Sleep and substance use disorder treatment: A preliminary study of subjective and objective assessment of sleep during an intensive outpatient program

Abstract

Background and Objectives:

Characteristics of sleep concerns and their relationship to mental health in heterogeneous substance use disorder (SUD) treatment settings are not well understood. The purpose of this preliminary study was to assess sleep using subjective and objective measures at two time points during SUD treatment and compare sleep changes to changes in mental health measures.

Methods:

Thirty adults completed Time 1 (the beginning of treatment) and 22 returned for Time 2 (approximately four weeks later). Majority of participants were white (80%), male (63%), and presenting for alcohol use disorder (60.0%), though almost half reported polysubstance abuse (43%). Comorbidity was common (53%). Sleep and mental health questionnaires with one week of actigraphy and sleep diaries were completed at both time points.

Results:

Most participants met criteria for a sleep disorder and mean scores on questionnaires showed poor sleep quality, insomnia symptoms, and frequent nightmares, with sleep quality and insomnia improving over time but remaining clinically significant. Nightmares did not improve. Actigraphy indicated poor sleep at both time points. Improvement in insomnia was related to improvement in measures of mental health while changes in actigraphy variables were not related to these measures.

Discussion and Conclusions:

Multiple types of sleep disturbance are prevalent in this population, with nightmares persisting throughout treatment and insomnia symptoms showing a relationship with mental health symptoms.

Scientific Significance:

This was the first study to longitudinally assess mental health with subjective and objective measures of sleep across multiple types of SUDs in a community SUD treatment setting.

Sleep disturbance during withdrawal and early treatment for substance use disorders (SUDs) is a common concern across all SUDs and has been linked to important treatment outcomes, including SUD treatment completion and substance use relapse^1,2. Information about the presentation and progression of sleep concerns during this period has largely come from studies focused on a single SUD type, including alcohol, opioids, cannabis, and stimulants³. Many of these had varying methodology with differing sleep targets (e.g., insomnia, sleep quality, sleep duration, nightmares), which makes it difficult to determine what, if any, aspects of sleep disturbance may be common across SUD types. Further, much of the information on sleep in SUD treatment comes from randomized clinical trials⁴ which often have strict inclusion criteria and may not be generalizable to heterogeneous community treatment programs. The need for better understanding the types of sleep disturbance in such a population is important because these are the most common treatment settings for SUDs⁵ and targeting the most prevalent and problematic aspects of sleep across SUDs could have a powerful impact on treatment outcomes.

There is also a high prevalence of psychiatric comorbidity in SUD treatment¹, though research examining both sleep and mental health (e.g., mood, anxiety, and posttraumatic stress disorder [PTSD]) in the same study is sparse, and again this is typically examined in the context of one SUD type. Studies utilizing validated sleep self-report measures alongside measures of mood and anxiety in treatment-seeking adults with alcohol⁶, cannabis⁷, cocaine⁸, methamphetamine⁹, and opioid use disorders¹⁰ demonstrated improvements in scores in all three areas throughout the first few weeks of treatment. However, these improvements often were not statistically significant or clinically meaningful (i.e., scores remained in the clinically significant range after improvement). These studies relied solely on subjective assessment of sleep and did not examine how changes in sleep related to changes in mood and anxiety over time. Two studies used self-report measures of nightmares and/or insomnia in a treatment program targeting both PTSD and SUD symptoms. Similar to the studies focused on mood and anxiety, there was improvement in PTSD and sleep, though statistical significance and clinical meaningfulness varied^11,12. One of these studies extended analyses to examine the relationship between improvement in PTSD and sleep over time and found they were significantly related¹². Most recently, Colvonan and colleagues used self-report measures and actigraphy (a wearable measuring rest and activity) with persons receiving treatment for PTSD and alcohol use disorder. PTSD and insomnia symptoms both significantly decreased, moving from the moderate range to the mild range¹³. Interestingly, most actigraphy variables did not significantly change. Additionally, when they examined changes over time, they found a reduction in PTSD symptoms was significantly related to reduction in insomnia symptoms but was not related to change in any actigraphy variables. The authors propose PTSD may be more strongly associated with negative perception of sleep than behavioral indicators of sleep, thus as PTSD improved so did perceivedn insomnia symptoms.

The few studies that have examined sleep in heterogeneous community treatment settings where individuals commonly have multiple SUDs and all types of SUDs are treated together have found high prevalence of self-reported sleep disturbance via sleep questionnaires such as the Pittsburgh Sleep Quality Index (PSQI)¹⁴ or the Insomnia Severity Index (ISI)^1,11,15–17. These findings are important but limited by lack of objective measures and longitudinal assessment of sleep and mental health symptoms over time.

