Friday, July 15, 2011

"I Can't Sleep": A Better Treatment for Insomnia


"Doctor: I am having trouble sleeping."

This has got to be one of the most common complaints we here from our older patients. Older patients often feel doctors don't take their sleep problems seriously. And perhaps this criticism that doctors dont take sleep complaints seriously enough is justified. After all, insomnia has major effects on quality of life in older persons. It is incredibly anxiety provoking to lay in bed night after night and not be able to fall asleep. And the general fatigue one feels after a sleepless night is awful.

One reason health providers may downplay sleep complaints is because they few sleep problems as less serious, and more minor than the diseases on a patient's list of diagnoses. But often, insomnia may have greater effects on quality of life than the diseases on the diagnosis list.

But another reason may be the discomfort health providers feel with the medicines used to treat insomnia. In particular many providers are concerned with the side effects of these medicines. Some of them are associated with higher rates of falling and hip fracture. And while most treatments are only recommended for short term use, once started, many patients end up on sleep medicines for prolonged periods of time.

A recent study published in the Archives of Internal Medicine, led by Dr. Daniel Buysse, from the University of Pittsburgh Department of Psychiatry suggests there may be a better way to treat insomnia. The study showed that a behavioral approach that did not use drugs was very effective.

A type of behavioral intervention known as cognitive behavioral therapy has long been known to be an effective treatment for insomnia. However, few patients ever actually receive this treatment. Few providers are trained in its use, and it is too cumbersome for many patients.

This pragmatic study examined whether a simplified brief behavioral treatment would work. They designed a brief treatment that could be administered by a nurse practioner with no previous training in sleep disorders.

The intervention focused on teaching patients to avoid common habits that conspire to promote insomnia and establish habits that promote normal sleep rhythms. These were the 4 key principles that were taught to patients:


  1. Reduce the amount of time spent in bed. Only stay in bed for sleep or sexual activity
  2. Get up at the same time everyday, regardless of sleep duration.
  3. Do not go to bed unless sleepy
  4. Do not stay in bed unless asleep

Patients were given a prescription for low stimulation activities they could do while not in bed that would promote sleepiness.

79 older persons (average age of 72) were randomly assigned to either receive this treatment or receive the equivalent of usual care. The main components of the intervention were administered by the nurse practioner in a 60 minute session. To boost the effect of this session, the nurse did a follow-up phone call one week later, met with the patient for 30 minutes at 2 weeks, and phoned the patient at 3 weeks. All in all, the intervention required about 2 hours of the nurse's time. So this is a really inexpensive treatment.

But while the intervention did not cost much, the benefits were very impressive:

  • After 4 weeks, 67% of patients getting treatment were significantly better. Only 25% of usual care patients were significantly better.
  • After 4 weeks, 55% of patients getting treatment no longer met diagnostic criteria for insomnia. In contrast, only 13% of usual care patients no longer had insomnia.
  • There was some evidence the benefits persisted for six months after the intervention. Of the patients who were examined at 6 months, 64% did not have insomnia. However, the number of patients not followed at 6 months was high. This limits the abilty to know for sure how durable the intervention was.

One interesting question posed by interventions like this is whehter they work if used by providers other than the study investigators. On the one hand, the approaches used here seem very pragmatic and easy to incorporate into clinical practice. It is easy to imagine large health systems like Kaiser or the VA implementing something like.

On the other hand, behavioral interventions such as this seem to always have a "black box." In this case, I believe the black box is the nurse who trained the patients to improve their sleep habits. It is one thing to have an intervention manual with a list of steps an intervention nurse should take. It is another thing to successfully deliver the intervention. I suspect that there was something very special about the nurse used in this study. I suspect she had a remarkable bed side manner, and an empathic approach that won the trust of her patients. This empathy and trust make it much more likely patients will successfully change their sleep habits.

This raises an important issue for both this and other behavioral interventions. In addition to describing the mechanical steps of the intervention, the studies need to tell us more about the people who delivered the intervention. It would be nice to know the qualities that made this nurse successful. In short, we also need a guide that tells us who to hire to administer the intervention.

Are we ready to recommend this program for general use? Probably not quite yet. The study was small, the exclusion criteria were too stringent, it was done at just one site, and the evidence of durability was somewhat limited. But we certainly have enough evidence to justify a large multicenter randomized trial that will provide definitive evidence. Hopefully the NIH will see the wisdom of sponsoring such a trial.

In the meantime, the study gives guidance to health providers and older patients about practical steps that can be taken to improve sleep in persons suffering from insomnia. And the study gives hope that we can effectively treat insomnia in many patients without the use of medicines.
by: Ken Covinsky

7 comments:

Jeanne Lahaie, RN-BC, MS, GCNS-BC said...

