Secondary Logo

Journal Logo

Utilization of an Employee Behavioral Health Program and Its Effects on Outcomes for Depression and Anxiety Disorders

Maeng, Daniel PhD; Cornell, Ann E. PsyD; Nasra, George S. MD, MBA

Journal of Occupational and Environmental Medicine: October 2019 - Volume 61 - Issue 10 - p 812–817
doi: 10.1097/JOM.0000000000001678

Objective: To examine the impact of an employer-sponsored behavioral health program on depression and anxiety by assessing dose effect of psychotherapy.

Methods: A retrospective data analysis of patients with baseline scores more than or equal to 10 on the Patient Health Questionnaire (PHQ9) or the Generalized Anxiety Disorder 7-item scale (GAD7). Survival analyses were conducted to assess whether those with a higher number of therapy sessions per episode (dose) achieved faster response (score reduction by 50% or below 10).

Results: Patients with medium (8 to 12 visits) or high (more than 12 visits) dose achieved faster response than those with low dose (less than eight visits; hazard ratios more than 1.5, P < 0.05). No significant difference was found between the medium and high dose.

Conclusion: Higher dose of psychotherapy is correlated with improved behavioral health outcomes, although there appears to be no incremental benefit beyond a certain level.

University of Rochester Medical Center, Rochester, New York.

Address correspondence to: Daniel Maeng, PhD, Department of Psychiatry, University of Rochester Medical Center, Box PSYCH, 300 Crittenden Blvd Rochester, NY 14642 (

The funding support for this study was provided by the interal funds within the Department of Psychiatry, University of Rochester Medical Center. No other external funding was provided.

Maeng, Cornell, Nasra have no other relationships/conditions/circumstances that present potential conflict of interest.

The JOEM editorial board and planners have no financial interest related to this research.

Statement of Clinical Significance: Depression and anxiety disorders are common behavioral health conditions affecting employees. This study finds that higher dose of psychotherapy provided via an employer-sponsored behavioral health program is correlated with improvements in PHQ8 and GAD7 among employees, although there appears to be no incremental benefit beyond a certain level.

Supplemental digital contents are available for this article. Direct URL citation appears in the printed text and is provided in the HTML and PDF versions of this article on the journal's Web site (

Back to Top | Article Outline

Learning Objectives

  • Discuss the rationale for and organization of the employer-sponsored behavioral health program described by Maeng et al.
  • Summarize the characteristics and outcomes of patients treated for anxiety and depression in the behavioral health program.
  • Describe the findings on analysis of the dose-response effect of psychotherapy in the treatment program.

Depression and anxiety disorder are two of the most common mental health conditions in the United States. A recent study1 estimates lifetime prevalences of 14.4% for major depressive disorder and 4.3% for generalized anxiety disorder (GAD) among the general population in the United States with age of 13 or older. The corresponding 12-month prevalences are 7.1% for major depressive disorder and 2.0% for GAD. These mental health conditions can lead to significant work productivity loss among working age adults.2–4 In particular, a recent study suggests that even low levels of depression severity are associated with noticeable work productivity loss.5 Although efficacious treatments for these conditions are available via psychotherapy, medications, or in combinations of both,6,7 use of such treatments is relative low.2,8 Consequently, some employers in the United States have sought ways to provide mental and behavioral health support for their employees, and several studies have looked at the impact of mental health program in workplace,9–13 showing some positive effects on absenteeism, presenteeism, psychological well-being, employee burnout, and turnover rates. However, more empirical evidence is needed to justify further expansion of similar programs.

This study examines a 5-year experience of a large academic employer, University of Rochester (UR), located in the Northeast region of the United States, in implementing its version of employer-sponsored behavioral health program. As the largest employer in Rochester, New York, UR has long recognized the importance of decreasing disability and turnover among its employees. At the same time, as a self-insured employer offering comprehensive health benefit packages to its employees, UR also has faced the dual challenge of meeting the employees’ healthcare needs while controlling the cost of care. As a result, UR has been active in adopting and implementing various employee wellness programs over the years, targeting improvements in healthy lifestyles, preventive care uitilization, and emotional well-being among its employees.

This study analyzes available retrospective administrative data obtained from electronic health records to describe the relationship between the frequency of participating employees’ visits to the employer-sponsored behavioral health clinic (referred to as behavioral health partners, or BHP) and the corresponding changes in their mental health outcomes over time. Note that this study does not directly estimate the treatment effects of the program relative to no intervention, because the available data lacked an acceptable control group. Instead, this study examines the correlation between the outcome of interest (ie, time to achieving goal) and a predictor (ie, frequency of visits); more specifically, we examined the dose–response effect of the intervention.