In this preliminary study we sought to expand previous findings by utilizing a multidimensional approach to assess the relationship between sleep and mental health in persons with a variety of SUD types who were engaged in early treatment in a community setting. Specifically, measures were collected at the beginning of an intensive outpatient (IOP) program (Time 1) and approximately four weeks later (Time 2, the average amount of time it takes to complete a course of treatment). The PSQI, ISI, and Nightmare Disorder Index (NDI)¹⁸, were administered to subjectively assess sleep, and actigraphy was used to objectively measure quantitative sleep parameters. Primary actigraphy variables of interest were those that are used to indicate good sleep quality^19,20: total sleep time (TST, total amount of sleep duration during the major sleep period), sleep onset latency (SOL, amount of time it takes to fall asleep), wake after sleep onset (WASO, amount of time awake during the major sleep period after first falling asleep), and sleep efficiency (SE, percentage of time spent asleep of the total amount of time spent in bed). Secondary actigraphy variables that are often used to characterize insomnia¹⁷ and sleep timing (an indicator of overall sleep health)²¹ were also collected: bedtime (BT, time of day one gets in bed for the major sleep period), rise time (RT, time of day one gets out of bed after the major sleep period), terminal wakefulness (TWAK, amount of time between final awakening of the major sleep period and RT), and time in bed (TIB, total amount of time between BT and RT). The Patient Health Questionnaire 9-item scale (PHQ-9)²², Generalized Anxiety Disorder 7-item scale (GAD-7)²³, and PTSD Checklist for the DSM-5 (PCL-5)²⁴ were administered to assess depression, anxiety, and PTSD, respectively.

The first hypothesis of this study was sleep disturbance would be highly prevalent at Time 1, defined as follows: 1) The majority of participants would meet criteria for a sleep disorder as determined by the Structured Clinical Interview for Sleep Disorders – Revised Edition (SCISD-R)²⁵; 2) Average scores on the PSQI, ISI, and the NDI would indicate clinically significant sleep disturbance; 3) Actigraphy and sleep diary measures of quantitative sleep parameters would be worse than those recommended for good sleep quality in the general population (TST < 7 hours, SOL > 30 minutes, WASO > 20 minutes, SE < 85%)^19,20. The second hypothesis of this study was PSQI, ISI, NDI, PHQ-9, GAD-7, and PCL-5 scores would significantly improve from Time 1 to Time 2, though remain clinically significant, and actigraphy and sleep diary parameters would not improve. The third hypothesis was improvement in PSQI, ISI, and NDI would be significantly related to improvement in PHQ-9, GAD-7, PCL-5 from Time 1 to Time 2, though primary actigraphy/sleep diary variables would not.

Methods

Materials

SCISD-R.

The SCISD-R is a brief, valid interview assessment of adult sleep disorders as defined by the DSM-5. It has been demonstrated to have good to excellent reliability for the diagnosis of insomnia, hypersomnolence, obstructive sleep apnea-hypopnea, circadian rhythm sleep-wake, nightmare, and restless legs syndrome disorders.

PSQI.

The PSQI is a 19-item questionnaire that assesses subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleep medications, and daytime dysfunction. The PSQI generates seven scores that correspond to the seven domains previously mentioned. The domain scores range from 0 (no difficulty) to 3 (severe difficulty) and combined produce a global score (0 – 21) with a score greater than 5 considered to be suggestive of significant sleep disturbance.

ISI.

The ISI is a 7-item self-report measure that assesses perceived severity of insomnia. Each item uses a Likert scale from 1 (not at all satisfied) to 5 (very much satisfied). The items are summed, yielding a minimum score of 7 and a maximum score of 35, with 8, 15, and 22 indicating subthreshold, moderate, and severe insomnia, respectively.

NDI.

The NDI is a 5-item questionnaire that assesses nightmare frequency and severity. The items are summed yielding a range of scores from 0 to 20, with higher scores indicating greater nightmare severity.

Actigraphy.

