Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Sleep Res. Author manuscript; available in PMC 2012 September 1.
Published in final edited form as:
PMCID: PMC3135706

Daytime symptom patterns in insomnia sufferers: is there evidence for subtyping insomnia?


The type and severity of daytime symptoms reported by insomnia sufferers may vary markedly. Whether distinctive daytime symptom profiles are related to different insomnia diagnoses has not been studied previously. Using profile analysis via multidimensional scaling, we investigated the concurrent validity of ICSD-2 insomnia diagnoses by analysing the relationship of prototypical profiles of daytime symptoms with a subset of ICSD-2 diagnoses, including insomnia associated to a mental disorder, psychophisiological insomnia, paradoxical insomnia, inadequate sleep hygiene, idiopathic insomnia, obstructive sleep apnea and restless legs syndrome. In a sample of 332 individuals meeting research diagnostic criteria for insomnia (221 women, Mage=46 yrs.), the profile analysis identified four prototypical patterns of daytime features. Pearson correlation coefficients indicated that the diagnoses of insomnia associated to a mental disorder and idiopathic insomnia were associated with a daytime profile characterized by mood disturbance and low sleepiness; whereas the diagnoses of psychophysiological insomnia and inadequate sleep hygiene were related to a profile marked by poor sleep hygiene, daytime tension and low fatigue. Furthermore, whereas paradoxical insomnia was consistently associated to lower daytime impairment, insomnia associated to a mental disorder was related to the most severe daytime impairment. This classification of insomnia sufferers along multiple defining dimensions provides initial validation for two basic insomnia subtypes, with a presumably distinct aetiology: insomnia characterized mainly by an “internal” component, and a “learned” insomnia. Research to determine which dimensions are critical for inclusion or differential weighting for defining a general typological system for insomnia sufferers is warranted.

Keywords: Insomnia, ICSD-2, Daytime symptoms, Concurrent validity


Current definitions of insomnia describe this condition as a disorder characterized not only by nighttime sleep difficulty, but also by daytime symptoms such as fatigue, sleepiness, mood disturbances, and cognitive difficulties (Edinger et al., 2004) Indeed, it has been well-documented that subjective wellbeing (Buysse et al., 2007), as well as psychomotor performance (Edinger et al., 2008), are compromised in insomnia sufferers when compared to normal sleepers. Furthermore, contemporary sleep disorder nosologies require some form of patient-reported daytime impairment for a clinical diagnosis of insomnia to be made (American Psychiatric Association, 2000, American Academy of Sleep Medicine, 2005).

As is the case for nighttime symptoms wherein patients may differ in terms of presenting complaint, the type and severity of daytime symptoms reported by insomnia sufferers may also vary markedly. It has been generally assumed that some daytime features accompanying the sleep-related complaint may provide clues to both the pathophysiology and the characterization of the specific insomnia disorder. Indeed, the International Classification of Sleep Disorders (ICSD & ICSD-2)(American Sleep Disorders Association, 1997, American Academy of Sleep Medicine, 2005) delineates various highly specific insomnia subtypes based, for the most part, on daytime-related features. For instance, the core and distinguishing hallmark for the diagnosis of inadequate sleep hygiene is the performance of daytime activities that are inconsistent with the maintenance of good-quality sleep. Likewise, one of the essential characteristics for the diagnosis of paradoxical insomnia is the report of severe sleep complaints without supporting daytime sequelae, such as excessive daytime sleepiness. Surprisingly, research to determine whether specific types of daytime complaints distinguish the insomnia diagnostic categories or support sharp boundaries between them has received scant attention.

Furthermore, although the appeal of defining different insomnia subtypes along one dimension (e.g. poor sleep hygiene behaviours, level of depression) is obvious from the standpoint of parsimony, dimensions may be highly correlated and it is likely that no single dimension can adequately separate insomnia sufferers into homogeneous subtypes. Under this framework, research geared toward identifying prototypical diurnal patterns fitting particular diagnoses, rather than merely detecting the presence/absence of “key” daytime features associated with a particular subtype, may be a more helpful approach. Indeed, integrated descriptions of patients’ daytime symptoms patterns may help advance our understanding of insomnia complaints generally, and possibly of diverse subtypes more specifically.

