This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Mental Health, is properly cited. The complete bibliographic information, a link to the original publication on http://mental.jmir.org/, as well as this copyright and license information must be included.
Cognitive behavioral therapy (CBT) is the gold standard treatment for adult anxiety disorders but is often not readily available in a scalable manner in many clinical settings.
This study examines the feasibility, acceptability, and effectiveness of a coach-facilitated digital cognitive behavioral program for anxious adults in primary care.
In an open trial, patients who screened positive for anxiety (General Anxiety Disorder-7 [GAD7] score ≥5) were offered the digital cognitive behavioral program (active group, n=593). Primary outcomes included anxiety, quality of life (QoL), and ambulatory medical use over 6 months. Intent-to-treat (ITT) and modified intent-to-treat (mITT) analyses were completed. Subsequently, we compared the outcomes of participants with those of a matched control group receiving primary care as usual (CAU; n=316).
More than half of the patients downloaded the cognitive behavioral mobile app program and about 60% of these were considered engaged, which was defined as completion of ≥3 techniques. The active group demonstrated medium size effects on reducing anxiety symptoms (effect size
A coach-facilitated digital cognitive behavioral program prescribed in primary care is feasible and acceptable. Primary care patients prescribed a digital cognitive behavioral program for anxiety experienced significant improvements in anxiety symptoms, QoL, and reduced medical utilization. This effect was observed even among patients with chronic medical conditions and behavioral health comorbidities. Although the primary outcomes in the active group did not improve significantly more than the CAU group, health care utilization declined, and some secondary outcomes improved in participants who engaged in the program compared to the CAU group.
ClinicalTrials.gov NCT03186872; https://clinicaltrials.gov/ct2/show/NCT03186872
Untreated anxiety disorders are common and their management is expensive, particularly in medical settings where they can often drive excessive health care utilization [
Despite the valuable CBT dissemination and implementation work and substantiated models of integrating CB approaches into primary care, scalability remains a problem [
This study is an open trial aimed at determining the feasibility, acceptability, and effectiveness of a digital coach-facilitated CB program for patients receiving routine primary care. In addition, we aimed to compare the primary care sites providing a digital CB program and sites practicing usual primary care. The pragmatic study design allows evaluation of a digital behavioral intervention in routine medical care with minimal disruption of the clinical ecosystem. Although symptom reduction is a key outcome, there is a critical need to address behavioral health factors that contribute to inappropriate medical utilization. Building a scalable approach to effectively manage anxiety symptoms and reducing the negative health and financial impact of unmanaged behavioral health conditions are essential in the era of population health care. We hypothesize that integration of a digital CB program for anxiety within primary care will allow for better access to evidence-based behavioral health care and be effective at reducing anxiety, improving quality of life (QoL), and decreasing high and potentially inappropriate medical utilization. We subsequently compare these outcomes with those of primary care as usual.
All study procedures complied with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, revised in 2008. This prospective pragmatic open trial (Trial Registration: ClinicalTrials.gov NCT03186872) evaluated a digital CB program over a 6-month time period within two primary care sites. The treatment group was then compared to a matched control group from two comparable primary care clinics. Details about the primary care clinics are provided in previous publications [
The digital CB program Lantern [
Lantern was developed in partnership with academic settings, treatment developers, and experts in digital mental health. A scientific advisory board reviewed the entirety of the program to ensure clinical validity. Ideally, users learn all six core skill components to equip themselves with the empirically supported skills to most effectively manage anxiety. The six core skill components of the anxiety program are delivered throughout a series of 40 brief 10- to 15-minute interactive units that introduce a total of 26 techniques, providing behavioral (eg, diaphragmatic breathing) and cognitive (eg, thought-challenging exercises) tools for practice, which users can complete quickly and apply immediately. Patients accessed Lantern via mobile phones (
The Lantern program includes integrated asynchronous texting with a human coach for personalized motivational behavioral coaching. The Lantern coaches provide human support to increase engagement by using motivational techniques, answering questions, monitoring progress throughout the intervention, facilitating goal setting, and reinforcing the content and skills presented to help shape the skill practice into the users’ daily lives. They provide skill coaching through brief text messages through the app. The coaches rely on an internal coaching portal that provides a dashboard for each user. The dashboard shows all user inputs in the program (direct messages to coaches and all content they have completed in the program). Coaches respond to each user a maximum of once per day.
