The importance of sleep is perhaps most realized when we become sick. When we are hospitalized and most in need of every ounce of health, though, hospital care practically guarantees that we won’t get good sleep. Fortunately, two approaches hold promise to improve sleep for patients: one organizational, and the other a common trick of the trade among those of us working in behavioral economics.

Recently I was all-too-miserably reminded of the challenges of hospital sleep when I spent a fitful night recovering from surgery to remove a small kidney tumor. Unlike some patients in that situation, my sleep was not disturbed by pain or nausea; I was lucky to avoid both of those postoperative complications. Instead, my sleep was interrupted, hourly, by clinicians taking care of me. There were vital sign checks every four hours, a frequency that makes sense given that I had just had part of my left kidney removed. Sometimes sleep interruptions are necessary in order to monitor patient conditions. But those vital sign checks, at midnight and 4 a.m., were not the only interruptions I experienced that night. At 3 a.m., if my very foggy memory serves me correctly, someone came into my room to draw blood for follow-up laboratory tests. Several other times that evening, the machine hovering near my left ear beeped to tell me that one of my IV medications had run out; I would push the nursing button and tell the person at the desk about the beeping, and eventually someone would come in and either replace the empty IV bag or turn the alarm off.

Between 10 p.m. and 6 a.m., I did not go more than an hour without some kind of interruption.

As I have already suggested, some of these interruptions are necessary. But many are not. And the consequence of too many sleep interruptions is that patients do not heal as quickly as they would otherwise, thereby not only reducing their quality of life but also driving up medical costs. Indeed, as I have written elsewhere: sleep disturbance is a leading cause of hospital complications, such as falls and delirium. Poor sleep has also been linked to reduced immune function,worsening blood pressure control and mood disorders. All of these problems potentially impair the ability of patients to recover from the acute illnesses that caused them to be hospitalized.

How do we improve hospital sleep?

First, hospitals could make simple organizational changes. During my recent hospital stay, for example, a major contributor to my interrupted sleeping was the specialization of tasks across different hospital personnel. When the IV machine beeped, it was the nurse who helped out, her training being necessary to monitor the IV lines and medications. When it came to measuring my vital signs, though, a nurse’s aide was sent to accomplish the task. And a phlebotomist came to draw my blood. Specialization matters. The doling out of these duties to different people — with different skills and different pay grades — makes great economic sense, and in many ways improves hospital quality of care. But such specialization interferes with sleep, because the different people performing each of these duties enter patient rooms at different times of the night.

There is a better way to coordinate these various clinicians to reduce sleep interruptions. For example, phlebotomists could coordinate their work with nursing aides. Imagine that instead of coming into patient rooms one hour apart from each other, the two came in together: “We are here to check your blood pressure and draw some blood,” they would say (maybe even in unison!). That little change would eliminate one interruption. A second change could also improve patient sleep: more flexibility in the timing of vital sign measures. If, for example, a patient’s IV machine beeps at 11 p.m. and the next check of her vital signs is due at midnight, the nurse could bump up the vital sign measures by an hour, since the patient is already awake.

Indeed, it was an 11 p.m. vital signs wake-up call that led to research that proves the value of my second approach to improving patient sleep: increasing the use ofsleep protocols designed to minimize unnecessary interruptions. More on sleep protocols in a bit, but first let me tell you about that 11 p.m. wake-up call.

Melissa Bartick is a hospitalist in the Harvard medical system, a physician who focuses mainly on inpatient rather than outpatient care, treating patients who have been hospitalized with acute or chronic illnesses. Spending as much time in hospitals as she does, Bartick has long recognized the problems created when patients have difficulty sleeping. But it took her own hospitalization to convince her how fixable this problem is.

Bartick had spent an exhausting evening being evaluated for an acute illness in her hospital’s emergency department. She finally made it up to a hospital bed around 10 p.m., where the nurse checked her vital signs and made sure she was receiving appropriate treatments. Finally allowed to rest, Bartick quickly fell asleep only to be awoken at 11 p.m. for … another vital sign check! She was not awoken because her illness was so acute that she needed hourly assessments. Instead, she was awoken because hospital protocol required nurses to check vital signs each shift, and the night shift began at 11 p.m.

“By the time I left the next morning, I had half of my research design worked out,” Bartick told me. When Bartick was healthy again, she conducted the clinical trial she had begun designing that night. In the trial, some patients, at random, were cared for under a sleep protocol, an order that alerts nursing staff to eliminate all unnecessary middle of the night intrusions — eight hours of quiet time for patients, with darkening of the room and avoidance of waking patients for nonurgent matters. The protocol reduced sleep disturbances by 38 percent.

How do we make sleep protocols the norm among stable hospitalized patients? We change hospital practice so that minimization of sleep disturbances becomes the default condition for how to care for non-critically ill patients, with more frequent sleep interruptions only occurring when physicians actively indicate that such interruptions are clinically necessary. Research in behavioral economics has demonstrated that people are strongly influenced by default options when making decisions. Employees are more likely to contribute to retirement funds when such contributions are automatic. People are more likely to donate their loved ones’ organs in countries where such donations are default policy.

Currently, the default setting in most hospitals is to prioritize testing over patient sleep. While patients in intensive care usually do need to have their vital signs monitored closely, and often need multiple blood draws each day to monitor rapidly changing clinical circumstances, many hospitalized patients do not change enough in their clinical course to require routine middle-of-the-night interruptions. Patient sleep would be improved overnight, literally, if hospitals established new default procedures — for instance: “all patients in non-ICU settings will be cared for under sleep protocols after 36 hours in the hospital, unless the physician indicates otherwise.”

There is one problem with making sleep protocols into default procedures — such protocols cut against tradition. “Hospital cultures are very difficult to change,” Bartick told me, “especially when there is so little incentive for hospitals to make those changes.” She explained that the problem is aggravated by low patient expectations: “We are trying to make hospital care more patient-centric, even looking closely at things like patient satisfaction measures. But sleep interruptions do not reduce patient satisfaction, because patients assume that all of those interruptions are medically necessary.” When Bartick and I were each hospitalized, it was not sleep interruptions per se that disturbed us. It was the fact that as physicians, we knew that those morning labs and middle of the night vital signs were not medically urgent.

The irony is that because of increased emphasis on “patient centric care,” hospitals sometimes act more concerned about patient satisfaction than patient health. Perhaps hospitals will change their practices when patients realize that hospitalization and sleep are not always mutually exclusive.

Peter Ubel is a physician and behavioral scientist who blogs at his self-titled site, Peter Ubel and can be reached on Twitter @PeterUbel.  He is the author of Critical Decisions: How You and Your Doctor Can Make the Right Medical Choices Together. This article originally appeared in The Atlantic.