The purpose of this exploratory study was to identify prototypical profiles of daytime features in a sample of individuals meeting research diagnostic criteria (RDC) for a general insomnia disorder (Edinger et al., 2004). The second aim was to use those profiles to study the concurrent validity of the most common ICSD-2 insomnia diagnoses (Buysse et al., 1994), such as insomnia associated to a mental disorder (IMD) and the “primary” insomnias, such as psychophysiological insomnia (PI), idiopathic insomnia (IDI), paradoxical insomnia (PXI) and inadequate sleep hygiene (ISH).



Data used herein come from a larger dual-site study conducted primarily to determine the reliabilities and validities of the DSM-IV-TR and ICSD-2 insomnia diagnoses that can be ascertained via interview (MH067057). This study was approved by the Institutional Review Boards of the collaborating study sites at Duke University Medical Center, Durham, NC, and Rush University Medical Center, Chicago, IL.

Participants were recruited between January 2004 and February 2009 through posted announcements and referrals by providers. For ecological validity, participant selection criteria were designed to allow enrolment of a diverse group of insomnia sufferers. Study candidates had to: (1) meet RDC for a general insomnia disorder (Edinger et al., 2004); (2) be 18 years of age or older; and (3) be fluent in English. Exclusion criteria were designed to prevent enrolment of those individuals who would be unable to fully participate in the trial completely and safely, such as those (1) suffering an unstable or life-threatening medical condition (2) imminently suicidal; (3) with cognitive impairment, as evidenced by a score of < 24 on the Mini Mental Status Exam; (4) being medical or psychiatric inpatients at the time of volunteering for the study. Study candidates who had previously been evaluated by any of the study clinicians were excluded.

Among the 332 participants included in this study (186 from Duke and 146 from Rush), there were about twice as many females as males (ratio 1.99:1) and the mean age was of 45.98 years (SD = 14.40 years). Table 1 includes the demographic characteristics of the sample.

Table 1
Sociodemographic characteristics of participants (N=332)


Study procedures included: (1) the Structured Clinical Interview for DSM-IV, SCID, (Spitzer et al., 1996); (2) two consecutive nights of laboratory polysomnography (PSG); (3) a sleep-history questionnaire (SHQ); (4) two weeks of sleep diary monitoring (SD); and (5) a computerized battery of self-report measures designed to assess sleep and daytime functioning. Once these procedures were completed, each participant was randomly assigned to one of the six possible orders of three different diagnostic interview methods. The first diagnostic method included solely the use of the Duke Structured Interview for Sleep Disorders (DSISD) (Edinger et al., 2009). The DSISD includes questions that incorporate criteria for ascertaining sleep disorders within both the DSM-IV-TR and ICSD-2 sleep disorder nosologies. The second diagnostic method comprised an unstructured clinical interview and a review of the participant’s SHQ and SDs. The third method included an unstructured clinical interview as well as a review of the patient’s SHQ, SDs and PSG summary data. More detailed descriptions of these instruments are provided below.

Given the objectives of the parent study, a total of six clinicians at each study site, MD and PhD sleep specialists with clinical experience in the diagnosis of sleep disorders, were identified to serve as diagnosticians. At each site these clinicians were divided into three pairs and then each dyad was randomly assigned to one of the assessment methods described above. Each dyad used the same diagnostic method during the entire study duration. Each participant was assessed, at different time periods, by the three different diagnostic methods.

All enrolled participants provided written informed consent and received parking expenses plus a maximum $400.00 payment if they completed all study procedures. Figure 1 provides a summary of the study flow.

Figure 1
Flow of participants and study procedures

Sleep Assessments

Participants were asked to complete some self-report instruments including a sleep diary for two weeks and the above-mentioned SHQ. Variables derived from the diaries included bedtime, sleep onset latency (SOL), number of nocturnal awakenings, time awake after sleep onset and prior to final awakening (WASO), time of final awakening, time of rising out of bed, total sleep time (TST), total time awake, time in bed (TIB), sleep efficiency (SE = total sleep time/time in bed * 100%). The SHQ is a 10-page paper and pencil instrument that solicited information about the participant’s demographics, current and past sleep complaints, medical and psychiatric history, and previous treatment history.

