Application of user profiling in mental health management: a scoping review

Abstract

Background

Mental health disorders affect 970 million people globally, yet treatment access remains limited by stigma and resource constraints. User profiling technology leverages AI and multi-source data to enable personalized mental healthcare.

Objective

This scoping review synthesizes evidence on user profiling applications in patient mental health management to inform evidence-based practice.

Methods

Following established scoping review methodology, we systematically searched seven databases (Wanfang, CNKI, PubMed, Embase, Web of Science, Scopus, Cochrane Library) from 1 January 1964 to 28 May 2025. Two researchers independently screened records against eligibility criteria and extracted data.

Results

Eighteen studies were included. User profiling construction encompassed three core phases: multidimensional data acquisition, feature extraction, and visualization. Applications spanned diverse populations including adolescents, young adults, and psychiatric patients, facilitating personalized intervention planning, resource allocation, and crisis management. Implementation demonstrated enhanced service accessibility, intervention precision, user experience optimization, and remote intervention efficacy.

Conclusion

User profiling technology delivers personalised mental health services through the integration of multidimensional data. This approach demonstrates significant advantages in enhancing intervention effectiveness, reducing healthcare costs, and improving user satisfaction, positioning it as a promising technology within the field of digital psychological interventions. However, its practical application still faces numerous challenges, including the long-term validation of current intervention outcomes and limitations in technological universality.

Introduction

Mental health disorders represent complex conditions characterized by behavioral or psychological patterns that cause significant distress or functional impairment [1]. These encompass anxiety, depression, substance use disorders, schizophrenia, eating disorders, bipolar disorder, and post-traumatic stress disorder (PTSD). These disorders present a growing global health challenge, with the World Health Organization (WHO) reporting that approximately 970 million individuals worldwide (one in eight people) experienced mental disorders in 2019, predominantly anxiety and depression [2]. Within the European Union (EU) alone, over one-third of the population contends with mental health issues annually [3].

Persistent demand for mental health services contrasts with insufficient intervention capacity. Pervasive social stigma frequently renders mental health conditions socially sensitive topics, creating fundamental barriers to help-seeking [4]. Additional obstacles identified in prior research include self-reliant coping tendencies, exclusive dependence on informal support networks, illness-related stigma, and structural barriers encompassing financial constraints, time limitations, transportation challenges, and restricted appointment availability [5].

Addressing these challenges requires expanded treatment modalities and scaled service delivery. Conventional psychological interventions face significant limitations including non-personalized approaches yielding heterogeneous outcomes, inequitable resource allocation restricting accessibility, static assessment protocols inadequate for capturing dynamic patient states, and suboptimal adherence compromising effectiveness [6, 7].

User profiling technology emerges as a promising solution, involving multi-source data integration (behavioral logs, physiological indicators, psychological assessments, sociodemographic characteristics, environmental parameters) to construct dynamic computational models characterizing individual psychological states, risk profiles, and needs/preferences [8]. Leveraging big data analytics and artificial intelligence (AI), this approach enables comprehensive analysis to deliver precise, personalized services, with expanding applications in mental health contexts [9, 10].

Internationally implemented user profiling-driven interventions demonstrate utility across treatment, training, and screening scenarios—particularly for anxiety, depression, postpartum depression, and psychotic disorders [10–12]. Synthesizing multidimensional patient data (demographic characteristics, medical records, mental health indicators, lifestyle factors) facilitates identification of individual needs and risks, enabling formulation of personalized interventions to enhance clinical outcomes. However, due to the strong subjectivity (reliance on self-report scales and interviews), high sensitivity (involving privacy and stigma), and multimodal nature (clinical records, wearable physiological metrics, online behavioral logs) of mental health data, constructing and applying user personas presents unique challenges [8, 13, 14].

