Abstract
Study Objectives:
Behavioral sleep medicine (BSM) is a subspecialty that combines behavioral psychology and sleep medicine specialties. The objective of this study was to analyze referral patterns to a BSM clinic. The 3 specific aims were: (1) describe factors that predict referral acceptance, (2) identify barriers to attending initial appointment, and (3) describe variables associated with the number of visits attended.
Methods:
Retrospective chart reviews were conducted as part of a quality improvement project by this study team’s clinical setting. Adults over 21 years of age who were referred to a BSM clinic in an urban Midwestern academic health care system between 2014 and 2019 were included in this study.
Results:
Sleep medicine was the main referral source for patients with BSM (74.2%), followed by internal medicine (9.3%) and neurology/psychiatry (7.3%). Thirty-eight percent of patients did not schedule an appointment after a referral for BSM was initiated. Younger age, longer distance from clinic, commercial insurance, and out-of-network insurance were all significantly greater for nonschedulers. Eighty-three percent of patients did attend the initial intake session with BSM providers. Older age was associated with lower likelihood of not attending scheduled BSM appointments.
Conclusions:
Patient characteristics of older age, closer distance from clinic, and in-network insurance coverage were found to significantly increase the likelihood of BSM scheduling, while younger age, Black race and not getting a primary sleep disorder diagnosis (vs a diagnosis of insomnia disorder) and shorter days from referral to appointment were associated with an increased likelihood of not attending the scheduled BSM treatment engagement.
Citation:
Chernyak Y, Ofner S, Williams MK, Bolarinwa C, Manchanda S, Otte JL. Patient accessibility and utilization of behavioral sleep medicine referrals in an academic center. J Clin Sleep Med. 2024;20(11):1793–1806.
Keywords: sleep, behavioral, referral, clinic
BRIEF SUMMARY
Current Knowledge/Study Rationale: Sleep disorders such as insomnia disorder are highly prevalent and are accompanied by a high symptom burden. Although nonpharmacological interventions through behavioral sleep medicine are recognized as effective treatments, little is known about the referral patterns to behavioral sleep medicine.
Study Impact: This study identified factors that influence patient acceptability of behavioral sleep medicine referrals. These findings highlight the need to address specific barriers in access to care to improve the referral process and ensure that patients with sleep disorders can receive high-quality treatment.
INTRODUCTION
Sleep insufficiency is the lack of quality time asleep that impacts daytime functioning.1 It is estimated that 14.5–15.5% of the United States population have reported trouble falling asleep and staying asleep.2 The negative impact of poor sleep includes decreased cognitive functioning and productivity, poor health outcomes, and accidents. The economic cost of poor sleep has been estimated at $299–433 billion in health care and related costs. The high prevalence of sleep insufficiency makes it critical that individuals get specialized care for contributing sleep disorders.
Behavioral sleep medicine (BSM) is a subspecialty that combines behavioral psychology and sleep medicine specialties.3,4 BSM offers nonpharmacological interventions for common sleep disorders including insomnia disorder, sleep apnea, and circadian rhythm disorders which have behavioral or psychological components.5 Insomnia disorder, defined as “dissatisfaction with sleep quantity or quality” is perhaps the most important target of BSM affecting 25% of the United States population over their lifetime, often becoming chronic and leading to years of sleep dysregulation and negatively impacting overall quality of life.6,7 Cognitive behavioral therapy for insomnia (CBT-I) is a psychotherapy falling under the umbrella of BSM which is the first line recommended treatment for insomnia disorder by various national organizations, including the American College of Physicians, the National Institutes of Health, the Department of Veterans Affairs, and the American Academy of Sleep Medicine, and has been found to be just as efficacious as prescription modalities over the long term.6–8 CBT-I can improve comorbid psychiatric conditions, reduce risk for other medical (eg, dementia and pain conditions) and psychiatric conditions (eg, anxiety and depression), and lower health care costs.6,7,9 Despite this, pharmacotherapy, typically sedative-hypnotics, continues to be the most commonly used treatment for insomnia disorder. A study on the difference between CBT-I referral and pharmacotherapy for insomnia disorder among over 5,000 veterans found that the majority received only pharmacotherapy (98%).6 This practice is problematic since pharmacological interventions are meant for short-term use and tend to be used longer than recommended to treat chronic insomnia disorder.4,10
In parallel with medical intervention, BSM can help address psychological and lifestyle factors including the sleep environment, sleep schedule, inactivity, light exposure, and caffeine use, among many others that impact not just insomnia but also circadian rhythm sleep wake disorders.5 BSM has also been shown to improve continuous positive airway pressure adherence in patients with sleep apnea with and without comorbid insomnia.4,10 As sleep disorders become more prevalent, with recognition of the adverse effects on health, mood, and quality of life, BSM is increasingly recognized as an important pathway for patients to receive nonpharmacologic treatment options.4
Despite the positive effects of BSM treatment, access to BSM centers is limited. There are various barriers to access and utilization. These include a lack of recognition of sleep disorders in primary care settings and a limited number of BSM specialists.6,11 Although there is limited information on referrals to BSM centers, studies have found that most referrals were made through primary care providers associated with government-run clinics followed by primary care in non-government-run clinics and self-referrals.9,11 Studies have found psychiatrists to be least likely to refer for CBT-I despite evidence that the most common diagnosis for a BSM referral was primary insomnia disorder with comorbid conditions of major depressive disorder, generalized anxiety disorder, and obsessive compulsive disorder.9,12 Despite being mental health professionals, psychiatrists may lack awareness of the importance of treating sleep disorders with behavioral interventions since they are not sleep specialists nor trained in nonpharmacological interventions and may view sleep outside the purview of traditional mental health care.13 Thus, many of the patients who would benefit from a BSM referral may be undiagnosed, untreated, or receive pharmacotherapy instead. Because BSM clinics provide the best front-line treatment for nonpharmacological sleep treatments, understanding referral practices is essential to determine how to shift the current paradigm. The purpose of this study was to analyze the referral patterns to an established BSM clinic at a midwestern academic health center. Factors selected for assessment included those theoretically most likely to impact successful intervention, including referral source, comorbidities, and proximity, all of which have also been previously shown to impact care access in other specialties.14,15 Results from the analysis will help guide clinical practice and potential quality improvement plans to enhance care to patients with sleep disorders. Our 3 specific aims were: (1) describe factors that predict accepting a BSM referral, (2) identify barriers to attending initial BSM appointment, and (3) describe variables associated with the number of BSM appointments attended.
METHODS
Sample selection
The following quality improvement study used existing clinical data and met exemption status as determined by the governing Institutional Review Board review. The retrospective clinical data were comprised of referral information to a BSM service at a psychiatry clinic in a large academic center in the Midwestern United States over a period of 5 years (2014–2019). Referrals were sent to the BSM clinic at which point patients were contacted by telephone and invited to schedule an appointment.
A coded formula was used to automate the process of identifying data specific to BSM referrals from a compiled Microsoft Excel file that housed information on all referrals on a university server. The search terms included: sleep, insomnia, hypersomnia, sleep medicine, narcolepsy, apnea, night, terrors, nightmares, rapid eye movement sleep behavior disorder, parasomnia, sleep walking, sleep talking. Duplicate patient data were removed.
