Healthcare providers’ knowledge, attitudes, and practices toward integrating evidence-based practice and artificial intelligence in Saudi Arabia: a cross-sectional study

Abstract

Background

Integrating evidence-based practice (EBP) with artificial intelligence (AI) has the potential to enhance healthcare quality, clinical decision-making, and system efficiency. Understanding healthcare providers’ knowledge, attitudes, and practices (KAP) toward this integration is essential to inform education and implementation strategies. This study aimed to evaluate healthcare providers’ knowledge, attitudes, and practices regarding the integration of EBP and AI in Saudi Arabia.

Methods

A descriptive cross-sectional study was conducted between October and December 2025 among licensed healthcare providers working in governmental and private institutions across Saudi Arabia. Data was collected using a validated, self-administered questionnaire. Descriptive and inferential analyses, including chi-square tests and logistic regression, were performed using SPSS, with p < 0.05 considered statistically significant.

Results

A total of 503 healthcare professionals participated. Overall mean scores indicated moderate knowledge of AI and EBP (mean = 2.45 ± 0.72), generally positive attitudes (mean = 2.46 ± 0.69), and lower levels of practice (mean = 2.22 ± 0.75). Knowledge was significantly associated with professional role (p = 0.009), years of experience (p = 0.014), institution type (p = 0.003), and academic qualification (p < 0.001). Practice was significantly associated with years of experience (p = 0.001) and educational qualification (p = 0.027), while attitudes differed significantly only by age group (p = 0.012). Significant positive correlations were observed among KAP domains (r = 0.37–0.45, p < 0.001), with practice showing the strongest association with the overall KAP score (r = 0.90).

Conclusion

Healthcare providers in Saudi Arabia demonstrate moderate knowledge and positive attitudes toward integrating AI and EBP; however, practical application remains limited. Targeted training and institutional support are needed to bridge the gap between knowledge and practice and promote effective integration into routine healthcare delivery.

Introduction

In the fast-growing global healthcare arena, two closely related trends have recently become prominent: evidence-based practice (EBP) and the widespread adoption of artificial intelligence (AI) across care pathways in clinical settings, driven by advances in digital technology. EBP entails using the best available research evidence, along with patient values and clinical expertise, to inform healthcare decisions. EBP has become not only a benchmark for quality clinical care but has also been consistently shown to be associated with increased patient satisfaction, optimal resource utilization, and safer, more cost-effective care [1]. At the same time, AI tools in healthcare settings are transforming how data is processed and analyzed, how decisions are made, and how care is provided. It also offers scalable pattern recognition, prediction, and decision support that can improve but not replace clinical judgment [2, 3]. Therefore, identifying best practices for maintaining EBP and AI is essential to enhance patient care outcomes, and the successful integration of EBP and AI in healthcare largely depends on professionals’ knowledge, attitudes, and practices (KAP) [4, 5].

Healthcare providers play a significant role in using up-to-date, rigorous scientific evidence to reduce medical errors and improve efficiency in healthcare systems. However, despite these advantages, EBP adoption remains inconsistent across health care systems worldwide [1, 6]. Research showed that 30–40% of patients still lack evidence-based care, highlighting a persistent evidence-practice gap [7]. In Saudi Arabia, EBP has been prioritized as a cornerstone of healthcare reform and quality improvement. Nevertheless, similar implementation challenges persist. For instance, fewer than half of Saudi primary care physicians in the northern region reported high levels of EBP knowledge or routine use of evidence in clinical decision-making [8]. Likewise, Saudi nurse managers demonstrated only moderate EBP competency, reflecting ongoing gaps in workforce training and institutional support [9].