The current literature on user profiling in mental health is characterized by a focus on primary development, validation, and evaluation, yet it lacks a scoping review that is critically needed to synthesize and map this highly heterogeneous body of work. Unlike a systematic review aimed at evaluating specific interventions, the scoping review methodology is ideally suited to map the heterogeneous landscape of this emerging field, clarify key concepts, and identify knowledge gaps, thereby providing a foundational overview to guide future research [15, 16]. Consequently, this study employs a scoping review methodology to systematically examine user profiling applications in psychological interventions. The research aims to elucidate optimal implementation strategies for patient care, thereby informing development of evidence-based, personalized interventions by clinical practitioners.

Methods

Materials and method

Research question formulation

Through preliminary literature review and team discussions, the following specific review questions were defined: ① How are user profiles constructed? ② To which populations have user profiles been applied? ③ What are the specific application modalities of user profiling in psychological interventions? ④ What are the primary focuses of psychological interventions based on user profile? ⑤ How effective is psychological management based on user profiling (e.g., user engagement and psychological intervention outcomes)?

Search strategy

A systematic search was conducted across the following databases: Wanfang Database, China National Knowledge Infrastructure (CNKI), VIP Database, SinoMed, PubMed, Embase, Scopus, Web of Science, CINAHL, and Cochrane Library. The search combined subject headings and free-text terms, covering records from 1 January 1964 to 28 May 2025 (with 1964 marking the creation of the Web of Science). Search terms: (“user profile” OR “user profiling” OR “health profiling” OR “profiling technology” OR “profile technology” OR “profile research” OR “profile construction” OR “user needs” OR “health profile” OR “health tag”)AND(“mental health” OR “mental health recovery” OR “psychiatric nursing” OR “psychiatric rehabilitation” OR “self psychology” OR “psychology*” OR “anxiety” OR “depression”).Reference lists of relevant publications were manually screened to ensure comprehensive coverage. The PubMed search strategy exemplifies the approach: ((“mental health“[Title/Abstract] OR “mental health“[MeSH Terms] OR “mental health recovery“[MeSH Terms] OR “psychiatric nursing“[MeSH Terms] OR “psychiatric rehabilitation“[MeSH Terms] OR “psychology“[MeSH Terms] OR “self psychology“[MeSH Terms] OR “psychology*“[Title/Abstract] OR “anxiety“[MeSH Terms] OR “depression“[MeSH Terms]) AND (“user profile“[Title/Abstract] OR “user profiling“[Title/Abstract] OR “health profiling“[Title/Abstract] OR “profiling technology“[Title/Abstract] OR “profile technology“[Title/Abstract] OR “profile research“[Title/Abstract] OR “profile construction“[Title/Abstract] OR “user needs“[Title/Abstract] OR “health profile“[Title/Abstract] OR “health tag“[Title/Abstract])

Literature inclusion and exclusion criteria

Inclusion criteria: (a) Focus on user profiling applications in psychological interventions; (b) Original research studies; (c) Chinese or English language publications.Exclusion criteria: (a) Unavailable full text; (b) Incomplete or ambiguous data; (c) Duplicate publications; (d) Non-peer-reviewed conference materials (abstracts, posters). This study operationally defines “user profiling” as: the segmentation of user groups based on one or more data dimensions, achieved through qualitative induction or quantitative analysis methods, resulting in a set of distinctive, generalisable characteristics [8].

Literature screening and data extraction

Based on the inclusion and exclusion criteria, an initial selection based on a scan of the titles and abstracts was made independently by three reviewers. The full texts of the retained articles were then screened by two reviewers to ascertain their eligibility. Any discrepancies in the selection of studies were resolved through a process of consensus and discussion. Extracted data elements included: publication year, country, study design, sample size, target population, profiling methodology, application format, intervention focus, and outcomes.

Results

Characteristics of included studies

The initial search yielded 9,885 records. After applying the inclusion and exclusion criteria, 18 studies were retained for analysis [8–14, 17–27], comprising 17 English-language publications and one publication in Chinese. Based on the PRISMA Statement and flowchart, four phases of review were conducted: identification, screening, eligibility assessment and final synthesis (see Fig. 1). Based on research focus, studies were categorized as: development studies (n = 4) focusing on profiling framework construction [13, 14, 23, 27], applied research with usability evaluation (n = 6) [8–11, 18, 26], applied intervention studies (n = 8) [12, 19–22, 24, 25]. Key characteristics of included studies are summarized in Tables 1 and 2.

