As physicians, we learn about the Starling curve in the first year of medical school. It expresses the relationship between left ventricular end-diastolic volume and cardiac output. The same can...
This is a burning question that every hospital CEO and medical director wants to know since most hospitals end up subsidizing hospitalists. And the answer is… it depends. Anyone who tells you categorically that the right number for every hospital is 15 patients a day is wrong.
What the hospital wants from the hospitalist is good value for the amount of money that the hospital pays to support the hospitalist. If the hospitalist is seeing too many patients per day, then there is a risk of bad things happening including medical errors, physician burn-out, increasing length of stay, worse patient satisfaction scores, and patient bottlenecks caused by later times of discharge. If your hospitalists are seeing too few patients a day, then you are not getting your money’s worth from them. Here are some of the variables that I look at when I’m deciding if our hospitalists are seeing the right number of patients.
- Patient case mix index (CMI). This is a pretty easy number to get from your hospital’s billing office. The higher the number above 1.0, the more medically complex the patients. It will give you an idea of the complexity of patients that the hospitalist is seeing and as a result, how much effort the hospitalist needs to put into the care of a given patient. Here is an example of 3 inpatient services from our own hospital. Service A is an attending-only (non-teaching) service that covers general medical admissions and the ICU – their CMI is 1.45. Service B is a teaching service with residents and a hospitalist attending that takes general medical admissions but does not cover the ICU – their CMI is 1.21. Service C is an attending-only (non-teaching) service that takes mainly cardiac admissions and a consequence, they have a high percentage of observation chest pain admissions – their CMI is 1.10.
- Teaching or non-teaching service. The ACGME limits the service census to 10 patients per intern. There is a time trade-off for hospitalists on teaching services: the residents will do a lot of the time-consuming work for the attending hospitalist but the hospitalist has to do uncompensated teaching time; in a healthy teaching service, these should balance out. A teaching service with a cap of 10 patients is rarely a full-day work for the attending hospitalist so he or she has to have some other income generating activity.
- Admitting service versus consultative service. Patients with medical illnesses requiring admission to the hospital are by definition sick. On the other hand, those coming in for an elective joint replacement generally have minimal medical conditions or their medical conditions are in good control. The hospitalist co-managing medically stable patients in for elective orthopedic surgeries can see considerably more patients per day than the hospitalist managing medical admissions coming in from the ER.
- Advanced practice providers. Physician assistants and nurse practitioners can allow a hospitalist to see more patients per day but they come at a cost, generally one-third to one-half the salary of a hospitalist. A physician assistant that allows a hospitalist to see 25 patients a day rather than 15 patients a day is probably worth it. However, if the use of a physician assistant only allows that hospitalist to see 18 patients per day rather than 15 patients per day, it may not be worth it.
- ICU or non-ICU. In the ICU, patients need to be re-assessed multiple times a day by the physician, there will be more bedside procedures to be done, and there will be more minute-by-minute orders to be placed. A hospitalist in the ICU may only be able to cover 12 patients a day whereas that same hospitalist may be able to take care of 20 patients a day on a general medicine nursing unit. That has to be tempered with the availability of additional consulting physicians – a general internal medicine hospitalist in the ICU may be able to see more patients if there is a critical care medicine consultant also rounding on the patients.
- Day shift versus night shift. There is a lot more work per patient on the census during the day than during the night. During the day, patients need to be rounded on, there are family meetings, and patients need to be discharged. During the night, the hospitalist does emergency admissions and deals with urgent/emergent inpatient issues. A night shift hospitalist may be able to cover 60 patients but a day shift hospitalist, only a quarter of that.
- Observation versus regular admission patients. This is a tricky one. On the one hand, observation patients are less medically complicated than regular admission patients and don’t have as much discharge complexity (need for home health, nursing homes, etc.). On the other hand, observation patients have a much shorter length of stay so a hospitalist with a lot of observation patients will be doing more time-consuming admissions and discharges per day than a hospitalist with mostly regular inpatient admissions. Currently in the U.S., the average hospital has 26% of their average patient hours being observation patients. Our hospitalist service that sees primarily cardiac patients has 50% of their patients in observation status; another hospitalist service that sees general medical admissions has 20% of their patients in observation status.
