ChatGPT’s Ability to Answer Cancer-Related Basic Questions in Urdu: A Comparative Study with English Responses

ChatGPT’s Ability to Answer Cancer-Related Basic Questions in Urdu: A Comparative Study with English Responses | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article ChatGPT’s Ability to Answer Cancer-Related Basic Questions in Urdu: A Comparative Study with English Responses Waqas Ahmed Khan, Misbah Soomro, Muhammad Afzal, Adeeba Zaki This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7472185/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Chat Generative Pre-Trained Transformer (ChatGPT) has become a valuable tool since its launch in 2022 and has proved to be useful in providing easily understandable conversational responses across various topics, including medical queries. This study aims to evaluate the efficacy of ChatGPT-4 in responding to basic cancer-related questions in both Urdu and English, investigating linguistic discrepancies that may affect the reliability of AI-generated medical advice. Methods We compiled a set of 68 distinct cancer-related questions, translated into both Urdu and English, and presented them to ChatGPT-4. Responses were independently evaluated by two physicians for accuracy and comprehensiveness, with discrepancies resolved by a third reviewer. The responses in the two languages were compared for accuracy. Results ChatGPT-4 provided comprehensive responses in 79% of the Urdu queries and 97% of the English queries. Accuracy assessment showed that 72% of Urdu responses were at least as accurate as their English counterparts. The treatment-related category had the highest comprehensiveness in Urdu responses at 92.3%. Conclusion While ChatGPT-4 performs proficiently in both Urdu and English, differences in the quality of responses show that there is a need for improvements in Urdu responses. Enhancing the model's training on Urdu datasets and medical terminology could bridge this existing gap and ensure equitable quality of medical information across languages. ChatGPT-4 Urdu English cancer accuracy linguistic discrepancies Figures Figure 1 Figure 2 Background Chat Generative Pre-Trained Transformer (ChatGPT) is a large language model (LLM), introduced by Open AI in November 2022, that has been trained on vast datasets covering a variety of topics. Since its launch, it has rapidly gained popularity worldwide. It comprehends user questions and responds in a conversational manner that is easy to understand ( 1 , 2 ). ChatGPT’s performance has been explored in multiple medical disciplines with queries related to diseases and their management over the last two years ( 3 – 5 ). According to a study assessing ChatGPT's understanding of cirrhosis and hepatocellular carcinoma, 79% of the model's responses were correct, potentially helping not only healthcare providers but also patients in understanding their disease ( 6 ). Cancer patients need comprehensive care and education about their disease, treatment options, and emotional needs ( 7 ). In the era of the internet and medical research, information on medical conditions is widely accessible but it is often not tailored to individual patient's needs. As a result, the patients are dependent entirely on information provided by physicians. The information can be influenced by various factors, including personal experiences, prevailing myths, and societal stigmas. This often results in misinformation, potentially impacting crucial decision-making ( 8 ). The digital divide further limits access, as disparities in technology and digital literacy prevent many patients, especially in underserved areas, from obtaining personalized, reliable information ( 9 ). This study aimed to evaluate the accuracy and reliability of ChatGPT-4 in addressing questions commonly asked by cancer patients, with a special focus on examining the disparities in responses between Urdu and English. Prior studies have evaluated the ability of ChatGPT-3.5 and ChatGPT-4 to provide accurate and culturally relevant health information in other languages such as Spanish, French, and Arabic, demonstrating their effectiveness in addressing medical conditions across different linguistic contexts ( 6 , 10 , 11 ). The lower literacy rates in Pakistan contribute to the prevalence of skewed information, which is often impacted by opinions and prone to serious misinterpretation ( 12 ). Reliable medical information in Urdu is scarce, making it difficult for the Pakistani population to access dependable resources. Many are unaware of where to find accurate information for self-education, leading to a knowledge gap that can negatively affect decision-making and patient outcomes. Given these challenges, ChatGPT has shown potential in addressing clinical questions, and this study sought to assess its accuracy in providing cancer-related information in both Urdu and English. Methods To generate a thorough list of questions covering various aspects of cancer, preliminary data comprising 115 questions was collected from multiple sources, such as blogs, social media platforms, FAQs from websites of hospitals and cancer institutes. Two physicians carefully screened each question, ensuring that it was relevant to a cancer patient and easy to understand. Duplicate and unclear questions were removed, and eight new, commonly asked questions that arise in day-to-day clinical practice were added. This resulted in a final list of 68 distinct questions. These questions were translated into Urdu and English and verified for accuracy by a third physician. The methodological framework is illustrated in Figure 1 and the full list of English-language cancer-related questions is provided in Supplementary File. The six main areas of oncology covered by the questions were: 1) Basic Questions, 2) Etiology and Risk Factors, 3) Diagnostics, 4) Treatment and Prevention, 5) Side Effects and Quality of Life aspects, and 6) Outcomes and Surveillance. ChatGPT was given a uniform prompt before each question as shown in Figure 2 . “You are an oncologist. Your patient asks you common questions related to their cancer and treatment. Please respond to the questions in the same language as the question. Respond with "OK" if the instructions are understood”. After ChatGPT initially confirmed by responding with "OK," the actual questions were displayed. For a total of 136 interactions, this process was repeated for every question, starting a fresh dialogue for each question after deleting the history of prior interaction. The responses generated by ChatGPT were independently evaluated by two physicians to assess accuracy, with any differences in evaluation resolved by a third reviewer. The assessment criteria were classified as follows: 1. Comprehensive: The response thoroughly addressed the question. 2. Correct but Inadequate or Unfocused: The response, while accurate, lacked depth, detail, or complete relevance. 3. Incorrect: The response did not accurately address the question. Additionally, the responses in Urdu and English were compared to determine any differences in accuracy and were scored as below: 1. Urdu Response is More Accurate: The Urdu response was deemed superior. 