Authors: Lance M. Rappaport, Emily Jerome, Michael Van Ameringen, Maureen Whittal, Carmen P. McLean
Categories: Original Research, Anxiety, MHealth, Smartphone app, Open-label trial, Cognitive behavioral therapy
Source: Journal of Mood and Anxiety Disorders
Authors: Lance M. Rappaport, Emily Jerome, Michael Van Ameringen, Maureen Whittal, Carmen P. McLean
Evidence-based treatments can effectively address anxiety and related conditions. However, new resources are needed to make psychological interventions accessible to the substantial and increasing North American population affected by anxiety and related psychopathology.
The present study evaluated whether use of the MindShift app (Anxiety Canada) may help reduce anxiety symptoms and related depressive symptoms, quality-of-life, and functional impairment among adults 18 years of age and older.
Adults ages 18–80 (N = 380) participated in an online open-label trial to evaluate change in anxiety and related distress while using the MindShift smartphone app. Inclusion residence in Canada or the USA and self-identification of anxiety or anxious distress to address during the study. Participants reported the severity of four primary outcomes at baseline and 2-, 4-, 8-, 12-, and 16-weeks after they began using the MindShift app.
All four primary outcomes improved over the 16-week participants reported reduced anxiety and depressive symptoms; reduced functional impairment; and improved quality-of-life. Improvements were unrelated to the frequency with which participants used the MindShift app. Effect sizes indicated moderate change in anxiety symptoms (d=0.61, p < 0.0001), depressive symptoms (d=0.50, p < 0.0001), functional impairment (d=0.55, p < 0.0001), and quality-of-life (d=0.31, p < 0.0001) at the end of the 16-week intervention; improvements were consistent with response to treatment but not remission.
Overall, the MindShift app may provide a ready to scale low-cost resource to assist in meeting the mental health needs of adults across North America, particularly those who report mild or moderate symptom severity.
Extensive research documents the efficacy and effectiveness of cognitive behavioural therapy for a myriad of psychological and psychiatric conditions [1], [2] including anxiety [3], [4] and related depressive disorders [5]. However, the substantial need for psychotherapy is also well-documented globally [6], [7], [8] as well as within Canada [9], [10], [11] and the United States of America [12], [13], [14]. Specifically, the number of individuals who warrant or may benefit from psychotherapy outpace the number of potential psychotherapy providers, particularly in geographically remote regions [15], [16]. Additional limitations further hinder the accessibility of evidence-based, efficacious cognitive behavioural therapy, including the cost of and limited private insurance or government coverage for psychotherapy; the need to take time off work to attend appointments regularly during providers’ traditional workday; and social stigma associated with seeking help for a mental health condition [16], [17].
Diverse novel strategies address the limitations associated with traditional modes of psychotherapy delivery [18], including time limited psychotherapy formulations [19], [20] and ‘train the trainer’ approaches [21] to increase the relative efficiency and number of treatment providers, respectively. However, issues of treatment cost [22], stigma [23], and access during traditional work hours or in remote geographical regions [24] may still present barriers to treatment for many who would benefit from it. To address these obstacles, internet- and smartphone-based psychotherapy leverages technological developments to deliver psychotherapeutic interventions privately and at a time convenient for each recipient [25], [26]. Internet- (e.g., internet-based cognitive behavioural therapy [iCBT]) and smartphone-based (e.g., mHealth) psychotherapy formulations vary in the degree to which a user is guided through the intervention. Interventions guided by a therapist show improved efficacy and effectiveness [27], particularly for individuals who report higher baseline symptomatology [28]. However, self-guided interventions may help individuals concerned about stigma in seeking help or for whom cost limits access to other treatment options as the relatively low per user cost for self-guided interventions allow them to be readily scaled to benefit many users. However, while many rely on evidence-based strategies and principles, relatively few smartphone-based apps empirically evaluated improvement in users’ mental health or well-being while using the app [29].
