https://orcid.org/0000-0003-3468-1732
Abstract
Digital interventions are increasingly in demand to address mental health concerns, with significant potential to reach populations that disproportionately face barriers to accessing mental health care. Challenges with user engagement, however, persist. The goal of this study was to develop user personas to inform the development of a digital mental health intervention (DMHI) for a perinatal population.
We used participatory User-Centered Design (UCD) methods to generate and validate personas (ie, representative profiles of potential users). We applied this methodology to a case example of an Anxiety Sensitivity Intervention. Phases included (1) Characteristic identification, (2) Persona generation, (3) Persona consolidation, (4) Persona validation, and (5) Persona refinement. Advisory Council members with lived expertise of perinatal mental health conditions generated 6 personas. We used cluster analysis and qualitative analysis to consolidate personas. We used participant interviews with perinatal individuals experiencing depression or anxiety and economic marginalization (n = 12) to qualitatively validate and refine these personas.
We identified 4 user personas with potentially unique design needs that we characterized as being “Resilient,” “Lonely,” “Overwhelmed,” and “Aware.”
Personas generated through this process had distinct characteristics and design implications including the need to prioritize (1) content personalization, (2) additional content describing support options and resources (eg, doulas, midwives), (3) careful consideration of the type of information provided by users, and (4) transparent options for information and data sharing.
DMHIs will need to be adapted for relevance for a perinatal population. The personas we developed are suggestive of the need for design considerations specific to distinct potential user groups within this population.
Lay Summary
There is increasing demand for digital solutions in mental health care, particularly for marginalized populations facing barriers to access. The goal of this study was to apply user-centered design approaches to guide the development of a digital mental health intervention (DMHI) for perinatal individuals (ie, pregnant and up to 1 year postpartum). Using participatory methods, we involved individuals with lived experience with perinatal mental health conditions to generate and refine 3 user personas. We validated personas with participant interviews, and identified a fourth distinct persona. Each persona suggests unique design needs. These personas highlighted the importance of considerations specific to perinatal mental health, including personalized content, access to support resources, and transparent data sharing options. Our findings emphasize the necessity of adapting DMHIs to meet the specific needs of perinatal populations, with consideration for diverse user groups.
Background and significance
Digital mental health interventions (DMHIs) include internet-delivered interventions, smartphone applications, and text message interventions. DMHIs can reduce depression and anxiety symptoms,1 and can be subject to regulating bodies such as the U.S. Food and Drug Administration.2 Increasingly proposed as an approach to making evidence-based interventions accessible at scale,3 DMHIs have the potential to remove barriers associated with multi-session in-person interventions. They also carry promise for reaching populations who are economically marginalized,4 or are excluded from health and financial services due to income and employment and less likely to receive high-quality mental health treatment.5,6 Despite this potential, DMHI engagement remains low outside the research context.7 Ensuring engagement is critical, as across DMHIs, greater engagement is associated with improvement in mental health symptoms.8
Designing DMHIs to address user needs is particularly important among perinatal populations. A study assessing a web-based treatment for perinatal anxiety found high dropout rates (82%), with users reporting that the intervention was not user-friendly and aspects of content were not relevant or appropriate.9 Many interventions developed for perinatal individuals lack end-user feedback in the design process.10 Once users initiate an intervention, engagement and sustainment can be low.10 More work is needed to develop interventions that are engaging and meet the needs of perinatal individuals.
User-Centered Design (UCD) is an approach that aims to increase intervention uptake and engagement by focusing on user needs throughout development.11,12 UCD is grounded in principles of cognitive psychology, acknowledging that human cognitive processes constrain and shape human-computer interactions.13 UCD incorporates concepts such as affordances, or the human-environment relations offering opportunities for perception and action.14 For instance, while the iPhone is a 2-dimensional surface, the interface employs 3-dimensional looking “buttons” to signify opportunities to swipe or click.13 A fundamental aspect of UCD is the recognition that users have pre-existing perceptions for interacting with their environments, which evolve over time due to factors such as technological advancements.13 For example, the rise of social media has significantly impacted how individuals engage with digital technologies.15 This theoretical foundation informs UCD methodology, which aims to develop a deep understanding and consideration of users and their evolving needs. UCD posits that this understanding results in a user interface that is engaging, addresses patient satisfaction, and meets user needs as they change. As such, UCD approaches are considered best practice in DMHI development.12
A lack of UCD is one key reason for low engagement with mental health apps, with apps omitting functions desired by their target populations and including features that users find burdensome.16 While more experimental work is needed to establish the impact of UCD on user engagement, several case studies have demonstrated how UCD approaches have identified and responded to unique user needs that may not otherwise have been identified.17,18
One UCD strategy is to develop personas, or user profiles of potential intervention users.11 A persona is a fictional character that represents a user group with shared behaviors, habits, and goals.18,19 Personas can clarify goals for and barriers to intervention use by creating a context for designers to empathize with potential users throughout the design process.11 Using personas can help ensure that design decisions reflect the needs of intended users rather than the assumption of the intervention developers. This approach seeks to understand user contexts to develop an intervention that is as relevant as possible to diverse members of the target population.20
Objectives
The goal of this study was to create user personas to inform the design of a DMHI for a perinatal population. We used the adaptation of an evidence-based intervention, called an Anxiety Sensitivity Intervention, as a case example that could inform the development of engaging DMHIs for perinatal populations more broadly.
