Epidemiological Study on Insomnia in the General Population

Sleep. 19(3):S7-S15 © 1996 American Sleep Disorders Association and Sleep Research Society Epidemiological Study on Insomnia in the General Population M.Ohayon Summary: This study was conducted with a representative sample of the French population of 5,622 subjects of 15 years old or more. The telephone interviews were performed with EVAL, an expert system specialized for the evaluation of sleep disorders. From this sample, 20.1% of persons said that they were unsatisfied with their sleep or taking medication for sleeping difficulties or anxiety with sleeping difficulties (UQS). A low family income, being a woman, being over 65 years of age, being retired and being separated, divorced or widowed are significantly associated with the presence of UQS. A sleep onset period over 15 minutes, a short night's sleep and regular nighttime awakenings are also associated with UQS. Medical consultations during the past 6 months and physical illnesses are more frequent among the UQS group. The consumption of sleep-enhancing medication and medication to reduce anxiety is important: in the past, 16% of subjects had taken a sleep-enhancing medication and 16.2% a medication to reduce anxiety. At the time of the survey 9.9% of the population were using sleep-enhancing medication and 6.7% were using a medication for anxiety. For most, hypnotic consumption was long-term: 81.6% had been using it for more than 6 months. Key words: Epidemiology-Hypnotics-Anxiolytics-Expert system. The frequency of sleep disorders, particularly insomnia, in the general population has been investigated by many authors (1-7). It is certain, however, that many differences emerge in the development of questionnaires and methods used to provide an appreciation of the presence and repercussions of these sleep disturbances. In fact, the prevalence of insomnia in the general population compels us to study this disorder because of its socioeconomic and therapeutic impacts. This is why a number of authors have insisted on the importance of determining their patients' hypnotic and anxiolytic consumption (1,4). It is all the more important because: 1) elderly subjects are the biggest consumers of this type of medication; 2) prescriptions of hypnotics are usually long-term; 3) consequences such as dependence, cognitive and memory disturbances and problems of vigilance are poorly documented in the general population; 4) hypnotic use instructions in different medical categories fluctuate and the question of a better understanding of sleep disorders remains unanswered. Finally, it is important to note that longitudinal studies about insomnia are rare even though a number of studies underscore the frequent associa- tion of this disorder to an organic or psychiatric pathology (3,5-7,9,10). In fact, most epidemiological studies examine the problem of insomnia either in the framework ofa much larger epidemiological survey that does not devote many questions to sleep disorders (1-3,5), or by focusing on a particular approach to insomnia (4,6,7) while encountering problems of data dissipation in the general population. This dissipation then leads to a combinatory explosion of cases that must be contemplated to construct a questionnaire for the study. For these reasons, we resorted to the use of an expert system that administers an epidemiological questionnaire and in which the unfolding of the interview and organization of questions is based upon the machine-subject interaction. The number of gathered variables thus becomes important (over 1,146). Furthermore, an advantage of this system is that it eventually allows the researcher to make a diagnosis as a function of the many reference classifications. In this paper, we will present only the general results of our study. The identification of the different sleep disorder diagnoses will be reviewed in a later publication. Accepted for publication December 1994. Address correspondence and reprint requests to Dr. M. Ohayon, Centre de Recherche de l'Institut Philippe Pinel de Montreal, 10905 Est, boulevard Henri-Bourassa, Montreal, Quebec HlC lHl, Canada. METHODS The present study was conducted on the whole French metropolitan territory between April 22, 1993 and July S7 Downloaded from https://academic.oup.com/sleep/article-abstract/19/suppl_3/S7/2749899 by guest on 20 May 2020 Centre de Recherche Philippe Pinel de Montreal, Canada S8 M.OHAYON I l Sleep, Vol. 19, No.3. 