CoChat
Sign Up
Cited by 10Artificial Intelligence in Healthcare and EducationMobile Health and mHealth ApplicationsAI in Service Interactions

The Usability and Experience of Artificial Intelligence‐Based Conversational Agents in Health Education for Cancer Patients: A Scoping Review

Open access
Li, Kun, Wu, Siyuan, Zhang, Ying, Zhu, Binbin, Qi, Ziyi, Hou, Sijia, Wang, Wei·2026·Journal of Clinical Nursing
DOI 10.1111/jocn.70020PMC PMC13569306
Open this paper in CoChat — free accountOpen PDF

Abstract

Background: Artificial intelligence‐based conversational agents (CAs) have shown transformative potential in healthcare, yet their application in cancer health education has remained underexplored, particularly regarding usability and patients' experiences. Existing reviews lack a dedicated focus on user perspectives, limiting insights into how CAs can be optimised for patient needs. Aim: To explore the usability and experience of artificial intelligence‐based conversational agents in health education for cancer from the user perspective. Design: A scoping review was conducted with the Joanna Briggs Institute Scoping Reviews conduct guidance and reported according to the preferred reporting items for systematic reviews and meta‐analyses extension for scoping reviews checklist. Methods: A search was performed in PubMed, Embase, CINAHL, Web of Science, PsycINFO, IEEE Xplore Digital Library and ACM Digital Library from their inception to March 6, 2024. The references to the articles included were also searched. The Pillar Integration Process was employed to chart data. Results: A total of 12 studies were included in this scoping review, which revealed that CAs supported diverse educational contexts, including cancer‐related knowledge (41.7%), pretest genetics (33.3%), self‐management (16.7%) and psychological skills (8.3%). Three studies reported that patients preferred interactions with multiple options or ‘read more’ functions. Patients were generally optimistic about the CAs and reported that CAs provided informational, physical, and psychological support for them. However, limitations such as insufficient customisation, lack of empathy, and defects in understanding free‐input questions were noted. Conclusion: This review demonstrated that CAs are promising complementary tools in cancer education, alleviating healthcare burdens while enhancing patient engagement, which was particularly critical in resource‐limited settings. However, clinical implementation requires more rigorous validation of safety protocols and high‐quality original studies. Relevance to Clinical Practice: Nurses and policymakers should consider CAs valuable tools to enhance cancer health education, provided that they align with patient needs and institutional safety standards.

AI snapshot

Written from the paper’s full text.

TL;DR

This scoping review of 12 studies found that cancer patients generally viewed AI-based conversational agents positively for health education, valuing usability, informational/physical/psychological support, and 24/7 access. However, evidence remains limited by heterogeneity, geographic concentration, and concerns about customisation, empathy, free-input understanding, and safety.

Key findings

  • Across 12 included studies and 1587 participants (1516 patients with cancer, 71 at risk), CAs addressed cancer-related knowledge (41.7%), pretest genetics (33.3%), self-management (16.7%), and psychological skills (8.3%).
  • Patients were generally positive: five studies reported satisfaction with CA information as valuable, informative, and reliable; five found few interaction barriers; over 90% in one study agreed ChemoFreeBot was easy to use, and over three-quarters of a median-age-68 group needed no assistance.
  • Three studies reported preferences for multiple options or “read more” functions; one study found smartphone preference, another computers/laptops, and lifestyle and breast biopsy side effects were the most reviewed topics among six offered.
  • CAs provided informational, physical, and psychological/emotional support; one RCT showed alleviated physical and psychological symptoms after chemotherapy and promoted self-care, but limitations included insufficient customisation, lack of empathy, and unstable understanding of free-input questions.

Methods

  • Scoping review following Joanna Briggs Institute Scoping Reviews conduct guidance and reported with PRISMA-ScR; searched PubMed, Embase, CINAHL, Web of Science, PsycINFO, IEEE Xplore Digital Library, and ACM Digital Library from inception to 6 March 2024, plus reference searching.
  • Inclusion criteria covered patients diagnosed with cancer or at high risk, CAs applied in health education, and primary studies such as RCTs, pilot studies, and qualitative investigations; no publication date or language restrictions. Exclusions included noncancer/healthy individuals, non-education CAs, development-only studies, protocols/reviews/theses/editorials/comments/conference abstracts, and unavailable full texts.
  • Data were charted in Excel, and the Pillar Integration Process (listing, matching, checking, pillar building) was used to integrate qualitative and quantitative findings; 12 studies were included.