Materials and methods
Digital mental health intervention
Anxiety Sensitivity Interventions are a prevention-oriented, cognitive-behavioral approach targeting anxiety sensitivity, or the maladaptive belief that physical symptoms of anxiety (eg, racing heart) are physically harmful. Anxiety sensitivity has been conceptualized as having physical, cognitive, or social components.21 This well-established risk-factor for anxiety disorders22,23 is also prospectively associated with panic attacks, mood disorders,23–27 suicidal ideation,28 and alcohol and substance use disorders.29–31 Anxiety sensitivity has been conceptualized as a cognitive vulnerability and transdiagnostic target for both prevention and treatment.32–34
Anxiety Sensitivity Interventions include <6 sessions and have been delivered in-person and through digital platforms.32 Psychoeducation focuses on beliefs about mental health by describing how physiological arousal (eg, heart racing, hyperventilation) accompanying anxiety is not physically dangerous but leads to avoidance of safe situations, which, in turn, reinforces and maintains anxiety.35,36 This approach has demonstrated improvements in anxiety and depression,22,32,35,37–41 and health concerns such as alcohol overuse and insomnia.32,42–47 While not yet examined in a perinatal population, Anxiety Sensitivity Interventions holds significant potential for improving mental health in perinatal populations, as these conditions negatively affect the health of perinatal individuals, birth outcomes, and the health and development of their children.43–47
Persona input
We used participatory UCD techniques to create personas representing DMHI users.11 We partnered with a national Advisory Council involved in the planning, completion, and dissemination of patient-centered outcomes research to address perinatal mental health inequities among perinatal individuals who have been marginalized.48 The Advisory Council includes 18 individuals with lived expertise of perinatal mood and/or anxiety disorders and experience serving in an advisory role in research. The group was developed from representative regions of the United States, inclusive of Black, Indigenous, Asian, multiracial, and gender minoritized groups.
We analyzed transcripts of interviews conducted with individuals who reported recent experiences of challenges with perinatal mood and/or anxiety disorders49 (henceforth: “participants”).
Finally, UCD principles suggest that Subject Matter Experts (SMEs; eg, clinicians serving a target population) should be included early on in intervention development in addition to end-users.50 We included input from SMEs who were researchers and clinicians in perinatal mental health (eg, perinatal psychiatrists).
By combining these 3 sources of input (Advisory Council, participants, SMEs), we sought to cover the breadth of possible users while also maximizing resources. Study phases included (1) Characteristic identification, (2) Proto-Persona generation, (3) Proto-Persona consolidation, (4) Persona validation, and (5) Persona refinement (Figure 1).
Participatory user-centered design process for creating personas to inform the design of a digital anxiety sensitivity intervention.
Phase I. Characteristic identification
We convened SMEs to identify characteristics that inform design considerations for the development of a DMHI. An expert in UCD (BT) held 2 workshops via videoconference with SMEs to develop characteristics for a DMHI. Workshop 1 (n = 3) focused on generating the characteristics. Workshop 2 (n = 4) focused on consolidating and finalizing characteristics. A modified Delphi approach51 was used in which SMEs responded to prompts independently prior to meeting. Results were shared in the meeting for discussion and approval. SMEs (n = 2) then met to adapt characteristics for relevance for the intervention context. See Appendix S1 for additional details. Characteristics were chosen to facilitate setting a context for each persona in the subsequent phases and to develop a common understanding to helps intervention designers to empathize with users. Criteria for selecting characteristics included relevance to the user context, potential impact on design decisions, and alignment with existing evidence. SMEs identified 9 characteristics and agreed upon the definition of these characteristics (Table 1).
Persona card template and user characteristics relevant to intervention design.
| Qualitative user characteristics—Persona card template | |
|---|---|
| Domain | Description |
| Demographics | Age, residence, background, income, marital and domicile status, children, faith, urban-suburban, rural, race, ethnicity, education |
| Facts and interactions | A brief history, basic information about the persona’s roles and responsibilities. Key information about the persona’s household and families |
| Challenges and feelings | Relevant health challenges that this persona faces, eg, worry about health is leading to fatigue, challenges in relationships and at work. |
| Goals, wants, and needs | What this persona wants from the intervention to meet their needs and goals and to overcome their challenges (eg, reduction in stress, more engaged with friends/family) |
| Qualitative user characteristics—Persona card template | |
|---|---|
| Domain | Description |
| Demographics | Age, residence, background, income, marital and domicile status, children, faith, urban-suburban, rural, race, ethnicity, education |
| Facts and interactions | A brief history, basic information about the persona’s roles and responsibilities. Key information about the persona’s household and families |
| Challenges and feelings | Relevant health challenges that this persona faces, eg, worry about health is leading to fatigue, challenges in relationships and at work. |
| Goals, wants, and needs | What this persona wants from the intervention to meet their needs and goals and to overcome their challenges (eg, reduction in stress, more engaged with friends/family) |
| Quantitative user characteristics | |
|---|---|
| Characteristic | Definition |
| Tech savviness | The comfort level of this person in using new technologies or technological products |
| Perceived need for mental health support | The degree to which lowering the likelihood of developing or worsening anxiety is important to this person |
| Mental health history | Previous experience receiving mental health services |
| Trauma history | Degree to which traumatic experiences currently impact persona’s daily life |
| Perceived social support | The degree to which supports are available or are not. Consider the number and intensity of relationships |
| Access to and continuity of care | The degree to which this person has access to care (low accessibility means high level of difficulty in getting care and defaulting to the emergency department for all healthcare). |
| Self-efficacy | The degree to which this person has access to care |
| Discrimination | The degree to which the user has experienced discrimination related to a marginalized identity |
| Trust in health care system | Degree to which person places trust in the health care system |
| Quantitative user characteristics | |
|---|---|
| Characteristic | Definition |
| Tech savviness | The comfort level of this person in using new technologies or technological products |
| Perceived need for mental health support | The degree to which lowering the likelihood of developing or worsening anxiety is important to this person |
| Mental health history | Previous experience receiving mental health services |
| Trauma history | Degree to which traumatic experiences currently impact persona’s daily life |
| Perceived social support | The degree to which supports are available or are not. Consider the number and intensity of relationships |
| Access to and continuity of care | The degree to which this person has access to care (low accessibility means high level of difficulty in getting care and defaulting to the emergency department for all healthcare). |
| Self-efficacy | The degree to which this person has access to care |
| Discrimination | The degree to which the user has experienced discrimination related to a marginalized identity |
| Trust in health care system | Degree to which person places trust in the health care system |
Note: Subject matter exports (SMEs) identified characteristics. Advisory Council members rated personas on each characteristic on a scale from 1 = low to 10 = high.
