1 Introduction

The International Classification of Sleep Disorder-Third Edition (ICSD 3), defines insomnia disorder as difficulty in initiating or maintaining sleep despite adequate time and circumstances available to sleep, with daytime consequences [1]. Insomnia disorder is classified in ICSD 3 as “chronic” (occurs 3 or more days per week for at least 3 months) [1], “short-term” (the sleep disturbance and associated daytime symptoms have been present for < 3 months) and “other” (does not meet the full criteria for either chronic insomnia disorder or short-term insomnia disorder) [2]. Individuals with insomnia have an increased incidence of cardiovascular morbidity, diabetes mellitus, obesity and depression, impaired work performance, work-related/motor vehicle accidents, and overall poor quality of life [3,4,5,6].

Unlike other sleep disorders, such as breathing-related and circadian rhythm disorders that are confirmable using relatively objective methods, chronic or transient insomnia is primarily diagnosed by self-reported sleep history obtained during clinical interview. Hence, screening tools like questionnaires have an important place in the management of insomnia as they help not only in diagnosis but also in post-intervention reassessment [7]. Some of the questionnaires available for insomnia include the Pittsburgh Sleep Quality Index [8], Oviedo Sleep Questionnaire [9], Athens Insomnia Scale (AIS) [10, 11], and Insomnia Severity Index [12].

These questionnaires for insomnia in general reflect the insomnia perception in the developed countries [13]. However, there are global variations in the symptoms of insomnia and their management [14]. Also, sleeping patterns have been observed to vary across cultures with cultural practices pertaining to sleep impacting sleep duration and quality [15]. Hence, there is a need for questionnaires validated to use in a local population.

In India, several studies have reported incidence of insomnia in the range of 13.8–33% of the population studied [16,17,18,19]. Yet, there are currently no validated questionnaires available for self-evaluation by those with insomnia. For any self-evaluation questionnaire, it is essential for the items to be written in a way that can be easily understood by most of the respondents [20]. Hence, we developed and validated a questionnaire based on the AIS but modified to enhance clarity and to be more specific to identify insomnia and to grade its severity in Indians.

2 Methods

2.1 Study Design

This prospective, comparative, multicentre study consisting of two phases was conducted from April 2019 to November 2019. Phase 1 of the study was conducted at 1 centre in Delhi and phase 2 was conducted at 4 centres (3 centres in Delhi; 1 centre in Mumbai).

In phase 1 of the study (face validity), the understandability and acceptability of the insomnia questionnaire modified from AIS (modified insomnia questionnaire [MIQ]) was assessed.

The AIS, used as the standard in this study, is a validated, self-assessment psychometric instrument designed for quantifying sleep difficulty based on the 10th revision of the International Statistical Classification of Diseases and Related Health Problems (ICD-10) criteria. The AIS-8, or the full version, consists of eight questions, of which the first five are related to nocturnal sleep, and the last three to daytime dysfunction. Each question is rated on a 0–3 scale (with 0 corresponding to “no problem at all” and 3 to “very serious problem”), and sleep is finally evaluated from the cumulative score of the factors (the total score ranges from 0 [denoting absence of any sleep-related problem] to 24 [representing the most severe degree of insomnia]). A total score of ≥ 6 is considered to indicate insomnia [10].

In comparison with the AIS, the MIQ was made more comprehensive with ten questions instead of eight. The additional questions in MIQ include the total duration of sleep complaint, keeping in mind the ICD-3 classification (question 5), and types of sleep aids used (question 10) (Table 1a). The eight questions from AIS were also rephrased to be more specific and thereby increase their understanding (e.g. “Final awakening earlier than desired” in AIS modified to “Do you wake up earlier then desired”). And the Likert scale for response categories was modified in MIQ to be more specific and was described in terms of number of days [(1) none (2) once a week (3) 2–4 days a week (4) 5–7 days a week]. All questions in the MIQ were in English.