Accumulating evidence supports the effectiveness of smartphone-based mobile mental health (mHealth) apps to manage a wide array of conditions, such as anxiety, depressive, posttraumatic stress, and substance use disorders [30]. However, the number of empirically supported resources pales in comparison to the proliferation of apps with limited scientific basis or empirical support [31], [32]. In recent reviews, Temkin and colleagues [33] and Firth and colleagues [34] synthesize the accumulating empirical support for mobile apps to assess and manage anxiety symptoms in youth and adults, respectively. For example, in one prior study of the MindShift app, 18 undergraduate students demonstrated reductions in generalized anxiety, depressive, and panic symptoms over three weeks though the specific trajectory of symptom change over time is unclear [35]. Similarly, Sharma et al. [36] suggest the feasibility and acceptability of the MindShift app for use with an adult inpatient population though a limited sample size precluded the estimation of effectiveness as compared to treatment-as-usual without use of the app. However, prior research relied on relatively small sample sizes and often integrated smartphone apps as adjunctive tools in the context of other, traditional interventions with treatment seeking samples [33]. Given their low cost to the user and ready scalability, mobile mental health apps may provide useful first-line resources to address subclinical psychosocial distress (e.g., anxiety symptoms) or to address nascent anxiety disorders while individuals wait for traditional treatments. However, to evaluate these potential uses, research is needed to evaluate the feasibility, acceptability, and potential helpfulness of mHealth apps in community samples outside of ongoing traditional treatment.
The present study evaluated potential change in anxiety symptom severity and related depressive symptom severity, quality-of-life, and functional impairment among adults 18 years of age and older while using the MindShift app (Anxiety Canada) for 16 weeks. Briefly, the MindShift app is a smartphone-based app developed to disseminate principles of cognitive-behavioral therapy in a scalable, accessible format. It provides 10 components (e.g., a thought record, daily mood check-in), with which users engage in an unguided fashion at their desired timing and frequency. Given the existence of other internet- and smartphone-based mental health resources [33], [35], it was deemed inconsistent with the ethical principle of equipoise to assign some participants to a control group assigned to abstain from using any resources to address anxious distress reported at baseline. Instead, the present study was designed as an open-label trial of the MindShift app to evaluate primary study outcomes (i.e., participants’ anxiety symptoms, depressive symptoms, quality-of-life, and functional impairment) over 16 weeks of app use relative to baseline. Specifically, participants in Canada and the United States of America completed surveys at baseline then after 2-, 4-, 8-, 12-, and 16- weeks of using the MindShift smartphone app. Based on the extensive evidence base for internet-based cognitive behavioural therapy, we hypothesized that participants would demonstrate improvement in all four primary study outcomes over time. Consistent with prior research on cognitive behavioural therapy and mHealth resources, we hypothesized that most improvement would be evident early in participants’ use of the app (e.g., within the first 4–8 weeks). To evaluate if helpfulness of the MindShift app was associated with enrolment in ongoing traditional pharmaco- or psychotherapy, planned moderation analyses examined if participants who received either pharmacotherapy or psychotherapy concurrently showed greater change in primary outcomes relative to participants who only used the MindShift app. This study was registered on ClinicalTrials.gov (NCT04562948).
Adult participants ages 18–80 years (M = 37.76, SD = 13.43) were recruited across Canada and the United States of America through advertisements on social media and the Anxiety Canada website. Inclusion criteria required that each participant a) experience anxiety or anxious distress; b) be sufficiently comfortable with the English language to complete study measures; c) have regular access to a smart phone or mobile device equipped with Android OS or iOS; and d) have semi-regular internet access to complete online study assessments. There were no exclusion criteria.
Following online informed consent, participants completed a baseline assessment of demographic information, anxiety symptoms, depressive symptoms, quality-of-life, and functional impairment. Participants then watched a video that provided instructions in the installation and use of the freely available MindShift smartphone application (Anxiety Canada) on iOS or Android OS. Participants were instructed to enter a study-specific ID code into the MindShift app to verify that it was downloaded and set up. Participants were encouraged to use the app regularly for the following 16 weeks. However, specific metrics for use were not provided to assess participants’ naturalistic use patterns and their associated outcomes. Two-, 4-, 8-, 12-, and 16- weeks following baseline assessment, participants completed follow-up assessments of anxiety symptoms, depressive symptoms, quality-of-life, and functional impairment via an emailed link to an online survey. Each follow-up assessment also asked about the frequency with which participants used the app and each feature of the app as well as their intentions to continue to use the app following the 16-week study period. The study was cleared by the Research Ethics Board of the University of Windsor (#20–127).
