Cost Effectiveness of a Web-Based Program for Residual Depressive Symptoms - Mindful Mood Balance

Abstract

Objective:

Mindful Mood Balance (MMB) is an effective web-based program for residual depressive symptoms that prevents major depressive relapse among patients who partially recovered from major depressive episodes. This is a cost-effectiveness analysis from the health plan perspective alongside a pragmatic randomized controlled trial of MMB.

Methods:

Adult patients were recruited and randomized to MMB + Usual Depression Care or Usual Depression Care (UCD) from behavioral health and primary care settings in a large integrated health system. Patients had >=1 prior major depressive episode, current PHQ-9 score of 5–9 indicating residual depressive symptoms, and internet access. Program costs included recruitment, coaching and MMB licensing. Center for Medicare and Medicaid fee schedules were applied to EHR utilization for all psychotropic medications, psychiatric and psychotherapy visits. Effectiveness was measured as Depression Free Days, converted from PHQ-9 scores collected every month during 1-year follow-up. Incremental Cost Effectiveness Ratios were calculated with different sets of cost-inputs.

Results:

There were 389 patients (210 UCD and 179 MMB+UDC) included with adequate follow-up PHQ-9 measures over 12 months. MMB patients had 28 more depression free days during the 12-month follow-up period. Overall, the incremental cost of MMB+UDC was $431.54 over 12-months. Incremental costs per depression free day gained ranged from $9.63 for program costs only to $15.04 when psychiatric visits, psychotherapy visits and psychotropic medications were included.

Conclusion:

MMB offers a cost-effective web-based program for reducing residual depressive symptoms and preventing major depressive relapse that health systems should consider adopting as adjunctive to traditional mental health care services.

Introduction

Major depression is a highly prevalent chronic condition with high societal cost and was the leading cause of disability in 2017 (1). Depression is associated with low education and economic attainment, marital disruption and suicide, each of which incur costs from lost productivity and health/mental health treatment (2–4). Residual depressive symptoms, which occur when patients partially recover from a major depressive episode, involve pervasive ongoing burden that is often left untreated and increases the risk for major depressive relapse (5–6).

Mindfulness-based cognitive therapy (MBCT) is an empirically supported program that teaches emotion-regulation skills for reducing residual depressive symptoms and confers a 43% reduction in relapse risk (7–10). Like many psychotherapies, MBCT faces dissemination challenges including shortages of MBCT-trained therapists, service cost, and logistical hurdles (11). These access barriers provided the rationale for a web-based program informed by MBCT, called Mindful Mood Balance (MMB). MMB was evaluated in a large RCT (N=460) and found effective for reducing residual depressive symptoms over a 12-month follow-up period. These data suggest that MMB may be a scalable solution for expanding availability of MBCT.

Among the few cost-effectiveness analyses of web-based depression treatments, results have been mixed. Some have shown cost-effectiveness within $20,000 or $50,000 per year willingness to pay thresholds (which are likely too high for U.S. health plans to consider), but none have shown significant reductions from averted health care utilization (12–15). Internet delivered cognitive behavioral therapy programs average 50% probability for cost-effectiveness at a $0 willingness to pay threshold (16). Web-based programs may promote help seeking behavior causing increased outpatient service utilization. Traditionally, this may be interpreted as a detrimental impact of web-based programs from a cost-effectiveness standpoint, but the context of resource and access limitations within the U.S. mental health care system must also be considered.

Cost effectiveness analyses from the health plan perspective, which directly examine actual costs associated with implementing the program as well as health care service utilization changes attributable to the program, are particularly important for health plan decision makers considering the multitude of digital programs available. As the Centers for Medicaid and Medicare explores the integration of behavior health services into capitated systems, investing in effective, low-cost approaches to mental health care is particularly important (17).

This cost-effectiveness analysis from the health plan perspective of MMB was deployed inside a randomized controlled trial. The effectiveness outcomes of the randomized trial indicated significant reductions in both residual depressive symptoms and major depressive relapse (18). The hypothesis was that MMB would have a low cost per depression-free-day (a measure of depression severity used in longitudinal follow-up studies) within 1-year, relative to comparable programs that are typically available in health systems to treat patients to remission.

