Scholarlink

Background <p>Artificial intelligence (AI), including generative AI and large language models, is entering hematology practice and research, but real-world integration depends on clinicians’ knowledge, attitudes, use, and governance rather than algorithm performance alone. Evidence specific to hematologists is scarce, and no national study in Türkiye has jointly assessed knowledge, clinical and research use, trust and governance, attitudes toward patients’ AI use, and educational needs.</p> Methods <p>We conducted a national, cross-sectional, self-administered online survey (Google Forms), reported per the CHERRIES guideline. Practicing adult and pediatric hematologists were invited through the verified member e-mail list of the Turkish Society of Hematology (a closed survey) and surveyed from 15–30 May 2026 after ethics approval. The instrument covered seven domains (participant profile; awareness/knowledge; clinical use; research/writing use; trust–risk–ethics–governance; attitudes to patients’ chatbot use; education and professional future). Categorical data are summarized as n(%) and ordinal items as median (IQR); groups were compared with non-parametric tests (effect sizes; Benjamini–Hochberg correction for exploratory analyses), with multivariable logistic and ordinal regression. Analyses used R.</p> Results <p>Of 639 invited members, 146 completed the survey (response rate 22.8%); 65.8% practiced adult hematology. Clinical AI use in the prior three months was reported by 87%, yet only 14% had received structured AI training. Generative AI for research/writing was used by 55.5%, of whom 81% had encountered fabricated references or incorrect information. Nearly half (48.6%) had entered patient data into non-approved public chatbots, most in anonymized form only (40.4%), while 8.2% entered identifiable personal or institutional information; written institutional AI policies were rarely reported or recognized (4.8% reported one, 64.4% reported none, and 30.8% were unaware). Responsible-use norms were strong (90% endorsed verification; 73% endorsed declaring AI use), and few felt professionally threatened (8.2%). Educational need was high and internally consistent (Cronbach’s α = 0.92). Higher digital literacy was independently associated with research/manuscript-writing AI use (adjusted OR 1.87 per point); an association with clinical AI use was also observed, but that model was underpowered and is reported as exploratory. Attitudes did not differ between adult and pediatric hematologists.</p> Conclusions <p>Hematologists have rapidly adopted AI while maintaining cautious, verification-oriented attitudes, but adoption is outpacing training and institutional governance. Hematology-specific policies, literacy-targeted training (especially safe generative-AI use), and supervised, institution-approved tools are needed for responsible integration.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Knowledge, attitudes, and use of artificial intelligence among adult and pediatric hematologists in Türkiye: a national cross-sectional survey

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月4日

查看全文
  • Olgu Erkin Çınar,
  • Hilmi Erdem Gözden,
  • Özlen Balta

摘要

Background

Artificial intelligence (AI), including generative AI and large language models, is entering hematology practice and research, but real-world integration depends on clinicians’ knowledge, attitudes, use, and governance rather than algorithm performance alone. Evidence specific to hematologists is scarce, and no national study in Türkiye has jointly assessed knowledge, clinical and research use, trust and governance, attitudes toward patients’ AI use, and educational needs.

Methods

We conducted a national, cross-sectional, self-administered online survey (Google Forms), reported per the CHERRIES guideline. Practicing adult and pediatric hematologists were invited through the verified member e-mail list of the Turkish Society of Hematology (a closed survey) and surveyed from 15–30 May 2026 after ethics approval. The instrument covered seven domains (participant profile; awareness/knowledge; clinical use; research/writing use; trust–risk–ethics–governance; attitudes to patients’ chatbot use; education and professional future). Categorical data are summarized as n(%) and ordinal items as median (IQR); groups were compared with non-parametric tests (effect sizes; Benjamini–Hochberg correction for exploratory analyses), with multivariable logistic and ordinal regression. Analyses used R.

Results

Of 639 invited members, 146 completed the survey (response rate 22.8%); 65.8% practiced adult hematology. Clinical AI use in the prior three months was reported by 87%, yet only 14% had received structured AI training. Generative AI for research/writing was used by 55.5%, of whom 81% had encountered fabricated references or incorrect information. Nearly half (48.6%) had entered patient data into non-approved public chatbots, most in anonymized form only (40.4%), while 8.2% entered identifiable personal or institutional information; written institutional AI policies were rarely reported or recognized (4.8% reported one, 64.4% reported none, and 30.8% were unaware). Responsible-use norms were strong (90% endorsed verification; 73% endorsed declaring AI use), and few felt professionally threatened (8.2%). Educational need was high and internally consistent (Cronbach’s α = 0.92). Higher digital literacy was independently associated with research/manuscript-writing AI use (adjusted OR 1.87 per point); an association with clinical AI use was also observed, but that model was underpowered and is reported as exploratory. Attitudes did not differ between adult and pediatric hematologists.

Conclusions

Hematologists have rapidly adopted AI while maintaining cautious, verification-oriented attitudes, but adoption is outpacing training and institutional governance. Hematology-specific policies, literacy-targeted training (especially safe generative-AI use), and supervised, institution-approved tools are needed for responsible integration.