Limitations

  • Stated: heterogeneity in study designs, populations, and educational contexts limited synthesis and generalisability; the review did not evaluate study quality; potential publication bias may have skewed findings toward favourable perceptions; insufficient qualitative studies and a small number of studies limit findings; scoping review only maps evidence and does not determine utility.
  • Stated: 11 of 12 studies originated from developed countries (USA 6, France 3, Germany 1, Norway 1; one from Egypt), limiting global generalisability; approximately 75% of studies were conducted on breast cancer patients (table reports 7/12, 58.3%), and participant ages spanned 19–90 years, which may influence feedback.
  • Evident: only one RCT demonstrated symptom and self-care benefits; safety, severity identification, and long-term outcomes remain inadequately evaluated.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1111/jocn.70020

Cites

67 of 67 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

55 of 90 cited sources checked · 29 supported · 0 flagged · accuracy not yet measured

  • Findingcitation

    Patients reported that CAs had an unstable understanding of self-written questions, failing to accurately provide the information they expected.

    “During interaction with CAs, patients reported that CAs had an unstable understanding of self-written questions, failing to accurately provide the information they expected (Siglen et al. 2023).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Siglen et al. 2023 source

      The Results (Emerging Concepts; page not available) say several interviewees reported that Rosa inconsistently understood self-composed questions and switched to suggested readings. The Discussion (Principal Findings; page not available) reports a high fallback rate for open-ended questions, but does not establish that the information Rosa provided was inaccurate or consistently failed to meet patients’ expectations. (LLM, conf=0.82)

      Experimental — accuracy not yet measured
  • Findingcitation

    Patients preferred interactions with multiple options or 'read more' functions.

    “Three studies reported that patients preferred interactions with multiple options or 'read more' functions.”
    • Also stated in: § 3.3.1 | Attitude Towards CAs · plain
      “In terms of interaction style, Chaix et al. found that patients preferred CAs with multiple options, making it easy to obtain relevant information by clicking buttons rather than typing responses (Chaix et al. 2019).”
    • Also stated in: § 4.1.1 | User-Friendly Tools for Patients · plain
      “Interestingly, the findings identified that multiple options and the function of 'read more' were the preferred interactions of patients (Chaix et al. 2019;Siglen et al. 2023).”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites Chaix et al. 2019 source
    • cites Siglen et al. 2023 source
  • Findingcitation

    Patients were not completely satisfied with CAs due to lacking customisation.

    “limitations such as insufficient customisation, lack of empathy, and defects in understanding free-input questions were noted.”
    • Also stated in: § 3.3.3 | Limitations of CAs · plain
      “Two studies found that patients are not completely satisfied with CAs due to lacking customisation (Goertz et al. 2023;Qiu et al. 2021;Siglen et al. 2023).”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites Goertz et al. 2023 source

      The Results (Patient evaluation; page unavailable) report that patients wanted answers and therapy recommendations tailored to their situations, supporting the concern about insufficient customization. They do not establish that patients were dissatisfied because of it; the same section reports that all participants wanted to use a medical chatbot again. (LLM, conf=0.82)

      Experimental — accuracy not yet measured
    • cites Qiu et al. 2021 source
    • cites Siglen et al. 2023 source

      The Results (Emerging Concepts; page unavailable) report participants’ concerns about Rosa’s lack of empathy and its difficulty understanding self-composed questions. The Introduction (Background; page unavailable) says personalized communication was not included, but the paper does not clearly say patients were dissatisfied because of insufficient customization; the Discussion (Principal Findings) instead describes an overall positive attitude toward Rosa. (LLM, conf=0.83)

      Experimental — accuracy not yet measured
  • Findingcitation

    Five studies showed that most patients were satisfied with the information provided by CAs.

    “Five studies showed that most patients were satisfied with the information provided by CAs CAs were valuable, informative, and reliable (Bibault, Chaix, Guillemassé, et al. 2019;Chetlen et al. 2019;Greer et al. 2019;Siglen et al. 2023;Tawfik et al. 2023;Visvanathan et al. 2023).”
    Quote in the paperNot a substantial claim1 of 3 sources checked
    • cites Bibault, Chaix, Guillemassé, et al. 2019

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Chetlen et al. 2019 source
    • cites Greer et al. 2019 source

      In Results > Perceived Helpfulness and Open-Ended Feedback (page unavailable), participants rated Vivibot helpful on average and some praised its lessons. The paper does not report that most participants were satisfied with information it provided or that they considered it reliable; its participants were young people after cancer treatment, not patients generally. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
  • Premisereasoning

    These existing reviews did not explicitly focus on the domain of health education for patients.

