Developing Dietary Self-Management Personas for mHealth Tools Among Patients with Kidney Failure Undergoing Hemodialysis: A Multi-Method Study

Abstract

Background

Dietary management remains a widespread challenge among patients with kidney failure undergoing hemodialysis. Although mobile health (mHealth) is considered a promising approach for providing self-management support, most existing tools lack the specificity to effectively address patients’ diverse behavioral patterns.

Objective

To develop user personas for dietary self-management among Chinese patients on hemodialysis and explore the needs of these personas regarding mHealth tools.

Methods

A sequential multi-method design was employed. In-depth qualitative interviews were conducted with 28 patients on hemodialysis recruited from a university-affiliated hospital between March and July 2025 to explore their dietary self-management experiences. Transcripts were analyzed using Colaizzi’s method, and dietary self-management personas were derived through manual thematic synthesis and iterative consensus among researchers. These personas were then presented with descriptive labels. In a subsequent quantitative phase, a cross-sectional survey was administered to a separate cohort of 167 patients on hemodialysis between October 2025 and February 2026. Participants were asked to self-identify with up to two personas that best reflected their current dietary behaviors, in order to assess the coverage and distribution of the developed personas.

Results

Qualitative analysis identified six distinct personas: (A) Proactive; (B) Occupation-Driven Maladaptive; (C) Family Care-Prioritizing Maladaptive; (D) Knowledge-Deficient; (E) Instant Gratification-Driven; and (F) Poor Execution Ability. Quantitative results showed that all participants identified with at least one of the six personas, with no patients reporting behaviors outside this classification. Persona A was the most prevalent (35.93%), and 19.76% of participants identified two personas.

Conclusion

The six personas reflect significant diversity in dietary self-management among patients undergoing hemodialysis and provide a preliminary, comprehensive typology that can inform the design of stratified, user-centered mHealth interventions. Future studies are needed to validate the feasibility and effectiveness of these personas.

Introduction

Kidney failure (KF), also known as end-stage kidney disease, is the most advanced stage of chronic kidney disease and constitutes a significant global health burden due to its substantial impact on human well-being.¹ Approximately 0.1% of the global population is suffering from KF.² Current treatment modalities include hemodialysis (HD), peritoneal dialysis (PD), and kidney transplantation (KT). Although KT is considered the optimal therapeutic option for eligible patients with KF, practical limitations—such as donor organ scarcity, significant financial demands, and complex donor–recipient compatibility—necessitate that many patients rely on dialysis for survival.³ The global prevalence underscores a dependence on HD, with an average of 323 patients per million population undergoing this therapy.⁴ This trend is particularly notable in China, where the population of patients on HD has increased at an annual rate of approximately 10% over the past decade, reaching 1,137,962 by the end of 2025.⁵

Although HD can prolong the lives of patients with KF, it does not fully restore homeostatic regulation due to its intermittent regimen (typically three 4-hour sessions per week) and the loss of native kidney metabolic clearance, making rigorous dietary restrictions essential to prevent uremic toxicity, volume overload, and electrolyte disturbances. However, inadequate dietary management is widely observed among HD recipients. A systematic review revealed that approximately 47.3–72.5% of individuals on HD exhibit poor dietary behaviors, including excessive fluid intake and consumption of foods high in potassium, sodium, and phosphorus, along with insufficient intake of protein, vitamins, and dietary fiber.⁶ Such behaviors directly jeopardize critical health outcomes, including uncontrolled blood pressure, compromised nutritional status, impaired quality of life, increased hospitalization rates, and elevated healthcare costs.⁷ Accordingly, enhancing patients’ dietary self-management is essential to mitigate these risks and improve prognosis.

With the advancement of technology, mobile health (mHealth) has emerged as a promising approach for providing timely, accessible, and interactive self-management support.⁸ However, patients’ self-management varies widely and is shaped by numerous contextual factors, including disparities in health literacy, socioeconomic constraints, cultural dietary norms, motivation, and competing life priorities.^9–11 Consequently, patients’ needs and preferences regarding mHealth tools can differ substantially. While users are naturally inclined toward tools that are tailored to their specific circumstances, developers often insufficiently account for this diversity during the design process.¹² This mismatch between users’ real-world needs and the generalized features of most mHealth applications can lead to diminished user engagement, loss of motivation, and ultimately, tool abandonment.¹³ Such limitations underscore the urgent requirement for strategies that can systematically capture and address the heterogeneity of patient needs within mHealth tool design.

User personas have been proposed as one such strategy. Derived from real-world user research, interviews, and data analysis, user personas synthesize the attributes of target users into virtual “persona profiles”, facilitating product teams’ understanding and empathy regarding users’ needs, routines, and pain points.¹⁴ Originally developed for consumer industries, user personas are now increasingly adopted in healthcare to translate patient diversity into actionable design guidance.¹⁵ Previous studies have developed patient personas for various chronic conditions, such as coronary heart disease,¹⁶ chronic obstructive pulmonary disease,¹⁷ and breast cancer.¹⁸ Additionally, persona-based interventions have been in chronic disease self-management,¹⁹ postoperative cancer rehabilitation,²⁰ and mental health,²¹ demonstrating encouraging improvements in patient engagement and clinical outcomes. However, the application of personas to dietary management in hemodialysis patients remains unexplored. This gap underscores the necessity for a condition-specific, empirically grounded characterization of patient heterogeneity in this context.

