تصورات الممرضين وتجاربهم ومعرفتهم بشأن الذكاء الاصطناعي: نتائج من دراسة استقصائية عبر الإنترنت في ألمانيا Nurses’ perceptions, experience and knowledge regarding artificial intelligence: results from a cross-sectional online survey in Germany

المجلة: BMC Nursing، المجلد: 23، العدد: 1
DOI: https://doi.org/10.1186/s12912-024-01884-2
PMID: https://pubmed.ncbi.nlm.nih.gov/38539169
تاريخ النشر: 2024-03-27

تصورات الممرضين وتجاربهم ومعرفتهم بشأن الذكاء الاصطناعي: نتائج من دراسة استقصائية عبر الإنترنت في ألمانيا

دومينيك سومر لوكاس شمدباور وفلوريان فاهل

الملخص

الخلفية تواجه التمريض ضغوطًا متزايدة بسبب التغيرات الديموغرافية ونقص العمالة الماهرة. تقدم الذكاء الاصطناعي (AI) فرصة لتخفيف العبء عن الممرضين وتقليل الضغط. إن تصور الممرضين للذكاء الاصطناعي أمر حاسم للتنفيذ الناجح. نظرًا لحالة البحث المحدودة، تهدف دراستنا إلى التحقيق في معرفة الممرضين وتصوراتهم حول الذكاء الاصطناعي. الطرق في يونيو 2023، أجرينا مسحًا عبر الإنترنت للممرضين في بافاريا، ألمانيا. تم استخدام عينة ملائمة من خلال مرافق الرعاية لاستبيان موجه نحو الاستطلاعات الحالية حول الذكاء الاصطناعي. تم إجراء تحليل البيانات بشكل وصفي، واستخدمنا تحليل القالب لتقييم الإجابات النصية الحرة. النتائج شارك 114 ممرضًا (%67.5، ở32.5 %). تظهر النتائج أن المعرفة حول الذكاء الاصطناعي محدودة، حيث يمكن وصف 25.2 % فقط بأنهم خبراء في الذكاء الاصطناعي. يرتبط الممرضون الألمان بشدة الذكاء الاصطناعي بـ (i) أجهزة الكمبيوتر والأجهزة، (ii) البرمجيات المعتمدة على البرمجة، (iii) أداة قاعدة البيانات، (iv) التعلم، و(v) اتخاذ القرارات. يعتقد ثلثا الممرضين أن الذكاء الاصطناعي يمثل فرصة. تظهر المخاوف حيث يُنظر إلى الذكاء الاصطناعي على أنه غير قابل للتحكم أو تهديد. يُعتبر موظفو الإدارة أكبر المستفيدين. الخاتمة على الرغم من عدم وجود فهم واضح لتكنولوجيا الذكاء الاصطناعي بين الممرضين، إلا أن الغالبية تعترف بالفوائد التي يمكن أن يجلبها الذكاء الاصطناعي من حيث التخفيف أو الدعم. نقترح أن يتم إعداد الممرضين بشكل أفضل للذكاء الاصطناعي في المستقبل، على سبيل المثال، من خلال التدريب وتدابير التعليم المستمر. الممرضون هم المجموعة العاملة التي تستخدم الذكاء الاصطناعي وهم حاسمون لتنفيذ الذكاء الاصطناعي في التمريض.

الكلمات الرئيسية: ممرض، الذكاء الاصطناعي، الرعاية الصحية، تعليم التمريض، ألمانيا
ساهم دومينيك سومر ولوكاس شميتباور بالتساوي في هذا العمل.
*المراسلة:
دومينيك سومر
دومينيك.سومر@th-deg.de
معهد ديغندورف للتكنولوجيا، ساحة ديتر-غورليتس 1، ديغندورف 94469، بافاريا، ألمانيا

الخلفية

لقد أدى صعود الذكاء الاصطناعي (AI)، المدفوع بتطورات مثل ChatGPT، إلى زيادة النقاش حول استخداماته المتنوعة وفوائده وتحدياته [1]. نشأ الذكاء الاصطناعي في الخمسينيات، وقد تطور مع التكنولوجيا، مما جذب اهتمامًا متزايدًا [2-4].

علم الوجود والمصطلحات

لفهم تفاصيل الذكاء الاصطناعي وحقوله الفرعية، من الضروري توضيح مصطلحات الذكاء الاصطناعي، وتعلم الآلة (ML)،
وتعلم العمق (DL). غالبًا ما تفتقر الأدبيات الحالية إلى تعريف هذه المصطلحات المختلفة، على الرغم من أن هذا الفصل ضروري وفقًا لكاستاغنو وآخرين [5] وغرازاني وآخرين [3].
  • الذكاء الاصطناعي يتعلق بالآلات التي تقوم بإجراء حسابات بحتة. إنها ليست مبنية على التفكير المنطقي وتقوم فقط بتقليد الخصائص البشرية [6].
  • تشمل التعلم الآلي والتعلم العميق، وهما فرعان من الذكاء الاصطناعي، التعلم والتكيف القائمين على البيانات، مما يعزز التنبؤات والقرارات بشكل مستقل.
الذكاء الاصطناعي، المصطلح الذي نركز عليه، يرمز إلى حدود جديدة في الذكاء الشبيه بالبشر. يجمع الذكاء الاصطناعي بين تقنيات متطورة متنوعة، بما في ذلك التعلم الآلي، ورؤية الكمبيوتر، ومعالجة اللغة الطبيعية. يصف الشيخ وآخرون [7] الذكاء الاصطناعي بأنه تقنية حديثة نسبيًا، وغالبًا ما يُنظر إليها على أنها صندوق أسود، مصممة لتنفيذ مهام تتطلب الذكاء البشري، بما في ذلك فهم الكلام، والتعرف على الأنماط، واتخاذ القرارات. عرّف المنتدى الاقتصادي العالمي الذكاء الاصطناعي بأنه “عمل من خلال الإحساس، وتفسير البيانات، والتعلم، والتفكير، والتوصية” [8]. ومع ذلك، يبدو أن هذه التصنيف معقد ومتعدد الأوجه، ويشمل العديد من الجوانب الاجتماعية والأخلاقية والقانونية والتقنية، دون أن يقدم تعريفًا عالميًا [3].
في دراستنا، نبسط الذكاء الاصطناعي كنظام ذكي يقوم بتحليل محيطه ويتصرف بشكل مستقل، لتحقيق أهداف محددة [7].

الأهمية والتطبيق في سياق التمريض

يواجه نظام الرعاية الصحية الأوروبي، الذي يتحدى من قبل شيخوخة السكان ونقص في طاقم التمريض، مطالب متزايدة وضغوطات متعلقة بالعمل [9، 10]. وبالتالي، تظهر الحلول المدعومة بالذكاء الاصطناعي كاستجابة رئيسية لهذه الضغوط [8، 10]. من الناحية الاقتصادية، يعتبر الذكاء الاصطناعي في الرعاية الصحية ذا أهمية كبيرة، حيث من المتوقع أن تصل الإنفاقات إلى 36.1 مليار دولار أمريكي بحلول عام 2025 [4].
تطبيقات الذكاء الاصطناعي، التي تتراوح من الوكلاء المحادثين في خدمة العملاء إلى الأدوات التحليلية في تحليل الصور والفيديو، تُحوّل مختلف القطاعات، بما في ذلك الرعاية الصحية، وتسهّل إنشاء المستشفيات الذكية. هذه التحولات تواجه ردود فعل عامة مختلطة، تتراوح بين التفاؤل والقلق بشأن فقدان الطابع الشخصي وفقدان الوظائف في الرعاية الصحية. التوسع السريع للذكاء الاصطناعي في الرعاية الصحية يعد بتقدم كبير، لا سيما في التحليل التنبؤي والمساعدة الافتراضية، مما يعيد تشكيل رعاية المرضى وممارسات التمريض في المستقبل القريب. إن دمج الذكاء الاصطناعي في الرعاية الصحية في طريقه…
لتخفيف الضغوط على القوى العاملة من خلال دعم طاقم التمريض في مجالات مثل التوثيق، وتحسين سير العمل، واتخاذ القرارات، مما يعزز جودة الرعاية. الهدف من تطوير تطبيقات الذكاء الاصطناعي هو دعم الممرضين في سير عملهم. بهذه الطريقة، يمكن الحفاظ على دور الممرضين، الذي هو في جوهره رعاية إنسانية. .

بحث حول تصورات الممرضين

الممرضات، وعددهن 27 مليون على مستوى العالم، هن أكبر مجموعة في مجال الرعاية الصحية ويلعبن دورًا حاسمًا في تقييم وتكامل تقنيات الذكاء الاصطناعي نظرًا لرؤاهن وتفاعلاتهن مع مجموعات مهنية مختلفة. إن الموقع المركزي للممرضات في رعاية المرضى وتنفيذ الذكاء الاصطناعي في البيئات السريرية هو مفتاح النجاح في اعتماد تقنيات الذكاء الاصطناعي. إن ارتفاع استخدام الذكاء الاصطناعي في الرعاية الصحية يزيد من الحاجة إلى أبحاث تفاعل الإنسان مع الكمبيوتر (HCI) ومشاركة الممرضات في تطوير الذكاء الاصطناعي. ومع ذلك، لا تزال تفاصيل مشاركة الممرضات بحاجة إلى تحديد. ].
البحث الدولي الحالي حول الذكاء الاصطناعي في التمريض متعدد الأبعاد، حيث يركز على الجوانب التقنية [17، 21] ويتعلق بتصورات الذكاء الاصطناعي، كما هو موضح في الجدول 1.
تشير معظم الدراسات إلى فهم محدود وتجربة مع الذكاء الاصطناعي بين الممرضين [5، 17، 22، 23]. وجد كاستاغنو وآخرون في عام 2020 [5] أن من الممرضات في بريطانيا لم يكن لديهن أي اتصال سابق بالذكاء الاصطناعي، و لم يكن يعرف الفرق بين مصطلحات الذكاء الاصطناعي. في ألمانيا، لا يفهم ثلثا الجمهور الذكاء الاصطناعي [22]. على الرغم من وجود فجوة في المعرفة، إلا أن هناك شعورًا بالتفاؤل تجاه الذكاء الاصطناعي واضح. يتفق عدد من الممرضين على أن الذكاء الاصطناعي يمكن أن يحفز التحول في التمريض من خلال تحسين الصحة، والعلاجات الشخصية، والأتمتة في الإدارة وكذلك المهام الروتينية [26]. يذكر الممرضون بشكل إيجابي أن الذكاء الاصطناعي يعزز نتائج الصحة، ويخفف العبء عن الموظفين، ويقلل التكاليف [5، 13، 15، 25-27]. وجد تشيو وآخرون [27] أن العاملين في الرعاية الصحية يرون أن الوصول إلى الذكاء الاصطناعي سهل. على الرغم من فوائد الذكاء الاصطناعي المتنوعة، لا يمكن للممرضين تقدير كيف ستتأثر أدوارهم [13]. بالإضافة إلى ذلك، يعتقد الممرضون أن الذكاء الاصطناعي ليس كافيًا، لأن التمريض يُعتبر مدفوعًا بالتفاعل البشري، وشاملاً، وقابلًا للتكيف [10، 15].
على عكس الافتراضات المحايدة والإيجابية، من بين 98 من المتخصصين في الرعاية الصحية يرون أن الذكاء الاصطناعي ‘قد يكون أكثر خطورة من الأسلحة النووية’ [5]. تشمل المخاوف عدم الوعي، وانعدام الثقة، والقلق، والقلق بشأن العلاقة مع المرضى واستبدال الوظائف [5، 8، 10، 13، 25-27]. على الرغم من فوائد الذكاء الاصطناعي، فإن التنفيذ العملي يواجه حواجز [29]. تعيق المخاوف، وعدم وجود التوعية، ونقص البيانات دمج الذكاء الاصطناعي [8، 22، 23، 25، 27، 30]. بالإضافة إلى ذلك، فإن الأخطاء تمثل مشكلة في استخدام الذكاء الاصطناعي، وتمنع لوائح حماية البيانات الممرضات من استخدام الذكاء الاصطناعي [20، 23، 27]. علاوة على ذلك، النتائج الإيجابية، والحلول المتعاطفة والشخصية،
الجدول 1 العمل المتعلق بآراء الممرضين حول الذكاء الاصطناعي
الفئات العناصر الفرعية المصادر
معرفة عدم اليقين في المصطلحات ونقص المعرفة بالذكاء الاصطناعي [٥، ١٧، ٢٢]
نقص الخبرة والتطبيق في التمريض [٥، ١٧، ١٩، ٢٣، ٢٤]
المواقف الذكاء الاصطناعي يعزز نتائج التمريض ويخفف عن الموظفين [٥، ١٣، ١٥، ٢٥، ٢٦]
الذكاء الاصطناعي يزيد الكفاءة ويقلل التكاليف [15، 26، 27]
الذكاء الاصطناعي متاح وسهل الاستخدام وودود للمستخدم [٢٧، ٢٨]
الذكاء الاصطناعي يغير المنظمات وسير العمل [5، 13، 15]
أل تغييرات القيادة [15، 29]
تؤثر الذكاء الاصطناعي على أدوار التمريض [5، 13، 15]
الذكاء الاصطناعي ليس قادرًا بما فيه الكفاية ليحل محل التفاعل البشري [10, 15]
التهديد المخيف وانعدام الثقة [5، 8، 25، 26]
عدم الوعي بالمزايا والتطبيقات [8، 29-31]
القلق بشأن علاقة المريض والسلامة [10، 13، 27]
حواجز قلق فقدان الوظيفة والأتمتة الكاملة [٢٥، ٢٧، ٣٠]
نقص التعليم والتوعية [٥، ٢٢، ٢٥]
نقص البيانات والواجهات لتدريب الذكاء الاصطناعي [8، 23، 27]
الأخطاء، النتائج غير المتوقعة وموثوقية الذكاء الاصطناعي [20, 25]
الأطر التنظيمية وحماية البيانات [٥، ٢١، ٢٣، ٢٧]
الميسرين النتائج الإيجابية تزيد من نية استخدام الذكاء الاصطناعي [1,25]
تعريف الذكاء الاصطناعي بشكل استباقي والدفاع عن المرضى [13, 21]
تطبيقات الذكاء الاصطناعي المتعاطفة والمخصصة [12, 27]
التطبيق في مراقبة الصحة، التوثيق، التواصل، ودعم القرارات السريرية [٢٠، ٢٣، ٢٤]
التدريب والمعلومات حول الذكاء الاصطناعي [٢٠، ٣٠، ٣١]
الممارسون الطبيون في تكنولوجيا المعلومات والبنية التحتية التقنية [8، 15، 20]
البحث الإضافي جميع التطبيقات والنتائج، خاصة تخفيف عبء الممرضات [1، 4، 10، 15، 17، 21، 23]
وجهة نظر التمريض، القبول، دور التمريض [12، 13، 18، 20، 27، 29]
القبول وتصميم موجه نحو المستخدم [٣١، ٣٢]
الآثار الأخلاقية والاجتماعية والقانونية [٨، ١٧، ٢٢، ٢٣، ٢٥]
القيود: الاستجابة، عينة الدراسة وإعدادات المختبر [1، 5، 15، 17، 23، 27]
صورة إيجابية صورة محايدة صورة سلبية
الممارسون في تكنولوجيا المعلومات الذين يدعمون العملية، وتعليم الذكاء الاصطناعي، والمشاركة الفعالة للممرضين في تطوير الذكاء الاصطناعي يسهلون استخدام الذكاء الاصطناعي في الرعاية الصحية. تلاحظ الممرضات تطبيقات لمراقبة الصحة، والتوثيق، والتواصل، ودعم القرار لتعزيز الذكاء الاصطناعي [20، 23، 24].
تكشف الأبحاث الحالية عن معرفة محدودة بالذكاء الاصطناعي ومواقف متنوعة بين الممرضين. تواجه الدراسات حول تصورات الممرضين للذكاء الاصطناعي تحديات مثل حجم العينات الصغيرة والتحيز في العينة. تُبرز الحاجة إلى أبحاث محدثة من خلال التغيرات التكنولوجية السريعة واختلافات بيئات الرعاية الصحية. مع تركيز معظم الأبحاث الدولية على الولايات المتحدة، فإن قابلية التطبيق على الرعاية الصحية في ألمانيا غير مؤكدة [18]. ومن ثم، تركز دراستنا على فهم تصور الذكاء الاصطناعي في سياق الرعاية الصحية الفريد في ألمانيا، بهدف توجيه تطوير الذكاء الاصطناعي الموجه نحو المستخدم في التمريض.