Participants underwent two consecutive nights of laboratory PSG. PSGs were scored using standard scoring criteria for sleep stages, apneas/hypopneas, periodic limb movements (PLMS) and related arousals (Rechtshaffen and Kales, 1968, Phillipson and Remmers, 1989, Coleman, 1982). PSG summary data included an apnea-hyponea index, PLMS index and a desaturation index (i.e., number of O2 declines ≥ 3% from baseline per hour of sleep), along with bedtime, SOL, number of nocturnal awakenings, WASO, time of final awakening, time of rising out of bed, TST, total time awake, TIB and SE.

Daytime measures

Study participants completed a battery of self-report measures. Unlike the PSG reports, SDs and SHQs, data derived from this battery were not made available to any of the study clinicians. For this specific study we selected measures of the daytime symptoms and practices most frequently reported by insomnia sufferers, such as sleepiness, fatigue, depression, anxiety, anger and poor sleep hygiene behaviours (Sateia et al., 2000). The instruments used included the Epworth sleepiness scale (ESS) (Johns, 1991), the Fatigue severity scale (FSS) (Krupp et al., 1989), the Inventory to diagnose depression (IDD) (Zimmerman and Coryell, 1987), the Tension/Anxiety (T) and Anger/Hostility (A) scales from the Profile of mood states questionnaire (POMS) (McNair et al., 1971), and the Sleep hygiene practice section from the Sleep hygiene awareness and practice scale (SHAPS) (Lacks and Rotert, 1986).

Diagnostic ratings

Immediately after interviewing each participant, the clinicians independently completed diagnostic rating forms to summarize their diagnostic impressions. The rating forms consisted of series of diagnoses presented on the screen of a specially programmed PDA hand-held computer. The first series of diagnostic choices consisted of 10 DSM-IV-TR sleep disorders diagnoses relevant to insomnia complaints. The second series of diagnoses consisted of 37 diagnostic choices selected from the ICSD-2 manual. Each diagnosis appeared on the PDA screen individually accompanied by a 100 pixel visual analogue scale (VAS) labelled “doesn’t fit at all” at its left extreme and “fits extremely well” at its right extreme. Clinicians were instructed to consider each diagnosis separately and rate how well that diagnosis “fit” the study participant. These ratings were converted into numeric values based on their location on the 100 pixel VAS line.

Since one of our objectives was testing the concurrent validity of the most frequently assigned ICSD-2 insomnia diagnoses, only a subset of these diagnoses were considered. Included among those were IMD, PI, IDI, PXI, and ISH. The diagnoses of obstructive sleep apnea (OSA) and restless legs syndrome (RLS) were also included so as to test the discrimination of insomnia disorders from other sleep disorders. Ratings assigned to each participant on each diagnostic category by the 6 clinicians were averaged. As a result, each participant had a mean rating on each of the 7 diagnoses included in this study, with higher values connoting the better the fit for a particular diagnosis.

Data analyses

In order to explore any relationship between basic demographics and insomnia diagnoses, we used T-test analyses to compare the mean ratings obtained by women and men on each diagnostic category. Likewise, we divided our sample in 3 age groups, i.e., younger than 40 years, between 40 and 60 years, and older than 60 years old, and conducted one-way ANOVAs with post-hoc tests (i.e. Student-Newman-Keuls) to examine whether ratings on each diagnostic category were equally prominent across the three age groups.

To identify the predominant daytime symptom profiles in our sample we used a procedure called profile analysis via multidimensional scaling (PAMS) (Davison et al., 1996, Kim et al., 2004, Frisby and Kim, 2008). Following the guidelines of Davidson et al. (1996), we submitted a 332-person × 6 daytime measure standardized-score matrix (scores on the ESS, FSS, IDD, SHAPS, Tension-anxiety and Anger-hostility POMS subscales) to a simple nonmetric multidimensional scaling analysis using the ASCAL program, available in SPSS version 15.0 (SPSS, Chicago, IL).