The present study represents the initial step of a larger research project aimed at developing a persona-based mHealth tool to support dietary self-management among patients on HD. Specifically, this study sought to develop user personas by examining patients’ dietary self-management experiences and to explore their distinct needs regarding mHealth tools. We hypothesized that distinct dietary self-management personas exist among patients undergoing hemodialysis. The findings are expected to offer a preliminary, empirically grounded typology that can inform the design of tailored mHealth interventions.

Methods

A conceptual framework was developed to guide this study, articulating the theoretical relationships among qualitative behavioral dimensions, patient persona typologies, and tailored mHealth design strategies (Figure 1). Within this framework, qualitative behavioral dimensions serve as the analytical foundation that characterizes and differentiates patient persona typologies. Each persona represents a distinct configuration of these behavioral dimensions, capturing the heterogeneity of patient self-management behaviors. In turn, the specific needs, barriers, and behavioral patterns associated with each persona inform and guide the design of corresponding mHealth strategies.

Figure 1.

Conceptual framework.

Study Design

A sequential multi-method design was employed, consisting of a qualitative phase for persona development followed by a quantitative phase for assessing the coverage and distribution of the derived personas within a broader patient cohort.

Sampling and Recruitment

Qualitative Phase

In-depth interviews were conducted with patients undergoing HD. Participants were recruited from the HD center at a tertiary hospital affiliated to a university in Shanghai, China, between March 2025 and July 2025, using purposive sampling and a maximum variation strategy. The inclusion criteria were as follows: (1) age ≥18 years; (2) a minimum of three months on HD treatment; (3) provision of informed consent; and (4) proficiency in smartphone usage. Exclusion criteria comprised: (1) psychiatric or cognitive disorders, (2) physical limitations hindering dietary self-care, (3) use of HD as a transitional therapy, and (4) communication barriers. Sample size determination was guided by data saturation, with interviews continuing beyond the initial ten participants until no new themes were identified across three successive cases.²²

Quantitative Phase

After completing the qualitative research phase, a cross-sectional questionnaire survey was administered from October 2025 to February 2026 to a separate cohort of patients receiving HD. The inclusion and exclusion criteria for this phase were identical to those used in the qualitative study. All eligible patients who met the inclusion criteria during the survey period at the study center were invited to participate.

Data Collection

Qualitative Data Collection

A semi-structured interview guide was systematically developed in accordance with the study objectives. The initial draft was formulated following an extensive literature review and discussions with a multidisciplinary team. Pilot interviews were subsequently conducted with three patients undergoing HD, after which the interview guide was iteratively refined based on feedback on question clarity, relevance, and emotional sensitivity.

The finalized interview protocol included open-ended questions such as: “How do you comprehend and implement dietary management in your daily life?” “What challenges have you faced in dietary management?” and “What are your perspectives on utilizing mHealth tools for dietary self-management?” Prior to each interview, participants were briefed on the study’s purpose, data collection procedures, and confidentiality protocols. Written informed consent was obtained from all participants.

Interviews were conducted in a quiet, private health education room within the HD unit to minimize environmental disturbances. Researchers systematically observed and documented nonverbal behaviors (eg, hesitation, emotional expressions) to enrich the interpretation of qualitative data. Each interview lasted 20–40 min, and audio recordings were transcribed verbatim within 24 h. Transcripts were cross-referenced with field notes to resolve ambiguities, and participants were contacted for clarification when required.

Quantitative Data Collection

To assess the coverage and distribution of qualitatively derived dietary management personas, a concise self-identification questionnaire was developed to enable participants to identify the persona(s) most consistent with their current dietary management patterns. The questionnaire comprised two sections: (1) Demographic and clinical information—age, sex, marital status, educational level, cohabitation status, duration of HD, and primary disease; (2) Persona selection framework—this section presented the six user personas derived from the qualitative analysis. Each persona was defined by a brief, plain-language description of its core behavioral characteristics, as identified in the thematic analysis. Recognizing the complexity of real-world behavior, participants were permitted to select up to two personas that best represented their current dietary management behaviors. The selection was based entirely on participants’ self-assessment of their own behaviors.

Data Analysis

Development of Persona Attribute Dimensions

Qualitative data were analyzed using NVivo 12.0 software, employing Colaizzi’s seven-step analysis method.²³ The process involved: (1) comprehensive and repeated review of each interview transcript to achieve an in-depth understanding and immersion in the data; (2) extraction of significant statements directly related to dietary self-management behaviors, with particular attention to core elements such as adherence challenges, decision-making processes, and psychosocial factors; (3) development of descriptive codes aligned with the extracted statements; (4) organization of codes into recurrent patterns and emergent themes; (5) generation of detailed descriptions for each theme, integrating representative raw participant statements; (6) iterative comparison of similar themes and their descriptions to identify and extract analogous perspectives, resulting in concise, meaningful thematic constructs; and (7) presentation of the emerging thematic structure to research participants for validation, ensuring accurate representation of their lived experiences. These themes constituted the analytical framework underpinning manual feature extraction and persona construction.