طرق

بين يونيو يونيو في عام 2023، أجرينا استطلاعًا عبر الإنترنت حول معرفة الممرضين وإدراكهم وتجربتهم مع الذكاء الاصطناعي في ألمانيا، بافاريا. تم تضمين الممرضين من جميع الأعمار وإعدادات الرعاية، وخاصة من التمريض geriatrie والتمريض الداخلي والتمريض السريري.

الهدف وسؤال البحث

تهدف دراستنا إلى تعزيز فهم تصورات الممرضين حول الذكاء الاصطناعي في ألمانيا، حيث يتزايد اعتماد الذكاء الاصطناعي. نهدف إلى جمع رؤى تجريبية لتوجيه تطوير الذكاء الاصطناعي الموجه نحو المستخدم في مجال التمريض. بالإضافة إلى ذلك، نهدف إلى تحديد المخاوف أو المفاهيم الخاطئة المحتملة لإبلاغ المبادرات التعليمية العملية وتقديم نظرة عامة على الوضع الحالي لتصور الذكاء الاصطناعي في التمريض لصانعي السياسات والتقنيين ومديري التمريض. الأسئلة البحثية (RQs) التي نحقق فيها هي:
RQ 1: إلى أي مدى يتم إبلاغ الممرضين في ألمانيا عن الذكاء الاصطناعي، وما هي المواقف التي يحملونها تجاه استخدام الذكاء الاصطناعي في التمريض؟
RQ 2: ما هي مجالات تطبيق الذكاء الاصطناعي التي يتم التعرف عليها بشكل أكبر من قبل الممرضين، وأي منها يرون أنها الأكثر وعدًا في مجال التمريض؟
RQ 3: بناءً على تقييم الممرضين الذاتي، ما هي المهن الصحية التي ستستفيد أكثر من دمج الذكاء الاصطناعي، ولماذا؟

أداة الدراسة

تم إبلاغ تطوير استبياننا من خلال مراجعة أدبية سردية، حيث تم جمع الأدبيات ذات الصلة من قواعد البيانات مثل PubMed وCINAHL وMedline وWeb of Science، بالإضافة إلى Google Scholar. كانت المصطلحات الرئيسية المتعلقة بوجهات نظر الممرضين وتجاربهم ومعرفتهم بالذكاء الاصطناعي هي التي وجهت البحث. كانت الدراسة الأكثر تأثيرًا في تصميم استبياننا هي دراسة BIDT [22]، مما سمح بالمقارنة لاحقًا مع بيانات السكان الأوسع. بينما تم النظر في نماذج مثل نموذج قبول التكنولوجيا (TAM) والنظرية الموحدة لقبول واستخدام التكنولوجيا (UTAUT)، بدا أن قابليتها للتطبيق محدودة بسبب تركيزها على استخدام التكنولوجيا في الحياة الواقعية [33]. بدلاً من ذلك، قمنا بتكييف نماذج مصممة خصيصًا للتمريض، كما وصفها غوهان وآخرون [34]. تم تشكيل الاستبيان النهائي، الذي تأثر بالدراسات الموجودة [5، 22، 25، 30]، من عناصر من دراسة BIDT الألمانية [22] وسوان وآخرون [26] من أجل القابلية للمقارنة. ضمنت اختبار مسبق شمل 6 ممرضات وضوح الاستبيان وموثوقيته، مما أدى إلى تعديلات طفيفة في الصياغة.
تشمل الاستبيان النهائي، المقدم في الملحق، اثني عشر سؤالًا مصنفة إلى ثلاثة أقسام: (1) البيانات الديموغرافية (فئة العمر، الجنس، مجال العمل، التعليم)، (2) المعرفة حول الذكاء الاصطناعي، و(3) التصورات حول الذكاء الاصطناعي. يستخدم الاستبيان بشكل أساسي مقاييس اسمية وترتيبية للبيانات الكمية (المغلقة). ولإثراء بياناتنا برؤى نوعية، قمنا بتضمين أسئلة مفتوحة. تم تشجيع المشاركين على وصف ما فهموه عن الذكاء الاصطناعي بكلماتهم الخاصة. في نهاية الاستبيان، تم دعوة المشاركين لتقديم تعليقات أو ملاحظات حول الذكاء الاصطناعي بصيغة نصية حرة. قدم هذا النهج المفتوح وجهات نظر متنوعة ودقيقة حول الذكاء الاصطناعي، كاشفًا عن مستويات متفاوتة من الفهم والمواقف بين الممرضين. كانت مزيج البيانات الكمية والنوعية من هذه الأسئلة فعالة في معالجة أسئلة البحث، مقدمة رؤية متوازنة حول معرفة الممرضين وتصوراتهم عن الذكاء الاصطناعي.

جمع البيانات

تم جمع البيانات لدراستنا المقطعية من خلال استبيان عبر الإنترنت تم تطويره ذاتيًا، باستخدام أداة Lime-Survey. تم توزيع الاستبيان بطريقة ملائمة.
تم أخذ عينات من خلال البريد الإلكتروني ومضاعفات إلى مرافق التمريض في بافاريا السفلى وبافاريا العليا. وافق المشاركون طواعية وبشكل مجهول، مع التزام الدراسة بالإرشادات الأخلاقية. اختار المشاركون في الدراسة الانضمام طواعية لموافقتهم بعد تلقي معلومات الدراسة. في جميع الحالات، استغرق إكمال الاستبيان أقل من خمس دقائق. كانت الأسئلة حول تجارب الممرضين السابقة مع الذكاء الاصطناعي في عملهم مستندة إلى أبعاد نتجت عن مراجعة شاملة [35]. كان تصميم الاستبيان وتوزيعه يهدفان إلى التقاط عينة واسعة وتمثيلية من وجهات نظر الممرضين حول الذكاء الاصطناعي، على الرغم من القيود مثل انحياز العينة والعناصر غير الموثقة، التي تم مناقشتها في قسم المناقشة.

تحليل البيانات

تتبع تقاريرنا عن النتائج هيكل استطلاعنا وتستند إلى المنهجيات المستخدمة في دراسة BIDT [22]. تم إجراء تحليل البيانات الكمية باستخدام IBM SPSS (الإصدار 29). قمنا بتحليل البيانات الديموغرافية (الأسئلة A.2.1-A.2.4) بما في ذلك العمر، الجنس، نوع المنشأة، ومستوى التعليم، بالإضافة إلى معرفة الذكاء الاصطناعي (A.3.1، A.3.3) واستجابات التصور (A.4.1، A.4.2)، باستخدام التكرارات والنسب المئوية.
استخدمنا الإحصائيات الثنائية، بما في ذلك ارتباط سبيرمان وCramer’s V، لتحليل تأثير العوامل الديموغرافية على تقييم الأفراد لفهمهم ومعرفتهم بالذكاء الاصطناعي. ساعدت هذه الطرق في تحديد الأنماط الرئيسية والارتباطات بين المتغيرات الديموغرافية ومعرفة/تصور الذكاء الاصطناعي. ومع ذلك، نظرًا لحجم العينة المحدود في دراستنا، يجب اعتبار هذه النتائج دالة، مما يستدعي الحاجة إلى مزيد من البحث [36]. تم تحديد مستويات الدلالة عند للأهمية و لأهمية عالية، تهدف إلى الثقة في تقليل -خطأ. تم تطبيق التحليل التبادلي واختبارات كاي-تربيع أيضًا لتقييم الفروق والعلاقات بين المتغيرات الديموغرافية والمتعلقة بالذكاء الاصطناعي.
بالإضافة إلى تحليل البيانات المغلقة، استخدمنا تحليل القالب النوعي للبيانات المفتوحة لإنشاء عروض مشتركة. لذلك، كانت الأسئلة A.3.2 و A.4.3 حقول نصية حرة تم تحليلها باستخدام تحليل قالب نوعي منظم وفقًا لمايرينغ [37] لتحديد الفئات للإجابة على سؤال البحث الخاص بنا. كانت وحدات التحليل هي الحقول النصية الحرة في الاستبيان. لتحديد نظام الفئات، اخترنا نهجًا استنتاجيًا من خلال التوجه نحو الفئات المحتملة المعروفة بالفعل بناءً على البحث في الأدبيات. في الخطوة التالية، تم البحث في وحدات النص عن المراجع وتم إجراء ترميز للمقاطع ذات الصلة. تم تحرير المراجع واستخراجها. استنادًا إلى الترميز، تم تعديل نظام الفئات الأصلي وبدأت جولة ثانية. تقدم الجدول 2 نظرة عامة على نظام الفئات النهائي. كشفت التحليل عن أربع فئات رئيسية. هذه هي (i) المعرفة
الجدول 2 فئات تحليل القالب النوعي
فئات فئات فرعية وصف العناصر
معرفة وتعريف الذكاء الاصطناعي مبني على الكمبيوتر الذكاء الاصطناعي كحل قائم على الأجهزة
مبني على البرمجة الذكاء الاصطناعي كبرنامج (لتوليد النصوص)
مستند إلى المدخلات/المعلومات/البيانات الذكاء الاصطناعي كنظام يعتمد على البيانات
التعلم الذكاء الاصطناعي كشيء يتطور
اتخاذ القرار الذكاء الاصطناعي كمساعد في اتخاذ القرار
شبيه بالبشر الذكاء الاصطناعي كتفاعل طبيعي يشبه الإنسان
شبيه بالروبوت الذكاء الاصطناعي كآلة، روبوت
تسمية مقدمي الخدمات والأدوات الذكاء الاصطناعي مثل ChatGPT أو Mindjourney
الفرص والفوائد المدركة الدعم والإغاثة في رعاية المرضى الذكاء الاصطناعي كمساعد لمهام رعاية المرضى
اكتساب المعرفة الذكاء الاصطناعي كأداة للعثور على المعلومات
المساعدة في المهام الإدارية الدعم في الكتابة/التوثيق
أنظمة المساعدة الذكاء الاصطناعي مع الروبوتات كمساعدة
القلق والعيوب مخاوف من استبدال الوظائف الذكاء الاصطناعي كتهديد لوظائف التمريض
بشكل غير شخصي الذكاء الاصطناعي كتهديد للعلاقات الاجتماعية
أخطاء المصدر الخطأ بسبب البيانات غير المكتملة
نتيجة غير كافية الذكاء الاصطناعي ليس شاملاً للإنتاج البشري
التكاليف وجهود التنفيذ الذكاء الاصطناعي مكلف في التنفيذ
الميسرين والاحتياجات المستقبلية تحسس المعلومات ضرورية للممرضين
إعادة تمويل استخدام الأل في المرافق الصحية يحتاج إلى تمويل
بيانات استخدام الذكاء الاصطناعي في المرافق الصحية يحتاج إلى بيانات
سهولة الاستخدام يجب أن يكون الذكاء الاصطناعي سهل الاستخدام وبسيطًا
حالات الاستخدام يجب أن تكون الذكاء الاصطناعي مفيدًا وقابلًا للتطبيق
وتعريف الذكاء الاصطناعي، (ii) الفرص والفوائد المتصورة، (iii) القلق والعيوب، و(iv) الميسرات والاحتياجات المستقبلية. حدد التحليل عدة فئات فرعية لكل فئة كما هو موضح في الجدول 2.

النتائج

تُعرض النتائج وفقًا لهيكل الاستطلاع، بما في ذلك (ط) البيانات السكانية، (2) فهم الذكاء الاصطناعي، (3) حالات الاستخدام، (4) التهديدات، و(5) الفوائد.

البيانات السكانية

أكمل 114 ممرضًا استبياننا. كما هو موضح في الجدول 3، كانت ثلثا المشاركين من الإناث، وكان معظم الممرضين ( كانوا تتراوح أعمارهم بين 31 و 40 عامًا. فيما يتعلق بشكل المرافق، من الممرضات يعملن في الرعاية طويلة الأمد للمرضى الداخليين، تليها المستشفيات ( ) ورعاية المرضى الخارجيين ( ). يعمل ربعهم في مرافق رعاية أخرى، مثل الرعاية المتخصصة. علاوة على ذلك، فإن عينة دراستنا تتمتع بمستويات تعليمية عالية، مثل الشهادات الجامعية ( ) وأكملوا فترات التدريب ( ).