1996 was the duration of consumption?; 9) treatment associated with sleep-enhancing medication and anxietyreducing medication, the name of the medication and instructions; 10) consumption of drugs, alcohol and tobacco. Following these questions, the system, based on expert knowledge of psychiatric classifications (DSM-II1R, DSM-IV, ICD-I0), asks a series of complementary questions leading to a diagnosis of sleep disorder and eventually to a possible associated psychiatric disorder. In total, there are 206 Boolean-type questions, 110 of a numeric or alphanumeric type, 153 "yes-no-unknown" -type questions, 80 of Likert type and 103 multiple choice questions. The system can deduce 1,146 variables in which 78 different diagnoses of sleep disorders can be found. The population of the study was divided into two distinct groups: 1) subjects satisfied with their sleep AND* not taking any medication for their anxiolytic or hypnotic effects (SQS). 2) subjects unsatisfied with their sleep OR* taking medication for sleeping difficulties or anxiety with sleeping difficulties (UQS). This group is made up of three subgroups: 1) subjects satisfied with their sleep AND* taking a sleep-enhancing medication; 2) subjects unsatisfied with their sleep AND* not taking any medication (UQSo); 3) subjects unsatisfied with their sleep AND* taking a sleep-enhancing medication or a medication to reduce anxiety (UQSa). Statistics The data were weighted to neutralize undersampling of men and young persons and compensate disparities between regions. All results are presented with weighted data and 95% confidence intervals (CI). Logistic regressions were calculated with the presence of UQS (dissatisfaction with sleep or taking sleep-enhancing medication) as the dependent variable. To determine which categories of the independent variables were significantly associated with the presence of UQS, the method of INDICATOR contrasts was used. For each variable, the reference category is that less contributes to the presence of UQS. The significant Odds ratios are interpreted according to these reference categories. To perform these multivariate analyses we verified first if the relation between the presence of UQS and a sociodemographic variable could be explained by the relations that maintained this variable with all other sociodemographic variables. Second, a similar analysis was performed in using only sleeping habits and health variables as independent variables. Finally, sociode- * The conjunction is used as a logical operator. Downloaded from https://academic.oup.com/sleep/article-abstract/19/suppl_3/S7/2749899 by guest on 20 May 2020 15, 1993. A representative sample of 5,622 subjects of 15 years of age or more was obtained following the solicitation of 6,966 persons (acceptance rate: 80.7%). The geographic distribution was determined using the method of quotas based on demographic data provided by the Institut National de Statistiques et Evaluation Economique (INSEE). During the telephone contact, the Kish selection method (11) was used to identify which person in the home was to be interviewed. Persons that had refused to participate in the study on the first phone call were subsequently contacted twice. A minimum of six calls with no answer was made before abandoning the phone number; these calls were made at different times of the day and week. Excluded from the study were those persons who did not speak French, who suffered from a hearing problem, who had speech impediments and those who were too sick to be interviewed. Telephone interviews were conducted by professional investigators using the expert system EVAL built around the inference engine Adinfer (©Ohayon, 1983-1993). Adinfer is a level 2 nonmonotonic expert