Persona card template and user characteristics relevant to intervention design.
| Qualitative user characteristics—Persona card template | |
|---|---|
| Domain | Description |
| Demographics | Age, residence, background, income, marital and domicile status, children, faith, urban-suburban, rural, race, ethnicity, education |
| Facts and interactions | A brief history, basic information about the persona’s roles and responsibilities. Key information about the persona’s household and families |
| Challenges and feelings | Relevant health challenges that this persona faces, eg, worry about health is leading to fatigue, challenges in relationships and at work. |
| Goals, wants, and needs | What this persona wants from the intervention to meet their needs and goals and to overcome their challenges (eg, reduction in stress, more engaged with friends/family) |
| Qualitative user characteristics—Persona card template | |
|---|---|
| Domain | Description |
| Demographics | Age, residence, background, income, marital and domicile status, children, faith, urban-suburban, rural, race, ethnicity, education |
| Facts and interactions | A brief history, basic information about the persona’s roles and responsibilities. Key information about the persona’s household and families |
| Challenges and feelings | Relevant health challenges that this persona faces, eg, worry about health is leading to fatigue, challenges in relationships and at work. |
| Goals, wants, and needs | What this persona wants from the intervention to meet their needs and goals and to overcome their challenges (eg, reduction in stress, more engaged with friends/family) |
| Quantitative user characteristics | |
|---|---|
| Characteristic | Definition |
| Tech savviness | The comfort level of this person in using new technologies or technological products |
| Perceived need for mental health support | The degree to which lowering the likelihood of developing or worsening anxiety is important to this person |
| Mental health history | Previous experience receiving mental health services |
| Trauma history | Degree to which traumatic experiences currently impact persona’s daily life |
| Perceived social support | The degree to which supports are available or are not. Consider the number and intensity of relationships |
| Access to and continuity of care | The degree to which this person has access to care (low accessibility means high level of difficulty in getting care and defaulting to the emergency department for all healthcare). |
| Self-efficacy | The degree to which this person has access to care |
| Discrimination | The degree to which the user has experienced discrimination related to a marginalized identity |
| Trust in health care system | Degree to which person places trust in the health care system |
| Quantitative user characteristics | |
|---|---|
| Characteristic | Definition |
| Tech savviness | The comfort level of this person in using new technologies or technological products |
| Perceived need for mental health support | The degree to which lowering the likelihood of developing or worsening anxiety is important to this person |
| Mental health history | Previous experience receiving mental health services |
| Trauma history | Degree to which traumatic experiences currently impact persona’s daily life |
| Perceived social support | The degree to which supports are available or are not. Consider the number and intensity of relationships |
| Access to and continuity of care | The degree to which this person has access to care (low accessibility means high level of difficulty in getting care and defaulting to the emergency department for all healthcare). |
| Self-efficacy | The degree to which this person has access to care |
| Discrimination | The degree to which the user has experienced discrimination related to a marginalized identity |
| Trust in health care system | Degree to which person places trust in the health care system |
Note: Subject matter exports (SMEs) identified characteristics. Advisory Council members rated personas on each characteristic on a scale from 1 = low to 10 = high.
Phase II. Proto-persona generation
Personas can be developed through interviews with end-users18 or by developing “proto-personas” through interviews with individuals who have indirect experience serving the population of interest such as mental health professionals.52 Incorporating proto-personas can minimize resources required to conduct qualitative interviews with a representative population.18,52
Advisory Council members attended a 90-minute convening via videoconference to engage in proto-persona generation. We presented the council members characteristics of our target population:
Currently pregnant
Experiencing symptoms of anxiety
Experiencing challenging financial circumstances
In 3 breakout rooms (5 council members; 2 moderators each), moderators asked each group to collaboratively create 2 personas. We asked group members to include the following categories of detail identified in Phase I as necessary and relevant to inform the design of the DMHI:
Demographics (eg, age)
Facts and Interactions (eg, basic information about roles, responsibilities)
Challenges and Feelings (eg, health challenges)
Goals, Wants, and Needs (eg, stress reduction)
Group facilitators used prompts when needed, then summarized the group’s contribution and asked members to rate the persona across the 9 relevant characteristics (Table 1; 1 = low to 10 = high). Groups then repeated the process with a second persona.
Following the meeting, we generated persona card templates representing the 6 personas generated by Advisory Council members (see Appendix S1). Each persona included qualitative information (ie, demographics, facts and interactions, challenges, and goals, wants and needs) and quantitative information (eg, degree of tech-savviness; Table 1).
Phase III. Proto-persona consolidation through cluster analysis and examination of qualitative data
We conducted hierarchical cluster analyses using Euclidean distance in the factoextra package53 in RStudio54 to examine quantitative characteristics (Table 1). The cluster analysis served as a starting point to organize qualitative data generated for each persona. Following the cluster analysis, we sought to identify commonalities among personas (eg, demographics, challenges, goals) of personas within the same cluster and compare across clusters. We created persona card templates of each persona displaying both the qualitative data (eg, demographics) and quantitative characteristic ratings (see Figure 2).
Sample persona card template developed following Advisory Council generated descriptions and characteristic ratings. Note: All 6 persona card templates are presented in Figure S4A.
