Development and validation of a chatbot for managing leprosy lesions

Abstract

The objective was to develop, validate, and test the usability of a chatbot as a support tool for healthcare professionals in the management of leprosy lesions. This methodological study involved content and appearance validation conducted by ten healthcare professionals and seven Information and Communication Technology (ICT) specialists, respectively. The study followed the recommendations of the Standards for Quality Improvement Reporting Excellence (SQUIRE 2.0) and the principles of the Object-Oriented Hypermedia Design Method (OHDM), encompassing the phases of textual corpus development (modeling), design of the chatbot’s structure and operation (navigational design and abstract interface), and implementation, including validation procedures and usability testing. The ChatLinoHans chatbot achieved satisfactory results across all validation assessments. Expert evaluation yielded a Content Validity Index (CVI) of 0.98, ensuring the credibility and legitimacy of the chatbot. Appearance validation performed by technical judges resulted in agreement rates ranging from 85.7% to 100%, confirming its adequacy in terms of appearance. Usability testing scores ranged from 77.5 to 100, indicating above-average performance.

Key words:
Leprosy; Technology; Artificial intelligence

Resumo

Objetivou-se desenvolver, validar e testar a usabilidade de um chatbot como ferramenta de suporte aos profissionais de saúde na abordagem de lesões da hanseníase. Trata-se de um estudo metodológico que teve o conteúdo e aparência validados por dez profissionais da área da saúde e sete Técnicos em Informática e Comunicação/Computação (TIC), respectivamente. O estudo seguiu as recomendações do Standards for Quality Improvement Reporting Excellence (SQUIRE2.0), além dos pressupostos do referencial Object-Oriented Hypermedia Design Method (OHDM), passando pelas fases de elaboração do corpus textual (modelagem), condução da estrutura e funcionamento do chatbot (projeto navegacional/ e interface abstrata) e realização das validações e teste de usabilidade (implementação). O chatbot ChatLinoHans obteve resultados satisfatórios nas avaliações de validações. A avaliação dos especialistas conferiu IVC de 0,98 garantindo credibilidade e legitimidade ao chatbot. A validação de aparência realizada pelos juízes técnicos, obteve uma avaliação de 85,7 a 100%, caracterizando como válida no quesito aparência. O teste de usabilidade obteve pontuação de 77,5 a 100, adquirindo resultado acima da média.

Palavras-chave:
Hanseníase; Tecnologia; Inteligência artificial

Resumen

El objetivo fue desarrollar, validar y evaluar la usabilidad de un chatbot como herramienta de apoyo para profesionales de la salud en el abordaje de lesiones de la lepra. Se trata de un estudio metodológico en el que la validación del contenido y de la apariencia fue realizada por diez profesionales del área de la salud y siete especialistas en Tecnologías de la Información y la Comunicación (TIC), respectivamente. El estudio siguió las recomendaciones de los Standards for Quality Improvement Reporting Excellence (SQUIRE 2.0), así como los principios del Object-Oriented Hypermedia Design Method (OHDM), abarcando las fases de elaboración del corpus textual (modelado), diseño de la estructura y funcionamiento del chatbot (diseño navegacional e interfaz abstracta) e implementación, incluyendo los procesos de validación y la prueba de usabilidad. El chatbot ChatLinoHans obtuvo resultados satisfactorios en las evaluaciones realizadas. La evaluación por expertos alcanzó un Índice de Validez de Contenido (IVC) de 0,98, lo que garantiza la credibilidad y legitimidad del chatbot. La validación de la apariencia realizada por los evaluadores técnicos obtuvo porcentajes de concordancia entre 85,7% y 100%, confirmando su adecuación en términos de apariencia. La prueba de usabilidad presentó puntuaciones entre 77,5 y 100, lo que indica un desempeño superior a la media.

Palabras clave:
Lepra; Tecnología; Inteligencia artificial

Introduction

Leprosy is a long-lasting and often ignored infectious disease¹, caused mainly by the bacterium Mycobacterium leprae (ML)². It mainly affects the skin and nerves in the face, arms, and legs³. Skin signs include light, red, or brown patches⁴, with changes in sensitivity, numbness, or loss of sensitivity². In the limbs, nerve damage can cause numbness, tingling, and weak muscles, leading to disability and deformity⁵.

Frequently associated with underreporting, stigmatization, and discrimination, it can produce significant biopsychosocial impacts, sustaining transmissibility due to late diagnoses and difficulties in clinical management⁶. As a result, this disease significantly affects populations worldwide, especially in developing countries⁷.