Actigraphy is a method for measuring sleep and activity patterns. In this study, sleep and wake patterns were measured using Respironics® Actiwatch Spectrum®²⁶, which are wrist-worn, battery-operated activity monitors that look similar to a small wristwatch. The Actiwatch utilizes a motion sensor known as an “accelerometer” to monitor the occurrence and degree of motion, Actigraphy is highly correlated with polysomnography (PSG), the gold standard for objectively measuring sleep²⁷. Epoch length was set to 30 seconds, wake threshold was set to 40 seconds, and participants were asked to push an event marker on the watch at two time points: 1) when they attempted to go to sleep for the night; 2) when they got out of bed for the day. Each variable of interest (TST, SOL, WASO, SE, BT, RT, TWAK, TIB) was calculated for each night, then averaged across all nights for that time point.

Consensus Sleep Diary (CSD).

The CSD²⁸ was developed by 25 experts to standardize prospective sleep self-monitoring. It contains 9 items that are considered the most critical parameters of self-reported sleep and are used to calculate TST, SOL, WASO, SE, BT, RT, TWAK, and TIB.

Timeline Follow-Back: 7 days (TLFB).

The TLFB²⁹ is a quantitative estimate of drug use-consumption variables that can be used to measure change in substance use via a brief interview It involves asking individuals their substance use prior to the interview date.

PHQ-9.

The PHQ-9 measures the severity of depression. Scores of each of the 9 items range from 0 (not at all) to 3 (nearly every day), yielding a minimum score of 0 and maximum of 27, with total scores of 5, 10, 15, and 20 suggesting mild, moderate, moderately severe, and severe depression, respectively.

GAD-7.

The GAD-7 assesses anxiety symptoms experienced during the previous 14 days. Each item uses a 4-point Likert scale from 0 to 3, yielding a minimum score of 0 and maximum score of 21, with higher scores more indicative of anxiety and a cut-off score of 10 suggesting clinically significant anxiety.

PCL-5.

The PCL-5 is a 20-item self-report measure that assesses the 20 DSM-5 symptoms of PTSD. Each item uses a 4-point Likert scale from 0 to 80, yielding a minimum score of 0 and maximum score of 80. A cut-off score of 31 to 33 suggests clinically significant posttraumatic stress disorder symptoms.

Procedure

Thirty participants were recruited from the IOP in the Center for Drug and Alcohol Programs (CDAP) at Medical University of South Carolina. The IOP provides treatment for individuals who meet criteria for at least one SUD and have completed detoxification or do not require hospitalization. The IOP program has been described in detail previously¹. Participants were considered for inclusion if they were at least 18 years old, diagnosed with at least one SUD during CDAP intake, able to read or speak English, and not currently experiencing psychosis or suicidal ideation. During the first week of IOP participants who were consented into the study attended an interview with a licensed psychologist who administered a structured interview for major mental health diagnoses, the SCISD-R, and all questionnaires (ISI, PSQI, NDI, PHQ-9, GAD-7, PCL-5). Following this visit participants were asked to wear an actiwatch and complete daily sleep diaries for one week. This is referred to as Time 1. Approximately 4 weeks later (i.e., the typical duration of IOP treatment), participants were asked to answer the same questionnaires and complete one week of actigraphy and sleep diaries. This is referred to as Time 2. Electronic medical records were reviewed at both time points to retrieve confirmation of a previous sleep diagnosis (if needed) and confirm TLFB self-reported abstinence via clinician documentation. Participants were compensated for each time point with money on a debit card. This study was approved by the Institutional Review Board for Human Research at the Medical University of South Carolina.

Data Analysis Plan

For the first hypothesis descriptive statistics were run on the sleep variables of interest: 1) SCISD-R results; 2) total scores on PSQI, ISI, and NDI; 3) TST, SOL, WASO, and SE from actigraphy and sleep diaries. For the second hypothesis, paired samples t-tests were run comparing Time 1 to Time 2 for all sleep measures (PSQI, ISI, NDI; actigraphy/sleep diary TST, SOL, WASO, SE, BT, RT, TWAK, TIB) and mental health questionnaires (PHQ-9, GAD-7, PCL-5). For the third hypothesis, Pearson r correlations were run to examine the relationship between the change in sleep questionnaires (PSQI, ISI, NDI) and primary actigraphy/sleep diary variables (TST, SOL, WASO, SE) with change in mental health measures (PHQ-9, GAD-7, PCL-5) from Time 1 to Time 2.

Self-report measures were included in analyses if participants answered all questions in the measure. Actigraphy and sleep diary variables were included in analyses if there were at least three recorded days for the week. Actigraphy rest intervals were autoscored, using event markers rather than the auto-score algorithm when the event markers were present. Additionally, paired samples t-tests were run comparing actigraphy variables to diary variables to assess consistency in reporting. All analyses were run using IBM SPSS Statistics (Version 27).