2. Similar Accuracy: The responses in both languages were judged to be equally accurate. 3. Urdu Response is Less Accurate: The Urdu response was less accurate than the English version. Results A total of 68 questions were included in our study. When examining the Urdu responses, the model provided 79% comprehensive and 21% correct but inadequate or unfocused answers. None of the responses were entirely incorrect. The grading of English responses showed significantly higher scores with 97% comprehensive answers (Table 1). When stratified by subgroups, ChatGPT performed equally well across all groups, with only two responses not meeting expectations. For Urdu responses, ChatGPT performed best in the treatment group with 92.3% comprehensive answers, followed by the basic questions group at 83.33%. Additional scores are detailed in Table 2. When comparing the accuracy of Urdu responses to English responses, 72% of responses were found to be at least as accurate as in English, with 4 (5.88%) responses even better than those in English. For 19 (27.94%) questions, Urdu responses were poorer compared to English (Table 3). Urdu responses were overall best in the side effects and quality of life group, with 2 responses better than English and 10 equally good as in English. In the rest of the groups, English responses were graded as better overall. The details of each subgroup are given in Table 4. Table 1. Grading responses by ChatGPT Overall All responses (n=68) Urdu Answers n (%) English Answers n (%) 1. Comprehensive 54 (79.41 %) 66 (97.05%) 2. Correct but Inadequate or Unfocused 14 (20.58%) 2 (2.94%) 3. Incorrect 0 0 Table 2: Grading of Responses Generated by ChatGPT-4 to Cancer Questions Categorized by Subgroup Response Scores Urdu Answers n (%) English Answers n (%) Basic Knowledge (n = 12) 1. Comprehensive 10 (83.33%) 12 (100%) 2. Correct but Inadequate or Unfocused 2 (16.66%) 0 3. Incorrect 0 0 Etiology/Risk Factors (n = 13) 1. Comprehensive 10 (76.92%) 13 (100%) 2. Correct but Inadequate or Unfocused 3 (23.07%) 0 3. Incorrect 0 0 Diagnostics (n = 6) 1. Comprehensive 4 (66.66%) 6 (100%) 2. Correct but Inadequate or Unfocused 2 (33.33%) 0 3. Incorrect 0 0 Treatment (n = 13) 1. Comprehensive 12 (92.30%) 13 (100%) 2. Correct but Inadequate or Unfocused 1 (7.69%) 0 3. Incorrect 0 0 Side Effects and Quality of Life Aspects (n = 15) 1. Comprehensive 12 (80%) 14 (93.3%) 2. Correct but Inadequate or Unfocused 3 (20%) 1 (6.66%) 3. Incorrect 0 0 Outcomes and Surveillance (n = 9) 1. Comprehensive 6 (66.67%) 8 (88.88%) 2. Correct but Inadequate or Unfocused 3 (33.33%) 1 (11.11%) 3. Incorrect 0 0 Table 3: Grading of Responses Comparing the Accuracy Between Urdu and English Responses Generated by ChatGPT to Cancer-Related Questions Total responses n=68 No. of Questions Percentage 1. Urdu response is more accurate 4 5.88% 2. Similar accuracy 45 66.17% 3. Urdu response is less accurate 19 27.94% Table 4: Grading of Responses Comparing the Accuracy Between Urdu and English Responses Generated by ChatGPT to Cancer-Related Questions Categorized by Subgroup Responses No of Questions Percentage Basic Knowledge (n=12) 1. Urdu response is more accurate 0 0% 2. Similar accuracy 7 58.33% 3. Urdu response is less accurate 5 41.66% Etiology and Risk Factors (n=13) 1. Urdu response is more accurate 0 0% 2. Similar accuracy 11 84.61% 3. Urdu response is less accurate 2 15.38% Diagnostics (n=6) 1. Urdu response is more accurate 1 16.66% 2. Similar accuracy 3 50% 3. Urdu response is less accurate 2 33.33% Treatment (n= 13) 1. Urdu response is more accurate 0 0 2. Similar accuracy 9 69.23% 3. Urdu response is less accurate 4 30.77% Side Effects and Quality of Life Aspects (n=15) 1. Urdu response is more accurate 2 13.33% 2. Similar accuracy 10 66.66% 3. Urdu response is less accurate 3 20% Outcomes and Surveillance (n=9) 1. Urdu response is more accurate 1 11.11% 2. Similar accuracy 5 55.55% 3. Urdu response is less accurate 3 33.33% Discussion The development of AI tools like ChatGPT-4 has made it convenient for everyone to quickly access reliable medical information ( 13 ). Unlike traditional search engines, such as Google that require sorting through large amounts of data, ChatGPT-4 offers a more user-friendly experience ( 14 ). However, the fact that English is not everyone's first language and that most online content is written in English presents a significant challenge for non-English speakers. There is a clear need for more inclusive language support and resources to ensure that everyone can benefit from new technologies ( 15 ). Our study evaluated ChatGPT's ability to answer common cancer-related questions in Urdu and English. The model performed exceptionally well in English, with 97% of responses being comprehensive, compared to 79% in Urdu. It was particularly effective for Urdu treatment-related questions, achieving 92.3% comprehensive answers, and scored 83.33% in basic questions. The performance difference is likely due to the more extensive English datasets used during training, though certain areas might have received more focused training in Urdu as well. Interestingly, 72% of the Urdu responses matched the accuracy of the English ones, with 5.88% even surpassing them. However, 28% of the Urdu responses were less accurate, pointing to a need for more tailored training data to improve overall accuracy in Urdu. While ChatGPT-4 provides adequate responses in both languages, there's room to enhance its Urdu capabilities to better serve millions of Urdu-speaking users. The lack of precise medical terms in Urdu complicates communication and indicates the need for further improvement. As technology advances, AI has the potential to improve medical care. However, to be truly effective, AI needs to work well in multiple languages, clearly communicate with patients in easy language, and provide accurate information to clinicians. Improving AI’s ability to function across different languages is crucial for ensuring equal access to accurate medical information and better outcomes for patients worldwide. Our study shows the importance of expanding non-English datasets to achieve the same level of accuracy seen in English, which is key to overcoming language barriers in healthcare. While ChatGPT’s ease of use alleviates some challenges of traditional search engines, it is crucial to acknowledge its limitations, such as data hallucinations, when the AI confidently provides inaccurate information ( 16 ). Therefore, AI should only be used as a supplementary tool in healthcare rather than a substitute for medical advice. Conclusions Although ChatGPT-4 performs well in both English and Urdu, there is still room for improvement in its Urdu responses to better serve the millions of Urdu-speaking users. To ensure all users have equal access to accurate and comprehensive medical information, AI must be multilingual and proficient across languages. Declarations Ethics approval and consent to participate This study did not