Phase IV. Persona validation through comparison with qualitative participant interviews
To validate the consolidated personas, we used transcripts of participant interviews from a preliminary study to examine perinatal individuals’ recommendations for interventions to prevent perinatal depression and anxiety (n = 12).49 The recruitment process is described elsewhere.49,55 We chose these interviews as being representative of the target population of an Anxiety Sensitivity Intervention for individuals at risk for perinatal anxiety disorders. We chose a sample size of 12, as this amount has been shown to be sufficient for thematic saturation in qualitative research.56–58 All study procedures were approved by the institution’s Institutional Review Board. Eligible interview participants were age 18 years or older, able to participate in an interview over phone or computer in English, experienced mental health concerns during or after pregnancy, and reported either (1) enrollment in services such as Women Infants Children (WIC) Program, (2) un-insured status or insured by Medicaid or public insurance, (3) food or housing insecurity and were pregnant or recently pregnant during the COVID-19 pandemic. Eligible participants provided informed consent via telephone. Participants completed a demographic survey, reporting race/ethnicity, employment, education, income, history of mental health treatment, impact of childcare (“Do problems getting childcare make it difficult for you to work or study?”), housing insecurity (“Are you worried about losing your housing?”), food insecurity, (“In the last 12 months how often were the following statements true: ‘We worried whether our food would run out before we got money to buy more.’”).
We designed semi-structured interview guides to assess patient experiences with mental health challenges, preventive care (eg, how risk was assessed, referrals for preventative counseling, resources), and preferences for prevention services. Illustrative questions include, “Were you worried about depression or anxiety when you became pregnant?”, “During your pregnancy, would you have been interested in information about mental health?”, “What kind of information would you have been interested in?” The interview guide included follow-up probes designed to explore deeper contextual details. The guide was reviewed and revised by a state-level Advisory Council of perinatal care professionals and individuals with lived experience of perinatal mood and anxiety disorders.
Participants engaged in 30-minute interviews conducted by lead author (M.Z.) via videoconference. Interviews were recorded and transcribed by an external transcription service. The sample and interview method is described in more detail elsewhere.49 We used a coding consensus, co-occurrence, and comparison approach to code interviews.59 We open-coded interviews, developed a codebook, and 2 researchers independently coded transcripts using DeDoose software.38 Discrepancies were resolved by a third reviewer in group meetings. Codes were developed to characterize DMHI preferences (see Table S2A).
One researcher (M.Z.) then used demographic surveys and coded interviews to assess participant fit with each category of the persona card templates (Table 1) using an excel template. A persona was considered to be validated if the majority (>50%) of persona demographics, facts and interactions, challenges and feelings, goals, wants and needs, and characteristics aligned with the persona cards generated from the interviews (see illustrative quotes in Table S3A). This allowed us to qualitatively assess fit of each participant with the 3 clusters of the personas identified in Phase III.
Phase V. Persona refinement
We revised our personas based on our Phase IV analyses of interviews. We presented the Advisory Council generated persona card templates (eg, Figure 2) at a second convening of the Advisory Council. Members were invited to provide feedback on personas and design needs. We asked advisory council members to identify any gaps or additional considerations not included in the presented personas. We revised personas based on feedback.
Results
Proto-persona consolidation
We identified 3 personas based on a cluster analysis and qualitative examination of advisory council-generated persona characteristics. Cluster analysis based on the quantitative characteristics (eg, tech-savviness) suggested 3 clusters (Figure 3), each comprised of 2 Advisory Council-generated personas.
Hierarchical cluster analysis of Advisory Council-generated persona characteristics.
Persona Cluster 1 was “Lonely.” This persona was characterized by high tech savviness and access to care, low perceived need for mental health services, minimal mental health history, and lack of social support. Personas included in this cluster were younger, not partnered, and living in temporary housing situations with little social support. It was their first pregnancy.
Persona Cluster 2 was “Resilient.” This persona was characterized by the presence of mental health history, trauma, and discrimination, coupled with minimal social support outside of a partner that may or may not be supportive. They had few resources and were concerned about job security and access to support.
Persona Cluster 3 was “Overwhelmed.” This persona was characterized by the presence of social support, a lack of access to care, and distrust in the health care system. Personas included in this cluster had at least 1 child, and had a supportive partner. They faced challenges such as childcare needs, health concerns, and experiences such as pregnancy loss.
Persona validation
Review of coded interviews provided support for the 3 persona clusters generated by the Advisory Council. Across all interviews, participants emphasized the need for support and education of stigma for perinatal individuals with mental health challenges.49
Generally, participants we identified as aligned with the “Resilient” persona (n = 2) reported that their experiences of discrimination led to a distrust of their health care providers and fears of being stigmatized. These participants faced stressors such as food insecurity and significant stigma experienced both in the context of their relationships and in health care settings. This group reported a strong preference for normalizing mental health and for support through peers, doulas, and midwives.
Participants we identified as aligned with the “Overwhelmed” persona (n = 4) emphasized a busy schedule and childcare needs that necessitated easy access and quick feedback and communication with providers. This group tended to report a preference for information and normalization of mental health challenges in the perinatal period.
Finally, a fourth group of users emerged that we identified as “Aware” of mental health (n = 2). For this group, mental health history was at the forefront of their experiences. “Aware” personas differed from Advisory Council-generated personas in that they reported generally high access to care and trust in at least some pregnancy care providers. They were already connected with mental health professionals and pregnancy support professionals such a midwives and doulas because of their previous experiences. This group was comfortable discussing mental health with their pregnancy care team but reported concern about their risk for perinatal mood and anxiety disorders as a result of their past experiences. This group reported interest in partner involvement in their mental health care and involvement of pregnancy care providers and mental health professionals.