فهم الذكاء الاصطناعي

يوضح الجدول 4 تقييم الممرضين الذاتي لمعرفتهم بالذكاء الاصطناعي، والذي تم قياسه من خلال خمسة عناصر.
الجدول 3 الخصائص العامة لمشاركي الدراسة (ذاتي النية، N=113 باستثناء الجنس)
متغير فئة ن (%)
جنس أنثى 77 (67.5 %)
ذكر 37 (32.5 %)
متنوع 0 (00.0 %)
فئة العمر تحت 21 سنة 1 (00.9 %)
21-30 سنة 23 (20.4 %)
31-40 سنة 33 (29.2 %)
41-50 سنة 24 (21.2 %)
51-60 سنة ٢٦ (٢٣.٠ ٪)
أكثر من 60 عامًا 6 (05.3 %)
نموذج المنشأة رعاية طويلة الأمد للمرضى الداخليين ٤٥ (٣٩,٨ ٪)
الرعاية الخارجية 18 (15,9 %)
التمريض (المستشفى/العيادة) 21 (18.6 %)
آخر ٢٩ (٢٥,٧ ٪)
مستوى التعليم شهادة الثانوية العامة 21 (18.6 %)
مدرسة ثانوية عليا أو مدرسة ثانوية متخصصة 16 (14.2 %)
التدريب المهني المكتمل 33 (29.2 %)
درجة جامعية (على الأقل درجة البكالوريوس) ٣٥ (٣١.٠ ٪)
آخر 6 (05.3 %)
الجدول 4 معرفة الممرضين حول الذكاء الاصطناعي (ذاتي المقصود، ترتيب الاستطلاع، N=113)
متغير فئة ن ( %)
كم تعرف عن الذكاء الاصطناعي؟ أصف نفسي كخبير في الذكاء الاصطناعي. 2 (01.8 %)
يمكنني أن أشرح جيدًا ما المقصود بذلك. ٢٦ (٢٣.٤ ٪)
أعرف تقريبًا ما المقصود بذلك. 67 (60.4 %)
أعرف المصطلح، لكن لا أعرف ما يعنيه. 14 (12.6 %)
لم أسمع بهذا المصطلح من قبل. 2 (01.8 %)
ملخص كـ ‘خبراء الذكاء الاصطناعي’ (25.2%)، ‘غير المتخصصين/-الخبراء في الذكاء الاصطناعي’ (74.8%)
وفقًا لـ Stürz وآخرون [22]، يمكن تلخيص الممرضات اللواتي يصفن أنفسهن كخبيرات في الذكاء الاصطناعي والممرضات اللواتي يمكنهن شرح معنى الذكاء الاصطناعي بشكل جيد على أنهن ‘خبراء الذكاء الاصطناعي’. باستخدام هذا التبسيط، أفادت 25.2% من الممرضات أن لديهن معرفة كبيرة بالذكاء الاصطناعي، وهن ‘خبراء الذكاء الاصطناعي’. بالمقابل، تفتقر 74.8% من الممرضات، بشكل عام، إلى فهم قوي للذكاء الاصطناعي. تم ملاحظة اختلافات بين الجنسين في المعرفة الذاتية المبلغ عنها حول الذكاء الاصطناعي: ( 15 من 35 ) من الرجال و ( 13 من 75 ) من النساء عرفن أنفسهن على أنهن ‘خبراء في الذكاء الاصطناعي’.
لفحص العلاقة بين معرفة الذكاء الاصطناعي والجنس، -اختبارات ( ) تم إجراء. أظهرت النتائج وجود ارتباط كبير بين الجنس ومعرفة الذكاء الاصطناعي ( قيمة )، مع ارتباط كرامر V بقيمة 0.335. بالإضافة إلى ذلك، نوع مرفق الرعاية الصحية ( قيمة كرايمر ) ومستوى التعليم ( قيمة معامل ارتباط سبيرمان كما أظهرت تأثيرات كبيرة على معرفة الذكاء الاصطناعي. في الوقت نفسه، لم يكن العمر عاملاً مهماً. نظراً لانخفاض عدد الخلايا في اختباراتنا، يجب تفسير هذه النتائج بحذر، مما يشير إلى الحاجة إلى مزيد من البحث مع عينات أكبر.
بالإضافة إلى ذلك، تقدم تعريفات المشاركين في الدراسة للذكاء الاصطناعي فهمًا متعدد الأبعاد يتجاوز الافتراضات الكمية، كما تم التأكيد عليه في الاقتباسات التالية. يُنظر إلى الذكاء الاصطناعي كأداة ديناميكية لديها القدرة على التعلم وتوسيع قدراتها (النص 1، الفقرة 55). تكشف التحليلات النوعية أن الممرضين يربطون الذكاء الاصطناعي بـ (1) الحواسيب والأجهزة، (2) البرمجيات، (3) الخصائص البشرية التي تتجلى من خلال معالجة اللغة الطبيعية، (4) أداة مدفوعة بالبيانات قادرة على (5) التعلم، و(6) اتخاذ القرارات. لقد أدركنا وجود ثنائية في وجهات نظر الممرضين، حيث يصفون الذكاء الاصطناعي بخصائص بشرية وكأداة أكثر ميكانيكية تشبه الآلات. غالبًا ما يرى المشاركون الذكاء الاصطناعي كأداة تحاكي التفاعلات البشرية أو تكرر السلوكيات البشرية. تتأثر هذه الرؤية أيضًا بارتباط الذكاء الاصطناعي بالروبوتات والآلات، مع الاعتراف بأن الذكاء الاصطناعي غالبًا ما يتم دمجه كبرمجيات في الأنظمة الروبوتية. كما يذكر الممرضون مزودي الذكاء الاصطناعي، مثل OpenAI (النص 1، الفقرة 48).

“الذكاء الاصطناعي هو آلة لديها ممارسات لتنفيذها والتعلم من البيانات” (النص 1، الفقرة 34)

“تم برمجة الذكاء الاصطناعي بواسطة البشر ويهدف إلى استخدامه في مجالات متعددة لاتخاذ القرارات، وأداء الأفعال، وأداء المهام اليومية البشرية ضمن برمجته، ولتحديد وحل المشكلات.” (النص 1، الفقرة 36)
“الذكاء الاصطناعي يعني استخدام الروبوتات التي يمكن أن تتفاعل مع الحركات أو الكلام.” (النص 1، الفقرة 83)
“الذكاء الاصطناعي […] يمكن أن يتصرف بشكل مشابه للإنسان، مثل كتابة النصوص، التصرف كإنسان، إلخ.” (النص 1، الفقرة 31)
“الذكاء الاصطناعي هو قدرة الآلة على تقليد القدرات البشرية.” (النص 1، الفقرة 72)
“يمكن للذكاء الاصطناعي أن يقرر ويتعلم بمفرده، مع الأخذ في الاعتبار العلامات التي تم تعليمه إياها.” (النص 1، الفقرة 15)
علاوة على ذلك، يدرك الممرضون أن الذكاء الاصطناعي ليس حلاً جاهزاً للاستخدام، بل يتطلب معالجة البيانات وإعدادها (النص 1، الفقرة 11). يفهم الممرضون الدور الحاسم لجودة البيانات في تشكيل الذكاء الاصطناعي كأداة لصنع القرار (النص 1، الفقرة 55).

حالات الاستخدام والتطبيقات

في مجالات تطبيق الذكاء الاصطناعي في التمريض (الجدول 5)، فإن الممرضين هم الأكثر دراية بمراقبة المرضى (55.7%) وتخطيط المسارات ( بالإضافة إلى توثيق التمريض المدعوم بالذكاء الاصطناعي ). يعرف الممرضون أقل شيء عن مجالات الذكاء الاصطناعي في توقع الرعاية ( تشخيص التمريض ) وإدارة الجروح ( ).
تكشف بياناتنا النوعية أن فائدة الذكاء الاصطناعي تمتد إلى مراقبة المرضى، والمهام التمريضية الروتينية، وتقييم ‘تأثيرات وتفاعلات توافق الأدوية’ (النص 2، الفقرة 124)، وتعزيز ‘رعاية المرضى’ (النص 2، الفقرة 127)، والمساعدة في ‘رعاية سلس البول الذكية’ (النص 2، الفقرة 130). بالإضافة إلى ذلك، يدرك الممرضون الدور الداعم للذكاء الاصطناعي في الوظائف الإدارية، مثل المساعدة في ‘تخطيط الموظفين’ (النص 2، الفقرة 125) و’جدولة الخدمات’ (النص 2، الفقرة 129). علاوة على ذلك، يتمتع الذكاء الاصطناعي بإمكانية في ‘أنشطة التنظيف’ (النص 2، الفقرة 134) من خلال الروبوتات التنظيفية المعتمدة على الذكاء الاصطناعي. كما تم تسليط الضوء على المساعدة التقنية المدعومة بالذكاء الاصطناعي، مثل الهيكليات الخارجية النشطة (النص 2، الفقرة 128).
الجدول 5 مجالات تطبيق الذكاء الاصطناعي في التمريض (مقصد ذاتي، ترتيب تنازلي، N=88)
متغير فئة ن (%)
ما هي مجالات التطبيق ما تعرفه عن الذكاء الاصطناعي في التمريض؟ مراقبة المرضى (مثل العلامات الحيوية، النوم) 49 (55.7 %)
تخطيط المسار 42 (47.7 %)
توثيق التمريض ٣٨ (٤٣.٢ ٪)
تنبؤ رعاية المرضى (مثل اكتشاف السقوط) 34 (38.6 %)
إجراء تشخيصات التمريض 28 (31.8 %)
إدارة الجروح 19 (21.6 %)
وفقًا لسيبرت وآخرين [23]، فإن جميع تطبيقات الذكاء الاصطناعي المعروضة موجودة ومفيدة للتمريض.
الجدول 6 تصور الممرضين حول الذكاء الاصطناعي (ذاتي النية، N=103)
متغير فئة ن (%)
هل ترى الذكاء الاصطناعي في التمريض كفرصة أكثر أم كتهديد أكثر؟ حصريًا كفرصة 14 (13.7 %)
بل كفرصة 53 (52.0 %)
بل كخطر 13 (12.7 %)
حصريًا كخطر 1 (01.0 %)
لا أعرف، لا أستطيع الحكم 22 (21.6 %)
ملخص كـ ‘إدراك إيجابي للذكاء الاصطناعي’ (65.7%) ملخص كـ ‘تصور سلبي للذكاء الاصطناعي’

الفرص والإدراك الإيجابي للذكاء الاصطناعي

تظهر الجدول 6 أن ما يقرب من ثلثي الأشخاص (65.7٪) يرون الذكاء الاصطناعي بشكل إيجابي كفرصة للتمريض، بينما يرى جزء أصغر (13.7٪) الذكاء الاصطناعي بشكل سلبي كتهديد أو خطر. بالإضافة إلى ذلك، لم يتمكن % من المستجيبين من تصنيف الذكاء الاصطناعي بأي شكل، مما يشير إلى أنهم لم يتمكنوا من الحكم بسبب نقص المعرفة.
تؤكد البيانات النوعية على وجهة نظر متفائلة بشأن إمكانيات الذكاء الاصطناعي، معترفة بقدرته على دعم وتحويل التمريض. كما يتضح من الاقتباسات، أعرب المستجيبون عن أن الذكاء الاصطناعي يمكن أن يكون مساعدة قيمة في التمريض، حيث يساعد في الأنشطة اليومية، ويعمل على تحسين عمليات العمل، ويدعم جمع البيانات وتفسيرها، مثل تحليل المشاعر أو المعايير الحيوية. كما تم الاعتراف بقدرة الذكاء الاصطناعي على تعزيز التفاعل بين الإنسان والكمبيوتر، خاصة من خلال قدرات التحكم الصوتي المعتمدة على الذكاء الاصطناعي.
“في مجال الذكاء الاصطناعي، نحن في بداية التغييرات في التمريض.” (النص 4، الفقرة 171)
لقد كانت الذكاء الاصطناعي جزءًا من حياتنا اليومية لفترة طويلة. سواء كان ذلك مثل سيري [أو] أليكسا، يحاول الذكاء الاصطناعي جعل الأمور اليومية أسهل أو تحسينها. ومع ذلك، هذه مجرد مجالات ثانوية بسيطة، يمكن للذكاء الاصطناعي التعرف على المشاعر وتسجيل المعايير الحيوية.
الذكاء الاصطناعي يعني تحسين ودعم العمل وتخفيف العبء عن الموظفين.
علاوة على ذلك، يقدّر الممرضون دور الذكاء الاصطناعي في اكتساب المعرفة، بما في ذلك استخدام الذكاء الاصطناعي للإجابة على الأسئلة، وتقديم الإرشادات، والمساعدة في التوثيق. التنبؤات التي يقدمها الذكاء الاصطناعي
تكون القدرات مفيدة، مما يسمح بالكشف المبكر عن احتياجات الرعاية والمخاطر، مثل الإنتان من الرعاية والمعايير المخبرية (النص 4، الفقرة 170، النص 3، الفقرة 152).

التهديدات والسلبية في تصور الذكاء الاصطناعي

بينما تقدم الذكاء الاصطناعي العديد من الفرص، مثل تبسيط المهام وجعل الحياة أسهل، يدرك الممرضون التهديدات والمخاطر المحتملة. تستند التصورات السلبية إلى مخاوف عملية، مثل خطر تعطل الذكاء الاصطناعي مما يؤدي إلى أخطاء. يخشى بعض الممرضين أن يحل الذكاء الاصطناعي محل المهام البشرية. وهذا يترافق مع الفهم بأن التمريض يعتمد أساسًا على العلاقات الاجتماعية بين الممرضين والمرضى، وهو بعد قد يجعل الذكاء الاصطناعي غير شخصي. تشمل المخاوف الأخرى فقدان السيطرة على الذكاء الاصطناعي، مما يمكن أن يسبب مشاعر العجز والتكاليف العالية المرتبطة بتنفيذ الذكاء الاصطناعي وصيانته.
“أنا خائف من غباء الذكاء الاصطناعي، نظام يقدم نتائج بعد التدريب غير المفهومة لأن المصادر مفقودة.” (النص 1، الفقرة 23) “الذكاء الاصطناعي يهدد وظائفنا ويجعل الرعاية أكثر عدم شخصية.” (النص 1، الفقرة 29)
“إنها أيضًا مسألة إعادة تمويل. للأسف، غالبًا ما يكون من غير الممكن لأسباب تتعلق بالتكاليف تنفيذ ابتكارات ذات مغزى أو شراؤها.” (النص 4، الفقرة 171)

المستفيدون ومجموعات المستخدمين المستفيدة

وفقًا للنتائج الكمية، توضح الجدول 7 وجهات نظر الممرضات حول أي عيادة خارجية
الجدول 7 أكثر الفئات استفادة من الذكاء الاصطناعي في التمريض (حسب النية الذاتية، ترتيب تنازلي، N=95)
متغير فئة ن (%)
أي مجموعة من الناس تستفيد أكثر من استخدام الذكاء الاصطناعي؟ موظفو الإدارة والإدارة ٤٧ (٤٩.٥ ٪)
الموظفين في التمريض والدعم 24 (25.3 %)
المرضى الذين يحتاجون إلى رعاية 17 (17.9 %)
الخدمات الاجتماعية والدعم 7 (07.4 %)
ويمكن أن تستفيد مجموعات الرعاية الداخلية بشكل كبير من الذكاء الاصطناعي. يعتقد ما يقرب من نصف الممرضات أن موظفي الإدارة والإدارة يمكن أن تدعم الذكاء الاصطناعي بشكل أكبر، تليه التمريض وموظفو الدعم )، المرضى الذين يحتاجون إلى الرعاية ( ) ، والخدمات الاجتماعية والدعم ( ).
بالإضافة إلى ذلك، حددت الممرضات عدة أطراف قد تستفيد من الذكاء الاصطناعي من خلال بيانات نوعية: (1) المستوى الميكروي – الأطباء ومشغلي المستشفيات، (2) المستوى الميزو – ممرضات الأسرة، والخدمات الاجتماعية، وموظفي الخدمة، و(3) المستوى الكلي – شركات التأمين الصحي والرعاية، والحكومة وصانعي السياسات، وعلماء التمريض. تشمل المزايا عمليات أكثر كفاءة مثل التحقق من الفواتير لشركات التأمين، والكشف المبكر عن المخاطر للحد من تكاليف الرعاية الصحية، وتوفير التوظيف لمشغلي المستشفيات وأرباب العمل، واحتياجات الرعاية التنبؤية وتطوير التقييم في علم التمريض، ودعم المشاركة الاجتماعية في الخدمات الاجتماعية.