Manual Feature Extraction to Develop Patient Personas

Building on the thematic analysis, personas were developed using a manual feature extraction process to retain the depth and contextual nuance inherent in the qualitative data, which is a recognized strength of interpretative qualitative methods. Two members of the research team independently reviewed the full set of interview transcripts and the emergent thematic structure. For each participant, they systematically assessed behavioral characteristics across the core dimensions identified in the thematic framework. Participants exhibiting similar patterns of characteristics were grouped into preliminary candidate personas. The two researchers then convened to compare their independently derived groupings. Any interpretative discrepancies were resolved through iterative discussion and consensus-building, with reference back to the original data when necessary. The final persona attributes were validated through a cyclical process of analysis, deliberation, and refinement.

Presentation of Patient Personas

The final step in developing user personas involved clearly and effectively presenting them with an intuitive, structured format. In this study, descriptive labels were used to represent user personas for dietary self-management among patients on HD, emphasizing the unique and diverse traits of their self-management behaviors.

Assessment of the Applicability of Patient Personas

Descriptive statistical analyses were performed using SPSS software (version 24.0) to examine the coverage and distribution of the patient personas identified from the qualitative analysis. Demographic and clinical characteristics of survey participants were summarized. Categorical variables were described using frequencies, whereas continuous variables were reported using standard deviations or medians with interquartile ranges. To examine the co-occurrence of persona affiliations, the number and proportion of participants selecting each possible combination of up to two personas were computed. This approach enabled the identification of predominant persona patterns and the extent of subgroup overlap, thereby providing evidence of the applicability of the constructed personas within a broader patient population. The persona development process is illustrated in Figure 2.

Figure 2.

Flowchart of the persona development process.

Ethical Considerations

This study received approval from the Ethics Committee of a tertiary hospital affiliated to a university in Shanghai, China (No. XHEC-D-2025-076), and was conducted in accordance with the Declaration of Helsinki. All participants provided written informed consent before enrollment. The consent form comprehensively delineated the study’s objectives, procedures, potential risks and benefits, voluntary participation, the right to withdraw at any stage without affecting clinical care, detailed data confidentiality measures, and the use of anonymized responses and direct quotes in publications.

For the participants in the qualitative phase, additional consent was obtained for audio recordings. Participants were informed of their right to pause or terminate the recording at any time, thereby ensuring autonomy and comfort throughout the interview process. All audio recordings, interview transcripts, and survey data were stored in password-protected systems with access restricted to authorized research team members. Data analysis involved de-identification to protect participant anonymity, and no identifiable information was incorporated into publications or presentations.

Results

Qualitative Findings

Participant Characteristics

A total of 28 eligible patients on HD were interviewed, including 17 men and 11 women. Participants ranged in age from 34 to 70 years, with a mean age of 54.11 years. The duration of HD treatment ranged from 1 to 15 years. Regarding educational attainment, 57.14% of patients had completed at least high school or received higher education. Detailed participant characteristics are provided in Table 1.

Table 1.

Participant Characteristics in the Qualitative Phase

Number	Sex	Age (Years)	Educational Level	Cohabitant	Duration of HD (Years)	Primary Illness
N1	Female	55	Senior high school	Spouse	15	Chronic glomerulonephritis
N2	Female	48	Senior high school	Parents	5	Chronic glomerulonephritis
N3	Female	36	Undergraduate	Relatives	3	Chronic glomerulonephritis
N4	Male	54	Junior college	Spouse	2	Diabetic nephropathy
N5	Female	65	Senior high school	Spouse and children	3	Hypertensive nephropathy
N6	Male	65	Junior high school	Spouse	2	Renal carcinoma
N7	Female	63	Junior high school	Lives alone	4	Polycystic kidney
N8	Male	46	Senior high school	Spouse	1	Chronic glomerulonephritis
N9	Male	44	Senior high school	Parents	4	Chronic glomerulonephritis
N10	Female	66	Primary school	Lives alone	5	Chronic glomerulonephritis
N11	Female	64	Junior high school	Spouse	6	Diabetic nephropathy
N12	Male	34	Undergraduate	Parents	10	Hypertensive nephropathy
N13	Male	52	Undergraduate	Spouse	5	Diabetic nephropathy
N14	Male	55	Senior high school	Spouse	1	Chronic glomerulonephritis
N15	Male	66	Junior college	Spouse and children	3	Diabetic nephropathy
N16	Female	50	Junior high school	Relatives	10	Chronic glomerulonephritis
N17	Male	48	Junior high school	Spouse	1	Diabetic nephropathy
N18	Male	70	Senior high school	Spouse	10	Polycystic kidney
N19	Female	35	Junior high school	Relatives	15	Chronic glomerulonephritis
N20	Female	65	Junior high school	Spouse	2	Diabetic nephropathy
N21	Male	60	Senior high school	Spouse	5	Diabetic nephropathy
N22	Male	64	Junior high school	Spouse	7	Hypertensive nephropathy
N23	Male	37	Undergraduate	Lives alone	2	Chronic glomerulonephritis
N24	Male	48	Junior college	Spouse	2	Hypertensive nephropathy
N25	Female	50	Junior high school	Spouse and children	6	Diabetic nephropathy
N26	Female	66	Junior high school	Spouse	3	Diabetic nephropathy
N27	Male	65	Junior high school	Spouse	4	Hypertensive nephropathy
N28	Female	44	Senior high school	Parents	8	Chronic glomerulonephritis

Abbreviation: HD, hemodialysis.