Coaches had backgrounds in health/wellness coaching or education or mental health treatment. All coaches had at least a bachelor’s degree, and the majority of coaches had an advanced degree in their respective fields (master’s level). A doctoral level licensed psychologist supervised the coaches. Coaches were trained in CBT theory and applied techniques, CBT-specific skill-coaching framework, and risk-management protocols. All user messages were reviewed daily for potential risk.
Core Lantern components.
Core component | Description |
Education | Education about anxiety and relationships between thoughts, emotions, and behaviors and what perpetuates anxiety |
Relaxation | Relaxation exercises such as diaphragmatic breathing, guided imagery, and progressive muscle relaxation |
Cognitive Restructuring | Guidance and techniques to challenge anxiety-maintaining thoughts in the service of developing more adaptive thoughts |
Exposure | Further education about the relationship between anxiety and avoidance and how systematic exposure can interrupt the anxiety cycle and facilitate new learning |
Mindfulness | Mindfulness explanation and exercises for observing thoughts and emotions without judgement |
Habit Formation and Skills Generalization | How to generalize Lantern skills to one’s life and sustain healthy habits |
Example of Lantern screenshots.
The comparison sites with similar site demographic characteristics to the active practices were offered primary care as usual (CAU), wherein the primary care physician (PCP) managed anxiety as per the standard protocol after positive screening. To enhance CAU at the comparison sites, staff were trained to offer patients a National Institute of Mental Health educational brochure about GAD and its treatment [
Patients were enrolled over a 1-year period (2016-2017). Eligibility criteria included age of 20-65 years; ability to speak English; a GAD-7 score ≥5, which is a screening measure part of routine primary care; and access to a mobile phone. At the clinic visit check-in, when a patient scored ≥5 points on the GAD-7, a best-practice alert was generated in the electronic medical record (EMR) to prompt the PCP that the patient is eligible for the digital CB intervention. The PCP decided whether it was appropriate to offer this behavioral health intervention to the patient. If it was appropriate, the PCP referred the patient to the program during the visit. After the patient consented to the study, the PCP ordered the digital CB program via the EMR. Participants could download the mobile app and sign up for the program during the office visit or after the medical appointment. Each participant had access to the digital CB program for 2 years after enrollment.
The comparison group (matched controls) included patients between 20 and 65 years of age who spoke English and had a GAD-7 score ≥5 at screening and a second GAD-7 score 6-months after screening.
Data including date of birth, gender, ethnicity, race, and insurance type were abstracted from the EMR.
The GAD-7 scale is a validated self-report questionnaire used to identify probable cases of evaluating the severity of seven diagnostic GAD symptoms occurring in the past 2 weeks [
The Short Form Health Survey (SF-12) is a 12-item validated self-report measure assessing health-related QoL (HRQoL) [
This 23-item self-report measure assesses 13 mental health domains across a range of psychiatric diagnoses [
Information on 10th revision of the International Statistical Classification of Diseases and Related Health Problems (ICD-10) code diagnoses during the lifetime, psychotropic medication history in the past 6 months, and medical utilization was obtained from the EMR. A medical complexity score was derived based on the total number of ICD-10 diagnoses in the EMR. Medical utilization includes primary care and specialty medical care ambulatory visits. Information on medical utilization was requested 6 months prior to and 6 months after study enrollment. High utilization was defined as ≥4 outpatient medical visits in the 6 months prior to the study, which was at the 75th percentile.
This 14-item self-report scale was administered to patients in the active condition. Respondents rated the helpfulness of and satisfaction with Lantern, length of the program, and the coach and reported the likelihood of recommending Lantern to family/friends.
Participant flow sheet.
Usage was assessed within the Lantern mobile app for each participant. Usage metrics included the number of units completed, the number of techniques practiced, and the number of days the participant logged into the Lantern mobile app over 6 months. Engagement was defined as completion of ≥3 techniques.
Data analytic methods were specified
Approximately 60% of the active group was missing the 6-month data (n=369) because of the manner in which the active group was enrolled and the naturalistic nature of this pragmatic trial. Consequently, we did not perform imputation as per published recommendations [
To test our first aim of a pre-post within-group trial, we characterized baseline characteristics (
An intent-to-treat (ITT) analysis was performed for the active group participants with baseline and 6-month GAD-7 scores; in addition, a modified ITT (mITT) analysis was performed for participants who completed 3 or more techniques. The mITT analysis facilitates interpretation of the ITT findings by providing better estimates of effects among the patients initiating the intervention [
Using the full (unmatched) sample, we performed descriptive statistics,
Using the final matched sample (N=316), we performed comparison analyses. We computed the Cohen
Matched group comparisons at baseline.
