1) a simple, low cost method for screening with adequate predictive power ( + and -) must be available.

In sleep clinics, at least some hallmark symptoms of the major sleep disorders (depending on severity) are successfully elicited by history-taking (e.g., frequency of inability to fall asleep, early morning awakening, daytime sleepiness, loud and disruptive snoring). Symptom checklists are often used to determine if more extensive testing, such as polysomnography, is indicated. It is reasonable that questions of this nature could be posed and interpreted by providers other than sleep specialists. Thus, a short sleep symptom questionnaire may be a valid tool for sleep disorders screening in primary care. There is ongoing research to determine if short symptom questionnaires can be used in the primary care setting as a first step to possible treatment or further evaluation by a sleep specialist.

2 ) the disorder must have a fairly high prevalence: the positive predictive value is a direct function of prevalence; case-finding and screening is most efficient for prevalent conditions or in groups of people at very high risk for less prevalent condition.

The prevalence of sleep disorders is high in the general population, and is likely to be even higher in primary care (1,4,5). Sleep disorders coexist with many other chronic diseases for which help is sought in primary care, thereby increasing the prevalence. Furthermore, underlying sleep disorders may be a driving force for seeking care for other conditions that appear to be the presenting problem ( e.g., headaches).

3) the disorder must pose a threat to life or health.

Sleep apnea syndrome causes daytime morbidity, including disabling sleepiness and accidents and is linked to hypertension and cardiovascular disease (6,7). Insomnia has been linked to depression and other disorders (8).

4) effective treatment must be available. For this criteria to be met, there must be efficacious treatment, the screener (primary care physician) must be prepared to provide care or referral and followup for the condition, and there must be a reasonable level of compliance with the treatment.

Treatment, with varying degrees of success, is available for insomnia, sleep apnea, narcolepsy, and restless legs syndrome. To be effective, a high proportion of those with the disorder must be given the treatment - a crucial step that depends on the provider. The degree to which primary care providers will take responsibility for cases they discover is unknown. Given the minimal medical school training or introduction to sleep disorders, there is clearly a need for education. The growth of sleep laboratories and increase in patients being evaluated for sleep apnea indicates an increase in primary care providers who will refer patients for this condition. In contrast, primary care providers historically have not been very willing to use hypnotics for insomnia therapy. Although compliance is not overwhelmingly high with CPAP, the most commonly used therapy for sleep apnea syndrome, it parallels compliance with antihypertensive medication use. Compliance with hypnotics use in insomniacs does not appear to have been investigated to date.

In summary, screening for sleep disorders in the primary care setting holds promise. For the major sleep disorders, there is a good likelihood that a simple, inexpensive tool can be developed with high positive and negative predictive value (the conditions are prevalent, and symptom questions are regularly used in sleep medicine clinics and in population surveys); untreated sleep disorders do cause significant morbidity, and effective treatment, known to be acceptable to patients, is available. The willingness of primary care providers to take responsibility for treatment or referral and followup, however, is unknown.

1. Sleep-Disordered Breathing.

Previously reported findings from this study of employed, middle-aged adults have established a prevalence of sleep-disordered breathing (SDB) (defined as at least 15 apneas and hypopneas per hour of sleep)of 4% in women and 9% in men. Cross-sectional analysis have established a statistically significant association of SDB and several outcomes.

a) SDB is associated with elevated blood pressure, independent of age, sex, body mass index, and other potential confounding factors (9). If the association is causal, SDB could result in hypertension in approximately 2.4 million men and women in the USA, in their most productive working years.

b) SDB is associated with quality of life (10). Using the SF-36 as the measure of quality of life, SDB at a severity level of 15 apneas and hypopneas per hour of sleep (AHI =15) was associated with decrements in SF-36 constructs equivalent to those of other chronic conditions.

c) SDB is associated with an increased risk of motor vehicle accidents (11). People with SDB (AHI =15) were 7 times as likely to have multiple crashes, indicated by state motor vehicle records, as were people without SDB.

