The probe-and-educate phase is in full swing, with hospitals receiving results of their probe audits and Medicare Administrative Contractors (MACs) conducting extensive education, performing both one-on-one sessions with hospitals and conducting general education via webinars and teleconferences.

The one-on-one sessions vary among MACs; in some instances the MAC representative presents general education about the two-midnight rule in a lecture format, some MACs present brief education sessions and encourage open dialogue about general concepts of the rule without allowing any specific case reviews from the hospital’s probe sample, and some review the actual cases from the hospitals that were audited and provide direct feedback (and will even reverse an adverse determination during the call).

The webinars also vary in content; all present a general overview of the rule. Most provide case examples of approved and denied cases, and fortunately, all have allowed an open question-and-answer period, either by “opening the phone lines” or allowing written questions to be put in the chat box. And so far the reviews of these sessions have been remarkably varied. As discussed on some of the online discussion groups, one of the MACs insists that the midnight clock starts only when a licensed provider first sees the patient, contrary to Centers for Medicare & Medicare (CMS) guidance indicating that the clock begins when symptom-directed care beyond basic triage begins. The issue of one-day inpatient stays (without a preceding outpatient night) also has produced varying guidance, with one MAC stating that it would be looking for documentation from the physician stating that the patient made an unexpectedly fast recovery. Another MAC indicated that such documentation is not necessary, because if the decision to admit was based on a reasonable expectation of a two-midnight stay, then the fact that the patient is able to be discharged after only one day is de facto evidence that the patient’s recovery was unexpectedly fast.

One of the MACs recently presented an educational webinar in which the medical review staff did an excellent job presenting the rule and providing examples, especially considering the technical difficulties they faced, but the question-and-answer period created confusion when their medical director joined in. Her answers seemed disjointed, and in some cases contrary to what was thought to be clear CMS guidance. In fact, the medical director herself realized that many of her answers were confusing and contradictory and personally called me at home the next evening to apologize for the confusion and provide some clarity.

During the presentation, it was stated that there are instances in which a patient requiring medically necessary care in the hospital could be kept in the facility as an outpatient with observation past a second midnight. A questioner asked for an example of such a patient, yet the presenter actually could not come up with an example. The medical director gave the example of a patient who appropriately was placed in observation for mild exacerbation of heart failure. The patient was treated beyond the first midnight and was stabilized but then developed dysarthria that resolved, but the physician felt that this represented a transient ischemic attack (TIA). The medical director stated that if the doctor felt that this TIA required evaluation in the hospital that would require another midnight in the hospital, the patient should be maintained on observation for the second midnight since it was a new diagnosis.

This interpretation is clearly incorrect; the fact that a second diagnosis has developed which warrants remaining hospitalized for a second midnight is irrelevant. Needing a second midnight in the hospital always warrants admitting the patient as an inpatient, no matter the diagnosis. Of course, if the second midnight is for convenience and is not medically necessary, the patient should not be admitted as an inpatient. The rules are clear that a patient should not remain in outpatient status past a second midnight if his or her care is medically necessary and only to be provided in the hospital (and there is actually no instance in which a patient should remain an outpatient past that second midnight if their care requires hospitalization). Besides the dramatic payment differential between inpatient and outpatient care, keeping a patient on observation for more than two midnights to avoid a possible denial is an intentional violation of beneficiary rights and subjects the patient to higher copays. It should be remembered that these long observation stays are one of the main reasons that the two-midnight rule was adopted.

In our conversation, the medical director explained that she was trying to convey that the TIA itself did not warrant inpatient admission, but rather that it was the medical need for a second night in the hospital that warranted admission. She said that since the two-midnight rule went into effect, they have seen a large number of TIA sufferers admitted as inpatients, kept one midnight, and discharged. Most of these admissions subsequently were denied payment because the expectation of a two-midnight stay was not supported by the presenting condition or the plans for the patient. Here she makes a very valid point; patients presenting to the ED after a TIA who are neurologically intact at presentation should be hospitalized as outpatients with observation, because they will be monitored and undergo testing and go home the next day if test results are normal and no recurrent symptoms present (unless the patient is determined to be at high risk for stroke within the next 48 hours via use of a recognized risk scoring tool and the doctor plans to monitor them in the hospital for the full 48 hours of risk).

The next moment of confusion occurred when a caller inquired about when the clock starts for counting observation hours. In response, the medical director stated that observation cannot be billed until the care is provided by “inpatient personnel,” such as a nurse coming down from an inpatient unit to care for the patient in the emergency department (ED). But the CMS rules are silent on location of care and type of personnel. There is no need for “inpatient personnel” to come to the ED to provide that service. In fact there is no such thing as “inpatient personnel,” technically speaking. Furthermore, observation is an outpatient service, so it makes no sense that this would require “inpatient personnel.” Once orders for observation services are written, the nurse (be it an ED nurse in the ED or a medical/surgical nurse on the med/surgical unit) will be carrying out those orders and the hours can be billed compliantly. In fact, a hospital can bill a whole inpatient stay of any length even  if the patient is stuck in a hallway bed in the ED and never makes it to a bed on an inpatient unit, as happens in some busy hospitals.