    “However, these reviews did not explicitly focus on the domain of health education for patients.”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites Bibault, Chaix, Nectoux, et al. 2019 source
    • cites Wang et al. 2023 source
  • May combine two claims. 35
  • Classified as probably not a substantial claim. 23

  • May combine two claims.

    Findingcitation

    Patients thought CAs lacked empathy and a human touch, which limited their ability to handle special and complex issues.

    “Patients thought that CAs lacked empathy and a human touch, which limited their ability to handle special and complex issues (Siglen et al. 2023).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Siglen et al. 2023 source

  • May combine two claims.

    Premisereasoning

    Existing reviews lack a dedicated focus on user perspectives, limiting insights into how CAs can be optimised for patient needs.

    “Existing reviews lack a dedicated focus on user perspectives, limiting insights into how CAs can be optimised for patient needs.”
    • Also stated in: § 1 | Introduction · plain
      “However, these reviews did not explicitly focus on the domain of health education for patients.”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites Bibault, Chaix, Nectoux, et al. 2019 source

      The Introduction (page not provided) describes this review as an overview of conversational agents in oncology, rather than a review dedicated to user perspectives. The Conclusion identifies a lack of objective evidence about chatbot relevance and efficacy, but the paper does not establish that existing reviews generally lack a user-perspective focus or that this limits insights into optimising CAs for patient needs. (LLM, conf=0.75)

      Experimental — accuracy not yet measured
    • cites Wang et al. 2023 source

      The Results section (‘Evaluation of Chatbots’; page unavailable) examines user acceptability, and the Conclusion says more user testing is needed to optimise satisfaction. The ‘Strengths and Limitations of Review’ section notes limited and heterogeneous acceptability data, but does not establish that existing reviews lack a dedicated focus on user perspectives. (LLM, conf=0.78)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    Only two existing reviews investigated the use of CAs in oncology: one examined the landscape of CAs and the other assessed applications, limitations, and challenges of employing CAs in cancer care.

    “There are only two existing reviews investigating the use of CAs in oncology: one examined the landscape of the use of CAs (Bibault, Chaix, Nectoux, et al. 2019) and the other assessed the applications, limitations and challenges of employing CAs in cancer care (Wang et al. 2023).”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites Bibault, Chaix, Nectoux, et al. 2019 source
    • cites Wang et al. 2023 source

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    The Pillar Integration Process was employed to map and synthesise qualitative and quantitative findings.

    “The pillar integration process (PIP), a transparent and rigorous four-stage method consisting of listing, matching, checking and pillar building for integrating and presenting qualitative and quantitative findings jointly (Johnson et al. 2017), was employed to map and synthesise the extracted data visually to produce meaningful results.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Johnson et al. 2017 source

  • May combine two claims.

    Findingcitation

    CAs were valuable, informative, and reliable.

    “CAs were valuable, informative, and reliable (Bibault, Chaix, Guillemassé, et al. 2019;Chetlen et al. 2019;Greer et al. 2019;Siglen et al. 2023;Tawfik et al. 2023;Visvanathan et al. 2023).”
    Quote in the paperSubstantial claim1 of 3 sources checked
    • cites Bibault, Chaix, Guillemassé, et al. 2019

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Chetlen et al. 2019 source
    • cites Greer et al. 2019 source

      In Results > Perceived Helpfulness and Open-Ended Feedback (page not provided), participants rated Vivibot as helpful and described its lessons as useful, supporting the claims that it was valuable and informative. The paper does not establish that Vivibot was reliable. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingcitation

    CAs provided informational, physical, and psychological support for cancer patients, potentially enhancing their health across various dimensions.

    “reported that CAs provided informational, physical, and psychological support for them.”
    • Also stated in: § 3.3.2 | Support by CAs · hedged
      “Six studies (Chaix et al. 2019(Chaix et al. , 2020;;Chetlen et al. 2019;Qiu et al. 2021;Soley et al. 2023;Tawfik et al. 2023) confirmed that CAs potentially enhanced patients' health across various dimensions.”
    Quote in the paperSubstantial claim1 of 3 sources checked
    • cites Chaix et al. 2019 source

      The Introduction (Objectives; page unavailable) says Vik supplied quality-checked information on breast cancer, treatments, side effects, and quality-of-life topics. The Results (The Way Patients Used Vik; page unavailable) report that patients felt supported and helped with treatment, and the Discussion (General Appreciation; page unavailable) notes emotional attachment. The paper does not establish that Vik provided distinct physical and psychological support or improved health across those dimensions. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites Chetlen et al. 2019 source
    • cites Qiu et al. 2021 source

  • May combine two claims.

