Hemşirelik ve Yapay Zekâ: Bir Ölçek Geliştirme Çalışması
Abstract
Yapay zekâ teknolojileri sağlık hizmetleri alanında geniş bir yelpazede uygulama potansiyeline sahiptir. Bu bağlamda yürütülen bu çalışmanın temel amacı, hemşirelerin yapay zekaya yönelik bilgi, tutum ve davranışlarını ölçmeye yönelik geçerli ve güvenilir ölçme aracı geliştirmeyi, ölçeğin toplam ve alt boyut puanlarının sosyodemografik değişkenler ile karşılaştırarak incelemeyi amaçlamaktır. Çalışmanın ilk evresinde, geliştirilmesi hedeflenen ölçeğin bağlamıyla uyumlu olacak şekilde, ilgili literatürdeki kuramsal çerçeveler ve araştırmalar detaylı bir şekilde incelenerek madde havuzu oluşturulmuştur. Geliştirilen madde havuzu, farklı uzmanlık alanlarından toplam 12 uzmanın görüşüne Lawshe yöntemi aracılığıyla sunulmuş ve neticesinde kapsam geçerlik indeksi 0,97 olarak hesaplanmıştır. Bu süreç sonucunda, beşli Likert tipi cevap formatına sahip 51 maddeden oluşan bir taslak ölçme aracı elde edilmiştir. Literatür doğrultusunda ölçek ve alt boyutlarına etki edeceği düşünülen 16 sosyodemografik soru ve hemşirelerin yapay zekayı kullanmak istedikleri alanları belirlemek için açık uçlu bir soru soru ile veri toplama formu hazırlanmıştır. Türkiye'nin 72 ilindeki kamu ve özel sağlık kuruluşlarında görev yapan hemşirelere 'Google Form' platformu üzerinden hazırlanan form hemşirelik dernekleri ve üyeleri aracılığıyla ulaştırılmıştır. Gönüllülük esasına dayalı olarak gerçekleştirilen bu çalışmaya 665 hemşire katılım sağlamıştır. Veriler AMOS, SPSS 23 ve MAXQDA programı kullanılarak analiz edilmiştir. Çalışmaya katılan katılımcıların %79,7 (n:530) kadın, %51,3 (n:341) evli, %17,7 (n:118) ön lisans, %48,6 (n:323) lisans, %26,5 (n:176) yüksek lisans, %7,2 (n:48) doktora mezunu olduğu saptanmıştır. Katılımcıların yalnızca %6,5(n:43) hemşirelik eğitimi sırasında, %5,3 (n:35) ise çalıştıkları kurumda yapay zekâ teknolojisine ilişkin eğitim aldığı saptanmıştır. Ölçeğin faktör analizi için uygunluğunu ve örneklem büyüklüğünü değerlendirmek için yapılan analizlerde, KMO değeri 0,96 ve Bartlett Küresellik Testi sonucu 34200,10 (p<0,001) olarak bulunmuştur. Yapılan AFA değerlendirmesinde faktör yüklerinin 0,493 ile 0,894 aralığında değiştiği ve özdeğeri 1'den büyük dört faktör açığa çıktığı belirlenmiştir. Bu dört faktörlü yapı, toplam varyansın %67,71'ini açıklamaktadır. Madde uyum indeksleri analiz sonuçları χ2/sd=2,867, RMSEA=,053, SRMR=,051, CFI=,934, IFI=,935, TLI=,930, NFI=,903, GFI=,827 olarak saptanmıştır. Ölçeğin güvenirliğine ilişkin elde edilen Cronbach's alfa 0,960 McDonald's Omega değerleri ise 0,970 olarak saptanmıştır. Ölçek maddelerinin toplam puanının ayırt edici düzeyini belirlemek amacıyla %27'lik alt ve üst grup karşılaştırmaları yapılmıştır. Analiz sonucunda, %27'lik alt ve üst gruplar arasında tüm maddelerde anlamlı farklılıklar saptanmıştır (p <0,001). Elde edilen bulgular neticesinde ölçeğin geçerliliğinin, güvenilirliğinin ve iç tutarlılığının yüksek olduğu sonucuna varılmıştır. Çalışma bulguları, ölçek ve alt boyut puanlarının katılımcıların yaşı, eğitim düzeyi, medeni durumu, çalıştıkları kurum türü, çalıştıkları birim türü, unvanları, yapay zekâ teknolojisi ve genel teknolojik yeterlilik algıları, yapay zekâ konusunda eğitim alma durumları ve eğitim alma istekleri ile alt boyutlar veya toplam puan düzeyinde istatistiksel olarak anlamlı farklılıklar gösterdiğini ve bazı sosyodemografik değişkenler ile pozitif yönde farklı derecelerde korelasyonlar ortaya koymuştur (p < 0,05). Çalışmanın nitel analizleri sonucunda, hemşirelerin dokuz farklı alanda yapay zekâ uygulamalarına yönelik kullanım isteği belirlenmiştir. Çalışma bulguları ışığında, sağlık yönetimi politikaları ve yönetimi açısından; yapay zekâ alanında eğitim olanaklarının artırılması, lisans ve lisansüstü eğitim müfredatlarına yapay zekâ içeriklerinin entegre edilmesi, hizmet içi eğitim programlarının düzenlenmesi, multidisipliner ekiplerle araştırma ve geliştirme faaliyetlerinin desteklenmesi, hemşirelerin yapay zekâ yeterliliklerini artırmaya yönelik stratejilerin geliştirilerek uygulamaya konulması önerilmektedir.