Dietary Self-Management Personas of Patients on HD

Utilizing thematic analysis of qualitative interviews, an analytical framework was established encompassing three core dimensions to characterize dietary self-management among patients undergoing HD: (1) characteristics of dietary management and (2) needs for mHealth tools. This framework guided the identification of six distinct user personas (Table 2).

Table 2.

Personas of Dietary Management for Patients on Hemodialysis

Name	A: Proactive	B: Occupation-Driven Maladaptive	C: Family Care-Prioritizing Maladaptive	D: Knowledge-Deficient	E: Instant Gratification-Driven	F: Poor Execution Ability
Representative individuals	N1, N2, N12, N13, N22	N3, N4, N14, N23, N24	N5, N15, N16, N25	N6, N7, N17, N18, N26	N8, N9, N19, N27	N10, N11, N20, N21, N28
Characteristics of Dietary Management	Comprehensive professional knowledge Strong self-discipline Diary-style dietary record Dynamic plan optimization Good family support	Time deprivation Takeout dependency Social compromise Behavioral disconnection	Time compression Sacrificing personal needs Maintaining the “all-round caregiver” image Avoidance of causing “trouble”	Inadequate Risk Awareness Low Health Literacy Reliance on Non-Professional Information Sources Insufficient Family Support	Instant gratification orientation Pain compensation Lack of dietary management motivation Absence of long-term dietary planning	Cognitive-behavioral disconnect Persistent behavioral inertia Often accompanied by self-blame Low self-efficacy Strong dependence on external support
Needs for mHealth tools	Advanced data analysis Personalized nutrition recommendations Access to cutting-edge information	Recommended quick recipes Privacy protection Dietary management reminders synchronized with work schedule	Time-efficient integration tool Family collaboration support	User-friendly operation Systematic knowledge dissemination Step-by-step task guidance	Gamified experience Sensory feedback Social competition elements Short-term goal orientation	Positive feedback Automatic diet plan scheduling Smart alert mechanism

Abbreviations: HD, hemodialysis; mHealth, mobile health.

Persona A: Proactive

Dietary Management Characteristics

This group comprises individuals aged 34–64 years, with equal representation of both sexes and substantial experience with HD treatment. Their persona is characterized by integrated professional knowledge and exceptional self-discipline. They systematically acquire HD-specific dietary management principles, implement accurate intake adjustments through daily quantitative monitoring, and proficiently incorporate medical feedback to refine their regimen. Through these practices, they maintain a comprehensive, self-directed approach to health management, further supported by the collaborative support of family members, a critical advantage.

Needs for mHealth Tools

They desire decision-making support through comprehensive analysis of their health metrics. Personalized nutrition plans tailored to their individual conditions are also highly valued. Additionally, they seek regular access to cutting-edge medical information and innovative management strategies to further optimize their dietary management.

N1: I keep detailed records of my daily meals, including the amount of each food item consumed, so I can ensure my diet aligns with the dietary guidelines for patients on HD. After all, I must take responsibility for myself.

N2: I regularly communicate with doctors and nurses, and I search for relevant online information on certain issues. Based on my most recent test results, I adjust my diet accordingly. I find this to be very important and helpful.

N13: I want a tool that can integrate my lab results with my daily food records and provide early warnings, so I can adjust my diet before problems arise.

Persona B: Occupation-Driven Maladaptive

Dietary Management Characteristics

Currently employed and predominantly male, patients in this group are aged 37–55 and have been on HD for a relatively short period (1–3 years). Their demanding workloads contribute to chronic energy depletion, which negatively impacts dietary management, leading to frequent reliance on takeout food and compromises during work-related social dining. These behaviors contrast with their self-disciplined dietary practices outside of the workplace. Additionally, occupational responsibilities require these patients to maintain the privacy of their health information, thereby complicating dietary management.

Needs for mHealth Tools

This group requires mHealth solutions that minimize additional cognitive burden while seamlessly integrating into their professional routines. Specifically, tools should offer time-efficient features such as instant access to HD-compliant quick recipes and unobtrusive reminders aligned with their work schedules. Equally critical is robust privacy protection to ensure that health management does not inadvertently expose their condition in professional settings.

N3: Work is quite hectic on weekdays, and sometimes I have to work overtime. Where would I find time or energy to manage my diet? I end up ordering takeout. Moreover, during social engagement, I have no choice but to eat with everyone else. Otherwise, I would seem like a real oddball.

N4: On weekends, I plan my meals carefully, but once the workweek hits, everything becomes haywire. Sometimes, I forget to watch my water intake.

N23: I hope the application can recommend quick and healthy recipes for me, preferably in sync with my work schedule, such as reminding me not to drink too much coffee before meetings. Privacy protection is also essential. I do not want too many people to know I am sick.

Persona C: Family Care-Prioritizing Maladaptive

Dietary Management Characteristics

This group consists mainly of females aged 50–66 who have been on HD for a moderate period (3–7 years). Most reside in multigenerational households, with two or three generations living together. They often assume caregiving responsibilities, prioritizing their family members’ needs over their own health. To maintain their caregiver roles, they frequently neglect their own dietary management and refrain from expressing their struggles. Their daily schedules are dominated by household tasks, and meal choices typically reflect family preferences. As a result, they focus on caring for others while neglecting themselves.