In parallel to the expansion of EBP, AI has become a transformative force in healthcare globally. AI applications in healthcare, such as predictive analytics, diagnostic imaging, clinical decision support, and patient monitoring, have shown potential to improve risk prediction, diagnostic accuracy, and workflow efficiency [5, 10]. In Saudi Arabia, AI adoption is a cornerstone of Vision 2030, which seeks to digitalize health services and improve the quality of life through innovative technologies [11]. The Saudi Data and Artificial Intelligence Authority (SDAIA) is leading strategic investments to build digital infrastructure and advance AI capabilities across hospitals and research institutions [12, 13]. Recent reports indicate that AI in Saudi healthcare remains at an early stage: only 5.88% of major hospitals have AI research centers, and 8.82% use AI in clinical care [14]. Surveys indicate that only 14.8% of healthcare providers are very familiar with AI, and approximately 31% use AI in decision-making, amid concerns about training, cost, data security, and reduced human interaction [15, 16]. A recent study indicates limited familiarity and use of AI among licensed dietitians in Saudi Arabia. Although 62.7% reported incorporating AI into practice, significant concerns remain about training, data security, costs, and the loss of human interaction, revealing gaps in preparedness for broader AI adoption [17].

Saudi Arabia’s healthcare transformation agenda similarly prioritizes strengthening EBP alongside AI to enhance care quality, system efficiency, and patient outcomes [13]. This transformation is especially critical in light of the Kingdom’s growing burden of non-communicable diseases, which account for approximately 73% of deaths [18]. Integrating EBP with AI in healthcare, therefore, empowers clinicians with advanced tools for prevention, early detection, and proactive management of chronic diseases [18]. Integrating AI into EBP aligns algorithm proposals with validated science and patient-centered principles, enhancing safety and trust [19]. However, several persistent barriers may hinder the adoption of both EBP and AI in Saudi clinical practice. These challenges include limited time, heavy workloads, insufficient managerial support, restricted access to scientific databases, and inadequate training in research and technology use [20–22]. Many healthcare professionals also report challenges with EBP and AI use, including limited digital literacy, disparities in digital skills, gaps in EBP knowledge, negative attitudes toward EBP use, and difficulty translating research into practice [13, 21, 22]. Furthermore, institutional concerns, such as data privacy, regulatory compliance, and insufficient leadership support, continue to impede the effective adoption of EBP and AI [6, 23]. These challenges collectively highlight the need for ongoing professional development, supportive leadership, and accessible training resources to ensure the successful integration of EBP and AI into clinical practice. Overcoming these challenges is essential to ensure that patient care aligns with global evidence-based standards [13, 24].

Despite numerous studies examining either EBP or AI adoption independently, limited research has investigated their combined integration in Saudi Arabia’s healthcare context. As both EBP and AI are essential pillars of modern healthcare, their integration represents a critical step toward improving overall system performance. Understanding healthcare providers’ KAP toward this integration is therefore vital for identifying readiness levels, training needs, and contextual barriers. Such insights can guide workforce development, inform curriculum revisions, and shape institutional strategies that promote evidence-driven digital transformation [8, 15]. Ultimately, this study advances Saudi Arabia’s digital health reforms by providing evidence for policymakers, clinical leaders, and educators, thereby supporting the creation of a modern, digital, evidence-based healthcare system aligned with Vision 2030 goals. Therefore, this study aimed to evaluate healthcare providers’ knowledge, attitudes, and practices regarding the integration of EBP and AI in Saudi Arabia.

Materials and methods

Study settings

This study was conducted among healthcare providers working across Saudi Arabia. We targeted both governmental and private healthcare institutions nationwide. Participants were selected from various health aspects, including curative, preventive, and academic professions. The study was conducted over three months, from October 2025 to December 2025.

Study design

A descriptive cross-sectional study design was adopted to assess healthcare providers’ knowledge, attitudes, and practices regarding the integration of evidence-based practice and artificial intelligence in Saudi Arabia.

Study population

The study population consisted of healthcare professionals currently working in the governmental or private health sectors across Saudi Arabia. Eligible participants included physicians, nurses, pharmacists, nutrition and dietetics specialists, public health specialists, faculty and academic staff, and other allied health professionals, such as laboratory technicians, physiotherapists, and radiologic technologists. All participants were required to hold a license from the Saudi Commission for Health Specialties (SCFHS) and to have at least two years of professional experience in their respective fields.

