The Development of a Brief Version of the Insomnia Severity Index (ISI-3) in Older Adult Veterans with Posttraumatic Stress Disorder

Data-driven shortened Insomnia Severity Index (ISI): a machine learning approach

BACKGROUND

The Insomnia Severity Index (ISI) is a widely used questionnaire with seven items for identifying the risk of insomnia disorder. Although the ISI is still short, more shortened versions are emerging for repeated monitoring in routine clinical settings. In this study, we aimed to develop a data-driven shortened version of the ISI that accurately predicts the severity level of insomnia disorder.

METHODS

We collected a sample of 800 responses from the EMBRAIN survey system. Based on the responses, seven items were grouped based on the similarity of their response using exploratory factor analysis (EFA). The most representative item within each group was selected by using eXtreme Gradient Boosting (XGBoost).

RESULTS

Based on the selected three key items, maintenance of sleep, interference with daily function, and concerns about sleep problems, we developed a data-driven shortened questionnaire of ISI, ISI-3 m (machine learning). ISI-3 m achieved the highest coefficient of determination (R 2 = 0.910 ) for the ISI score prediction task and the accuracy of 0.965, precision of 0.841, and recall of 0.838 for the multiclass-classification task, outperforming four previous versions of the shortened ISI.

CONCLUSION

As ISI-3 m is a highly accurate shortened version of the ISI, it allows clinicians to efficiently screen for insomnia and observe variations in the condition throughout the treatment process. Furthermore, the framework based on the combination of EFA and XGBoost developed in this study can be utilized to develop data-driven shortened versions of the other questionnaires.

Psychometric Properties of the Insomnia Severity Index and Its Comparison With the Shortened Versions Among the General Population

OBJECTIVE

The aim of this study was to explore the psychometric properties of the Insomnia Severity Index (ISI) based on modern test theory, such as item response theory (IRT) and Rasch analysis, with shortened versions of the ISI among the general population.

METHODS

We conducted two studies to evaluate the reliability and validity of the shortened versions of the ISI in a Korean population. In Study I, conducted via online survey, we performed an exploratory factor analysis (n=400). In Study II, confirmatory factor analysis (CFA) was conducted (n=400). IRT and Rasch analysis were performed on all samples. Participants symptoms were rated using the ISI, Dysfunctional Beliefs and Attitudes about Sleep-16 items, Dysfunctional Beliefs about Sleep-2 items, Patient Health Questionnaire-9 items, and discrepancy between desired time in bed and desired total sleep time.

RESULTS

CFA showed a good fit for the 2-factor model of the ISI (comparative fit index=0.994, Tucker-Lewis index=0.990, root-meansquare-error of approximation=0.039, and standardized root-mean-square residual=0.046). The 3-item versions also showed a good fit for the model. All scales showed good internal consistency reliability. The scale information curve of the 2-item scale was similar to that of the full-scale ISI. The Rasch analysis outputs suggested a good model fit.

CONCLUSION

The shortened 2-factor ISI is a reliable and valid model for assessing the severity of insomnia in the Korean population. The results are needed to be explored further among the clinical sample of insomnia.

The simplification of the insomnia severity index and epworth sleepiness scale using machine learning models

Insomnia and excessive daytime sleepiness (EDS) are the most common complaints in sleep clinics, and the cost of healthcare services associated with them have also increased significantly. Though the brief questionnaires such as the Insomnia Severity Index (ISI) and Epworth Sleepiness Scale (ESS) can be useful to assess insomnia and EDS, there are some limitations to apply for large numbers of patients. As the researches using the Internet of Things technology become more common, the need for the simplification of sleep questionnaires has been also growing. We aimed to simplify ISI and ESS using machine learning algorithms and deep neural networks with attention models. The medical records of 1,241 patients who examined polysomnography for insomnia or EDS were analyzed. All patients are classified into five groups according to the severity of insomnia and EDS. To develop the model, six machine learning algorithms were firstly applied. After going through normalization, the process with the CNN+ Attention model was applied. We classified a group with an accuracy of 93% even with only the results of 6 items (ISI1a, ISI1b, ISI3, ISI5, ESS4, ESS7). We simplified the sleep questionnaires with maintaining high accuracy by using machine learning models.

Validation of the Three-Item Insomnia Severity Index Short Form in Young Adult Cancer Survivors: Comparison with a Structured Diagnostic Interview

Chronic insomnia affects ∼25% of young adult cancer survivors (YACS) but is often overlooked in routine follow-up. A recently introduced three-item version of the Insomnia Severity Index (ISI-3) was compared with a diagnostic interview (SCID-5) in 250 YACS (ages 18-40) to evaluate its validity in this population. The ISI-3 had good discrimination compared with the SCID-5 (area under the receiver operating characteristic curve = 0.88). Although no ISI-3 cutoff met study criteria for both sensitivity (≥0.85) and specificity (≥0.75), an ISI-3 cutoff of ≥4 had high sensitivity (94%) and moderate specificity (70%), and is recommended as the first step in a two-step screening procedure.

Validation of insomnia questionnaires in the general population: The Nord-Trøndelag Health Study (HUNT)

The primary aim was to validate questionnaire-based insomnia diagnoses from a modified Karolinska Sleep Questionnaire (KSQ) and the Insomnia Severity Index (ISI), by age category (< or >65 years), against a semi-structured face-to-face interview. Secondary aims were to split validity by diagnostic certainty of the interview and to compare prevalence estimates of questionnaire- and interview-based diagnoses. A total of 232 out of 1,200 invited (19.3%) from the fourth Nord-Trøndelag Health Study (HUNT4) completed questionnaires, including the KSQ and ISI, shortly before attending a face-to-face diagnostic interview for insomnia based on the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5). Both a tentative (DSM-5 criteria A-E) and a definite (criteria A-H) interview diagnosis was evaluated. Cohen's kappa statistic quantified questionnaire validity. In all, 33% (95% confidence interval 27-39%) of participants had definite insomnia: 40% of women and 21% of men. The ISI (cut-off 12) and several KSQ-based diagnoses showed very good validity (κ ≤0.74) against the tentative, versus good validity (κ ≤0.61) against the definite interview diagnosis. Short questionnaires, requiring a daytime symptom at least three times a week, may underestimate insomnia prevalence. Validity was consistently higher for persons aged below versus above 65 years (definite insomnia: κ ≤0.64 vs. κ ≤0.56). Our results have implications for epidemiological population-based studies utilising insomnia questionnaires.