Two scales - the Ullanlinna Narcolepsy scale developed by a Finnish team (Hublin et al., 1994) and the Stanford Sleep Inventory which was validated at the Stanford Sleep Disorders Clinic - are currently in use:

• The Ullanlinna Narcolepsy scale is a paper-pencil questionnaire, designed to be used with the general population.
• The Stanford Sleep Inventory is also a paper-pencil, self-reporting questionnaire that explores symptoms that constitute narcolepsy, as well as other sleep disorders.

There is also the Sleep-EVAL Expert System. It is a computerized diagnostic tool based on the collection of sleep and psychiatric symptomatology. The diagnostic procedure is continually refined during the interview allowing for the adjustment of the decision-making process by the Sleep-EVAL system. This encourages the detection of rare syndromes and their manifestations in the general population.

METHODS

SUBJECTS

N = 284 subjects.

A sample of 96 narcoleptic subjects were recruited by the Center for Narcolepsy Research at Stanford University. Each narcoleptic patient was asked to provide the name and phone numbers of his/her mother, father, brother(s) and/or sister(s). Eighty-five people, either living with a narcoleptic (spouse or roommate) or a close friend, constituting the control group, were also interviewed.

Narcoleptic patients, their family members and control subjects were interviewed by telephone using the Sleep-EVAL system. Almost all contacted individuals (96%) accepted to participate in the study. Information collected by the system included a complete description of narcolepsy symptoms:

• daytime sleepiness
• cataplexy
• hypnagogic and hypnopompic hallucinations
• sleep paralysis
• daytime functioning
• sleep habits
• sleep and mental disorders diagnoses according to the DSM-IV and ICSD classifications

The results of the Stanford Sleep Inventory (SSI) were available for the 284 subjects:

• 82 narcoleptics
• 62 fathers
• 65 mothers
• 41 sisters
• 12 brothers
• 22 controls

RESULTS

The narcoleptic group was 62.2% women. The mean age was 30.4 (± 10.1) years. Most of them were single (56.1%).

The control group was 63.6% women. The mean age for this group was 39.2 (± 12.8) years. Most of them were married (68.1%).

Control subjects were significantly older than narcoleptics (p < 0.05) but the age discrepancy was not significant with the narcoleptics' sibling.

Table 1. Demographic characteristics of the sample

	Proband	Father	Mother	Sister	Brother	Control
	% (n=82)	% (n=62)	% (n=65)	% (n=41)	% (n=12)	% (n=22)
Gender
Male	37.8	100			100	36.4
Female	62.2		100	100		63.6
Age (yr) s.d.	30.4 (±10.1)	58.4 (±11.1)	56.5 (±11.7)	33.1 (±10.6)	37.9 (±13.2)	39.2 (±12.8)
Marital st.
Single	56.1	-	4.6	34.1	25.0	27.3
Married	36.6	87.1	81.6	60.9	66.7	68.1
Divorced	7.3	11.3	12.3	4.9	8.3	4.5
Widowed	-	1.6	1.5	-	-	-

COMPARISONS SLEEP-EVAL VS. STANFORD SLEEP INVENTORY

For about a third of the sample, the Sleep-EVAL interview was administered before the Stanford Sleep Inventory (SSI).

In 88.3% of cases both questionnaires were administered within a six-month period.

Sleep-EVAL administered before SSI		Same month	SSI administered before Sleep-EVAL
n=9	n=85	n=7	n=159	n=11	n=15
7-12 months	1-6 months	< 1 month	1-6 months	7-12 months	> 12 months

CATAPLEXY

The information was missing on the SSI for 14 subjects. 37.8% of subjects were identified with cataplexy on the SSI. 31.1% of subjects reported cataplexy during the Sleep-EVAL interview.

Regardless of the elapsed time between the administration of the 2 questionnaires, the sensitivity and specificity of Sleep EVAL's cataplexy questions were very good and the correlation between the 2 questionnaires was positive and significant.

		SSI		Sensitivity	Specificity	r
		-	+
Sleep-EVAL	-	161	7	75.5%	95.8%	.75
	+	25	77

When the delay between the two questionnaires is about 6 months or less:

• The specificity of Sleep EVAL's cataplexy questions did not improve (95.7%) but the sensitivity lightly increased (77.9%)
• the correlation between the cataplexy questions of the 2 questionnaires increased (Spearman Corr. r =0.77).

SLEEP PARALYSIS

According to the SSI, 27.2% of the subjects reported to ever have experienced sleep paralysis. During the Sleep-EVAL interview, 33.6% of subjects reported experiencing sleep paralysis.

Specificity and sensitivity of Sleep-EVAL on these questions were good and the correlation between the two instruments was positive and high.