    Findingcitation

    Only one study mentioned user data security, and it reported that no one was concerned about personal data safety when using CAs.

    “Only one study mentioned user data security, and it reported that no one was concerned about personal data safety when using CAs (Goertz et al. 2023).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Goertz et al. 2023 source

  • May combine two claims.

    Findingcitation

    Tawfik et al. conducted an RCT proving that CAs could effectively alleviate both physical and psychological symptoms after chemotherapy and promote self-care behaviours.

    “Tawfik et al. conducted an RCT proving that CAs could effectively alleviate both physical and psychological symptoms after chemotherapy and promote self-care behaviours (Tawfik et al. 2023).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Tawfik et al. 2023 source

      The Methods describe a three-arm randomized controlled trial, and the Results report less frequent, less severe physical and psychological symptoms and more effective self-care behaviors in the ChemoFreeBot group than in the nurse-led education and routine-care groups. The claim overstates this as proof about conversational agents generally: the trial tested one chatbot in women receiving chemotherapy for breast cancer, and measured the effectiveness of self-care behaviors rather than simply whether those behaviors increased. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingcitation

    CAs were available 24/7 and integrated into daily routine, allowing users to get instant consultations via mobile devices anytime and anywhere.

    “Moreover, CAs were available 24/7 (Siglen et al. 2023) and integrated into daily routine (Qiu et al. 2021), allowing users to get instant consultations via their mobile devices anytime and anywhere.”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Siglen et al. 2023 source
    • cites Qiu et al. 2021 source

  • May combine two claims.

    Findingcitation

    After using CAs, many patients intended to re-use CAs in the future and supported applying CAs in clinical routine.

    “After using CAs, many patients intended to re-use CAs in the future and supported applying CAs in the clinical routine (Goertz et al. 2023).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Goertz et al. 2023 source

  • May combine two claims.

    Findingcitation

    One study found patients preferred using CAs on their smartphones, while another reported they frequently used computers or laptops to access CAs.

    “One study found that patients preferred using CAs on their smartphones (Chetlen et al. 2019), while the other one reported they frequently used computers or laptops to access CAs (Goertz et al. 2023).”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Chetlen et al. 2019 source
    • cites Goertz et al. 2023 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingcitation

    Siglen et al. found patients tended to use the 'read more' function to continue conversations with CAs.

    “Siglen et al. found patients tended to use the 'read more' function to continue conversations with CAs (Siglen et al. 2023).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Siglen et al. 2023 source

  • May combine two claims.

    Findingcitation

    Technical limitations identified by the review include a lack of customisation and weakness in natural language processing.

    “Technical limitations identified by this review, including a lack of customisation and weakness in natural language processing (Qiu et al. 2021;Siglen et al. 2023)”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites Qiu et al. 2021 source
    • cites Siglen et al. 2023 source

  • May combine two claims.

    Interpretationcitation

    With a global nursing shortage, CAs can take on some of the health education for cancer survivors and allow nurses and doctors to spend more time with patients who truly need a consultation.

    “In particular, with a global nursing shortage (Drennan and Ross 2019), CAs can take on some of the health education for the huge numbers of cancer survivors and allow nurses and doctors to spend more time with patients who truly need a consultation.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Drennan and Ross 2019 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingcitation

    A patient described the conversational agent as a friend who remained with her throughout her disease journey.

    “A patient described the CAs as a friend who remained with her throughout her disease journey (Siglen et al. 2023).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Siglen et al. 2023 source

      The Discussion’s Principal Findings say a participant described Rosa as “having a friend beside them, one who is there for the whole ride.” The paper places that description in the genetic testing process, not a journey through a diagnosed disease. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Interpretationcitation

    Unlike face-to-face interactions with nurses, conversational agents just answered questions and would not evaluate patients' true feelings, which may encourage patients to share more openly during interactions to ease anxiety.