Sample method and size

A non-probability snowball sampling design was employed to recruit participants via social media platforms, including WhatsApp and X (formerly Twitter), as well as professional healthcare groups. Firstly, we distribute the questionnaire link to selected healthcare providers working in curative, preventive, and academic institutions across Saudi Arabia. Secondly, we requested the participants to share the questionnaire link with their professional peers and colleagues. Due to the snowball sampling methodology, participants forwarded the link to unknown others, and therefore, the total number of individuals who received the survey link cannot be determined. As a result, a true response rate cannot be calculated. Since the total number of healthcare providers in Saudi Arabia is not precisely known, the sample size was calculated using the formula for an infinite population, which is appropriate when the population is large or undefined. The following equation was applied:

Sample size calculation, (n) = Z²*p*q / d², Where (n) is the required sample size, (Z = 1.96) corresponds to a 95% confidence level, (p = 0.5 or 50%) reflects the expected prevalence rate, and (d = 0.05) denotes a 5% margin of error. Based on this, the minimum sample size required was 385 participants. This approach is widely accepted in health research when population parameters are unknown or difficult to enumerate. In the present study, a total of 503 complete responses were obtained, exceeding the minimum required sample size. The larger sample size enhances the study’s statistical power, reduces sampling error, and increases the robustness, precision, and external validity of the findings. Therefore, the final sample size is considered more than sufficient to support reliable and scientifically sound conclusions [25, 26].

Validity of the questionnaire

The questionnaire’s validity was ensured through expert review and statistical assessment. Content validity was established by a multidisciplinary panel that evaluated item clarity, relevance, and alignment with the study objectives [27]. Construct validity was supported by strong internal consistency across domains (Cronbach’s alpha: 0.763 for knowledge, 0.815 for attitudes, and 0.938 for practices) and by significant positive correlations among the knowledge, attitude, and practice domains (r = 0.374–0.452, p < 0.001). The strong correlation between the practice domain and the overall KAP score (r = 0.899, p < 0.001) further confirmed the instrument’s ability to capture practical engagement. Overall, the questionnaire demonstrated good validity and reliability for assessing healthcare providers’ KAP toward integrating evidence-based practice and artificial intelligence [28].

Data collection procedure

Data were collected using a self-administered questionnaire designed explicitly for this study. It was developed by the research team following an extensive review of current literature and previously validated KAP instruments addressing EBP and AI in healthcare [4, 5, 8]. In addition, the development process was guided by the World Health Organization (WHO), which provided a standardized methodological framework for constructing KAP surveys [29]. The questionnaire was originally developed in English. A forward-backward translation method was used to produce an Arabic version. Two independent bilingual experts translated the English version into Arabic, and a third expert back-translated it into English. Discrepancies were resolved by consensus. The final Arabic version was used for data collection. It was then evaluated by a panel of academic and subject-matter experts to ensure clarity, relevance, and content validity. The reviewers included specialists in Evidence-Based Practice EBP, AI in healthcare, public health, biostatistics, and language.

The questionnaire consisted of four main sections. The first section contains demographic information, including age, gender, occupation, and years of experience. The second section assessed participants’ knowledge about evidence-based practice and artificial intelligence in healthcare. It considers whether they have received formal instruction, recognizes how AI assists in clinical decision-making, differentiates between traditional statistical approaches and AI-based models, and is aware of real-world applications of AI in healthcare. Overall, it assesses their awareness and foundational knowledge of EBP and AI integration in professional practice. The third section measures participants’ attitudes toward incorporating EBP and AI into healthcare. It explores their views on the inclusion of these topics in health education, their confidence in using AI tools, and the ethical issues that may influence their use. It also examines their interest in further training, beliefs about AI’s potential to improve healthcare performance, concerns about job replacement, and comfort in interpreting AI-generated results. Additionally, it examines the extent to which participants trust AI recommendations supported by scientific data and their concerns regarding data privacy. The fourth section focused on the practical application of these concepts in daily professional activities, assessing how healthcare professionals use AI to support EBP. It examines the frequency of AI use for finding and managing research evidence, evaluates the quality and credibility of studies, and applies AI tools in daily decision-making processes. Participants indicate whether they use AI to search literature, appraise evidence, interpret data patterns, and apply findings to clinical or public health settings. Additionally, it identifies the primary sources of evidence they rely on, such as scientific journals, official guidelines, institutional protocols, expert opinions, health data systems, AI-driven tools, and educational resources. Overall, this section measures the practical extent to which AI and EBP are integrated into professional healthcare activities.