Methods

Participants

Participants were recruited from Kaiser Permanente Colorado (KPCO), an integrated health system serving approximately 630,000 patients from the Denver, Boulder, and surrounding metropolitan areas. Participants were randomly assigned to receive Mindful Mood Balance and Usual Depression Care (MMB+UDC: N= 230) or Usual Depression Care alone (UDC: N=230), with recruitment spanning from March 2, 2015 to November 30, 2018. Recruitment occurred through clinic flyer and provider referral, but primarily by administratively identifying potentially eligible patients with a depression diagnosis and a qualifying Patient Health Questionnaire-9 (PHQ-9) score during a routine visit. Patients were then sent an invitation email for the trial that included a consent website (19). Inclusion criteria were minimal and included: >=18 years of age (determined by medical record), at least one prior episode of Major Depressive Episode (medical record diagnosis of depression plus phone interview), current PHQ9 score ≥ 5 and ≤ 9 (initially completed during health care visit and confirmed via phone interview) (20), internet access and intention to continue KPCO membership for at least 6 months (phone interview). Exclusion criteria were determined by medical record review and included: history of ICD9/10 diagnoses within the past 2 years of schizophrenia or current psychosis, organic mental disorder, or pervasive developmental delay.

Study design

The incremental cost-effectiveness of MMB + UDC versus UDC was determined. The study was conducted alongside a single-blind, two-group, randomized controlled trial that reported significant reductions in both depressive symptom severity and relapse of major depression (18). Briefly, through eight self-administered sessions, MMB uses vicarious learning (videos), experiential practice, and didactic information to teach mindfulness and meditation skills and increases awareness of dysfunctional automatic thought processes that contribute to depression.

The study was approved by the Institutional Review Boards of Kaiser Permanente Colorado, University of Toronto, and the University of Colorado Boulder and registered as NCT02190968 at https://clinicaltrials.gov/. All patients provided informed consent and signed a HIPAA authorization for medical record information.

Effectiveness measure

Depression Free Days (DFDs) is a standardized measure for cost-effectiveness studies of cumulative depression severity when there are multiple measurements over time. Patients were assessed for depression severity via monthly PHQ-9 web-based questionnaire for 12 months following the 8 session (12 weeks) MMB program. PHQ-9 scores were converted to DFDs using a method that was previously developed (21). Because the first DFD score was constrained to a 5–9 range, it required a minimum of two measurement points (baseline and 1 follow-up). Further, in prior study, one value for PHQ9 in a 12-month period skews results substantially compared to at least 2 values (21). Sensitivity analyses (not shown) were conducted to examine differences in DFD calculations between groups with different requirements for follow-up PHQ-9 completeness: 1.) One PHQ-9 at anytime, 2.) One initial and one late PHQ-9, 3.) One initial plus ≥ 3 PHQ-9s. Differences of <2 DFDs were observed across these requirements, but more restrictive approaches excluded more patients, so the most inclusive approach was employed of only requiring one PHQ-9 during follow-up (in addition to the baseline PHQ-9 measurement).

Cost measures

Using instruments previously employed for other behavioral health inventions (22, 23), labor costs for recruitment and coaching (MMB only) were collected via task/time tracking logs that were recorded by all study staff for four weeks at the start of the trial, four weeks in the middle of the trial, and four weeks at the end of the trial. Each participant received an average of 2.34 hours of coaching support. A description of coaching activities is provided in Segal et al. 2020. Recruitment tasks included population-based outreach to patients via email and enrollment website tracking. Tasks associated with eligibility screening, survey data collection or other “research only” tasks were excluded. The mean hourly wage for a Health Education Specialist from the U.S. Bureau of Labor Statistics was used to estimate staff salaries (24). Overhead costs of 10% were applied for workspace and equipment (i.e. computers). While patients were not charged for MMB as part of this research trial, MMB developers estimated a cost of $180 per patient in the future, which was included.