d) SDB is associated with a significant decrease in "psychomotor function", measured by a factor from a 1-hour battery of common neuropsychologic tests (12)

2. Insomnia

A sample of participants in the Wisconsin Sleep Cohort Study, ages 30-65, who had at least one primary care visit in a 2-year period was used to estimate the prevalence of insomnia in HMO- primary care patient populations and investigate health outcomes of insomnia

a). Occurrence of insomnia in a sample of primary care users
Insomnia was defined as the self-reported "almost always" ( from a scale of 0-5, category 5) occurrence of 1 or more of the following:

• difficulty getting to sleep
• waking up during the night and difficult to go back to sleep
• wake up repeatedly in the night
• wake up too early in the AM and cannot get back to sleep.

  Prevalence of Habitual Insomnia, %
  

  age, yrs	Women	Men
  30-39	11.5	12.3
  40-49	14.4	10.7
  50-65	18.3	11.2

The occurrence of insomnia was also investigated longitudinally. 6.5% of women and 5.7% of men reported insomnia (defined as above) at baseline and 4 years later at followup. b). Association of habitual insomnia and quality of life measured by the SF36; mean construct scores below are adjusted for age, sex, and body mass index.

Mean SF-36 score ( scores range from 0-100 higher score is higher quality of life)

SF-36 construct	Habitual Insomnia	Not habitual insomnia (all p<.001)
Physical functioning	81.7	87.5
Role performance - physical	71.2	82.7
general health perception	64.5	74.0
vitality	49.9	60.1
pain	63.6	73.8
social functioning	72.5	85.1

(1) Cross-sectional data:
Zung score, adjusted mean ( age, sex, BMI)

Insomnia	44.4
No Insomnia	40.4 p<.0001

(2) Longitudinal data

Development of new depression (measured by Zung self-rating scale>=60 or use of antidepressants) over a 4-year period in people with habitual insomnia versus those without habitual insomnia was estimated using logistic regression.

Among participants free of depression at baseline, those with habitual insomnia at baseline were 2.5 times more likely ( p=.004) to develop depression at 4-year followup. The association is independent of age, sex, and BMI.

• 13% ( 139/1064) had ever gone to a doctor for a sleep problem
• most sought help from primary care:

  Type of provider help sought from:

  family physician, general
  practitioner-- 50.5%
  nurse practitioner-- 0.7%
  sleep specialist-- 20.9%
  psychiatrist or psychologist-- 10.8%
  ENT or pulmonologist-- 10.1%
  other specialist-- 7.0%
  

  Of those who sought help for a sleep disorder,

  48.2 % (67/139) were told they had a sleep disorder;
  29.5% (41/139) were told they did not have a sleep disorder
  22.3 % (31/139) reported getting no help at all

• Of those with a sleep disorder diagnosed, most were told they had sleep apnea

  Sleep disorder diagnosed

  sleep apnea-- 65.6%
  insomnia-- 13.4%
  chronic fatigue synd, fibromy-- 6.0%
  restless legs-- 3.5%
  narcolepsy-- 3.0%
  other-- 7.5%
  

The help-seeking behavior in this sample indicates that providers cannot rely alone on having their patients with sleep problems actually report them during a visit. For example, in the case of insomnia, only a very small proportion of those who reported having habitual insomnia (on our questionnaires) actually sought help for insomnia from a care provider. The prevalence of insomnia, defined as having difficulty getting to sleep or maintaining sleep always or almost always by self-report in this sample is 12.5%, while the prevalence of diagnosed (or doctor-declared) insomnia is 0.8%. Thus, even if primary care providers were sufficiently responsive to their patients' sleep complaints, relying on unsolicited sleep problems from the patients will only identify a small number of those with sleep problems. Consequently, a screening tool as well as increased responsiveness to sleep problems on the part of primary care providers would be needed to reduce the high proportion of undiagnosed sleep disorders in adults.