Needs for mHealth Tools

For this group, mHealth tools should prioritize time-efficient integration—such as automated meal logging and shared meal planning that accommodate both dietary restrictions and family preferences. Equally important are features that facilitate family collaboration, enabling family members to offer gentle support and reminders, thereby transforming dietary management into a shared activity rather than an individual burden.

N5: I have to take care of two children every day, so I simply do not have the time to think about my meals. When it comes to eating, I always have to accommodate the whole family. I cannot simply consider my own needs. If food is not suitable for me, I eat a little less. What else can I do? Cook separately? (bitter laugh).

N15: I do not want my family to worry about my illness, so even when they ask what I need, I do not want to say too much. I have already caused them enough trouble, and I do not want them to think I am a fragile patient who requires constant care.

N16: I hope the application can recommend recipes suitable for patients on HD and healthy individuals, and allow me to quickly log my meals. It would be ideal if my family could view my entries and send me reminders. Please keep this simple. I really do not have much time.

Persona D: Knowledge-Deficient

Dietary Management Characteristics

Older adults (mean age: 62.4 years) dominate this group, with HD durations ranging widely from 1 to 10 years. Their health literacy is generally low; many do not fully understand the importance of dietary management or recognize the serious consequences of poor dietary choices. They often lack basic nutritional knowledge and struggle to identify foods that pose risks for patients undergoing HD. Easily swayed by unreliable sources, they often make dietary decisions based on personal experience or intuition and receive limited support from their families, making health management and knowledge acquisition even more challenging.

Needs for mHealth Tools

They need tools that are extremely simple to use, with intuitive interfaces and step-by-step guidance. Clear, systematic dietary knowledge presented in plain language is essential. Basic reminders and straightforward instructions can help them build confidence in making correct food choices.

N7: Honestly, I do not really know which foods are safe to eat and which are not. Sometimes, when I hear other patients say something is off-limits, I just avoid it. However, I do not actually know why.

N26: He (my spouse) already does a lot with grocery shopping and cooking. Whatever he cooks, I eat. I cannot really ask for more. No one helps me keep track of these things (my eating habits) and after trying to log them myself for a few days, I gave up. It does not seem like a big deal anyway.

N18: I would love a simple, user-friendly application that lets me watch short videos, preferably as effortlessly as scrolling through TikTok, that walks me through step-by-step instructions. It should assign me a few easy tasks each day that I can follow.

Persona E: Instant Gratification-Driven

Dietary Management Characteristics

This patient group is predominantly young and male. Their dietary choices are driven by immediate sensory perceptions, often relying on highly palatable foods to alleviate distress. They generally avoid systematic dietary planning and display low responsiveness to health-monitoring feedback. This manifests as a cyclical pattern of short-term satisfaction, with limited intrinsic motivation to sustain dietary management.

Needs for mHealth Tools

They value engaging, rewarding user experiences that provide immediate gratification. Gamified features, such as point systems and achievement badges, can help sustain their motivation. Short-term goal setting with positive feedback is more appealing than long-term abstract planning.

N8: HD is just so painful. Sometimes after treatment, I really crave something delicious, like a piece of cake, to treat myself. I know it is not good for me, but I just cannot resist it.

N9: Sticking to low-sodium/low-potassium diets long-term? Impossible. Health matters, but so does joy. Fried chicken tastes heavenly, and Coke hits the spot. Who cares about the future? Comfort first.

N19: If the application could be like a game with rewards for completing tasks and the ability to compete with fellow patients, I would be willing to use it. Each entry can feature pleasant sound effects and animations, making it more engaging and encouraging for regular use.

Persona F: Poor Execution Ability

Dietary Management Characteristics

The patient group, predominantly female and aged between 44 and 66 years, has undergone HD for a moderate duration and demonstrates sufficient cognitive understanding. Their most prominent characteristic was knowing what to do, but struggling to follow through. Despite clear dietary guidelines, implementation proves challenging due to entrenched behavioral patterns. Persistent failures in adhering to guidelines often result in intense self-criticism and diminish self-efficacy. Although they express willingness to improve dietary management, a lack of structured external support hinders their ability to consistently maintain effective behaviors independently.

Needs for mHealth Tools

Their needs centered on external support. They emphasize the importance of automated, simplified action plans that minimize the need for personal decision-making, real-time alerts to promptly address high-risk dietary behaviors, and ongoing microrewards as performance feedback.

N11: I understand what the doctor says, but I just cannot follow through. Every time I eat something I should not, I regret it and feel terrible about myself, but I still cannot resist the next time (sigh).

N20: I need someone to constantly remind me and keep me in check; otherwise, I easily slip back into my old eating habits, even though I know they are unhealthy.

N21: I wish for a tool that could just plan my meals for me every day — you know, tell me what to eat — and warn me if I am about to do something wrong. And maybe it could also encourage me, like say “Hey, you did great today!” That would really help me feel better about myself.

Quantitative Findings

Participant Characteristics

A total of 167 patients undergoing HD were included in the survey (Table 3). The mean age was 61.77 ± 11.3 years, and the average duration of HD treatment was 5.98 ± 5.28 years.

