Dramatic decreases in hospital lengths of stay can occur. Here's how

I recently wrote about my plan to reduce hospital readmissions. Now I will discuss the problem of reducing length of stay.

The recent hurricane in New York City and the closures of some hospitals requiring the transfer of a large number of patients reminded me of something that happened on 9/11/2001.

I was working at a hospital near New York. You may recall that among the many problems that day was a breakdown in communications. Reliable information on the number of casualties and extent of injuries was hard to determine.

Late on the morning of 9/11, a meeting was held at my hospital. We canceled all elective surgery and decided to discharge as many patients as possible in preparation for the injured who, sadly, never arrived. Victims either got out of the World Trade Center and walked away or perished. Injured patients were few and were cared for by hospitals in New York.

Unaware of what was really happening in the city, we made rounds on every patient and discharged nearly 50 who otherwise would have stayed a day or two longer. As far as I know, there were no complications related to what seemed to be a premature departure from the hospital for many.

The next day someone wondered why, if we were able to discharge so many patients on the day of a disaster, could we not do so more often?

Granted, once a wholesale cleanout took place, there would probably not be 50 patients eligible for discharge every day. But it might be 10 or 15. Multiply that by a few thousand hospitals and you might see quite a savings in the cost of medical care.

Will it happen? I doubt it. For one thing, ours was not the only hospital to have that experience. If it was going to happen, it would have caught on by now.

Why not? On 9/11, the inpatients were motivated to leave. They were scared. They wanted to be with their families. They felt like they were helping others—the potential victims who never materialized. It would be hard to muster those feelings every day.

I have written before that hospital length of stay is not simply a matter of the physician deciding that a patient can go home. The patient may not want to leave. There may be no support at home. There may be no one to drive the patient home. The nursing home or rehab center may not have an available bed.

Still, it is interesting to contemplate what occurred on 9/11/2001 and why we can’t be more proactive in discharging patients.

Not news: Shorter hospital lengths of stay = higher readmission rates

A study in the Journal of the American Medical Association from researchers at Duke University reveals that 14.5% of American patients who were admitted for acute myocardial infarction were readmitted within 30 days. Patients in other countries were readmitted only 9.9% of the time, a significant difference. (p = 0.001)

Patients in the US had significantly shorter initial hospitalization lengths of stay [LOS], 3 days vs. 6 days. The authors concluded that the short initial hospitalization LOS is highly likely to be the cause of the increased rate of readmission.

Why are patients being readmitted in such high numbers? In a press release the senior author of the MI study, Dr. Manesh R. Patel said, “In the United States, care is episodic, not always coordinated, and it's not clear in many cases whether the patient is seen again by the same doctor or care team within the first seven days after discharge.” Can you say “hospitalist”? The same doctor may not even see the patient every day in the hospital. That’s how the hospitalist model works.

Here’s an interesting fact. The 30-day readmission rate of 14.5% for US MI patients is actually less than that of all medical discharges (21.0%) found in a large study of Medicare patients published in the New England Journal of Medicine in 2009. That study also showed that readmissions are very costly.

I’m a mere surgeon, not a statistician or epidemiologist, but I can tell you that it is not surprising that the shorter an initial hospital stay is, the higher the readmission rate will be.

Here’s my theory on this subject. In case you haven’t heard, we are under tremendous pressure to discharge patients quickly. In every hospital, utilization review [UR] is conducted by squads of clipboard-wielding nurses, who are trained to prod doctors into sending patients home as soon as possible. Third party payers demand short LOSs. Digression: Lengths of stay are often based on the Milliman care guidelines. But remember, guidelines are just that. They aren’t meant to apply to 100% of patients.

I used to be intimidated by the pressure to send patients home. It was very uncomfortable to go into a patient’s room and explain to him that he had to be discharged because the some arbitrary authority had mandated it. No matter how I spun it, the patient felt that I was the villain for “kicking him out” before he was ready.

I have reached an age where I really don’t care what the UR nurse says. I refuse to jeopardize my relationship with a patient to please the bureaucrats. When I feel a patient is medically ready, I discharge him, but I do listen to the patient. If he has a valid reason for not wanting to go home, I will usually acquiesce.

I advise my medical colleagues to keep their MI patients in the hospital for what they believe is a reasonable time, UR be damned. Show them the JAMA article if they give you a hard time.

Robotic gastrectomy: Is it better?