Table 1 (a) Modified insomnia questionnaire (MIQ), (b) Indian Insomnia Rating Scale (IIRS)
Full size table

In this phase, six participants with insomnia and six healthy controls were included. All participants were asked to fill MIQ at the start of the study, and after 1 week they were asked to fill the AIS [10, 11]. The participant responses to both questionnaires were assessed, and based on their responses, MIQ was modified to further enhance its ease of understanding and use, and the final version (Indian Insomnia Rating Scale [IIRS]) was arrived at Table 1b.

In phase 2, IIRS was administered in a larger participant population (104 participants with insomnia and 34 healthy controls) to assess reliability, validity, sensitivity and specificity. In this phase, participants were asked to fill both IIRS and AIS at study start (first evaluation) and again after 1 week (second evaluation). The participants’ identity was masked by identifying them in the database using only a unique identification number and not including their names.

In IIRS, each question is rated on a 0–3 scale (with 0 corresponding to “no insomnia” and 3 to “severe insomnia”), and insomnia was evaluated from the cumulative score of the factors. A total score from all 10 questions of 0–4 denotes “no insomnia”, a score of 5–10 denotes “mild insomnia’, a score of 11–20 denotes “moderate insomnia” and a score of 21–30 denotes “severe insomnia”.

The study endpoints were the difference in mean scores between IIRS and AIS, and the difference in mean scores between test and retest of IIRS.

The study protocol and its amendments were approved by an independent ethics committee. The study was performed in accordance with the protocol, ethical principles originating from the Declaration of Helsinki, which were consistent with International Conference on Harmonization Good Clinical Practices, and applicable government regulations. All participants provided written informed consent before entering the study.

2.2 Participants

Individuals with insomnia included in the study were between 18 and 60 years of age, of either sex, having difficulty in initiating or maintaining sleep as per ICSD-3 criteria [1] which lasted for either 3 days per week for three months or less than 3 months or for unspecified period. Healthy adults without any sleep disorder were included as controls.

Individuals excluded were those with cognitive impairment or serious illness and those on concomitant medications such as anticonvulsants (phenytoin and lamotrigine), beta-blockers (acebutolol, atenolol, etc.), antipsychotics (sulpiride), antidepressants (selective serotonin reuptake inhibitors or monoamine oxidase inhibitors) and non-steroidal anti-inflammatory drugs (indomethacin, diclofenac, naproxen, and sulindac) that are likely to cause insomnia.

2.3 Statistical Methods

Based on guideline recommendations [21], and assuming at least eight items in the questionnaire that are each rated on a scale of 0–3, the minimum sample size required was 100.

In phase 1, the responses to AIS and MIQ were correlated, and necessary corrections were made to IIRS. In phase 2, the parameters assessed were: (1) reliability, tested by measuring internal consistency and test–retest reliability; (2) test of validity; (3) comparison of means for different sample groups; and (4) sensitivity and specificity.

For assessing internal consistency, questions from AIS and IIRS that measured the same criteria were loaded onto same factors and combined and compared using Cronbach’s alpha test. A Cronbach’s alpha > 0.7 for 10 items or less and CA > 0.8 for 11–30 items is considered to indicate adequate internal consistency. Test–retest reliability was evaluated using Pearson’s product moment correlation coefficient (Pearson’s r varies between 0 and 1, and larger stability coefficient indicates stronger test–retest reliability).

Validity was measured by calculating Pearson’s product–moment correlation coefficient (Pearson’s r) between the final added score of all questions with each questions’ response. The calculated correlation value was compared with a standard table of “Pearson’s r” with n-2 degree of freedom, where n = total sample size. If the “Pearson’s r” was ≥ critical value, then the question was considered as valid, or else it was considered as statistically invalid. In the current study, based on the final sample size of 138, the degree of freedom was 136 and hence applicable critical value was 0.1946 at 5% level of significance in a two-tailed test or non-directional test.

The means at different time points for responses from the same individual was compared using Wilcoxon signed-rank test at a 5% level of significance, to evaluate the ability of IIRS to discriminate between participants with or without insomnia. The sensitivity and specificity analysis determined the external validity of IIRS measured against the total score of AIS.