Table 3.

Characteristics of the Participants in the Quantitative Phase by Persona (n = 167)

Variable	All (n = 167)	Persona A (n = 71)	Persona B (n = 10)	Persona C (n = 9)	Persona D (n = 37)	Persona E (n = 44)	Persona F (n = 29)
Age (years)	61.77 ± 11.3	63.76 ± 10.61	44.60 ± 6.77	57.44 ± 13.30	65.51 ± 7.69	57.70 ± 11.43	61.07 ± 9.03
Sex
Male	100(59.9%)	46(64.8%)	6(60.0%)	5(55.6%)	19(51.4%)	30(68.2%)	14(48.3%)
Female	67(40.1%)	25(35.2%)	4(40.0%)	4(44.4%)	18(48.6%)	14(31.8%)	15(51.7%)
Marriage
No	10(6.0%)	4(5.6%)	1(10.0%)	2(22.2%)	1(2.7%)	3(6.8%)	1(3.4%)
Yes	149(89.2%)	65(91.6)	9(90.0%)	6(66.7%)	34(91.9%)	38(86.4%)	28(96.6%)
Divorced/widowed/other	8(4.8%)	2(2.8%)	0(0.0%)	1(11.1%)	2(5.4%)	3(6.8%)	0(0.0%)
Educational level
Primary school	6(3.6%)	2(2.8%)	0(0%)	0(0%)	3(8.1%)	1(2.3%)	1(3.4%)
Junior high school	35(21.0%)	15(21.1%)	1(10.0%)	2(22.2%)	10(27.0%)	11(25.0%)	3(10.3%)
Senior high school	84(50.3%)	36(50.7%)	1(10.0%)	3(33.3%)	20(54.1%)	19(43.2%)	21(72.4%)
Junior college or above	42(25.1%)	18(25.4%)	8(80.0%)	4(44.5%)	4(10.8%)	13(29.5%)	4(13.8%)
Cohabitant
Lives alone	31(18.6%)	14(19.8%)	3(30.0%)	1(11.1%)	5(13.5%)	9(20.4%)	3(10.4%)
Relatives	16(9.6%)	4(5.6%)	2(20.0%)	4(44.4%)	2(5.4%)	5(11.4%)	1(3.4%)
Children	8(4.8%)	4(5.6%)	0(0.0%)	1(11.1%)	3(8.1%)	0(0.0%)	2(6.9%)
Spouse	112(67.0)	49(69.0%)	5(50.0%)	3(33.3%)	27(73.0%)	30(68.2%)	23(79.3%)
Monthly income (yuan)
<3000	4(2.4%)	2(2.8%)	0(0.0%)	1(11.1%)	0(0.0%)	2(4.5%)	0(0.0%)
3000–5000	39(23.4%)	16(22.5%)	0(0.0%)	2(22.2%)	8(21.6%)	12(27.3%)	9(31.0%)
≥5000	124(74.3%)	53(74.7%)	10(100%)	6(66.7%)	29(78.4%)	30(68.2%)	20(69.0%)
Duration of HD (years)	5.98 ± 5.28	6.78 ± 5.71	3.25 ± 3.67	5.44 ± 6.09	4.22 ± 3.59	5.85 ± 4.33	6.62 ± 5.58
Primary illness
Chronic glomerulonephritis	68(40.7%)	35(49.3%)	3(30.0%)	5(55.6%)	16(43.3%)	16(36.4%)	6(20.7%)
Diabetes	39(23.4%)	11(15.5%)	1(10.0%)	1(11.1%)	13(35.1%)	16(36.4%)	6(20.7%)
Hypertension	42(25.1%)	18(25.3%)	3(30.0%)	2(22.2%)	6(16.2%)	6(13.6%)	14(48.3%)
Other	18(10.8%)	7(9.9%)	3(30.0%)	1(11.1%)	2(5.4%)	6(13.6%)	3(10.3%)

Abbreviation: HD, hemodialysis.

Proportions of Persona Combinations

Out of the 167 participants, 134 (80.24%) selected one of the six personas to represent their dietary management behaviors (Figure 3). The largest proportion of participants (35.93%) self-identified as persona A, followed by personas D (16.17%) and E (13.77%). Additionally, 33 participants (19.76%) self-identified with two distinct personas; the proportions selecting personas A and F, or D and E, were identical (5.39% each).

Figure 3.

Distribution and overlap of persona prevalence (n = 167).

Discussion

This study employed a multi-method desigh to explore the personas of dietary self-management behaviors among patients undergoing HD. Through qualitative interviews, six distinct personas were developed. In the subsequent quantitative survey, every participant identified with one or two of the six personas, with no cases falling outside the system. This comprehensive coverage further supports its applicability for designing targeted interventions.