system endowed with a causal reasoning mode. There are many advantages of using such a system for this type of interview: 1) minimal training time is required (approximately 3 hours); 2) the questionnaire is adjusted to the particular case being interviewed; 3) missing data are practically nonexistent; 4) no data entry needs to be carried out following the interview. The system formulates its own questions. The interviewer thus simply has to read them to the subject as they appear on the screen. The system also decides which questions are to be asked to the particular case being interviewed. Its knowledge base allows the sequencing of a series of questions concerning: 1) sociodemographic information; 2) medical consultations during the past 6 months, including treated illnesses; 3) psychiatric and medical history; 4) quantity and quality of sleep, bedtime, time at which the subject falls asleep, time of awakening, length of the night, nighttime awakenings and their frequency, daytime sleepiness, number of daily naps, satisfaction with sleep; 5) the repercussions of sleep disorders assessed on a IS-item scale providing an evaluation of the night, daytime functioning, cognitive effects and affective tone; 6) manifestations reported by persons close to the subject, bruxism, snoring, nocturnal myoclonus, agitated sleep, respiratory pauses; 7) time at which the subject hopes to fall asleep and awaken; 8) present or past use of sleep-enhancing medication and anxiety-reducing medication (prescribed or not), the name of the product, dosage, number of tablets, the specialist who prescribed the medication (if applicable) and duration of consumption. When it is a treatment that has ended, for how long has the treatment been stopped and what EPIDEMIOLOGICAL STUDY ON INSOMNIA S9 TABLE 1. Prevalence of UQS subjects according to marital status, educational level, monthly family incomes and employment status Women n % Marital status Single Married or living with someone Divorced-separated-widowed 671 1,607 651 17.6 21.1 39.4 a 66 834 875 520 333 201 98 618 543 491 256 174 162 107 431 Men n % ±2.88 ± 1.99 ±3.75 758 1,722 214 12.1 15.2 30.4a ±2.32 ±1.70 ±6.16 47.2a 33.9 a 21.6 21.1 15.6 19.4 10.2 ±12.04 ±3.21 ±2.73 ±3.51 ±3.90 ±5.47 ±5.99 30 550 985 482 301 145 200 23.8 24.8 a 13.4 13.0 11.3 7.8 17.1 ± 15.24 ±3.61 ±2.13 ±3.00 ±3.58 ±4.37 ±5.22 30.6 a 26.5 19.6 14.7 17.9 18.8 24.4 18.3 ±3.63 ±3.71 ±3.51 ±4.34 ±5.70 ±6.02 ±8.14 ±3.65 361 472 416 346 204 233 197 375 14.5 18.7 15.1 12.4 13.0 15.7 12.9 12.1 ±3.63 ±3.52 ±3.44 ±3.47 ±4.62 ±4.67 ±4.68 ±3.30 28.4 a 1,657 1,271 19.1 a p < 0.05 with categories in boldface type in the same column. ±2.17 ±2.16 1,161 1,535 18.7a 13.2 ±2.24 ±1.69 Educational level None CEP CAP-BEPC BACC BACC + 2 BACC + 4 BACC + 5 and more Monthly family incomes ;::6,000 francs 6,000-9,000 francs 9,000-12,000 francs 12,000-15,000 francs 15,000-18,000 francs 18,000-24,000 francs ,.;24,000 francs Refusal Employment status Without work Working mographic, sleeping habits and health variables were put together in the same analysis to verify in what the effect of sociodemographic variables was modified when the effect of sleeping habits and health variables was taken into account. RESULTS Subjects were between the ages of 15 and 96 years old. The weighted sample contained 52.1 % women and 47.9% men. A dissatisfaction with sleep or taking medication for sleeping difficulties or anxiety with sleeping difficulties was found in 20.1 % of subjects (UQS group). In this group, women were significantly more numerous than men (24.4%; CI 95% ± 1.6% vs. 15.6%; CI 95% ± 1.3%). Subjects satisfied with their sleep and taking medication for sleeping difficulties or anxiety with sleeping difficulties represent 8.2% of the population (subgroup SQSa); those who were unsatisfied with their sleep but took no medication (subgroup UQSo) 7.3% and those who were both unsatisfied with their sleep and taking medication (subgroup UQSa) represent 4.6% of the population. For 1.1 % of subjects, the sleep disturbances had disappeared at the time of the interview due to their medication