3 Results

All 150 individuals who participated in the study, completed the study. All participants in both phase 1 and phase 2 of the study were proficient in English and were able to understand and respond to the questions in English in the MIQ.

In phase 1, the mean (SD) age of participants was 35.83 (12.55) years, and 58% were male (Table 2). Comorbidities observed were liver disease and thyroid disorder in one participant each. Alcohol consumption and smoking were observed in two participants.

Table 2 Participants demographic and baseline characteristics
Full size table

In phase 2, the mean age was 40.55 (11.81) years, and 52% were male (Table 2); most (61.59%) had a graduate degree.

3.1 Phase 1

The mean score for MIQ and AIS was higher for participants with insomnia, as expected, than for healthy controls (Fig. 1).

Fig. 1
figure 1

Mean scores for MIQ and AIS

Full size image

For face validity, the participant responses to MIQ (Table 1a) and to AIS were compared. Based on their responses, MIQ was modified to arrive at the final version, IIRS (Table 1b). The modifications to questions in MIQ related to better capturing the frequency and/or duration of events. For example, the question in MIQ of “Do you feel difficulty in falling asleep?” was changed in IIRS to “How often do you find difficulty in falling asleep?”, which improved the clarity of the question. Similarly, modifications were made to the other questions in MIQ to enhance clarity, and the resultant changes in IIRS are highlighted in Table 1b. In both MIQ and IIRS, the response options provided for the individual questions generally remained the same, except the duration of insomnia for a score of 3 was changed from “5–7 days a week” in MIQ to “4–7 days a week” in IIRS.

3.2 Phase 2

The mean (SD) scores for each individual question in IIRS remained nearly consistent between first and second evaluation, for both participants with insomnia and healthy controls (Fig. 2).

Fig. 2
figure 2

Mean IIRS individual question scores. a First evaluation. b Second evaluation. Second evaluation occurred 1 week after the first evaluation

Full size image

At first evaluation, the mean scores for each individual question ranged between 0.24 and 0.59 for healthy controls and between 1.56 and 2.39 for participants with insomnia; at the second evaluation/retest, the mean scores for each question ranged between 0.26 and 0.59 for healthy controls and between 1.14 and 1.97 for those with insomnia. Similarly, the mean (SD) total score of IIRS and AIS was consistent between first and second evaluation, for both participants with insomnia and healthy controls (Table 3).

Table 3 Mean (SD) total IIRS and AIS scores
Full size table

The proportion of participants with mild, moderate, and severe insomnia was 10.48%, 51.43% and 38.10%, respectively, at first evaluation and was 15.00%, 67.00% and 18.00%, respectively, at second evaluation. The mean (SD) scores for IIRS and AIS increased with increasing severity of insomnia (Table 4).

Table 4 Mean (SD) total IIRS and AIS scores, by insomnia grade
Full size table

3.2.1 Internal Consistency

The Cronbach’s alpha for IIRS was 0.959 at first evaluation and 0.934 at second evaluation. The reliability measure for the internal consistency value (Cronbach’s alpha), based on standardized items was 0.961 and 0.939 for first and second evaluation, respectively, indicating excellent internal consistency. The Cronbach’s alpha value obtained by deleting each of the 10 questions from IIRS at first and second evaluation, indicated the importance of each question in the questionnaire, and also showed that the alpha value did not majorly vary if any item from the questionnaire was deleted (Table 5).

Table 5 Cronbach’s alpha score obtained after deletion of each of the questions in IIRS
Full size table

3.2.2 Test–retest Reliability

The mean total score for test–retest reliability for each of the questions in IIRS ranged from 1.23 to 1.88 at first evaluation and from 0.93 to 1.57 at second evaluation (Fig. 3).

Fig. 3
figure 3

Mean scores for test–retest reliability of IIRS

Full size image

The values of item-item correlation between first and second evaluation ranged from 0.674 to 0.749 with mean of 0.712 (0.0215), indicating acceptable reliability. Item-total correlation ranged from 0.694 to 0.869 with mean of 0.826 (0.0483) for first evaluation (good reliability) and from 0.535 to 0.800 with mean of 0.752 (0.0759) for second evaluation (acceptable reliability).