Data
Once files were identified, a retrospective chart review was conducted as part of a quality improvement project. Demographic and clinical data were pulled from the electronic medical records of patients from 2014–2019 including date of birth, sex, race, zip code, and insurance, diagnosis, and dates of service. A referral was noted as “accepted” if an outpatient encounter was found indicating that an appointment was scheduled and as “attended” if the encounter included a start and end date/time. Diagnosis was based on the International Classification of Diseases, tenth revision and Diagnostic and Statistical Manual of Mental Disorders, fifth edition. The number of visits was determined based on number of completed encounters in the BSM program. Zip code distance was calculated as driving distance using the free CDXTech (Randolph, NJ) GeoData template with an acquired Bing application programming interface key code. Data on adjusted gross income in 2016 by zip code were obtained and used as a proxy for patients’ income.
Statistical analysis
Demographic and clinical characteristics were summarized as counts and percentages for categorical variables and as means and standard deviation for continuous variables. Percentages were compared between those who accepted the referral, defined as scheduling the initial BSM appointment, and those who did not by χ2 or Fisher’s exact tests and means were compared by t tests. Bivariate logistic models of scheduling an appointment were fit with each characteristic as the independent variable. Characteristics which had an association at P value of .3 were considered eligible for inclusion in a multivariable logistic regression model.16 Backward variable selection was used to remove nonsignificant variables until all remaining variables were significant at the 0.05 level. Similar methods were used for comparing not attending an appointment to attending the appointment from among those who scheduled an appointment. Poisson regression was used to model the number of appointments for those who attended appointments with similar entry criteria and variable selection method for obtaining the final multivariable regression model. SAS/STAT Software version 9.6 (SAS Institute Inc., Cary, North Carolina) was used for analysis.
RESULTS
The progression of patient engagement in the health care setting from referral to scheduled and completed appointments is described in Figure 1. Sleep medicine providers were the main referral source for patients with BSM (74.2%), followed by internal medicine (9.3%) and neurology/psychiatry (7.3%). Referrals were predominantly initiated for a diagnosis of Insomnia Disorder (85.0%) in the vast majority of cases. Associated mental health diagnoses were prevalent with 39.5% of the population having a concurrent mood disorder diagnosis, defined as those meeting Diagnostic and Statistical Manual of Mental Disorders, fifth edition criteria for one of the disorders listed under “Depressive Disorders” or “Bipolar and Related Disorders” categories, and another 15.8% with an anxiety disorder diagnosis.
Figure 1. Referral accrual data.
Aim 1: predicting likelihood of scheduling a BSM appointment
Thirty-eight percent of patients did not schedule an appointment after a referral for BSM was initiated (Table 1). Mean age was older for those who scheduled an appointment (P < .001). Mean distance in miles from home to clinic was significantly less for those who scheduled an appointment (P = .01). The percentage of patients with noncommercial insurance, defined as insurance provided by a government entity such as Medicaid, Medicare, or Veterans Administration was significantly greater for those who scheduled an appointment (P = .02). Of the patients with commercial insurance, a greater percentage of those with in-network coverage scheduled appointments (P < .001).
Table 1.
Summaries of patient characteristics by scheduling an appointment.
| Overall (n = 450) | Appointment Scheduled | P | |||
|---|---|---|---|---|---|
| Yes (n = 277, 61.6%) | No (n = 173, 38.4%) | ||||
| Age (in years) | Mean ± SD | 47.7 ± 14.8 | 49.8 ± 14.9 | 44.4 ± 14.1 | .0002 |
| Sex, n (%) | Male | 185 (41.1) | 121 (43.7) | 64 (37.0) | .1607 |
| Female | 265 (58.9) | 156 (56.3) | 109 (63.0) | ||
| Race, n (%) | White | 364 (80.9) | 220 (79.4) | 144 (83.2) | .7854 |
| Black | 60 (13.3) | 39 (14.1) | 21 (12.1) | ||
| Asian | 11 (2.4) | 8 (2.9) | 3 (1.7) | ||
| Unknown | 15 (3.3) | 10 (3.6) | 5 (2.9) | ||
| Referral source, n (%) | Sleep med | 334 (74.2) | 211 (76.2) | 123 (71.1) | .2130 |
| PCP/internal medicine | 42 (9.3) | 21 (7.6) | 21 (12.1) | ||
| Neurology/psychiatry | 33 (7.3) | 22 (7.9) | 11 (6.4) | ||
| Self/other | 33 (7.3) | 17 (6.1) | 16 (9.2) | ||
| Unknown | 8 (1.8) | 6 (2.2) | 2 (1.2) | ||
| Referral reason, n (%) | Sleep/insomnia | 363 (85.0) | 221 (82.5) | 142 (89.3) | .6902 |
| Other | 31 (7.3) | 20 (7.5) | 11 (6.9) | ||
| Unknown | 33 (7.7) | 27 (10.1) | 6 (3.8) | ||
| Primary sleep diagnosis, n (%) | Insomnia | 371 (82.4) | 234 (84.5) | 137 (79.2) | .6882 |
| Hypersomnia | 11 (2.4) | 7 (2.5) | 4 (2.3) | ||
| Other | 32 (7.1) | 18 (6.5) | 14 (8.1) | ||
| Unknown | 36 (8.0) | 18 (6.5) | 18 (10.4) | ||
| Primary mental health diagnosis, n (%) | Depression | 151 (34.1) | 93 (34.2) | 58 (33.9) | .8566 |
| Bipolar disorder | 24 (5.4) | 16 (5.9) | 8 (4.7) | ||
| Anxiety | 58 (13.1) | 38 (14.0) | 20 (11.7) | ||
| PTSD | 12 (2.7) | 10 (3.7) | 2 (1.2) | ||
| Schizophrenia | 3 (0.7) | 2 (0.7) | 1 (0.6) | ||
| Substance abuse | 1 (0.2) | 1 (0.4) | 0 (0.0) | ||
| Other | 22 (5.0) | 14 (5.1) | 8 (4.7) | ||
| Unknown | 172 (38.8) | 98 (36.0) | 74 (43.3) | ||
| Distance (in miles) | Mean ± SD | 22.8 ± 32.3 | 19.7 ± 21.8 | 27.9 ± 44.0 | .0098 |
| Insurance type, n (%) | Commercial/private | 224 (49.8) | 125 (45.1) | 99 (57.2) | .0150 |
| Noncommercial | 224 (49.8) | 151 (54.5) | 73 (42.2) | ||
| Unknown | 2 (0.4) | 1 (0.4) | 1 (0.6) | ||
| Insurance network, n (%) | In-network | 409 (90.9) | 274 (98.9) | 135 (78.0) | < .0001 |
| Out-of-network | 41 (9.1) | 3 (1.1) | 38 (22.0) | ||
| Adjusted gross income (in dollars) | Mean ± SD | $896,164 ± $623,175 | $917,575 ± $646,262 | $860,522 ± $582,905 | .3547 |
PCP = primary care doctor, PTSD = post-traumatic stress disorder, SD = standard deviation.
Bivariate logistic models showed that the odds of scheduling an appointment increased 13% for each 5-year increase in age, a 64% increase if the patient had noncommercial insurance vs commercial insurance and a 96% reduction if the patient’s insurance was out-of-network vs in-network. Thirty-three patients (7.7%) had unknown referral reason and were 2.9 times more likely to schedule an appointment than those with a referral reason of sleep/insomnia. Furthermore, patients’ distance from the clinic significantly influenced the likelihood of scheduling an appointment, such that the likelihood of scheduling an appointment decreased by 8% for every 10-mile increase in the distance between home and clinic.