احتياجات لمزيد من تنفيذ الذكاء الاصطناعي

تحدد بياناتنا النوعية العوامل الميسرة المحتملة، والعوائق، والمتطلبات اللازمة لتنفيذ الذكاء الاصطناعي بنجاح في التمريض. الحاجة الأساسية هي توفر بيانات موثوقة، منظمة، وكاملة (النص 1، الفقرة 55). بدون هذه البيانات، قد تكون نتائج الذكاء الاصطناعي غير مرضية (النص 1، الاقتباس 41). كلما كانت البيانات أكثر سهولة وموثوقية، كانت القرارات التي يتخذها نظام الذكاء الاصطناعي أكثر قوة. بالإضافة إلى اعتبارات البيانات، فإن المشاركة البشرية ضرورية في تشكيل الذكاء الاصطناعي. وهذا يعني المشاركة النشطة في تصميمه وتنظيمه، والحفاظ على علاقات المرضى، وتعزيز الثقة من خلال التوعية والتعليم حول الذكاء الاصطناعي (النص 4، الفقرتان 165، 166). من الضروري أيضًا معالجة المخاوف التي تؤدي إلى مشاعر العجز، مثل القلق بشأن السيطرة وإمكانية تصحيح النتائج غير المتوقعة (النص 4، الفقرة 164).
“يجب إطعام الذكاء الاصطناعي […] بالمدخلات للحصول على مخرجات متCorresponding.” (النص 1، الفقرة 11)
ستستفيد جميع فئات المجتمع فقط إذا احتضنوا هذا التغيير بنشاط. من المحتمل أن يكون هذا هو أكبر تحدٍ، إلى جانب إعادة التمويل المعقولة.
(النص 4، الفقرة 165)
“يجب تنظيم الذكاء الاصطناعي، ولا ينبغي لأحد أن يفقد وظيفته بسبب الذكاء الاصطناعي، ويجب الحفاظ على العلاقة مع المرضى.” (النص 4، الفقرة 159)
“يجب أن يكون بإمكان البشر الوصول إلى و”إعادة برمجة” في أي وقت.” (النص 4، الفقرة 164)
“الذكاء الاصطناعي في تخطيط المسار يعتمد حاليًا على العديد من المعايير ليعمل بشكل صحيح. يجب دائمًا تطوير الذكاء الاصطناعي لتقديم أقصى دعم للممرض.” (النص 4، الفقرة 167)
كما تشير الاقتباسات، فإن تجربة المستخدم هي عامل حاسم آخر. يجب أن تكون حلول الذكاء الاصطناعي سهلة الاستخدام، وسهلة، وتوفر الوقت (النص 4، الفقرة 161). علاوة على ذلك، تظهر الاقتباسات أن حلول الذكاء الاصطناعي لا تزال غير مرضية بما فيه الكفاية، مما يؤدي إلى إمكانية تحسينها. وأخيرًا، يجب ألا تُغفل الاعتبارات المتعلقة بالتكاليف والجهد المطلوبين للتنفيذ (النص 4، الفقرة 165).

نقاش

للإجابة على أسئلة البحث، تم إجراء مسح عبر الإنترنت عابر للقطاعات بين الممرضات في بافاريا، ألمانيا، في يونيو 2023، باستخدام عينة ملائمة. لقد قدم نهجنا رؤى شاملة حول وعي الممرضات ومواقفهن تجاه الذكاء الاصطناعي.

القيود

تقدم دراستنا رؤى قيمة حول إدراك الممرضين للذكاء الاصطناعي ولكنها تخضع لعدة قيود. أولاً، مفهوم ‘الإدراك’ نفسه غير محدد بوضوح، مما قد يشمل جوانب الاستخدام والقبول، مما يبرز الحاجة إلى أبحاث ميدانية. قد لا تعكس منهجية العينة المريحة والتركيز على بافاريا بدقة السكان الأوسع من الممرضين، مما قد يقدم تحيزًا. يحد حجم العينة الصغيرة المكونة من 114 مشاركًا من القوة الإحصائية للدراسة، خاصة عند إجراء تحليلات متعددة المتغيرات الشاملة. استخدمنا اختبارات ثنائية المتغيرات وحسبنا مربع كاي. ومع ذلك، كان العدد المتوقع في بعض الخلايا أقل من 5 في من الخلايا، مما يحد من موثوقيتنا ويشير إلى الحاجة إلى مزيد من البحث مع عينات أكبر [36]. باختصار، يعيق قيد حجم العينة الصغيرة قدرتنا على استكشاف العلاقات المتبادلة وتكون التعميمات محدودة.
بالإضافة إلى ذلك، قد يؤثر اعتماد الدراسة على الأدوات الموجودة مثل دراسة BIDT [22]، مع تجاهل النماذج الأوسع مثل TAM أو UTAUT على عمق الاستكشاف النظري. قد تؤدي قصر الاستبيان إلى تقييم غير كافٍ لخصائص الممرضين الديموغرافية وتجاربهم مع الذكاء الاصطناعي. علاوة على ذلك، فإن معايير الشمول والاستبعاد محددة، تركز على الممرضين الرسميين. لم تأخذ دراستنا في الاعتبار المهنيين الصحيين الآخرين، مما يفوت رؤى مقارنة حول إدراك الذكاء الاصطناعي عبر أدوار الرعاية الصحية. قد لا تلتقط تصميمنا العرضي الإدراكات المتطورة للذكاء الاصطناعي، خاصة في سياق السرد الإعلامي المتغير بسرعة. يجب أن تأخذ الأبحاث المستقبلية في الاعتبار أحجام عينات أكبر وأكثر تنوعًا، وتدمج أطر نظرية شاملة وتتبنى طرقًا طولية.

المعرفة والمواقف

في معالجة السؤال البحثي 1 حول معرفة الممرضين ومواقفهم تجاه الذكاء الاصطناعي، وجدت دراستنا أن فقط من الممرضين لديهم معرفة كبيرة بالذكاء الاصطناعي، مع وجود الغالبية ( ) لديهم فهم محدود. يتماشى هذا مع الأبحاث الحالية التي تشير إلى نقص عام في الألفة مع الذكاء الاصطناعي في الرعاية الصحية [17، 19]. الحاجة إلى زيادة الوعي والتدريب على الذكاء الاصطناعي واضحة [1،20]. علاوة على ذلك، أكدنا عدم شمولية تعريف الذكاء الاصطناعي من قبل الممرضين [7]. غالبًا ما يدرك الممرضون الذكاء الاصطناعي من الناحية التقنية، مما قد يسهم في الشك ويعيق القبول. تظهر دراستنا أيضًا أن بعض المشاركين يساوون الذكاء الاصطناعي بالبشر، مما يؤدي إلى الحاجة إلى مزيد من التوعية حول الذكاء الاصطناعي.
على الرغم من أن العديد من موظفي التمريض يعرفون القليل جدًا عن الذكاء الاصطناعي، إلا أن ثلثيهم ينظرون إلى الذكاء الاصطناعي بتفاؤل، مما يشير إلى انفتاحهم على تنفيذ الذكاء الاصطناعي على الرغم من المخاوف بشأن علاقات المرضى والأخطاء والتكاليف وأمان الوظائف. هذا أمر مفاجئ ويحتاج إلى مزيد من البحث، حيث يمكن أن يكون نقص المعرفة أحد الأسباب المستمرة للإدراك السلبي والقبول المنخفض [34]. علاوة على ذلك، تدعو بياناتنا المفتوحة إلى تطبيقات الذكاء الاصطناعي الوظيفية والآمنة وتؤكد على أهمية التواصل الواضح ومشاركة الممرضين في تطوير الذكاء الاصطناعي لتعزيز الفهم وتخفيف المخاوف. يتم مناقشة هذا أيضًا في الأدبيات الحالية كأمر مهم [1،17،19،20]. بالإضافة إلى ذلك، لاحظت الدراسات حول إدارة التغيير في التمريض معدل نجاح أعلى وقبول للتقنيات الجديدة إذا كان الممرضون مشاركين وكانوا جزءًا من عملية اتخاذ القرار [38].

مجالات التطبيق

بالنسبة للسؤال البحثي 2، فإن أكثر تطبيقات الذكاء الاصطناعي شهرة هي (i) مراقبة المرضى (55.7 %)، (ii) تخطيط المسارات (47.7 %)، و(iii) توثيق التمريض ( ). تؤكد بياناتنا النوعية هذه الإمكانيات. ومع ذلك، فإن البيانات المفتوحة ونقص المعرفة تجعلنا نفترض أن استخدام الذكاء الاصطناعي في رعاية التمريض في بافاريا لم يحقق حتى الآن تقدمًا كبيرًا
وأن الممرضين لم يتعرضوا، بالتالي، تقريبًا للذكاء الاصطناعي في سياق العمل.
تسلط دراسات الوقت في التمريض الضوء على الوقت الكبير الذي يقضى في المهام الروتينية والإدارية، مما يدعم استخدام الذكاء الاصطناعي في مجالات مثل المراقبة والتوثيق [39، 40]. تركز الأبحاث الحالية، بما في ذلك أعمال مثل Cho et al. [41]، على حلول الذكاء الاصطناعي لمراقبة المرضى ودعم اتخاذ القرار السريري. تتماشى نتائج استبياننا مع الأدبيات الوطنية [23]، مما يشير إلى أن الممرضين على دراية بتطبيقات الذكاء الاصطناعي في المهام الإدارية، وهو اتجاه من المتوقع أن ينمو في العقد القادم. يمكن أن يؤدي استخدام الذكاء الاصطناعي لتبسيط هذه العمليات إلى تقليل العبء البيروقراطي.

النتائج والفوائد

في معالجة السؤال البحثي 3 حول النتائج المتوقعة وفوائد الذكاء الاصطناعي، تشير دراستنا إلى إدراكات إيجابية في الغالب. يتوقع الممرضون أن يخفف الذكاء الاصطناعي الأعباء المتزايدة في التمريض، خاصة في ظل نقص الموظفين الحالي وزيادة الأعباء [9، 10]. كما يتصورها مستجيبونا، فإن المستفيدين الرئيسيين هم الإدارة والإدارة ( ) وموظفو التمريض ( ). تؤكد هذه النتيجة على الأدوار المتطورة في الرعاية الصحية المتكاملة بالذكاء الاصطناعي وتتوافق مع Swan et al. [26].
تشير نتائجنا إلى بدء تنفيذ الذكاء الاصطناعي في المجالات الإدارية، دعمًا لتوصيات Chew et al. [27]، والتوسع إلى الرعاية المباشرة. ومع ذلك، فإن المخاوف بشأن الذكاء الاصطناعي التي تم التعبير عنها في بعض الردود المفتوحة تسلط الضوء على الحاجة إلى زيادة الوعي والتعليم عبر جميع مجموعات المستخدمين المحتملة لتخفيف المخاوف وتعزيز فهم شامل لدور الذكاء الاصطناعي في التمريض.

الآثار على الممرضين والفنيين وصانعي السياسات

تؤدي نتائج دراستنا إلى عدة آثار وتوصيات حاسمة لممارسة التمريض وصانعي السياسات وإدارة مرافق الرعاية الصحية:
(i) الحملات التعليمية: زيادة النقاش العام وتقديم تدريب متخصص وتعليم إضافي لموظفي التمريض لتعزيز فهمهم للذكاء الاصطناعي.
(ii) سياسات التعويض والأطر القانونية: وضع سياسات تعويض واضحة وأطر قانونية تدعم استخدام الذكاء الاصطناعي في التمريض.
(iii) الاستثمار في الذكاء الاصطناعي في التمريض: تخصيص الموارد لتطوير وتنفيذ الذكاء الاصطناعي في التمريض، مع ضمان سهولة الاستخدام.
(iv) مشاركة الممرضين في تطوير الذكاء الاصطناعي: إشراك الممرضين في تطوير الذكاء الاصطناعي.
(v) توفير الدعم الفني: نشر دعم فني إضافي وموظفي برمجة لمساعدة الممرضين في استخدام أدوات الذكاء الاصطناعي في مرافق الرعاية الصحية.
(vi) أكاديميات التمريض: دمج تعليم الذكاء الاصطناعي في المناهج الدراسية للتمريض لإعداد الممرضين المستقبليين للتقدم التكنولوجي في الرعاية الصحية.
نؤكد على أن صانعي السياسات يجب أن يعطوا الأولوية لتطوير وتنفيذ (i) حملات تعليمية مستهدفة، (ii) سياسات تعويض وأطر قانونية تدعم دمج الذكاء الاصطناعي في التمريض، و(iii) الاستثمار في تطوير وتنفيذ الذكاء الاصطناعي في التمريض المخصص للمهنيين في التمريض. يجب أن تشمل الاستثمارات تمويل أبحاث الذكاء الاصطناعي، ووضع إرشادات واضحة لتطبيق الذكاء الاصطناعي في الرعاية الصحية، وضمان أن تكون الاعتبارات الأخلاقية مركزية في نشر الذكاء الاصطناعي في بيئات التمريض. يجب أن تساعد هذه الإرشادات الواضحة في القضاء على عدم الثقة والإدراكات السلبية، مثل رؤية عدم الأمان والتهديدات لرعاية المرضى كما تم تصورها نادرًا من خلال بياناتنا المفتوحة. يجب أن تركز الاستثمارات والتطورات على الذكاء الاصطناعي لمراقبة المرضى (بما في ذلك تخصيص المرضى واكتشاف السقوط)، وتخطيط المسارات، وتوثيق التمريض، حيث يعرفها الممرضون بالفعل ويعتبرونها مجالات واعدة. يمكن أن يكون موظفو الإدارة وإدارة التمريض هم أصحاب المصلحة الأكثر استفادة، يجب على مطوري الذكاء الاصطناعي أن يسألوا أنفسهم كيف يمكن دعم الممرضين.
علاوة على ذلك، يجب على الفنيين (iv) إشراك الممرضين في تطوير تطبيقات الذكاء الاصطناعي من خلال، على سبيل المثال، مجموعات التركيز. نوصي (v) بأن يحصل الممرضون على دعم فني في استخدام أدوات الذكاء الاصطناعي بفعالية وتوافقها مع احتياجاتهم. يجب أن تدمج المناهج الدراسية للممرضين (vi) تعليم الذكاء الاصطناعي وتعد الممرضين للتقدم التكنولوجي.

الخاتمة

استنتاجات إدراك الممرضين للذكاء الاصطناعي

باختصار، 74.8 % من الممرضين الذين شملهم الاستطلاع يفتقرون إلى فهم قوي للذكاء الاصطناعي. يعتبر الذكاء الاصطناعي فرصة إيجابية. استنتاجنا الرئيسي بشأن استخدام الذكاء الاصطناعي هو أن الممرضين يعرفون مراقبة المرضى، تخطيط المسارات، وتوثيق التمريض كحقول تطبيق. مع الضغط المتزايد على التمريض، من المتوقع أن يلعب الذكاء الاصطناعي دورًا حيويًا في تخفيف عبء العاملين في الرعاية الصحية وتلبية الطلبات المتزايدة على التمريض في غرب أوروبا [10، 23]. المجموعات المستفيدة بشكل رئيسي هي موظفو الإدارة والإدارة ( ) بالإضافة إلى الممرضين وموظفي الدعم الخاص بهم (25.3%).
على الرغم من أن دراستنا أكثر دلالة من كونها حاسمة، فإن النتيجة التي تشير إلى أن فقط من المستجيبين يرون أنفسهم كخبراء في الذكاء الاصطناعي تبرز الحاجة إلى زيادة التعليم والتوعية بالذكاء الاصطناعي في التمريض. ردود الفعل المختلطة، بما في ذلك الآراء الإيجابية إلى جانب المخاوف النادرة بشأن أمان الوظائف وتغيرات رعاية المرضى، تؤكد أكثر
أهمية مشاركة الممرضين وتعليمهم في تطوير الذكاء الاصطناعي.
نتوقع زيادة في المعرفة والقبول بالذكاء الاصطناعي في التمريض خلال العقد المقبل، مما يشير إلى مستقبل واعد لدمج الذكاء الاصطناعي. تشير نتائجنا إلى أن المهام الإدارية والروتينية هي مجالات رئيسية للذكاء الاصطناعي، مما يجعلها نقاط انطلاق استراتيجية لتعظيم إمكانيات الذكاء الاصطناعي. بالإضافة إلى ذلك، تكشف الردود المفتوحة عن الحاجة إلى التنظيم، والوصول، والتركيز على تخفيف أعباء الممرضين. نؤكد على أهمية تطوير ذكاء اصطناعي سهل الاستخدام يوفر الوقت في المهام المتكررة والإدارية، مما يسمح للممرضين بالتركيز على رعاية المرضى المباشرة.