involve human participants, identifiable data, or clinical interventions. According to the Aga Khan University Ethics Review Committee (ERC) Policy on Research Ethics Review (ORGS/008-2018, Section 1.2) and the National Bioethics Committee of Pakistan Operational Manual (2023, Section 4.2), such research is exempt from formal ethics review. Ethics approval was therefore not required, and no reference number was issued. This exemption is consistent with the policies of the Aga Khan University ERC. No informed consent was required, as the study involved only AI-generated content and all procedures were conducted solely by the authors. The study adhered to the ethical principles outlined in the Declaration of Helsinki. Consent for publication Not applicable. This AI-based study did not involve human participants or personal data; therefore, informed consent was not required.. Availability of data and materials The datasets and materials generated or analyzed during this study are available upon request. For access to the raw data analyzed in this study, please contact the corresponding author at [email protected] . Competing interests The authors declare no competing interests, financial or otherwise, related to this research. Funding No funding was received for the design, execution, or publication of this study. Authors’ contributions The idea was conceived by WAK. Question generation, translation, and prompt creation were handled by WAK and MS, and verified by MA and AZ. Data collection and analysis were performed by WAK and MA. All authors contributed to the writing. Acknowledgements Not applicable. References Roumeliotis K, Tselikas N. ChatGPT and Open-AI Models: A Preliminary Review. Future Internet. 2023;15:192. Kasneci E, Sessler K, Küchemann S, Bannert M, Dementieva D, Fischer F, et al. ChatGPT for good? On opportunities and challenges of large language models for education. Learning and Individual Differences. 2023;103:102274. Yeo YH, Samaan JS, Ng WH, Ting P-S, Trivedi H, Vipani A, et al. Assessing the performance of ChatGPT in answering questions regarding cirrhosis and hepatocellular carcinoma. Clin Mol Hepatol. 2023;29(3):721-32. Johnson SB, King AJ, Warner EL, Aneja S, Kann BH, Bylund CL. Using ChatGPT to evaluate cancer myths and misconceptions: artificial intelligence and cancer information. JNCI Cancer Spectr. 2023;7(2):pkad015. Nedbal C, Naik N, Castellani D, Gauhar V, Geraghty R, Somani BK. ChatGPT in urology practice: revolutionizing efficiency and patient care with generative artificial intelligence. Curr Opin Urol. 2024;34(2):98-104. Yeo YH, Samaan JS, Ng WH, Ting PS, Trivedi H, Vipani A, et al. Assessing the performance of ChatGPT in answering questions regarding cirrhosis and hepatocellular carcinoma. Clin Mol Hepatol. 2023;29(3):721-32. McCorkle R, Ercolano E, Lazenby M, Schulman-Green D, Schilling LS, Lorig K, et al. Self-management: Enabling and empowering patients living with cancer as a chronic illness. CA Cancer J Clin. 2011;61(1):50-62. Ajith K, Sarkar S, Sethuramachandran A, Manghat S, Surendran G. Myths, beliefs, and attitude toward cancer among the family caregivers of cancer patients: A community-based, mixed-method study in rural Tamil Nadu. J Family Med Prim Care. 2023;12(2):282-8. Fareed N, Swoboda CM, Jonnalagadda P, Huerta TR. Persistent digital divide in health-related internet use among cancer survivors: findings from the Health Information National Trends Survey, 2003-2018. J Cancer Surviv. 2021;15(1):87-98. Soto-Chávez MJ, Bustos MM, Fernández-Ávila DG, Muñoz OM. Evaluation of information provided to patients by ChatGPT about chronic diseases in Spanish language. Digit Health. 2024;10:20552076231224603. Guillen-Grima F, Guillen-Aguinaga S, Guillen-Aguinaga L, Alas-Brun R, Onambele L, Ortega W, et al. Evaluating the Efficacy of ChatGPT in Navigating the Spanish Medical Residency Entrance Examination (MIR): Promising Horizons for AI in Clinical Medicine. Clin Pract. 2023;13(6):1460-87. Khan S, Jalees S, Jabeen Z, Khan M, Qadri RH, Adnan H, et al. Myths and Misconceptions of Breast Cancer in the Pakistani Population. Cureus. 2023;15(6):e40086. Reynolds K, Tejasvi T. Potential Use of ChatGPT in Responding to Patient Questions and Creating Patient Resources. JMIR Dermatol. 2024;7:e48451. Xu R, Feng Y, Chen H. ChatGPT vs. Google: A Comparative Study of Search Performance and User Experience2023. Pelicioni PHS, Michell A, Santos PCRd, Schulz JS. Facilitating Access to Current, Evidence-Based Health Information for Non-English Speakers. Healthcare. 2023;11(13):1932. Emsley R. ChatGPT: these are not hallucinations – they’re fabrications and falsifications. Schizophrenia. 2023;9(1):52. Additional Declarations No competing interests reported. Supplementary Files CancerQuestionsTableEnglish.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7472185","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":524583138,"identity":"b0edb080-c9e7-47f0-80a7-3e8b78fedfce","order_by":0,"name":"Waqas Ahmed Khan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+UlEQVRIiWNgGAWjYBCDBCA2ZgYScgw8ID4bCVqMSdeS2EBIi8Hxwxs//mCwy9Ntb95sXFBzJ72/54wBw4eyw7i1nEkrluZhSC42O3OsOHnGsWe5M872GDDOOIdbi9mBHANpBgbmxG03cowP87Adzm04z2PAzNuGR8v5N8Y/fzDUJ267/wao5d/hdHmQlr/4tNzIMZPgYTgMtIXHOBloeIIB0GHMjHi02N94VmbNY3A8cRvQU8a8fYcNN545VnCw51w6Ti2S/cmbb/6oqE7cdvzwZmmeb4fl5c4kb3zwo8wapxYgMAAjFHAAn3oGTPWjYBSMglEwCtAAALJaW1HwiSOLAAAAAElFTkSuQmCC","orcid":"","institution":"Aga Khan University Hospital","correspondingAuthor":true,"prefix":"","firstName":"Waqas","middleName":"Ahmed","lastName":"Khan","suffix":""},{"id":524583139,"identity":"759781c2-aa60-4051-8302-a89c2f4ae25c","order_by":1,"name":"Misbah Soomro","email":"","orcid":"","institution":"Aga Khan University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Misbah","middleName":"","lastName":"Soomro","suffix":""},{"id":524583140,"identity":"83cbdfa0-9fd5-47cd-8d05-014607ba08b5","order_by":2,"name":"Muhammad Afzal","email":"","orcid":"","institution":"Aga Khan University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Muhammad","middleName":"","lastName":"Afzal","suffix":""},{"id":524583142,"identity":"579ce462-54f4-4093-a95f-257f1dd7a374","order_by":3,"name":"Adeeba Zaki","email":"","orcid":"","institution":"Aga Khan University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Adeeba","middleName":"","lastName":"Zaki","suffix":""}],"badges":[],"createdAt":"2025-08-27 13:38:16","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7472185/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7472185/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":92943995,"identity":"24f02f63-99bd-4e82-9fe3-4cadb325f827","added_by":"auto","created_at":"2025-10-07 12:02:48","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":170856,"visible":true,"origin":"","legend":"","description":"","filename":"KhanUrduChatGPT2025revision3.docx","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/6d09703a9fd7fb51b6f92146.docx"},{"id":92943433,"identity":"9666292c-1be3-42eb-83c9-59cd387e7c90","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5899,"visible":true,"origin":"","legend":"","description":"","filename":"aabe8cb7f1fc4fed94b2ad0ed0341683.json","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/6536e3f75ccd56597f88cc87.json"},{"id":92943996,"identity":"fca7773e-afb2-494d-b5a1-fa906b56819b","added_by":"auto","created_at":"2025-10-07 