To assess anxiety symptom severity, participants completed the Generalized Anxiety Disorder (GAD)−7 [37], on which participants rate the frequency of 7 anxiety symptoms over the past 2-week period from 0 (Not at all) to 3 (Nearly every day). Prior research extensively documents strong psychometric properties of the GAD-7 to assess anxiety severity in community [38], [39] and clinical samples [40], [41] including to index change in anxiety severity over the course of an intervention [42], [43]. In this study, interitem reliability (α = 0.91 – 0.94) and unidimensionality (ω = 0.89 – 0.92) were high at all timepoints.
To assess depressive symptom severity, participants completed the Patient Health Questionnaire (PHQ)−9 [44], [45], on which participants rate the frequency of 9 depressive symptoms over the past 2-week period from 0 (Not at all) to 3 (Nearly every day). Prior research extensively documents strong psychometric properties of the PHQ-9 to assess depressive symptom severity in community [46], [47] and clinical samples [48] including to index change in depressive severity over the course of an intervention [49], [50], [51]. In this study, interitem reliability (α = 0.90 – 0.93) and unidimensionality (ω = 0.87 – 0.91) were high at all timepoints.
To assess quality-of-life, participants completed the short form of the Quality of Life Enjoyment and Satisfaction Questionnaire [Q-LES-Q-SF; 52], on which participants rate their satisfaction with 16 life domains (e.g., physical health, family relationships) form 1 (very poor) to 5 (very good). Prior research documents strong psychometric properties of the Q-LES-Q-SF in community [53] and clinical samples [54], [55]. In this study, interitem reliability (α = 0.91 – 0.96) and unidimensionality (ω = 0.90 – 0.95) were high at all timepoints.
To assess functional impairment in daily life, participants completed the Work and Social Adjustment Scale [WSAS; 56], on which participants rate their degree of impairment in 5 life domains (e.g., work, social activities) from 0 (not at all impaired) to 8 (very severely impaired). Prior research documents strong psychometric properties of the WSAS in community and clinical samples [56], [57], [58] including to index change in functional impairment [59], [60]. In this study, interitem reliability (α = 0.88 – 0.94) and unidimensionality (ω = 0.87 – 0.93) were high at all timepoints.
To assess participants’ use of the smartphone app, follow-up surveys asked participants to rate their frequency of app use (“How often do you use the MindShift app”) from “I used it only once” to “Daily” as well as their intention to continue to use the app following the 16-week study (“Which statement best represents you”) from “I am not going to use the MindShift app after my participation in this research study.” to “I am highly likely to use the MindShift app after my participation in this research study.” Participants also rated the frequency with which they used each component of the app (e.g., self-check-in, coping cards) from “I haven’t used this component” to “3 or more times per day” and their satisfaction with each component used from 0 (bad) to 4 (great). Finally, participants were asked whether they use the app regularly with support from a mental health provider, or when experiencing acute anxiety; whether they would be willing to share information provided in the app with a healthcare provider (e.g., physician or psychologist) or counsellor; and provided space to provide qualitative feedback on each app component.
Developed by Anxiety Canada, the MindShift app is a freely available smartphone app that provides cognitive behavioral therapy principles through, currently, 10 components to help users manage anxiety and related distress. For example, components allow users to log current distress on a 5-point scale from “awful” to “great”; record and reflect on thoughts; complete guided relaxation exercises; and complete guided exposure to anxiogenic stimuli. Immediately after app installation, users complete a brief introduction designed to familiarize users to the app layout and tools. All 10 app components are then simultaneously available to users who freely use any components when desired. The app provides no structured, scheduled feedback to users about their progress in use of the app.
Substantial prior research indicates both linear and non-linear change during psychosocial mental health interventions [61]. Therefore, using multilevel, mixed effects regression, each primary outcome (e.g., anxiety symptoms) was regressed on time and successively higher order polynomial transformations of time to identify linear and non-linear change in symptoms, quality-of-life, and functional impairment over time. Successively higher order polynomials were added until an additional term indicated limited improvement to the model, at which point successively higher order polynomials were added as random effects to model between-person heterogeneity in change over time. Models were fit with full information maximum likelihood (FIML) estimation for robustness to data missing conditional on included covariates [62]. Analytic models included correlations among all random effects to adjust for any potential associations of baseline symptom severity with change over time [63]. Given the examination of four correlated outcomes, we note the need to consider an adjusted test-wise alpha to limit the possible inflation of family-wise alpha or Type I error rate. Specifically, empirical p-values of change in each outcome were considered against a test-wise alpha adjusted using the traditional Bonferroni procedure for the number of tests conducted on correlated outcomes (e.g., anxiety and depressive symptoms). However, we note that adjustments to test-wise alpha largely assume that tests are independent of one another. Therefore, consistent with statistical recommendations [64], unadjusted p-values are provided below to present the outcome of each statistical test clearly and transparently as Bonferroni adjustment may provide an overly conservative test of study results.