Results

Sample Characteristics

Thirty participants were consented into the study (see Table 1 for sample characteristics). For the first hypothesis, prevalence of sleep disturbance at Time 1, 29 participants completed the SCISD-R and questionnaires (one participant left the initial appointment early and subsequently dropped out of the study), and 23 participants had at least three nights of actigraphy and sleep diary data. For the second and third hypotheses, examining changes in sleep and mental health measures over time, 22 participants completed the questionnaires and 18 participants had at least three nights of actigraphy and sleep diary data. TLFB responses indicated participants remained abstinent from illicit substances and, for those not in treatment for alcohol use disorder, drank alcohol within national recommended limits³⁰.

Table 1.

Sample characteristics

Primary substance, N (%)
Alcohol	18(60.0)
Opioid	7(23.3)
Cannabis	1(3.3)
Sedative hypnotics	1(3.3)
Stimulants	3(10.0)
Polysubstance abuse, N(%)	13(43.3)
Comorbidity, N(%)	16(53.3)
SCISD-R Sleep disorder, N(%)	19(63.3)
Age M(SD)	40.36(15.0)
Gender, N (%)
Male	19(63.3)
Female	11(36.7)
Marital Status, N(%)
Single	13(43.3)
Widowed	1(3.3)
Living with partner	3(10.0)
Married	9(30.0)
Separated/Divorced	4(13.3)
Ethnicity, N(%)
Hispanic or Latino	2(6.7)
Not Hispanic or Latino	27(90.0)
Not reported	1(3.3)
Race, N(%)
American Indian	1(3.3)
Black	3(10.0)
White	24(80.0)
More than one race	1(3.3)
Not reported	1(3.3)
Education Level, N(%)
Some high school	2(6.7)
High school (or equivalent)	1(3.3)
Some college	10(33.3)
Associate degree	4(13.3)
Bachelor degree	10(33.3)
Masters degree	3(10.0)
Household Income, M(SD)	$57,637.93($52070.20)
Employment/Student Status, N(%)
Employed full-time	12(40.0)
Employed part-time	1(3.3)
Student full-time	2(6.7)
Employed/student full-time	1(3.3)
Unemployed/non-student	12(40.0)

Note: Polysubstance abuse = documented use of one or more additional substances used regularly in the last week that have impacted functioning; Comorbidity = documented diagnosis of PTSD, mood, or anxiety disorder

Time 1 Prevalence of Sleep Disturbance

Sixty-six percent (n = 19) of participants met criteria for a sleep disorder on the SCISD-R. Of these, 16 met criteria for insomnia disorder and the other three were hypersomnia, circadian rhythm sleep-wake disorder (shift work type), and obstructive sleep apnea. Of those with insomnia, two had comorbid restless leg syndrome and two had comorbid sleep apnea. Sleep apnea was confirmed via medical record review. Mean scores indicated clinically significant sleep disturbance as measured by the PSQI (M = 10.3, SD = 3.6) and mild to moderate insomnia symptoms as measured by the ISI (M = 12.3, SD = 5.1). Twenty participants reported nightmares at least once per week on the NDI. Actigraphy showed a mean TST of 7.1 hours (SD = 2.0), with 30.2 minutes of SOL (SD = 26.2), 53.7 minutes WASO (SD = 22.4), and 78.2% SE (SD = 9.6). Sleep diary averages were not significantly different from actigraphy (p > .05) in TST (M = 6.6, SD = 1.7), SOL (M = 22.7, SD = 23.1) or SE (M = 80.6%, SD = 15.0). WASO, however, was shorter as measured by sleep diaries (M = 30.4; SD = 35.0; t[22] = 4.28, p < .001) than actigraphy. Of note, one participant slept significantly longer than the others (2 standard deviations above the next highest TST). When this person was removed average TST decreased to 6.75 hours. Other variables did not change significantly when this participant was removed.

Changes in Sleep, Mood, Anxiety, and PTSD Symptoms

Independent samples t-tests revealed no significant differences (all p > .05) on any Time 1 measures (questionnaires/actigraphy/sleep diaries) between those who completed Time 2 assessments (n = 22) and those who did not (n = 7). As seen in Table 2, questionnaire scores at Time 2 showed statistically significant improvement from Time 1 on the PSQI, ISI, PHQ-9, and GAD-7, but not on the NDI or PCL-5. Regarding actigraphy and sleep diaries, there were no significant differences between Time 1 and Time 2 on any of the sleep parameters measured. As with Time 1, the only significant difference between actigraphy and sleep diaries at Time 2 was that of WASO, with sleep diaries indicating less time awake at night than actigraphy (t[17] = 4.51, p < .001).