3.2.3 Test of Validity

The final scores for each of the questions in IIRS ranged from 0.757 to 0.898 at first evaluation and between 0.616 and 0.841 at second evaluation. Since the scores for all questions at first and second evaluation were higher than the calculated critical value of 0.1946 (details of critical value calculation in statistical methods section above), all questions in IIRS were considered as valid for inclusion.

3.2.4 Sensitivity and Specificity

IIRS was observed to have 100% sensitivity (all participants with insomnia were correctly identified as such) at first and second evaluation in the overall population, and when graded by severity of insomnia. The specificity was 95% at first evaluation (6 healthy controls were categorised under insomnia based on responses) and 90% at second evaluation (11 healthy controls were categorised under insomnia). By insomnia grading, specificity of IIRS was 79%, 95%, and 100% in mild, moderate, and severe category, respectively, at first evaluation (three healthy controls each were categorised under mild and moderate insomnia) and was 65%, 96%, and 100%, respectively, at second evaluation (eight and three healthy controls were categorised under mild and moderate insomnia, respectively).

4 Discussion

Patient reported outcome measurements are becoming increasingly important in clinical practice as it helps physicians identify patient concerns and facilitates patient involvement in healthcare decisions. However, prior to using these instruments in clinical practice, it is essential that these instruments are validated, by analysing whether the questionnaire measures what it is intended to measure.

The results from this study demonstrated excellent internal consistency for IIRS, based on Cronbach’s alpha values of > 0.9 at first and second evaluation. These values also point towards the very high level of homogeneity of the questionnaire, which is further supported by the proof that Cronbach’s alpha does not vary much if any item from the questionnaire is deleted, and by the high correlation coefficients of the item-total correlations. The test–retest reliability of the questionnaire was also within the acceptable range. The sensitivity of the questionnaire was 100% at first and second evaluation, indicating the questions in the questionnaire adequately and accurately capture the condition of insomnia. The specificity was > 90% at both evaluations in the total participants with insomnia; it was 100% in those with severe insomnia at both evaluations, and in those with moderate insomnia, it was 95% and 96%, respectively, at the first and second evaluation.

The AIS, chosen as the standard in this study is based on the ICD-10 criteria, and is a comprehensive tool as it assesses all top four sleep domains considered most important (sleep adequacy, sleep maintenance, sleep initiation, daytime functioning), and is also ranked high on content validity and feasibility, considering it has only eight items [10, 22,23,24]. At a score of 6 or higher, the scale presents with 93% sensitivity and 85% specificity for diagnosis of insomnia. At a cutoff score of 6, in the general population, the scale has a positive predictive value of 41% and a negative predictive value of 99% [11]. Similar to AIS, IIRS also addresses the important sleep domains, but is made more comprehensive by addition of two more questions, one related to total duration of sleep complaints and the other related to use of sleeping aids. Also, the questions in AIS were rephrased in MIQ and further refined in IIRS to enable better understanding in an Indian context.

In general, studies in India have highlighted the lack of awareness regarding insomnia or sleep-related problems among the general population. Panda et al. observed a low perception of sleep related problems among the study participants, reflecting on the poor knowledge and awareness about sleep disorders and their health-related negative consequences in India [16]. Yardi et al. [18] observed that among office employees diagnosed with insomnia, nearly 96.4% of them had been previously undiagnosed highlighting the lack of awareness among the general population. Hence, there is a need for creating better awareness among the general population as well as developing instruments specific to Indian settings for better diagnosis of insomnia.

5 Conclusions

The newly developed insomnia questionnaire, IIRS, demonstrated excellent internal consistency, test–retest reliability, validity, sensitivity and specificity for the diagnosis of insomnia. Results suggest the new questionnaire to be an acceptable and valid instrument to diagnose insomnia and to measure its severity in Indian clinical and research settings.