- Ease of documentation. If a hospitalist has a really good electronic medical record with vital signs, medication records, progress notes, lab reports, etc. then it can be far more efficient to take care of patients than if medical records are fragmented. For example, at one hospital in our community, the physician progress notes are handwritten in a paper chart, the vital signs and medication records are on one computer system, and the lab and radiology reports are on another computer system. It is neither possible or safe for a hospitalist to see as many patients in this environment as they can in a hospital with a single, integrated electronic medical record.
- Patient captivity in the electronic medical record. By this, I mean whether the hospital and the primary care physicians caring for the patients who get admitted to that hospital use the same electronic medical record. If they do, then it is much easier for the hospitalist to do admissions and discharges since much of medical history documentation is already in the electronic medical record. It is much faster to do an H&P if you can draw in the entire past medical, surgical, family, and social history plus all of the patient’s current medications and doses with one click on the computer rather than having to manually enter all of the information.
- Non-clinical duties. A hospitalist that is spending 2 hours a day in committee meetings cannot see as many patients per day as a hospitalist who has no committee assignments.
- Shift duration. A hospitalist working a 12-hour shift may be able to see 20 patients a day (1.7 encounters per hour) comfortably but that same hospitalist working an 8-hour shift may only be able to see 14 patients a day (1.7 encounters per hour) comfortably. Shift duration also affects the number of shifts per month you should expect your hospitalists to work: if you expect your hospitalists to work 2,300 hours per year, then that is 16 12-hour shifts per month but 24 8-hour shifts per month.
- Hospitalist experience. All hospitalists are not equal. A new hospitalist right out of residency is not going to be as efficient and see as many patients as a hospitalist with 20 years of practice experience. High hospitalist turn-over means more new physicians who cannot see as many patients per day as experienced hospitalists. If you force your hospitalists to see too many patients per day, they will quit and you will end up with excessive hospitalist turn-over.
- Hospital geography. It can take a hospitalist caring for 15 patients on 6 different nursing stations more time per day to manage than a hospitalist caring for 20 patients on a single nursing station.
- Encounters versus census. We often focus on the hospital midnight census to measure hospital capacity. But that only measures the patients who are in a bed at midnight and over the course of the day, there is going to be bed turnover as patients are admitted and discharged. If the patient length of stay is long, then the midnight census will be close to the number of daily patient encounters per physician. If the length of stay is short, then the hospitalists will have a lot more patient encounters per day than the midnight census.
- Census variability. Too often, we look at census averages and although this is useful, it doesn’t tell the whole story. For example, last Monday, we had 109 medical/surgical beds occupied and by Thursday we had 140 – that is a 31-patient swing in just 3 days. This means that the hospitalist services all had more patients per hospitalist on Thursday than they did on Monday. So, if your hospitalist census averages 15 patients per physician but the census fluctuates between 8 and 25, there are going to be days that the hospitalists will have a hard time safely caring for those higher numbers of patients. If there is not a surge plan to bring in “risk call” hospitalists on those high census days, you may need to settle for a lower average daily census per hospitalist in order to accommodate those unpredictable days when the hospital census is usually high.
- RVU productivity. This is also a tricky metric because it does not capture all of the work done by a single hospitalist but at least it gives you a ballpark comparative to determine if your hospitalist program as a whole is meeting productivity benchmarks. The MGMA reports that the median total RVUs generated by a hospitalist is 5,900 and the work RVUs are 4,100. These numbers are affected by day versus night shift and other variables.
- Robustness of case management. Case management has to happen whether or not a hospital has case managers. A hospitalist who has to do a lot of the discharge planning because of a lack of case managers cannot see as many patients per day.
- The local market. If your town has several competing hospitals, then each hospital will be competing with the others for hospitalist and if the hospital down the street has an expectation of 15 patients per day and your hospital’s expectation for the same patient population is 20 patients per day, then you are going to lose valuable hospitalists.
- The patient demographic. If your hospital mainly sees patients with good commercial insurance and good primary care providers, then it is easier for the hospitalist to focus on the acute problem that brings the patient into the hospital and it is easier to make discharge arrangements. On the other hand, if you have a high percentage of uninsured or Medicaid patients, then the hospitalist taking care of a patient with pneumonia is likely going to also be spending time tuning up that patients diabetes, heart failure, or hypertension since the only time the patient sees a doctor each year is when he/she is in the hospital.
So what does a medical director or hospital CEO do? I recommend starting with an assumption of 15-18 patients per hospitalist and then working up or down from that number based on the unique features of your own hospital, community, and hospitalist program structure by taking into account the variables I mention above.
August 13, 2016