User-Centered Design of a Digitally Enabled Care Pathway in a Large Health System: Qualitative Interview Study

Introduction

Background

Depression is a leading cause of disability worldwide, affecting nearly 300 million people [1,2]. Major depressive disorder (MDD) is a growing problem in the United States, with the total number of US adults with MDD increasing by 12.9%, from 15.5 million to 17.5 million between 2010 and 2018, and it is associated with a significant economic burden [3]. Primary care centers are the largest mental health service providers for people living with MDD, with up to two-thirds of visits to health care providers (HCPs) for depression occurring in a primary care setting [3-6]. Time constraints and the need to frequently manage multiple conditions during a single visit to a primary care setting can make it difficult for HCPs to fully engage with patients when it comes to their treatment for MDD [4]. Frequent communication and engagement between people living with MDD and their care teams may improve therapeutic outcomes, especially for chronic conditions such as MDD [4].

Measurement-Based Care

The American Psychiatric Association clinical guidelines recommend measurement-based care (MBC) for treating depression. MBC includes the routine use of standardized outcome measures to assess changes in depression symptoms, level of functioning, and quality of life across the treatment course [7]. MBC for MDD is effective because it allows the primary care provider to quantify clinical outcomes. This provides guidance for timely treatment modifications that may better meet the needs of the person being treated [8]. Furthermore, treatment decisions based on MBC give people living with MDD a better understanding of how their condition is changing over time and therefore potentially empowers them in terms of their own care [8]. In addition, compared with usual care, MBC in the management of depression has been shown to improve treatment adherence, thereby leading to improvements in clinical outcomes [8,9]. Despite the demonstrated benefits of MBC in improving treatment outcomes and patient engagement, adoption of MBC in routine clinical practice has been slow, with only 20% of HCPs using it in their practice [8]. Increased consultation times are often cited as a barrier to MBC implementation.

Shared Decision-Making

Along with MBC, another critical factor shown to improve treatment outcomes is the involvement of people living with MDD in the decision-making processes of their treatment journey [10]. Several studies have found that people experiencing a mental illness want to play a large role in the treatment decision-making process [10]. A shared decision-making (SDM) model of interaction can foster patient-provider engagement by empowering patients to play a great role in the decision-making process, thus creating an opportunity for them to have their voices, beliefs, values, goals, experiences, and preferences reflected in the treatment planning and monitoring process. This, in turn, can help increase treatment satisfaction and overall treatment adherence [10].

Among people receiving treatment for mental health disorders (including depression), increased adherence to treatments (including both psychopharmaceutical and psychotherapeutic) has been reported when they recognize that their treatments reflect their unique needs and preferences. A strong alignment between the treatment goals of the person living with depression and their treatment provider is another critical factor shown to be important in promoting adherence [10,11]. Several studies have reported misalignment between what patients and their treatment providers consider to be the most important treatment goal [10]. Thus, involvement in SDM is associated with a high probability of receiving quality care and improvement in symptoms [12] through increased adherence to drug treatment [10].

Digital Tool Development in MDD

Digital tool development that enhances patient engagement, MBC, and SDM has the potential to improve treatment outcomes in MDD [13]. Digital communications and information technologies have previously been shown to improve health care delivery by improving communication between providers, and by decreasing the need for face-to-face appointments, thus helping to alleviate the workload of HCPs [14,15]. With rapid advances and the adoption of smartphone technology, mobile health apps have generated interest from both the public and medical communities [16]. For the HCP, digital technology platforms, such as mobile apps, can offer low-cost interventions to monitor and improve services for patient populations that are difficult to retain during treatment [15]. A mobile app could help save time because people living with MDD could engage with symptom assessments (such as the 9-item Patient Health Questionnaire [PHQ-9]) and other clinical instruments outside their visit, whereas HCPs would only need to review these results instead of administering the instrument during the visit itself [8]. A mobile app could therefore promote MBC by allowing patients to remotely engage with validated quantitative measures of assessments, which can then be uploaded into their electronic health records (EHRs) for in-office visits and physician monitoring [8]. Collection of PHQ-9 results through a mobile app has been shown to be as sensitive as, or even more sensitive than, the traditional (in-person and paper-based) PHQ-9 data collection method [17]. This may be because people living with MDD might feel more comfortable revealing their symptoms in remote settings through a mobile app rather than in traditional in-person settings.