for sleeping difficulties. CI95% CI95% Demographic characteristics of the population Among persons under 45 years of age we find the least number of UQS subjects. For women, the two highest age group categories (over 54 years old) have significantly greater rates of UQS subjects than those found in younger subjects (less than 45 years old). For men, only rates observed in subjects over 65 years old differ significantly from those of the other age groups (Fig. 1). Marital status was divided into three groups: 1) single; 2) married or living with someone; 3) separated, divorced or widowed. A significantly higher rate of UQS subjects was found in the third group. UQS women represent 39.4% ofthe group, which is significantly higher than other groups of women. This rate is also higher than that observed in men of the same group (30.4%). For men alike, the highest rate of UQS subjects was found in the separated, divorced or widowed group (Table 1). Women with little schooling (none or CEP, i.e., less than 7 years) are found more frequently in the UQS group compared to women of other educational levels. For men, only those at the CEP level are significantly more numerous in the UQS group compared to men with more schooling, with the exception at the highest levels of education (Table 1). The rates ofUQS subjects Sleep. Vol. 19. No.3. 1996 Downloaded from https://academic.oup.com/sleep/article-abstract/19/suppl_3/S7/2749899 by guest on 20 May 2020 Variables -----, M.OHAYON S10 TABLE 2. Results of the logistic regression for the whole sample regarding the effect of sociodemographic characteristics, sleep habits and health variables on the presence of UQS ~.-, 50 D E:I Women Men Coefficient 40 Variables t'! ~ "l 30 N Q\ aO - aO 10 15-24 25-34 35-44 45-54 55-64 <! 65 Age Groups FIG. 1. Prevalence ofUQS according to age groups. are similar across categories of monthly family income for men. Only women whose income is less than 6,000 francs per month (30.6%) often demonstrate UQS compared to women with monthly incomes of 9,000 to 24,000 francs (Table 1). UQS subjects are also frequently found in the unemployed group. More specifically, retired women seem most affected by sleep disturbances (35.8%). As for men, it is also the retired (23.2%) or those without work for unknown reasons (44.6%)-other than the unemployed (11.8%), students (8.7%) or those in military service (18.9%)-who are most affected. Among people who work, the rates are similar between different categories of professions. Working schedules (daytime, evening, nighttime or variable schedule) were also examined; the rates ofUQS subjects are comparable between these different categories. For the whole sample, results of multivariate analyses using only sociodemographic variables show that low family income (p < 0.05), being a woman (p < 0.0001), being over 65 years of age (p < 0.001) and being separated, divorced or widowed (p < 0.000 1) are significantly associated with the presence of UQS. For women, a low educational level (p < 0.01), a low family income (p < 0.05) and being separated, divorced or widowed (p < 0.000 1) are significantly associated with UQS. As for men, it is being retired (p < 0.01), over 55 years of age (p < 0.02) and separated, divorced or widowed (p < 0.01) that is significantly associated with UQS. When sleeping habits and health variables are added to the model, we found for the whole sample that being a woman (p < 0.0001), being separated, divorced or widowed (p < 0.0001), having little schooling (p < Sleep, Vol. 19, No.3, 1996 OR CI95% Sex 0.2961 a 1.34 [1.16; 1.57] Woman Marital status 0.5334b 1.70 [1.40; 2.07] Divorced-separated-widowed Educational level 0.2523< 1.29 [1.06; 1.56] Low Length of sleep onset period 0.7617 b 2.14 [1.82; 2.52] > 15 minutes Duration of the night 0.539Jb 1.71 [1.44; 2.04] <7 hours Nighttime awakenings 1.263 Jb 3.53 [3.05; 4.11] Regular Medical consultation 0.927Jb 2.53 [2.08; 3.06] Presence Physical illness 0.3661b 1.44 [1.21; 1.71] Presence Reference categories: man; married or living with someone; high; :::; 15 minutes; 7-8 hours; no nighttime awakenings, medical consultations or physical illnesses. a p < 0.001. b P < 0.0001. 