Trial Outcomes & Findings for Oregon Military Employee Sleep and Health Study (NCT NCT02946736)

n=215 were identified as supervisors in the intervention condition and were excluded (i.e., focus is on service members only for analysis)

Missing data

Total number of hours calculated from reported bed time and wake time. Minimum 0 hours, Maximum 24 hours. Longer duration indicates longer sleep duration. Ideal range is 7-9 hours of sleep per night.

Total number of hours calculated from reported bed time and wake time. Minimum 0 hours, Maximum 24 hours. Longer duration indicates longer sleep duration. Ideal range is 7-9 hours of sleep per night.

Four item subscale of the larger 8 item Sleep Disturbance scale. Likert-type responses: 1 = Not at all to 5 = Very much). T-Scores were created using HealthMeasures Scoring service, where data are uploaded into the HealthMeasures scoring service website, which generates T-Scores. This is considered the most accurate option. T-Scores estimates were developed using a population of 2,252 participants, 259 of which had clinical sleep disorders, and was intended to be representative of the US population. This is the reference population that the T-score means and SDs were calibrated and centered with. The T-Scores that HealthMeasures generates are not precisely at a mean of 50 and standard deviation of 10 because they are based on the unique MESH samples. A higher PROMIS T-score represents more of the concept being measured (worse).

Four item subscale of the larger 8 item Sleep Disturbance scale. Likert-type responses: 1 = Not at all to 5 = Very much). T-Scores were created using HealthMeasures Scoring service, where data are uploaded into the HealthMeasures scoring service website, which generates T-Scores. This is considered the most accurate option. T-Scores estimates were developed using a population of 2,252 participants, 259 of which had clinical sleep disorders, and was intended to be representative of the US population. This is the reference population that the T-score means and SDs were calibrated and centered with. The T-Scores that HealthMeasures generates are not precisely at a mean of 50 and standard deviation of 10 because they are based on the unique MESH samples. A higher PROMIS T-score represents more of the concept being measured.

Four item subscale of the larger 8 item Sleep Disturbance scale. Likert-type responses: 1 = Not at all to 5 = Very much. T-Scores were created using HealthMeasures Scoring service, where data are uploaded into the HealthMeasures scoring service website, which generates T-Scores. This is considered the most accurate option. T-Scores estimates were developed using a population of 2,252 participants, 259 of which had clinical sleep disorders, and was intended to be representative of the US population. This is the reference population that the T-score means and SDs were calibrated and centered with. The T-Scores that HealthMeasures generates are not precisely at a mean of 50 and standard deviation of 10 because they are based on the unique MESH samples. A higher PROMIS T-score represents more of the concept being measured (worse).

Four item subscale of the larger 8 item Sleep Disturbance scale. Likert-type responses: 1 = Not at all to 5 = Very much). T-Scores were created using HealthMeasures Scoring service, where data are uploaded into the HealthMeasures scoring service website, which generates T-Scores. This is considered the most accurate option. T-Scores estimates for Dissatisfaction with Sleep were developed using a population of 2,252 participants, 259 of which had clinical sleep disorders, and was intended to be representative of the US population. This is the reference population that the T-score means and SDs were calibrated and centered with. The T-Scores that HealthMeasures generates are not precisely at a mean of 50 and standard deviation of 10 because they are based on the unique MESH samples. A higher PROMIS T-score represents more of the concept being measured (worse).

8 item scale. Likert-type responses: 1 = Not at all to 5 = Very much). T-Scores were created using HealthMeasures Scoring service, where data are uploaded into the HealthMeasures scoring service website, which generates T-Scores. This is considered the most accurate option. T-Scores estimates for SRI were developed using a population of 2,252 participants, 259 of which had clinical sleep disorders, and was intended to be representative of the US population. This is the reference population that the T-score means and SDs were calibrated and centered with. The T-Scores that HealthMeasures generates are not precisely at a mean of 50 and standard deviation of 10 because they are based on the unique MESH samples. A higher PROMIS T-score represents more sleep impairment (worse).

8 item scale. Likert-type responses: 1 = Not at all to 5 = Very much). T-Scores were created using HealthMeasures Scoring service, where data are uploaded into the HealthMeasures scoring service website, which generates T-Scores. This is considered the most accurate option. T-Scores estimates for SRI were developed using a population of 2,252 participants, 259 of which had clinical sleep disorders, and was intended to be representative of the US population. This is the reference population that the T-score means and SDs were calibrated and centered with. The T-Scores that HealthMeasures generates are not precisely at a mean of 50 and standard deviation of 10 because they are based on the unique MESH samples. A higher PROMIS T-score represents more sleep impairment (worse).

Actigraphic measurements obtained using Actiwatch2 worn for 3 weeks Average sleep duration in hours Great duration generally better, with target range of 7-9 hours each sleep period.

Actigraphic measurements obtained using Actiwatch2 worn for 3 weeks Average number of minutes spent awake during the sleep period. More minutes indicates worse outcome.

Service member satisfaction with current job; Likert-type scale 1 = Strongly disagree to 5 = Strongly agree, mean created from the 3 items. Scores could range from 0-5 with higher levels indicating greater satisfaction. Source: Cammann et al., 1983

Service member intention to quit current job; Two items with responses on Likert-type scale 1 = Strongly disagree to 5 = Strongly agree. Overall score created with a mean of the two items, with higher levels indicating greater intention to leave one's job. Source: Boroff \& Lewin, 1997

Service member self-reported stress; Likert-type scale 0 = Never to 4 = Very often, combined to a mean score, with a possible range from 0 to 4. Higher scores indicate greater stress (worse). Source: Cohen \& Williamson, 1988

Service member difficulty with completing work tasks and quality; 6 items with Likert-type responses: 1 = No difficulty at all to 5 = Extreme difficulty, mean score created from all items, with possible score range from 1 to 5. Higher levels indicating greater impairment (worse). Source: Subscale of Overall Functional Impairment by Herrell et al., 2014

Service member difficulty with getting personal life skills completed; 2 items with Likert-type responses: 1 = No difficulty at all to 5 = Extreme difficulty, mean score created from all items, with possible score range from 1 to 5. Higher levels indicating greater impairment (worse). Source: Subscale of Overall Functional Impairment by Herrell et al., 2014

Service member difficulty functioning in social situations; 4 items with Likert-type responses: 1 = No difficulty at all to 5 = Extreme difficulty, mean score created from all items, with possible score range from 1 to 5. Higher levels indicating greater impairment (worse). Source: Subscale of Overall Functional Impairment by Herrell et al., 2014

Perceived supervisor support for work-family integration as reported by service member, 4 items, with responses on Likert-type scale 1=Strongly disagree to 5= Strongly agree, overall score created by mean score, with possible scores ranging from 1 to 5.. Higher scores indicating higher levels of support (better) Source: Hammer et al., 2013

Perceived supervisor support for sleep health as reported by service member, 8 items, with responses on Likert-type scale 1=Never to 5= Always, overall score created by mean score, with possible scores ranging from 1 to 5.. Higher scores indicating higher levels of support (better) Source: Modified version of Gunia et al., 2015