Multiple regression models allow us to see adjusted associations between explanatory variables and the outcome. The multiple regression logistic model of scheduling an appointment included age, distance, and out-of-network insurance (Table 2). The adjusted odds of scheduling an appointment increased 11% for each 5-year increase of age, decreased 10% for each 10-mile increase in the distance from home to clinic, and were 96% smaller if patients’ insurance was out-of-network vs in-network.
Table 2.
Logistic model of scheduling an appointment.
| Bivariate Logistic Models | Multiple Logistic Regression Model* | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Beta (Std Error) | P (Overall Test) | P | Odds Ratio | 95% CI | Beta (Std Error) | P | Odds Ratio | 95% CI | ||
| Intercept | −0.711 (0.331) | .0318 | −0.058 (0.364) | .8738 | ||||||
| Age (per 5 years) | 0.025 (0.007) | .0002 | 1.134 | (1.061, 1.212) | 0.108 (0.037) | .0033 | 1.114 | (1.037, 1.196) | ||
| Intercept | 0.637 (0.155) | < .0001 | ||||||||
| Sex | Female | −0.278 (0.199) | .1611 | .1611 | 0.757 | (0.513, 1.117) | ||||
| Male | 0.000 | Reference | ||||||||
| Intercept | 0.424 (0.107) | .0001 | ||||||||
| Race | Black | 0.195 (0.291) | .7474 | .5025 | 1.216 | (0.687, 2.151) | ||||
| Asian | 0.557 (0.685) | .4167 | 1.745 | (0.455, 6.686) | ||||||
| Unknown | 0.269 (0.558) | .6294 | 1.309 | (0.438, 3.909) | ||||||
| White | 0.000 | Reference | ||||||||
| Intercept | 0.540 (0.113) | < .0001 | ||||||||
| Referral source | PCP/internal medicine | −0.540 (0.329) | .2839 | .1007 | 0.583 | (0.306, 1.110) | ||||
| Neurology/psychiatry | 0.153 (0.386) | .6912 | 1.166 | (0.547, 2.486) | ||||||
| Self/other | −0.479 (0.366) | .1910 | 0.619 | (0.302, 1.270) | ||||||
| Unknown | 0.558 (0.824) | .4983 | 1.747 | (0.347, 8.788) | ||||||
| Sleep medicine | 0.000 | Reference | ||||||||
| Intercept | 0.442 (0.108) | < .0001 | ||||||||
| Referral reason | Other | 0.156 (0.390) | .0706 | .6905 | 1.168 | (0.543, 2.511) | ||||
| Unknown | 1.062 (0.464) | .0221 | 2.891 | (1.165, 7.178) | ||||||
| Sleep/insomnia | 0.000 | Reference | ||||||||
| Intercept | 0.535 (0.108) | < .0001 | ||||||||
| Sleep diagnosis | Hypersomnia | 0.024 (0.636) | .4311 | .9696 | 1.025 | (0.295, 3.563) | ||||
| Other | −0.284 (0.372) | .4454 | 0.753 | (0.363, 1.561) | ||||||
| Unknown | −0.535 (0.350) | .1264 | 0.585 | (0.295, 1.163) | ||||||
| Insomnia | 0.000 | Reference | ||||||||
| Intercept | 0.673 (0.124) | < .0001 | ||||||||
| Distance (per 10 miles) | −0.008 (0.003) | .0154 | 0.922 | (0.864, 0.985) | −0.100 (0.037) | .0065 | 0.904 | (0.841, 0.972) | ||
| Intercept | 0.233 (0.135) | .0831 | ||||||||
| Insurance type | Noncommercial | 0.494 (0.196) | .0398 | .0118 | 1.638 | (1.116, 2.406) | ||||
| Unknown | −0.233 (1.421) | .8696 | 0.792 | (0.049, 12.822) | ||||||
| Commercial/private | 0.000 | Reference | ||||||||
| Intercept | 0.708 (0.105) | .0000 | ||||||||
| In-network (yes/no) | Out-of-network | −3.247 (0.609) | < .0001 | < .0001 | 0.039 | (0.012, 0.128) | −3.218 (0.613) | < .0001 | 0.040 | (0.012, 0.133) |
| In-network | 0.000 | Reference | ||||||||
| Intercept | 0.376 (0.174) | .0309 | ||||||||
| Adjusted gross income (per $100,000) | 0.015 (0.016) | .3546 | 1.015 | (0.983, 1.048) | ||||||
To reach final multiple regression logistic model, backwards variable selection was used on full model which included age, sex, referral source, referral reason, insurance type, insurance in-network (y/n) and distance from home to clinic. CI = confidence interval, PCP = primary care doctor, Std Error = standard error, y/n = yes/no.
Aim 2: predicting likelihood of not attending BSM appointment
Once a BSM appointment was generated by the patient, 82.5% of those patients attended the initial intake session with BSM providers. The mean age was significantly younger (P = .01) for those who did not attend their appointment. The percentage of patients with primary sleep diagnosis of other (not insomnia disorder or hypersomnia) was significantly larger for those who did not attend their appointment (Table 3).
Table 3.
Patient characteristics by attended appointment.