Fig. 1.

PRISMA flowchart describing the study identification and selection

Table 1.

Characteristics of User Profiles in the Included Digital Mental Health Intervention Studies (n=18)

Author	Country	Target Population	User Profile Type
Ospina-Pinillos L 2025[8]	Colombia	Adolescents	High emotional distress, Moderate emotional distress, Mild emotional distress
Meng Guangting 2025[17]	China	Undergraduate students	Type A (Normal), Type B (Track), Type C（Alert)
Martínez-Vispo C 2024[18]	Spain	Adults without anxiety disorder diagnosis	low-risk, moderate-to-high-risk
Yoo I-J 2024[19]	South Korea, USA	Older adults aged ≥70 years	helpers, friends, short-term users, and long-term users
Giulia Rocchi 2024[9]	Italy, Switzerland	Young adults	Cluster 1 (high notification frequency+active texting+ insomnia+low mood), Cluster 2(social media dependency+attention/energy issues), Cluster 3 (workers+elderly), Cluster 4 (students+ obsessive thoughts+ physical pain)
Franco P 2024[10]	Chile	Women in the postpartum period (0-6 months)	Individual user profiles
M.D. Romael Haque 2023[20]	USA	Individuals aged ≥12 years	Seeking immediate support, alternative social interaction, therapeutic assistance
O'Sullivan S 2024[12]	Australia	Young adults in mental illness recovery	low use, maintained use of social components,maintained use of both therapy and social components
Huberty J 2022[13]	USA	Cancer patients	proactive integrators, cautious avoiders, and professionally guided users
Rachel Kornfield 2022[11]	USA	Young adults with depression/anxiety	Exploratory, Self-Regulating, Feedback-Driven
Judith Borghouts 2022[21]	USA	Deaf and hard-of-hearing individuals	Language diversity needs, privacy-sensitive users, resource-constrained users
Andreas Balaskas 2021[22]	Ireland, USA	Individuals ≥3 years old	Self-monitoring dependents, structured course seekers, and emergency support seekers
Buck B 2021[23]	USA	Caregivers of young adults with early-stage mental illness	Information-driven, altruistic, authority-trusting, and efficient communication-preferring
Stephen M Schueller 2021[24]	USA	Current or former users of mood tracking apps	Stress-driven, Self-reflective, Social support-seeking
Matthew Louis 2021[25]	USA	Healthy adults experiencing daily stress	Low-complexity stressors, high-complexity stressors
Emily Widnall 2020[26]	UK	Adolescents <18 years old	Pragmatists, Personalization Seekers, Tech-Savvy Users
Theresa Fleming 2019 [27]	New Zealand	Young adults	Players, Participants, Skeptics, and Outspoken
Stephen M Schueller 2018[14]	USA	Smartphone users with mental health needs	Socially dependent, self-exploratory, price-sensitive

Table 2.

Effectiveness, Core Features, and Platform Types of the Included Digital Mental Health Interventions(n=18)