    “Unlike face-to-face interactions with nurses, CAs just answered questions and would not evaluate patients' true feelings, which may encourage patients to share more openly during interactions to ease anxiety (Laymouna et al. 2024;Marita et al. 2021).”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Laymouna et al. 2024 source

      The Results section (‘Health Care Chatbot Benefits’; page unavailable) says friendly chatbot interactions could enhance self-disclosure, supporting the possibility that patients may share more openly. It does not compare chatbots with face-to-face nursing interactions, say that chatbots do not evaluate patients’ feelings, or link this disclosure to easing anxiety. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites Marita et al. 2021

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim.

    Findingown data

    Three Pillar themes were determined following PIP analysis: Attitude towards CAs, Support by CAs, and Limitations of CAs.

    “Three Pillar themes were determined following PIP analysis: Attitude towards CAs, Support by CAs, and Limitations of CAs (Figure 2).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    All conversations employed text output.

    “These conversations all employed text output”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingnone given

    Patients wanted to know more details about sexual health, which they felt uncomfortable asking nurses.

    “For example, they wanted to know more details about sexual health, which they felt uncomfortable asking nurses.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingnone given

    All included studies were published between 2019 and 2023, and half were pilot studies or evaluation surveys.

    “All included studies were published between 2019 and 2023, and half were pilot studies or evaluation surveys.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingcitation

    Most patients thought CAs helped them better understand their health conditions and treatment, enhancing their patient experience.

    “Most patients thought CAs helped them better understand their health conditions and treatment, enhancing their patient experience (Chetlen et al. 2019).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Chetlen et al. 2019 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingcitation

    Lifestyle and side effects of breast biopsy were the most frequently reviewed topics of 6 topics offered by CA.

    “On topics of interest, lifestyle and side effects of breast biopsy were the most frequently reviewed topics of 6 topics offered by CA (Chetlen et al. 2019).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Chetlen et al. 2019 source

  • May combine two claims.

    Findingcitation

    The visual design of CAs was user-friendly, allowing patients to navigate the interface with ease.

    “The visual design of CAs was user-friendly, allowing patients to navigate the interface with ease (Visvanathan et al. 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Visvanathan et al. 2023 source

  • Probably not a substantial claim: may not pass S4 (based).

    Findingnone given

    This is the first review focusing on the usability and patients' experience of CAs in health education for cancer.

    “This is the first review focusing on the usability and patients' experience of CAs in health education for cancer.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingnone given

    Half of the included studies were pilot studies or surveys, and only four were randomised controlled trials.

    “Half of the included studies were pilot studies or surveys, and only four were randomised controlled trials.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    It is unclear whether conversational agents can accurately identify the severity of patient issues, define appropriate severity criteria, or prevent potential adverse consequences.

    “It is unclear whether CAs can accurately identify the severity of patient issues, define appropriate severity criteria, or prevent potential adverse consequences.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingnone given

    RCTs demonstrated CAs' efficacy, while qualitative studies highlighted patients' subjective preferences but lacked generalisability.

    “Methodologically, RCTs demonstrated CAs' efficacy, while qualitative studies highlighted patients' subjective preferences but lacked generalisability.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The included studies exhibited significant heterogeneity in both methodology and demographic characteristics.

    “Notably, the included studies exhibited significant heterogeneity in both methodology and demographic characteristics.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Findingnone given

    The security of conversational agents is still uncertain.

    “The security of CAs is still uncertain; future research should focus on testing and updating the ability of CAs to identify serious health problems.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    Conversational agents functioned as digital companions, offering emotional support to cancer patients.

    “Beyond self-management, CAs also functioned as digital companions, offering emotional support to cancer patients.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingnone given

    Patients report that conversational agents relieve negative emotions and make them feel more comfortable.

    “relieve negative emotions and feel more comfortable.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    Participants' ages ranged from 19 to 90 years, and three studies recruited patients with an average or median age over 60.

    “Participants' ages ranged from 19 to 90 years, and three studies (Al-Hilli et al. 2023; Goertz et al. 2023; Visvanathan et al. 2023) recruited patients with an average or median age over 60.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingcitation

    Diverse health education contexts were observed, including cancer-related knowledge education (n = 5), pretest genetic education (n = 4), self-management education (n = 2), and positive psychological skills education (n = 1).