From the patient’s perspective, several studies using smartphone app–based interventions for depressive disorders have shown that depressive symptoms were reduced significantly more with smartphone apps [16,18,19] through motivating some users to consult medical professionals for diagnosis and management [20,21] and through self-observation [22,23]. Furthermore, apps dedicated to the caregivers of people living with MDD have also been shown to help caregivers in better supporting their loved ones and to destigmatize mental health care [24].

User-Centered Design Approach

Achieving this potential of digital tools to improve treatment outcomes requires a nuanced understanding of the desirability and usability of patient-facing and care team−facing digital interfaces alongside the practical requirements for patient and care team adoption. Perceived utility and overall value of the product to care teams and privacy and confidentiality concerns are often cited by care teams as examples of barriers that can limit the adoption of these types of apps [25]. Various mobile apps are available for depression management; however, many are patient facing only and do not include a care team interface [26]. The lack of guidance or feedback from the care team has often been cited by people living with depression as a barrier to the adoption of many apps that are dedicated to mental health care [25,27,28]. Patients might, for example, perceive the lack of feedback or engagement from the care team as indicating that the information they are providing through the app is not being monitored or integrated into their care processes, thus disincentivizing their engagement with these products [29].

A user-centered design approach can improve the understanding of the patient journey (Multimedia Appendix 1) [30]. Patient insights help develop an understanding of user needs, with iterative designs and prototypes playing a key role in how these insights and needs can be unearthed [30]. However, studies describing the user-centered design approach in health-related technology remain limited [30]. Takeda, Lundbeck, and Advocate Aurora Health (AAH) partnered to cocreate a digitally enabled care experience with users (patients living with MDD and care teams), software developers, and health-technology product development specialists (Ctrl Group and Fora Health) [31]. Specifically, a digital mobile patient interface, the Pathway App, was designed with a conversational interface and tested via a pilot feasibility study with 40 patients living with MDD [31]. The study showed a trend toward high patient activation and patient-provider engagement for people who used the app in addition to usual care compared with those who were assigned to usual care only [31]. Building on these results, a new iteration of the app was created and incorporated into a new digital platform, Pathway Platform, which includes the Pathway App; EHR-integrated, real-time, patient-level data sharing; and educational programming that is both care team and patient facing. This was guided by a more comprehensive understanding of care team workflows and patient and care team insights [32]. In this paper, we describe how user-centered design was applied to develop a digitally enabled MDD care platform that is optimized for the needs of patients living with MDD and their primary care teams.

Methods

Overview

This report describes the qualitative research undertaken to understand, iterate, and integrate Pathway Platform into primary care in the AAH system within and around Chicago, Illinois. In stage 1, interviews were conducted with 20 patients with MDD who participated in the pilot feasibility study and 15 HCPs (from September 19, 2018, to January 30, 2019). The stage-1 interview sought to understand the experiences with and perceptions about the initial functionality of the Pathway app (Multimedia Appendix 2) while also exploring perceptions about potential additional features such as HCP visit preparation, patient education, and goal setting and tracking. Interviews with patients lasted up to 60 minutes, and interviews with HCPs lasted up to 30 minutes. Another round of follow-up feedback was collected from 36 HCPs at 8 primary care sites through stage-2 interviews to understand care team workflows in the treatment of patients with MDD, the extent to which Pathway Platform can help optimize care, and what support the care team will need to make Pathway Platform work at their respective sites. HCPs were included if they were involved in primary care and specifically managed the care of patients with MDD. The semistructured, in-person interviews were conducted by a study team with expertise in qualitative research, user-centered design methods, workflow assessment, and educational support (Ctrl Group and PRIME Education LLC). Care teams were interviewed about topics related to workflow, perceptions about and experiences with MBC and SDM, and educational needs of patients with MDD and their care team members. The interviews were conducted until observational and analytical saturation was achieved. Saturation was defined as when no new inductive themes emerged during analysis and a priori or deductive themes were exemplified in the data [33].