Analyses were conducted in R version 4.1.2 [65] using the nlme package [66] to estimate multilevel models and ggplot2 package [67] for figure generation.
Of 380 participants who completed baseline assessment, 316 (83.16%) identified as women, 63 (16.58%) identified as men, and 1 (0.26%) did not disclose their gender identity. Most of the sample identified as White (n = 301; 79.21%); 19 (5.00%) participants identified as South Asian, 11 (2.89%) as Chinese, 10 (2.63%) as Latinx, 10 (2.63%) as Arab or West Asian, 6 (1.58%) as Black. Another 6.06% of the sample identified as another racial category, reports of each category were so low that providing specific information may risk publishing data about an individual participant. Two hundred and ninety participants (76.32%) resided in Canada; 90 (23.68%) resided in the USA. Two hundred and two (53.16%) participants reported a history of psychotherapy; 64 (16.84%) reported participating in psychotherapy in the past 2 months. Similarly, 233 participants (61.32%) reported a history of pharmacotherapy; 170 (44.74%) reported receiving pharmacotherapy in the past 2 months.
Of 380 participants who initially enrolled in the study, 231 (60.79%) completed the 2-week follow-up assessment, 235 (61.84%) the 4-week follow-up assessment, 234 (61.58%) the 8-week assessment, 224 (58.95%) the 12-week assessment, and 223 (58.68%) completed the final, 16-week follow-up assessment. Participants who completed the final, 16-week follow-up assessment were younger (M=36.46) than those who dropped out of the study (M=39.6), t(301.76) = −2.19, p = 0.029; reported lower baseline depressive symptoms (M=11.35) than those who dropped out of the study (M=13.21), t(344.90) = −3.06, p = 0.002; and reported a higher quality of life (M=31.80) than those lost to follow-up (M=29.24), t(320.01) = 2.50, p = 0.013. Missing data was also associated with participant race, χ^2^(12) = 21.27, p = 0.047, and more likely among participants who reported past use of psychotropic medication, χ^2^(1) = 5.54, p = 0.019, or recent use of psychotropic medication (i.e., in the past 2 months), χ^2^(1) = 5.39, p = 0.02. Participant gender, baseline anxiety symptomatology, baseline functional impairment, employment, and past or current psychotherapy use were unrelated to study attrition (p > 0.09). As compared against the largest group (i.e., participants who self-identified as White) in a logistic regression of missing data at the final, 16-week follow-up assessment, identification as part of another racial group was not associated with missing data (ps > 0.05). However, the available sample sizes were small for many groups. Therefore, participant age, race, and past and current use of psychotropic medication were retained as covariates in subsequent analyses to adjust for any potential correlates of missingness. As described above (see Data Analysis), modeled random effects adjust for baseline depressive symptoms and quality-of-life in analysis of depressive symptoms and quality-of-life, respectively.
Mean anxiety symptoms (M=11.40, SD=5.12), depressive symptoms (M=12.12, SD=5.95), quality-of-life (M=30.76, SD=9.77), and functional impairment (M=21.31, SD=9.76) at baseline were largely consistent with outpatient samples of individuals seeking treatment for anxiety and depressive disorders [68], [69], [70], [71]. Self-reported app usage was high but declined moderately over the 16-week study period. After two-weeks of app use, 14.1% of the sample reported using the app daily; 73.6% reported using the app at least weekly or more often. At the 16-week follow-up assessment, 7.3% of the sample reported daily app use; 45.9% reported use weekly or more often. Similarly, participants’ self-reported likelihood of using the MindShift app following the study period was high but decreased from 62.9% who endorsed likely or highly likely “to use the MindShift app after my participation in this research study” at two-week follow-up to 52.0% at 16-week follow-up.