Along with the prototypical profiles, PAMS determines, for each individual included in the data set, the so called person parameters. These are quantitative variables indicating (1) the degree of “matching” of each individual observed profile and each one of the identified prototypical profiles, and (2) the individual observed profile’s elevation, i.e. if the scores on the measures considered overall are above or below average. The former variables are called weights and the latter is called the level parameter.

To ascertain if participants with higher symptom severity (as indicated by the level parameter), showed a tendency toward a particular profile’s shape, individual differences in the profile level were correlated with individual differences in profile weights by using Pearson correlation coefficients (Kim et al., 2004). Additionally, we tested the concurrent validity of the selected ICSD-2 diagnoses by evaluating how the participants’ patterns of daytime symptoms mapped onto their clinically assigned diagnoses. We computed Pearson correlations between participants’ loadings/weights on each of the prototypical daytime profiles and their average ratings on each of the above mentioned ICSD-2 diagnostic categories included in this study. Likewise, we conducted correlations between the individuals’ level parameters and their average diagnostic ratings. All statistical analyses were conducted using the SPSS software, version 15.0 (SPSS, Chicago, IL). Statistical significance was set at ≤ 0.05.


Nightime and daytime characteristics of participants

Table 2 summarizes descriptive statistics on selected self-reported sleep variables and daytime measures.

Table 2
Self-reported nighttime and daytime characteristics of participants (N=332)

Ratings on ICSD-2 diagnoses and demographic variables

As shown in Table 3, women had significantly higher ratings than men on the ISH and PI diagnostic categories, whereas men obtained higher ratings on the OSA diagnosis. Likewise, older individuals obtained significantly higher ratings on the ISH diagnosis and lower ratings on the IMD diagnostic category.

Table 3
Mean diagnostic ratings (SD) across gender and age groups

Prototypical daytime profiles obtained with PAMS analysis

In PAMS analysis, the goodness of fit of the dimensional solution is evaluated by means of the Stress and R2 fit statistics. The Stress is a normalized, least squares measure of fit, which ranges from 0 to 1; the closer the Stress value is to 0, the better the fit (Davison et al., 1996). The R2 value reflects the proportion of variance in the rescaled data that is accounted for by the reproduced distances. The PAMS model fits for the 2-dimensional solution were that Stress = 0.11 and R2 = 0.93, whereas the fits for the unidimensional model were that Stress = 0.33 and R2 = 0.66. Hence, based on the fit statistics, we retained the 2-dimensional solution.

The 2-dimensional solution yielded 4 prototypical profiles. The prototypical profile 1 and its mirror image are plotted in Figure 2. As can be seen, profile 1 is characterized by high positive scores on sleepiness and low scores on mood disturbance, i.e. depression, anger and tension. Mirror image profile 1 is therefore marked by negative affect coupled with low scores on daytime sleepiness.

Figure 2
Dimension 1: Prototypical profile 1 and mirror image profile 1

Figure 3 depicts the prototypical profile 2 and its mirror image. Profile 2 shows high levels of fatigue combined with the absence of poor sleep hygiene practices. The mirror image of profile 2 is thus marked by low levels of fatigue but by relatively high levels of tension-anxiety linked with poor sleep hygiene practices.

Figure 3
Dimension 2: Prototypical profile 2 and mirror image profile 2

Person parameters: correspondence between observed and prototypical profiles and their relationship