Characteristics | Original unmatched samples | Matched samples | |||||||||||
Comparison group (n=219) | Active group (n=593) | Comparison group (n=158) | Active group (n=158) | ||||||||||
Gender - female, n (%) | 143 (65.3) | 426 (71.8) | .09 | 100 (63.3) | 100 (63.3) | 1.00 | |||||||
Age in years, mean (SD) | 47.76 (11.94) | 41.49 (13.06) | <.001 | 45.86 (12.64) | 44.03 (13.38) | .21 | |||||||
Race - white, n (%) | 212 (96.8) | 537 (90.6) | .01 | 151 (95.6) | 149 (94.3) | .80 | |||||||
GAD-7a, mean (SD) | 10.40 (4.9) | 11.24 (4.90) | .03 | 10.46 (5.10) | 10.91 (4.97) | .43 | |||||||
Mental health | 35.51 (10.18) | 35.55 (9.90) | .97 | 36.04 (10.27) | 35.23 (9.44) | .46 | |||||||
Physical health | 37.04 (12.36) | 42.61 (11.59) | <.001 | 40.21 (11.93) | 40.88 (12.27) | .63 | |||||||
Medical complexity score | 3.19 (2.44) | 2.26 (2.07) | <.001 | 2.75 (2.28) | 2.73 (2.28) | .94 | |||||||
High utilizer prior to studyd, n (%) | 82 (37.44) | 166 (27.99) | .01 | 43 (27.22) | 43 (27.22) | 1.00 | |||||||
Any inpatient encounter in the past 6 months | 30 (13.70) | 60 (10.12) | .19 | 14 (8.86) | 20 (12.66) | .36 | |||||||
Any emergency room visit in the past 6 months | 53 (24.20) | 116 (19.56) | .18 | 34 (21.52) | 31 (19.62) | .78 | |||||||
Medicare | 44 (20.09) | 48 (8.36) | <.001 | 23 (14.56) | 15 (9.49) | .23 | |||||||
Medicaide | 64 (29.22) | 104 (18.12) | .001 | 45 (28.48) | 29 (18.35) | .05 | |||||||
Commerciale | 111 (50.68) | 422 (73.52) | <.001 | 90 (56.96) | 114 (72.15) | .01 | |||||||
Other antidepressants | 102 (46.6) | 309 (52.11) | .19 | 80 (50.6) | 85 (53.8) | .65 | |||||||
Tricyclic antidepressants | 15 (6.9) | 31 (5.23) | .47 | 8 (5.1) | 10 (6.3) | .81 | |||||||
Benzodiazapines | 57 (26.0) | 122 (20.6) | .12 | 42 (26.6) | 42 (26.6) | 1.00 | |||||||
Sleep medications | 20 (9.1) | 49 (8.3) | .80 | 10 (6.3) | 15 (9.5) | .40 | |||||||
Stimulants | 5 (2.3) | 6 (1.0) | .29 | 4 (2.5) | 1 (0.6) | .37 | |||||||
Mood stabilizers | 46 (21.0) | 56 (9.4) | <.001 | 26 (16.5) | 22 (14.0) | .64 | |||||||
Buspirone | 7 (3.2) | 11 (1.9) | .38 | 4 (2.5) | 4 (2.5) | 1.00 | |||||||
Atypical antipsychotics | 11 (5.0) | 26 (4.4) | .84 | 4 (2.5) | 7 (4.4) | .54 | |||||||
Opioids | 82 (37.4) | 100 (16.9) | <.001 | 31 (19.6) | 33 (20.9) | .89 | |||||||
Depression | 91 (41.6) | 230 (38.8) | .526 | 55 (34.8) | 70 (44.3) | .11 | |||||||
Anxiety | 92 (42.0) | 248 (41.8) | 1 | 61 (38.6) | 73 (46.2) | .21 | |||||||
Substance abuse | 7 (3.2) | 22 (3.7) | .89 | 4 (2.5) | 6 (3.8) | .75 | |||||||
ADHDg | 4 (1.8) | 6 (1.0) | .56 | 4 (2.5) | 2 (1.3) | .68 | |||||||
Hypertension | 83 (37.9) | 118 (19.9) | <.001 | 54 (34.2) | 44 (27.9) | .27 | |||||||
Chronic GIi condition | 66 (30.1) | 131 (22.1) | .02 | 43 (27.2) | 45 (28.5) | .90 | |||||||
Hyperlipidemia | 60 (27.4) | 93 (15.7) | <.001 | 39 (24.7) | 38 (24.1) | 1.00 | |||||||
Obesity | 48 (21.9) | 91 (15.4) | .04 | 28 (17.7) | 28 (17.7) | 1.00 | |||||||
Asthma | 45 (20.6) | 71 (12.0) | .003 | 27 (17.1) | 21 (13.3) | .43 | |||||||
Diabetesb | 44 (20.1) | 41 (6.9) | <.001 | 28 (17.7) | 14 (8.9) | .03 | |||||||
Chronic pain | 35 (16.0) | 83 (14.00) | .55 | 19 (12.0) | 23 (14.6) | .62 | |||||||
Arthritis | 37 (16.9) | 54 (9.1) | .003 | 22 (13.9) | 15 (9.5) | .29 |
aGAD-7: Generalized Anxiety Disorder-7.