Artificial intelligence technologies have the potential to be applied in a wide range of areas in healthcare. In this context, the main purpose of this study is to develop a valid and reliable measurement tool to measure nurses' knowledge, attitudes and behaviors towards artificial intelligence and to examine the total and sub-dimension scores of the scale by comparing them with sociodemographic variables. In the first phase of the study, the theoretical frameworks and research in the relevant literature were examined in detail in a way that would be compatible with the context of the scale targeted for development, and an item pool was created. The developed item pool was presented to the opinions of a total of 12 experts from different fields of expertise via the Lawshe method, and as a result, the content validity index was calculated as 0,97. As a result of this process, a draft measurement tool consisting of 51 items with a five-point Likert-type response format was obtained. A data collection form was prepared with 16 sociodemographic questions thought to affect the scale and its sub-dimensions in line with the literature and an open-ended question to determine the areas where nurses wanted to use artificial intelligence. The form prepared via the 'Google Form' platform was delivered to nurses working in public and private health institutions in 72 provinces of Türkiye through nursing associations and their members. In addition, 665 nurses participated in this study, which was conducted on a voluntary basis. The data were analyzed using the AMOS, SPSS 23 and MAXQDA programs. It was determined that 79.7% (n: 530) of the participants in the study were female, 51.3% (n: 341) were married, 17.7% (n: 118) had an associate degree, 48.6% (n: 323) had a bachelor's degree, 26.5% (n: 176) had a master's degree, and 7.2% (n: 48) had a doctorate. It was determined that only 6.5% (n: 43) of the participants received training on artificial intelligence technology during their nursing education and 5.3% (n: 35) in the institution they worked at. In the analyses conducted to evaluate the suitability of the scale for factor analysis and sample size, the KMO value was found to be 0,96 and the Bartlett Test of Sphericity result was found to be 34200,10 (p<0.001). In the EFA evaluation, it was determined that the factor loadings ranged from 0.493 to 0.894 and four factors with eigenvalues greater than 1 were revealed. This four-factor structure explained 67.71% of the total variance. The item fit index analysis results were determined as χ2/sd=2.867, RMSEA=0.053, SRMR=0.051, CFI=0.934, IFI=0.935, TLI=0.930, NFI=0.903, GFI=0.827. Regarding the reliability of the scale, Cronbach's alpha was found to be 0.960 and McDonald's Omega values were found to be 0.970, In order to determine the discriminatory level of the total score of the scale items, 27% subgroup and upper group comparisons were made. As a result of the analysis, significant differences were found in all items between the 27% subgroup and upper group (p <0.001). As a result of the findings obtained, it was concluded that the validity, reliability and internal consistency of the scale were high. The study findings revealed that the scale and sub-dimension scores showed statistically significant differences at the level of participants' age, level of education, marital status, type of institution they work in, type of unit they work in, their titles, perceptions of artificial intelligence technology and general technological competence, status of receiving training in artificial intelligence and their desire to receive training, and sub-dimensions or total score, and positive correlations with some socio-demographic variables at different levels (p < .05). As a result of the qualitative analyses of the study, nurses' desire to use artificial intelligence applications in nine different areas was determined. In light of the findings of the study, in terms of health management policies and management; it is recommended to increase educational opportunities in the field of artificial intelligence, integrate artificial intelligence content into undergraduate and graduate education curricula, organize in-service training programs, support research and development activities with multidisciplinary teams, and develop and implement strategies to increase nurses' artificial intelligence competencies.