البحث الإضافي

تتطلب قيود دراستنا والفجوة البحثية الوطنية مزيدًا من التحقيقات المتعمقة في فهم الممرضين وقلقهم وتوقعاتهم بشأن الذكاء الاصطناعي. يجب أن تبني الأبحاث اللاحقة على نتائجنا لتعميق فهم دور الذكاء الاصطناعي في الرعاية الصحية. تشمل المجالات الأساسية للبحث المستقبلي تطبيق تقنيات جديدة في بيئات واقعية ودمج المعرفة بالذكاء الاصطناعي في التعليم التمريضي، لضمان صلتها واستدامتها [42]. يجب أن تركز الدراسات على دمج المعرفة بالذكاء الاصطناعي في تدريب الممرضين، خاصة بالنظر إلى مستويات معرفة الممرضين المتفاوتة. يجب تمكين الممرضين من استخدام التقنيات الجديدة، ويجب أن تأخذ الحلول في الاعتبار الاحتياجات الواقعية ويجب أن تدمج في الممارسة دون عتبة منخفضة، كما يشير مشروعنا ‘الغابة الذكية – عيادات 5G’.
يجب أن تكون التحليلات الإضافية مع عينات دراسية أكثر شمولاً وطويلة الأمد لتشمل بيانات أوسع ومقارنة تصورات الذكاء الاصطناعي عبر المهن الصحية. يجب إجراء تحليل متعدد المتغيرات مع بيانات جديدة، للإجابة على سؤال ما هي العوامل المؤثرة في فهم جيد للذكاء الاصطناعي والموقف الإيجابي تجاهه. بالإضافة إلى ذلك، فإن تحليل الاختلافات المحتملة في التصورات، مثل بين الممرضين والأطباء، أمر حيوي. علاوة على ذلك، فإن استكشاف التنفيذ الواقعي للذكاء الاصطناعي وقبول المستخدمين له أمران حاسمان، خاصة في ضوء نتائجنا التي تشير إلى استعداد الممرضين لتبني الذكاء الاصطناعي كفرصة بدلاً من خطر.

معلومات إضافية

تحتوي النسخة الإلكترونية على مواد إضافية متاحة على https://doi. org/10.1186/s12912-024-01884-2.
الملف الإضافي 1. المواد الإضافية متاحة في الملحق. مجموعة البيانات متاحة من مؤلفي النشر عند الطلب.

الشكر والتقدير

نود أن نشكر جميع المشاركين في الدراسة.

مساهمات المؤلفين

ساهم DS و LS بالتساوي في هذه المخطوطة، بما في ذلك تصور الدراسة، التصميم، جمع البيانات، التحليل، والكتابة. ساهم FW بشكل كبير في تصور الدراسة، والمراجعة، والتمويل.

التمويل

تم تمكين وتنظيم تمويل الوصول المفتوح بواسطة مشروع DEAL. تم دعم هذا البحث من قبل أجندة التكنولوجيا العالية في بافاريا.

توفر البيانات والمواد

مجموعات البيانات التي تم إنشاؤها خلال الدراسة غير متاحة للجمهور لأسباب تتعلق بحماية البيانات ولكن يمكن تقديمها عند الطلب المعقول من المؤلفين.

الإعلانات

بعد تقييم ذاتي من قبل جمعية الجامعات البافارية للعلوم التطبيقية (UAS) المعايير، تم التأكد من أن تصميم بحثنا يضمن عدم وجود ضرر محتمل للمشاركين. النتائج من تقييم جمعية الجامعات البافارية للعلوم التطبيقية (UAS) أعفت من الحاجة إلى موافقة أخلاقية إضافية. قدم كل مشارك في دراستنا موافقته المستنيرة. علاوة على ذلك، تم حماية البيانات الشخصية للأفراد بإدارة حماية البيانات (DPM) وفقًا للائحة الاتحاد الأوروبي 2016/679. حصلنا على موافقة مستنيرة من جميع المشاركين في الدراسة.
غير قابل للتطبيق.

المصالح المتنافسة

يعلن المؤلفون عدم وجود مصالح متنافسة.
تاريخ الاستلام: 4 سبتمبر 2023 تاريخ القبول: 20 مارس 2024
تم النشر عبر الإنترنت: 27 مارس 2024

References

  1. McGrow K. Artificial intelligence: Essentials for nursing. Nursing. 2019;49(9):46-9. https://doi.org/10.1097/01.NURSE.0000577716.57052.8d.
  2. TURING AM. I.-COMPUTING MACHINERY AND INTELLIGENCE. Mind. 1950;LIX(236):433-460. https://doi.org/10.1093/mind/lix.236.433.
  3. Graziani M, Dutkiewicz L, Calvaresi D, Amorim JP, Yordanova K, Vered M, et al. A global taxonomy of interpretable AI: unifying the terminology for the technical and social sciences. Artif Intell Rev. 2022;56(4):3473-504. https://doi.org/10.1007/s10462-022-10256-8.
  4. Robert N. How artificial intelligence is changing nursing. Nurs Manag. 2019;50(9):30-9. https://doi.org/10.1097/01.NUMA.0000578988.56622.21.
  5. Castagno S, Khalifa M. Perceptions of Artificial Intelligence Among Healthcare Staff: A Qualitative Survey Study. Front Artif Intell. 2020;3:578983. https://doi.org/10.3389/frai.2020.578983.
  6. Kouassi Konan JC. A Comprehensive Overview of Artificial Intelligence. In: 12th International Conference on Artificial Intelligence, Soft Computing and Applications. 2022. https://doi.org/10.5121/csit.2022.122314.
  7. Sheikh H, Prins C, Schrijvers E. In: Artificial Intelligence: Definition and Background. Cham: Springer International Publishing; 2023. p. 15-41. https://doi.org/10.1007/978-3-031-21448-6_2.
  8. Reiss D, Spina A. Scaling smart solutions with AI in health: Unlocking impact on high potential use cases. 2023. https://www3.weforum.org/ docs/WEF_Scaling_Smart_Solutions_with_Al_in_Health_Unlocking_ Impact_on_High_Potential_Use_Cases.pdf. Accessed 25 Mar 2024.
  9. Huhtala , Geurts , Mauno , Feldt . Intensified job demands in healthcare and their consequences for employee well-being and patient satisfaction: A multilevel approach. J Adv Nurs. 2021;77(9):3718-32. https://doi.org/10.1111/jan.14861.
  10. Watson D, Womack J, Papadakos S. Rise of the robots: Is artificial intelligence a friend or foe to nursing practice? Crit Care Nurs Q. 2020;43(3):303-11.
  11. Kwon H, An S, Lee HY, Cha WC, Kim S, Cho M, et al. Review of smart hospital services in real healthcare environments. Healthc Inform Res. 2022;28(1):3-15.
  12. Gao S, He L, Chen Y, Li D, Lai K. Public Perception of Artificial Intelligence in Medical Care: Content Analysis of Social Media. J Med Internet Res. 2020;22(7):e16649. https://doi.org/10.2196/16649.
  13. Buchanan C, Howitt ML, Wilson R, Booth RG, Risling T, Bamford M. Predicted Influences of Artificial Intelligence on the Domains of Nursing: Scoping Review. JMIR Nurs. 2020;3(1):e23939. https://doi.org/10.2196/ 23939.
  14. Linz K, Stula S. Demographic change in Europe-an overview. Obs Sociopolitical Developements Eur. 2010;4(1):2-10.
  15. Laukka E, Hammarén M, Kanste O. Nurse leaders’ and digital service developers’ perceptions of the future role of artificial intelligence in specialized medical care: An interview study. J Nurs Manag. 2022;30(8):383846. https://doi.org/10.1111/jonm.13769.
  16. McCarthy C, Boniol M, Daniels K, Cometto G, Diallo K, Deen Lawani A, et al. State of the world’s nursing 2020: investing in education, jobs and leadership. Genève: World Health Organization; 2020.
  17. von Gerich H, Moen H, Block LJ, Chu CH, DeForest H, Hobensack M, et al. Artificial Intelligence -based technologies in nursing: A scoping literature review of the evidence. Int J Nurs Stud. 2022;127:104153. https://doi.org/ 10.1016/j.jjnurstu.2021.104153.
  18. Sodeau A, Fox A. Influence of nurses in the implementation of artificial intelligence in health care: a scoping review. Aust Health Rev Publ Aust Hosp Assoc. 2022;46(6):736-41. https://doi.org/10.1071/AH22164.
  19. Frith KH. Artificial Intelligence: What Does It Mean for Nursing? Nurs Educ Perspect. 2019;40(4):261. https://doi.org/10.1097/01.NEP.0000000000 000543.
  20. Lambert SI, Madi M, Sopka S, Lenes A, Stange H, Buszello CP, et al. An integrative review on the acceptance of artificial intelligence among healthcare professionals in hospitals. NPJ Digit Med. 2023;6(1):111. https://doi.org/10.1038/s41746-023-00852-5.
  21. O’Connor S, Yan Y, Thilo FJS, Felzmann H, Dowding D, Lee JJ. Artificial intelligence in nursing and midwifery: A systematic review. J Clin Nurs. 2023;32(13-14):2951-68. https://doi.org/10.1111/jocn.16478.
  22. Stürz R, Stumpf C, Mendel U. Bayerisches Forschungsinstitut für Digitale Transformation, editor. Künstliche Intelligenz verstehen und gestalten: Ergebnisse und Implikationen einer bidt-Kurzbefragung in Deutschland: Analysen und Studien. Berlin; 2020. https://www.bidt.digital/wp-content/ uploads/sites/2/2022/08/bidt_Kl-Studie_Juni-2020.pdf. Accessed 25 Mar 2024.
  23. Seibert K, Domhoff D, Bruch D, Schulte-Althoff M, Fürstenau D, Biessmann F, et al. Application Scenarios for Artificial Intelligence in Nursing Care: Rapid Review. J Med Internet Res. 2021;23(11):e26522. https://doi.org/10. 2196/26522.
  24. Ng ZQP, Ling LYJ, Chew HSJ, Lau Y. The role of artificial intelligence in enhancing clinical nursing care: A scoping review. J Nurs Manag. 2021;30(8):3654-74.
  25. Kwak Y, Ahn JW, Seo YH. Influence of AI ethics awareness, attitude, anxiety, and self-efficacy on nursing students’ behavioral intentions. BMC Nurs. 2022;21(1):267. https://doi.org/10.1186/s12912-022-01048-0.
  26. Swan BA. Assessing the Knowledge and Attitudes of Registered Nurses about Artificial Intelligence in Nursing and Health Care. Nurs Econ. 2021;39(3):139-43.
  27. Chew HSJ, Achananuparp P. Perceptions and Needs of Artificial Intelligence in Health Care to Increase Adoption: Scoping Review. J Med Internet Res. 2022;24(1):e32939. https://doi.org/10.2196/32939.
  28. Rodriguez-Arrastia M, Martinez-Ortigosa A, Ruiz-Gonzalez C, RoperoPadilla C, Roman P, Sanchez-Labraca N. Experiences and perceptions of final-year nursing students of using a chatbot in a simulated emergency situation: A qualitative study. J Nurs Manag. 2022;30(8):3874-84. https:// doi.org/10.1111/jonm.13630.
  29. Ronquillo CE, Peltonen LM, Pruinelli L, Chu CH, Bakken S, Beduschi A, et al. Artificial intelligence in nursing: Priorities and opportunities from an international invitational think-tank of the Nursing and Artificial Intelligence Leadership Collaborative. J Adv Nurs. 2021;77(9):3707-17. https:// doi.org/10.1111/jan.14855.
  30. Abdullah R, Fakieh B. Health Care Employees’ Perceptions of the Use of Artificial Intelligence Applications: Survey Study. J Med Internet Res. 2020;22(5):e17620. https://doi.org/10.2196/17620.
  31. Elsayed W, Sleem W. Nurse Managers’ Perspectives and Attitude Toward Using Artificial Intelligence Technology in Nursing Settings. Assiut Sci Nurs J. 2021;0(0):0. https://doi.org/10.21608/asnj.2021.72740.1159.
  32. Seo YH, Ahn JW. The validity and reliability of the Korean version of the General Attitudes towards Artificial Intelligence Scale for nursing students. J Korean Acad Soc Nurs Educ. 2022;28(4):357-67. https://doi.org/ 10.5977/jkasne.2022.28.4.357.
  33. Shachak A, Kuziemsky C, Petersen C. Beyond TAM and UTAUT: Future directions for HIT implementation research. J Biomed Inform. 2019;100:103315. https://doi.org/10.1016/j.jbi.2019.103315.
  34. Gaughan MR, Kwon M, Park E, Jungquist C. Nurses’ Experience and Perception of Technology Use in Practice: A Qualitative Study Using an Extended Technology Acceptance Model. CIN Comput Inform Nurs. 2022;40(7):478-86. https://doi.org/10.1097/cin.0000000000000850.
  35. Zi Qi P, Li Ying JL, Han Shi JC, Ying L. The role of artificial intelligence in enhancing clinical nursing care: A scoping review. J Nurs Manag. 2022;30(8):3654-74.
  36. Bewick V, Cheek L, Ball J. Statistics review 8: Qualitative data – tests of association. Crit Care. 2004;8(1):46. https://doi.org/10.1186/cc2428.
  37. Mayring P. Qualitative content analysis. London: SAGE Publications; 2021.
  38. Nilsen P, Seing I, Ericsson C, Birken SA, Schildmeijer K. Characteristics of successful changes in health care organizations: an interview study with physicians, registered nurses and assistant nurses. BMC Health Serv Res. 2020;20(1). https://doi.org/10.1186/s12913-020-4999-8.
  39. Westbrook JI, Duffield C, Li L, Creswick NJ. How much time do nurses have for patients? a longitudinal study quantifying hospital nurses’ patterns of task time distribution and interactions with health professionals. BMC Health Serv Res. 2011;11(1). https://doi.org/10.1186/ 1472-6963-11-319.
  40. Roche MA, Friedman S, Duffield C, Twigg DE, Cook R. A comparison of nursing tasks undertaken by regulated nurses and nursing support workers: a work sampling study. J Adv Nurs. 2017;73(6):1421-32. https://doi. org/10.1111/jan.13224.
  41. Cho I, Cho J, Hong JH, Choe WS, Shin H. Utilizing standardized nursing terminologies in implementing an AI-powered fall-prevention tool to improve patient outcomes: a multihospital study. J Am Med Inform Assoc. 2023;30(11):1826-36. https://doi.org/10.1093/jamia/ocad145.
  42. Kunhardt H, Juergen Zerth, Eberl I, Flemming D, Hilbert J, Weber K, et al. Hochschulinitiative für eine personenzentrierte und nachhaltige Gestaltung von Technik und Digitalisierung in der Pflege – Wege von der Anwendungsforschung zur Implementierung und Evaluierung in der Praxis und Lebenswelt. Preprint. 2023. https://doi.org/10.13140/RG.2.2. 32095.64160.

ملاحظة الناشر

تظل Springer Nature محايدة فيما يتعلق بالمطالبات القضائية في الخرائط المنشورة والانتماءات المؤسسية.

  1. تبسيط مصطلح الذكاء الاصطناعي يعترف بتعقيده ويساعد في إنشاء إطار بحث عملي، مما يجسر الفجوة بين النظرية والممارسة في استكشاف تطبيقات الذكاء الاصطناعي.