| Overall (n = 275) | Attended Appointment | P | |||
|---|---|---|---|---|---|
| No (n = 48, 17.4%) | Yes (n = 227, 82.5%) | ||||
| Age (in years) | Mean ± SD | 49.9 ± 14.8 | 44.7 ± 13.6 | 51.0 ± 14.9 | .0078 |
| Sex, n (%) | Male | 119 (43.3) | 17 (35.4) | 102 (44.9) | .2266 |
| Female | 156 (56.7) | 31 (64.6) | 125 (55.1) | ||
| Race, n (%) | White | 218 (79.3) | 35 (72.9) | 183 (80.6) | .0894 |
| Black | 39 (14.2) | 12 (25.0) | 27 (11.9) | ||
| Asian | 8 (2.9) | 0 (0.0) | 8 (3.5) | ||
| Unknown | 10 (3.6) | 1 (2.1) | 9 (4.0) | ||
| Referral source, n (%) | Sleep med | 209 (76.0) | 41 (85.4) | 168 (74.0) | .4185 |
| PCP/internal medicine | 21 (7.6) | 3 (6.3) | 18 (7.9) | ||
| Neurology/psychiatry | 22 (8.0) | 2 (4.2) | 20 (8.8) | ||
| Self/other | 17 (6.2) | 1 (2.1) | 16 (7.0) | ||
| Unknown | 6 (2.2) | 1 (2.1) | 5 (2.2) | ||
| Referral reason, n (%) | Sleep/insomnia | 219 (82.3) | 42 (87.5) | 177 (81.2) | .7738 |
| Other | 20 (7.5) | 3 (6.3) | 17 (7.8) | ||
| Unknown | 27 (10.2) | 3 (6.3) | 24 (11.0) | ||
| Primary sleep diagnosis, n (%) | Insomnia | 234 (85.1) | 34 (70.8) | 200 (88.1) | .0158 |
| Hypersomnia | 7 (2.5) | 3 (6.3) | 4 (1.8) | ||
| Other | 18 (6.5) | 6 (12.5) | 12 (5.3) | ||
| Unknown | 16 (5.8) | 5 (10.4) | 11 (4.8) | ||
| Primary mental health diagnosis, n (%) | Depression | 92 (34.1) | 15 (31.9) | 77 (34.5) | .0925 |
| Bipolar disorder | 16 (5.9) | 6 (12.8) | 10 (4.5) | ||
| Anxiety | 38 (14.1) | 4 (8.5) | 34 (15.2) | ||
| PTSD | 10 (3.7) | 1 (2.1) | 9 (4.0) | ||
| Schizophrenia | 2 (0.7) | 0 (0.0) | 2 (0.9) | ||
| Substance abuse | 1 (0.4) | 1 (2.1) | 0 (0.0) | ||
| Other | 14 (5.2) | 1 (2.1) | 13 (5.8) | ||
| Unknown | 97 (35.9) | 19 (40.4) | 78 (35.0) | ||
| Distance | Mean ± SD | 19.7 ± 21.8 | 18.9 ± 21.9 | 19.9 ± 21.9 | .7635 |
| Insurance type, n (%) | Commercial/private | 124 (45.1) | 23 (47.9) | 101 (44.5) | .6305 |
| Noncommercial | 150 (54.5) | 24 (50.0) | 126 (55.5) | ||
| Unknown | 1 (0.4) | 1 (2.1) | 0 (0.0) | ||
| Insurance network, n (%) | In-network | 272 (98.9) | 46 (95.8) | 226 (99.6) | .0797 |
| Out-of-network | 3 (1.1) | 2 (4.2) | 1 (0.4) | ||
| Adjusted gross income (by zip code) | Mean ± SD | $918,211 ± $644,792 | $822,985 ± $707,444 | $938,191 ± $630,740 | .2662 |
| Days from referral to appointment | Mean ± SD | 35.9 ± 50.6 | 25.0 ± 16.8 | 38.2 ± 54.9 | .1012 |
PCP = primary care doctor, PTSD = post-traumatic stress disorder, SD = standard deviation.
Bivariate logistic models showed that the odds of not attending an appointment decreased 14% for every 5-year increase in age. Interpreted as the likelihood of attending an appointment, the estimate shows that the odds of attending a scheduled appointment increased 16% for every 5 years increase in age: odds ratio = 1.16, 95% confidence interval = (1.04, 1.29). Black patients were 2.3 times more likely to not attend an appointment than White patients. Patients with other sleep diagnosis were 2.9 times more likely to not attend an appointment than those with diagnosis of insomnia disorder. Those with an unknown sleep diagnosis were 2.7 times more likely not to attend an appointment than those with diagnosis of insomnia disorder, although this was not statistically significant. Those with out-of-network insurance were 9.8 times more likely not to attend their appointment than if insurance was in-network, was found to approach significance (P = .06).
The multiple regression logistic model included age, race, sleep diagnosis and the number of days from the referral to the appointment. The adjusted odds of not attending an appointment decreased 18% for every 5 years increase in age, such that older patients were more likely to attend appointments, which is consistent with the results of the bivariate model. The adjusted odds ratio of not attending the appointment were 2.9 times larger for Black patients than White patients, 4.2 times larger for diagnosis other than insomnia disorder, 6.4 times larger for those with an unknown sleep diagnosis than those with insomnia disorder, and decreased 1.4% for each additional day between the referral and the appointment (Table 4).
Table 4.
Multivariable model of did not attend appointment.
| Bivariate Logistic Models | Multiple Logistic Regression Models* | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Beta (Std Error) | P (Overall Test) | P | Odds Ratio | 95% CI | Beta (Std Error) | P (Overall Test) | P | Odds Ratio | 95% CI | ||
| Intercept | −0.168 (0.532) | .7519 | 0.373 (0.616) | .5448 | |||||||
| Age (per 5 years) | −0.029 (0.011) | .0089 | 0.865 | (0.776, 0.964) | −0.203 (0.063) | .0013 | 0.817 | (0.721, 0.924) | |||
| Intercept | −1.792 (0.262) | < .0001 | |||||||||
| Sex | Female | 0.397 (0.330) | .2285 | .2285 | 1.488 | (0.779, 2.841) | |||||
| Male | 0.000 | Reference | |||||||||
| Intercept | −1.654 (0.184) | < .0001 | |||||||||
| Race | Black | 0.843 (0.393) | .0424 | .0319 | 2.324 | (1.076, 5.020) | 1.068 (0.430) | .0126 | .0129 | 2.91 | (1.254, 6.755) |
| Asian/unknown | −1.179 (1.045) | .2594 | 0.308 | (0.040, 2.387) | −1.528 (1.074) | .1547 | 0.217 | (0.026, 1.780) | |||
| White | 0.000 | Reference | 0.000 | Reference | |||||||
| Intercept | −1.410 (0.174) | < .0001 | |||||||||
| Referral source | PCP/internal medicine | −0.381 (0.647) | .5280 | .5559 | 0.683 | (0.192, 2.429) | |||||
| Neurology/psychiatry | −0.892 (0.762) | .2415 | 0.410 | (0.092, 1.824) | |||||||
| Self/other | −1.362 (1.045) | .1926 | 0.256 | (0.033, 1.987) | |||||||
| Unknown | −0.199 (1.109) | .8576 | 0.820 | (0.093, 7.206) | |||||||
| Sleep medicine | 0.000 | Reference | |||||||||
| Intercept | −1.438 (0.172) | < .0001 | |||||||||
| Referral reason | Other | −0.296 (0.649) | .5588 | .6484 | 0.744 | (0.208, 2.655) | |||||
| Unknown | −0.641 (0.636) | .3135 | 0.527 | (0.151, 1.832) | |||||||
| Sleep/insomnia | 0.000 | Reference | |||||||||
| Intercept | −1.772 (0.186) | < .0001 | |||||||||
| Sleep diagnosis | Hypersomnia | 1.484 (0.786) | .0291 | .0590 | 4.412 | (0.945, 20.589) | 0.954 (0.863) | .0045 | .2690 | 2.596 | (0.478, 14.100) |
| Other | 1.079 (0.533) | .0431 | 2.941 | (1.034, 8.365) | 1.441 (0.579) | .0129 | 4.227 | (1.358, 13.159) | |||
| Unknown | 0.984 (0.570) | .0846 | 2.674 | (0.874, 8.178) | 1.852 (0.657) | .0048 | 6.372 | (1.758, 23.090) | |||
| Insomnia | 0.000 | Reference | 0.000 | Reference | |||||||
| Intercept | −1.505 (0.215) | < .0001 | |||||||||
| Distance (per 10 miles) | −0.002 (0.008) | .7627 | 0.977 | (0.842, 1.134) | |||||||
| Intercept | −1.480 (0.231) | < .0001 | |||||||||
| Insurance type | Noncommercial | −0.179 (0.321) | .8565 | .5778 | 0.836 | (0.446, 1.569) | |||||
| Commercial/private | 0.000 | Reference | |||||||||
| Intercept | −1.592 (0.162) | < .0001 | |||||||||
| In-network | Out-of-network | 2.285 (1.235) | .0644 | .0644 | 9.825 | (0.873, 110.633) | |||||
| In-network | 0.000 | Reference | |||||||||
| Intercept | −1.290 (0.284) | < .0001 | |||||||||
| Adjusted gross income (per $100,000) | −0.031 (0.028) | .2669 | 0.970 | (0.918, 1.024) | |||||||
| Intercept | −1.260 (0.220) | < .0001 | |||||||||
| Days from referral to appointment | −0.009 (0.006) | .1040 | 0.991 | (0.979, 1.002) | −0.014 (0.007) | .0467 | 0.986 | (0.973, 1.000) | |||
To reach final multiple regression logistic model, backwards variable selection was used on full model included age, sex, race, sleep diagnosis, insurance in-network (y/n), adjusted gross income, number of days from referral to appointment. CI = confidence interval, PCP = primary care doctor, Std Error = standard error, y/n = yes/no.