Data Analysis

The qualitative data were first coded and analyzed independently by 2 Ctrl Group researchers using a set of foundational structural codes. A structural coding framework, based on the agreed-upon discussion guide from the qualitative interviews, was created for the “top-down” codes; “bottom-up” codes that organically arose from the data were also used. The top-down and bottom-up coding was then refined collaboratively as the researchers progressed through the data set by, for example, splitting preexisting codes into more specific subcodes and combining existing subcodes where appropriate. Once refined, the codes enabled thematic analysis and identification of recurring themes across participants’ accounts and perceptions. Guided by prior research findings [34-36], we adopted thematic analysis to generate an overarching understanding of shared experiences, thoughts, and behaviors across a broad, qualitative data set, to allow both inductive and deductive insights to emerge. This was particularly important because our study aimed to validate preexisting product features (through deductive reasoning) while also allowing respondents to provide suggestions that can inform the development of completely new features (inductive reasoning). This type of mixed reasoning would have been very difficult to achieve if the participants had only been surveyed about their thoughts around specific product features. The flexibility allowed by the thematic analyses was also important because it enabled engagement with both personal accounts of patients’ experiences and understandings and broad social constructs (eg, “workflow efficiency” or “medication adherence”) in slightly different social contexts (ie, different primary care environments) in the same research [36].

Thematic analysis of interviews was supported by Dedoose (SocioCultural Research Consultants, LLC), a qualitative data analysis software tool that enables systematized coding. Audio recordings from the interviews were uploaded to the web-based transcription service, Rev. Transcripts were uploaded to Dedoose [37], the top-down codes created during the coding framework phase were applied to appropriate excerpts for each transcript, and new bottom-up codes were created and applied as they arose over the course of the analysis. High-level analyses were conducted by reviewing the codes that had been applied to the interview transcripts. This was done to identify key and recurring themes from the insight-gathering stage. Interpretive insights were formulated through subsequent in-depth analyses. Findings and insights were presented based on code frequency, salience, and relevance to the research project.

The Standards for Reporting Qualitative Research recommendations were followed in the reporting of the study [38].

Ethics Approval

Ethics approval from the AAH institutional review board was obtained by making amendments to the existing Advocate Pathway study protocol (approval number: AHC-6680-75000249).

Results

Participant Characteristics

A total of 37 patients living with MDD completed the 18-week primary follow-up period in the pilot feasibility study—19% (n=7) were Black and 38% (n=14) were Hispanic [31]. Overall, 54% (20/37) of patients from the pilot feasibility study (9/18, 50% from the Pathway App arm and 11/19, 58% from the usual care arm) participated in the qualitative interviews in stage 1. In addition, 15 HCPs also participated in the stage-1 interviews. Overall, 53% (8/15) of HCPs in the group were previously involved in the pilot study, 25% (2/8) of whom were assigned to the PathwayApp arm and had used the app to derive patient reports. A total of 36 HCPs from primary care medicine with experience in managing people with depression participated in the stage-2 interviews (Multimedia Appendix 3). These HCPs were sampled for diversity of roles both within and across sites, and it was found that 33% (12/36) were physicians, followed by certified medical assistants (8/36, 22%), registered nurses (6/36, 17%), and licensed practical or advanced practice nurses (6/36, 17%).

Users’ Experiences With and Perceptions of the Pathway App

Overview

Thematic analysis identified major themes to describe the users’ experiences with and perceptions about the Pathway App, which included functionality, support, and patient-provider engagement, along with subthemes such as ease of use, utility, reliability, motivation, reducing burden, communication, understanding, and shared vision (Tables 1 and 2).

Functionality

Functionality was a major theme that emerged from the interviews, with 3 associated subthemes—ease of use, utility, and reliability. Patients living with MDD and HCPs emphasized ease of use and overall value as 2 of the most beneficial app features. Participants appreciated the simplicity of the design, which made it easy to understand and use the app effectively, and many reported that the app provided clinically meaningful information, such as tracking of symptoms, mood, and medication adherence, which helped them manage their symptoms more effectively. HCPs viewed mood, medication, and side effect tracking to be clinically valuable, owing in part to the continuous and direct input from the patient.

Support

Support was an important theme that emerged, which could be split into 2 subthemes—motivation and burden reduction. Patients living with MDD reported that the goal-setting and goal-tracking features provided emotional support and motivation through regular reminders and encouragement, which helped them stay engaged in their treatment.