'p < 0.01. 0.01), having consulted a physician within the last 6-month period (p < 0.0001), suffering from a physical illness (p < 0.0001), having a sleep onset period greater than 15 minutes (p < 0.0001), sleeping less than 7 hours per night (p < 0.0001) and to waking up regularly at night (p < 0.0001) are significantly associated with the presence ofUQS. These later results show that after taking into account sleeping habits and health variables, the effect of family income and age groups disappeared; observed relations were explained by sleeping habits and health variables. The schooling variable is marginally significant because the inferior limit of the odds ratio is close to 1.00 (Table 2). For women, being separated, divorced or widowed (p < 0.0001), having consulted a physician within the last 6-month period (p < 0.0001), suffering from a physical illness (p < 0.00 1), having a sleep onset period greater than 15 minutes (p < 0.0001), sleeping less than 7 hours per night (p < 0.0001) and to wake up regularly at night (p < 0.0001) are significantly associated with UQS. For men, being separated, divorced or widowed (p < 0.0001), having consulted a physician within the last 6-month period (p < 0.0001), suffering from a physical illness (p < 0.05), having a sleep onset period greater than 15 minutes (p < 0.0001), sleeping less than 7 hours per night (p < 0.05), to wake up regularly at night (p < 0.0001) and to wake up early in the morning (p < 0.05) are significantly associated with UQS. Downloaded from https://academic.oup.com/sleep/article-abstract/19/suppl_3/S7/2749899 by guest on 20 May 2020 20 - - (3 Odds ratio ------------------------------------------------------------------------- ---- Sll EPIDEMIOLOGICAL STUDY ON INSOMNIA 40T-______________________________________________________________-, o o r:a SI.Dif. W AnxietyW SI. Dif. and Anx. W [1 SI. Dif. M [3 AnxietyM • SI. Dif. and Anx. M 30 4.4 25-34 35-44 45-54 55-64 ~65 Age Groups FIG. 2. Consumption of sleep-enhancing and anxiety-reducing medication according to age groups. Consumption of sleep-enhancing or anxiety-reducing medication Consumption of medication is important for the whole sample: 1) 16% of subjects had already taken a medication for sleeping difficulties, and 16.2% had taken a medication for anxiety but had stopped that consumption at the time of the survey. 2) At the time of the survey, 9.9% of the population were using a sleep-enhancing medication and 6.7% were using a medication for anxiety; 21.2% used both. Twice as many women than men used medication for sleeping difficulties (12.7% vs. 6.8%) or medication for anxiety (8.7% vs. 4.5%). For most, consumption of sleep-enhancing medication is chronic; 30.9% have been using it for more than 5 years, 41.7% between 1 and 5 years and 9.0% between 6 months and 1 year. For 27% of the cases, consumption of medication for anxiety had occurred for less than 6 months, 30.4% had been using it for more than 5 years, 32.8% between 1 and 5 years and 9.8% between 6 months and 1 year. Consumption of sleep-enhancing medication significantly increases in subjects 45 years of age and older, particularly those over 65 years old. This is true for both men and women. Consumption of medication for anxiety is significantly higher in women over 55 years of age compared to those between 15 and 34 years old (p < 0.05). For men, rates are comparable between age groups (see Fig. 2). The double consumption of medication for sleeping difficulties and for anxiety was more prevalent in women than in men (4.1% vs. 1.5%); the highest prevalence was in the 55-64-year age group. Sleep, Vol. 19, No.3, 1996 Downloaded from https://academic.oup.com/sleep/article-abstract/19/suppl_3/S7/2749899 by guest on 20 May 2020 10 M.OHAYON S12 100 60 *QIO QIO 0 lSJ *t-- = 50 QIO 80 WOMEN MEN N ~ QIO =' IC = 40 Q\ 60 r...: N ~ ~ ~ ~ E-o ~ 40 30 N = N 20 Q\ !'"i .... No illness Women D UQS group ~ Men Illness FIG. 4. Physical illness. SQS group * p < 0.0001, compared with those SQS FIG. 3. Utilization