Journal: BMC Nursing, Volume: 23, Issue: 1
DOI: https://doi.org/10.1186/s12912-024-01884-2
PMID: https://pubmed.ncbi.nlm.nih.gov/38539169
Publication Date: 2024-03-27

Nurses’ perceptions, experience
Check for updates and knowledge regarding artificial intelligence: results from a cross-sectional online survey in Germany

Domenic Sommer , Lukas Schmidbauer and Florian Wahl

Abstract

Background Nursing faces increasing pressure due to changing demographics and a shortage of skilled workers. Artificial intelligence (AI) offers an opportunity to relieve nurses and reduce pressure. The perception of AI by nurses is crucial for successful implementation. Due to a limited research state, our study aims to investigate nurses’ knowledge and perceptions of AI. Methods In June 2023, we conducted a cross-sectional online survey of nurses in Bavaria, Germany. A convenience sample via care facilities was used for the questionnaire oriented on existing AI surveys. Data analysis was performed descriptively, and we used a template analysis to evaluate free-text answers. Results 114 (%67.5%, ở32.5 %) nurses participated. Results show that knowledge about AI is limited, as only 25.2 % can be described as AI experts. German nurses strongly associate AI with (i) computers and hardware, (ii) program-ming-based software, (iii) a database tool, (iv) learning, and (v) making decisions. Two-thirds of nurses report AI as an opportunity. Concerns arise as AI is seen as uncontrollable or threat. Administration staff are seen as the biggest profiteers. Conclusion Even though there is a lack of clear understanding of AI technology among nurses, the majority recognizes the benefits that AI can bring in terms of relief or support. We suggest that nurses should be better prepared for AI in the future, e.g., through training and continuing education measures. Nurses are the working group that uses Al and are crucial for implementing nursing AI.

Keywords Nurse, Artificial intelligence, Healthcare, Nursing education, Germany
Domenic Sommer and Lukas Schmidbauer contributed equally to this work.
*Correspondence:
Domenic Sommer
domenic.sommer@th-deg.de
Deggendorf Institute of Technology, Dieter-Görlitz-Platz 1, Deggendorf 94469, Bavaria, Germany

Background

The rise of artificial intelligence (AI), propelled by advancements like ChatGPT, has heightened discourse around its varied uses, benefits, and challenges [1]. Originating in the 1950s, AI has evolved with technology, attracting increasing interest [2-4].

Ontology and terminology

To understand the nuances of AI and its subfields, it’s crucial to delineate the terms AI, machine learning (ML),
and deep learning (DL). The definition of these different terms is often lacking in current literature, although this separation is crucial according to Castagno et al. [5] and Graziani et al. [3].
  • AI is about machines that perform pure calculations. They are not based on logical thinking and merely imitate human characteristics [6].
  • ML and DL, subsets of AI, encompass data-driven learning and adaptation, enhancing predictions and decisions autonomously [4, 7].
AI, the term we focus on, symbolizes a new frontier in human-like intelligence. AI amalgamates various cuttingedge technologies, including ML, Computer Vision, and Natural Language Processing. Sheikh et al. [7] describe AI as a relatively recent technology, often perceived as a black box, designed to execute tasks requiring human intellect, including speech comprehension, pattern recognition, and decision-making. The World Economic Forum defined AI as an ‘act by sensing, interpreting data, learning, reasoning, and recommending’ [8]. However, this taxonomy appears intricate and multifaceted, encompassing numerous social, ethical, legal, and technical aspects, without yielding an universal definition [3].
In our study, we simplify AI as an intelligent system that analyses its surroundings and autonomously acts, to fulfill specific objectives [7].

Relevance and appliance in nursing context

European healthcare, challenged by an aging population and nursing staff shortages, faces increasing demands and work-related stress [9, 10]. Consequently, AI-supported solutions are emerging as key responses to these pressures [8, 10]. Economically, AI in healthcare is significant, with spending projected to reach 36.1 billion USD by 2025 [4].
AI applications, ranging from conversational agents in customer service to analytical tools in image and video analysis, are transforming various sectors, including healthcare, and facilitating smart hospitals [11]. This transformation is met with mixed public reactions, from optimism to concerns over depersonalization and job displacement in healthcare [12]. The rapid expansion of AI in healthcare promises significant advancements, notably in predictive analysis and virtual assistance, reshaping patient care and nursing practices in the near future [5, 13]. AI’s integration into healthcare is poised
to alleviate workforce strains by supporting nursing staff in areas like documentation, workflow optimization, and decision-making, enhancing care quality [10,14]. The goal in developing AI applications is to support nurses in their workflow. In this way, the role of the nurses, which is essentially human care, can be retained .

Research about nurses perceptions

Nurses, numbering 27 million globally, are the largest group in healthcare and play a critical role in evaluating and integrating AI technologies due to their insights and interfaces with various professional groups [16, 17]. Nurses central position in patient care and implementation of AI in clinical settings is key to the successful adoption of AI technologies [5, 13, 18]. The rise of AI in healthcare heightens the need for human-computer interaction ( HCI ) research and nurses involvement in AI development [17]. However, the specifics of nurses involvement are yet to be defined [ ].
The current international research on nursing AI is multidimensional, focusing on technical aspects [17, 21] and envolved on perceptions of AI, as shown in Table 1.
Most studies indicate a limited understanding and experience with AI among nurses [5, 17, 22, 23]. Castagno et al. 2020 [5] found that of nurses in Britain had no previous contact with AI, and didn’t know the difference between AI terms. In Germany, two-thirds of the public don’t understand AI [22]. Even though there is a knowledge gap, a sense of optimism toward AI is evident. Over of nurses agree that AI can catalyze transformation in nursing through better health promotion, personalized treatments, and automation in administration as well as routine tasks [26]. Nurses positively state that AI enhances health outcomes, relieves staff, and reduces costs [5, 13, 15, 25-27]. Chew et al. [27] found out, that healthcare workers perceive low-threshold access to AI. Although various AI benefits, nurses can’t estimate how their roles will be affected [13]. In addition, nurses think, that AI isn’t capable enough, because nursing is seen as human-interaction driven, holistic, and adoptable [10, 15].
Contrary to neutral and positive assumptions, of 98 healthcare professionals see AI as ‘potentially more dangerous than nuclear weapons’ [5]. Concerns include unawareness, mistrust, anxiety, worries about the patient relationship and job replacement [5,8,10, 13, 25-27]. Despite AI’s benefits, the practical implementation encounters barriers [29]. Integrating AI is hindered by fears, missing sensitization, and a lack of data [8, 22, 23, 25, 27, 30]. In addition, errors are problematic in AI usage, and data protection regulations stop nurses from using AI [20, 23, 27]. Moreover, positive outcomes, empathetic and personalized solutions,
Table 1 Related work about nurses’ perceptions on AI
Al Categories Subitems Sources
Knowledge Uncertainties in terminology and lack of AI knowledge [5, 17, 22]
Lack of experience & application in nursing [5, 17, 19, 23, 24]
Attitudes AI enhances nursing outcomes & reliefs staff [5, 13, 15, 25, 26]
Al increases efficiency & reduces costs [15, 26, 27]
AI is available, user-friendly & easy to use [27, 28]
Al changes organizations & workflows [5, 13, 15]
Al changes leadership [15, 29]
AI influences nursing roles [5, 13, 15]
AI isn’t capable enough to replace human interaction [10, 15]
Al as frightening threat & mistrust [5, 8, 25, 26]
Unawareness of advantages & applications [8, 29-31]
Worries about patient relationship & safety [10, 13, 27]
Barriers Anxiety of job loss and full automation [25, 27, 30]
Missing education and sensitization [5, 22, 25]
Lack of data and interfaces to train AI [8, 23, 27]
Errors, unexpected results and AI trustworthiness [20, 25]
Regulatory Frameworks and Data Protection [5, 21, 23, 27]
Facilitators Positive Outcomes increases intention to use AI [1,25]
Proactive define AI & advocate for patients [13, 21]
Empathetic & personalised AI applications [12, 27]
Application in health monitoring, documentation, communication, & clinical decisions support [20, 23, 24]
Training & Information about AI [20, 30, 31]
IT clinicians & technical infrastructure [8, 15, 20]
Further Research Al applications & outcomes, esp. relief of nurses [1, 4, 10, 15, 17, 21, 23]
Nursing perspective, acceptance, nursing role [12, 13, 18, 20, 27, 29]
Acceptance & user-centered design [31, 32]
Ethical, social & legal implications [8, 17, 22, 23, 25]
Limitations: response, study sample & lab settings [1, 5, 15, 17, 23, 27]
positive picture neutral picture negative picture
IT clinicians that support the process, AI education, and active involvement of nurses in AI development facilitate AI in healthcare [ ]. Nurses notice applications for health monitoring, documentation, communication, and decision support to foster AI [20, 23, 24].
Current research reveals limited AI knowledge and varied attitudes among nurses. Studies on nurses’ perceptions of AI face challenges like small samples and sampling bias [ ]. The need for updated research is underscored by rapid technological changes and healthcare setting variances . With most international research centered on the US, the applicability to German healthcare is uncertain [18]. Hence, our study focuses on understanding AI perception in Germany’s unique healthcare context, aiming to guide user-oriented AI development in nursing.

Methods

Between June and June , 2023, we conducted an online survey regarding nurses’ knowledge, perception, and experience of AI in Germany, Bavaria. Nurses of all ages and care settings were included, mainly from geriatric, inpatient, and clinical nursing.

Objective and research question

Our study seeks to enhance the understanding of nurses’ perceptions of AI in Germany, where AI adoption is emerging. We aim to gather empirical insights to guide user-centric AI development in nursing. Additionally, we aim to identify potential anxieties or misconceptions to inform practical educational initiatives and provide an overview of the current status of AI perception in nursing for policymakers, technicians and nursing managers. The research questions (RQs) we investigate are:
RQ 1: To what extent are nurses in Germany informed about AI, and what attitudes do they hold towards using AI in nursing?
RQ 2: Which areas of AI application are most recognized by nurses, and which ones do they perceive as most promising in nursing?
RQ 3: Based on nurses’ self-assessment, which healthcare professions would reap the greatest benefits from AI integration, and why?

Study instrument

Our questionnaire development was informed by a narrative literature review, sourcing relevant literature from databases like PubMed, CINAHL, Medline, and Web of Science, supplemented by Google Scholar. Key terms related to nurses’ perceptions, experience, and knowledge of AI guided the search. The most influential study for our questionnaire design was the BIDT study [22], allowing later comparison with broader population data. While models like the Technology Acceptance Model (TAM) and the Unified Theory of Acceptance and Use of Technology (UTAUT) were considered, their applicability seemed limited due to their focus on real-life technology use [33]. Instead, we adapted models tailored for nursing, as described by Gaughan et al. [34]. The final survey, shaped by existing studies [5, 22, 25, 30], integrated items from the German BIDT study [22] and Swan et al. [26] for comparability. A pre-test involving 6 nurses ensured the survey’s clarity and reliability, leading to minor adjustments in wording.
The final survey, provided in the Appendix, comprises twelve questions categorized into three sections: (i) demographic data (age group, gender, working field, education), (ii) knowledge about AI, and (iii) perceptions of AI. The survey primarily uses nominal and ordinal scales for quantitative (close-ended) data. To enrich our data with qualitative insights, we included open-ended questions. Participants were prompted to describe in their own words what they understood by AI. At the end of the survey, participants were invited to offer comments or feedback on AI in a free-text format. This open-ended approach provided diverse, nuanced perspectives on AI, revealing varying levels of understanding and attitudes among nurses. The combination of quantitative and qualitative data from these questions effectively addressed the research questions, offering a balanced view of nurses’ knowledge and perception of AI.

Data collection

Data for our cross-sectional study was collected via a self-developed online survey, using the tool Lime-Survey. The questionnaire was distributed in a convenience
sample through emails and multipliers to nursing facilities in Lower Bavaria and Upper Palatinate. Participants consented voluntarily and anonymously, with the study adhering to ethical guidelines. Study participants opt-in voluntarily for their consent after study information. In all cases, completing the survey took under five minutes. Questions about nurses’ prior experiences with AI in their work were based on dimensions resulting from a scoping review [35]. The survey design and its distribution aimed to capture a broad and representative sample of nurses’ perspectives on AI, despite limitations like sampling bias and non-validated items, discussed under Discussion section.

Data analysis

Our reporting of results adheres to our survey structure and is informed by methodologies used in the BIDT Study [22]. Quantitative data analysis was performed using IBM SPSS (Version 29). We analyzed demographic data (questions A.2.1-A.2.4) including age, gender, facility type, and education level, along with AI knowledge (A.3.1, A.3.3) and perception (A.4.1, A.4.2) responses, using frequencies and percentages.
We utilized bivariate statistics, including Spearman’s rank correlation and Cramer’s V, to analyze the impact of demographic factors on self-rated AI understanding and knowledge. These methods helped identify key patterns and associations between demographic variables and AI knowledge/perception. However, due to our study’s limited sample size, these results should be considered indicative, leading to the need of further research [36]. The significance levels were set at for significance and for high significance, aiming for a confidence in minimizing the -error. Cross-tabulation and chi-square tests were also applied to assess differences and relationships among demographic and AI-related variables.
In addition to the analysis of close-ended data, we employed a qualitative template analysis for the openended data to create joint displays. Therefore, questions A.3.2 and A.4.3 were free text fields that were analyzed with a structured qualitative template analysis according to Mayring [37] to determine categories for answering our RQ. The units of analysis were the free text fields of the questionnaire. To determine the category system, we chose a deductive approach by orienting ourselves on already known possible categories based on the literature research. In the next step, the text modules were searched for references and a coding of the relevant passages was carried out. The references were edited and extracted. Based on the coding, the original category system was adapted and a second run was started. Table 2 gives an overview of the final category system. The analysis revealed four main categories. These are (i) Knowledge
Table 2 Categories of the qualitative template analysis
Categories Subcategories Description for Items
Knowledge and definition of AI Computer-based Al as a hardware-based solution
Programming-based Al as a (text-generating) program
Input-/information-/data-based AI as system relying on data
Learning AI as a developing thing
Decision-making AI as a decision helper
Human-like Al as a natural, human-like interaction
Robot-like AI as a machine, a robot
Naming Al providers and tools AI as e.g. ChatGPT or Mindjourney
Opportunities and perceived benefits Support and relief in patient care AI as helper for patient care tasks
Knowledge acquisition Al as a tool to find information
Helping in administrative tasks Al as writing/ documentation support
Assistance systems Al with robots as assistance
Anxiety and disadvantages Fears of job replacement Al as threat for nursing jobs
Impersonally AI as a threat to social relationships
Mistakes Al as error source due to incomplete data
Insufficient Outcome Al as not comprehensive to human output
Costs and implementation efforts AI as expensive in implementing
Facilitators and future needs Sensitization Al information is necessary for nurses
Refinancing Al usage in health facilities needs financing
Data AI usage in health facilities needs data
User-friendliness Al should be easy to use and simple
Use cases AI should be helpful and practicable
and definition of AI, (ii) Opportunities and perceived benefits, (iii) Anxiety and disadvantages, and (iv) Facilitators and future needs. The analysis identified several subcategories for each category as shown in Table 2.

Results

The results are presented following the survey structure, including (i) demographics, (ii) an understanding of AI, (iii) use cases, (iv) threats, and (v) benefits.

Demographic data

A total of 114 nurses completed our survey. As Table 3 shows, two-thirds of our participants were female, and most nurses ( ) were 31 to 40 years old. Regarding the form of the facilities, of nurses work in inpatient long-term care, followed by hospitals ( ) and outpatient care ( ). One-quarter work in other care settings, like specialized care. Moreover, our study population has high education levels, like university degrees ( ) and completed apprenticeships ( ).

