When The Numbers Don’t Add Up: What Are We Missing?

A lot of emphasis in medicine is placed on outcomes, meaning the changes or results that emerge after someone uses a specific treatment. The most common sleep outcomes are the changes in sleepiness or tiredness or insomnia. Also, some sleep centers inquire about changes in sleep quality.

If you have been working with a higher quality sleep center, there’s a 99% chance you have had to fill in some questionnaires about sleepiness, tiredness, insomnia or sleep quality. Typically, you fill these forms in at the beginning of your treatment program, and then periodically through the course of treatment you would be asked to fill in the same forms at various points in follow-up as a way to monitor your progress.

As one of the best examples, the Epworth Sleepiness Scale (ESS) is widely used, and some insurance carriers request the ESS as part of their own monitoring programs for sleep apnea patients. Another scale widely in use among centers that specialize in insomnia treatment is the Insomnia Severity Index (ISI), which provides very useful information about both insomnia and sleep quality. Still another common tool is the visual analogue scale (VAS) where the form shows either an unmarked horizontal line with a 0 at the left end and a 10 at the right end, delineating respectively low levels or high levels of a symptom. Another variation of a VAS lists all the numbers between 0 and 10, as when someone might ask of you, “how much pain do you feel on a scale from 0 to 10. VASs are used to measure numerous medical symptoms, and pain symptoms may be the one type that most people are familiar with. 

In sleep medicine, VASs are most commonly used to gauge changes in sleepiness or tiredness. For all the scales mentioned here, like the VAS concept, lower numbers mean you are doing better, and higher numbers mean you are doing worse. So, obviously, for most sleep apnea or insomnia cases, we anticipate seeing higher numbers at intake and then hopefully lower numbers as patients progress through therapy.

At this point, the concepts of using scales or questionnaires probably appear straightforward, and considerable information can be gained by asking patients to complete these surveys at the start of treatment and thereafter. But, there is a possibility of serious flaws in the use of these sleep-related scales, as with any scale, in that we cannot always be exactly sure how accurately we are measuring the thing we are measuring.

Sleepiness in fact is a much more complex symptom than can be measured by the Epworth, which is why so many sleep professionals complain about the obsessive reliance on this scale. The Epworth asks how likely are you to doze off in an assortment of eight behavioral situations (e.g. sitting and reading, watching tv, inactive in a public place, passenger in a car for more than an hour to name some scenarios). Many people are confused by the questionnaire, because they have trouble distinguishing between feeling like they would want to doze off but recognizing that they could not doze off because of other obligations or responsibilities.

Moreover, if these situations do not apply to your personal lifestyle, it becomes too much guesswork to imagine the situation. Or, extraneous factors could influence a patient’s perception of sleepiness in ways unaccounted for by the scale. A classic example of this paradox would be trauma survivors or other anxiety patients, because their agitated state is sufficient to overcome sleepiness in many of the situations described in the ESS. Trauma survivors often look at me funny when I ask inquire as to whether or not they could fall asleep at a movie or church or some other public place with many people in attendance. On the other hand, sitting in front of television set during the daytime in their own home with the doors locked and a guard dog nearby might relax the individual sufficiently to doze off. Scales do not ask about these sorts of circumstances.

A further problem with sleepiness scales (irrespective of their inconsistent results in mental health patients) is some sleep apnea patients simply do not suffer the anticipated degree of sleepiness that would have been expected by the severity of their sleep apnea. For example, there are many patients without any other co-occurring medical conditions where the phenomenon of unexpectedly low sleepiness scores emerges. Some explanations are very obvious for this finding as when a patient drinks 5 to 10 cups of coffee each day. But, other situations seem to offer no clear scientific explanation for the discrepancy. These patients might suffer hundreds of breathing events and simultaneously hundreds of EEG arousals causing severe sleep fragmentation, yet they do not complain of daytime sleepiness to any appreciable degree. Some report no sleepiness. Among this group, you may find some who report a strong measure of tiredness but not sleepiness. Still, you can find some patients who report no sleepiness or tiredness or fatigue, the latter term typically used as a definition of chronic tiredness.

Now, in some of these cases, the ESS scale will be low, because the individual does not experience problems during the behavioral problems listed on the scale. However, these same individuals may report high levels of sleepiness on a VAS, because the question is not about behaviors; the scale simply measures the sensation of sleepiness or drowsiness irrespective of anything else going on. This particular discrepancy is very common in some people whose life style is built on the premise that “idle hands make the devil’s work.” In other words, they are constantly very busy throughout waking hours, and therefore by never letting down, there is no time to experience sleepiness or tiredness. Thus, they may have actual awareness of the feeling of sleepiness or tiredness on the VAS, and yet never doze off on the items listed on the ESS.