The issue of counting midnights for transfers also has resulted in incorrect interpretations. The medical director first said that CMS has given conflicting guidance and recommended that hospitals keep treating transfers as they always have by reassessing the patient upon arrival at the receiving hospital to determine if inpatient care is warranted. She later retracted that response, however, and correctly stated that receiving hospitals should count the time spent at the other hospital in determining if the two-midnight benchmark was or will be reached. This means that if a patient spends two midnights in the hospital after a heart attack and is transferred to another hospital for cardiac catheterization, the hospital should admit that patient as an inpatient even though he or she may go home in four hours.

To demonstrate the importance of every single word of some guidance, one very astute listener (a shout-out to Justin, whoever you are) asked about one of their case scenarios. This patient presented to the ED in rapid atrial fibrillation and was started on intravenous Cardizem and converted to sinus rhythm. The patient then was admitted as an inpatient, was monitored and started on amiodarone, and discharged the next day. On the slide it read, “this case should be billed outpatient.” Justin properly pointed out that because the patient was already discharged, the stay could not be billed outpatient, but rather the hospital must self-deny the admission and rebill as inpatient part B. Now if that slide had read “this patient should have been placed in outpatient observation,” it would not have created this confusion, because the issue was that the original status determination was incorrect; the plan for the patient was to monitor overnight, start a medication, and then discharge the next day if the patient remained stable.


The last source of confusion related to the need for a physician to explicitly state his or her expectation of the length of stay in numerical form (as in two days, five days, etc.). It was asked whether an admission would be denied if the physician did not indicate a number of days. Here, the medical director said that it was required and even went so far as to state that it was good medical practice to state the expectation of the length of stay, even in cases in which the two-midnight expectation is clear. The caller asked if the doctor must predict how much time a patient with septic shock will require in the hospital, being as patients with septic shock clearly meet the two-midnight expectation, and the medical director stated that it is required that the physician document the expected length of stay in all cases. She further stated that in this situation, the doctor should note that “I expect this patient with septic shock to need five days in the hospital.”

Fortunately, in this instance one of the educators on the call from the MAC intervened, stating that this is not a requirement and that cases would not be denied if there was no numerical value. This view is clearly supported by CMS, which indicated very clearly in its Feb. 27 Open Door Forum that a doctor never has to forecast a number of days. “The next major point is the estimated or actual required time of hospital time. Just to clarify, this does not require a presumption or a complete estimate based at the time of the order,” it was said at the time. “A physician does not have to state how much time they expect a patient to be in the hospital. It would be acceptable for a physician to – or ordering practitioner to state that a patient – that time as an inpatient was medically necessary, and that can be documented by the medical record.” Remember, if the doctor did not expect the patient to require two midnights of care, the doctor would not have written the admission order in the first place.

Here the medical director informed me that once again her comments were based on the contradictory guidance from CMS that frustrates even the MACs. When the actual rule states that certification must indicate the expected length of stay and then CMS states in a phone call that the physician does not have to state an expectation in days, it is very difficult for them to make a definitive statement on this issue. She did stress that whether an expectation is documented or not, the clinical documentation must describe a patient that reasonably would be expected to require two midnights in the hospital.

The medical director also wanted me to tell readers that they are seeing quite a few claims of patients being kept for long periods in outpatient with observation awaiting placement in a hospice inpatient bed. She stated that the two-midnight rule applies to these cases; if there is a need to keep the patient hospitalized for symptom management while awaiting an inpatient hospice bed, the patient should be admitted as inpatient and not kept as an outpatient. On the other side of the coin, she noted that patients who require care that can be provided safely in a skilled nursing facility should not be admitted as inpatient or kept hospitalized in an acute-care hospital.

The two-midnight rule is actually quite easy to understand; reasonable estimates of anticipated length of hospitalization are acceptable and physicians just need to do a better job documenting their thought processes, concerns, short-term risks, and plans so auditors can read the record and identify a patient who likely will need at least two midnights in the hospital. But with the MACs having such trouble interpreting and teaching the rule, what will happen once the Recovery Auditors resume work in 2015? It is anyone’s guess. Until then, we all need to attend as many educational sessions as we can, ask a lot of questions, provide CMS references when their advice differs from CMS guidance, and remember to provide CMS feedback on MAC performance by sending comments, both good and bad, to the mailbox.

The fact that this medical director took the time to call me in the evening at home shows that the MACs are doing their best to be partners with hospitals in complying with these new rules.

About the Author

Ronald Hirsch, MD, is vice president of the Regulations and Education Group at Accretive Physician Advisory Services at Accretive Health. Dr. Hirsch’s career in medicine includes many clinical leadership roles at healthcare organizations ranging from acute care hospitals and home health agencies to long-term care facilities and group medical practices. In addition to serving as a medical director of case management and medical necessity reviewer throughout his career, Dr. Hirsch has delivered numerous peer lectures on case management best practices and is a published author on the topic. He is a member of the American Case Management Association and a Fellow of the American College of Physicians.

Contact the Author

To comment on this article go to

Share This Article