12:02:48","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17301,"visible":true,"origin":"","legend":"","description":"","filename":"CancerQuestionsTableEnglish.docx","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/a5d104ccbd21a306bf44e177.docx"},{"id":92943437,"identity":"317e60ff-eb85-4919-91c6-5dfd5f290bff","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":60692,"visible":true,"origin":"","legend":"","description":"","filename":"aabe8cb7f1fc4fed94b2ad0ed03416831enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/986dc417dbeca1956f6715e0.xml"},{"id":92943439,"identity":"636b85f4-173c-48d8-b27e-c4cd0a7429f4","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":88146,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/e0cc697bc601f85d67e3a0d6.jpeg"},{"id":92943442,"identity":"3c5400d1-69bb-421f-80c3-0095ecf512af","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"jpeg","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":156664,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/adecab71a410f627ec7a5717.jpeg"},{"id":92943999,"identity":"741e1068-391e-4fba-ae20-096a6969a4b0","added_by":"auto","created_at":"2025-10-07 12:02:48","extension":"jpeg","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":87571,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/3463aff48562baab95911a6a.jpeg"},{"id":92943441,"identity":"a95ca07d-c5bc-4727-a22b-933fc31bdd57","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":72381,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/7178a2df909172493fadde5b.jpeg"},{"id":92944519,"identity":"116765d4-4abf-46ed-82db-231744049fb9","added_by":"auto","created_at":"2025-10-07 12:10:48","extension":"jpeg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":111678,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/e0961871986256126ffb4110.jpeg"},{"id":92943443,"identity":"64407ea7-155c-454f-995a-fe4c8a4a839b","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"jpeg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1074,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/d4e2886e51ce2f25ff52e321.jpeg"},{"id":92943452,"identity":"b03f2770-d926-418d-8496-ef0413083c26","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"jpeg","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":140415,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/c5ac80ba03ba06ea1403410e.jpeg"},{"id":92944518,"identity":"fc61c4e7-4e16-4a7e-bcf3-3867f9ee1f0c","added_by":"auto","created_at":"2025-10-07 12:10:48","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":15901,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/bd733b3756df3efa34aafe65.png"},{"id":92943446,"identity":"b3b02e44-e5d1-4ddc-b8d6-9ef4f546301e","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":27985,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/0fa5ea00c0ec587894e05fc9.png"},{"id":92944003,"identity":"1b3ad504-e298-45fc-91a7-7df1dcacd2ab","added_by":"auto","created_at":"2025-10-07 12:02:48","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":15967,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/ba72ac11634552384661944e.png"},{"id":92943449,"identity":"7f6f5789-8c09-4d4d-bd9f-e50fbf57fe25","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13461,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/2ef6c55a89416215af4af8e3.png"},{"id":92944001,"identity":"1738474b-de98-4849-bf34-f34afabdd9f7","added_by":"auto","created_at":"2025-10-07 12:02:48","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":18533,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/fb647d06fb5003d37f8ecfc4.png"},{"id":92943454,"identity":"b74d5c0e-c85e-40c7-bc25-f76809486c47","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":935,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/8b33612ab7ba97a0125d2d30.png"},{"id":92943997,"identity":"8ca13dc4-6974-49be-aff3-42f30ebaeee1","added_by":"auto","created_at":"2025-10-07 12:02:48","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":31599,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/e9542b7a233ca32d187c98ca.png"},{"id":92943447,"identity":"7e0ea04d-0392-4754-8df3-b17d7fc2aa00","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"xml","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":60739,"visible":true,"origin":"","legend":"","description":"","filename":"aabe8cb7f1fc4fed94b2ad0ed03416831structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/22450af4d1240a6dbc9a1210.xml"},{"id":92943450,"identity":"f595ea3e-b05d-46a0-9c7f-dd248161e5f1","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"html","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":66193,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/543167ece3c561af19a25756.html"},{"id":92943431,"identity":"f379d08e-f0f5-4399-9f6a-026e9258f04f","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":63232,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eOverview of Methodology\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/b2a62865edd2a4d4c2da0ce2.png"},{"id":92943994,"identity":"a32b232d-9de3-472d-81f6-872e2c55047c","added_by":"auto","created_at":"2025-10-07 12:02:48","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":101160,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePrompt given before each query to ChatGPT-4\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/7bf80269fff7933a8ef7a4e1.png"},{"id":101496747,"identity":"2a3ff2e3-297d-4b54-8f65-e990aff5ba34","added_by":"auto","created_at":"2026-01-30 12:42:37","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1154603,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/6f7c9bed-fa3e-4d87-838a-1d0ad9b2bc85.pdf"},{"id":92943434,"identity":"c6022e55-e0ec-4ddd-8576-a44535a3598b","added_by":"auto","created_at":"2025-10-07 11:54:48","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":17301,"visible":true,"origin":"","legend":"","description":"","filename":"CancerQuestionsTableEnglish.docx","url":"https://assets-eu.researchsquare.com/files/rs-7472185/v1/78d6001935401e1f0153e78d.