We defined “dose” as the frequency of visits in a given period and “response” as more than or equal to 50% improvement in health outcome measures. It is hypothesized that higher frequency of visits (ie, higher dose) is correlated with a more rapid improvement in mental health outcomes (ie, higher response). We propose that this dose–response approach to evaluating program impact is appropriate in this context, even though the optimal dose level has not been established,14,15 because the magnitude of the impact would crucially depend on treatment dose.

Back to Top | Article Outline


In 2013, initial discussions began with the aim of designing and implementing an employer-sponsored mental health clinic within UR that would be funded the Human Resources (HR) Department and offered as part of a larger health benefit package to all employees enrolled in the UR health plan. An analysis of UR's 2010 health plan claims data had indicated that 21,707 mental health services were delivered that year to approximately 2500 UR employees and their adult (18 years of age or older) dependents. Of those beneficiaries, the majority of the diagnoses was related to mood and anxiety disorders, with a smaller proportion related to substance use disorders. This indicated to UR leadership that there was sufficient need for creating an in-house psychiatric clinic focusing on depression and anxiety disorders, with a specific focus on access and quality.

BHP opened its doors in January of 2014. It was designed to offer short-term, evidence-based mental health treatment for UR employees and their adult dependents enrolled in UR health plan. The vision was to provide in-house mental health services with the goal of better influencing utilization, outcomes, and cost by leveraging the existing behavioral health care resources and expertise available within UR. To this end, BHP was staffed by a multidisciplinary team of licensed mental health therapists and providers including psychologists, social workers, a registered nurse, nurse practitioner, a psychiatrist, and a care manager, all of whom were full-time staff and faculty with clinical and academic priviledges through the Department of Psychiatry at the UR Medical Center. In its first year of operation, BHP served 480 patients—since then, it has experienced steady growth in patient volume.

BHP clinics are physically located in multiple locations at or near UR campus to ensure convenience and accessibility for employees. The program continues to explore marketing, staffing, and operational innovations to expand its pool of patients served. Such innovations include outreach efforts to reach specific employee groups within UR, after-hour scheduling, adding more locations, as well as a marketing campaign to expand the use of telehealth services. BHP augments the traditional Employee Assistance Program (EAP) that provides short term assessment, brief counseling, and referral services16 by adding condition-specific, multidisciplinary psychiatric treatment, including medication management, for up to 18 sessions for individuals suffering from depression and anxiety disorders. The range of BHP services includes different forms of evidence-based psychotherapies, medication consultation, and management, along with coordination and outreach by an embedded care manager.

Conditions such as bipolar disorder, psychosis, eating disorder, severe obsessive compulsive disorder, primary substance use disorder, or other conditions that are classified as severe and persistent mental illness fall outside the scope of the BHP program. A triage phone call by the care manager prior to enrolling the patient in care at BHP helps determine eligibility. Few cases requiring further discussion are presented at team meetings and eligibility is determined at that time. All new patients are offered an intake appointment within 10 business days, although most appointments are schedule within 4 days of the initial call. Patients with conditions that fall outside the program's scope are provided with appropriate community referrals.

No referrals from primary care clinicians, or prior authorizations from insurers, are necessary prior to BHP enrollment. The majority of BHP patients are self-referred or referred from other UR benefit programs, such as EAP. Patients’ out-of-pocket spending related to BHP visits can vary depending on the patient's health plan type. For those employees with a high-deductible plan, the corresponding deductible must be met, and for those with the traditional managed care copay plan, there is no out-of-pocket cost for BHP services. Surveys have shown a high degree of satisfaction among BHP patients and therapists. Comments have consistently mentioned the convenient location, flexibility of services offered, the quality of service, and the welcoming office atmosphere.

Back to Top | Article Outline


This study has been reviewed and approved by University of Rochester's Institutional Review Board. The analysis focuses on a subset of BHP with moderate to severe depression or anxiety disorders as defined by elevated scores on the Patient Health Questionnaire (PHQ9) or the Generalized Anxiety Disorder (GAD7) scores. PHQ9 and GAD7 are clinically validated patient-reported instruments that measure the severity of depression and anxiety disorder, respectively.17,18 Because each instrument typically takes less than 5 minutes to complete, and a single score is generated that allows clinicians to categorize the respondent into one of the ordinal levels of severity (ie, mild, moderate, or severe), the PHQ9 or GAD7 instruments have been administered as a matter of routine to all BHP patients at every visit, enabling the clinicians to monitor each patient's progress over time.