Author	Platform Type	Core Features & Interventions	Effectiveness evaluation
Ospina-Pinillos L 2025[8]	Responsive Website	①Mental health education library;②Automated screening (Kessler Scale);③ Remote counseling & crisis support;④ Customized resource delivery	Practicality, participation, detection rate of psychological distress, and usage of various functions
Meng Guangting 2025[17]	Service Package	①Routine education (courses & activities);②Four-tier intervention network (school-dormitory);③ Crisis response with family collaboration	Image type recognition accuracy and early warning effect
Martínez-Vispo C 2024[18]	Web/Mobile App	①Mental health screening & dynamic tracking (13 factors);②Customized education library & health plans;③Remote counseling & emergency service links	Practicality, participation, and content validity
Yoo I-J 2024[19]	Robot	①Affective interaction (touch/handshake);②Assistive activities (exercise, puzzles, reminders)	Participation, practicality, satisfaction
Giulia Rocchi 2024[9]	Chatbot	①Personalized push notifications;②Recommended exercises (Mindfulness/CBT);③ Mental health manuals & crisis support	Subjective well-being, life satisfaction,practicality, and participation,
Franco P 2024[10]	Web Platform	①Depression & anxiety psychoeducation;②CBT techniques (cognitive restructuring, behavioral activation)	Practicality, participation, and content validity
M.D. Romael Haque 2023[20]	Mobile App	①CBT & mindfulness exercises;②Emotion tracking & personalized conversations;③Crisis resource links	Participation, Practicality, Satisfaction
O'Sullivan S 2024[12]	Online Platform	①Online therapy & clinical support;②Peer-to-peer social networking & interactions	Social function, Psychotic symptoms, Depression/ anxiety level
Huberty J 2022[13]	Framework (Prototype)	①Content: Cancer journey scenarios, symptom management;②Functions: Emergency toolkit, curated community, recommendation algorithm	Participation, Acceptance, and Practicality
Rachel Kornfield 2022[11]	SMS Text System	①Multi-strategy content (CBT, ACT, Positive Psychology);②Two-way interaction & on-demand support;③Context-aware personalized messaging Customized resource delivery	Participation, Help-seeking behavior, Depression /anxiety level
Judith Borghouts 2022[21]	Mobile App	①Accessibility: Language support, short videos;②Support: Peer chat, remote counseling, forums, blogs;③Crisis: Suicide prevention hotlines	Participation
Andreas Balaskas 2021[22]	Mobile App	①Psychoeducation & relaxation training;②Cognitive & behavioral techniques in interactive formats	Practicality, content validity
Buck B 2021[23]	mHealth Framework	①Core: Symptom monitoring, info database, communication training;②Support: Caregiver community, resource navigation	/（In development）
Stephen M Schueller 2021[24]	Mobile Platform	Recommendations: Dynamic multi-entry tracking, data interpretation, simplified input, trend graphs over aggregates	/（In development）
Matthew Louis 2021[25]	Chatbot	①CBT (e.g., problem-solving, worst-case analysis);②Positive psychology (e.g., reframing) & distraction techniques;③Daily impact assessment	Depression/ anxiety level, stress, sleep, self-rated social interaction, acceptance,participation
Emily Widnall 2020[26]	Mobile App	①Customizable emotion tags & free-text entry;②Visualized emotion trends & reminder modules	Participation, satisfaction
Theresa Fleming 2019 [27]	No intervention implemented, but recommendations provided	Recommendations: Multi-module app (meditation, info, interaction), social media integration for crisis keywords	/（In development）
Stephen M Schueller 2018[14]	No intervention implemented, but recommendations provided	Recommendations: User-friendliness, strong privacy protection, low cost	/（In development）

User profiling construction process

User profiling development encompasses data acquisition, feature extraction, and visualization. Data collection primarily utilized surveys [8, 12–14, 23, 25], interviews [8, 10, 11, 13, 14, 18, 21, 24–27], and big data platforms [8, 9, 12, 14, 17, 19, 20, 22, 24, 26]. Feature extraction addressed three domains: (1) Demographic characteristics (age, gender) [8–14, 18–25, 27]; (2) Socioeconomic attributes (occupation, income) [8, 9, 18, 19]; and (3) Mental health dimensions including emotional states (anxiety, depression [8, 10–14, 17–20, 22, 23, 25]), behavioral manifestations [9, 10, 17], sleep patterns [9, 20], psychiatric symptoms [9, 12], social relationships [8, 18], health status [18, 19], lifestyle factors [18, 19], psychological needs [13, 27], digital therapy attitudes [11, 25, 27], technology usage patterns [9, 13, 14, 20–23, 25], and intervention preferences [11, 13, 14, 20–22, 24, 26, 27]. Machine learning algorithms predominated feature extraction [9, 11, 13, 14, 17, 19–21], complemented by discursive methods [10, 12, 22, 23, 27] and thematic analysis [8, 18, 24, 25]. Visualization techniques included word clouds [8, 17], tag lists [10, 11, 13, 14, 21–24, 26], temporal charts [12, 19], SHAP plots [9], classification matrices [27], and end-user diagrams [18, 25].

