO audits up and running as of September, it is more important than ever to take stock of how your hospital is tracking and addressing the short stay admission. The work done on the front end by the UM team and the attending physicians, will set you up for success with downstream denials. Review the admission orders for clarity.  Make sure the process screens for, and correct any deficiencies in documentation of CMS-recognized exceptions to the 2MN benchmark, such as the patient who recovers faster than expected. Define your criteria for which cases will be flagged as a short stay and when to escalate cases for physician advisor review. Be able to identify cases giving the false appearance of a short stay, and consider ways to mitigate the risk of their being denied. And, if you are unsure where to turn next, consider reaching out to your peers through AHDAM.