Back to Top | Article Outline


BHP's electronic health record database was queried to identify all individuals enrolled in the BHP program and had at least one completed visit with the BHP behavioral healthcare providers at any point during the study period, which spanned from January 1, 2014 to August 15, 2018. From the same database, each patient's demographic information, appointment dates, visit type (“new patient” or “follow-up”), status (“cancelled” or “completed”), and duration (in minutes), as well as all available PHQ9 or GAD7 scores were obtained.

In this study, a patient classified as having moderate to severe depression or anxiety disorder at baseline (ie, PHQ9 or GAD7 more than or equal to 10 at the initial BHP visit of an episode of care) was considered to have achieved his/her response goal if their PHQ9 or GAD7 score falls below 10 or is reduced by at least 50% relative to baseline.19,20 For the purposes of this study, an episode of care was defined as the period between the date of the first visit labeled as “new patient visit” (ie, index date) for a given patient and either the date of the last follow-up visit prior to a subsequent “new patient visit” for the same patient or the last follow-up visit observed in the data during the study period (ie, right-censored observations). The maximum length of follow-up period for a given episode was limited to 365 days since the index date.

Psychotherapy dose for each patient was measured by the counts of completed visits scheduled for at least 45 minutes with a BHP therapist within each episode of care. In practice, appointments could have been scheduled for different durations, depending on the purpose of the visit, for example, 30 minutes for medication management-only visits. In addition, the data indicated wide variations in the length and frequency of visits with the BHP therapists. For the purposes of this study, those patients who had fewer than eight 45-minute psychotherapy visits during their episode of care were considered to be in the “low dose” category; those who had 8 to 12 visits were considered to be in the “medium dose” category; and those who had more than 12 visits were considered to be in the “high dose” category.

Furthermore, episodes that had durations of less than 90 days were excluded from the analysis for two reasons: first, for such short episodes, the dose variable would be confounded, that is, for such short episodes, the dose variable would reflect the short duration of time rather than low treatment intensity of the program. Second, such short episodes might reflect the patients’ disengagement from the program (due to, for instance, financial, or other burdens) rather than low treatment intensity.

Back to Top | Article Outline

Statistical Analysis

The outcome variable was the duration of time in a given episode between the initial visit and the first visit in which the patient had achieved their response goal. The predictor was the dose categories described above. Furthermore, patients in the high dose category were also compared against those in the medium dose category to examine whether there is a non-linear relationship between dose and the time to response, that is, the hypothesized correlation between psychotherapy dose and response might not hold after a certain dose level due to the diminishing incremental value of additional visits.

These hypotheses were tested first using non-parametric comparisons across the three psychotherapy dose levels using the Kaplan–Meier (KM) estimators.21 KM estimator plots the probability of the “survival” of the BHP patients, defined as not yet having achieved response at a given point in the follow-up period, across the three dose levels, yielding three distinct downward sloping curves. Separation of the curves illustrate the magnitude of any differences across the dose categories. Log-rank tests were conducted to assess the statistical significance of the separation of the curves. Each episode was observed until its completion or for a maximum of 365 days after the initial visit, whichever was earlier (ie, the follow-up period). Those episodes for which the patients had not achieved a response by the end of the follow-up period were considered to be right-censored and treated accordingly in the analysis.

In addition, Cox proportional hazard regression models21 were estimated to examine whether the adjusted hazard ratios associated with the dose level categories were different from one another, controlling for a set of covariates including patients’ sex, age, race/ethnicity, baseline PHQ9 or GAD7 scores, and calendar year (to account for the potential effect of BHP program maturity over time), as well as a set of interaction terms between sex and age (to account for any sex-age specific confounders, such as pregnancy). Because the unit of analysis was each episode and there could be multiple episodes associated with a single patient, clustered standard errors by the patient ID were obtained to account for the possible correlation of the error terms among the episodes associated with the same patient.21

Back to Top | Article Outline


Table 1 summarizes the characteristics of the overall sample, representing all patients ever treated by the BHP program as well as the subsample of the episodes for which the patient has been identified as having moderate to severe depression or anxiety disorder at baseline. On average, a BHP program patient had approximately two episodes of care, and each episode included approximately 11 visits over roughly a 4-month period. The corresponding standard deviations were larger than the mean values, suggesting that there was wide variation in these variables. The sample indicates a population that is on average younger than 40 years old and predominantly white female.