Understanding of AI

Table 4 shows nurses’ self-assessment of AI knowledge, measured by five items.
Table 3 General characteristics of study participants (selfintended, N=113 except gender)
Variable Category n (%)
Gender Female 77 (67.5 %)
Male 37 (32.5 %)
Diverse 0 (00.0 %)
Age group under 21 years 1 (00.9 %)
21-30 years 23 (20.4 %)
31-40 years 33 (29.2 %)
41-50 years 24 (21.2 %)
51-60 years 26 (23.0 %)
over 60 years 6 (05.3 %)
Facility form Inpatient long-term care 45 (39,8 %)
Outpatient care 18 (15,9 %)
Nursing (hospital/clinic) 21 (18.6 %)
Other 29 (25,7 %)
Education level Secondary school diploma 21 (18.6 %)
Higher School or specialized secondary school 16 (14.2 %)
Completed vocational training 33 (29.2 %)
University degree (at least Bachelor’s degree) 35 (31.0 %)
Other 6 (05.3 %)
Table 4 Knowledge of nurses about AI (self-intended, survey order, N=113)
Variable Category n ( %)
How much do you know about AI? I would describe myself as an AI expert. 2 (01.8 %)
I can explain well what is meant by it. 26 (23.4 %)
I know roughly what is meant by it. 67 (60.4 %)
I know the term, but I don’t know what it means. 14 (12.6 %)
I have not heard the term before. 2 (01.8 %)
Summarized as ‘AI Connoisseurs/ Experts’ (25.2%), ‘AI Non-Connoisseurs/-Experts’ (74.8%)
According to Stürz et al. [22], nurses who describe themself as an AI expert and nurses who can explain well what AI means can be summarized as ‘AI Connoisseurs’. Using this simplification, 25.2 % nurses state to have substantial AI knowledge, are ‘AI Connoisseurs’. Conversely, 74.8 % nurses, in summary, lack a solid AI understanding. Gender differences in self-reported AI knowledge were observed: ( 15 out of 35 ) of men and ( 13 out of 75 ) of women identified themselves as ‘AI Connoisseurs’.
To examine the association between AI knowledge and gender, -tests ( ) were conducted. The results indicated a significant association between gender and AI knowledge ( value ), with a Cramer’s V correlation of 0.335 . Additionally, the type of healthcare facility ( value , Cramer’s ) and educational level ( value , Spearman’s correlation ) also showed significant influences on AI knowledge. At the same time, age was not a significant factor. Due to low cell counts in our tests, these results should be interpreted cautiously, indicating the need for further research with larger samples.
In addition, study participants’ definitions of AI offer a multidimensional understanding that extends beyond the quantitative assumptions, as emphasized in the following quotes. AI is seen as a dynamic tool with the potential to learn and expand its capabilities (Text 1, para 55). The qualitative analysis reveals that nurses associate AI with (i) computers and hardware, (ii) software, (iii) human-like characteristics exemplified by natural language processing, (iv) a data-driven tool capable of (v) learning, and (vi) decision-making. We recognized a dichotomy in nurses’ perspectives, describing AI with human-like attributes and as a more mechanical, machine-like tool. Participants often see AI as a tool that simulates human interactions or replicates human-like behaviour. This view is further influenced by AI’s association with robots and machines, recognizing that AI is frequently integrated as software into robotic systems. Nurses also name AI providers, such as OpenAI (Text 1, para. 48).

‘AI is a machine with practices to perform and learn from data’ (Text 1, para. 34)

‘AI is programmed by humans and is intended to be used in multiple domains to make decisions, perform actions, and perform everyday human tasks within its programming and to identify and solve problems.’ (Text 1, para. 36)
‘AI means using robots which can react to movements or speech.’ (Text 1, para. 83)
‘AI […] can act similarly to a human, e.g., write text, act human, etc.’ (Text 1, para. 31)
‘AI is a machine’s ability to imitate human abilities.’ (Text 1, para. 72)
‘AI can decide and learn alone, considering markers taught to it.’ (Text 1, para. 15)
Furthermore, nurses recognize that AI is not a ready-touse solution but requires data processing and preparation (Text 1, para. 11). Nurses understand the critical role of data quality in shaping AI as a decision-making instrument (Text 1, para. 55).

Use-cases and applications

For the fields of application of AI in nursing (Table 5), nurses are most familiar with patient monitoring (55.7 %) and route planning ( ) as well as AI-aided nursing documentation ( ). Nurses know least about the AI areas of care prediction ( ), nursing diagnosis ( ), and wound management ( ).
Our qualitative data reveals AI’s utility extends to patient monitoring, routine nursing tasks, assessing ‘drug compatibility effects and interactions’ (Text 2, para. 124), enhancing ‘patient care’ (Text 2, para. 127), and assisting with ‘intelligent incontinence care’ (Text 2, para. 130). Additionally, nurses recognize AI’s supportive role in administrative functions, such as assistance in ‘staff planning’ (Text 2, para. 125) and ‘service scheduling’ (Text 2, para. 129). Furthermore, AI has potential in ‘housekeeping activities’ (Text 2, para. 134) through AI-based cleaning robots. AI-supported technical assistance, like active exoskeletons, were also highlighted (Text 2, para. 128).
Table 5 Application fields of AI in nursing (self-intended, descending order, N=88)
Variable Category n (%)
Which application areas of AI in nursing do you know? Patient monitoring (e.g. vital signs, sleep) 49 (55.7 %)
Route planning 42 (47.7 %)
Nursing documentation 38 (43.2 %)
Patient care prediction (e.g. fall detection) 34 (38.6 %)
Making nursing diagnoses 28 (31.8 %)
Wound management 19 (21.6 %)
According to Seibert et al. [23] all displayed AI applications exist and are beneficial for nursing
Table 6 Perception of nurses about AI (self-intended, N=103)
Variable Category n (%)
Do you see Al in nursing more as an opportunity or more as a threat? Exclusively as an opportunity 14 (13.7 %)
Rather as an opportunity 53 (52.0 %)
Rather as a danger 13 (12.7 %)
Exclusively as a danger 1 (01.0 %)
I do not know, I cannot judge 22 (21.6 %)
Summarized as ‘positive Al perception’ (65.7%) Summarized as ‘negative Al perception’

Opportunities and positive AI perception

Table 6 displays that nearly two-thirds (65.7 %) view AI positively as an opportunity for nursing, while a smaller fraction (13.7 %) perceives AI negatively as a threat or danger. Additionally, of respondents could not categorize AI either way, indicating that they could not judge due to a lack of knowledge.
The qualitative data confirm an optimistic view of AI’s potential, recognizing its ability to support and transform nursing. As visualized by the quotes, respondents expressed that AI can be a valuable relief in nursing, aiding everyday activities, optimizing work processes, and supporting data collection and interpretation, such as analyzing feelings or vital parameters. The ability of AI to enhance human-computer interaction is also recognized, particularly through AI-based voice control capabilities.
‘In the area of AI, we are at the beginning of the changes in nursing.’ (Text 4, para. 171)
‘AI has been part of our everyday lives for a long time. Whether as Siri [or] Alexa, AI tries to make everyday things easier or improve them. However, these are only minor peripheral areas, AI can recognize feelings and record vital parameters.’ (Text 1, para. 42)
‘AI means to optimize and support work and relieve employees.’ (Text 1, para. 74)
Moreover, nurses appreciate the role of AI in knowledge acquisition, including AI to answer questions, provide guidance, and aid in documentation. AI’s predictive
capabilities are beneficial, allowing for early detection of care needs and risks, such as sepsis from care and laboratory parameters (Text 4, para. 170, Text 3, para. 152).

Threats and negative Al perception

While AI presents numerous opportunities, such as simplifying tasks and making life easier, nurses recognize potential threats and risks. Negative perceptions are grounded in practical concerns, such as the risk of AI malfunction leading to mistakes. Some nurses fear that AI could replace human tasks. This is coupled with the understanding that nursing is fundamentally grounded in social relationships between nurses and patients, a dimension that AI might render impersonal. Other concerns include losing control over AI, which can cause feelings of helplessness and the high costs associated with AI implementation and maintenance (Text 4, para. 164).
‘I am afraid of AI stupidity, a system that delivers results after training that is not comprehensible because the sources are missing.’ (Text 1, para. 23) ‘AI threatens our jobs and makes care even more impersonal.’ (Text 1, para. 29)
‘It is also a question of refinancing. Unfortunately, it is often not possible for cost reasons to implement meaningful innovations or purchase them.’ (Text 4, para. 171)

Profiteers and beneficial user groups

According to the quantitative findings, Table 7 illustrates the nurses’ perspectives on which outpatient
Table 7 Most benefiting group of AI in nursing (self-intended, descending order, N=95)
Variable Category n (%)
Which group of people benefits most from the use of AI? Administration and management staff 47 (49.5 %)
Nursing and support staff 24 (25.3 %)
Patients in need of care 17 (17.9 %)
Social services and support 7 (07.4 %)
and inpatient care groups could benefit most from AI. Nearly half of the nurses think that administration and management staff ( ) can be supported most by AI, followed by nursing and support staff ( ), patients in need of care ( ), and social services and support ( ).
In addition, nurses identified several parties who may profit from AI through qualitative data: (i) Micro Level – Physicians and hospital operators, (ii) Meso Level – Family nurses, social services, and service personnel, and (iii) Macro Level- Health and care insurers, government and policymakers, and nursing scientists. The advantages include more efficient processes such as invoice verification for insurers, early risk detection to contain healthcare costs, staffing savings for operators and employers, predictive care needs and assessment development in nursing science, and support for social participation in social services (Text 3, para 136-150).

Needs for further AI implementation

Our qualitative data identifies potential facilitators, barriers, and requirements for successful AI implementation in nursing. The foremost necessity is the availability of reliable, structured, and complete data (Text 1, para. 55). Without such data, AI outcomes may be unsatisfactory (Text 1, Quote 41). The more accessible and trustworthy the data, the more robust the decisions the AI system makes. In addition to data considerations, human involvement is essential in shaping AI. This means active participation in its design and regulation, maintaining patient relationships, and fostering trust through sensitization and education about AI (Text 4, paras. 165, 166). Addressing fears that lead to feelings of helplessness, such as concerns over control and the possibility of rectifying unexpected outcomes, is also vital (Text 4, para. 164).
‘AI […] must be fed with input to get a corresponding output.’ (Text 1, para. 11)
‘All groups of people will benefit only if they actively embrace this change. This is likely the greatest challenge, along with sensible refinancing.’
(Text 4, para. 165)
‘AI should be regulated, no one should lose her job through AI, and the relationship with patients must be preserved.’ (Text 4, para. 159)
‘Humans should be able to access and “reprogram” at any time.’ (Text 4, para. 164)
‘AI in route planning currently depends on too many parameters to work properly. AI should always be developed to provide maximum support to the nurse.’ (Text 4, para. 167)
As the quotes also conduct, user experience is another critical factor. AI solutions must be user-friendly, easy to use, and time-saving (Text 4, para. 161). Furthermore, the quotes show that AI solutions are still not satisfying enough, leading to the potential for improvements. Lastly, considerations regarding the costs and effort required for implementation should not be overlooked (Text 4, para. 165).

Discussion

To answer the RQs, a cross-sectional online survey was conducted among nurses in Bavaria, Germany, in June 2023, utilizing a convenience sample. Our approach has provided comprehensive insights into nurses’ awareness and attitudes towards AI.

Limitations

Our study provides valuable insights into nurses’ AI perception but is subject to several limitations. First, the concept of ‘perception’ itself is not clearly defined, potentially encompassing aspects of usage and acceptance, which visualizes the need for ground research. Our convenience sampling approach and the focus on Bavaria may not accurately reflect the broader nursing population, potentially introducing bias. The small sample size of 114 participants limits the study’s statistical power, particularly for conducting comprehensive multivariate analyses. We used bi-variate tests and calculated the chisquare. Still, nevertheless, the expected count in some cells was under 5 in of the cells, limiting our reliability and implicating the need for further research with bigger samples [36]. In summary, the small study sample constraint hinders our ability to explore interrelations and generalizations are limited.
Additionally, the study’s reliance on existing instruments like the BIDT study [22], while omitting broader models such as TAM or UTAUT may have affected the depth of theoretical exploration. The survey’s brevity could have resulted in an under-assessment of detailed nurse demographics and AI experiences. Furthermore, the inclusion and exclusion criteria are specific, focusing on formal nurses. Our study did not consider other healthcare professionals, missing out on comparative insights of AI perception across healthcare roles. Our cross-sectional design might not capture the evolving perceptions of AI, especially in the context of rapidly changing media narratives. Future research should consider more extensive and more varied sample sizes, incorporate comprehensive theoretical frameworks and adopt longitudinal methods.

Al knowledge and attitudes

In addressing RQ 1 on AI knowledge and attitudes in nursing, our study found that only of nurses have substantial AI knowledge, with the majority ( ) having limited understanding. This aligns with existing research indicating a general lack of AI familiarity in healthcare [17, 19]. The need for increased AI awareness and training is evident [1,20]. Furthermore, we confirmed the non-encompassing AI definition by nurses [7]. Nurses often perceive AI technically, which might contribute to skepticism and hinder acceptance. Our study also shows that some participants equate AI with humans, leading to a need to sensitize further for AI.
Although many nursing staff know too little about AI, two-thirds view AI optimistically, suggesting openness to AI implementation despite concerns about patient relationships, errors, costs, and job security. This is surprising and needs further research, as a lack of knowledge can be one of the constant causes of negative perception and low acceptance [34]. Furthermore, our open-ended data calls for functional, safe AI applications and emphasizes the importance of clear communication and nurses’ involvement in AI development to foster understanding and alleviate. This is also discussed in current literature to be important [1,17,19,20]. In addition, studies about change management in nursing observed a higher success rate and acceptance of new technologies if nurses are involved and are part of the decision-making [38].

Al application areas

For RQ 2 the most known AI applications are (i) patient monitoring (55.7 %), (ii) route planning (47.7 %), and (iii) nursing documentation ( ). Our qualitative data confirm these potentials. Nevertheless, open-ended data and the lack of knowledge let us assume that the use of AI in nursing care in Bavaria has so far made little inroads
and that nurses have, therefore, hardly been confronted with AI in the working context.
Time studies in nursing highlight the considerable time spent on routine and administrative tasks, thereby supporting the use of AI in areas like monitoring and documentation [39, 40]. Current research, including works like Cho et al. [41], focuses on AI solutions for patient monitoring and clinical decision support. Our survey findings align with national literature [23], indicating that nurses are familiar with AI applications in administrative tasks, a trend expected to grow in the coming decade. Utilizing AI to streamline these processes can reduce bureaucratic workload.

Al outcome and benefits

Addressing RQ 3 on the expected outcomes and benefits of AI, our study indicates mostly positive perceptions. Nurses anticipate AI to alleviate the increasing burdens in nursing, particularly amidst current staff shortages and escalating workloads [9, 10]. As perceived by our respondents, the primary beneficiaries are administration and management ( ) and nursing staff ( ). This finding underscores the evolving roles in AI-integrated healthcare and aligns with Swan et al. [26].
Our results suggest starting AI implementation in administrative areas, supporting the recommendations of Chew et al. [27], and extending to direct care. However, the apprehensions about AI expressed in some open-ended responses highlight the need for increased awareness and education across all potential user groups to mitigate concerns and foster a comprehensive understanding of AI’s role in nursing.