As part of the new Hospital Elder Life Program (HELP) at CPMC, I provide education to patients aged 70+ regarding good sleep hygiene if there is a self report of insomnia or report from night shift RN's 1-hr rounding. I use our online medical record system to "order" a nonpharm sleep protocol either by our specially trained HELP volunteers or our RNs if the volunteers are unavailable or if the patient does not meet HELP criteria. Patients seem to be very interested in what will help them sleep without medication, and I'm always surprised that what seems like basic information about such a common problem seems to be such a revelation. The HELP protocol calls for warm milk or decaf tea before bed and a shoulder/hand rub. We also supplement that by "ordering" a warm blanket before bed, keeping the room light during the day and dark at night, limiting the stimulation of television and/or caffeine before bed, and promoting mobility during the day. Mobility in this case can range from only ROM exercises to walking a patient around the unit, depending on the patient's condition and discretion of the MD/PT/RN. One of the criteria for enrolling a patient in HELP is insomnia or a sleep med on the MAR. It's too early to tell if our practice of flagging a med used for sleep (such as Benadryl or >7.5mg Restoril, or >50 mg Trazodone) has truly lowered overall use of sleep meds, but we have been successful in lowering doses of Restoril and Trazodone as well as using less Ambien and Benadryl. So far, we haven't had any falls in our HELP patients (that I'm aware of), and though there were 2 falls on another unit just prior to our study when Trazodone was administered during hours typically associated with wakefulness, we haven't seen that on our unit since we launched HELP in May 2011. As a board certified pain management RN and gerontological clinical nurse specialist, I can't underscore the importance of the Goldilocks approach to pain management in older adults in particular -- not too much, not too little. Morrison et al. (2003) found a 9x greater risk of delirium in older adults with hip fracture if they were undermedicated for pain. Concern for the patient's overmedication can be alleviated by taking into account whether the patient is opioid naive or opioid tolerant and whether the medication is likely to be eliminated or accumulated. If renal function is an issue, for example, morphine is likely not the best choice. If pain is adequately treated, thereby lowering discomfort and subsequent agitation, a good night's sleep is more likely.

Eric Widera said...

Great post Ken and great comment Jeanne about HELP and what you are doing at CPMC. It is amazing how simple interventions can impact patient care so significantly.

I also want to point out how common this is in the population that we see on our palliative care service. The Journal of Clinical Oncology published a study in January by Carney, Koetters, and Cho et al that estimated 40-50% of cancer patients undergoing radiation therapy and their caregivers had clinically significant sleep disturbances. The Journal of Pain and Symptom Management also just published a study of advanced cancer patients showing 85% had sleep disturbances (JPSM 2911;41:819-827).

ken covinsky said...

Thanks Jeanne for describing your great work at CPMC. It is great to hear that the HELP protocol has been implemented in San Francisco.

Reading your comment, it struck me that essentially all the standard recommendations for sleep hygiene are subverted by standard care in hospitals. Sure provides a case for better processes of care.

Lindsey Yourman said...

he he, my night float laughed out loud when I described nonpharmacologic sleep protocols for elders (and that I'd dc'd the prn ambien for our older patient). So, if your night float insists on a pharmacologic prn for insomnia, is there a class of sleep-meds that tends to pose less risk for delirium? I imagine anything shorter-acting but beyond that I'm not sure . . .

Eric Widera said...

A good study to look up is the following:

Glass J, Lanctôt KL, Herrmann N, Sproule BA, Busto UE. Sedative hypnotics in older people with insomnia: meta-analysis of risks and benefits. BMJ. 2005;331(7526):1169.

This was a meta-analysis of sedative hypnotics. Basically what they found was a number needed to treat for improved sleep quality of 13 and the number needed to harm for any adverse event of 6. They also try to seperate newer benzodiazepine receptor agonists and benzodiazepine. Basically they found no difference in adverse effects. So to some it up, you can use on of these agents for sleep but they are more likely to cause more problems than good, especially in the elderly. Also, none of these trials would also even consider someone for entry into their study if they are hospitalized for a medical condition.

Ramelteon, a new melatonin receptor agonist, may have less adverse effects, although efficacy may be less too. Plus, it really hasnt been studied in this population.

A glass of warm milk may be a nice alternative with less risk (unless they are lactose intolerant)

ken covinsky said...

Sharon Inouye's HELP protocol for insomnia in the hospital is a glass of warm milk and a back rub. When I show her paper to our residents, they always seem to like this part of the intervention.

Lindsey Yourman said...

Thank you so much Eric and Ken! I'm going to have the Glass and Inouye resources handy on the wards . . . "I'm just an intern but I guess what the expert Geriatricians and peer-reviewed research are indicating these days is that . . ." perfect!