Each participant was assigned weights on the two identified dimensions and a level parameter. PAMS also provided a R2 value, indicating the proportion of variance in the participant’s observed profile that was explained by the prototypical profiles. To illustrate the meaning of these person parameters, which are represented as z values, a group of 3 individuals was selected from the 332 participants on which PAMS prototypical profiles were obtained. As can be seen in table 4, participant No. 1102 had a high negative weight, −1.46, on dimension 1 and a considerably lower weight on dimension 2, 0.49. We thus surmised that this individual’s observed profile tended to reflect the mirror image of profile 1. This person’s observed profile was elevated, as reflected by the level parameter estimate, 1.47, indicating a global score on all the measures far above average overall. Conversely, participant No. 2002 had a notably positive weight on dimension 1, 0.75, and a small weight on the second dimension, −0.05. This observed profile thus tends to parallel the shape of the prototypical profile 1. In addition, the level parameter close to 0, i.e., 0.5, indicates that the participant scored around average overall. Finally, participant No. 2046 had a negligible weight on dimension 1, −0.05, but a substantially negative weight on dimension 2, −0.65. This person’s observed profile would then display a trend toward the mirror image of profile 2. The negative level parameter, −0.86, indicates that this participant scored below average overall. Inasmuch as we selected these specific examples to illustrate somewhat close adherence to the prototypical profiles, all reported R2 values were high.

Table 4
Person parameters and R2 estimated by PAMS

Figure 4 graphically compares the observed scores (presented as z values) of participants No. 1102 and 2002 with the prototypical profile 1.

Figure 4
Observed profile patterns of two participants superimposed to the prototypical profile 1

With respect to the relationship between weights on dimensions and the level parameter, we found a significant negative correlation between the participants’ level parameter and their weights on dimension 1, r = −0.240, p < 0.001. Since the level parameter shows whether individuals have scored above or below average overall and, hence, a higher value would suggest more severe daytime symptoms overall, this negative correlation indicates that individuals with more severe daytime symptomatology show a tendency towards negative weights on dimension 1. Translating that into profiles, this statistically significant correlation suggests that individuals getting higher level parameters tend to exhibit observed profiles of scores emulating the particular pattern of peaks and valleys characterizing mirror image profile 1. By contrast, the Pearson correlation coefficient between the level parameter and weights on dimension 2 was not significant, r = 0.046, p = 0.404. This non-significant value indicated that neither paralleling profile 2 nor its mirror image was consistently associated with a higher symptom severity.

Relationship between person parameters and ICSD-2 diagnostic ratings

Table 5 shows that statistically significant negative correlations were found between dimension 1 weights and the clinicians’ ratings for the diagnoses of IMD, r = −0.344, p < 0.001, and IDI, r = −0.130, p = 0.018. These negative correlations suggest that participants obtaining higher ratings on those diagnoses tended to exhibit daytime profiles resembling mirror image profile 1. In contrast, ratings for the diagnosis of OSA were positively associated with dimension 1 weights, r = 0.184, p = 0.001. In other words, the better the fit for the diagnosis of OSA, the better the match for the profile 1.

Table 5
Pearson’s product moment correlation coefficients between PAMS-derived person parameters and clinicians’ ratings on ICSD-2 diagnostic categories

Our correlations also showed that the clinicians’ ratings for the diagnoses of ISH and PI were significantly associated to dimension 2 weights. Again, these correlations were both negative, ISH r= −0.122, P = 0.026, and PI r = −0.115, P = 0.036, suggesting that participants with higher mean ratings on these diagnostic categories tended to exhibit profiles of scores matching the mirror image profile 2, the one characterized by poor sleep hygiene and anxiety/tension.

The level parameter also demonstrated some association with clinical ratings. We found significant positive correlations between the level parameter values and the ratings for the diagnoses of IMD, OSA and RLS, with IMD displaying the strongest correlation, r = 0.395, p < 0.001. Thus, there was a tendency for greater daytime symptom severity in individuals getting higher ratings on these sleep disorders categories. Conversely, a better fit for the diagnosis of PXI was negatively associated with the level parameter, r = −0.156, p = 0.004. That is, individuals getting higher ratings on this diagnostic category tended to present less severe daytime symptomatology.