We presented the 6 proto-personas, resultant 3 persona clusters, and proposed additional persona at the second Advisory Council convening. The Advisory Council recommended including a non-cisgender persona to better represent the spectrum of perinatal people, in addition to greater representation of multiple pregnancy loss and/or In Vitro Fertilization (IVF) pregnancies. We included these characteristics in our revised personas. Our resulting set of 4 personas included “Resilient,” “Lonely,” “Overwhelmed,” and “Aware,” users (see Table 2).
Brief descriptions of final personas.
| “Lonely” | “Resilient” | “Overwhelmed” | “Aware” |
|---|---|---|---|
| This group of users tends to be younger, experiencing their first pregnancy, and most importantly may not feel represented in by materials representing “traditional” pregnancy experiences or motherhood (eg, representations of planned/wanted pregnancy, partnered pregnancy, heterosexual relationships). This group tends to be tech-savvy and uses apps. They are looking for virtual connection but often prefer to connect anonymously. They have low social support and want to feel less alone. They may be living in temporary housing, and searching seeking more stability. | This group of users has faced trauma and discrimination with little support or positive experiences from the health care system. They may have struggled with depression, anxiety, or PTSD that was never recognized or treated. They may have a partner but lack other sources of support or may not feel comfortable talking about mental health with support system. They may face challenges with food insecurity and are looking for better job security to provide for their families. They are looking for social and emotional support. They do not feel trusting of the health care system and prefer other sources of support.a | This group of users may have one or more children already, health complications, history of pregnancy loss, or fertility concerns. While they may have a social support system (eg, they have supportive partner, family) they have limited access to care and limited internet access. They tend to use the internet for health information when available. They are seeking help with daily demands of childcare. They are seeking better employment and/or educational opportunities. They are looking for information and support delivered in a way that can meet the need of their busy schedules. | This group of users has a history of depression, anxiety, or other mental health concerns and has received mental health services. They feel generally trusting of the health care system but are worried that pregnancy will be a stressor that exacerbates or leads to a recurrence of their mental health challenges. This group is comfortable discussing mental health with service providers and is actively seeking strategies to support mental health during pregnancy. |
| “Lonely” | “Resilient” | “Overwhelmed” | “Aware” |
|---|---|---|---|
| This group of users tends to be younger, experiencing their first pregnancy, and most importantly may not feel represented in by materials representing “traditional” pregnancy experiences or motherhood (eg, representations of planned/wanted pregnancy, partnered pregnancy, heterosexual relationships). This group tends to be tech-savvy and uses apps. They are looking for virtual connection but often prefer to connect anonymously. They have low social support and want to feel less alone. They may be living in temporary housing, and searching seeking more stability. | This group of users has faced trauma and discrimination with little support or positive experiences from the health care system. They may have struggled with depression, anxiety, or PTSD that was never recognized or treated. They may have a partner but lack other sources of support or may not feel comfortable talking about mental health with support system. They may face challenges with food insecurity and are looking for better job security to provide for their families. They are looking for social and emotional support. They do not feel trusting of the health care system and prefer other sources of support.a | This group of users may have one or more children already, health complications, history of pregnancy loss, or fertility concerns. While they may have a social support system (eg, they have supportive partner, family) they have limited access to care and limited internet access. They tend to use the internet for health information when available. They are seeking help with daily demands of childcare. They are seeking better employment and/or educational opportunities. They are looking for information and support delivered in a way that can meet the need of their busy schedules. | This group of users has a history of depression, anxiety, or other mental health concerns and has received mental health services. They feel generally trusting of the health care system but are worried that pregnancy will be a stressor that exacerbates or leads to a recurrence of their mental health challenges. This group is comfortable discussing mental health with service providers and is actively seeking strategies to support mental health during pregnancy. |
PTSD: Posttraumatic Stress Disorder.
Brief descriptions of final personas.
| “Lonely” | “Resilient” | “Overwhelmed” | “Aware” |
|---|---|---|---|
| This group of users tends to be younger, experiencing their first pregnancy, and most importantly may not feel represented in by materials representing “traditional” pregnancy experiences or motherhood (eg, representations of planned/wanted pregnancy, partnered pregnancy, heterosexual relationships). This group tends to be tech-savvy and uses apps. They are looking for virtual connection but often prefer to connect anonymously. They have low social support and want to feel less alone. They may be living in temporary housing, and searching seeking more stability. | This group of users has faced trauma and discrimination with little support or positive experiences from the health care system. They may have struggled with depression, anxiety, or PTSD that was never recognized or treated. They may have a partner but lack other sources of support or may not feel comfortable talking about mental health with support system. They may face challenges with food insecurity and are looking for better job security to provide for their families. They are looking for social and emotional support. They do not feel trusting of the health care system and prefer other sources of support.a | This group of users may have one or more children already, health complications, history of pregnancy loss, or fertility concerns. While they may have a social support system (eg, they have supportive partner, family) they have limited access to care and limited internet access. They tend to use the internet for health information when available. They are seeking help with daily demands of childcare. They are seeking better employment and/or educational opportunities. They are looking for information and support delivered in a way that can meet the need of their busy schedules. | This group of users has a history of depression, anxiety, or other mental health concerns and has received mental health services. They feel generally trusting of the health care system but are worried that pregnancy will be a stressor that exacerbates or leads to a recurrence of their mental health challenges. This group is comfortable discussing mental health with service providers and is actively seeking strategies to support mental health during pregnancy. |
| “Lonely” | “Resilient” | “Overwhelmed” | “Aware” |
|---|---|---|---|
| This group of users tends to be younger, experiencing their first pregnancy, and most importantly may not feel represented in by materials representing “traditional” pregnancy experiences or motherhood (eg, representations of planned/wanted pregnancy, partnered pregnancy, heterosexual relationships). This group tends to be tech-savvy and uses apps. They are looking for virtual connection but often prefer to connect anonymously. They have low social support and want to feel less alone. They may be living in temporary housing, and searching seeking more stability. | This group of users has faced trauma and discrimination with little support or positive experiences from the health care system. They may have struggled with depression, anxiety, or PTSD that was never recognized or treated. They may have a partner but lack other sources of support or may not feel comfortable talking about mental health with support system. They may face challenges with food insecurity and are looking for better job security to provide for their families. They are looking for social and emotional support. They do not feel trusting of the health care system and prefer other sources of support.a | This group of users may have one or more children already, health complications, history of pregnancy loss, or fertility concerns. While they may have a social support system (eg, they have supportive partner, family) they have limited access to care and limited internet access. They tend to use the internet for health information when available. They are seeking help with daily demands of childcare. They are seeking better employment and/or educational opportunities. They are looking for information and support delivered in a way that can meet the need of their busy schedules. | This group of users has a history of depression, anxiety, or other mental health concerns and has received mental health services. They feel generally trusting of the health care system but are worried that pregnancy will be a stressor that exacerbates or leads to a recurrence of their mental health challenges. This group is comfortable discussing mental health with service providers and is actively seeking strategies to support mental health during pregnancy. |
PTSD: Posttraumatic Stress Disorder.