bSF-12: Short Form Health Survey.
cFor the original unmatched samples, n=216 in the comparison group and n=559 in the active group.
dDefined as four or more outpatient medical encounters in 6 months prior to study enrollment.
e
fICD-10: International Statistical Classification of Diseases and Related Health Problems, 10th revision.
gADHD: attention-deficit hyperactivity disorder.
hReporting only those medical conditions with at least 10% prevalence.
iGI: gastrointestinal.
Baseline characteristics are presented in
One of two feasibility/acceptability
A total of 310 participants (of 593, 52.3%) downloaded the Lantern mobile app and completed an average of 6.2 units (SD 9.4, median=2.0, range=0-40 units). Completion of the first 6 units indicated that the participants were exposed to psychoeducation, tracking anxiety cues, motivational interviewing, readiness to change, self-assessment of barriers to change and personal strengths, grounding exercises, applied muscle relaxation, diaphragmatic breathing, mindful walking, and psychoeducation about worry. On average, participants completed 10.4 techniques (SD 17.4, median=4.0, range=0-137) and used Lantern for 16.6 days (SD 21.3, median=9.0, range=0-116) over 6 months. Of the 310 patients who downloaded the app, 27% (n=84) never started the program (ie, completed 0 techniques). Almost 60% (n=183) of participants who downloaded Lantern were considered to be “engaged” in the program (defined as completing ≥3 techniques). Engaged Lantern participants completed an average of 10.3 (SD 10.43) units, which covers automatic thoughts and cognitive reframing. Engaged Lantern participants used Lantern on an average of 26.0 (SD 23.17) days over 6 months.
Participants who downloaded Lantern were administered a questionnaire assessing the helpfulness of and satisfaction with Lantern. The satisfaction and helpfulness survey was completed by 122 participants (
The ITT sample included active group participants who completed the 6-month GAD-7 assessment. In the ITT sample (N=224), anxiety symptoms (beta=–2.61; standard error [SE]=0.34;
Using the same mixed-effects model, an mITT analysis was performed using engaged Lantern participants. The engaged Lantern subgroup demonstrated significantly reduced anxiety symptoms with a medium effect size (beta=–3.74; SE=0.49;
We examined clinically meaningful improvement and deterioration among active group participants who met the clinical threshold for anxiety (GAD-7 score ≥8; n=150). Such an improvement or deterioration was defined as at least a 4-point reduction or increase, respectively, on the GAD-7 score. Among the ITT sample, clinically meaningful improvement was observed in 54% (81/150) of participants. Clinically meaningful deterioration occurred in 8.7% (13/150) of participants. Intervention response (attaining at least 50% reduction in anxiety) was observed in 30% (n=45) of participants. Among the engaged Lantern patients (mITT), 60.8% (45/74) experienced clinically meaningful improvement and 5.4% (4/74) experienced deterioration.