    “Diverse health education contexts were observed, with cancer-related knowledge education (n = 5) (Bibault, Chaix, Guillemassé, et al. 2019; Chaix et al. 2019, 2020; Chetlen et al. 2019; Goertz et al. 2023), pretest genetic education (n = 4)”
    • Also stated in: § 5 | Conclusion · plain
      “self-management education (n = 2) (Qiu et al. 2021; Tawfik et al. 2023), and positive psychological skills education (n = 1) (Greer et al. 2019).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    This scoping review involved a total of 1587 participants, comprising 1516 (95.5%) patients confirmed with cancer and 71 (4.5%) at risk of cancer but not yet diagnosed.

    “This scoping review involved a total of 1587 participants, comprising 1516 (95.5%) patients confirmed with cancer and 71 (4.5%) were at risk of cancer but not yet diagnosed.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Regular clinical consultation and effective health education are essential to optimise the outcomes of cancer patients by enhancing cancer knowledge, promoting self-management, providing psychological support, and ultimately improving quality of life.

    “Regular clinical consultation and effective health education are essential to optimise the outcomes of cancer patients by enhancing cancer knowledge, promoting self-management, providing psychological support, and ultimately improving quality of life (Chang et al. 2020).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Chang et al. 2020

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Discrepancies between the two reviewers were resolved through discussion with a third reviewer.

    “If discrepancies arose between the decisions of the two reviewers, a consensus was reached through discussion with the third reviewer (Y.Z.).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Data extraction items were pretested by two reviewers with an additional reviewer sampling two included studies before formal extraction.

    “Before formal data extraction, two reviewers (K.L. and S.W.) conducted a pretest, with an additional reviewer (Y.Z.) sampling two from the included studies.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The scoping review was reported according to the PRISMA-ScR checklist.

    “This scoping review was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) checklist (see Appendix S1).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    CAs are promising complementary tools in cancer education, alleviating healthcare burdens while enhancing patient engagement, particularly in resource-limited settings.

    “This review demonstrated that CAs are promising complementary tools in cancer education, alleviating healthcare burdens while enhancing patient engagement, which was particularly critical in resource-limited settings.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    The user's experience and requirements are integral to determining the successful development and improvement of digital products.

    “Although the design method is important to developing digital products, the user's experience and requirements are also integral to determining their successful development and improvement (Fu et al. 2022).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Fu et al. 2022

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Premisecitation

    Jiang et al. focused on the development and evaluation of embodied conversational agents for chronic diseases including cancer, and offered insights into the development and implementation of CAs from the design process perspective.

    “Jiang et al. focused on the development and evaluation of embodied conversational agents for chronic diseases including cancer, and offered technicians and healthcare professional's insights into the development and implementation of CAs from the design process perspective (Jiang et al. 2024).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Jiang et al. 2024

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    The retrieved studies were screened independently by two reviewers.

    “The studies retrieved were imported into Endnote X9 to eliminate duplicates and then screened independently by two reviewers (K.L. and B.Z.).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    AI-based CAs served as user-friendly and accessible tools for health education among cancer patients.

    “AI-based CAs served as user-friendly and accessible tools for health education among cancer patients.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning

    Potential publication bias should be acknowledged in this review.

    “Second, potential publication bias should be acknowledged.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Studies with positive outcomes or significant results are more likely to be published.

  • May combine two claims.

    Interpretationcitation

    Conversational agents empower healthcare providers to work more effectively and improve the experience of patients.

    “they empower healthcare providers to work more effectively and improve the experience of patients (Greene et al. 2019).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Greene et al. 2019 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Studies screening and data extraction were conducted independently after two reviewers reached a strong agreement to minimise selection bias.

    “Additionally, studies screening and data extraction were conducted independently after two reviewers reached a strong agreement to minimise selection bias.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductreasoning

    The review employed the Pillar Integration Process to integrate mixed-methods data, but differences in methodological rigour and focus may affect the comparability of results.

    “While we employed the PIP to integrate mixed-methods data, differences in methodological rigour and focus may affect the comparability of results.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given

    Especially in developing countries with limited medical resources, there may be a greater need to integrate conversational agents into primary care to provide healthcare support for cancer patients and share the educational burden of nurses.

    “Especially in developing countries with limited medical resources, there may be a greater need to integrate CAs into primary care to provide healthcare support for cancer patients and share the educational burden of nurses.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Developing countries have limited medical resources.

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    Developed countries have solid financial and technological support to develop CAs, potentially leading to the exacerbation of medical inequity.

    “The reason may be that developed countries have solid financial and technological support to develop CAs, potentially leading to the exacerbation of medical inequity.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationnone given

    Provided that patients possess a mobile device, they can access the system directly via a link or app.