We then identified and applied design considerations relevant for adapting an Anxiety Sensitivity Intervention for mobile health with a perinatal population using stated preferences from qualitative interviews (Table 3) and consideration of qualitative persona characteristics. Design considerations included: (1) content personalization, (2) additional content describing support options and resources (eg, doulas, midwives), (3) careful consideration of the type of information provided by users, and (4) transparent options for information data sharing (see Table 4). We provide examples of how we modified content of the Anxiety Sensitivity Intervention based on these considerations (Figure 4).
Sample persona-informed design applications. Note: (A) Highlighting privacy and transparency in a way that emphasizes confidentiality (“Resilient,” “Lonely” personas) but also provides option for sharing information if desired (“Aware” persona) (B) Providing options for support outside the pregnancy care team (“Resilient persona”) (C) Emphasizing destigmatizing messages, (“Resilient,” “Lonely” personas) (D) Describing unsafe situations as separate from other anxiety-provoking situations (“Resilient persona”), (E) Providing option to engage at a later time to accommodate schedule (“Overwhelmed” persona). Content was built in the Computerized Intervention Authoring System version 3.0 (CIAS 3.0; cias.app).
Persona characteristics and associated preferences generated from qualitative interviews with perinatal individuals with lived expertise of perinatal mood and anxiety disorders who experienced economic marginalization (n = 12).
| Characteristics | Characteristics illustrative quote | Reported preferences | Preferences illustrative quotes | |
|---|---|---|---|---|
| “Lonely” |
|
|
|
|
| “Resilient” |
|
|
|
|
| “Overwhelmed” |
|
|
|
|
| “Aware” |
|
|
|
|
| Characteristics | Characteristics illustrative quote | Reported preferences | Preferences illustrative quotes | |
|---|---|---|---|---|
| “Lonely” |
|
|
|
|
| “Resilient” |
|
|
|
|
| “Overwhelmed” |
|
|
|
|
| “Aware” |
|
|
|
|
Persona characteristics and associated preferences generated from qualitative interviews with perinatal individuals with lived expertise of perinatal mood and anxiety disorders who experienced economic marginalization (n = 12).
| Characteristics | Characteristics illustrative quote | Reported preferences | Preferences illustrative quotes | |
|---|---|---|---|---|
| “Lonely” |
|
|
|
|
| “Resilient” |
|
|
|
|
| “Overwhelmed” |
|
|
|
|
| “Aware” |
|
|
|
|
| Characteristics | Characteristics illustrative quote | Reported preferences | Preferences illustrative quotes | |
|---|---|---|---|---|
| “Lonely” |
|
|
|
|
| “Resilient” |
|
|
|
|
| “Overwhelmed” |
|
|
|
|
| “Aware” |
|
|
|
|
Persona-informed design considerations.
| Persona 1 “Lonely” | Persona 2 “Resilient” | Persona 3 “Overwhelmed” | Persona 4 “Aware” | |
|---|---|---|---|---|
| Content personalization | May have mixed feelings about pregnancy, may not be partnered, may not identify as woman or have same-gender partner → Content should have options to consider nature of pregnancy (planned versus unplanned) presence of partner, gender identity of user and partner | History of depression or anxiety without treatment, discrimination, or negative experiences w/health care professionals, stigma related to mental health → Focus on destigmatizing language | Overwhelmed by busy schedule (ie, child care, work, managing health concerns) → Brief content with option to learn more | Experience with mental health services, comfortable talking about mental health → Mental health terminology included |
| Additional content | First-time parent → Addition of more general information about what to expect |
|
|
|
| Information provided by users | Does not perceive high need but may be curious about intervention → Anonymity (eg, username) | Mental health stigma → Anonymity (eg, username) | Overwhelmed by busy schedule → Ensure intervention does not feel burdensome (eg, frequent asks for symptom monitoring) | Perceived need is high → Wants to provide information, interested in connecting with others |
| Data and information sharing | Challenges with phone bill → Option for downloadable content that does not require data | Experiences with discrimination have led to mistrust → Anonymity (eg, username) and transparency | Lack of reliable internet access → Option for downloadable content that does not require data | Already connected with therapist → Options for sharing information with therapist |
| Persona 1 “Lonely” | Persona 2 “Resilient” | Persona 3 “Overwhelmed” | Persona 4 “Aware” | |
|---|---|---|---|---|
| Content personalization | May have mixed feelings about pregnancy, may not be partnered, may not identify as woman or have same-gender partner → Content should have options to consider nature of pregnancy (planned versus unplanned) presence of partner, gender identity of user and partner | History of depression or anxiety without treatment, discrimination, or negative experiences w/health care professionals, stigma related to mental health → Focus on destigmatizing language | Overwhelmed by busy schedule (ie, child care, work, managing health concerns) → Brief content with option to learn more | Experience with mental health services, comfortable talking about mental health → Mental health terminology included |
| Additional content | First-time parent → Addition of more general information about what to expect |
|
|
|
| Information provided by users | Does not perceive high need but may be curious about intervention → Anonymity (eg, username) | Mental health stigma → Anonymity (eg, username) | Overwhelmed by busy schedule → Ensure intervention does not feel burdensome (eg, frequent asks for symptom monitoring) | Perceived need is high → Wants to provide information, interested in connecting with others |
| Data and information sharing | Challenges with phone bill → Option for downloadable content that does not require data | Experiences with discrimination have led to mistrust → Anonymity (eg, username) and transparency | Lack of reliable internet access → Option for downloadable content that does not require data | Already connected with therapist → Options for sharing information with therapist |
Persona-informed design considerations.
