Improving patient sleep by shaking up routines

Melissa Bartick, ACP Member, found that simple changes can reduce patients' requests for sedatives.


Often one of the more unpleasant aspects of a hospital stay is the inability to get a good night's sleep. Melissa Bartick, ACP Member, a hospitalist at Cambridge Health Alliance in Cambridge, Mass., sought to change this by crafting an intervention that minimized sleep disruption during “quiet time”—between the hours of 10 p.m. and 6 a.m. Results of her study were published in the Sept. 15, 2009 Journal of Hospital Medicine.

Dr. Bartick found that interruption by hospital staff to take vital signs or give medication was the biggest cause of sleep disruption—greater than noise, roommates or pain. Her intervention involved a ban on these activities during quiet time, except under rare circumstances. To accomplish this, medication orders were changed from every X number of hours to a certain number of times a day (for example, BID), which allowed flexibility of administration. Checks on telemetry patients, which had to be done every four hours, were shifted to involve one night awakening instead of two. Noise monitors were installed in staff areas that flashed lights when the din went above 80 decibels.

The intervention decreased the proportion of patients who reported having sleep disturbed by hospital staff from 42% to 26%, a 38% drop (P=0.009). Patients were also less likely to request sedatives to help with sleep: 32% asked for them pre-intervention, compared to 16% during the intervention, a 49% drop (P=0.004). Subjective sleep scale scores were unchanged, however. Dr. Bartick talked to ACP Hospitalist about the study and intervention.

Q: Where did the idea for this study come from?

A: It was conceived when I got admitted to my own unit. I got sick at work with gastroenteritis, and it was so bad I had to be admitted. I got to my unit around 10 p.m., was checked in, got vital signs taken, and finally fell asleep. I was awoken one to two hours later for my vitals to be checked. I told the aides, “But I just had my vitals done about an hour ago.” They said, “Well, it's every shift, and we're the night shift.”

I thought, ‘This makes no sense at all.’ By morning, I had much of the study designed in my head.

Q: Can you explain a bit about the setting where this study was conducted?

A: Ours was a 28-bed medical/surgical unit. Most of the beds were used for medical patients covered by the hospitalist service. It had three U-shaped pods, so it was very different from the traditional long hallway. The rooms were less exposed to noise, so in a sense the unit design helped control for that potential variation between patient experiences you'd get with a traditional unit setup. Of note: After the study was accepted, that unit of the hospital closed, and we have since moved to our Cambridge campus. We're hoping to transfer the intervention there, but haven't yet.

Q: How did you decide which interventions to use? Did you adapt from previous studies, or create them on your own?

A: We basically thought of them on our own. As the first step, I did pilot interviews with 18 patients in order to figure out which questions to ask on the survey about sleep, and to think about the interventions we might design later. Then [my co-authors and I] solicited opinions from hospital staff at staff meetings. We ran our ideas about the interventions by the staff and invited them to comment and to think of their own ideas.

Q: How were patients informed of the option of taking as-needed sedatives to sleep?

A: The staff didn't go around asking whether patients needed or wanted a sleeping pill; they basically waited until the patient asked for one. The staff also didn't know we were measuring for this in our study. So it was done in a way that reflected usual practice.

Q: Why do you think the intervention reduced sedative use more dramatically among patients older than 65 years?

A: It's hard to say. I can only guess, as a true answer would require a separate study or literature search. They may have a harder time getting back to sleep after being woken up, or they may use more sedatives at baseline.

Q: According to the modified Verran Snyder-Halpern sleep scale you used, patients didn't subjectively report an improvement in sleep. Why?

A: Another way of looking at the results is to say that patients were able to get the same quality of sleep, but with fewer drugs. But I have to point out that 75% of our patients were unable to use the scale, usually because they felt too sick to use it. When the VSH scale was developed and validated, it was done so at a time when inpatients were much less sick than they are today, so I wouldn't be surprised if what it measures is no longer valid. A more objective measure would be EEG monitoring or wrist actigraphy.

Q: Did you have anecdotal reports from patients of improved sleep?

A: Yes. In the pilot phase, we asked a patient what kept her awake. She said she was up all night going to the bathroom because she had a catheter in her bladder that was removed at night. She got diuretics at 10 a.m. and 10 p.m., then they took out her catheter and she was up all night. This is crazy, if you think about it, but it happens all the time. Her experience led us to design, as part of the intervention, no standing doses of diuretics after 4 p.m.