Typically, the usage of digital behavioral health interventions is not linearly related to clinical outcomes [
Among the matched active group ITT sample, reduction in anxiety symptoms did not significantly differ from that in the matched comparison group (CAU;
Using the same mixed-effects model, an mITT analysis was performed using engaged Lantern participants (n=69;
Among the patients with baseline GAD-7 scores ≥8, clinically meaningful improvement was observed in 53.5% of the active group (54/101) and 45.8% of the comparison group (44/96). Clinically meaningful deterioration occurred in 11.9% of the active group and 16.7% of the comparison group. Among engaged Lantern patients (mITT), 56.3% (27/48) experienced clinically meaningful improvement. Substantially lower rates of deterioration occurred among engaged Lantern participants (6.3%, 3/48) as compared to 16.7% (16/96) of comparison participants. Intervention response was observed in 27.7% of the active group and 26% of the matched comparison group.
Anxiety, Quality of Life, and Medical Utilization. Significant findings after adjusting for diabetes and insurance type. Effect sizes of 0.2, 0.5, and 0.8 indicate small, moderate, and large clinically meaningful effects, respectively.
Measures | Mean change (SD) | Effect size (Cohen |
Mixed-effects model | ||||||||||
Within group (95% CI) | Between group (95% CI) | Beta | SEa | ||||||||||
Active | –1.97 (6.06) | –0.32 (–0.55 to –0.10) | 0.001 (–0.22 to 0.22) | –0.01 | 0.66 | 0.99 | |||||||
Control | –1.96 (5.75) | –0.34 (–0.56 to –0.12) | |||||||||||
Active | 5.57 (11.08) | 0.50 (0.26 to 0.74) | 0.198 (–0.03 to 0.43) | 2.27 | 1.27 | 0.07 | |||||||
Control | 3.4 (10.77) | 0.32 (0.09 to 0.54) | |||||||||||
Active | 2.81 (9.07) | 0.31 (0.07 to 0.55) | 0.205 (–0.03 to 0.44) | 1.92 | 1.06 | 0.07 | |||||||
Control | 0.95 (9.12) | 0.10 (–0.12 to 0.33) | |||||||||||
High utilizer at 6 months (yes/no) | —b | — | Odds ratio=0.09 | –2.40 | 0.49 | <.001 |
aSE: standard error.
bNot applicable.
Anxiety, Quality of Life, and Medical Utilization among engaged Lantern participants. Significant findings after adjusting for diabetes and insurance type.
Measures | Mean change (SD) | Effect size (Cohen |
Mixed-effects model | ||||
Within group (95% CI) | Between group (95% CI) | Beta | SEa | ||||
Active | –3.16 (5.38) | –0.59 (–0.93 to –0.24) | 0.22 (–0.07-0.50) | –1.20 | 0.81 | .14 | |
Control | –1.96 (5.75) | –0.34 (–0.56 to –0.12) | |||||
Active | 6.47 (11.22) | 0.58 (0.18 to 0.98) | 0.28 (–0.04-0.60) | 3.23 | 1.71 | .06 | |
Control | 3.4 (10.77) | 0.32 (0.09 to 0.54) | |||||
Active | 4.72 (9.4) | 0.50 (0.10 to 0.90) | 0.40 (0.08-0.72) | 3.79 | 1.46 | .01 | |
Control | 0.95 (9.12) | 0.10 (–0.12 to 0.33) | |||||
High utilizer at 6 months (yes/no) | —b | — | Odds ratio=0.04 | –3.20 | 0.69 | <.0001 |
aSE: standard error.
bNot applicable.
This pragmatic open trial demonstrated partial feasibility and adequate acceptability among patients who were prescribed Lantern by their PCPs. Although conversion rates from eligibility to consent improved after resolving clinic electronic screening issues, it is plausible that more patients would have enrolled if Lantern was offered systematically to all patients meeting the eligibility criteria. PCPs used their discretion to prescribe Lantern, which impeded our ability to differentiate provider adoption and willingness to prescribe Lantern from patient willingness to participate.