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"ChatGPT’s Ability to Answer Cancer-Related Basic Questions in Urdu: A Comparative Study with English Responses","fulltext":[{"header":"Background","content":"\u003cp\u003eChat Generative Pre-Trained Transformer (ChatGPT) is a large language model (LLM), introduced by Open AI in November 2022, that has been trained on vast datasets covering a variety of topics. Since its launch, it has rapidly gained popularity worldwide. It comprehends user questions and responds in a conversational manner that is easy to understand (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). ChatGPT\u0026rsquo;s performance has been explored in multiple medical disciplines with queries related to diseases and their management over the last two years (\u003cspan additionalcitationids=\"CR4\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). According to a study assessing ChatGPT's understanding of cirrhosis and hepatocellular carcinoma, 79% of the model's responses were correct, potentially helping not only healthcare providers but also patients in understanding their disease (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eCancer patients need comprehensive care and education about their disease, treatment options, and emotional needs (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). In the era of the internet and medical research, information on medical conditions is widely accessible but it is often not tailored to individual patient's needs. As a result, the patients are dependent entirely on information provided by physicians. The information can be influenced by various factors, including personal experiences, prevailing myths, and societal stigmas. This often results in misinformation, potentially impacting crucial decision-making (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e). The digital divide further limits access, as disparities in technology and digital literacy prevent many patients, especially in underserved areas, from obtaining personalized, reliable information (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThis study aimed to evaluate the accuracy and reliability of ChatGPT-4 in addressing questions commonly asked by cancer patients, with a special focus on examining the disparities in responses between Urdu and English. Prior studies have evaluated the ability of ChatGPT-3.5 and ChatGPT-4 to provide accurate and culturally relevant health information in other languages such as Spanish, French, and Arabic, demonstrating their effectiveness in addressing medical conditions across different linguistic contexts (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). The lower literacy rates in Pakistan contribute to the prevalence of skewed information, which is often impacted by opinions and prone to serious misinterpretation (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). Reliable medical information in Urdu is scarce, making it difficult for the Pakistani population to access dependable resources. Many are unaware of where to find accurate information for self-education, leading to a knowledge gap that can negatively affect decision-making and patient outcomes. Given these challenges, ChatGPT has shown potential in addressing clinical questions, and this study sought to assess its accuracy in providing cancer-related information in both Urdu and English.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eTo generate a thorough list of questions covering various aspects of cancer, preliminary data comprising 115 questions was collected from multiple sources, such as blogs, social media platforms, FAQs from websites of hospitals and cancer institutes. Two physicians carefully screened each question, ensuring that it was relevant to a cancer patient and easy to understand. Duplicate and unclear questions were removed, and eight new, commonly asked questions that arise in day-to-day clinical practice were added. This resulted in a final list of 68 distinct questions. These questions were translated into Urdu and English and verified for accuracy by a third physician. The methodological framework is illustrated in\u003cstrong\u003e\u0026nbsp;Figure 1\u0026nbsp;\u003c/strong\u003eand the full list of English-language cancer-related questions is provided in \u003cstrong\u003eSupplementary File.\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u0026nbsp;The six main areas of oncology covered by the questions were: 1) Basic Questions, 2) Etiology and Risk Factors, 3) Diagnostics, 4) Treatment and Prevention, 5) Side Effects and Quality of Life aspects, and 6) Outcomes and Surveillance.\u0026nbsp;\u003cbr\u003eChatGPT was given a uniform prompt before each question as shown in \u003cstrong\u003eFigure 2\u003c/strong\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026ldquo;You are an oncologist. Your patient asks you common questions related to their cancer and treatment. Please respond to the questions in the same language as the question. Respond with \u0026quot;OK\u0026quot; if the instructions are understood\u0026rdquo;.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;After ChatGPT initially confirmed by responding with \u0026quot;OK,\u0026quot; the actual questions were displayed. For a total of 136 interactions, this process was repeated for every question, starting a fresh dialogue for each question after deleting the history of prior interaction.\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u0026nbsp;The responses generated by ChatGPT were independently evaluated by two physicians to assess accuracy, with any differences in evaluation resolved by a third reviewer. The assessment criteria were classified as follows:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e1. Comprehensive:\u003c/strong\u003e The response thoroughly addressed the question.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2. Correct but Inadequate or Unfocused:\u003c/strong\u003e The response, while accurate, lacked depth, detail, or complete relevance.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3. Incorrect:\u003c/strong\u003e The response did not accurately address the question.\u003c/p\u003e\n\u003cp\u003eAdditionally, the responses in Urdu and English were compared to determine any differences in accuracy and were scored as below:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e1. Urdu Response is More Accurate:\u003c/strong\u003e The Urdu response was deemed superior.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2. Similar Accuracy:\u003c/strong\u003e The responses in both languages were judged to be equally accurate.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3. Urdu Response is Less Accurate:\u003c/strong\u003e The Urdu response was less accurate than the English version.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eA total of 68 questions were included in our study. When examining the Urdu responses, the model provided 79% comprehensive and 21% correct but inadequate or unfocused answers. None of the responses were entirely incorrect. The grading of English responses showed significantly higher scores with 97% comprehensive answers \u003cstrong\u003e(Table 1).\u003c/strong\u003e When stratified by subgroups, ChatGPT performed equally well across all groups, with only two responses not meeting expectations. For Urdu responses, ChatGPT performed best in the treatment group with 92.3% comprehensive answers, followed by the basic questions group at 83.33%. Additional scores are detailed in \u003cstrong\u003eTable 2.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWhen comparing the accuracy of Urdu responses to English responses, 72% of responses were found to be at least as accurate as in English, with 4 (5.88%) responses even better than those in English. For 19 (27.94%) questions, Urdu responses were poorer compared to English \u003cstrong\u003e(Table 3).