		SSI		Sensitivity	Specificity	r
		-	+
Sleep-EVAL	-	180	6	92.7%	89.6%	.79
	+	21	76

When the delay between the administration of the two questionnaires was about 6 months or less:

• The specificity of Sleep EVAL's sleep paralysis questions remained unchanged (95.7%) but the sensitivity increased (89.6%).
• The correlation between the sleep paralysis questions of the 2 questionnaires is about the same (Spearman Corr. r = 0.80).

HYPNAGOGIC AND/OR HYPNOPOMPIC HALLUCINATIONS

Using the SSI, 33.2% of subjects reported hypnagogic and/or hypnopompic hallucinations. Using Sleep-EVAL, 25.4% of subjects reported having these hallucinations at least 2 times per month.

The specificity of Sleep-EVAL on this topic was very high (92.6%) but the sensitivity was low although acceptable.

		SSI		Sensitivity	Specificity	r
		-	+
Sleep-EVAL	-	175	36	61.7%	92.6%	.59
	+	14	58

When the delay between the administration of the two questionnaires was about 6 months or less:

• The specificity of Sleep EVAL's hypnagogic/hypnopompic questions is higher (93.6%) and the sensitivity increased (63.3%).
• The correlation between hypnagogic/hypnopompic questions of the 2 questionnaires increased (Spearman Corr. 0.62).

AUTOMATIC BEHAVIORS

The information was not available for 37 subjects on the SSI.

Using the SSI, 35.2% of subjects reported their mind was going blank while driving or at work (or other similar situations, e.g. working on computer, studying). Using Sleep-EVAL, 37.7% of subjects reported having these episodes at least 1 time per month.

Specificity and sensitivity of Sleep-EVAL on these questions were good and the correlation between the two instruments was positive and acceptable.

		SSI		Sensitivity	Specificity	r
		-	+
Sleep-EVAL	-	135	25	73.1%	87.7%	.62
	+	19	68

The specificity and sensitivity changed by less than 1% when the delay between the administration of the two questionnaires was about 6 months or less. The correlation remained the same.

CORRELATION WITH NARCOLEPSY DIAGNOSIS

For both instruments, cataplexy and sleep paralysis exhibited the highest correlation with the diagnosis of Narcolepsy.

On sleep paralysis and automatic behaviors: correlation was higher between Sleep-EVAL and Narcolepsy diagnosis than between SSI and narcolepsy diagnosis. On cataplexy, correlation was higher between SSI and Narcolepsy diagnosis than between Sleep-EVAL and narcolepsy diagnosis. Many(NEED %) subjects initiated medication treatment for cataplexy during the time from completion of the SSI and Sleep-EVAL interview. Thus, cataplexy may have remitted when they were interviewed with Sleep-EVAL.

Correlation between symptoms and narcolepsy diagnosis
All = Regardless the elapsed time between the administration of the 2 questionnaires
≤ 6 ms = Delay of 6 months or less between the administration of the 2 questionnaires

	SSI		Sleep-EVAL
	All	≤ 6 ms	All	≤ 6 ms
Cataplexy	.89	.90	.79	.81
Sleep paralysis	.75	.75	.78	.83
Hallucinations	.68	.68	.64	.69
Automatic behaviors	.66	.67	.72	.75

CONCLUSIONS

Narcolepsy questions from the Sleep-EVAL system are comparable to those in the SSI.

Agreement on narcolepsy symptoms between the two instruments is auspicious even when the elapsed time between the two administrations exceed 6 months. The Stanford Sleep Inventory is an effective tool to identify narcoleptic individuals. However, the lack of guidelines necessary to interpret it and the limitation in the choice of answers (many items are only yes/no answers) pose a limit to its use by non-sleep specialists.

On the other hand, the use of a computerized diagnostic tool such as Sleep-EVAL allows overcoming of these limits: non-sleep specialists can use it. It is able to achieve a diagnosis using fuzzy answers (e.g., always, often, sometimes, rarely, never) with results superior to those obtained with a paper-pencil questionnaire.

REFERENCES

1. Hublin C, Kaprio J, Partinen M, Koskenvuo M, Heikkila K. The Ullanlinna Narcolepsy Scale: validation of a measure of symptoms in the narcoleptic syndrome. J Sleep Res 1994;3(1):52-59
2. Anic-Labat S, Guilleminault C, Kraemer HC, Meehan J, Arrigoni J, Mignot E. Validation of a cataplexy questionnaire in 983 sleep-disorders patients.Sleep 1999; 22:77-87
3. Ohayon MM. Improving decision making processes with the fuzzy logic approach in the epidemiology of sleep disorders. J Psychosom Res 1999;47:297-311.