In 2023, 172,717 new leprosy cases were reported globally, with most occurring in India, Brazil, and Indonesia⁷. Notably, Brazil accounts for about 90% of leprosy cases in the Americas⁸. From 2013 to 2022, 316,182 new cases were identified there⁹. Following previous declines, Brazilian case numbers have risen since 2020, increasing from 18,318 in 2020 to 22,773 in 2023. Borderline disease and grade 2 disability were the most common presentations, mainly affecting males aged 30-59 years⁹.

To deal with this, the Global Leprosy Strategy (2021-2030) sets goals like making health services better, stopping disability, fighting stigma, and keeping health workers trained¹⁰. An understanding that highlights the need for effective alternatives to aid professional practice and clinical reasoning in the face of this condition.

Artificial intelligence (AI) is now used more in healthcare and can help doctors make better diagnoses¹¹, reduce errors¹², and possibly achieve lower costs¹³. AI includes rule-based systems, which organize and present information through rules, flows and content^14,15, up to machine learning and content generation, which represent a greater degree of adaptation and autonomy of these software programs¹⁶, learning new patterns from data and continuously adapting their behavior¹⁷.

In leprosy care, AI can support early detection¹⁸, diagnostic support^19,20, and clinical management²¹. Notably, recent uses have focused on diagnostic support¹⁹, particularly on symptom identification using image-based classification with methods such as neural networks, convolutional neural networks, and support vector machines²².

Even though machine learning is helping with leprosy, these tools can be hard to use in daily care because they require lots of data²³, are hard to understand²⁴, and mostly focus on diagnosis instead of full support for deciding treatment²⁵.

Because most tools are still about diagnosing leprosy, there is still needs for help with decisions and lesion management in everyday care^19,26. Thus, given the identified gap and the need for support in addressing leprosy lesions, this study developed and tested a chatbot to support health workers managing leprosy lesions.

Methods

This methodological study describes the development, validation, and usability testing of a chatbot. The process followed the Standards for Quality Improvement Reporting Excellence (SQUIRE 2.0)²⁷ and targeted AI systems for health decision-making. The Object-Oriented Hypermedia Design Method (OHDM) guided four phases: conceptual modeling, navigational design, abstract interface design, and implementation²⁸. This study is available at www.scielo.org.

The chatbot used a hybrid architecture. It combined an explicit rule-based clinical system, which set interaction flow, content, and response limits, with the GPT-4 model to generate natural language answers.

The chatbot was developed using information from a narrative review of leprosy lesions. We used articles from the Virtual Health Library and the Ministry of Health (Leprosy)²⁹ for data. We organized the information with the IRaMuTeQ software through classical lexical statistics analysis, basic lexicography and similarity analysis, ensuring thematic coherence and terminological consistency³⁰.

The chatbot’s knowledge base described the four types of leprosy and the skin signs for each type. It also had common user questions and answers. This information only helped the system make answers. It did not train, retrain, or fine tune any AI models.

During navigational design, the system began by identifying a natural language input. It processed the information and triggered rules using recognized keywords and semantic patterns tied to the four leprosy lesion groups. The system retrieved content from the knowledge base, then built a structured prompt sent to GPT-4 to create a natural language response. For each flow, pre-structured responses offered various forms of information presentation and linguistic variation. This process preserved clinical coherence and alignment with the predefined informational scope, as shown in Chart 1.

Chart 1
Examples of flows followed by the chatbot.

Entry	Rule	Flow	Response type
Presence of a spot with altered sensitivity	“spot”, “numbness”, “loss of sensation”	Paucibacillary leprosy	Clinical description, diagnostic differentiation, and guidance for professional evaluation
Report of numbness or tingling	“tingling”, “numbness”, “shock”	Neural impairment	General clinical guidelines and referral alert
Diffuse lesion without sensation	“Extensive injury” “No sensation”	Lepromatous leprosy	Clinical information and reinforcement of the need for specialized evaluation
Question about treatment	“treatment”, “medicine”, “polychemotherapy”	Treatment	Educational information
General complaints	Uncategorized questions	General information	Guiding response

• Source: Authors.

In designing the interface, we included spaces for users to ask questions in their own words, pick from set responses, and see helpful messages. We also defined feedback to user mechanism and added ways for users to keep talking to the chatbot, switch topics, or end the chat.

The implementation phase combined all components into a working chatbot. This included the GPT-4 language model, knowledge base, navigation rules, and preset prompts. The authors conducted repeated tests to find and fix technical issues, conversational flow flaws, and out-of-scope responses. Navigation rules, prompts, and response presentation were refined as needed.