of health services in the past 6 months. Medical consultations in the past 6 months Medical consultations during the preceding 6 months were prevalent for both men and women. However, for both men and women, significantly more UQS than SQS subjects consulted a physician (Fig. 3). Most of them had consulted a general practitioner; this included 91.2% of women and 88.8% of men from the UQS group and 85.7% of women and 87.9% of men from the SQS group. A small number of persons had consulted a psychiatrist; there were less than 1% in the SQS group and 5.6% of women and 8% of men from the UQS group. Nearly half had consulted another specialist; this included 47.4% and 43.4% of UQS women and men, respectively, and 44.8% and 37.6% ofSQS women and men, respectively. Only the fact of having consulted a psychiatrist was associated with being in the UQS group, both for men and women (p < 0.001). being treated. Thus, 40.2% (CI 95% ± 3.88%) of women who had a physical illness were also UQS subjects. This rate is also significantly higher than that observed in men (27.9%; CI 95% ± 4.05%). The observed prevalence rate of UQS men among those suffering from a physical illness is significant compared to those with no physical illness (Fig. 4). Sleeping habits The rates of UQS subjects vary little according to bedtime. Only women going to bed very early (before 9:00 p.m.) are found more often to be UQS than women whose bedtime is between 10:00 p.m. and 11 :00 p.m. For men, the prevalence rates of UQS subjects are stable across different categories of bedtime (Table 3). A sleep onset period over 15 minutes is associated with a significantly higher UQS prevalence, both for men and women. This rate significantly increases as a function of the duration of the sleep onset period: the longer the time needed to fall asleep, the higher the prevalence of UQS subjects (Table 3). A short night's sleep is also associated with a significantly higher prevalence ofUQS subjects. A clearly higher proportion of UQS women was found among those sleeping less than six hours per night than among Treated physical illness women sleeping over 6 hours per night. For men, only At the time of the survey, a significantly higher prev- those sleeping less than 5 hours seemed to be found alence of UQS was found among subjects suffering UQS significantly more often compared to those sleepfrom a physical illness for which they were presently ing over 5 hours per night (Table 3). Sleep, Vol. 19, No.3, 1996 Downloaded from https://academic.oup.com/sleep/article-abstract/19/suppl_3/S7/2749899 by guest on 20 May 2020 ~ S13 EPIDEMIOLOGICAL STUDY ON INSOMNIA Regarding time of awakening, significantly more subjects who wake up early, before 5:00 a.m., are found to be UQS (Table 3). of those with at least three nighttime awakenings can be found to be UQS. Daytime sleepiness Nighttime awakenings The prevalence ofUQS subjects is significantly higher among those who regularly sleep during the day compared to those who never or rarely have daytime naps. Among women who regularly sleep during the day, 35.1 % (el 95%: ± 5.64) are also UQS. That is the case for 26.3% eCI 95%: ±3.36) of those who occasionally sleep and 22.3% (el 95%: ± 1.83) of those who never sleep during the day. As for the men, the same pattern prevails with 25.2% (el 95%: ±4.31), 16.2% (el 95%: ±2.69) and 12.9% (el 95%: ± 1.65) of those who regularly, occasionally or do not sleep at all during the day, respectively, found to be UQS men. Among those who regularly sleep during the day, 95.8% of women and 97.7% of men only have one nap during the day. DISCUSSION AND CONCLUSION The common criteria for the different definitions suggested for insomnia include difficulty in initiating or maintaining sleep or early morning awakenings (1,6,12-15). Daily fatigue is also a common notion TABLE 3. Prevalence of UQS subjects according to bedtime, length of sleep onset period, duration of the night and time of awakening Women Variables n Bedtime 226 s21:00 640 21:01-22:00 1,523 22:01-23:00 409 23:01-24:00 124 >00:00 Length of sleep onset period 1,193 s15 minutes 