Target populations

Profiling applications spanned diverse clinical populations: Four studies evaluated efficacy in anxiety/depression management [11, 12, 14, 23]; One focused on perinatal mental health [10]; Multiple investigations addressed adolescent interventions [8, 9, 17, 23, 26]; And one study developed cancer-specific digital tools through needs-based profiling [13].

Implementation modalities

Implementation featured group-level profiling for categorical interventions (behavioral patterns [11, 24], clinical characteristics [8, 18], needs-resource alignment [21]), supplemented by individualized approaches in one study [10]. Adaptive optimization mechanisms included: Schueller’s data-driven iteration using engagement metrics [14]; Rocchi’s extensible clustering frameworks [9]; and Meng’s lifecycle-sensitive prioritization for student populations [17]. Digital delivery platforms comprised responsive web/mobile applications [8, 10, 12, 18, 20–22, 24, 26], chatbots [9, 19, 25], and SMS systems [11].

Intervention applications

Core applications included: (1) Automated screening with tailored resource allocation (8 studies [8–12, 17, 18, 20]), exemplified by Ospina’s adolescent programs [8] and Giulia’s chatbot-integrated CBT/mindfulness delivery [9]; (2) Multimodal support including peer networks, teletherapy, and psychoeducation videos (4 studies [13, 21, 25, 27]); (3) Needs-based functional optimization via comment analysis (3 studies [14, 22, 26]); and (4) Dynamic profiling enabling real-time adaptation to preference shifts (2 studies [17, 18]).

Intervention efficacy

This study included 18 research projects, primarily evaluating user experience and the efficacy of psychological interventions. Firstly, regarding user experience, 11 studies focused on assessments of user engagement, satisfaction, practicality, and content effectiveness [8–11, 13, 18–21, 25, 26]. Three studies [8, 20, 26] revealed low sustained engagement, with next-day attrition rates reaching up to 81.2%. Conversely, the remaining eight studies [9–11, 13, 18, 19, 21, 25] demonstrated that user profiling technology significantly enhanced both acceptance and sustained engagement with digital psychological interventions. By precisely matching user needs and preferences, customised services not only improved service quality but also effectively enhanced user retention. Three studies [9, 20, 26] indicated that user profiling technology significantly optimised the user experience, making it more aligned with individual needs and habits, thereby increasing user satisfaction. Features such as emotion tracking, psychological education, and automated screening have gained user recognition, particularly among adolescents and special needs groups (e.g., hearing-impaired individuals). Furthermore, regarding intervention effectiveness evaluation, three studies [9, 12, 25] confirmed that providing precise digital psychological resources (e.g., customised CBT modules) based on user profiling effectively improves psychological symptoms (e.g., reducing anxiety and depression levels) and social functioning. Six studies [8, 17–21] designed timely crisis intervention modules for high-risk populations, all indicating low utilisation rates for crisis intervention. Furthermore, evidence suggests that user profiling enhances the detection rate of psychological issues [8, 24]. Finally, regarding theoretical framework development, four studies [14, 23, 24, 27] did not construct specific remote management platforms. However, based on in-depth profiling analyses, they proposed targeted remote management frameworks or prototype designs for cancer patients, young people experiencing psychological distress, and carers of individuals with mental disorders, respectively.

Discussion

Principal findings

This review analyses 18 studies, systematically examining the current application and developmental trends of user profiling in mental health management. Findings indicate that user profiling has been extensively applied across diverse mental health intervention scenarios, including stress management [24], emotional monitoring [26], anxiety intervention for cancer patients [13], postnatal depression intervention [10], and adolescent psychological counselling [8]. These studies predominantly employ multi-dimensional data to construct user profiles, encompassing behavioural patterns, psychological characteristics, usage preferences, and socio-demographic information [13]. Based on these profiles, multiple functionalities have been realised, including personalised content delivery, intervention pathway recommendations, and risk alerts [9].