Multilevel, mixed effects regression models indicated that anxiety symptoms, depressive symptoms, and functional impairment decreased over 16-weeks in a downward, decelerating fashion, in which most improvement occurred in the first 4 weeks (see Table 1 and Fig. 1). At baseline, 63.47% and 62.96% of participants met criteria for moderate anxiety or depressive symptoms, respectively. At 16-week follow-up, 33.94% and 37.10% of participants met criteria for moderate anxiety or depressive symptoms, respectively. Similarly, mean anxiety (M=8.06, SD=5.25) and depressive symptoms (M=8.65, SD=6.14) at 16-week follow-up indicated residual symptoms and moderate but clinically meaningful change. Quality-of-life also improved in an upward, decelerating fashion, in which most improvement occurred in the first 4 weeks. Dependent t-tests of change from baseline to 16-week follow-up indicate moderate effect sizes for change in anxiety (d=0.61, p < 0.0001) and depressive symptoms (d=0.50, p < 0.0001) as well as quality-of-life (d=0.31, p < 0.0001) and functional impairment (d=0.55, p < 0.0001).Table 1Change in Primary Outcomes over 16-Weeks.Table 1Anxiety SymptomsDepressive SymptomsQuality-of-LifeFunctional ImpairmentNumber of People375378222375Number of Assessments1518152112511517Fixed EffectsIntercept11.47+(9.99, 12.95)11.37+(9.65, 13.09)34.39+(30.34, 38.45)18.91+(15.94, 21.88)Age-0.04*(−0.07, −0.007)-0.05*(−0.08, −0.006)-0.01(−0.11, 0.09)-0.03(−0.10, 0.04)Prior Pharmacotherapy2.68+(1.50, 3.85)3.27+(1.90, 4.64)-3.91*(−7.11, −0.70)4.68***(2.30, 7.06)Current Pharmacotherapy-0.55(−1.71, 0.61)0.65(−0.70, 2.00)0.69(−2.55, 3.93)1.34(−1.01, 3.68)Linear Change over Time-0.79+(−1.01, −0.56)-1.41+(−1.81, −1.01)0.51+(0.29, 0.74)-1.49+(−1.90, −1.07)Quadratic Change over Time0.07***(0.03, 0.10)0.24***(0.12, 0.37)-0.02**(−0.03, −0.008)0.15+(0.09, 0.21)Cubic Change over Time-0.002*(−0.003, −0.0005)-0.02**(−0.03, −0.006)---0.005***(−0.007, −0.002)Quartic Change Over Time--0.0005*(0.0001, 0.0009)----Random EffectsIntercept3.99(3.62, 4.40)4.93(4.49, 5.40)8.01(6.93, 9.24)8.01(7.28, 8.82)Linear Change over Time0.22(0.18, 0.26)0.68(0.54, 0.84)0.37(0.27, 0.51)1.23(0.96, 1.57)Quadratic Change Over Time--0.04(0.03, 0.05)--0.06(0.05, 0.09)Covariance of Intercept withLinear Change over Time-0.35(−0.51, −0.18)-0.35(−0.52, −0.17)0.06(−0.47, 0.56)-0.20(−0.37, −0.02)Covariance of Intercept withQuadratic Change over Time--0.25(0.05, 0.44)--0.15(−0.06, 0.34)Covariance of Linear withQuadratic Change over Time---0.93(−0.96, −0.89)---0.97(−0.98, −0.93)Note. *p< 0.05, **p< 0.01, ***p< 0.001, + p< 0.0001. Race was included as a covariate in all four models. However, coefficients are excluded above to avoid providing baseline mental health data on marginalized populations who are, in this sample, represented by relatively few participants.Fig. 1Change in Primary Outcomes over 16-Weeks. Note. Each line reflects a loess smoothing algorithm applied to data on anxiety symptomatology, depressive symptomatology, quality-of-life, and functional impairment assessments conducted at baseline assessments and 2-, 4-, 8-, 12-, and 16-week follow-up assessments while participants used the freely available MindShift smartphone app.Fig. 1
Similar patterns of change in all four primary outcomes were evident when the sample was restricted to the 156 participants for whom entrance of their study identification code into the MindShift smartphone app verified that they downloaded the smartphone app. Secondary moderation analyses indicated that improvement in all four primary outcomes did not differ as a function of whether participants engaged in psychotherapy (ps>0.28), pharmacotherapy for any mental health condition (ps>0.37), or pharmacotherapy for anxiety (ps>0.26) in the past 2 months nor did primary outcomes differ as a function of the frequency with which participants reported using the app over the 16-week study period (ps>0.17). Similarly, improvement in all four primary outcomes did not vary as a function of participant gender (ps>0.50) or race/ethnicity (ps>0.03) after adjustment of alpha for potential inflation of type I error given the large number of race/ethnicity categories evaluated.