Aim 3: associations with number of BSM visits
Data on number of visits were available for 220 of the 227 patients who attended at least 1 appointment. Five patients had an unknown number of visits and 2 patients with extreme values were excluded from this analysis. The mean number of visits completed with a BSM provider was 4.4 with standard deviation of 6.1 and median of 3.0 (Table 5). Bivariate Poisson models showed that sleep diagnosis was significantly associated with the number of visits since those with hypersomnia (excessive daytime sleepiness) had significantly greater number of visits than those with insomnia disorder (P < .01). Age, race, referral source, referral reason, sleep diagnosis, having only 1 mental health diagnosis, distance from home to clinic, and adjusted gross income were eligible for inclusion in the multiple regression model. Referral reason and sleep diagnosis were strongly related to each other. Since sleep diagnosis was statistically significant in the bivariate model, that variable was used for the multiple regression Poisson model and not referral reason. Using backward variable selection, we removed variables from the full model starting with the most nonsignificant. The final multiple regression model included age, referral source, sleep diagnosis, and having only 1 mental health diagnosis. Increasing age, having unknown referral source, hypersomnia and more than 1 mental health diagnosis were associated with greater number of visits. Having an unknown sleep diagnosis was associated with fewer number of visits (Table 6).
Table 5.
Associations with number of visits completed.
| Overall (n = 220) | ||
|---|---|---|
| Age | Mean ± SD | 50.8 ± 15.0 |
| Median (min, max) | 51.0 (18.0, 83.0) | |
| Sex | Male | 100 (45.5) |
| Female | 120 (54.5) | |
| Race | White | 179 (81.4) |
| Black | 25 (11.4) | |
| Asian | 8 (3.6) | |
| Unknown | 8 (3.6) | |
| Referral source | Sleep med | 165 (75.0) |
| PCP/internal medicine | 17 (7.7) | |
| Neurology/psychiatry | 18 (8.2) | |
| Self/other | 15 (6.8) | |
| Unknown | 5 (2.3) | |
| Referral reason | Sleep/insomnia | 173 (81.6) |
| Other | 17 (8.0) | |
| Unknown | 22 (10.4) | |
| Primary sleep diagnosis | Insomnia | 195 (88.6) |
| Hypersomnia | 4 (1.8) | |
| Other | 12 (5.5) | |
| Unknown | 9 (4.1) | |
| Two sleep diagnoses | No | 144 (65.5) |
| Yes | 76 (34.5) | |
| Primary mental health diagnosis | Depression | 73 (33.8) |
| Bipolar disorder | 9 (4.2) | |
| Anxiety | 33 (15.3) | |
| PTSD | 9 (4.2) | |
| Schizophrenia | 2 (0.9) | |
| Other | 13 (6.0) | |
| Unknown | 77 (35.6) | |
| Two mental health diagnoses | No | 160 (74.1) |
| Yes | 56 (25.9) | |
| Unknown | 4 | |
| Distance | Mean ± SD | 19.9 ± 21.8 |
| Median (min, max) | 14.3 (0.0, 130.0) | |
| Insurance type | Commercial/private | 98 (44.5) |
| Noncommercial | 122 (55.5) | |
| Insurance network | In-network | 219 (99.5) |
| Out-of-network | 1 (0.5) | |
| Adjusted gross income (by zip code) | Mean ± SD | $938,300 ± $639,267 |
| Median (min, max) | $751,583 ($19,642, $3,452,148) | |
| Number of appointments | Mean ± SD | 4.4 ± 6.1 |
| Median (min, max) | 3.0 (1.0, 45.0) | |
max = maximum, min = minimum, PCP = primary care doctor, PTSD = post-traumatic stress disorder, SD = standard deviation.
Table 6.
Poisson model of number of visits.
| Bivariate Poisson Models | Multiple Regression Poisson Model* | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Beta (Std Error) | 95% CI | P (Overall Test) | P | Beta (Std Error) | 95% CI | P (Overall Test) | P | ||
| Intercept | 1.079 (0.260) | (0.570, 1.588) | < .0001 | 0.688 (0.274) | (0.150, 1.226) | .0122 | |||
| Age per 5 years | 0.040 (0.024) | (−0.007, 0.086) | .0965 | 0.067 (0.024) | (0.021, 0.114) | .0048 | |||
| Intercept | 1.470 (0.106) | (1.262, 1.678) | < .0001 | ||||||
| Sex | Female | 0.032 (0.143) | (−0.248, 0.312) | .8224 | .8225 | ||||
| Male | 0.000 | ||||||||
| Intercept | 1.471 (0.079) | (1.316, 1.625) | < .0001 | ||||||
| Race | Black | −0.055 (0.231) | (−0.508, 0.398) | .2286 | .8127 | ||||
| Asian | 0.609 (0.287) | (0.047, 1.171) | .0337 | ||||||
| Unknown | −0.218 (0.424) | (−1.049, 0.613) | .6075 | ||||||
| White | 0.000 | ||||||||
| Intercept | 1.479 (0.082) | (1.319, 1.639) | < .0001 | ||||||
| Referral source | PCP/internal medicine | 0.070 (0.259) | (−0.438, 0.578) | .1696 | .7872 | 0.192 (0.252) | (−0.302, 0.686) | .0195 | .4459 |
| Neurology/psychiatry | −0.195 (0.285) | (−0.753, 0.363) | .4939 | −0.393 (0.288) | (−0.957, 0.172) | .1728 | |||
| Self/other | −0.180 (0.308) | (−0.783, 0.423) | .5594 | −0.076 (0.298) | (−0.660, 0.508) | .7982 | |||
| Unknown | 0.824 (0.322) | (0.193, 1.454) | .0104 | 1.055 (0.312) | (0.442, 1.667) | .0007 | |||
| Sleep medicine | 0.000 | 0.000 | |||||||
| Intercept | 1.402 (0.083) | (1.239, 1.565) | < .0001 | ||||||
| Referral reason | Other | 0.447 (0.228) | (0.000, 0.894) | .1242 | .0502 | ||||
| Unknown | 0.243 (0.223) | (−0.194, 0.680) | .2754 | ||||||
| Sleep/insomnia | 0.000 | ||||||||
| Intercept | 1.477 (0.074) | (1.331, 1.623) | < .0001 | ||||||
| Sleep diagnosis | Hypersomnia | 0.898 (0.340) | (0.232, 1.564) | .0143 | .0082 | 1.259 (0.351) | (0.572, 1.947) | .0006 | .0003 |
| Other | 0.181 (0.284) | (−0.375, 0.738) | .5230 | 0.064 (0.280) | (−0.485, 0.612) | .8195 | |||
| Unknown | −1.035 (0.586) | (−2.183, 0.113) | .0773 | −1.316 (0.579) | (−2.452, −0.181) | .0231 | |||
| Insomnia | 0.000 | 0.000 | |||||||
| Intercept | 1.457 (0.089) | (1.282, 1.631) | < .0001 | ||||||
| Only 1 sleep diagnosis | No | 0.088 (0.147) | (−0.201, 0.376) | .5542 | .5526 | ||||
| Yes | 0.000 | ||||||||
| Intercept | 1.440 (0.085) | (1.272, 1.607) | < .0001 | ||||||
| Only 1 mental health diagnosis | No | 0.191 (0.157) | (−0.116, 0.498) | .2284 | .2222 | 0.311 (0.154) | (0.009, 0.613) | .0477 | .0436 |
| Yes | 0.000 | 0.000 | |||||||
| Intercept | 1.391 (0.095) | (1.204, 1.578) | < .0001 | ||||||
| Distance per 10 miles | 0.048 (0.028) | (−0.008, 0.103) | .0946 | ||||||
| Intercept | 1.455 (0.108) | (1.243, 1.667) | < .0001 | ||||||
| Insurance type | Noncommercial | 0.058 (0.143) | (−0.223, 0.339) | .6862 | .6865 | ||||
| Commercial/private | 0.000 | ||||||||
| Intercept | 1.491 (0.071) | (1.352, 1.630) | < .0001 | ||||||