These findings suggest that some insomnia subtypes can be distinguished in terms of the more predominant patterns and severity of daytime symptoms they present. One of our core daytime symptom profiles, characterized by mood disturbance, was associated with the diagnoses of IMD and IDI, whereas another symptom profile, marked by poor sleep hygiene and tension, was related to the diagnoses of ISH and PI. These associations were not only statistically significant, but also appeared clinically meaningful. Moreover, they may prove useful in developing etiological explanations for different types of insomnia sufferers. Indeed, our results suggest the distinction of two basic insomnia presentations which could have, presumably, distinct underlying pathophysiologies. Within this framework, the similar daytime features shared by IMD and IDI may be interpreted as indicative of a common underlying factor, which is distinguished from the diagnoses of PI and ISH. This category would include a form of insomnia arising from “internal” factors, with a predominantly neurophysiological/neurochemical basis. Likewise, ISH and PI would denote another basic form of insomnia; one that has mainly arisen from maladaptive behavioural factors, a “learned” insomnia.

However, it should be noted that, although the correlation coefficients between the PI and ISH diagnoses and dimension 1 weights didn’t reach statistical significance, their absolute value was close to 0.10 and, therefore, didn’t seem to differ much from their correlations with the poor seep hygiene and tension/anxiety profile. Actually, this finding is related to one of the most appealing features of PAMS analysis. Unlike other profile analysis techniques, which group individuals into single relatively homogeneous groups or clusters, PAMS assumes that individuals may share features of the global core prototypical profiles defining the larger sample. Indeed, our findings suggest that individuals getting higher ratings on the ISH diagnostic category may also show observed profiles of scores resembling the prototypical profile characterized by sleepiness (r=0.105, p=0.056). By contrast, higher ratings on the PI diagnostic category seemed to be marginally related to the prototypical profile marked by negative affect (r=−0.096, p=0.081). These observations are, indeed, clinically meaningful. Although PI and ISH could be both considered forms of “learned” insomnia, some distinctiveness between these two insomnia subtypes seems to exist. Likewise, our findings suggest that PI should not be simply regarded as a milder form of IMD, notwithstanding PI individuals may be prone to show features of mood disturbance. Indeed, IMD ratings were not related to the daytime profile most prominently associated to PI, the one marked by poor sleep hygiene behaviours.

Our daytime profile characterized by elevated sleepiness was associated with the OSA diagnostic category. Epidemiological studies have suggested a link between insomnia and subjective daytime sleepiness (Johns and Hocking, 1997), although insomnia patients typically do not display an irresistible proneness to fall asleep in the daytime. Since the sleepiness scale used herein (i.e. ESS) measures the latter tendency, it is not surprising that the sleepiness profile was not strongly linked to any of our insomnia groups.

The innovative profile analysis technique used herein allowed us to both distinguish individuals in terms of the more predominant patterns of daytime features they presented and to locate those individuals along a continuum of severity of daytime impairment, regardless the type of symptoms. Individuals getting better fits for the diagnoses of IMD, OSA and RLS seemed to display greater daytime impairment. In contrast, individuals obtaining higher ratings in the subjective insomnia diagnostic category, i.e. PXI, were those exhibiting less severe daytime features. Once again, our results support a distinction of insomnia subtypes, with IMD representing a more impairing form of insomnia. Moreover, since objective short duration has been proposed as an objective marker of the biological severity of insomnia (Fernández-Mendoza et al, 2010), and PXI is theoretically characterized by normal sleep duration, the milder severity of daytime features distinctively associated with the diagnosis of PXI seems congruous with this characterisation.

One methodological advantage of this study was the diagnostic imputation method used. Our participants were not assigned a single ICSD-2 diagnosis, nor were they assigned a “first diagnosis” followed by a “second diagnosis” and so forth. Instead, each participant obtained a continuous score on each of the diagnostic categories considered in this study. As such, a participant could obtain a high mean rating for the PXI diagnostic category, a more modest mean rating for the PI diagnostic category, a lower mean rating for the OSA diagnostic category, etc. The individual thus fulfilled some criteria for other insomnia-related diagnoses, but the best-fitting diagnostic label for this specific individual would be that of PXI. It is important to note that higher mean scores in a particular diagnostic category not only indicated that the patient better mapped onto that diagnosis, but also better diagnostic agreement between clinicians.