Table 2.

Measures at Time 1 (first week of treatment) and Time 2 (post-treatment)

Measure	Time 1 M(SD)	Time 2 M(SD)	t	p	d
Questionnaire (N = 22)
PSQI	10.5(3.7)	8.2(4.2)	2.68	0.014	0.65
ISI	12.7(5.0)	8.4(5.3)	3.04	0.006	0.57
NDI*	8.7(4.1)	8.5(4.5)	0.24	0.815	0.07
PHQ-9	11.7(6.1)	7.8(5.7)	3.37	0.003	0.72
GAD-7	9.7(5.9)	6.2(5.0)	3.16	0.005	0.67
PCL-5	29.5(17.4)	23.1(16.8)	1.80	0.087	0.38
Actigraphy (N = 18)
TST (hours)	7.1(2.0)	6.76(2.2)	1.23	0.235	0.29
SOL (minutes)	30.2(26.2)	30.40(23.5)	−0.56	0.586	0.13
WASO (minutes)	53.7(22.4)	59.18(28.8)	−0.52	0.611	0.12
SE (percentage)	78.2(9.6)	76.82(10.2)	0.48	0.640	0.11
BT (hh:mm)	11:52pm(1:41)	11:19pm(1:32)	1.46	0.164	0.34
RT (hh:mm)	7:28am(2:13)	7:44am(2:22)	−0.71	0.485	0.17
TWAK (minutes)	14.4(16.2)	22.04(15.6)	0.15	0.882	0.04
TIB (hours)	8.7(2.2)	8.4(2.3)	0.93	0.368	0.77
Sleep Diary (N = 18)
TST (hours)	6.59(1.7)	6.55(1.4)	−0.37	0.714	0.09
SOL (minutes)	22.72(23.1)	19.11(25.5)	0.95	0.354	0.22
WASO (minutes)	30.37(35.0)	23.38(25.7)	1.25	0.228	0.30
TWAK (minutes)	44.03(61.0)	44.47(81.5)	0.03	0.979	0.01
SE (percentage)	80.61(15.0)	82.53(14.7)	−0.88	0.391	0.21
TIB (hours)	8.1(1.8)	8.0(1.5)	0.31	0.762	0.07
BT	11:15pm(1:28)	11:35pm(1:49)	1.18	0.254	0.28
RT	7:21am(2:05)	7:35am(2:03)	−0.84	0.415	0.20

^*N = 12 (only those that endorsed > 1 nightmare/week at both time points)

Relationship Between Sleep, Mood, Anxiety, and PTSD Symptoms

As seen in Table 3, improvement in ISI was significantly correlated with improvement in PHQ-9, GAD-7, and PCL-5. Improvement in PSQI and NDI were also significantly correlated with improvement in PCL-5. Aside from these there were no other significant associations.

Table 3.

Correlations between change in sleep measures and depression, anxiety, and PTSD questionnaires from Time 1 to Time 2

		Questionnaires			Actigraphy (Mean)				Sleep Diary (Mean)
		PSQI	ISI	NDI	TST	SOL	WASO	SE	TST	SOL	WASO	SE
PHQ-9	Pearson r	0.412	0.437 *	0.298	0.120	−0.196	0.076	0.158	0.032	0.161	−0.116	−0.181
	Sig.	0.057	0.042	0.347	0.634	0.435	0.764	0.531	0.899	0.523	0.646	0.473
	N	22	22	12	18	18	18	18	18	18	18	18
GAD-7	Pearson r	0.060	0.511 *	0.504	0.365	−0.412	−0.036	−0.041	0.192	−0.123	−0.287	−0.048
	Sig.	0.792	0.015	0.094	0.137	0.089	0.888	0.873	0.446	0.627	0.248	0.849
	N	22	22	12	18	18	18	18	18	18	18	18
PCL-5	Pearson r	0.470 *	0.623 *	0.590 *	0.339	−0.012	−0.172	0.119	0.112	0.142	0.102	−0.438
	Sig.	0.027	0.002	0.006	0.168	0.961	0.495	0.638	0.628	0.574	0.686	0.069
	N	22	22	12	18	18	18	18	18	18	18	18

Note:

^*= p < .05

Discussion

The purpose of this study was to longitudinally assess sleep and mental health using a multi-method approach in persons receiving SUD treatment in a community setting. The first hypothesis, high prevalence of sleep disturbance at Time 1, was supported as the majority of these treatment-seeking participants met criteria for a sleep disorder, mean ISI and PSQI scores exceeded clinical cutoffs, and 66% reported at least one nightmare per week. Further, actigraphy and sleep diaries revealed TST and SOL were on the border of the minimum values recommended for good sleep quality, and WASO and SE were worse than recommended. This high prevalence of sleep disturbance across all measurements is consistent with previous studies that have used various methodologies to assess differing aspects of sleep in a single SUD type, including alcohol³¹, opioids³², cannabis³³, and stimulants³⁴.