Here is yet another paper, this time from Archives of Surgery, extolling the virtues of robotic surgery. Thus time the subject is gastrectomy for cancer. Surgeons in Korea retrospectively looked at 827 gastrectomies for cancer; 591 of which were done with standard laparoscopic technique and 236 were done robotically. Preoperative co-morbidities were similar but the robotic patients were an average of 4 years younger, which was statistically significant, p < 0.001.

The main results were that the complications, deaths, extent of lymph nodes removed were not significantly different between the two groups. The robotic surgery took significantly longer (49 minutes) to perform, p < 0.001. The average estimated blood loss was statistically significantly less in the robotic patients, (91.6 mL vs 147.9 mL, p = 0.002). Hospital length of stay (LOS) was significantly shorter for the standard laparoscopic group, 7.0 vs. 7.7 days, p= 0.004.

The authors concluded “robotic gastrectomy [has] better short-term and comparable oncologic outcomes compared with laparoscopic gastrectomy.”

Is this conclusion valid? Let’s take a closer look.

The study was not randomized nor was it prospective. Despite the similar number of patients with co-morbidities in both groups, patients chosen for robotic surgery were obviously selected for suitability. Other confounding factors may not have been unaccounted for. The only short-term advantage for robotic surgery was in the estimated blood loss. The authors themselves admit, “The statistically significant difference in 56.3 mL of blood loss between the robotic and laparoscopic groups may not translate into much clinical benefit for every individual patient.” This is certainly true. In addition, estimated operative blood loss is notoriously inaccurate. A study involving spine surgery showed that estimated blood loss exceeded measured blood loss by a mean of 248 mL (p = 0.0001). And since the study was not blinded, the blood loss estimates could easily have been biased.

Hospital LOS was actually longer for the robotic patients, amounting to 0.7 of a day or 17 hours. The authors tried to explain away the difference in LOS by pointing out that the robotic group had a couple of outliers who had really long LOSs. As I have blogged before, LOS is a soft endpoint which can be affected by many things other than the clinical state of the patient.

The study did not mention readmission rates for either group. Long-term follow-up was not included in the study, meaning that the oncologic outcome has yet to be determined. The issue of cost was neatly avoided by a convoluted explanation of the uniqueness of the Korean national insurance program and individual hospital differences. However, the methods section of the paper did note that the patients would have to pay for the extra costs of robotic surgery themselves. This suggests that the robotic patients may have been from a higher socioeconomic group. Robotic surgery is unquestionably more expensive as the robot itself costs $1.5-2.0M with an annual service contract of at least $150K and disposable instrument costs of about $2K per case. A New England Journal article estimated the actual additional cost of each robotic procedure at $3.2K.

So you tell me, does robotic gastrectomy have better short-term results than standard laparoscopic gastrectomy?

Uninsured patients released from hospital sooner than insured: Significant? You tell me.

There’s a bit of excitement on Twitter today with a number of tweets and retweets about a paper just published in the Annals of Family Medicine which shows that uninsured patients are being released from  hospitals significantly sooner than insured patients. The numbers don’t lie.

From the abstract: “Across all hospital types, the mean length of stay … was significantly shorter for individuals without insurance (2.77 days) than for those with either private insurance (2.89 days, P = .04) or Medicaid (3.19, P <.01).” These are statistically significant differences.

The authors conclude, “Future research should examine whether patients without insurance are being discharged prematurely.”

Let’s look a little closer at these numbers. The difference between the uninsured length of stay (2.77 days) and those with private insurance (2.89 days) is 0.12 days or to put it another way, 2.9 hours.

Do you really think that a difference in hospital length of stay of less than 3 hours is really clinically significant? I don’t.

Here’s another problem with the paper. Length of stay is what is called a “soft” endpoint. Having practiced surgery for 40 years, I can assure you that length of stay is very often not determined by the type of illness, treatment rendered, skill of the physician or any other parameter you can think of.

Here is what I mean. Just yesterday, a patient told me he could not go home on the day he had his laparoscopic cholecystectomy because his sister, whom he lives with, gets upset whenever he comes home from the hospital. He felt she needed another day to adjust. Patients have told me, “No one can come and pick me up today.” The care manager says, “The bed at the nursing home isn’t available today.” Three weeks ago we couldn’t send some patients home because there was a massive power outage in our area. This list of excuses goes on and on.

I have written before about the problem of things being statistically significant but not clinically significant.

The paper is another example of statistical significance not corresponding to clinical significance.