About 73% of patients who downloaded the app started Lantern and nearly 60% engaged in the Lantern program. The average of ≥10 usage days among those who downloaded the Lantern app is comparable to the 10 distinct mobile app use benchmarks followed in digital CBT intervention literature; very few coach-guided CB mobile apps reach this level [
Although there was a significant reduction in anxiety and QoL improvement in the active group receiving Lantern, this reduction did not significantly outperform the matched comparison group in the ITT analysis. One reason for this lack of difference is that patients in the comparison group received CAU, wherein behavioral health resources were provided to PCPs in addition to psychotropic medications prescribed by PCPs. The reduction in anxiety symptoms, the magnitude of effect sizes (│
Interestingly, Lantern had a greater impact on HRQoL than anxiety. Although anxiety has a significant impact on HRQoL, they are sufficiently distinct [
HRQoL improved with patients who used the Lantern anxiety program, which could have occurred via reduction of PTSD symptoms, comorbid depressive symptoms, or other symptoms of anxiety conditions rather than GAD symptoms measured by the GAD-7. In this sample, 44% of patients had a diagnosis of depression in their medical record, and among the subgroup with scores above the anxiety clinical threshold, 67% reported current depression symptoms, 62% reported current sleep difficulties, and 58% reported current anger. Although a specific measure of depressive symptom severity was not used, an improvement in depressive symptoms may have occurred, as prior Lantern evaluations have found significant reductions in depressive symptoms [
The most robust finding is that with Lantern, the likelihood of high outpatient medical utilization was significantly lower than that in a matched comparison group. Both groups showed an average of about three outpatient medical visits within 6 months prior to the study. This is consistent with the average outpatient utilization among patients with anxiety disorders who are untreated or undertreated [
This was an open pre-post pragmatic trial. Careful propensity matching allowed for the most rigorous comparison in the absence of randomization. However, it was important to show what results were possible in a pragmatic trial with little research infrastructure to be able to generalize our findings for real-world settings. Information provided to PCPs was minimal. PCPs may benefit from better guidelines for selecting appropriate candidates for Lantern. Due to the absence of a research infrastructure to guide study recruitment and retention, there was a high percentage of missing outcome data in the active group, which limits interpretation of the findings but still offers value as a feasibility trial. We are not aware of how frequently PCPs referred to psychotherapy or prescribed psychotropic medications at the comparison sites. When comparing the minimally engaged subsample (mITT group) to the matched controls, it is important to remember that the comparison group was matched to the full ITT sample and not the mITT sample. Thus, it may be that more likely that patients who showed improvements continued to use the app as a source of survival bias. However, the purpose of the mITT evaluation was to facilitate interpretation of the effect size in relation to the ITT group. There is also a possibility of a “digital placebo” effect, whereby the benefits may be due to placebo engagement in the mobile app intervention rather than the active components of the CB program itself [
Lantern demonstrated moderate effectiveness for addressing anxiety symptoms over 6 months, but not more than primary CAU. Lantern led to a robust decrease in the number of high outpatient medical utilizers among primary care patients with other morbidities such as depression, obesity, hypertension, chronic pain, chronic gastrointestinal conditions, and use of psychotropic medications. The surprising effectiveness seen in primary CAU may have suppressed group differences typically observed with standard CAU. However, with the minimal research infrastructure of this pragmatic study and limited resources to optimize PCP uptake and effectiveness of prescribing Lantern, these outcomes are meaningful from a population health and medical cost-offset perspective. This may be the first study to show that a coached CB program delivered via a mobile app provided in primary care reduces outpatient medical utilization compared to a matched control group. Whether this decreased utilization was diverted to other types of medical utilization (eg, emergency department visits) remains unknown. Future analyses of 12-month follow-up data will provide information about the extent to which Lantern has an impact on improving appropriate healthcare utilization.
Lantern satisfaction and helpfulness questionnaire (n=77) results.
care as usual
cognitive behavioral therapy
electronic medical record
Generalized Anxiety Disorder Questionnaire
health-related quality of life
International Statistical Classification of Diseases and Related Health Problems, 10th revision
intent to treat
mental composite score
Mental health composite
modified intent-to-treat
primary care physician
physical composite score
physical health composite
posttraumatic stress disorder
Research and Development
standard error
Short Form Health Survey
quality of life
This research was funded by UPMC Enterprises. This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. The authors would like to thank all the participants who took part in this research study. We would like to acknowledge the valuable contributions of Meredith Strassburger, Katheryn Mcauliff, Katie Scott, Michelle Sloane, Dina Perry, and the primary care physicians and practices that participated in this study.
EMS received grants from NIH and consultant fees from AbbVie and APPI. MO was a shareholder in Thrive Network, Inc (DBA, Lantern) at the time this study was conducted. UPMC Enterprises has a financial interest in Thrive Network, Inc, which develops and commercializes Lantern’s products and services. EMS is a scientific advisor and consultant to UPMC Enterprises. MLW and FS have no competing financial interests. MLW is an independent statistician.