| In-network | Out-of-network | −1.491 (2.214) | (−5.830, 2.848) | .3727 | .5005 | ||||
| In-network | 0.000 | ||||||||
| Intercept | 1.616 (0.128) | (1.366, 1.866) | < .0001 | ||||||
| Adjusted gross income per $100,000 | −0.013 (0.012) | (−0.037, 0.010) | .2603 | ||||||
To reach final multiple regression Poisson model, backwards variable selection was used on full model which included: age, race, referral source, sleep diagnosis, having 1 mental health diagnosis, distance from home to clinic and adjusted gross income. CI = confidence interval, PCP = primary care doctor, Std Error = standard error.
DISCUSSION
Nearly 40% of patients did not schedule an appointment after a referral for BSM was initiated, most of which were initiated for insomnia disorder. These findings are similar to a recently published study from this team where the majority of cancer survivors referred to BSM for chronic insomnia disorder also did not complete the scheduled referral process.17 The lack of BSM follow-up represents a gap in implementing recommended care, especially in light of the American College of Physicians 2016 guideline that the gold standard treatment for insomnia disorder should be CBT-I, delivered predominantly by trained BSM specialists.18 Approximately the same percentage of individuals who did not schedule BSM appointment had a concurrent mood disorder diagnosis, however, this was not found to be a significant predictor. Once the initial intake session was completed with a BSM specialist, 96.9% of patients continued ongoing care, defined by completing further appointments.
Not surprisingly, barriers to scheduling an appointment included commercial insurance, out-of-network insurance, and distance from clinic. These findings suggest that systematic barriers to scheduling BSM evaluations may be logistical in nature, such that improving access to providers and reducing costs associated with treatment would improve follow-through. One method to improve access is offering several CBT-I modalities (eg, in-person and telehealth).19 Notably, telehealth has emerged as an effective resource to improve access to care20 and there is preliminary evidence that the efficacy of CBT-I delivered via telehealth is comparable to that of in-person CBT-I.21,22 Another method could be to dedicate efforts to streamlining insurance panels across specialty clinics to help control costs. CBT-I sessions can be costly depending on various factors (eg, clinician and number of sessions) and some insurance may limit the number of telehealth sessions that are eligible for reimbursement.23 Additional barriers often include gaps in insurance coverage for behavioral health and high rates of behavioral health claim denials due to inconsistent medical necessity criteria.24 As an additional method to reduce barriers, providers could discuss the value of CBT-I with their patients. A recent systematic review found that CBT-I is a cost-effective treatment for insomnia disorder.19 Thus, efforts should also be made to increase patients’, providers’, and insurers’ awareness of CBT-I as a practical, effective, and cost-saving treatment option for insomnia disorder.
Although the majority of BSM referrals in this study originated with sleep medicine specialists, this may be a reflection of greater awareness of BSM and an established professional relationship between providers at this institution, which may be translatable to other academic medical centers who have in-house BSM expertise. A significant minority of referrals came from internal medicine, which is where many patients with insomnia disorder or other BSM needs are seen in the community. Referrals for BSM were predominantly for a diagnosis of Insomnia Disorder, which is consistent with its high base rates of nearly 30% lifetime prevalence compared to other sleep disorders.9
The likelihood of attending an appointment with a BSM specialist was influenced by sleep diagnosis list on the referral, such that those with an “unknown” diagnosis were 6 times and those with an “other” diagnosis 4 times more likely to not attend the appointment compared to those with an insomnia diagnosis. The authors posit that those with an “unknown” or “other” referral reason represented either more complex and/or less appropriate medical cases for a BSM referral. Patients and/or referring providers may be less certain of benefit from BSM for these complex or nonspecific diagnoses and therefore be less likely to follow-through with the referral, which is consistent with the lack of literature supporting BSM outside of evidence-based interventions such as CBT for insomnia disorder or hypersomnia.
Although most patients who scheduled an intake appointment attended it, it is noteworthy that 38% of patients with referrals did not schedule an appointment. This suggests that patients may have encountered significant barriers during the referral process. Reasons why patients may not engage in CBT-I include barriers such as time constraints, long distances, and a lack of awareness about the effectiveness of nonpharmacological treatment options for insomnia disorder.25 Another potential barrier may have been the scheduling system itself. As appointment systems in large health care organizations move to automated scheduling, patients, especially vulnerable or underrepresented minorities, may find it burdensome.26–28 The low rate of BSM evaluation scheduling represents a significant loss to an access of care point where providers could potentially recommend effective treatment plans, as patients are generally more likely to attend a scheduled BSM evaluation than to schedule one.
Once a BSM appointment was scheduled by the patient, over 80% of patients attended the initial intake session with providers. It is noteworthy that this referral completion rate is greater than that reported in a previous study in which only 57% of patients who were referred to specialty services in a large health care system attended their scheduled appointment.29 Surprisingly, shorter duration between referral and appointment date also decreased the chances of attending the appointment. As with scheduling appointments, the chances of attending also improved with older age.