    “Provided that patients possess a mobile device, they can access the system directly via a link or app.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    Compared to mixed information from broad internet sources, patients would feel comfortable and trustworthy with conversational agents who are admitted professional virtual educators.

    “Compared to the mixed information from broad sources on the internet, patients would feel comfortable and trustworthy with CAs who are admitted professional virtual educators.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Younger patients may exhibit higher digital literacy and thus be more likely to use digital tools and have positive experiences.

    “Younger patients may exhibit higher digital literacy and thus be more likely to use digital tools and have positive experiences (Paillaud et al. 2025).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Paillaud et al. 2025

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Interpretationreasoning

    This delivery relieves the need for additional expenditure on access devices or extensive time in learning how to use CAs.

    “This delivery relieves the need for additional expenditure on access devices or extensive time in learning how to use CAs.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Previous research showed that compared to typing on the keyboard, patients may feel more comfortable through voice interaction.

    “In terms of interactive modality, previous research showed that compared to typing on the keyboard, patients may feel more comfortable through voice interaction (Car et al. 2020).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Car et al. 2020

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Study design and conductreasoning

    The study designs and a small number of studies limit the findings of this review.

    “The study designs and a small number of studies limit the findings of this review.”
    Quote in the paperSubstantial claim
55 of 55

  • The paper relies on this without support.

    Traditional face-to-face consultation is restricted by time and location, which can lead to insufficient or delayed health education for cancer patients.

    “However, traditional face-toface consultation is restricted by time and location.For example, patients may face difficulties in accessing healthcare providers due to geographical distance or the unavailability of timely appointments, which could lead to insufficient or delayed health education for cancer patients.”

  • The paper relies on this without support.

    Developing countries have limited medical resources.

    “Especially in developing countries with limited medical resources, there may be a greater need to integrate CAs into primary care to provide healthcare support for cancer patients and share the educational burden of nurses.”
    “especially for those developing countries and areas with limited medical resources.”

  • The paper relies on this without support.

    Studies with positive outcomes or significant results are more likely to be published.

    “Studies with positive outcomes or significant results are more likely to be published, which may have skewed the findings towards favourable perceptions of CAs.”
3 of 3

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Abstract counts three interaction-preference studies; results identify two

    Minor

    The abstract reports three studies on preferences for multiple options or "read more," but the results identify only Chaix and Siglen for those interaction preferences.

  • Customisation claim counts two studies but cites three

    Minor

    The statement attributes the finding to two studies, while its citation list names Goertz, Qiu, and Siglen.

  • Population categories total 13 studies, not 12

    Minor

    The population counts in Table 2 sum to 13, although the table says it covers 12 included studies.

    Reported 12 · recomputed 13

  • Satisfaction claim counts five studies but cites six

    Minor

    The statement says five studies, but its supporting citation list names six distinct included studies.

4 of 4

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Li, K., Wu, S., Zhang, Y., Zhu, B., Qi, Z., Hou, S., & Wang, W.
Affiliations
First Affiliated Hospital Zhejiang University (CN)
Corresponding author
Wang, Wei
ORCID
Venue
Journal of Clinical Nursing
Publisher
Wiley
Published
2025-07-14
Volume
35
Issue
10
Pages
4038-4048
ISSN
  • 0962-1067 (ISSN-L)
  • 1365-2702
DOI
10.1111/jocn.70020
PMC
PMC13569306
OpenAlex
W4412388442
Semantic Scholar
e27bc9b9df48edecfaa12e725ce8855e9d54e130
Cited by
10
References
50
Citations by year
2026: 8 · 2025: 2
FWCI
2.02
Type
journal article
Language
en
Open access
Licence
Locations
  • Journal of Clinical Nursing (publishedVersion)
  • PubMed (publishedVersion)
  • PubMed Central (submittedVersion)
Funders
Zhejiang University (HL202402)
Topics
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Health Sciences › Health Professions › General Health Professions › Mobile Health and mHealth Applications
  • Physical Sciences › Computer Science › Artificial Intelligence › AI in Service Interactions
Keywords
chatbots, artificial intelligence, usability, patient experience, patient engagement, self-management, psychological support, genetic testing, resource-limited settings, nurse-led care
MeSH terms
Artificial Intelligence, Communication, Health Education, Humans, Neoplasms, Patient Education as Topic

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-23.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-23.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review