\u003c/strong\u003e Urdu responses were overall best in the side effects and quality of life group, with 2 responses better than English and 10 equally good as in English. In the rest of the groups, English responses were graded as better overall. The details of each subgroup are given in \u003cstrong\u003eTable 4.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1. Grading responses by ChatGPT Overall\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 43.4169%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAll responses\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n=68)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 28.2132%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUrdu Answers\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003en (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 28.3699%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eEnglish Answers\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003en (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 43.4169%;\"\u003e\n \u003cp\u003e1. Comprehensive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 28.2132%;\"\u003e\n \u003cp\u003e54 (79.41 %)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 28.3699%;\"\u003e\n \u003cp\u003e66 (97.05%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 43.4169%;\"\u003e\n \u003cp\u003e2. Correct but Inadequate or Unfocused\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 28.2132%;\"\u003e\n \u003cp\u003e14 (20.58%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 28.3699%;\"\u003e\n \u003cp\u003e2 (2.94%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 43.4169%;\"\u003e\n \u003cp\u003e3. Incorrect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 28.2132%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 28.3699%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2: Grading of Responses Generated by ChatGPT-4 to Cancer Questions Categorized by Subgroup\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eResponse Scores\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUrdu Answers\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003en (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eEnglish Answers\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003en (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBasic Knowledge (n = 12)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e1. Comprehensive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e10 (83.33%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e12 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e2. Correct but Inadequate or Unfocused\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e2 (16.66%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e3. Incorrect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eEtiology/Risk Factors (n = 13)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e1. Comprehensive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e10 (76.92%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e13 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e2. Correct but Inadequate or Unfocused\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e3 (23.07%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e3. Incorrect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDiagnostics (n = 6)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e1. Comprehensive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e4 (66.66%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e6 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e2. Correct but Inadequate or Unfocused\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e2 (33.33%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e3. Incorrect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTreatment (n = 13)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e1. Comprehensive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e12 (92.30%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e13 (100%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e2. Correct but Inadequate or Unfocused\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e1 (7.69%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e3. Incorrect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSide Effects and Quality of Life Aspects (n = 15)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e1. Comprehensive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e12 (80%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e14 (93.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e2. Correct but Inadequate or Unfocused\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e3 (20%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e1 (6.66%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e3. Incorrect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOutcomes and Surveillance (n = 9)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e1. Comprehensive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e6 (66.67%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e8 (88.88%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e2. Correct but Inadequate or Unfocused\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e3 (33.33%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e1 (11.11%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.7189%;\"\u003e\n \u003cp\u003e3. Incorrect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.2019%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 29.0792%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3: Grading of Responses Comparing the Accuracy Between Urdu and English Responses Generated by ChatGPT to Cancer-Related Questions\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTotal responses n=68\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNo. of Questions\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePercentage\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e1. Urdu response is more accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e5.88%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e2. Similar accuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e66.17%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e3. Urdu response is less accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e27.94%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 4: Grading of Responses Comparing the Accuracy Between Urdu and English Responses Generated by ChatGPT to Cancer-Related Questions Categorized by Subgroup\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eResponses\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNo of Questions\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePercentage\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBasic Knowledge (n=12)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e1. Urdu response is more accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e0%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e2. Similar accuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e58.33%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e3. Urdu response is less accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e41.66%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eEtiology and Risk Factors (n=13)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e1. Urdu response is more accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e0%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e2. Similar accuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e84.61%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e3. Urdu response is less accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e15.38%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDiagnostics (n=6)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e1. Urdu response is more accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e16.66%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e2. Similar accuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e50%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e3. Urdu response is less accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e33.33%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTreatment (n= 13)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e1. Urdu response is more accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e2. Similar accuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e69.23%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e3. Urdu response is less accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e30.77%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSide Effects and Quality of Life Aspects (n=15)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e1. Urdu response is more accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e13.33%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e2. Similar accuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e66.66%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e3. Urdu response is less accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e20%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOutcomes and Surveillance (n=9)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e1. Urdu response is more accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e11.11%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e2. Similar accuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e55.55%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 53.9967%;\"\u003e\n \u003cp\u003e3. Urdu response is less accurate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 21.5334%;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.4698%;\"\u003e\n \u003cp\u003e33.33%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe development of AI tools like ChatGPT-4 has made it convenient for everyone to quickly access reliable medical information (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). Unlike traditional search engines, such as Google that require sorting through large amounts of data, ChatGPT-4 offers a more user-friendly experience (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). However, the fact that English is not everyone's first language and that most online content is written in English presents a significant challenge for non-English speakers. There is a clear need for more inclusive language support and resources to ensure that everyone can benefit from new technologies (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eOur study evaluated ChatGPT's ability to answer common cancer-related questions in Urdu and English. The model performed exceptionally well in English, with 97% of responses being comprehensive, compared to 79% in Urdu. It was particularly effective for Urdu treatment-related questions, achieving 92.3% comprehensive answers, and scored 83.33% in basic questions. The performance difference is likely due to the more extensive English datasets used during training, though certain areas might have received more focused training in Urdu as well.\u003c/p\u003e\u003cp\u003eInterestingly, 72% of the Urdu responses matched the accuracy of the English ones, with 5.88% even surpassing them. However, 28% of the Urdu responses were less accurate, pointing to a need for more tailored training data to improve overall accuracy in Urdu. While ChatGPT-4 provides adequate responses in both languages, there's room to enhance its Urdu capabilities to better serve millions of Urdu-speaking users. The lack of precise medical terms in Urdu complicates communication and indicates the need for further improvement.\u003c/p\u003e\u003cp\u003eAs technology advances, AI has the potential to improve medical care. However, to be truly effective, AI needs to work well in multiple languages, clearly communicate with patients in easy language, and provide accurate information to clinicians. Improving AI\u0026rsquo;s ability to function across different languages is crucial for ensuring equal access to accurate medical information and better outcomes for patients worldwide. Our study shows the importance of expanding non-English datasets to achieve the same level of accuracy seen in English, which is key to overcoming language barriers in healthcare.\u003c/p\u003e\u003cp\u003eWhile ChatGPT\u0026rsquo;s ease of use alleviates some challenges of traditional search engines, it is crucial to acknowledge its limitations, such as data hallucinations, when the AI confidently provides inaccurate information (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). Therefore, AI should only be used as a supplementary tool in healthcare rather than a substitute for medical advice.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eAlthough ChatGPT-4 performs well in both English and Urdu, there is still room for improvement in its Urdu responses to better serve the millions of Urdu-speaking users. To ensure all users have equal access to accurate and comprehensive medical information, AI must be multilingual and proficient across languages.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;This study did not involve human participants, identifiable data, or clinical interventions. According to the Aga Khan University Ethics Review Committee (ERC) Policy on Research Ethics Review (ORGS/008-2018, Section 1.2) and the National Bioethics Committee of Pakistan Operational Manual (2023, Section 4.2), such research is exempt from formal ethics review.\u003c/p\u003e\n\u003cp\u003eEthics approval was therefore not required, and no reference number was issued. This exemption is consistent with the policies of the Aga Khan University ERC. No informed consent was required, as the study involved only AI-generated content and all procedures were conducted solely by the authors.\u003c/p\u003e\n\u003cp\u003eThe study adhered to the ethical principles outlined in the Declaration of Helsinki.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;Not applicable. This AI-based study did not involve human participants or personal data; therefore, informed consent was not required..