Interestingly, race was a predictive factor in the likelihood of attending a BSM appointment, such that Black patients were less likely to attend a scheduled BSM appointment. This finding is consistent with previous studies which found that Black patients had lower appointment completion rates across various specialties.30,31 This suggests that there may be systematic barriers for this demographic population that warrant further investigation. One barrier may have been limited access to specialty care. As Black Americans are at greater risk of being uninsured or underinsured,32 some Black patients may have chosen to forgo a BSM appointment due to the potential financial burden of out-of-pocket expenses. Other barriers associated with socioeconomic status may also explain this finding. A systematic review found that factors such as work conflicts and lower education were associated with lower appointment completion in Black patients referred to cardiac rehabilitation.33
Several systematic changes may promote sleep health equity. One, providers could consider offering BSM services through multiple modalities that can further support underserved groups. In addition to telehealth, which has been endorsed by the American Academy of Sleep Medicine as one method to improve access to sleep care for underserved groups,34 group CBT-I is considered an efficacious treatment for insomnia that could increase access to sleep care.35 Notably, recruitment of minoritized patients who can support each other has been previously identified as a potential method to improve continuous positive airway pressure adherence.36 Second, health care leaders could consider adopting a collaborative care model in their health care systems. The collaborative care model, a team-based approach in which mental health services are provided within a medical setting, has been identified as a cost-effective option that may improve access to behavioral health care for underserved groups by providing direct access through primary care.37
Streamlining scheduling methods so that a referral appointment could be scheduled before the patient leaves the referring provider’s office could circumvent scheduling challenges and increase the likelihood of referral completion, given that most patients follow-through once the appointment is scheduled. Second, patients may benefit from sleep education (ie, insomnia disorder and CBT-I, BSM for continuous positive airway pressure) before a referral is placed. Targeted sleep education has been proposed as one way to address some factors (ie, beliefs about sleep and health literacy) that contribute to sleep health disparities.36 Lastly, advocacy efforts to promote insurance coverage for BSM specialty care could be another solution to address sleep disparities and improve access to care.36
CBT-I represents the vast majority of BSM referrals. Average treatment duration with CBT-I is 6+ sessions,38 which is greater than this study’s mean number of 4.4 visits. This finding could be explained by various potential factors. The session average in this study could have been brought down by low adherence rates, such that patients did not attend recommended interventions to completion or full symptom resolution. It could also be influenced by lower treatment duration for other BSM presenting problems such as continuous positive airway pressure desensitization, which may require fewer visits.39,40 Age, hypersomnia, having an unknown referral source, and more than 1 mental health diagnosis were all associated with greater number of visits. The unknown referral source could indicate a motivated self-referred patient, who might have different characteristics or expectations that contribute to both initiation and completion of BSM treatment. Possible explanations for those with hypersomnia and more than 1 mental health diagnosis are that they could represent more treatment resistant cases who also need more intensive intervention. Given that this study did not examine the reasons behind the number of visits attended, a discussion of this finding should be interpreted with caution and tested in future studies.
Limitations
Although the study has many strengths, there were several notable limitations. First, this was a retrospective chart review limiting the study team’s ability to verify medical history and other documentation with the patient. The absence of patient feedback also limited the ability to determine patient specific reasons for referral process outcomes. Second, the data were based on a Midwestern urban medical center limiting possible generalizability to other more rural areas without sleep clinic support. Third, while sleep medicine provided the majority of BSM referrals, BSM was not embedded within the sleep medicine clinic. This may have presented as a barrier to care for patients who received BSM referrals. While integrating BSM services into sleep medicine clinics is a logical choice, other tertiary care centers that have clinical populations with elevated rates of insomnia disorder (eg, psychiatry and pain management) are also considered appropriate places to provide these services.38 Lastly, this study may not have fully accounted for all the factors that contribute to referral completion and retention rates (eg, sleep disorder chronicity and subjective reports).
CONCLUSIONS
In closing, this study analyzed the referral patterns to BSM and highlighted a need to improve the referral process for BSM services. Since BSM is a direct access point to providing nonpharmacological interventions that are as effective and safer than pharmacological interventions, a paradigm shift in treatment of insomnia disorder and other sleep disorders is needed. This study found that most referrals were made by departments most familiar with BSM and its targets. Therefore, an effective strategy to improve referral rates may be by increasing awareness among health care providers of available BSM services. Given that many referrals were not completed, efforts should also be made to bolster the referral process for patients. Strategies may include increasing patient awareness of the benefits of BSM services, integrating BSM services into departments with high rates of sleep disorders, streamlining insurance panels across specialty clinics, and increasing advocacy efforts for patients with barriers to treatment. To further illuminate the main barriers to the BSM referral process, future studies are encouraged to include patient-reported data (eg, reasons for appointment completion and barriers related to incomplete appointments). Together, these efforts could help ensure that patients with sleep disorders receive high quality care.
DISCLOSURE STATEMENT
All authors have seen and approved the manuscript. The authors report no conflicts of interest.
ABBREVIATIONS
- BSM
behavioral sleep medicine
- CBT-I
cognitive behavioral therapy for insomnia
REFERENCES
- 1. Cirelli C . Insufficient sleep: definition, epidemiology, and adverse outcomes . UptoDate. Updated October 17, 2023. . https://www.uptodate.com/contents/insufficient-sleep-definition-epidemiology-and-adverse-outcomes . Accessed January 20, 2024.
- 2. CDC . Sleep Difficulties in Adults: United States, 2020. Centers for Disease Control and Prevention; . https://www.cdc.gov/nchs/data/databriefs/db436-tables.pdf . Accessed January 10, 2024. [Google Scholar]
- 3. Haynes PL . The role of behavioral sleep medicine in the assessment and treatment of sleep disordered breathing . Clin Psychol Rev. 2005. ; 25 ( 5 ): 673 – 705 . [DOI] [PubMed] [Google Scholar]
- 4. Stepanski EJ, Perlis ML . Behavioral sleep medicine. An emerging subspecialty in health psychology and sleep medicine . J Psychosom Res. 2000. ; 49 ( 5 ): 343 – 347 . [DOI] [PubMed] [Google Scholar]
- 5. Roth T, Drake C, Roehrs T . Behavioral sleep medicine . J Clin Sleep Med. 2013. ; 9 ( 9 ): 981 – 982 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Bramoweth AD, Renqvist JG, Hanusa BH, Walker JD, Germain A, Atwood CW Jr . Identifying the demographic and mental health factors that influence insomnia treatment recommendations within a veteran population . Behav Sleep Med. 2019. ; 17 ( 2 ): 181 – 190 . [DOI] [PubMed] [Google Scholar]