The chatbot was validated from December 2024 to January 2025 by 10 health experts and 7 Information and Communication Technology/Computing (ICT) professionals. After initial adjustments, the chatbot remained stable. No structural or content changes occurred during validation and usability testing.

Judges were chosen through non-probabilistic convenience snowball sampling³¹, employing Jasper’s eligibility criteria³². We aimed to recruit 6-20 experts^33,34, a sufficient range for stable content-validity indices^35,36. Eligibility required judges to have at least two of these elements: a master’s thesis or doctoral dissertation in the field, teaching or care experience in the field, research output, serving on examining committees, or receiving honors at scientific events³².

Health experts validated the chatbot for content, relevance, comprehension, appropriateness, and clarity. They tested responses by asking about lesions of the four clinical forms of leprosy: indeterminate, tuberculoid, borderline, and lepromatous. Experts interacted directly with the chatbot, freely simulating relevant clinical questions.

The experts then answered six questions: “Does this content appear clear and understandable for the chatbot’s target audience?”; “Is the way the content was addressed in the chatbot appropriate?”; “Is the content retrieved relevant to the chatbot? “Does the chatbot content contribute to closing the diagnosis of leprosy?” “Is this chatbot pertinent for health professionals?” and “Is the chatbot content relevant?” Responses were recorded on a Likert-type scale adapted from Mota et al.³⁷, with the following options: “Agree”, “Strongly agree”, “Disagree”, “Strongly disagree”, and “not applicable”.

Content validity was measured using the Content Validity Index (CVI). For each item, the I-CVI was the proportion of evaluators scoring 3 or 4. Overall validity was measured by S-CVI/Ave, the mean of all I-CVIs, and S-CVI/UA, the proportion of items with full evaluator agreement. A CVI≥0.80 was required for acceptance³⁸.

To analyze content validity, the binomial test was used at a 5% significance level. The null hypothesis was agreement below 0.80; the alternative, above 0.80. Statistical analyses were conducted using R. Results at or above cutoffs were deemed satisfactory, showing adequate content, technical structure, and usability for the chatbot.

In addition, the ICT questionnaire assessed functionality, reliability, usability, efficiency, and performance using the response options “completely appropriate”, “very appropriate”, “moderately appropriate”, “slightly appropriate”, and “not at all appropriate”, adapted from Tannure³⁹. Percentage agreement analysis was also applied, and items reaching at least 75% agreement were considered valid⁴⁰.

Chatbot usability was assessed by seven health professionals (5 nurses and 2 doctors). Professionals were selected from one Family Health Strategy (ESF) unit in each health district of Juazeiro do Norte with the highest number of leprosy cases in the assigned catchment area. Usability was assessed using a questionnaire adapted from Brooke⁴¹.

For the usability questionnaire, the sum of the values obtained for the 10 questions was multiplied by 2.5. The resulting score was interpreted as follows: 20.5 (worst imaginable), 21-38.5 (poor), 39-52.5 (fair), 53-73.5 (good), 74-85.5 (excellent), and 86-100 (best imaginable)^41-43.

Thus, the study was limited to content, appearance, and usability validation of the chatbot and did not aim to evaluate clinical effectiveness or impact on care practice. The Research Ethics Committee of the State University of Ceará approved the study under Opinion No. 7.108.334.

Results

The “ChatLinoHans” chatbot was hosted on a landing page using a content management system (CMS) in WordPress mode for hosting and management. We opted for this paid version because it offered easier control and more efficient administration⁴⁴. OpenAI Incorporated was used as the provider, and ChatGPT-4 served as the interface⁴⁵.

The chatbot is openly accessible at https://hanseniase.contratesolutions.com.br and is organized into four interfaces: the first introduces the technology; the second displays a QR code for access; the third provides a description of the tool; and the fourth opens the chatbot for conversation, as shown in Figure 1. The lesion images available in the technology are in the public domain and were part of the publications retrieved in the narrative review that informed the knowledge base.

Figure 1
ChatLinoHans interfaces.

Chart 2 presents an example of interaction between a user and the chatbot, illustrating how the system operates through questions formulated in natural language and the corresponding responses generated.

Chart 2
User-chatbot interaction example.

User question

Chatbot answer

The patient displays a single hypopigmented spot, with partially preserved sensitivity.