| Persona 1 “Lonely” | Persona 2 “Resilient” | Persona 3 “Overwhelmed” | Persona 4 “Aware” | |
|---|---|---|---|---|
| Content personalization | May have mixed feelings about pregnancy, may not be partnered, may not identify as woman or have same-gender partner → Content should have options to consider nature of pregnancy (planned versus unplanned) presence of partner, gender identity of user and partner | History of depression or anxiety without treatment, discrimination, or negative experiences w/health care professionals, stigma related to mental health → Focus on destigmatizing language | Overwhelmed by busy schedule (ie, child care, work, managing health concerns) → Brief content with option to learn more | Experience with mental health services, comfortable talking about mental health → Mental health terminology included |
| Additional content | First-time parent → Addition of more general information about what to expect |
|
|
|
| Information provided by users | Does not perceive high need but may be curious about intervention → Anonymity (eg, username) | Mental health stigma → Anonymity (eg, username) | Overwhelmed by busy schedule → Ensure intervention does not feel burdensome (eg, frequent asks for symptom monitoring) | Perceived need is high → Wants to provide information, interested in connecting with others |
| Data and information sharing | Challenges with phone bill → Option for downloadable content that does not require data | Experiences with discrimination have led to mistrust → Anonymity (eg, username) and transparency | Lack of reliable internet access → Option for downloadable content that does not require data | Already connected with therapist → Options for sharing information with therapist |
| Persona 1 “Lonely” | Persona 2 “Resilient” | Persona 3 “Overwhelmed” | Persona 4 “Aware” | |
|---|---|---|---|---|
| Content personalization | May have mixed feelings about pregnancy, may not be partnered, may not identify as woman or have same-gender partner → Content should have options to consider nature of pregnancy (planned versus unplanned) presence of partner, gender identity of user and partner | History of depression or anxiety without treatment, discrimination, or negative experiences w/health care professionals, stigma related to mental health → Focus on destigmatizing language | Overwhelmed by busy schedule (ie, child care, work, managing health concerns) → Brief content with option to learn more | Experience with mental health services, comfortable talking about mental health → Mental health terminology included |
| Additional content | First-time parent → Addition of more general information about what to expect |
|
|
|
| Information provided by users | Does not perceive high need but may be curious about intervention → Anonymity (eg, username) | Mental health stigma → Anonymity (eg, username) | Overwhelmed by busy schedule → Ensure intervention does not feel burdensome (eg, frequent asks for symptom monitoring) | Perceived need is high → Wants to provide information, interested in connecting with others |
| Data and information sharing | Challenges with phone bill → Option for downloadable content that does not require data | Experiences with discrimination have led to mistrust → Anonymity (eg, username) and transparency | Lack of reliable internet access → Option for downloadable content that does not require data | Already connected with therapist → Options for sharing information with therapist |
Discussion
We used participatory UCD techniques to elucidate characteristics and preferences of perinatal individuals using a DMHI. We first developed proto-personas with an Advisory Council and then validated and refined these personas with a small number of participant interviews. This approach was less resource-intensive than conducting in-depth interviews with a large number of end-users. Moreover, this process allowed us to foster a community-research partnership early on in intervention development, an approach that aligns with recent calls to advance health equity.60 Results suggested that proto-personas generated by the Advisory Council were aligned with participant interviews, but that they did not encompass all personas, as we added an additional persona that was evident from participant interviews. The resulting 4 distinct personas among our intended population of users can inform the prioritization of features and functionalities in a DMHI for perinatal individuals.
Overall, DMHIs for perinatal mental health will need to include examples relevant to perinatal individuals, such as anxiety related to health, health of the baby, and stress related to social norms and expectations.61 Content should also be adapted to reflect culturally responsive principles, and use of de-stigmatizing language.62,63 In addition to these adaptations, the personas we developed are suggestive of the need for design considerations specific to distinct potential user groups of perinatal individuals.
First, content should be personalized to address the varied experiences and circumstances of pregnancy as much as possible. For instance, the “Lonely” users emphasized that hearing about pregnancy in the context of a supportive partner relationship contributed to feelings of loneliness and isolation. Similarly, these individuals reported feeling excluded by the assumption that pregnancies were planned. Incorporating options for users to specify these key experiences and have these inform subsequent messages could increase user engagement and satisfaction.