Over 16 weeks of using the MindShift smartphone app, participants demonstrated improvements in reported anxiety symptoms, depressive symptoms, quality-of-life, and functional impairment. Results are consistent with extensive research on in-person cognitive behavioral therapy [2], including cognitive behavioral therapy for anxiety disorders [3], as well as internet- and smartphone-based formulations of cognitive behavioral therapy of anxiety disorders for adult [27], [34] and youth samples [33]. Crucially, results expand initial data on the MindShift app from relatively small samples [35], [36] to a larger, North American community sample and to broad functional outcomes beyond improved symptomatology including quality-of-life and functional impairment.
The results revealed both residual symptomatology at 16-week follow-up and change from moderate to mild anxiety and depressive symptom severity over the 16-week study period. Both findings are consistent with moderate contributions of the MindShift app (i.e., moderate effect sizes for change in primary study outcomes), which may provide one resource in users’ set of tools to manage anxiety and related distress. Moderate within-person change while using the MindShift app, particularly early in one’s use of the app, is consistent with other within-person research on smartphone-based mental health resources [72], [73]. The MindShift app leverages traditional benefits of smartphone apps to offer a scalable resource freely available to users. Following a stepped care model, app use is likely insufficient for many people [16]. However, it may provide some additive benefit at relatively low cost to both the user and health-care system. Moreover, the unique advantages of smartphone apps may be particularly well-suited to partially address long-standing health disparities, such as those that arise from the cost to access traditional treatments [22], stigma [23], and access during traditional work hours or in remote geographical regions [24]. To that end, research is critically needed to empirically evaluate the proliferation of smartphone apps that purport to support users’ mental health [29]. While comparison against a sample who abstained from accessing mental health resources was deemed incompatible with the ethical principle of equipoise, the present data indicate that users’ anxiety and depressive symptom severity, quality-of-life, and functional impairment improved over 16-weeks of using the MindShift app relative to baseline assessment.
Findings from the present study should be considered in light of several limitations. First, the lack of a comparison group who abstained from using the MindShift app limits the causal inferences one can draw about the contribution of the MindShift app to participants’ improved well-being over time. For example, participants who reported increased app use did not demonstrate greater improvements in well-being. However, participants may titrate their use of the app to be sufficient for their unique needs such that additional use may not be beneficial. Similarly, results may indicate regression-to-the-mean. However, it is unclear that regression-to-the-mean would fully explain improved well-being during the specific 16-week period that participants used the MindShift app. Future research is needed to evaluate the optimal frequency and duration with which to use the MindShift app to guide users in its optimal use, particularly during the critical first four weeks of use when most gains occurred. Future research is also needed to compare use of the MindShift app to other mental health resources. For example, it might be compared against those currently in use for individuals waiting for traditional treatment to establish whether the MindShift app can be recommended as a superior stop-gap solution for adults waiting for traditional, more intensive treatment options [74].
Second, the present study included no follow-up assessments after participants ceased using the MindShift app. Future research is needed to understand the durability of the change demonstrated here. Extensive research documents that improvements from traditional cognitive behavioral therapy are durable for several years following treatment [75], [76]. However, given the low cost and little effort required to continue using the MindShift app, users may be encouraged to continue using it or to use it again as needed. Similarly, given the current evidence that most change occurred over the first four weeks of app use, future research is needed to clarify how participants can stagger use of different apps or app features to maximize their benefit from each. Such research may guide the development of smartphone apps to prompt users to try specific novel features every four weeks [77], [78].
Third, White women were overrepresented in the present sample. While there was no evidence that gender or race/ethnicity moderated change in well-being over time, future research is needed to ensure that mHealth resources are equitably effective for a diverse range of users. Ensuring equitable benefits for users from diverse gender, racial, and socioeconomic backgrounds is particularly crucial if mHealth resources are used to address longstanding health disparities that disproportionately affect marginalized communities. Similarly, study inclusion criteria provided a low threshold to include any participants who self-identified the presence of anxiety or anxious distress that they hoped to address through use of the MindShift app. This threshold may help to reflect the large number of individuals who use smartphone-based apps to address subclinical, mild, or moderate psychosocial distress [79]. However, future research is needed to evaluate the potential for mHealth resources in other settings, such as inpatient psychiatry [36]. The present study represents an important first step to evaluate the utility of, and feasibility of research on, the MindShift app in large, multinational community samples. However, future research is needed to guide the development of mHealth resources to ensure equitable benefit for users from diverse backgrounds and clinical presentations.