Hello, doctor! This is Dr. Hansen. Based on the description: “single hypopigmented spot, with partially preserved sensitivity”, we are dealing with a lesion suggestive of leprosy in its INDETERMINATE form. ● Lesion characteristics: Hypochromic spot (lighter than normal skin), single, with poorly defined borders, generally without infiltration or elevation. ● Sensitivity: Partially preserved, which is typical of this form. There may be some decrease in thermal and/or tactile sensitivity, but not necessarily abolished. ● Operational classification: This presentation corresponds to the PAUCIBACILLARY group (1 to 5 lesions). ● Sensitivity test: It is essential to perform sensitivity tests (thermal, tactile and painful) on the lesion to assist in diagnostic confirmation. ● Clinical examination: Always examine all skin segments, including areas normally covered by clothing, such as buttocks and back, to rule out other lesions. ● If there is no contraindication, start treatment for paucibacillary leprosy: Rifampicin, Dapsone and Clofazimine, according to the SUS protocol for 6 months. ● Instruct the patient on the importance of follow-up and monitor for possible adverse reactions to treatment.

• Source: Authors.

Content was validated by 10 health experts from December 2024 to January 2025. Most were women (80.0%), with a mean age of 41.20 years (SD±7.50) and a mean time since graduation of 17.10 years. Master’s degree holders predominated (60%). Half (50.0%) had experience in leprosy, and 40% were so in Collective/Public Health; 40.0% were faculty members, and 40.0% worked in Primary Health Care. All participants had publications and research experience in the fields of interest (infectious/parasitic diseases, leprosy, and Public/Collective health).

For expert content validation, the CVI (I-CVI and S-CVI/Ave) was calculated for the evaluative items of the instrument. The content and information-reliability domains both reached an I-CVI of 1.00. The overall CVI, calculated from the mean of all items, was 0.98, as shown in Chart 3.

Chart 3
Content validation (n=10).

Explored clinical form	Evaluated domain	Evaluated items	I-CVI	p**	Mean & SD
Indeterminate leprosy lesion characteristics	Content	Does this content seem clear and understandable to the chatbot’s target audience?	1.00	0.001	4.5±0.53
	Structure and presentation	Is the way this content was presented in the chatbot appropriate?	0.9	0.021	4.2±0.92
	Information reliability	Is the content triggered in the scripts relevant to the chatbot?	1.00	0.001	4.5±0.53
Tuberculoid leprosy lesion characteristics	Content	Does this content seem clear and understandable to the chatbot’s target audience?	1.00	0.001	4.5±0.53
	Structure and presentation	Is the way this content was presented in the chatbot appropriate?	1.00	0.001	4.3±0.48
	Information reliability	Is the content triggered in the scripts relevant to the chatbot?	1.00	0.001	4.5±0.53
Dimorphic leprosy lesion characteristics	Content	Does this content seem clear and understandable to the chatbot’s target audience?	1.00	0.001	4.6±0.52
	Structure and presentation	Is the way this content was presented in the chatbot appropriate?	1.00	0.001	4.4±0.52
	Information reliability	Is the content triggered in the scripts relevant to the chatbot?	1.00	0.001	4.4±0.52
Lepromatous leprosy lesion characteristics	Content	Does this content seem clear and understandable to the chatbot’s target audience?	1.00	0.001	4.7±0.48
	Structure and presentation	Is the way this content was presented in the chatbot appropriate?	1.00	0.001	4.4±0.52
	Information reliability	Is the content triggered in the scripts relevant to the chatbot?	1.00	1.00	4.4±0.52
General	Relevance	Does the chatbot’s content contribute to a confirmed leprosy diagnosis?	0.9	0.021	4.3±0.95
		Is this chatbot relevant for health professionals?	1.00	1.00	4.6±0.52
		Is the chatbot’s content relevant?	1.00	1.00	4.7±0.48
S-CVI/UA§			0.86
S-CVI/Ave †			0.98

• Notes: Five-point Likert scale (1=strongly disagree; 5=strongly agree), SD: sample standard deviation, I-CVI: Item Content Validity Index, S-CVI/UA§: Universal agreement proportion, S-CVI/Ave†: Mean of I-CVI.
• Source: Authors, based on research data.

All items yielded p-values indicating statistically significant validation, reinforcing the reliability of participant assessments, as well as a 98% agreement rate, reflecting strong agreement between the findings and the reference criteria. All suggestions made by the judges were incorporated. These suggestions included adding more images, reorganizing the text, replacing terminology, and redirecting responses related to clinical leprosy conditions.

The ICT judges were aged 22-38 years. All were male, held degrees in information systems, and were experts. Professional experience ranged from 3 to 11 years, and all had experience in software development. ChatLinoHans showed adequate face validity for all 26 evaluative items across the five analyzed attributes, with agreement percentages of 85.7% or 100%^37,40.