Figures 1 and 2 show the KM survival curves of the subsample of the BHP episodes of care for which the patients were identified as having moderate to severe depression (Fig. 1) or anxiety disorder (Fig. 2), stratified by the three psychotherapy dose levels. Both figures indicate that those patients who had the low dose level were consistently less likely to achieve response than those in either the medium or the high dose levels at any point in time during the follow-up period. Furthermore, the figures also suggest that there is little difference between the medium and high dose levels of psychotherapy, as indicated by the corresponding curves that are in close proximity of each other.





Table 2 shows the adjusted hazard ratios associated with the three psychotherapy dose levels using the Cox proportional hazard models (see Appendix [], for the full regression model outputs). These regression-adjusted estimates are consistent with the patterns observed in the KM results shown in Figures 1 and 2. Specifically, the results suggest that the patients who were in the medium to high dose levels had nearly twice the hazard rates of achieving response as those of the patients who were in the low dose level—differences that are statistically significant at P = 0.05 level in all cases. Furthermore, the differences between the medium and high dose levels are not statistically significant, as the corresponding 95% confidence intervals overlap with each other.



Back to Top | Article Outline


Psychotherapy dose–response relationship has long been recognized as an important predictor of improved health outcomes,14,15 and the results of this study suggest that such a correlation also exists in the case of this employer-sponsored behavioral health program. Approximately one-third of the BHP patients in the sample fall in the low psychotherapy dose category (less than eight session per episode of care). This implies that a significant number of BHP patients may not be reaping the full benefits of the program. The reason for this is not obvious. The program has been reaching out to different categories of employees throughout UR, raising awareness about the program and providing education about the services during employee health fairs and new employee orientation, via blogs published on line, as well as other targeted outreach to vulnerable groups such as medical and surgical residents who have been identified as having higher levels of stress and burnouts. Still, barriers to participation in the program are likely to remain and may include employees’ reluctance to seeking mental health care at a program sponsored by their employer. This may be a reflection of the employees’ concerns about confidentiality and privacy as well as the overall stigma related to seeking mental health care in general.22 As such, more research is needed to better understand such barriers and to develop services and strategies that would mitigate them.

Furthermore, the finding that the high dose category was not correlated with better outcomes relative to the medium dose category has an important implications for further optimizing the BHP program to achieve better efficiency. In a separate analysis (shown in the Appendix,, the average number of visits among the patients in the low dose category was approximately four, among those in the medium dose category approximately 10, and among those in the high dose categor approximately 22. This implies that the 12 additional visits in the high dose category relative to the medium dose category did not have any significant incremental impact in terms of additional improvements in the outcome. On the other hand, getting the patients in the low dose category to make six additional visits may have a significant impact in helping them achieve response more quickly. If this holds true, it would suggest that BHP could improve its operational efficiency and its impact on the health and wellness of its patients without necessarily requiring additional resources. However, it is important to note that this calculation does not take into consideration the possibility that further psychotherapy sessions may provide other benefits to the individuals enrolled beyond the improvement in PHQ9 or GAD7 scores. Such benefits may include additional functional improvement, psychological well-being or improvements in interpersonal relationships, all of which were not examined as part of this study.

Several limitations are important to consider in reviewing the results of this study. This study has not attempted to identify an optimal psychotherapy dose level, as this would require more research, taking into consideration the type of therapy, the unique patient characteristics, as well as disease severity. Also, this study does not constitute a true program impact analysis, because a pure comparison group consisting of individuals who would be eligible for the BHP program but were not given access to the same services was not available. Nevertheless, the results are consistent with the expectation that the BHP program had a positive impact on the targeted individuals. Lastly, the study has relied on a retrospective secondary data collected for administrative purposes. Consequently, it is not possible to conclude that there is necessarily a causal relationship between the interventions provided and improved patient outcomes, as other confounding variables may have had an effect on the outcomes reported.

Back to Top | Article Outline


Higher numbers of psychotherapy visits per episode of care among patients in an employer-sponsored behavioral health program were found to be correlated with improved behavioral health outcomes. However, there appears to be no incremental benefit to the patients beyond a certain number of visits. Further research is needed to identify the optimal treatment intensity that is likely to vary by patient characteristics and disease severity.

Back to Top | Article Outline


The authors wish to acknowledge and thank LouAnn Jaeger for providing the data elements needed for this study and Vanessa Mace for providing administrative support. The authors also thank Susan McDaniel, PhD, for reviewing and providing constructive feedback on the manuscript.