Artificial intelligence technologies have the potential to be applied in a wide range of areas in healthcare. In this context, the main purpose of this study is to develop a valid and reliable measurement tool to measure nurses' knowledge, attitudes and behaviors towards artificial intelligence and to examine the total and sub-dimension scores of the scale by comparing them with sociodemographic variables. In the first phase of the study, the theoretical frameworks and research in the relevant literature were examined in detail in a way that would be compatible with the context of the scale targeted for development, and an item pool was created. The developed item pool was presented to the opinions of a total of 12 experts from different fields of expertise via the Lawshe method, and as a result, the content validity index was calculated as 0,97. As a result of this process, a draft measurement tool consisting of 51 items with a five-point Likert-type response format was obtained. A data collection form was prepared with 16 sociodemographic questions thought to affect the scale and its sub-dimensions in line with the literature and an open-ended question to determine the areas where nurses wanted to use artificial intelligence. The form prepared via the 'Google Form' platform was delivered to nurses working in public and private health institutions in 72 provinces of Türkiye through nursing associations and their members. In addition, 665 nurses participated in this study, which was conducted on a voluntary basis. The data were analyzed using the AMOS, SPSS 23 and MAXQDA programs. It was determined that 79.7% (n: 530) of the participants in the study were female, 51.3% (n: 341) were married, 17.7% (n: 118) had an associate degree, 48.6% (n: 323) had a bachelor's degree, 26.5% (n: 176) had a master's degree, and 7.2% (n: 48) had a doctorate. It was determined that only 6.5% (n: 43) of the participants received training on artificial intelligence technology during their nursing education and 5.3% (n: 35) in the institution they worked at. In the analyses conducted to evaluate the suitability of the scale for factor analysis and sample size, the KMO value was found to be 0,96 and the Bartlett Test of Sphericity result was found to be 34200,10 (p<0.001). In the EFA evaluation, it was determined that the factor loadings ranged from 0.493 to 0.894 and four factors with eigenvalues greater than 1 were revealed. This four-factor structure explained 67.71% of the total variance. The item fit index analysis results were determined as χ2/sd=2.867, RMSEA=0.053, SRMR=0.051, CFI=0.934, IFI=0.935, TLI=0.930, NFI=0.903, GFI=0.827. Regarding the reliability of the scale, Cronbach's alpha was found to be 0.960 and McDonald's Omega values were found to be 0.970, In order to determine the discriminatory level of the total score of the scale items, 27% subgroup and upper group comparisons were made. As a result of the analysis, significant differences were found in all items between the 27% subgroup and upper group (p <0.001). As a result of the findings obtained, it was concluded that the validity, reliability and internal consistency of the scale were high. The study findings revealed that the scale and sub-dimension scores showed statistically significant differences at the level of participants' age, level of education, marital status, type of institution they work in, type of unit they work in, their titles, perceptions of artificial intelligence technology and general technological competence, status of receiving training in artificial intelligence and their desire to receive training, and sub-dimensions or total score, and positive correlations with some socio-demographic variables at different levels (p < .05). As a result of the qualitative analyses of the study, nurses' desire to use artificial intelligence applications in nine different areas was determined. In light of the findings of the study, in terms of health management policies and management; it is recommended to increase educational opportunities in the field of artificial intelligence, integrate artificial intelligence content into undergraduate and graduate education curricula, organize in-service training programs, support research and development activities with multidisciplinary teams, and develop and implement strategies to increase nurses' artificial intelligence competencies.
Description
Keywords
Sağlık Yönetimi, Healthcare Management
Turkish CoHE Thesis Center URL
WoS Q
Scopus Q
Source
Volume
Issue
Start Page
End Page
213