This is a really great illustration of the kind of thing we don't think about. You don't need the diuretics to be ordered every 12 hours; they can instead be given twice a day at any time, as long as the doses are far enough apart to make sense with the metabolism of the drug—which, for her diuretic, was a six-hour span.

Q: Do most medications have that flexibility, where they can be changed from, say, “every six hours” to “four times a day”?

A: Yes, in most cases. That was another big part of the intervention: training providers to change their instructions for medication to BID or QID, etc., from every 12 or six hours. A lot of the providers, both physicians and nurses, didn't know there was a difference. We had to explain that QID was not the same as every six hours.

Q: Are there some medications which can't be changed in this way?

A: Yes, and we had to exempt some medications from the list. Basically, anything that required monitoring of drug levels, like aminoglycosides, we had to exempt. But our pharmacist, who is the second author on the paper, helped us think through which medications would be exempt. So it's important to involve a pharmacist in the intervention. Change is usually interdisciplinary.

Q: How did the staff respond to the placement of noise monitors in their work area?

A: They didn't like it, initially. We had to play with the settings, because if they flash red all the time, people just zone out on what it means. So we had to set them to alarm at a noisier level than is recommended by guidelines. The staff did say that by seeing the lights going off all the time, it was helpful in making them aware of the noise they were making. But it's interesting that noise wasn't a big factor reported by patients as affecting sleep. There was a trend toward improvement with noise reduction in the intervention, but it didn't reach significance.

Q: Do you think the effectiveness of the measures might vary with patients' disease severity?

A: Possibly. You know, we didn't use any ICU patients. They were all in a medical/surgical unit, and within that some people might be sicker than others. I think you'd need to design another study using APACHE scores or something to get at the sickness levels. I will say that we got these significant results even though there were quite a few telemetry patients who had to be awoken every four hours to do vitals. We changed the times, though, so they were only woken once a night instead of twice. So instead of going in at 12 a.m., 4 a.m., 8 a.m., and so on, we changed it to 10 p.m., 2 a.m., 6 a.m., etc.

Q: How easy was it to get staff buy-in for the interventions?

A: We had a very strong nurse manager who was very invested in the success of the study and in improving patient-centered care. She basically said, “We are changing all the vital signs to two hours earlier, and night shift people, you'll do it at the end of shift instead of the beginning, and that's how it is.” And of course this was after the meeting where we discussed these things as possible ideas. So in this case, the staff just did it. Now, we have had some comments on the study, with people saying, “But the night nurses won't have anything to do now,” which is sort of the point, but is hard for some to accept. The whole concept of being patient-centered around sleep involves, on some level, admitting that the way you've been doing things for your whole career might have been wrong. And that is hard.

Q: What would you recommend to others who might want to try this intervention, in terms of getting staff buy-in?

A: I would really recommend involving staff in designing the process, and that includes the nurses, nurses' aides and secretaries. A lot of people have something to do with the process and work flow. Also, at our meetings, we had the respiratory therapist there as well, and this was the person who pointed out that prescriptions should be written as “every four hours while awake for treatments”—otherwise, the patient had to be awoken if it said only “every four hours.” So, do things in a way where the people involved have some ownership. And, like anything, if you have people who are particularly resistant to the idea, make sure they are on the committee that helps change things.

Q: How are hospitalists, in particular, affected by this intervention?

A: One very important thing is for them to learn to write flexible orders—like changing “every twelve hours” to BID—whenever possible, and to really consider their medication choices. So, let's say I'm trying to choose between a couple of different drugs. I might be more inclined to order something that can be given fewer times a day. You might also think about the fact that running IV fluids overnight can wake people up, because if the bag runs out, or if the patient moves an arm in a certain way, it can trigger a loud alarm. So is that IV necessary overnight? Also, hospitalists can help train new residents to be aware of flexible order writing and patient-centered care. And, of course, it also helps to have hospitalist champions to support the intervention generally, and get the whole staff excited about it.

Q: How do you train staff on the intervention?

A: You do the usual things—you use posters and storyboards and staff meetings. You also make sure you have a couple of champions on each shift who will talk one-on-one with all the people. That really works, and is how I've reached most people. It's not a terrific amount of work, because you'll be talking to them anyway. You also collect feedback through staff meetings.