Appearance validity was assessed by seven ICT professionals who evaluated the following attributes: functional adequacy, reliability, usability, performance efficiency, and maintainability, as described in Chart 4.

Chart 4
Appearance validation (n=7).

Evaluated attribute	Agreement percentage
Functional suitability attribute
Does the chatbot have the necessary functions to operate?	100%
Does the chatbot function correctly?	85.7%
Do user actions work in the chatbot?	100%
Reliability attribute
Does the chatbot have errors that frequently hinder its performance?	85.7%
After errors/failures, does the chatbot resume functioning correctly?	100%
Is there any data loss in case of failures?	100%
Is the chatbot accessible for use?	100%
Usability attribute
Is the chatbot easy to implement?	100%
Is it easy to learn how to use the chatbot?	100%
Is it easy to control the use of the chatbot?	85.7%
Can the chatbot be used for the target audience (health professionals)?	100%
Does the chatbot’s structure help prevent execution errors?	85.7%
Is the chatbot’s appearance (colors, images, and overall design) pleasing?	100%
Performance efficiency attribute
Is the chatbot's execution time adequate?	100%
Does the chatbot have adequate features?	85.7%
Does the chatbot perform well with the features it has?	100%
Does the chatbot seem to operate well in the presence of other systems?	85.7%
Does the chatbot interfere with the operation of other coexisting systems?	100%
Does the chatbot seem secure regarding the stored data?	85.7%
Does the chatbot keep user records?	85.7%
Does the chatbot record user data/information?	100%
Maintainability attribute
Is it easy to identify a flaw in the chatbot?	100%
Is it easy to test the chatbot if changes are made?	100%
Can the chatbot be used in at least one other different system?	100%
Is it easy for the chatbot to adapt to other environments?	100%
Is it possible to adjust the chatbot for a new update?	100%

• Source: Authors, based on research data.

After validation by experts and ICT professionals, usability testing was performed. Seven professionals participated in this stage (2 doctors and 5 nurses). The purpose of usability testing was to assess the chatbot’s efficiency, ease of use, satisfaction, and intuitiveness. Scores ranged from 77.5 to 100 points. According to Brooke⁴¹, ChatLinoHans scored above average and was qualitatively classified as excellent/best imaginable as per Martins et al.⁴².

Discussion

Chatbots have become increasingly common in healthcare delivery and are associated with clinical practice and the promotion of continuing education for health professionals⁴⁶. This is largely due to their ability to sustain a communicative process similar to a conversation, perform tasks, and retrieve information, which together can meaningfully support clinical judgment and decision-making²⁴.

These technologies can take different forms and vary in domain configuration, purpose, and type of interaction. Their use has been linked to streamlined patient management by providing precision, speed, and the ability to process large data volumes in a short time⁴⁶. Accordingly, the use of such technologies in medicine has grown substantially over recent decades, particularly in psychiatry, cardiology, dermatology, and diagnostic imaging²⁵.

The broad validation process undertaken in this study supports the reliability and applicability of the ChatLinoHans chatbot for the clinical management of leprosy lesions, highlighting its alignment with the needs imposed by the disease’s clinical complexity, which requires accurate and contextualized responses tailored to the patient’s condition. The OHDM-supported technology was developed through precise and well-defined stages, which contributed to its quality and reliability.

Notably, ChatLinoHans was designed as an accessible, multiplatform, and user-oriented architecture. The decision to adopt a chatbot format stemmed from these systems’ ability to work with natural language and simulate a human-like conversation⁴⁷.

In addition, landing-page hosting was chosen because it offers greater centralization and broader access to the technology⁴⁸, while providing users with orientation before interaction, allowing them to understand the tool’s purpose and limitations.

Consistent with this, the high validation scores obtained for the tool indicate expert agreement regarding the clarity, relevance, and appropriateness of both the content and the appearance of ChatLinoHans. This, in turn, reinforces the chatbot’s ability to support health professionals in the assessment of leprosy lesions.

Other studies on the development and validation of chatbot-based technologies, such as those by Ferreira et al.⁴⁹ and Cavalcanti et al.⁵⁰, have likewise documented the usability of these tools and their potential contribution to clinical support. These findings underscore the pressing need for initiatives aimed at integrating technology into PHC, given the documented contributions of resources such as telehealth and electronic health records to work management and the reduction of barriers between professionals and patients^49,50.

As identified in the systematic review conducted by Fernandes et al.²², other tools described in previous studies have been proposed to assist in leprosy diagnosis through AI, especially in clinical and diagnostic procedures. However, the authors also highlight the need for strategies capable of incorporating these instruments into routine clinical practice.