All ambulatory and acute health care services utilization were extracted from the EHR and claims databases. Medical costs were estimated using the Standard Relative Resource Cost Algorithm (SRRCA) for all services provided and paid for by KPCO (25, 26). SRRCA applies Centers for Medicare and Medicaid Services fee schedules to standardized claims or EHR derived utilization data to ensure that observed differences are not a result of differing pricing methods and billing rules (bundling of services). These estimates do not however, account for variation in patient cost-sharing or copays. Costs were divided into groups including: outpatient behavioral health visits, psychotropic medication use, emergency and inpatient department visits with primary or secondary psychiatric or substance abuse diagnoses and inflation adjusted using the medical care component of the Consumer Price Index to 2015 US dollars (27, 28) . Total and average costs were estimated during the 12-month follow-up period.

Statistical analysis

Incremental Cost Effectiveness Ratios (ICERs) were calculated with depression free days as the effectiveness measure. Separate ICERs were calculated for different sets of cost inputs to highlight the impact of costs from different sources because health system stakeholders may be more interested in specific cost impacts. Initially, ICERs only considered the labor costs, then other costs were incrementally added for mental health and substance abuse therapy visits, antidepressant medications, and all psychotropic medications. The trial was not powered on cost-outcomes, rather on the effectiveness measures for changes in depression symptoms, therefore, analysis of inpatient and emergency costs is exploratory since this study was not powered to handle large cost fluctuations. Confidence intervals (95%) were constructed using Fieller’s method (29, 30). A sensitivity analyses was conducted with all costs for the 230 patients, which added costs back in for those that were lost to follow-up.

Role of the Funder/Sponsor:

This work was funded by grant MH102229 (PI Segal) from the National Institute of Mental Health.

Results

Among the 460 patients recruited for the trial, 389 (85%) of patients (210 UCD and 179 MMB+UDC) had a PHQ-9 completed during the follow-up period and were included in the main analysis. Table 1 illustrates baseline demographic characteristics and shows that there are no significant differences on any self-reported demographic, risk factor or treatment history variables. Patients in the MMB+UDC group had an average of 28.72 more DFDs in the 12-month follow-up period compared to UDC (IQR = 40.91).

Table 1.

Baseline characteristics of patients

Variable	Level	UDC (value)	UDC (% or IQR)	MMB+ UDC (value)	MMB+ UDC (% or IQR)	p-val
N		210		179
Sex	Female	159	76.4%	136	76.0%	0.91
Age	Median, IQR	49	35–61	47	35–60	0.99
Education	No HS	3	1.4%	2	1.1%	0.78
	HS	21	10.1%	21	11.9%
	College	85	40.9%	64	36.2%
	Grad School +	99	47.6%	90	50.8%
Race	Asian	4	1.9%	3	1.7%	0.95
	Black	4	1.9%	4	2.2%
	White	187	90.3%	163	91.6%
	AI/AN, Hawaiian, Other	15	7.1%	9	5%
Hispanic Ethnicity		16	7.9%	15	8.6%	0.79
Marital Status	Never married	44	21.3%	47	26.4%	0.63
	Married/ Domestic Partnership	105	50%	77	43%
	Divorced	47	22.7%	44	24.7%
	Separated	4	1.9%	4	2.2%
	Widowed	7	3.4%	6	3.4%
Annual income	$0–29,999	22	10.6%	18	10.2%	0.94
	$30,000 – $69,999	81	39.1%	70	39.8%
	$70,000 – $99,999	53	25.6%	41	23.3%
	$100,000+	51	24.6%	47	26.7%
Employment	Disabled	6	2.9%	4	2.2%	0.84
	Student	10	4.8%	8	4.5%
	Full Time Work	108	51.4%	97	54.2%
	Homemaker	7	3.3%	4	2.2%
	Other	10	4.8%	6	3.3%
	Part Time Work	28	13.3%	16	8.9%
	Retired	36	17.1%	40	22.3%
	Unemployed	5	2.4%	4	2.2%
Ever hospitalized for depression		31	15.0%	28	15.8%	0.82
Ever suicide attempt		37	18.0%	27	15.3%	0.48
Ever treated for drug/alcohol		18	8.7%	11	6.2%	0.37
Frequency Marijuana	None	159	76.4%	139	78.5%	0.83
	Less than 1x/month	20	9.6%	11	6.2%
	1–2x per month	10	4.8%	12	6.8%
	At least 1x per week	19	9%	15	8.4%
Current or previous counseling		174	83.7%	151	84.8%	0.75
Currently taking antidepressants		160	77.3%	141	78.8%	0.73