Although this diagnosing method is clearly not practical under routine clinical conditions it can be very useful for research purposes. Indeed, it reduces the problem of diagnostic reliability by addressing two fronts: first, the bias related to the background of the clinician on diagnosis assignment is attenuated by considering a combination of the 6 clinicians’ ratings. Second, since there is some overlap between ICSD-2 diagnostic categories, and differences between subtypes may be more quantitative than qualitative, such a dimensional representation of the insomnia diagnoses may be much more valid than the traditional categorical distinction (i.e., a yes/no decision about the presence of a particular diagnosis). Consequently, this method of diagnosis assignment doesn’t “force” the clinician to rigidly choose one diagnosis in favour of the other diagnoses.

Admittedly, this investigation had some limitations. Our findings are influenced by the self-report instruments used herein. Although they were intended to measure a comprehensive array of daytime features commonly reported by insomnia sufferers, they do not represent an exhaustive, all-inclusive battery of clinically relevant daytime symptoms. It is possible that the inclusion of additional measures of different daytime characteristics (e.g. hyperarousal, beliefs about sleep) could have resulted in the identification of somewhat different core symptom profiles.

Furthermore, the statistically significant correlations between insomnia diagnoses and daytime profiles were not overly strong. This may be related to the poor reliability of the majority of the insomnia diagnoses under study (Edinger, Wyatt, Olsen et al., 2010). Indeed, an unreliable measure of a construct, i.e. insomnia diagnosis, will likely attenuate the “true” association of the construct with other variables. Interestingly, the more unreliable diagnoses, according to the results reported by Edinger et al. (2010), i.e. PI and ISH, were the ones obtaining smaller correlation values in the present study, whereas the most reliable ICSD-2 insomnia diagnosis, i.e., IMD, showed the higher correlation value herein. On the other hand, our sample was comprised largely of adult research volunteers so the findings may not generalize to clinical patients or younger age groups. Moreover, only a subset of the clinicians had access to participants’ PSGs so it is possible, if not likely, that the clinicians lacking this information made inaccurate judgements about the presence vs. absence of diagnoses such as OSA. However, since two of the clinicians did have PSG information and all clinicians’ assessments were integrated in a mean rating, these diagnostic errors may have been offset to some degree.

Despite its limitations, this study provides empirical support for greater integration of daytime symptoms with nighttime symptoms in assessments of insomnia. Arguably, an important area for research will be to determine which dimensions are critical for inclusion or differential weighting for defining a general typological system for insomnia sufferers. Since insomnia diagnosis relies primarily on self-report, rather than on confirmatory laboratory tests, the complementary use of psychometrically sound self-report measures of patients’ daytime symptoms could perhaps serve the role of validating the clinical impressions and discriminating the putatively distinctive insomnia subtypes. Nonetheless, the development of a simple algorithm to classify multidimensional subtypes for enhancing diagnostic reliability and validity requires further research.


SUPPORT: National Institute of Mental Health, Grant # R01MH067057


Disclosure statement

Dr. Edinger has consulting relationships with Philips-Respironics, Inc. & Kingsdown, Inc. He has also received grant support from Philips-Respironics, Inc. & Helicor, Inc.

Dr. Wyatt has received grant support from the Respironics Sleep and Respiratory Research Foundation

Dr. Sánchez-Ortuno has no conflicts of interest.