Additional content is needed to meet the needs of diverse user groups. While the “Aware” users may have experience with mental health treatment and language and be interested in learning strategies to address mental health, users like the “Resilient” users reported stigma and discrimination related to identity and mental health that requires a different approach. Stigma is a barrier for engagement with mental health interventions among many perinatal individuals.64,65 This group of users may be more likely to engage with strategies after having received more information to destigmatize mental health. While perinatal individuals in all persona groups reported a need for emotional support and social connection, the “Resilient” users reported a lack of information and access to resources and support. Providing information about doulas could address this need.
The type of user information collected by DMHIs should also be considered. “Lonely” users reported interest in using apps to obtain information and connect virtually with others but had concerns about anonymity. Privacy issues are particularly relevant to perinatal populations, who may have concerns about reporting requirements and child protective services involvement.66 This finding is consistent with recent calls to better protect data privacy in the context of mobile health interventions for mental health to address privacy concerns and lack of trust.67–69 Options to use an anonymous username and navigate the app without providing identifying information could facilitate comfort among this group of users.
Similarly, preferences for data sharing differed across the user groups. Users like the “Overwhelmed” users may have some interest in sharing information with pregnancy care team members, if those individuals demonstrated trusting relationships and ability to discuss mental health. “Aware” users were interested in sharing information with both mental health and health care professionals. These sharing preferences could be offered, while being explicit that users can elect not to share information to better engage users with privacy concerns. Past work has suggested that the incorporation of options may increase acceptability for all users.70
While personas have been used to develop digital interventions for perinatal populations (ie, individuals affected by miscarriage71; perinatal individuals with gestational diabetes72) their application to DMHIs is understudied. This gap is important, as individuals with mental health concerns in the perinatal period face unique challenges.65 The present study offers a starting place for DMHI developers to understand potential user needs and design for optimal user engagement. These personas developed in the context of UCD can also be used to document changing needs in a way that is replicable and systematic. This work provides the opportunity to test an assumption of UCD approaches, which is that understanding users at this level will improve engagement with DMHIs. Future work could evaluate whether end-users of an intervention with specific personas find persona-relevant design features to be more useful, engaging, and if they use them more. Future research can also examine the relationship between these personas and relevant psychological characteristics in order to inform tailoring. For instance, when adapting an Anxiety Sensitivity Intervention, future research could assess whether dimensions of anxiety sensitivity (eg, physical, social, or cognitive concerns) are more relevant for particular personas. It is also important to note that user needs change over time. Thus, personas should be iteratively refined to reflect current needs.
Limitations
A limitation of the present study was that all Advisory Council members and interview participants had access to internet to participate. It is possible that perinatal individuals without access to a smartphone or consistent internet may have generated additional user personas and relevant design considerations (eg, texting-only interventions for individuals without access to a smartphone). In addition, we were only able to interview individuals who were English speaking. This approach significantly limits our ability to understand design needs of individuals might use this intervention in another language.
While potentially useful for developing interventions for a target population, qualitative development of personas also has limitations. Qualitative interviews can be resource intensive and personas can be difficult to validate.73 While the present study sought to address the first concern, including Advisory Councils to reduce the number of participant interviews, we did not empirically validate the personas with a user population.73 The development of proto-personas is assumption based,52 and it is possible that expert characteristics could differ from the literature or from end-users. To address the limitations of developing the initial personas using an Advisory Council, we also investigated responses from naïve users (participants) to validate and refine the initial personas. Our validation process was subjective, as the research team made decisions regarding closeness of fit between qualitative data and persona card templates. Concerns have also been raised about the reproducibility of this approach to qualitative persona development.73 It is possible that the same methods could lead to the generation and validation of different personas if different users from the target population had been sampled. However, personas are living documents change over time and need refinement through multiple iterations. Our study provides valuable knowledge for future studies targeting this population for DMHIs. Future studies can use a mixed-methods or quantitative methods to validate personas in a larger sample to refine our understanding of the population while also capturing the changing needs of users.
Conclusion
DMHIs have the potential to impact the landscape of perinatal mental health by offering high-quality interventions at scale. Understanding diverse user needs through persona generation could promote greater population reach and generate sustained user engagement that is necessary to impact mental health. These UCD approaches have the potential to better meet the unique needs of perinatal individuals who face barriers to accessing behavioral health interventions.
Acknowledgments
We wish to express our deep appreciation for the invaluable insights and guidance provided by the Lifeline for Moms Postpartum Mental Health Advisory Council.
Author contributions
Martha Zimmermann (Conceptualization, Formal analysis, Investigation, Writing—Original Draft), Kimberly A. Yonkers (Methodology, Writing—Review and Editing), Karen M. Tabb (Resources, Writing—Review and Editing), Ana Schaefer (Resources, Writing—Review and Editing), Elizabeth Peacock-Chambers (Writing—Review and Editing), Camille A. Clare (Writing—Review and Editing), Edwin D. Boudreaux (Writing—Review and Editing), Stephenie C. Lemon (Writing—Review and Editing), Nancy Byatt (Writing—Review and Editing), and Bengisu Tulu (Conceptualization, Resources, Supervision, Writing—Review and Editing).
Supplementary material
Supplementary material is available at JAMIA Open online.
Funding
This work was supported by the National Center for Advancing Translational Sciences, National Institutes of Health, through Grant KL2TR001454. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Conflicts of interest
C.A.C sits on the Board of Trustees of the National Medical Association and on the New York Statewide Breastfeeding Coalition. N.B. has received salary and/or funding support from Massachusetts Department of Mental Health via the Massachusetts Child Psychiatry Access Program for Moms (MCPAP for Moms). N.B. is also the Medical Director of Research and Evaluation for MCPAP for Moms and the Executive Director of the Lifeline for Families Center at UMass Chan Medical School. She has received honoraria from Global Learning Collaborative. She has also served as a consultant for The Kinetix Group, VentureWell, and JBS International. All other authors have no competing interests to declare. All other authors have no interests to declare.
Data availability
Data will be made available upon request.