Among the six personas, persona A (Proactive) was the most prevalent, accounting for 35.93% of the sample. While this suggests that a considerable proportion of patients have established effective self-management routines, it also indicates that the majority—nearly two-thirds—continue to face varying degrees of difficulty in dietary management, highlighting the demand for intensified support. Notably, the remaining five personas reflect distinct types of difficulty. Personas D, E, and F—representing knowledge deficits, motivational deficiencies, and execution failures, respectively—each accounted for substantial proportions (16.17%, 13.77%, and 7.19%), jointly underscoring the multifaceted nature of intrapersonal barriers to dietary self-management.^10,24 In contrast, Personas B and C, though smaller in proportion (each < 5%), reveal context-dependent barriers rooted in specific life roles—active employment and intensive caregiving responsibilities—highlighting the role of environmental and social circumstances in shaping dietary behaviors. Together, these personas demonstrate that dietary self-management difficulties in patients on HD arise from a complex interplay of intrapersonal and contextual factors, rather than a single underlying cause.

Additionally, 19.76% of participants identified with two personas. Rather than indicating insufficient boundary distinctiveness of the six personas, this co-occurrence likely reflects the multidimensional nature of real-world dietary behaviors, wherein some patients experience multiple coexisting barriers or behavioral tendencies. Notably, the overlaps were concentrated—A (Proactive) and F (Poor Execution Ability), and D (Knowledge-Deficient) and E (Instant Gratification-Driven)—instead of being uniformly distributed across all possible combinations. This clustering suggests that these overlaps represent meaningful behavioral dimensions that can coexist within individuals, such as the conflict between knowing what to do and struggling to execute it, or between lacking knowledge and seeking immediate rewards. These findings underscore the need for mHealth interventions with functionally tailored modules that correspond to individual patient personas, addressing the limitations inherent in conventional “one-size-fits-all” health management approaches.²⁵

Proactive Patients

This group represents patients who have successfully integrated dietary management into their daily routines, demonstrating strong knowledge, consistent self-monitoring, and the ability to dynamically adjust their plans based on clinical feedback. Their high level of behavioral consistency and autonomy, reinforced by robust family support, suggests that they have achieved a relatively stable and effective self-management state.

Despite their proficiency, these patients still expresse a substantial need for personalized nutrition guidance and advanced data analytics, marking a shift from basic tracking toward data-driven decision-making. This self-identified need for more proactive, predictive tools—rather than remedial support—suggests that traditional static information may be insufficient to adequately support their need for adaptive, forward-looking management. Tools that combine longitudinal physiological data with algorithmic models could offer a means to address this gap by providing real-time, customized dietary adjustments, which could in turn enable precision nutrition.²⁶ Notably, this cohort’s active exploration of mHealth tools positions them as innovators within the diffusion of innovation framework.²⁷ Engaging them as co-designers of mHealth solutions may enhance clinical relevance and user adoption of new features.²⁸ Future research should explore their potential as peer mentors or advocates to facilitate broader dissemination of digital interventions across diverse patient groups.

Occupation-Driven Maladaptive Patients

This population face a conflict between professional demands and rigorous self-management required. Their dietary behavior exhibits pronounced situational duality: discipline often lapses at work because of time constraints, social dining pressures, and reliance on takeout, resulting in inadvertent intake of high-phosphorus and high-potassium meals;²⁹ however, compliance improves outside of work. This pattern suggests that the primary barrier is not knowledge or motivation deficits but contextual behavioral disconnection driven by workplace conditions.

Accordingly, rather than providing general dietary education, which addresses a deficit these patients do not appear to have, it may be more effective to support streamlined decision-making in high-pressure settings. For example, integrating location-based technology with HD dietary algorithms could provide real-time, compliant quick-dining recommendations. Privacy protection features may also warrant consideration, reflecting concerns about “health stigma” in the workplace, particularly among younger working individuals with chronic illness.³⁰ Thus, for these patients, mHealth tools should be designed to safeguard their professional identity, ensuring that health management does not become an additional source of workplace vulnerability.

Family-Care Prioritizing Maladaptive Patients

This group illustrates how ingrained caregiving roles can consistently undermine dietary self-management, particularly among female patients living in multigenerational households. Their tendency to prioritize family needs over personal health reflects a role conflict in which dietary adherence is deprioritized to maintain a caregiver identity. This pattern not only leads to dietary compromises but also perpetuates a cycle of self-neglect that may worsen clinical outcomes over time.

Given that the primary challenges stem from limited time and conflicting roles, it may be more effective to design mHealth interventions that shift from individual-focused models toward strategies that incorporate family involvement. Specifically, tools should automate data collection (eg, via smart scales or cups) to minimize the burden of manual logging.³¹ More critically, these solutions should facilitate family coordination with minimal added effort. For instance, shared recipe algorithms could generate meal options that accommodate dialysis-related restrictions and the broader family’s preferences, such as “one-dish, two-preparations” options. Furthermore, collaborative features that enable family members to monitor progress or receive gentle reminders—while preserving the patient’s primary autonomy—may help transform health management into a collective, supportive activity.

Knowledge-Deficient Patients

Lack of HD-specific dietary knowledge constitutes the fundamental barrier for these patients. Without a clear understanding of what to eat and why, they often rely on nonprofessional information sources, leading to misconceptions that increase the risk of complications such as hyperkalemia and volume overload. Compounding this is insufficient family support, which further limits their access to consistent monitoring or encouragement.³²

Given their older age and lower educational background, these patients require knowledge delivered in simple, digestible formats and tools that are easy to operate. To address the knowledge gap, a “food traffic light” system³³—using color-coded indicators to signal food suitability—may help patients quickly identify safe and unsafe options. Supplemented by short, plain-language videos explaining core dietary principles, this approach could replace fragmented, unreliable online information with structured, accessible education. At the same time, the interface should minimize technical barriers such as voice input, icon-based navigation, large-font displays, and step-by-step guidance.³⁴ By simplifying knowledge delivery and lowering operational demands, such tools may empower these patients to move from passive reliance on uncertain information toward more informed and confident dietary decisions.