- 7. Byars K, Simon S . Practice patterns and insomnia treatment outcomes from an evidence-based pediatric behavioral sleep medicine clinic . Clin Pract Pediatr Psychol. 2014. ; 2 ( 3 ): 337 – 349 . [Google Scholar]
- 8. Edinger JD, Arnedt JT, Bertisch SM, et al . Behavioral and psychological treatments for chronic insomnia disorder in adults: an American Academy of Sleep Medicine clinical practice guideline . J Clin Sleep Med. 2021. ; 17 ( 2 ): 255 – 262 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Morin CM, LeBlanc M, Daley M, Gregoire JP, Mérette C . Epidemiology of insomnia: prevalence, self-help treatments, consultations, and determinants of help-seeking behaviors . Sleep Med. 2006. ; 7 ( 2 ): 123 – 130 . [DOI] [PubMed] [Google Scholar]
- 10. Sweetman A, Lack L, Catcheside PG, et al . Cognitive and behavioral therapy for insomnia increases the use of continuous positive airway pressure therapy in obstructive sleep apnea participants with comorbid insomnia: a randomized clinical trial . Sleep. 2019. ; 42 ( 12 ): zsz178 . [DOI] [PubMed] [Google Scholar]
- 11. Mahendran R, Chan YH . Pathways to specialist care in an insomnia clinic at a psychiatric hospital: a comparative analysis of two periods . Ann Acad Med Singap. 2008. ; 37 ( 9 ): 733 – 737 . [PubMed] [Google Scholar]
- 12. Bartlett DJ, Marshall NS, Williams A, Grunstein RR . Predictors of primary medical care consultation for sleep disorders . Sleep Med. 2008. ; 9 ( 8 ): 857 – 864 . [DOI] [PubMed] [Google Scholar]
- 13. Khawaja IS, Dickmann PJ, Hurwitz TD, et al . The state of sleep medicine education in North American psychiatry residency training programs in 2013: chief resident’s perspective . Prim Care Companion CNS Disord. 2017. ; 19 ( 4 ): 17br02167 . [DOI] [PubMed] [Google Scholar]
- 14. Chung J, Aguila F, Harris O . Validity assessment of referral decisions at a VA health care system polytrauma system of care . Cureus. 2015. ; 7 ( 1 ): e240 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Brathwaite R, Hutchinson E, McKee M, Palafox B, Balabanova D . The long and winding road: a systematic literature review conceptualising pathways for hypertension care and control in low- and middle-income countries . Int J Health Policy Manag. 2022. ; 11 ( 3 ): 257 – 268 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Hosmer DW, Lemeshow S . Applied Logistic Regression. 2nd ed . Hoboken, NJ: : John Wiley & Sons, Inc; .; 2000. . [Google Scholar]
- 17. Otte JL, Chernyak Y, Johns SA, et al . Referral process to further evaluate poor sleep in breast cancer survivors . Cancer Med. 2022. ; 11 ( 8 ): 1891 – 1901 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Qaseem A, Kansagara D, Forciea MA, Cooke M, Denberg TD ; Clinical Guidelines Committee of the American College of Physicians . Management of chronic insomnia disorder in adults: a clinical practice guideline from the American College of Physicians . Ann Intern Med. 2016. ; 165 ( 2 ): 125 – 133 . [DOI] [PubMed] [Google Scholar]
- 19. Natsky AN, Vakulin A, Chai-Coetzer CL, et al . Economic evaluation of cognitive behavioural therapy for insomnia (CBT-I) for improving health outcomes in adult populations: a systematic review . Sleep Med Rev. 2020. ; 54 : 101351 . [DOI] [PubMed] [Google Scholar]
- 20. Gajarawala SN, Pelkowski JN . Telehealth benefits and barriers . J Nurse Pract. 2021. ; 17 ( 2 ): 218 – 221 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Arnedt JT, Conroy DA, Mooney A, Furgal A, Sen A, Eisenberg D . Telemedicine versus face-to-face delivery of cognitive behavioral therapy for insomnia: a randomized controlled noninferiority trial . Sleep. 2021. ; 44 ( 1 ): zsaa136 . [DOI] [PubMed] [Google Scholar]
- 22. Gehrman P, Gunter P, Findley J, et al . Randomized noninferiority trial of telehealth delivery of cognitive behavioral treatment of insomnia compared to in-person care . J Clin Psychiatry. 2021. ; 82 ( 5 ): 20m13723 . [DOI] [PubMed] [Google Scholar]
- 23. Vargas I, Egeler M, Walker J, Benitez D . Examining the barriers and recommendations for integrating more equitable insomnia treatment options in primary care . Front Sleep. 2023. ; 2 : 1279903 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Modi H, Orgera K, Grover A . Exploring barriers to mental health care in the U.S. https://www.aamcresearchinstitute.org/our-work/issue-brief/exploring-barriers-mental-health-care-us . Accessed April 7, 2023.
- 25. Koffel E, Bramoweth AD, Ulmer CS . Increasing access to and utilization of cognitive behavioral therapy for insomnia (CBT-I): a narrative review . J Gen Intern Med. 2018. ; 33 ( 6 ): 955 – 962 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Woodcock E, Sen A, Weiner J . Automated patient self-scheduling: case study . J Am Med Inform Assoc. 2022. ; 29 ( 9 ): 1637 – 1641 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Luxenburg O, Myers V, Ziv A, et al . Factors affecting the patient journey in scheduling a specialist appointment in a public healthcare system . J Patient Exp. 2022. ; 9 : 23743735221092547 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Woodcock EW . Barriers to and facilitators of automated patient self-scheduling for health care organizations: scoping review . J Med Internet Res. 2022. ; 24 ( 1 ): e28323 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Patel MP, Schettini P, O’Leary CP, Bosworth HB, Anderson JB, Shah KP . Closing the referral loop: an analysis of primary care referrals to specialists in a large health system . J Gen Intern Med. 2018. ; 33 ( 5 ): 715 – 721 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Ebrahimzadeh JE, Long JM, Wang L, et al . Associations of sociodemographic and clinical factors with gastrointestinal cancer risk assessment appointment completion . J Genet Couns. 2020. ; 29 ( 4 ): 616 – 624 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Mathews L, Akhiwu O, Mukherjee M, Blumenthal RS, Matsushita K, Ndumele CE . Disparities in the use of cardiac rehabilitation in African Americans . Curr Cardiovasc Risk Rep. 2022. ; 16 ( 5 ): 31 – 41 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Hughes AJ, Gunn H, Siengsukon C, et al . Eliminating sleep health disparities and achieving health equity: seven areas for action in the behavioral sleep medicine community . Behav Sleep Med. 2023. ; 21 ( 5 ): 633 – 645 . [DOI] [PubMed] [Google Scholar]
- 33. Koehler Hildebrandt AN, Hodgson JL, Dodor BA, Knight SM, Rappleyea DL . Biopsychosocial-Spiritual factors impacting referral to and participation in cardiac rehabilitation for African American patients: a systematic review . J Cardiopulm Rehabil Prev. 2016. ; 36 ( 5 ): 320 – 330 . [DOI] [PubMed] [Google Scholar]
- 34. Shamim-Uzzaman QA, Bae CJ, Ehsan Z, et al . The use of telemedicine for the diagnosis and treatment of sleep disorders: an American Academy of Sleep Medicine update . J Clin Sleep Med. 2021. ; 17 ( 5 ): 1103 – 1107 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Koffel EA, Koffel JB, Gehrman PR . A meta-analysis of group cognitive behavioral therapy for insomnia . Sleep Med Rev. 2015. ; 19 : 6 – 16 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Billings ME, Cohen RT, Baldwin CM, et al . Disparities in sleep health and potential intervention models: a focused review . Chest. 2021. ; 159 ( 3 ): 1232 – 1240 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Mongelli F, Georgakopoulos P, Pato MT . Challenges and opportunities to meet the mental health needs of underserved and disenfranchised populations in the United States . Focus (Am Psychiatr Publ). 2020. ; 18 ( 1 ): 16 – 24 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Muench A, Vargas I, Grandner MA, et al . We know CBT-I works, now what? Fac Rev. 2022. ; 11 : 4 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Medalie L . CBT for CPAP adherence . In: Medalie L , ed. Behavioral Sleep Medicine: A Practical Guide for Adult and Pediatric Providers. Cham, Switzerland: : Springer International Publishing; ; 2022. : 53 – 63 . [Google Scholar]
- 40. Dettenmeier P, Ordaz E, Espiritu J . Evaluation of a continuous positive airway pressure desensitization protocol for CPAP-intolerant patients: a pilot study . Chest. 2013. ; 144 ( 4 ): 979A . [Google Scholar]