One limitation of this work is that, although the chatbot has no machine-learning mechanisms, its operation depends on manual maintenance and updating of the knowledge base, clinical rules, and prompts, as well as ongoing monitoring of updates to natural-language platforms, which may represent a long-term operational challenge²².

In addition, the small sample size and the use of a convenience sample are limitations. Also, we should note that the present study received no funding and that no funding organizations influenced the development of the study or its results.

Conclusions

The ChatLinoHans chatbot showed evidence of reliable validity for its intended purpose of supporting professionals in the assessment of leprosy lesions. Its development was grounded in prior theoretical references and followed the OHDM methodological model, ensuring scientific underpinning from conceptualization through expert validation. This innovative technology uses AI through an easily accessible interface. It is important because it offers support to health professionals and may contribute to improving the epidemiological situation of the disease. Although the study highlights the limitation posed by the need for more frequent updates and maintenance because of the nature of AI, further effect-evaluation studies are recommended to confirm its effectiveness.

References

• ¹ World Health Organization (WHO). Global leprosy (Hansen disease) update, 2022: new paradigm - control to elimination [Internet]. 2022 [cited 2025 abr 20]. Available from: https://www.who.int/publications/i/item/who-wer9837-409-430
  » https://www.who.int/publications/i/item/who-wer9837-409-430
• ² Chen KH, Lin CY, Su SB, Chen KT. Leprosy: a review of epidemiology, clinical diagnosis, and management. J Trop Med 2022; 2022:8652062.
• ³ Brasil. Ministério da Saúde (MS). Secretaria de Vigilância em Saúde. Departamento de Vigilância e Doenças Transmissíveis. Guia prático sobre a hanseníase. Brasília: MS; 2017.
• ⁴ Chandran SL, Tiwari A, Lustosa AA, Demir B, Bowers B, Albuquerque RGR, Prado RBR, Lambert S, Watanabe H, Haagsma J, Richardus JH. Revised estimates of leprosy disability weights for assessing the global burden of disease: a systematic review and individual patient data meta-analysis. PLoS Negl Trop Dis 2021; 15(3):e0009209.
• ⁵ Brasil. Ministério da Saúde (MS). Secretaria de Ciência, Tecnologia, Inovação e Insumos Estratégicos em Saúde. Protocolos clínicos e diretrizes terapêuticas da hanseníase. Brasília: MS; 2022.
• ⁶ Van Brakel WH, Sihombing B, Djarir H, Beise K, Kusumawardhani L, Yulihane R, Kurniasari I, Kasim M, Kesumaningsih KI, Wilder-Smith A. Disability in people affected by leprosy: the role of impairment, activity, social participation, stigma and discrimination. Glob Health Action 2012; 5:18394.
• ⁷ Organização Mundial da Saúde (OMS). Global leprosy (Hansen disease) update, 2023: elimination of leprosy disease is possible - time to act! Wkly Epidemiol Rec 2024; 99(37):501-521.
• ⁸ Organização Pan-Americana da Saúde (OPAS). Leprosy (Hansen disease) [Internet]. 2024 [cited 2025 dez 13]. Available from: https://www.paho.org/en/topics/leprosy-hansen-disease
  » https://www.paho.org/en/topics/leprosy-hansen-disease
• ⁹ Brasil. Ministério da Saúde (MS). Boletim Epidemiológico: hanseníase [Internet]. 2024 [acessado 2025 abr 23]. Disponível em: https://www.gov.br/saude/pt-br/centrais-de-conteudo/publicacoes/boletins/epidemiologicos/especiais/2024/be_hansen-2024_19jan_final.pdf/view
  » https://www.gov.br/saude/pt-br/centrais-de-conteudo/publicacoes/boletins/epidemiologicos/especiais/2024/be_hansen-2024_19jan_final.pdf/view
• ¹⁰ World Health Organization (WHO). Global leprosy strategy 2021-2030: towards zero leprosy. Geneva: WHO; 2021.