Instant Gratification-Driven Patients

This group is characterized by a tendency to prioritize short-term sensory pleasure over long-term health goals, as reflected in their frequent consumption of highly palatable foods to alleviate hemodialysis-related distress. This pattern, known as pain compensation, functions as an emotional coping strategy and often undermines their motivation for sustained dietary management.³⁵

Since this behavior is driven by a need for immediate emotional relief rather than a lack of knowledge or skills, interventions that offer immediate positive feedback align more directly with the underlying behavioral driver than educational or planning approaches would. Accordingly, mHealth interventions for this group could be designed to reframe health management as a source of instant positive feedback. Gamified strategies—such as earning points for dietary compliance, unlocking virtual achievements, or receiving tangible incentives—may make routine tasks more engaging and satisfying.³⁶ Additionally, integrating elements of social comparison, such as compliance leaderboards or team challenges, could leverage peer influence to sustain engagement.³⁷ Adopting a short-term goal orientation—breaking management tasks into weekly challenges rather than indefinite commitments—may address their preference for tangible, immediate feedback.

Patients with Poor Execution Ability

These patients know what to do but struggle to follow through, illustrating the central dilemma of cognitive-behavioral disconnect in chronic disease management. They understand dietary recommendations and intend to adhere, yet persistent behavioral inertia prevents sustained action. Repeated failures often lead to self-blame and diminished self-efficacy, creating a cycle of inactivity, distress, and dependency.³⁸ This suggests that effective management for these patients relies heavily on external, structured support rather than further increasing their internal motivation or knowledge.

Accordingly, mHealth interventions should function as an external executive assistant, compensating for self-regulatory deficits rather than adding to their cognitive burden. This could be achieved through three complementary strategies. First, automatic diet plan scheduling—such as pregenerated, simple daily menus—can reduce the paralysis of initiation and free patients from complex decision-making. Second, smart alert mechanisms can provide timely, proactive reminders—for instance, a push notification when a patient attempts to log a high-potassium food.³⁹ Finally, acknowledging incremental achievements through continuous positive feedback, such as a progress badge after a week of consistent adherence, can help restore self-efficacy.⁴⁰ By offering structured planning, timely guidance, and positive reinforcement, mHealth tools may help bridge the gap between intention and action, enabling these patients to build greater internal control over time.

Limitations

Several limitations of this study should be acknowledged. First, the personas were generated through qualitative analysis of interview data and iterative researcher consensus, which inherently involves subjective interpretation. While we applied systematic analytic procedures (Colaizzi’s method) and conducted independent assessments by two team members followed by consensus-building, the resulting personas still reflect the researchers’ perspectives and may not capture all aspects of dietary self-management experienced by patients. Second, the representativeness of the sample was limited, as recruitment was confined to a single-center hospital, restricting geographic, cultural, and socioeconomic diversity. Consequently, the proposed personas should be regarded as conceptual design tools that require validation in diverse healthcare settings and cultural contexts. Third, the lack of longitudinal behavioral tracking represents another limitation. The cross-sectional nature of the data precluded analysis of temporal changes in dietary behaviors or whether patients transition between personas over time. Future research should adopt longitudinal tracking methods to explore how user personas evolve and adapt. Fourth, technological validation was insufficient. Although this study identified mHealth functionality needs for six distinct personas, no prototype was developed or empirically tested. To address this gap, interdisciplinary collaboration among software engineers, clinicians, and researchers is essential for developing and testing prototypes in randomized controlled trials.

Conclusion

This study employed both qualitative and quantitative methods to develop, for the first time, six distinct user personas of dietary self-management among patients undergoing HD, underscoring significant differences in patients’ behaviors and their corresponding functional needs for mHealth design. These findings provide a behavioral evidence base for tailoring dietary self-management support to meet the diverse needs of patients on HD. A stratified design approach—delivering core functions to all patients while incorporating customized features for specific subgroups—may serve as a valuable framework for future mHealth development. Future research should validate these findings in multi-center cohorts across diverse regions and dialysis centers, and further translate these personas into prototype tools to evaluate their impact on clinical outcomes.

Funding Statement

This study was supported by the Nursing Research Project of Shanghai Jiao Tong University School of Medicine (No.Jyhz2312) and the Nursing Talent Project of Xinhua Hospital, Shanghai Jiao Tong University School of Medicine (No.Xhlhb009).

Abbreviations

HD, hemodialysis; KF, Kidney failure; KT, kidney transplantation; mHealth, mobile health; PD, peritoneal dialysis.

Data Sharing Statement

All raw data and code are available from the corresponding author upon reasonable request.

Ethics Statement

Ethical approval was granted bythe Ethics Committee of Xinhua Hospital, Shanghai Jiao Tong University School of Medicine (Approval No. XHEC-D-2025-076).

Consent for Publication

Consent for publication was obtained from the participants.

Disclosure

The authors have no conflicts of interest.