Similarly, HCP visit preparation was noted as very useful by patients living with MDD. Many reported that the ability to prepare for HCP visits using the Pathway App reduced their anxiety and improved the ease and accuracy of communication with their HCP.

HCPs felt that the visit preparation feature reduced their administrative burden and improved the efficiency and effectiveness of appointments. Similarly, HCPs stated that patient education features reduced the education burden on them, providing a constructive focus for patients living with MDD and improving patient-provider interactions.

Patient-Provider Engagement

Patients living with MDD reported that the app helped to improve their interactions with their HCP by providing easy-to-use tools for tracking symptoms and progress. They also appreciated the ability to share their data with their HCP, which they said helped to facilitate better communication and collaboration. Patients and HCPs found that the goal-setting and goal-tracking features of the app supported clinical conversations while also helping patients focus on their treatment goals. HCPs reported that patients may remain engaged in their care by knowing that their input and concerns are being taken seriously, thereby promoting SDM.

Both patients living with MDD and HCPs expressed positive feedback about the patient education feature of the Pathway App. They found the feature to be visually appealing and engaging and deemed the content to be suitable in length and depth. Furthermore, HCPs reported that the education feature would facilitate productive conversations with patients.

Concerns

Concerns about the Pathway App were that it was difficult to use and had a lack of interactivity, with subthemes such as being confusing, repetitive, and time consuming and lacking workflow integration (Table 3).

The initial well-being tracker, a visual analog scale from 0 to 100, received the most critical feedback from patients living with MDD, as many of them found the construct and scale to be unnecessarily complicated. HCPs also expressed skepticism about the clinical utility of the well-being tracker feature and had concerns about the ease with which they could interpret the responses. Both HCPs and patients suggested changes to include a simple response format and visual indicators to support interpretation. The cognition 2-back feature also generated negative reviews, with both patients and HCPs expressing their frustration, as many found it to be confusing and anxiety inducing, and they expressed skepticism about the usefulness and interpretation of data. Participants also expressed frustration about the daily repetition of certain questions and expressed concerns about disengagement.

Recommendations

For the well-being tracker, participants suggested a simple response format and the inclusion of visual indicators to support interpretation (Table 3). A better explanation and provision of an alternative cognitive exercise were suggested by participants for the cognition 2-back feature. Providing examples of goals and assisting patients living with MDD with setting their own goals independently of their care teams were among the HCP-suggested changes to the goal-setting feature. Both HCPs and patients suggested changes to goal tracking, such as options to set goal reminders, record goal progress, and provide rewards to motivate goal attainment.

Many of the HCP-suggested changes included an emphasis on the value of EHR integration for easy access and to save time. They suggested that features such as HCP visit preparation, goal setting and tracking, and the Pathway report would benefit from integration with the EHR to be more useful and effective for HCPs. In addition to workflow integration, HCP-suggested changes included increasing the interactivity and variety of the given content for patient education, while also making the value and purpose of the educational materials clear.

Optimization of the Integration of the Pathway App Into a Primary Care Clinical Workflow

Overview

Care team members made several recommendations regarding how to facilitate the integration of the Pathway App into clinical workflows. These included product-specific recommendations and suggested changes to existing workflow scopes. Key themes that emerged were the importance of MBC, SDM, educational needs for the care team, and patient education (Table 4).

Importance of MBC

Although many primary care professionals within Advocate Aurora primary care perceived MBC, collaborative care, and SDM as important, various barriers limited their inclusion in day-to-day clinical workflows. Care teams agreed that Pathway administration of the PHQ-9 builds on and improves current clinical practices. HCPs viewed the PHQ-9 as a key component of current clinical practice; therefore, its continuous use to track patients’ treatment progress outside visits, as enabled by Pathway Platform, was perceived as clinically useful.

Shared Decision-Making

Tracking collaboratively devised goals was perceived as useful for the care team, with many also stating that this would help engage people living with MDD in their own care. HCPs reported that the presentation of patient-reported outcome (PRO) trajectories for medication adherence, PHQ-9 scores, and side effects provided a clinically useful view of how each patient’s condition has changed over time, which, in turn, supports better clinical decision-making. HCPs suggested including more information about patient care in the Pathway-EHR interface, such as medication refill data for comparison with reported adherence.