Implications for nurses, technicians and policymakers

Our study’s findings lead to several critical implications and recommendations for nursing practice, policymakers, and healthcare facility management:
(i) Educational Campaigns: Increase public discourse and offer specialized training and further education for nursing staff to enhance their understanding of AI.
(ii) Reimbursement and Legal Frameworks: Set up clear reimbursement policies and legal frameworks that support the use of AI in nursing.
(iii) Investing in Nursing AI: Allocate resources for developing and implementing AI in nursing, ensuring user-friendliness.
(iv) Involvement of Nurses in AI Development: Engage nurses in the AI development.
(v) Provision of Technical Support: Deploy additional technical support and programming staff to assist nurses using AI tools in healthcare facilities.
(vi) Academization of Nursing: Integrate AI education into nursing curricula to prepare future nurses for the technological advancements in healthcare.
We emphasize that policymakers should prioritize the development and implementation of (i) targeted educational campaigns, (ii) reimbursement and legal frameworks that support the integration of AI in nursing, and (iii) investment into the development and implementation of nursing AI tailored to nursing professionals. Investments should include funding for AI research, establishing clear guidelines for AI application in healthcare, and ensuring ethical considerations are central to AI deployment in nursing settings. These straightforward guidelines should help to eliminate mistrust and negative perceptions, like seeing insecurities and threats to patient care as visualized rarely by our open-ended data. Investment and developments should focus on AI for patient monitoring (including patient allocation and fall detection), route planning, and nursing documentation, as nurses know them already and perceive them as promising areas. Administration and nursing management staff can be the most beneficial stakeholders, AI developers should ask themselves how to support nurses.
Moreover, technicians should (iv) involve nurses in developing AI applications through, e.g., focus groups. We recommend (v) that nurses get technical support in using AI tools effectively and aligning them with their needs. Nurses’ curricula should (vi) integrate AI education and prepare nurses for technological advancements.

Conclusion

Nurses AI perception takeaways

In summary, 74.8 % of our surveyed nurses lack a solid AI understanding. perceive AI positive as an opportunity. Our primary conclusion regarding the usage of AI is that nurses know patient monitoring, route planning, and nursing documentation as application fields. With the mounting pressure on nursing, AI is poised to play a vital role in alleviating healthcare staff and meeting the escalating nursing demands in Western Europe [10, 23]. Beneficial user groups are mostly administration and management staff ( ) as well as nurses and their support staff (25.3 %).
Although our study is more indicative than conclusive, the finding that only of respondents view themselves as AI experts highlights the need for increased AI education and sensitization in nursing. Mixed reactions, including positive views alongside rare concerns about job security and patient care changes, further underscore
the importance of nurse involvement and education in AI development.
We anticipate a rise in AI knowledge and acceptance in nursing over the next decade, suggesting a promising future for AI integration. Our results indicate that administrative and routine tasks are prime areas for AI, making them strategic starting points for maximizing AI’s potential. Additionally, open-ended responses reveal a need for regulation, accessibility, and a focus on alleviating nurses’ workloads. We emphasize the importance of developing user-friendly AI that saves time on repetitive and administrative tasks, allowing nurses to focus on direct patient care.

Further research

The limitations of our study and the national research gap necessitate further in-depth investigations into nurses’ comprehension, apprehensions, and expectations of AI. Subsequent research should build on our findings to deepen the understanding of AI’s role in healthcare. Essential areas for future research include applying new technologies in realistic settings and integrating AI knowledge into nursing education, ensuring its relevance and sustainability [42]. Studies should focus on assimilating AI knowledge into nursing training, especially considering nurses’ varying levels of familiarity. Nurses should be enabled to use new technologies, and the solutions should note real-world needs and should be integrated without a low threshold into practice, like our project ‘Smart Forest – 5G Clinics’ implies.
Further analysis should be with more extensive study samples and longitudinal to encompass broader data and compare AI perceptions across healthcare professions. Multivariate analysis should be conducted with new data, answering the question of what are influential factors to a good AI understanding and positive attitude towards AI. Besides, analyzing potential differences in perceptions, e.g., between nurses and doctors, is vital. Furthermore, exploring the real-world implementation of AI and its user acceptance are crucial, especially in light of our findings indicating nurses’ readiness to embrace AI as an opportunity rather than a risk.

Supplementary Information

The online version contains supplementary material available at https://doi. org/10.1186/s12912-024-01884-2.
Additional file 1. Supplementary Material is available in the Appendix. The data set is available from both publication authors on request.

Acknowledgements

We would like to acknowledge all study participants.

Authors’ contributions

DS and LS contributed equally to this manuscript, including study conceptualization, design, data collection, analysis, and writing. FW contributed significantly to the study conceptualization, review, and funding.

Funding

Open Access funding enabled and organized by Projekt DEAL. This research has been supported by the Hightech Agenda Bavaria.

Availability of data and materials

The data sets generated during the study are not publicly available for data protection reasons but can be provided upon reasonable request from the authors.

Declarations

Following a self-assessment by the Association of the Bavarian Universities of Applied Sciences (UAS) standards, it was established that our research design ensured no potential harm to participants. The results from the assessment of the Association of the Bavarian Universities of Applied Sciences (UAS) waiving the need for further ethical approval. Every participant in our study provided their informed consent. Furthermore, individuals’ personal data was protected with Data protection management (DPM) according to the EU-Regulation 2016/679. We obtained informed consent from all study participants.
Not applicable.

Competing interests

The authors declare no competing interests.
Received: 4 September 2023 Accepted: 20 March 2024
Published online: 27 March 2024

References

  1. McGrow K. Artificial intelligence: Essentials for nursing. Nursing. 2019;49(9):46-9. https://doi.org/10.1097/01.NURSE.0000577716.57052.8d.
  2. TURING AM. I.-COMPUTING MACHINERY AND INTELLIGENCE. Mind. 1950;LIX(236):433-460. https://doi.org/10.1093/mind/lix.236.433.
  3. Graziani M, Dutkiewicz L, Calvaresi D, Amorim JP, Yordanova K, Vered M, et al. A global taxonomy of interpretable AI: unifying the terminology for the technical and social sciences. Artif Intell Rev. 2022;56(4):3473-504. https://doi.org/10.1007/s10462-022-10256-8.
  4. Robert N. How artificial intelligence is changing nursing. Nurs Manag. 2019;50(9):30-9. https://doi.org/10.1097/01.NUMA.0000578988.56622.21.
  5. Castagno S, Khalifa M. Perceptions of Artificial Intelligence Among Healthcare Staff: A Qualitative Survey Study. Front Artif Intell. 2020;3:578983. https://doi.org/10.3389/frai.2020.578983.
  6. Kouassi Konan JC. A Comprehensive Overview of Artificial Intelligence. In: 12th International Conference on Artificial Intelligence, Soft Computing and Applications. 2022. https://doi.org/10.5121/csit.2022.122314.
  7. Sheikh H, Prins C, Schrijvers E. In: Artificial Intelligence: Definition and Background. Cham: Springer International Publishing; 2023. p. 15-41. https://doi.org/10.1007/978-3-031-21448-6_2.
  8. Reiss D, Spina A. Scaling smart solutions with AI in health: Unlocking impact on high potential use cases. 2023. https://www3.weforum.org/ docs/WEF_Scaling_Smart_Solutions_with_Al_in_Health_Unlocking_ Impact_on_High_Potential_Use_Cases.pdf. Accessed 25 Mar 2024.
  9. Huhtala , Geurts , Mauno , Feldt . Intensified job demands in healthcare and their consequences for employee well-being and patient satisfaction: A multilevel approach. J Adv Nurs. 2021;77(9):3718-32. https://doi.org/10.1111/jan.14861.
  10. Watson D, Womack J, Papadakos S. Rise of the robots: Is artificial intelligence a friend or foe to nursing practice? Crit Care Nurs Q. 2020;43(3):303-11.
  11. Kwon H, An S, Lee HY, Cha WC, Kim S, Cho M, et al. Review of smart hospital services in real healthcare environments. Healthc Inform Res. 2022;28(1):3-15.
  12. Gao S, He L, Chen Y, Li D, Lai K. Public Perception of Artificial Intelligence in Medical Care: Content Analysis of Social Media. J Med Internet Res. 2020;22(7):e16649. https://doi.org/10.2196/16649.
  13. Buchanan C, Howitt ML, Wilson R, Booth RG, Risling T, Bamford M. Predicted Influences of Artificial Intelligence on the Domains of Nursing: Scoping Review. JMIR Nurs. 2020;3(1):e23939. https://doi.org/10.2196/ 23939.
  14. Linz K, Stula S. Demographic change in Europe-an overview. Obs Sociopolitical Developements Eur. 2010;4(1):2-10.
  15. Laukka E, Hammarén M, Kanste O. Nurse leaders’ and digital service developers’ perceptions of the future role of artificial intelligence in specialized medical care: An interview study. J Nurs Manag. 2022;30(8):383846. https://doi.org/10.1111/jonm.13769.
  16. McCarthy C, Boniol M, Daniels K, Cometto G, Diallo K, Deen Lawani A, et al. State of the world’s nursing 2020: investing in education, jobs and leadership. Genève: World Health Organization; 2020.
  17. von Gerich H, Moen H, Block LJ, Chu CH, DeForest H, Hobensack M, et al. Artificial Intelligence -based technologies in nursing: A scoping literature review of the evidence. Int J Nurs Stud. 2022;127:104153. https://doi.org/ 10.1016/j.jjnurstu.2021.104153.
  18. Sodeau A, Fox A. Influence of nurses in the implementation of artificial intelligence in health care: a scoping review. Aust Health Rev Publ Aust Hosp Assoc. 2022;46(6):736-41. https://doi.org/10.1071/AH22164.
  19. Frith KH. Artificial Intelligence: What Does It Mean for Nursing? Nurs Educ Perspect. 2019;40(4):261. https://doi.org/10.1097/01.NEP.0000000000 000543.
  20. Lambert SI, Madi M, Sopka S, Lenes A, Stange H, Buszello CP, et al. An integrative review on the acceptance of artificial intelligence among healthcare professionals in hospitals. NPJ Digit Med. 2023;6(1):111. https://doi.org/10.1038/s41746-023-00852-5.
  21. O’Connor S, Yan Y, Thilo FJS, Felzmann H, Dowding D, Lee JJ. Artificial intelligence in nursing and midwifery: A systematic review. J Clin Nurs. 2023;32(13-14):2951-68. https://doi.org/10.1111/jocn.16478.
  22. Stürz R, Stumpf C, Mendel U. Bayerisches Forschungsinstitut für Digitale Transformation, editor. Künstliche Intelligenz verstehen und gestalten: Ergebnisse und Implikationen einer bidt-Kurzbefragung in Deutschland: Analysen und Studien. Berlin; 2020. https://www.bidt.digital/wp-content/ uploads/sites/2/2022/08/bidt_Kl-Studie_Juni-2020.pdf. Accessed 25 Mar 2024.
  23. Seibert K, Domhoff D, Bruch D, Schulte-Althoff M, Fürstenau D, Biessmann F, et al. Application Scenarios for Artificial Intelligence in Nursing Care: Rapid Review. J Med Internet Res. 2021;23(11):e26522. https://doi.org/10. 2196/26522.
  24. Ng ZQP, Ling LYJ, Chew HSJ, Lau Y. The role of artificial intelligence in enhancing clinical nursing care: A scoping review. J Nurs Manag. 2021;30(8):3654-74.
  25. Kwak Y, Ahn JW, Seo YH. Influence of AI ethics awareness, attitude, anxiety, and self-efficacy on nursing students’ behavioral intentions. BMC Nurs. 2022;21(1):267. https://doi.org/10.1186/s12912-022-01048-0.
  26. Swan BA. Assessing the Knowledge and Attitudes of Registered Nurses about Artificial Intelligence in Nursing and Health Care. Nurs Econ. 2021;39(3):139-43.
  27. Chew HSJ, Achananuparp P. Perceptions and Needs of Artificial Intelligence in Health Care to Increase Adoption: Scoping Review. J Med Internet Res. 2022;24(1):e32939. https://doi.org/10.2196/32939.
  28. Rodriguez-Arrastia M, Martinez-Ortigosa A, Ruiz-Gonzalez C, RoperoPadilla C, Roman P, Sanchez-Labraca N. Experiences and perceptions of final-year nursing students of using a chatbot in a simulated emergency situation: A qualitative study. J Nurs Manag. 2022;30(8):3874-84. https:// doi.org/10.1111/jonm.13630.
  29. Ronquillo CE, Peltonen LM, Pruinelli L, Chu CH, Bakken S, Beduschi A, et al. Artificial intelligence in nursing: Priorities and opportunities from an international invitational think-tank of the Nursing and Artificial Intelligence Leadership Collaborative. J Adv Nurs. 2021;77(9):3707-17. https:// doi.org/10.1111/jan.14855.
  30. Abdullah R, Fakieh B. Health Care Employees’ Perceptions of the Use of Artificial Intelligence Applications: Survey Study. J Med Internet Res. 2020;22(5):e17620. https://doi.org/10.2196/17620.
  31. Elsayed W, Sleem W. Nurse Managers’ Perspectives and Attitude Toward Using Artificial Intelligence Technology in Nursing Settings. Assiut Sci Nurs J. 2021;0(0):0. https://doi.org/10.21608/asnj.2021.72740.1159.
  32. Seo YH, Ahn JW. The validity and reliability of the Korean version of the General Attitudes towards Artificial Intelligence Scale for nursing students. J Korean Acad Soc Nurs Educ. 2022;28(4):357-67. https://doi.org/ 10.5977/jkasne.2022.28.4.357.
  33. Shachak A, Kuziemsky C, Petersen C. Beyond TAM and UTAUT: Future directions for HIT implementation research. J Biomed Inform. 2019;100:103315. https://doi.org/10.1016/j.jbi.2019.103315.
  34. Gaughan MR, Kwon M, Park E, Jungquist C. Nurses’ Experience and Perception of Technology Use in Practice: A Qualitative Study Using an Extended Technology Acceptance Model. CIN Comput Inform Nurs. 2022;40(7):478-86. https://doi.org/10.1097/cin.0000000000000850.
  35. Zi Qi P, Li Ying JL, Han Shi JC, Ying L. The role of artificial intelligence in enhancing clinical nursing care: A scoping review. J Nurs Manag. 2022;30(8):3654-74.
  36. Bewick V, Cheek L, Ball J. Statistics review 8: Qualitative data – tests of association. Crit Care. 2004;8(1):46. https://doi.org/10.1186/cc2428.
  37. Mayring P. Qualitative content analysis. London: SAGE Publications; 2021.
  38. Nilsen P, Seing I, Ericsson C, Birken SA, Schildmeijer K. Characteristics of successful changes in health care organizations: an interview study with physicians, registered nurses and assistant nurses. BMC Health Serv Res. 2020;20(1). https://doi.org/10.1186/s12913-020-4999-8.
  39. Westbrook JI, Duffield C, Li L, Creswick NJ. How much time do nurses have for patients? a longitudinal study quantifying hospital nurses’ patterns of task time distribution and interactions with health professionals. BMC Health Serv Res. 2011;11(1). https://doi.org/10.1186/ 1472-6963-11-319.
  40. Roche MA, Friedman S, Duffield C, Twigg DE, Cook R. A comparison of nursing tasks undertaken by regulated nurses and nursing support workers: a work sampling study. J Adv Nurs. 2017;73(6):1421-32. https://doi. org/10.1111/jan.13224.
  41. Cho I, Cho J, Hong JH, Choe WS, Shin H. Utilizing standardized nursing terminologies in implementing an AI-powered fall-prevention tool to improve patient outcomes: a multihospital study. J Am Med Inform Assoc. 2023;30(11):1826-36. https://doi.org/10.1093/jamia/ocad145.
  42. Kunhardt H, Juergen Zerth, Eberl I, Flemming D, Hilbert J, Weber K, et al. Hochschulinitiative für eine personenzentrierte und nachhaltige Gestaltung von Technik und Digitalisierung in der Pflege – Wege von der Anwendungsforschung zur Implementierung und Evaluierung in der Praxis und Lebenswelt. Preprint. 2023. https://doi.org/10.13140/RG.2.2. 32095.64160.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

  1. Simplifying the term AI acknowledges its complexity and aids in creating a practical research framework, bridging theory and practice in exploring AI’s applications.