During the 12-month follow-up period, patients in MMB+UDC had slightly higher outpatient mental health costs, mostly driven by more psychotherapy visits ($27.05) (Table 2). Second generation anti-psychotics are substantially higher for the MMB+UDC group, which was investigated further and found to be driven by 18 more dispensings for Aripiprazole (Abilify) among <5 patients. Abilify was a high-cost medication during the study and made up over 90% of the costs in that category for both groups. Overall, the incremental cost of MMB+UDC was $431.94.

Table 2 –

Average costs per patient during 12-month follow-up

Cost type	Cost description	UDC Total visits/dis pensings	UDC (N=210)	UDC+MMB Total visits/dispens ings	MMB+UDC (N = 179)	Cost differences
Outpatient mental health costs
	Med management visits	39	$23.32	40	$27.38	$4.06
	Psychotherapy visits	364	$197.22	432	$224.27	$27.05
Prescription costs
	Antidepressants	1132	$1,342.80	933	$1,350.97	$8.17
	Antipsychotics - 2nd gen.	26	$74.68	40	$225.05	$150.37
	Anxiety & Sleep	164	$48.27	203	$58.36	$10.09
	Anticonvulsants	92	$160.07	83	$131.18	$−28.90
	Stimulants (Attention)	164	$125.67	93	$138.64	$12.97
	Other psychotropic meds	22	$40.00	17	$11.45	$−28.54
Emergency and inpatient	Emergency department	43	$220.44	39	$363.64	$−143.20
	Inpatient visit	19	$1350.51	11	$774.27	$−576.24
Labor
	Recruitment costs	n/a	$83.48	n/a	$83.48	$ -
	Coaching costs	n/a	$ -	n/a	$96.67	$96.67
MMB licensing		n/a	$ -	n/a	$180	$180

Notes: Anxiety and sleep medications include benzodiazepines, Z-drugs, and other hypnotic medications. Other psychotropic meds include non-stimulant attentional drugs, first generation antipsychotics, combination drugs with two agents, lithium, and injectable antipsychotics.

Incremental costs per depression free day gained ranged from $9.63 to $15.04 for different cost inputs (Table 3). When only MMB licensing and labor costs for recruitment and coaching are included, the incremental costs were $9.63 for each additional DFD. Negligible increases to $10.72 and $11.00 occurred when adding in outpatient mental health services and antidepressants, which represent the highest overall service costs for this sample. A larger increase in cost per depression free day ($15.04) was observed when adding in all psychotropic costs, which was driven by more dispensings of Abilify to the MMB+UDC group. Figure 1 shows that at $15.04 willingness to pay for an additional depression free day, MMB is cost neutral.

Table 3.

Incremental Cost Effectiveness of MMB Compared to Usual Care

Description of cost	UDC costs	MMB+UDC Costs	Incremental costs per DFD gained	Incremental cost 95% CI
Labor & MMB costs only	$83.48	$ 360.15	$ 9.63	*n/a
Labor, MMB, MH Services	$304.02	$ 611.79	$ 10.72	$10.23 – 11.21
Labor, MMB, MH Services, Antidepressants	$1,646.82	$ 1,962.76	$ 11.00	$10.77 – 11.24
Labor, MMB, MH Services, All Psychotropics	$2,095.52	$ 2,527.45	$ 15.04	$13.8–16.71
Labor, MMB, MH Services, All Psychotropics, Emergency, Inpatient	$ 3666.46	$ 3665.35	$ (0.04)	$−1.32 to 1.25

Notes: DFD = Depression Free Day. UDC = Usual Depression Care. MMB = Mindful Mood Balance. The incremental effectiveness for all calculations was 28.72 Depression Free Days. Labor costs for UDC includes recruitment costs only, MMB includes licensing, recruitment and coaching costs. 95% CI calculated using Fieller’s Theorem (29). Labor and MMB licensing costs were not collected at patient level, but averaged across all patients within group from staff logs of cost on specific tasks. Standard deviation is therefore 0 and confidence intervals aren’t calculatable.