• ¹¹ Teles AS, Fernandes TRS, Balhara S, Gupta N, Teixeira S. Artificial intelligence on diagnostic aid of leprosy: a systematic literature review. J Clin Med 2024; 13(1):180.
• ¹² Ouanes K, Farhah N. Effectiveness of artificial intelligence (AI) in clinical decision support systems and care delivery. J Med Syst 2024; 48(1):74.
• ¹³ Alowais S, Alghamdi S, Alsuhebany N, Alqahtani T, Alshaya A, Almohareb S, Aldairem A, Alrashed M, Bin Saleh K, Badreldin HA, Al Yami MS, Al Harbi S, Albekairy AM. Revolutionizing healthcare: the role of artificial intelligence in clinical practice. BMC Med Educ 2023; 23(1):689.
• ¹⁴ Ronanki R, Nguyen A. Artificial intelligence for the real world. IRJMETS 2023; 5(6):3526-3537.
• ¹⁵ Xu Y, Wang Q, An Z, Wang F, Zhang L, Wu Y, Dong F, Qiu CW, Qiu J, Hua K, Su W, Wu J, Xu H, Han Y, Fu C, Yin Z, Liu M, Roepman R, Dietmann S, Virta M, Kengara F, Zhang Z, Zhang L, Zhao T, Dai J, Yang J, Lan L, Luo M, Liu Z, An T, Zhang B, He X, Cong S, Liu X, Zhang W, Lewis JP, Tiedje JM, Wang Q, An Z, Wang F, Zhang L, Huang T, Lu C, Cai Z, Wang F, Zhang J. Artificial intelligence: a powerful paradigm for scientific research. Innovation 2021; 2(4):100179.
• ¹⁶ Banh L, Strobel G. Generative artificial intelligence. Electron Mark 2023; 33:1-17.
• ¹⁷ Kühl N, Schemmer M, Goutier M, Satzger G. Artificial intelligence and machine learning. Electron Mark 2022; 32:2235-2244.
• ¹⁸ Souza M, Lopes G, Branco A, Fairley J, Fraga L. Leprosy screening based on artificial intelligence: development of a cross-platform app. JMIR Mhealth Uhealth 2021; 9(4):e23718.
• ¹⁹ Andrade HGV, Rocha ES, Monteiro KHC, Morais CM, Santos DC, Nascimento DC, Dourado RA, Lynn T, Endo PT. On the usage of artificial intelligence in leprosy care: a systematic literature review. PLoS Comput Biol 2025; 21(6):e1012550.
• ²⁰ Simões MMR, Lima FR, Lugão HB, Paula NA, Silva CML, Ramos AF, Frade MAC. Development and validation of a machine learning approach for screening new leprosy cases based on the leprosy suspicion questionnaire. Sci Rep 2025; 15(1):6912.
• ²¹ Deps P, Yotsu R, Furriel B, Oliveira B, Lima S, Loureiro R. The potential role of artificial intelligence in the clinical management of Hansen's disease (leprosy). Front Med (Lausanne) 2024; 11:1338598.
• ²² Fernandes JRN, Teles AS, Fernandes TRS, Lima LDB, Balhara S, Gupta N, Teixeira S. Artificial intelligence on diagnostic aid of leprosy: a systematic literature review. J Clin Med 2023; 13(1):180.
• ²³ Elshawi R, Al-Mallah MH, Sakr S. On the interpretability of machine learning-based model for predicting hypertension. BMC Med Inform Decis Mak 2019; 19:146.
• ²⁴ Xu L, Sanders L, Li K, Chow J. Chatbot for health care and oncology applications using artificial intelligence and machine learning: systematic review. JMIR Cancer 2021; 7(4):e27850.
• ²⁵ Paixão GMM, Santos BC, Araújo RM, Ribeiro MH, Moraes JL, Ribeiro AL. Machine learning na medicina: revisão e aplicabilidade. Arq Bras Cardiol 2020; 118(1):95-102.
• ²⁶ Castor ECS, Fernandes AL, Motta ACGD, Garcia RB, Lima AF. Chatbot: impactos no ambiente acadêmico de uma universidade do Rio de Janeiro. P2P Inovação 2021; 8(1):71-92.
• ²⁷ Ogrinc G, Davies L, Goodman D, Batalden P, Davidoff F, Stevens D. Standards for Quality Improvement Reporting Excellence 2.0. J Surg Res 2016; 200(2):676-682.
• ²⁸ Schwabe D, Rossi G. The object-oriented hypermedia design model. Commun ACM 1995; 38(8):45-46.
• ²⁹ Brasil. Ministério da Saúde (MS). Biblioteca Virtual em Saúde - Hanseníase [Internet]. 2022 [acessado 2024 jun 10]. Disponível em: https://hansen.bvs.br/colecao-ministerio-da-saude/.
  » https://hansen.bvs.br/colecao-ministerio-da-saude
• ³⁰ Camargo BV, Justo AM. Tutorial para uso do software IRAMUTEQ. Florianópolis: Universidade Federal de Santa Catarina; 2018.