So far, in this discussion, it might seem as if these inconsistencies are trivial with little clinical impact. To some extent, this perspective contains some truth, but we often find these scales provide invaluable information when looking for certain subtleties that manifest in conjunction with the discrepancies.

Let’s looks at some case examples to show how to make more effective use of the scales, whether you are a sleep professional or a patient.

One of the more striking examples of misleading numbers arises when the sleep professional is not thinking about possibilities outside the realm of sleep medicine. This problem arises frequently among depressed patients. They may have low scores on insomnia, sleepiness, and tiredness, yet still report nonrestorative sleep (poor sleep quality). When we look at their objective sleep data on a retitration study in the sleep lab, in some cases we might find leg jerks, after which we strongly recommend their treatment. However, in more cases, we find nothing that seems problematic. The airflow signal looks very good, REM sleep is well consolidated, and only mild NREM sleep fragmentation is noted. The missing link turns out to be either the patient suffers from untreated depression that is a possible cause for the nonrestorative sleep, or ironically, the patient’s antidepressants and perhaps other psychotropic medications are lightening the patient’s sleep. This latter scenario creates the obvious clinical conundrum of whether or not to taper the patient off medications to see whether sleep quality improves, or whether the goal should be to add or substitute new medications in the hope of improving the depression.

Where the above scenario gets really tricky is when the patient cannot reliably discern what’s going on to provide the sleep professional with clarifying information. Many people with depression cannot always detail with precision the changes or improvements they have received from using psychotropic medications. This problem arises when the patient’s experience with an antidepressant leads him or her to say something like, “I don’t think I have as much depression as I did before, but I do not know whether or not I am feeling better.” This odd statement arises among depressed individuals who are “numbed” by the psychotropic medication. This numbing effect of psychotropic meds is widespread and is helpful in so far as it helps someone feel “not depressed,” but long-term it may be harmful, if the person continues in this manner without trying to work on other techniques to help with depression, e.g. exercise, diet, and behavioral modification strategies.

In sum, regarding the issue of depression, these kinds of patients are commonly seen in sleep centers, and it takes quite a long time to sort out the possibilities. As you might imagine, most patients and their prescribing physicians or psychologists are reluctant to taper a patient off medication if there is the sense the drugs are effectively treating the depression. The motivation to taper therefore must come from the patient who continues to complain about nonrestorative sleep. What follows is a back and forth between the sleep doctor and the psychiatrist that might drag on for months if not a year or longer, because it is such a tricky and stressful experience for patients to consider tapering off medications that appear to be working. When the medications are not working that well, then the discussion about tapering or changing meds is a lot more practical for the patient to consider.

From the sleep professional perspective, we feel as if our hands are tied. We want the patient to receive better sleep quality, and the patient desires the same; but the prescribing physician, rightly so, is worried about the patient descending to the depths of a bad depression, which was the initiating factor in prescribing the medication at some point in the past. This point by the way could have been months or years, even decades in some patients.

One would like to think that using specific scales designed to measure depression symptoms would make this problem more easily resolvable. But, here again, many patients with depression do not consistently mark depression scales in a definitive fashion to nail down the diagnosis; and, the patients themselves may have started the whole conversation with the complaint of too much fatigue or tiredness, leaving any type of doctor they encounter wondering whether it’s a psychiatric or a sleep condition. Still another variation regarding depression are those patients who present with very high scores on tiredness, yet the tiredness scores never go down no matter how much seemingly effective sleep treatment the patient is receiving.

When patients report these paradoxical results on their questionnaires, it behooves both the patient and the sleep professionals to take a few steps back to garner a fresh look at the nature and severity of the sleep problems. I am not suggesting that the scales should be thrown out when the data look confusing. Rather, it’s when the data look confusing, we should be presuming we are missing something. While it may turn out we are dealing with a level of complexity that will take months or years to resolve, it is still more important to begin formulating theories and related strategies on how to tackle the problems of one of these more complex problems. Just like so many sleep professionals who only know how to prescribe sleeping pills for insomniacs, our field of sleep medicine would be better served by recognizing the need to go beyond the scales for sleepiness, tiredness, insomnia and sleep quality when the numbers do not add up.

Barry Krakow MD


Dr Krakow’s 27 years of sleep research have focused on the complex relationship between physiological and psychological sleep disorders. Dr Krakow currently operates private sleep medical center, Maimonides Sleep Arts & Sciences, Ltd., and serves as Classic SleepCare’s paid Medical Director.

Leave a comment

Comments have to be approved before showing up.