896 16-30 minutes 519 31-60 minutes 308 >60 minutes Duration of the night 95 s5 hours 245 5-6 hours 565 6-7 hours 1,056 7-8 hours 702 8-9 hours 254 >9 hours % Men CI95% n % CI95% 34.125.1 22.6 23.9 27.9 ±6.18 ±3.36 ±2.1O ±4.13 ±7.89 180 483 1,372 463 192 20.8 16.2 15.1 15.4 12.0 ±5.93 ±3.29 ± 1.89 ±3.29 ±4.60 13.3 23.034.654.9- ±1.93 ±2.76 ±4.09 ±5.56 1,288 856 371 171 9.2 16.522.642.5- ±1.58 ±2.49 ±4.26 ±7.41 52.143.3" 28.4 19.2 19.4 22.8 ±10.05 ±6.20 ±3.72 ±2.38 ±2.93 ±5.16 132 227 662 959 517 191 38.623.9 14.5 12.5 13.6 13.1 ±8.31 ±5.55 ±2.68 ±2.09 ±2.95 ±4.79 ±7.12 ±3.58 ±2.24 ±3.34 ±5.70 219 632 1,165 439 236 22.319.2 12.0 17.0 14.2 ±5.51 ±3.07 ±1.87 ±3.51 ±4.45 Time of awakening 180 38.7" s5:00 567 25.3 5:01-6:00 21.7 1,300 6:01-7:00 640 24.6 7:01-8:00 231 26.6 >8:00 "p < 0.05 with categories in boldface type in the same column. Sleep. Vol. 19. No.3. 1996 Downloaded from https://academic.oup.com/sleep/article-abstract/19/suppl_3/S7/2749899 by guest on 20 May 2020 In total, 30% of subjects reported awakening regularly all nights. Among these persons, the UQS subject rate is significantly higher than the two other categories (none or sometimes). For women and men, a significant increase in the number ofUQS subjects as a function of the three categories of sleep awakenings (none, sometimes, regularly) was observed, with the category of regular nighttime awakenings having the highest rate ofUQS subjects (45.0% ± 3.12% for women and 32.8% ± 3.44% for men). Among subjects reporting regular awakenings during the night, the proportion of UQS subjects is significantly higher among those who report waking up at least twice a night compared to those waking once a night. For women, 33.7% (el 95%: ±4.9) of those who wake up once in the night, 48.6% (el 95%: ±5.33) of those waking twice a night and 58.5% (el 95%: ±6.0l) of those who experience at least three awakenings per night are found to be UQS. As for the men, 24% (el 95%: ±6.4) of those who wake up once during the night, 37.2% (el 95%: ±6.4) of those waking up twice during the night and 51.8% eCI 95%: ±8.25) M.OHAYON S14 70 60 o WOMEN [SJ MEN SO ~ ~ 30 20 ..... ..... N 10 None Sometimes Regularly Nighttime awakenings FIG. 5. Nighttime awakenings. included in definitions of insomnia (13,14,16,17). RUther (8), as well as the ICD-l 0, regard an increased mental activity before the main sleeping period as constituting a characteristic of insomniacs. It is generally accepted that an individual suffers from insomnia when he or she perceives the quantity or quality of sleep as being unsatisfactory and that this insomnia leads to repercussions on the individual's daily functioning (13,14,17-19). On this basis, we have divided our population into two distinct groups according to the subjects' reports at the moment of the survey: satisfied or not with the quantity or quality of their sleep. These two groups were then further subdivided to take into account the therapeutics used. The quantity of information gathered in the population forces us to present only some of the main epidemiological themes. Thus, the prevalence of dissatisfaction with sleep or use of sleep-enhancing or anxiety-reducing medication at the time of the survey was 20.1%; 39.7% of these subjects report that this problem has been present for over 5 years. Compared to studies conducted by Mellinger et al. (1), Klink and Qhan (2), Bixler et al. (5) and Quera-Salva et al. (4), comparable rates on the criteria used in these studies were found with ours: 30% of subjects had regular nighttime awakenings; 24.4% had a sleep onset period over 30 minutes (among which 8.6% were over 1 hour); 7.1 % had early morning awakenings (before 5:00 a.m.). Women and elderly persons are those who are most affected by sleep disturbances. In the present study, a Sleep, Vol. 19, No.3, 1996 Acknowledgement: This research was granted by Synthelabo and the F.R.S.O. REFERENCES 1. Mellinger GD, Balter MB, Uhlenhuth EH. Insomnia and its treatment. Arch Gen Psychiatry 1985;42:225-32. 2. Klink M, Quan SF. Prevalence of reported sleep disturbances in a general adult population and their relationship to obstructive airways diseases. Chest 1987;:54~6. 