Workflow Integration and Data Visibility

Interviews also highlighted the importance of understanding care team needs, such as interpreting PRO measures and trend visualizations, to ensure that Pathway Platform can support care team workflows. Data about medication adherence, PHQ-9 scores, and side effects were perceived to be the most clinically important PROs. However, there were also concerns that Pathway data would be overwhelming and may lead to key data being missed.

Educational Needs

Many expressed low familiarity and desire for education regarding PROs other than PHQ-9 and clinical use of cognitive tests. In addition to their own education, the care team members also expressed a desire for patient-directed educational materials. Many care team members spoke about patients living with MDD being provided with educational handouts at their respective practices (including handouts printed from web-based searches, those that have been externally printed, and those taken from web-based medical reference platforms and the EHR). Many care team members stated that, from the list of educational topics (which included understanding one’s diagnosis, treatment options, side effect management, goal setting, and patient engagement), they were most interested in accessing materials about helping patients understand their diagnosis and their available treatment options. Care team members also thought that patients would benefit from educational resources about the types of local behavioral health resources that are available to them and information about what type of role they can play in their depression care and how often they should follow-up with their physician.

Discussion

Principal Findings

Qualitative research and thematic analyses conducted in this study allowed us to capture the user experiences and perceptions of patients living with MDD and their care teams. These results can guide researchers and app developers in designing effective digital health interventions that will be readily accepted by their intended end users [30].

Most of the currently available apps developed for depression management have only been assessed for effectiveness in a research setting and have not been integrated within clinical workflows; this has resulted in a lack of adoption by care teams and broad health care systems [26]. Incorporating the voices of people living with MDD and their care teams into the product development process aligns with the broad paradigm shift toward patient-focused decision-making and SDM between patients and their providers, ultimately leading to high-quality, fully informed, and preference-based treatment plans [39]. Our analyses revealed that both patients and their care teams perceived the remote tracking of PROs via digital tools to be clinically useful and reliable. Other highlights of our study included the need to enhance the flow of real-time data shared with the care team and the need to integrate within the care team workflow, including real-time sharing of the patient’s app data within the EHR. Results from our analyses also highlighted the need for care team education about MBC and SDM and about how to use the Pathway App to improve these processes by using features such as visit preparation and collaborative goal setting and tracking. Using the broad insights gathered from our thematic analyses, we were able to understand, iterate, and integrate a digitally enabled platform, Pathway Platform, into a primary care setting in the United States. The first iteration of the Pathway App included PRO measures related to depression, well-being, cognitive symptom tracking, medication adherence, and side effects [31]. Pilot results confirmed the feasibility of using the Pathway App among patients living with MDD and showed a trend in enhanced patient activation in the app arm, albeit in a small sample size [31]. Building on the results from the pilot study, Pathway Platform was developed to consist of 3 components: the latest iteration of the Pathway App; EHR-integrated, real-time, patient-level data sharing; and educational programming that is both care team and patient facing (including a web-based educational resource center that describes the utility of Pathway Platform to the care team through reading materials, presentations, and videos; Multimedia Appendix 4). The current version of the app prompts patients living with MDD to complete the following scales every 2 weeks: PHQ-9 [40] and Perceived Deficits Questionnaire—Depression [41] to assess depression status and subjective cognitive impairment, World Health Organization Well-being Index [42] to assess quality of life and emotional well-being, and Digit Symbol Substitution Test [43] as an objective measure of cognition to assess working memory and processing speed.

The Pathway App also includes a daily “evening check-in” to collect information about medication adherence and side effects. Data collected by the Pathway App are electronically transmitted and stored in an EHR-integrated web interface. These data are accessible to the care team and provide a longitudinal summary that may assist them in clinical decision-making and overarching depression management. Care team members can view these data either before or during the clinical visit and then use the data to collaboratively discuss future treatment decisions with the people they are treating. A web-based educational training program for primary care team members was also developed by using evidence-based medicine, building on the concepts of MBC and SDM, as they relate to depression management.