Figure 1.

Incremental Net Benefit ($) at Different Willingness to Pay Thresholds

In the sensitivity analysis, patients that were excluded because they did not have follow-up PHQ-9s, had negligible differences for medication, psychiatric and psychotherapy services (Table 4). Overall, this shows that those who were lost to follow-up did not seek more outpatient care, but their clinical outcomes remain unknown. There were higher emergency and inpatient stay utilization for the UDC group (17 vs. 10 in-patient stays between UDC/MMB+UDC) in the exploratory analyses which showed −0.04 per depression free, however this cost difference disappeared when lost to follow-up patients were included ($13.84 per depression free day).

Table 4.

Sensitivity Analyses for Cost Effectiveness of MMB+UDC Compared to Usual Care

Description of costs	Incremental costs per DFD (sensitivity)	Incremental costs per DFD (main)
Labor and MMB costs only	$ 9.63	$ 9.63
Labor, MMB, MH Services	$ 10.37	$ 10.72
Labor, MMB, MH Services, Antidepressants	$ 9.09	$ 11.00
Labor, MMB, MH Services, All Psychotropics	$ 12.71	$ 15.04
Labor, MMB, MH Services, All Psychotropics, Emergency and Inpatient	$ 13.84	$ (0.04)

Notes: DFD = Depression Free Day. Sensitivity analysis includes all patients regardless of PHQ-9 follow-up information. Main analysis eliminated patients who did not have at least one PHQ-9 reported during the 12-month follow-up.

Discussion

Health insurance companies are now offering several mental health digital programs to their members as a benefit (31); however, few have undergone rigorous evaluation to understand costs (32). This cost effectiveness study compared the MMB program to usual depression care by examining program, outpatient mental health services, and psychotropic medication cost differences within a large integrated health system. MMB had a cost per depression free day of $9.63 for program costs only and $15.04 when all medication and services costs were included. Most of the increase over program costs was attributed to high-cost medication choice and very little to differences in outpatient psychotherapy or psychiatric visits. Results show that MMB offers an effective and low-cost adjunctive service to traditional mental health services for treatment of residual depressive symptoms that health systems should consider adopting.

There were no significant cost offsets for outpatient mental health care or antidepressant medication use during the follow-up period. However, this is not surprising based on prior studies of MBCT and antidepressant medications (33). While the meditation and mindfulness skills learned in MMB may help avert mental health services utilization, MMB also teaches recognition of warning signs to know when to engage services, which may increase treatment seeking.

The incremental costs per patient for MMB were $431.54 for a gain of 29 depression free days. This is respectable compared to other prominent programs for depression offered through integrated health systems. For example, cost effectiveness studies of collaborative care management for depression, which typically involve both in-person and telephone support, have incremental costs of $454 - $960 per patient over usual care with 18 – 58 incremental depression free days gained over a year (32, 34–37). The incremental costs of $431.54 are comparable to other effective web-based depression programs reporting incremental costs of $52 with little coaching/therapist cost, but limited engagement (15) to $703 (13) for synchronous therapist contact and high engagement. This would indicate that MMB may be equivalent to collaborative care in overall cost-effectiveness and within typical cost ranges for other web-based depression programs with some coaching support.

The literature is sparse on robust cost estimates for in-person Mindfulness Based Cognitive Therapy (MBCT) compared to usual care, but two studies offer estimates of the labor cost for in-person MBCT program. Cost per patient for an altered version called MBCT with support to taper antidepressant medication (MBCT-TS) of approximately £112 (which converts to $168 in 2015) (33). An earlier study of traditional MBCT estimates that the labor costs per patient of the in-person program was the equivalent of 3 contact hours with a therapist, which would be approximately $300 based on the Medicare fee schedule (7). The equivalent from the current study to these estimates is the coaching costs, which were $96.67 per patient over the course of MMB. Importantly, the hazard ratios for depressive relapse from in-person MBCT in a meta-analysis compared to (hazard ratio, 0.69; 95% CI, 0.58–0.82) (10) web-based MMB were nearly the same (hazard ratio, 0.61; 95% CI, 0.39–0.95) (18). This indicates that web-based MMB with email/phone coaching offers a substantially lower-cost alternative than in-person MBCT with nearly equivalent outcomes.