Previous reviews indicate [28] that while AI-based technologies (such as chatbots and predictive modelling) can enhance early identification and intervention for mental health issues and improve patient engagement, their interventions often lack specificity and personalisation. In contrast, this review finds that user profiling methodologies exhibit considerable diversity, spanning traditional approaches like questionnaires and clinical assessments to emerging techniques such as digital behavioural analysis, natural language processing, and machine learning clustering [11, 12]. Some studies further integrate user profiling with recommendation systems, conversational bots, and telemedicine platforms to establish closed-loop personalised service systems [19, 26]. For instance, by segmenting users based on usage patterns (e.g., high/low frequency users, socially dependent types, self-regulatory types), differentiated intervention content can be delivered [25]. This ‘profiling + recommendation’ model not only enhances service precision but also creates conditions for achieving sustainable user engagement.

However, most current profiling models still rely on static or single-source data, lacking the capacity to capture dynamic changes in user states [17]. Future research should focus on exploring dynamic profiling construction methods that integrate multimodal data (such as physiological signals, speech sentiment analysis, and daily behavioural trajectories), while incorporating artificial intelligence technologies to enable real-time profile updates and automatic optimisation of intervention strategies [22].

Studies included in this review consistently report positive effects of personalised interventions based on user profiling in enhancing engagement, satisfaction, and perceived usefulness. Research indicates that profile-driven push notifications and content recommendations contribute to sustained long-term user participation [20]. However, Ospina-Pinillos L [8] found that next-day engagement rates were merely 18.8%, suggesting challenges in maintaining initial engagement momentum. Huberty J [13] further emphasised that fostering sustained long-term system usage requires deeper exploration. Evidence suggests that strategies such as delivering more compelling notifications to ‘low-usage’ users, enhancing gamification elements, strengthening social features, integrating offline services, or activating community functions for ‘socially dependent’ users may effectively reduce churn rates [8].

Regarding improvements in psychological well-being, only three studies included in this review demonstrated positive effects from personalised interventions based on user personas. Giulia Rocchi [9] identified four distinct user personas with varying behavioural and psychological characteristics through clustering. By implementing personalised recommendations and interventions, the experimental group’s subjective well-being scores rose significantly from a baseline average of 3.06 to 4.1 at the conclusion of the intervention. O’Sullivan S [12] constructed user personas for young adults in psychiatric recovery, delivering targeted online therapy, peer-to-peer social networking, and personalised clinical support via a web platform. Results indicated significant improvements in social functioning and overall severity of psychiatric symptoms within the ‘Maintenance Therapy and Social Support’ group. Matthew Louis [25] similarly observed more pronounced reductions in depressive symptoms (assessed via PHQ-4) among high-frequency users. These findings preliminarily support the practical value of user profiling in mental health interventions.However, existing studies predominantly focus on developing user profiling-based remote psychological management platforms, while systematic validation of psychological intervention efficacy and in-depth clinical efficacy analysis remain insufficient. Therefore, rigorous research designs are essential to further validate the practical effectiveness of such interventions in preventing and improving psychological well-being. Whilst a few studies have tentatively explored intervention efficacy, there remains a general lack of long-term follow-up assessments tracking mental health changes post-use. Future research necessitates extended follow-up studies to evaluate the sustainability of intervention effects.

Furthermore, six studies in this review [8, 17–21] incorporated timely crisis intervention modules for high-risk populations, such as providing suicide prevention helplines [21]. However, one study reported a low completion rate for remote consultations, at merely 26.5% [8]. Potential reasons include systems failing to accurately and promptly identify crisis signals in user language, alongside user concerns regarding data usage and crisis management protocols [8, 19]. Future efforts should focus on developing more intelligent algorithms to enhance crisis signal recognition and matching accuracy, while clarifying data usage regulations and crisis intervention procedures to improve system usability and user trust [20].