\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003cbr\u003eThe datasets and materials generated or analyzed during this study are available upon request. For access to the raw data analyzed in this study, please contact the corresponding author at \u003cem\[email protected].\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;The authors declare no competing interests, financial or otherwise, related to this research.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;No funding was received for the design, execution, or publication of this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe idea was conceived by WAK. Question generation, translation, and prompt creation were handled by WAK and MS, and verified by MA and AZ. Data collection and analysis were performed by WAK and MA. All authors contributed to the writing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;Not applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eRoumeliotis K, Tselikas N. ChatGPT and Open-AI Models: A Preliminary Review. Future Internet. 2023;15:192.\u003c/li\u003e\n\u003cli\u003eKasneci E, Sessler K, K\u0026uuml;chemann S, Bannert M, Dementieva D, Fischer F, et al. ChatGPT for good? On opportunities and challenges of large language models for education. Learning and Individual Differences. 2023;103:102274.\u003c/li\u003e\n\u003cli\u003eYeo YH, Samaan JS, Ng WH, Ting P-S, Trivedi H, Vipani A, et al. Assessing the performance of ChatGPT in answering questions regarding cirrhosis and hepatocellular carcinoma. Clin Mol Hepatol. 2023;29(3):721-32.\u003c/li\u003e\n\u003cli\u003eJohnson SB, King AJ, Warner EL, Aneja S, Kann BH, Bylund CL. Using ChatGPT to evaluate cancer myths and misconceptions: artificial intelligence and cancer information. JNCI Cancer Spectr. 2023;7(2):pkad015.\u003c/li\u003e\n\u003cli\u003eNedbal C, Naik N, Castellani D, Gauhar V, Geraghty R, Somani BK. ChatGPT in urology practice: revolutionizing efficiency and patient care with generative artificial intelligence. Curr Opin Urol. 2024;34(2):98-104.\u003c/li\u003e\n\u003cli\u003eYeo YH, Samaan JS, Ng WH, Ting PS, Trivedi H, Vipani A, et al. Assessing the performance of ChatGPT in answering questions regarding cirrhosis and hepatocellular carcinoma. Clin Mol Hepatol. 2023;29(3):721-32.\u003c/li\u003e\n\u003cli\u003eMcCorkle R, Ercolano E, Lazenby M, Schulman-Green D, Schilling LS, Lorig K, et al. Self-management: Enabling and empowering patients living with cancer as a chronic illness. CA Cancer J Clin. 2011;61(1):50-62.\u003c/li\u003e\n\u003cli\u003eAjith K, Sarkar S, Sethuramachandran A, Manghat S, Surendran G. Myths, beliefs, and attitude toward cancer among the family caregivers of cancer patients: A community-based, mixed-method study in rural Tamil Nadu. J Family Med Prim Care. 2023;12(2):282-8.\u003c/li\u003e\n\u003cli\u003eFareed N, Swoboda CM, Jonnalagadda P, Huerta TR. Persistent digital divide in health-related internet use among cancer survivors: findings from the Health Information National Trends Survey, 2003-2018. J Cancer Surviv. 2021;15(1):87-98.\u003c/li\u003e\n\u003cli\u003eSoto-Ch\u0026aacute;vez MJ, Bustos MM, Fern\u0026aacute;ndez-\u0026Aacute;vila DG, Mu\u0026ntilde;oz OM. Evaluation of information provided to patients by ChatGPT about chronic diseases in Spanish language. Digit Health. 2024;10:20552076231224603.\u003c/li\u003e\n\u003cli\u003eGuillen-Grima F, Guillen-Aguinaga S, Guillen-Aguinaga L, Alas-Brun R, Onambele L, Ortega W, et al. Evaluating the Efficacy of ChatGPT in Navigating the Spanish Medical Residency Entrance Examination (MIR): Promising Horizons for AI in Clinical Medicine. Clin Pract. 2023;13(6):1460-87.\u003c/li\u003e\n\u003cli\u003eKhan S, Jalees S, Jabeen Z, Khan M, Qadri RH, Adnan H, et al. Myths and Misconceptions of Breast Cancer in the Pakistani Population. Cureus. 2023;15(6):e40086.\u003c/li\u003e\n\u003cli\u003eReynolds K, Tejasvi T. Potential Use of ChatGPT in Responding to Patient Questions and Creating Patient Resources. JMIR Dermatol. 2024;7:e48451.\u003c/li\u003e\n\u003cli\u003eXu R, Feng Y, Chen H. ChatGPT vs. Google: A Comparative Study of Search Performance and User Experience2023.\u003c/li\u003e\n\u003cli\u003ePelicioni PHS, Michell A, Santos PCRd, Schulz JS. Facilitating Access to Current, Evidence-Based Health Information for Non-English Speakers. Healthcare. 2023;11(13):1932.\u003c/li\u003e\n\u003cli\u003eEmsley R. ChatGPT: these are not hallucinations \u0026ndash; they\u0026rsquo;re fabrications and falsifications. Schizophrenia. 2023;9(1):52.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"ChatGPT-4, Urdu, English, cancer, accuracy, linguistic discrepancies","lastPublishedDoi":"10.21203/rs.3.rs-7472185/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7472185/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003eChat Generative Pre-Trained Transformer (ChatGPT) has become a valuable tool since its launch in 2022 and has proved to be useful in providing easily understandable conversational responses across various topics, including medical queries. This study aims to evaluate the efficacy of ChatGPT-4 in responding to basic cancer-related questions in both Urdu and English, investigating linguistic discrepancies that may affect the reliability of AI-generated medical advice.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eWe compiled a set of 68 distinct cancer-related questions, translated into both Urdu and English, and presented them to ChatGPT-4. Responses were independently evaluated by two physicians for accuracy and comprehensiveness, with discrepancies resolved by a third reviewer. The responses in the two languages were compared for accuracy.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eChatGPT-4 provided comprehensive responses in 79% of the Urdu queries and 97% of the English queries. Accuracy assessment showed that 72% of Urdu responses were at least as accurate as their English counterparts. The treatment-related category had the highest comprehensiveness in Urdu responses at 92.3%.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e\u003cp\u003eWhile ChatGPT-4 performs proficiently in both Urdu and English, differences in the quality of responses show that there is a need for improvements in Urdu responses. Enhancing the model's training on Urdu datasets and medical terminology could bridge this existing gap and ensure equitable quality of medical information across languages.\u003c/p\u003e","manuscriptTitle":"ChatGPT’s Ability to Answer Cancer-Related Basic Questions in Urdu: A Comparative Study with English Responses","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-07 11:54:43","doi":"10.21203/rs.3.rs-7472185/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"3b0151dc-c52e-4e6d-8359-8614bcc4bb61","owner":[],"postedDate":"October 7th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-01-30T12:41:12+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-07 11:54:43","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7472185","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7472185","identity":"rs-7472185","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}