The second hypothesis, improvement in questionnaires but not quantitative parameters from Time 1 to Time 2, was partially supported, as ISI, PSQI, PHQ-9, and GAD-7 scores improved from Time 1 to Time 2, with only GAD-7 falling below the clinical threshold. Quantitative sleep parameters (TST, SOL, WASO, SE, BT, RT, TWAK, TIB) did not improve. This is consistent with Colvonen and colleagues¹³, who found ISI scores significantly improved (but remained clinically significant) in veterans receiving treatment for PTSD and alcohol use disorder while most actigraphy variables did not. Contrary to the second hypothesis, PCL-5 did not signficantly change. This differs from previous studies that found improvement in PTSD scores in SUD treatment, however, the treatment programs for these previous studies targeted both PTSD and SUD, which is not the case in the present study. These findings may also be a floor effect, as mean scores at Time 1 indicated low severity of PTSD symptoms. Considering this low severity, it is interesting there was such a high prevalence of nightmares, which are often a prominent symptom in PTSD. NDI scores also failed to improve over time. This is similar to a study by McHugh and colleagues¹², who did not find significant changes in presence of nightmares in women receiving treatment for PTSD and SUD. This may indicate nightmares are a prevalent and residual symptom requiring more tailored intervention in this population.

The third hypothesis, changes in mental health questionnaires would be related to changes in sleep questionnaires but not quantitative sleep parameters, was partially supported. The ISI was the only sleep questionnaire associated with all three mental health measures (PHQ-9, GAD-7, PCL-5), while the PSQI and NDI were only correlated with the PCL-5. This is again comparable to the study of Colvonen and colleagues¹³ that found a reduction in PTSD symptoms was related to improvement on the ISI. These findings point to insomnia symptoms as an important sleep component tied to improvement in multiple measures of mental health in early treatment.

To our knowledge, this is the first study to use subjective and objective measures to longitudinally assess sleep and its relationship with measures of mental health in a heterogeneous sample of SUD treatment-seeking adults. In this preliminary study nightmares were a prevalent sleep concern that did not improve with treatment as usual, and insomnia stood out as a sleep problem that is common, malleable, and strongly tied to changes in mental health symptoms. These findings suggest nightmares and/or insomnia may be important to target in this population. Though sleeping medications are often avoided in SUD populations due to their addictive properties, evidence-based behavioral interventions exist for both nightmares (e.g., exposure, relaxation, and rescription therapy) and insomnia (e.g., cognitive behavioral therapy for insomnia) and would likely lend themselves well to integration in SUD community programs, which generally have a significant portion of treatment time committed to individual and/or group therapy. It is possible improving these factors to a clinically meaningful degree would further bolster improvements in other treatment outcomes, such as mental health, though more studies are needed to explore these relationships.

The lack of improvement in actigraphy variables and their lack of correlations with improvement in PHQ-9, GAD-7, and PCL-5 are not surprising in the context of past studies. It is common in insomnia treatment for self-reported insomnia symptoms to remit while seeing very few changes in sleep measured via actigraphy or PSG, which may point to an overlap in perceived changes in daytime dysfunction (which may or may not be perceived to relate to sleep changes) that is asked in questionnaires but not captured via objective sleep³⁵. Despite the lack of statistical significance in these areas it is still important to note the parameters measured by actigraphy and sleep diaries indicated poor overall sleep with marked room for improvement. It is possible targeting and improving these variables may be more associated with objective measures in this population, such as attention tasks or daily activity level, though more studies are needed to explore this.

Though these preliminary findings are important, interpretation is limited by small sample size. A larger sample would allow the ability to correct for family wise error generated from running multiple t-tests, provide more demographic diversity to improve generalizability of results, and allow comparisons across SUD type. Despite the sample size, this study has demonstrated insomnia and nightmares are highly prevalent across SUD types and important factors to address in community SUD treatment settings.