  • American Academy of Sleep Medicine. The international classification of sleep disorders : diagnostic & coding manual. 2 American Academy of Sleep Medicine; Westchester, IL: 2005.
  • American Psychiatric Association. Diagnostic and statistical manual of mental disorders : DSM-IV-TR. 4 American Psychiatric Association; Washington, DC: 2000. Task Force on DSM-IV.
  • American Sleep Disorders Association. The international classification of sleep disorders, revised : diagnostic and coding manual. American Sleep Disorders Association; Rochester, MN: 1997.
  • Buysse D, Reynolds CR, Kupfer D, et al. Clinical diagnoses in 216 insomnia patients using the International Classification of Sleep Disorders (ICSD), DSM-IV and ICD-10 categories: a report from the APA/NIMH DSM-IV Field Trial. Sleep. 1994;17:630–637. [PubMed]
  • Buysse D, Thompson W, Scott J, et al. Daytime symptoms in primary insomnia: a prospective analysis using ecological momentary assessment. Sleep Med. 2007;8:198–208. [PMC free article] [PubMed]
  • Coleman R. Periodic movements in sleep (nocturnal mioclonus) and restless leg syndrome. In: Guillerminault C, editor. Sleeping and waking disorders: Indications and techniques. Addison-Wesley; Menlo Park: 1982. pp. 265–295.
  • Davison ML, Gasser M, Ding SA. Identifying major profile patterns in a population: An exploratory study of WAIS and GATB patterns. Psychol Assess. 1996;8:26–31.
  • Edinger J, Bonnet M, Bootzin R, et al. Derivation of research diagnostic criteria for insomnia: report of an American Academy of Sleep Medicine Work Group. Sleep. 2004;27:1567–1596. [PubMed]
  • Edinger JD, Means MK, Carney CE, Krystal AD. Psychomotor performance deficits and their relation to prior nights’ sleep amona individuals with primary insomnia. Sleep. 2008;31:599–607. [PubMed]
  • Edinger JD, Wyatt JK, Olsen MK, et al. Reliability and validity of the DUKE structured interview for sleep disorders for insomnia screening. Sleep. 2009;32:0810. (Abstract)
  • Edinger JD, Wyatt JK, Olsen MK, et al. Utility, validity, and methodological factors influencing assignment of ICSD-2 insomnia diagnoses: considerations for future insomnia nosologies. J Sleep Res. 2010;19:43. (Abstract)
  • Fernández-Mendoza J, Calhoun S, Bixler EO, et al. Insomnia with objective short duration is associated with deficits in neuropsychological performance: a general population study. Sleep. 2010;33:459–465. [PubMed]
  • Frisby CL, Kim SK. Using profile analysis via multidimensional scaling (PAMS) to identify core profiles from the WMS-III. Psychol Assess. 2008;20:1–9. [PubMed]
  • Johns M, Hocking B. Daytime sleepiness and sleep habits of Australian workers. Sleep. 1997;20:844–849. [PubMed]
  • Johns MW. A new method for measuring daytime sleepiness - The Epworth sleepiness scale. Sleep. 1991;14:540–545. [PubMed]
  • Kim SK, Frisby CL, Davison ML. Estimating cognitive profiles using profile analysis via multidimensional scaling (PAMS) Multivariate Behav Res. 2004;39:595–624.
  • Krupp LB, Larocca NG, Muirnash J, Steinberg AD. The fatigue severity scale - application to patients with multiple-sclerosis and systemic lupus-erythematosus. Arch Neurol. 1989;46:1121–1123. [PubMed]
  • Lacks P, Rotert M. Knowledge and practice of sleep hygiene techniques in insomniacs and good sleepers. Behav Res Ther. 1986;24:365–368. [PubMed]
  • Mcnair DM, Lorr M, Droppleman LF. Profile of Mood States Manual. Educational and Industrial Testing Service; San Diego, CA: 1971.
  • Phillipson EA, Remmers JE. American Thoracic Society Consensus Conference on Indications and Standards for Cardiopulmonary Sleep Studies. Am Rev Respir Dis. 1989;139:559–568. [PubMed]
  • Rechtshaffen A, Kales A. A manual of standardized terminology, techniques, and scoring systems of sleep stages of human subjects. UCLA Brain Information Service/Brain Research Institute; Los Angeles: 1968. [PubMed]
  • Sateia M, Doghramji K, Hauri P, Morin C. Evaluation of chronic insomnia. An American Academy of Sleep Medicine review. Sleep. 2000;23:243–308. [PubMed]
  • Spitzer R, Williams J, Gibbons M, First M. Instruction Manual for the Structured Clinical Interview for DSM-IV (SCID-IV). (SCID 1996 Revision) Biometrics Research Department, New York Psychiatric Institute; New York: 1996.
  • Zimmerman M, Coryell W. The inventory to diagnose depression (IDD) - A self-report scale to diagnose major depressive disorder. J Consult Clin Psychol. 1987;55:55–59. [PubMed]