In addition, audit and feedback sessions will be conducted to benchmark performance measures, reflect on current clinical practice and improvement strategies, and set team-based action plans. Specific training sessions were conducted for the care team members about how to instruct patients to use Pathway Platform and how to use EHRs to view data collected via Pathway Platform. A training manual was also developed for patients living with MDD that describes the functionality of Pathway Platform and how to use and interpret their data.

Pathway Platform was cocreated with input from all users (people living with MDD, care teams, health system information technology personnel, and study collaborators), along with input from software developers and health-technology product development researchers, to optimize usability, utility, iteration speed, and integrated system performance and to ultimately enable nuanced care focused on SDM and MBC (Multimedia Appendix 5). Continued reassessment based on user feedback has allowed for fast iterations, optimized system performance, and sustainability [39]. This user-centric approach has, in turn, led to an enhanced digital platform to improve treatment outcomes by supporting an expanded understanding of MDD treatment, bolstering care team workflows, and providing patients with additional support throughout their treatment journey.

Limitations and Future Directions

A limitation of our study is its small sample size of patients with MDD and their care teams. In addition, among the care teams, only 33% (12/36) were clinicians. Future studies would benefit from an even split between clinicians and nurses in the sample. Moreover, this was a single-provider network study, and the results of this study may have limited generalizability. In addition, although the design and development of Pathway Platform were guided by a deep understanding of care team workflows, the extent to which clinical workflows can be modified to accommodate the adoption of Pathway Platform will ultimately depend on the clinical team.

Furthermore, although thematic analyses offered the necessary tools for organizing, interpreting, and transforming data without the need for separate theories, the depth of our conclusions may have been increased by additional methods such as modeling and theory building [36]. However, creating an overarching or generalizable theory to explain the way people reacted to the product’s feature set as a whole [44] would have gone beyond the primary interests of the study, which were focused on understanding and analyzing specific feedback such that it could be mindfully applied to iterations of the product features. We also used a highly systematized and enumerative approach to coding and generating themes, as recommended by the guidelines [34]. In addition, we minimized interpretive inconsistencies by using a single code tree and 2 analysts to evaluate each other’s work for analytical consistency. The improved iteration of Pathway Platform is being evaluated in an ongoing large-scale implementation study (Use of a Digitally Enabled App With Clinical Team Interface in the Management of Depression; NCT04891224). The study will include up to 200 patients at 20 primary care clinics. The implementation study aims to test the scaling and integration of Pathway Platform, along with educational interventions, at multiple primary care sites within the AAH system, with the primary objective of determining improvement in adherence to MBC practices [32]. Results are expected to provide insights into the improvements in clinical workflows that are necessary to enhance collaborative care, depression management, clinician and patient experience, adherence to medication, patient-provider engagement, and depression outcomes in the primary care setting [32]. EHR integration and how it enables decision-making, and efficiencies with current AAH information technology platforms, such as ease of access of data in real time by the care team, will also be assessed. Together, insights from this study will allow further amendment of workflows to ensure the optimal use of Pathway Platform [32].

Providing effective care for MDD has become more important than ever, with the prevalence of symptoms of anxiety disorder and depressive disorder having increased more than 3-fold in the United States during the COVID-19 pandemic [45,46]. Furthermore, people with few social and economic resources had high likelihood of exhibiting depression symptoms during this time [46]. Digital tools are therefore increasingly relevant in the era of COVID-19, owing to increased use of telehealth services to facilitate access to care [47]. In addition, low-cost interventions, such as digital tools, could provide increased monitoring and improved services to at-risk populations [15,48].

Moreover, future research methodologies, analysis protocols, and publications should provide a more explicit account of the impact of people’s social intersectionality on their perceptions about remote MDD monitoring. The imperative for this is heightened by the increasing attention that reimbursement entities are paying to the way any given intervention can help minimize the negative effects that social determinants have on treatment outcomes [49].

Conclusions

Ongoing communication with patients with MDD and their care teams (cocreation) provided an opportunity for continued reassessment and developmental iteration of Pathway Platform. These insights included the need for rapid communication of updated and current patient data with the care team, integration of the app into the MDD care pathway via the EHR, and education of the care team about the interpretation and use of these data. This cocreation model using qualitative research findings has resulted in fast iterations and optimized system performance and will allow for eventual sustainability outside the research environment. Future development of Pathway Platform will continue, consistent with the evolving needs of people living with MDD and their care teams.