• ³¹ Polit DF, Beck CT. Fundamentos de pesquisa em enfermagem: avaliação de evidências para a prática de enfermagem. 7ª ed. Porto Alegre: Artmed; 2011.
• ³² Jasper MA. Expert: a discussion of the implications of the concept as used in nursing. J Adv Nurs 1994; 20(4):769-776.
• ³³ Pasquali L. Psicometria: teoria e aplicações. Brasília: UnB; 1997.
• ³⁴ Vianna HM. Testes em educação. São Paulo: Ibrasa; 1982.
• ³⁵ Alexandre NMC, Coluci MZO. Validade de conteúdo nos processos de construção e adaptação de instrumentos de medidas. Cien Saude Colet 2011; 16(7):3061-3068.
• ³⁶ Rubio DM, Berg-Weger M, Tebb SS, Lee ES, Rauch S. Objectifying content validity. Soc Work Res 2003; 27(2):94-104.
• ³⁷ Mota L, David R, Botelho D, Oliveira V. Pesquisa UX aplicada à transformação digital da atenção primária no Brasil. Rev Eur Saude Publica 2020; 30.
• ³⁸ Polit DF, Beck CT. Fundamentos de pesquisa em enfermagem: avaliação de evidências para a prática de enfermagem. 7ª ed. Porto Alegre: Artmed; 2011.
• ³⁹ Tannure MC. Construção e avaliação da aplicabilidade de um software com o processo de enfermagem em UTI [tese]. Belo Horizonte: Universidade Federal de Minas Gerais; 2012.
• ⁴⁰ Felipe GF, Lima FET, Barbosa LP, Moreira TMM, Joventino ES, Freire VS, Mendonça LBA. Evaluation of user embracement software with pediatric risk classification. Rev Bras Enferm 2020; 73(3):e20180677.
• ⁴¹ Brooke J. SUS: a quick and dirty usability scale [Internet]. 1996 [acessado 2025 abr 23]. Disponível em: https://www.researchgate.net/publication/319394819_SUS_--_a_quick_and_dirty_usability_scale
  » https://www.researchgate.net/publication/319394819_SUS_--_a_quick_and_dirty_usability_scale
• ⁴² Martins AI, Rosa AF, Queirós A, Silva A, Rocha NP. European Portuguese validation of the System Usability Scale (SUS). Procedia Comput Sci 2015; 67:293-300.
• ⁴³ Sousa LB, Sousa IM, Costa EC, Freitas APF, Santos LVF, Ferreira AGN. IST Nurse: development and validity evidence of a mobile application. Rev Latino Am Enferm 2024; 32:e4317.
• ⁴⁴ WordPress. WordPress sem complicação [Internet]. 2024 [acessado 2025 abr 23]. Disponível em: https://wordpress.com/pt-br/.
  » https://wordpress.com/pt-br
• ⁴⁵ OpenAI. Introducing ChatGPT [Internet]. 2025 [acessado 2025 abr 23]. Disponível em: https://openai.com/index/chatgpt/.
  » https://openai.com/index/chatgpt
• ⁴⁶ Wah J. Revolutionizing e-health: the transformative role of AI-powered hybrid chatbots in healthcare solutions. Front Public Health 2025; 13:1530799.
• ⁴⁷ Casheekar A, Lahiri A, Rath K, Prabhakar K, Srinivasan K. A contemporary review on chatbots and AI-powered virtual conversational agents. Comput Sci Rev 2024; 52:100632.
• ⁴⁸ Santiago NO, Leite Filho CAP. O uso de chatbots como ferramenta de comunicação online em lojas de varejo brasileiras. Rev Uni RN 2019; 19(1):208-235.
• ⁴⁹ Ferreira DS, Oliveira BSB, Paula RNF, Jorge MSB. Validação de um chatbot para o cuidado em saúde. Bol Conjuntura 2024; 18(53):324-346.
• ⁵⁰ Cavalcanti AOR, Cavalcanti ARS, Nascimento JWA, Souza RR, Roque GSL, Queiroz SRM, Santos ICRV. Chatbot de orientações acerca do aleitamento materno. In: Simpósio Brasileiro de Computação Aplicada à Saúde (SBCAS), 2021. Porto Alegre: SBC; 2021. p. 422-427.

• Data availability statement
  The data sources adopted in the research are indicated in the article’s body.

• Chief editors:
  Maria Cecília de Souza Minayo, Romeu Gomes, Antônio Augusto Moura da Silva, Vania de Matos Fonseca

Data availability

The data sources adopted in the research are indicated in the article’s body.

Publication Dates

• Publication in this collection
  29 June 2026
• Date of issue
  May 2026

History

• Received
  16 May 2025
• Accepted
  26 Jan 2026
• Published
  28 Jan 2026