Back to Top | Article Outline


1. Kessler RC, Petukhova M, Sampson NA, Zaslavsky AM, Wittchen H-U. Twelve-month and lifetime prevalence and lifetime morbid risk of anxiety and mood disorders in the United States. Int J Methods Psychiatr Res 2012; 21:169–184.
2. Stewart WF, Ricci JA, Chee E, Hahn SR, Morganstein D. Cost of lost productive work time among US workers with depression. JAMA 2003; 289:3135–3144.
3. Wittchen H-U. Generalized anxiety disorder: prevalence, burden, and cost to society. Depress Anxiety 2002; 16:162–171.
4. Goetzel RZ, Long SR, Ozminkowski RJ, Hawkins K, Wang S, Lynch W. Health, absence, disability, and presenteeism cost estimates of certain physical and mental health conditions affecting U.S. employers. J Occup Environ Med 2004; 46:398–412.
5. Beck A, Crain AL, Solberg LI, et al. Severity of depression and magnitude of productivity loss. Ann Fam Medicine 2011; 9:305–311.
6. Khan A, Faucett J, Lichtenberg P, Kirsch I, Brown WA. A systematic review of comparative efficacy of treatments and controls for depression. PLoS One 2012; 7:e41778.
7. Craske MG, Roy-Byrne PP, Stein MB, Sullivan G, Sherbourne C, Bystritsky A. Treatment for anxiety disorders: efficacy to effectiveness to implementation. Behav Res Ther 2009; 47:931–937.
8. Putnam K, McKibbin L. Managing workplace depression: an untapped opportunity for occupational health professionals. AAOHN J 2004; 52:122–129. quiz 130-121.
9. Lerner D, Adler D, Hermann RC, et al. Impact of a work-focused intervention on the productivity and symptoms of employees with depression. J Occup Environ Med 2012; 54:128–135.
10. Rothermund E, Kilian R, Rottler E, et al. Improving access to mental health care by delivering psychotherapeutic care in the workplace: a cross-sectional exploratory trial. PLoS One 2017; 12:e0169559.
11. Lam RW, Lutz K, Preece M, Cayley PM, Walker AB. Telephone-administered cognitive-behavioral therapy for clients with depressive symptoms in an employee assistance program: a pilot study. Ann Clin Psychiatry 2011; 23:11–16.
12. LaMontagne AD, Martin A, Page KM, et al. Workplace mental health: developing an integrated intervention approach. BMC Psychiatry 2014; 14:131.
13. Goetzel RZ, Ozminkowski RJ, Sederer LI, Mark TL. The business case for quality mental health services: why employers should care about the mental health and well-being of their employees. J Occup Environ Med 2002; 44:320–330.
14. Hansen NB, Lambert MJ, Forman EM. The psychotherapy dose-response effect and its implications for treatment delivery services. Clin Psychol 2002; 9:329–343.
15. Carey T. Estimating treatment duration for psychotherapy in primary care. J Public Mental Health 2006; 5:23–28.
16. Lam RW, Wolinsky D, Kinsella C, Woo C, Cayley PM, Walker AB. The prevalence and impact of depression in self-referred clients attending an employee assistance program. J Occup Environ Med 2012; 54:1395–1399.
17. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 2001; 16:606–613.
18. Spitzer RL, Kroenke K, Williams JBW, Löwe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med 2006; 166:1092–1097.
19. McMillan D, Gilbody S, Richards D. Defining successful treatment outcome in depression using the PHQ-9: a comparison of methods. J Affect Disord 2010; 127:122–129.
20. Robinson E, Titov N, Andrews G, McIntyre K, Schwencke G, Solley K. Internet treatment for generalized anxiety disorder: a randomized controlled trial comparing clinician vs. technician assistance. PLoS One 2010; 5:e10942.
21. Cleves M, Gould W, Gutierrez R, Marchenko Y. An Introduction to Survival Analysis Using Stata. College Station, TX: Stata Press; 2010.
22. Magaard JL, Seeralan T, Schulz H, Brutt AL. Factors associated with help-seeking behaviour among individuals with major depression: a systematic review. PLoS One 2017; 12:e0176730.

anxiety; behavioral health; depression; dose–response; employee wellness; the Generalized Anxiety Disorder; mental health; Patient Health Questionnaire

Supplemental Digital Content

Back to Top | Article Outline
Copyright © 2019 by the American College of Occupational and Environmental Medicine