Despite these challenges, we remain optimistic about the potential of user-profile-based remote management models to improve mental health. With the increasing prevalence of digital health technologies and ongoing advancements in artificial intelligence, models integrating evidence-based psychotherapy, digital content, and personalised management strategies hold considerable promise. We believe that accumulating evidence will further validate the efficacy of user profiling techniques in mental health interventions, thereby providing more cost-effective and accessible healthcare services for individuals with mental health needs.

Future research direction

Based on a review of existing research, future breakthroughs in this field may be pursued across four key directions. Firstly, the precision and dynamism of profiling models must be enhanced. This involves incorporating multimodal data and constructing temporal models capable of capturing real-time shifts in user states [29], thereby overcoming current limitations such as insufficient sample representativeness and static modelling approaches. Secondly, the efficacy validation and personalisation mechanisms of interventions should be strengthened. This entails rigorously evaluating clinical outcomes through randomised controlled trials and long-term follow-up assessments, alongside developing profile-based adaptive recommendation algorithms to achieve precise matching of intervention content with delivery timing. Thirdly, attention must be directed towards the needs of special populations and cross-cultural adaptation. It is imperative to develop inclusive profiling models and intervention protocols specifically tailored for groups such as those with hearing impairments or distinct cultural backgrounds, thereby ensuring the universality and equity of the technology [29–31]. Future designs should incorporate user-friendly, portable devices [11] and develop symptom-adaptive interfaces (e.g., simplified operations during manic episodes, enhanced guidance during depressive phases) to enhance the accessibility of mental health services. Finally, exploring viable clinical integration pathways and ethical frameworks requires research into effectively embedding user profiling technologies within real-world healthcare systems and service workflows. Proactive development of governance frameworks encompassing data privacy protection, algorithmic fairness, and crisis response mechanisms is essential [32, 33] to ensure responsible and sustainable technological advancement.

Limitations

This study presents several limitations. Firstly, the majority of included literature originates from developed nations. In low- and middle-income countries, where digital healthcare infrastructure remains underdeveloped and user digital literacy and cultural backgrounds vary considerably, constructing and disseminating user profiles may encounter applicability challenges. Secondly, as this is a scoping review, it did not conduct systematic quality assessments or risk of bias evaluations of the included studies, nor did it perform publication bias analyses. Consequently, caution is warranted when interpreting these findings, as their inferential validity is constrained by the methodological limitations inherent to scoping reviews. Thirdly, this review was not prospectively registered on international systematic review registries such as PROSPERO, potentially introducing selection and reporting bias risks. Finally, the included studies encompass both experimental and observational designs, presenting methodological heterogeneity—particularly concerning causal inference limitations. Interpreting findings requires careful consideration of the impact of differing research methodologies.

Conclusion

This review highlights how user profiling technology can be applied to mental health management. As an emerging health information integration technique, user profiling synthesises multidimensional health data for analysis, delivering personalised, precise and dynamic management plans for mental health service users. It demonstrates significant value in predicting psychological risks, enabling early prevention and therapeutic interventions, while also enhancing service accessibility, reducing healthcare costs, improving treatment engagement and boosting patient satisfaction. Nevertheless, practical implementation faces multiple challenges, including insufficient representativeness of current research samples, limitations of static modelling, the need for long-term validation of intervention efficacy, constraints on technological universality (particularly for groups with low digital literacy), and risks to user data privacy and security (such as leakage of sensitive information). Future research must conduct in-depth exploration of these critical bottlenecks.

Authors’ contributions

CangMei Fu and Ya Hu wrote the original manuscript draft. XiTing Huang and Jie Zhou contributed to the conception of the study and the protocol. CangMei Fu and Ya Hu planned the initial search strategy and conducted a pilot study to test the protocol. CangMei Fu and Ya Hu developed the data extraction form. All authors, including CangMei Fu, Yi Zhang, XiTing Huang, Jie Zhou, and Ya Hu, contributed to the scoping review protocol, screening methodology, data analysis, and interpretation, and have read and approve the final version of this manuscript.

Funding

This study did not receive any funding.

Data availability

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.