Fourth, assessment of participants’ symptomatology, well-being, and app use relied on self-report so that participants could be recruited online from across Canada and the United States of America. Empirical results would benefit for corroboration by clinician observation. In particular, app usage data would benefit from comparison against objective data on app use. However, the study employed self-report measures (e.g., PHQ-9) extensively validated to assess and track symptomatology and well-being in community adult samples. Similarly, empirical results were unchanged when examined in the subset of participants for whom we can verify that the app was installed correctly.
Fifth, data for the present study were collected during the COVID-19 pandemic. Considerable extant research documented increased psychosocial distress during the pandemic among children, adolescents, and adults [80], [81], [82]. It is unclear how increased distress may contribute to study findings or how well the present findings may generalize to address adult anxiety and related distress outside the context of a chronic stressor, such as that posed by the COVID-19 pandemic [83], [84]. However, several empirical results support the potential generalizability of study findings beyond the COVID-19 pandemic. Specifically, we note that mean anxiety symptoms, depressive symptoms, and functional impairment at baseline were consistent with data from outpatient samples of individuals seeking treatment for anxiety and depressive disorders collected prior to the COVID-19 pandemic [68], [69], [70], [71]. Similarly, as discussed above, empirical evidence of change in anxiety and depressive symptoms is consistent with prior research on internet- and smartphone-based resources for adults [27], [34] and youth [33] including prior research on the MindShift app [35], [36]. Additionally, the present empirical results suggest the potential benefit of scalable e-mental health apps, such as MindShift, to address anxiety and related distress during periods of heightened stress, such as disasters, that may substantially increase the need for mental health services.
Finally, study attrition is consistent with other research on digital interventions [85] largely due to a considerable number of participants who chose to complete only the baseline assessment. However, retention was high from the two-week follow-up to the 16-week follow-up and correlates of attrition are modelled in primary analyses using FIML estimation, which is robust to data missing conditionally at random.
The present study recruited a community sample of adults residing in Canada and the U.S. to evaluate change in anxiety and depressive symptoms as well as quality-of-life and functional impairment over 16-weeks of using the MindShift smartphone app. On average, participants demonstrated improvement on all four study outcomes. The degree of improvement was unrelated to the reported frequency with which participants used the MindShift app. Specifically, improvement in anxiety and depressive symptoms was modest and consistent with response to intervention but not remission. Consistent with extensive prior research on in-person and online cognitive behavioral therapy, most improvement on all four study outcomes occurred in the first four weeks of app use. Overall, results suggest that the MindShift app may be one helpful, low-cost component in a set of tools used to manage anxiety and related distress.
The authors declare the following financial interests/personal relationships which may be considered as potential competing Lance M. Rappaport reports financial support was provided by Anxiety Canada. Michael Van Ameringen reports a relationship with AbbVie Inc that consulting or advisory and speaking and lecture fees. Michael Van Ameringen reports a relationship with Bausch Health Companies Inc Canada that consulting or advisory. Michael Van Ameringen reports a relationship with Biogen Inc that consulting or advisory. Michael Van Ameringen reports a relationship with Boehringer Ingelheim Canada Ltd that consulting or advisory. Michael Van Ameringen reports a relationship with Elvium (Purdue) that consulting or advisory, funding grants, and speaking and lecture fees. Michael Van Ameringen reports a relationship with Jazz Pharmaceuticals Inc that consulting or advisory. Michael Van Ameringen reports a relationship with Vistagen that consulting or advisory. Michael Van Ameringen reports a relationship with Lundbeck that speaking and lecture fees. Michael Van Ameringen reports a relationship with Otsuka that speaking and lecture fees. Michael Van Ameringen reports a relationship with Pfizer that speaking and lecture fees. Michael Van Ameringen reports a relationship with Sunovion Pharmaceuticals Canada Inc that speaking and lecture fees. Michael Van Ameringen reports a relationship with Takeda that speaking and lecture fees. Michael Van Ameringen reports a relationship with Biron that speaking and lecture fees. Lance M. Rappaport, Michael Van Ameringen, Maureen Whittal, and Carmen P. McLean volunteer on the Scientific Advisory Committee of Anxiety Canada.