3. Ford DE, Kamerow DB. Epidemiologic study of sleep disturbances and psychiatric disorders: an opportunity for prevention? JAMA 1989;262: 1479-84. 4. Quera-Salva MA, Orluc A, Goldenberg F, Guilleminault C. Insomnia and use of hypnotics: study of a french population. Sleep 1991; 14:386-91. 5. Bixler EO, Kales A, Siodatos CR, Kales JD, Healey S. Prevalence of sleep disorders in the Los Angeles metropolitan area. Am J Psychiatry 1979; 136: 1257-62. 6. Gislason T, Almqvist M. Somatic diseases and sleep complaints: an epidemiological study of3,201 Swedish men. ActaMed Scand 1987;221:475-581. 7. Houyez F, Degoulet P, Cittee J, Fouriaud C, J Lang T, Aimee F. Sleep and hypertension: an epidemiologic study in 7,901 workers. Arch Mal Coeur Vaiss 1990;83:1085-8. 8. RUther E. Depression, circadian rhythms and trimipramine. Drugs 1989;38 Suppll:I-3. 9. Mosko S, Zetin M, Glen S, et al. Self-reported depressive symptomatology, mood ratings, and treatment outcome in sleep disorders patients. J Clin PsychoI1989;45:51-60. 10. Souetre E. Troubles du sommeil lies Ii l'anxiete. Presse Med 1990; 19: 1839-41. Downloaded from https://academic.oup.com/sleep/article-abstract/19/suppl_3/S7/2749899 by guest on 20 May 2020 ~ 40 9.9% rate of consumption of medication for sleep difficulties was found. This was comparable to that found in another French study (10%) (4). One important pitfall in the epidemiological recognition of hypnotic prescriptions is found in the intricacy of anxiety, depression and insomnia leading to the prescription of different psychotropic molecules whose effects on sleep are well known. We will be taking up this theme in more detail in a further article. Another question is that of prescriptions associated with a hypnotic therapeutic: 39.2% of these subjects take 2 to 13 types of medication in association (13.1 % take 5 types of medication or more). Of subjects who use medication for sleeping difficulties or anxiety, 38.2% are treated for a physical illness. This latter conclusion and the great number of elderly subjects in this population largely explain the associated prescriptions. In conclusion, our study shows the prevalence of insomnia (in relation to dissatisfaction with sleep) in the French population. Recourse to medication for sleeping difficulties or anxiety is frequent. The number of elderly subjects affected is large and the association with an organic or psychiatric pathology is also frequent. Certain sociodemographic elements can be identified. However, it appears that although there is agreement with other studies as to the prevalence of insomnia in the general population, few studies have been carried out to assess the socio-professional consequences of insomnia. EPIDEMIOLOGICAL STUDY ON INSOMNIA 16. Infante-Rivard C, Dumont M, Montplaisir J. Sleep disorder symptoms among nurses and nursing aides. Int Arch Occup Environ Health 1989;61:353-8. 17. American Psychiatric Association (Task Force on DSM-IV). DSM-IV Draft criteria. Washington, DC: The American Psychiatric Association, 1993. 18. Lisi DM. Comment: temazepam and diphenhydramine in nursing homes. DICP, The Annals of Pharmacotherapy 1989;23: 178-80. 19. Hauri PI, Esther MS. Insomnia. Mayo Clin Proc 1990;65:86982. Downloaded from https://academic.oup.com/sleep/article-abstract/19/suppl_3/S7/2749899 by guest on 20 May 2020 11. Kish L. Survey sampling. New York: John Wiley and Sons Inc, 1965. 12. Adam K, Oswald I. Effects of repeated ritanserin on middleaged poor sleepers. Psychopharmacology 1989;99:219-21. 13. Organisation Mondiale de la Sante. CIM-IO: classification internationale des troubles mentaux et des troubles du comportement; descriptions cliniques et directives pour Ie diagnostic. Paris: Masson, 1993. 14. American Psychiatric Association. Diagnostic and statistical manual of mental disorders, third edition, revised (DSM-III-R). Washington, DC: The American Psychiatric Association, 1987. 15. Reynolds CF. The implications of sleep disturbance epidemiology. JAMA 1989;262:1514. SI5 Sleep, Va!. 19, No.3, 1996