It is important to consider that patients who decide to enroll in a study for MMB may have different preferences from those choosing in-person MBCT or another in-person program like collaborative care for depression. Web-based MMB offers an alternative for patients with logistical barriers, social anxiety, or simply those who do not prefer in-person group-based programs (38). Conversely, there are patients who may learn best in an in-person group-environment or those without stable internet access, who would not be able to do a web-based program. These barriers may explain the reason that 37% of patients did not complete at least 4 sessions of MMB (considered an adequate dose), but this attrition rate is smaller than 50+% reported for prior web-based depression programs (39). Regardless, future studies of MMB and other web-based depression treatments should examine mediators of successful treatment response and attrition to improve delivery of these programs, especially since treatment failures and delay exacerbate the severity and chronicity of illness course (40). To address the burden of depression, it is important that multiple modalities of treatment be offered that address differences in patients’ preferences and access. This analysis indicates that MMB offers a highly feasible and scalable model for health systems looking to expand their offering of evidence-based programs for depression.

This study has several advantages and some limitations compared to prior cost-effectiveness studies of web-based depression programs. Many studies have omitted capture of program costs (e.g staff time to engage and retain patients) or used patient self-report measures for service utilization, whereas this study captures the costs associated with implementing the program and estimates actual service costs collected from electronic health records (EHR) and claims within a capitated health plan. One limitation is the potential for missing mental health service costs paid privately outside the health system; however, these should be minimal because private mental health care would have higher out of pocket costs compared to the covered benefits reported here. The sample was predominantly white females, which may impact generalizability with respect to race, ethnicity, and gender. Data for patient time costs, presenteeism and absenteeism were not available, so an analysis from a societal perspective was not possible. Another limitation is that the sample size was too small and the costs variability too high for emergency department and inpatient costs associated with mental health, substance abuse, suicide ideation and behavior to be interpretable. This information was included for transparency, but it is not likely accurate to say that MMB is cost-saving considering the inclusion of a handful of lost to follow-up patients changed the cost per DFD from −.04 to $13.84. The latter is close to results when only outpatient and medication costs were included. This is also illustrated in the large cost differences associated with Abilify for a small additional number of dispensings in the MMB group. Abilify is prescribed to patients as adjunctive to antidepressants when antidepressants alone lack adequate effectiveness (41). This study was powered on the effectiveness outcomes and not on cost outcomes. Therefore, this study is not equipped to absorb these differences associated with high cost-variance from emergency, inpatient and high cost medications. The total visits/dispensings are included in Table 2 to show proportionally that most of the utilization lies within antidepressants and outpatient visits. When only antidepressants medications were incorporated, the incremental cost per depression free day was $11.00. Despite this, the more conservative estimate that incorporates Abilify is reported because it is possible that MMB caused more treatment-seeking behavior resulting in the additional dispensings. This study does not report QALYs, as this metric is rarely used by policy makers (e.g CMS, USPSTF) or by health system decision makers in the U.S and there is controversy about their use due to potential violation of the Americans with Disability Act (42).

Conclusions

From a health plan perspective, this study evaluated the cost-effectiveness of a web-based program for patients with residual depressive symptoms, Mindful Mood Balance (MMB), over a 1-year follow-up, considering all program and mental health service utilization costs. At a $0 willingness to pay, MMB would cost $15.04 per depression free day. MMB is a cost-effective web-based program for reducing residual depressive symptoms and preventing major depressive relapse that health systems should consider adopting as adjunctive to traditional mental health care services.

Highlights.

• This is a cost-effectiveness analysis of Mindful Mood Balance, a web-based program for reducing residual depressive symptoms and preventing major depressive relapse that is closely aligned with mindfulness based cognitive therapy.

• Compared to usual depression care, Mindful Mood Balance produced 29 more depression free days over 1-year costing $10-$15 per depression free day with an incremental cost of $432.

• Mindful Mood Balance offers a cost-effective web-based program for reducing residual depressive symptoms and preventing major depressive relapse that has comparable or lower costs to other web-based depression programs.