Pilot study

A pilot test was conducted among 50 healthcare professionals to evaluate the clarity and reliability of the questionnaire. These 50 participants were excluded from the final study sample. The pilot results showed that the questionnaire items were clear, culturally appropriate, and well accepted by the target healthcare professionals. Minor revisions were made to improve item clarity and logical flow. Internal consistency reliability was assessed using Cronbach’s alpha for the translated Arabic version of the questionnaire. Cronbach’s alphas demonstrated good internal consistency across all domains: 0.74 for the knowledge domain, 0.81 for the attitude domain, and 0.90 for the practice domain.

Data analysis

Data were checked for completeness and analyzed using SPSS software. Descriptive statistics, including frequencies, percentages, means, standard deviations, and reliability measures, were used to summarize the data. Associations between categorical variables and demographic characteristics were assessed using the Chi-square test and Fisher’s exact test. Correlation analysis was also performed. Logistic regression analysis was conducted to identify associations between attitude and independent predictors. Furthermore, ordinal regression analysis was used to determine associations between ordinal outcome variables and independent variables. Attitude scores were dichotomized into “negative” (< 60%) and “positive” (≥ 60%) categories because only 19 participants (3.8%) had negative attitudes, causing a ceiling effect that could violate the proportional odds assumption of ordinal regression and potentially impact model stability. Knowledge and practice scores were retained in three categories (low, moderate, high) due to sufficient distribution across levels. Results were presented as odds ratios (ORs) with 95% confidence intervals (CIs), and p-values < 0.05 were considered statistically significant for all analyses.

For data cutoff and scoring, each section was scored using the typical KAP methodology employed in previous studies. Knowledge items (six questions) were scored as 3 for ‘Yes,’ 2 for ‘Not sure,’ and 1 for ‘No,’ with total scores ranging from 6 to 18. Participants scoring ≥ 75% were considered to have high knowledge, those scoring 60–75% were considered to have moderate knowledge, and those scoring < 60% were considered to have poor knowledge. Attitude items (eleven questions) were measured using a three-point Likert scale (Agree = 3, Neutral = 2, Disagree = 1), yielding total scores ranging from 11 to 33. Scores < 60% indicated a negative attitude, whereas scores ≥ 60% reflected a positive attitude. Practice items consisted of a total of sixteen questions, of which fifteen were included in the scoring analysis, and one was analyzed separately. The fifteen practice questions were scored on a three-point scale (Yes = 3, Sometimes = 2, No = 1), with total scores ranging from 15 to 45. Participants scoring ≥ 75% were classified as having high practice, those scoring 60–75% as having moderate practice, and those scoring < 60% as having poor practice. These cutoff values align with commonly accepted standards in KAP research for distinguishing satisfactory from unsatisfactory levels of knowledge, attitudes, and practices.

Results

Table 1 shows the demographic profile of the 503 participants. More than one-third of respondents were aged 25–34 years (36.0%), and nearly one-third were aged 35–44 years (32.4%), whereas only a minority were aged 55 years or older (3.0%) or 24 years or younger (5.6%). Most of the sample were male (62.0%), and nearly three-quarters were Saudi nationals (74.8%). Most participants worked in urban areas (61.0%), while more than one-third were based in small towns (33.2%) and only a minority in rural areas (5.8%). Nearly one third of participants were public health specialists (28.6%), around one fifth were physicians (21.1%) and nurses (20.7%), less than one fifth were allied health professionals (18.5%), and only about one tenth were faculty members (11.1%). Nearly half worked in government hospitals (45.9%), more than one quarter in primary healthcare centers (25.4%), less than one fifth in private hospitals (17.7%), and about one tenth in universities (10.9%). Nearly two-fifths had 10 or more years of work experience (38.4%), one-third had 2–6 years (32.2%), and almost one-third had 7–10 years (29.4%). Less than half held a bachelor’s degree (44.9%), more than one quarter had a master’s degree (26.8%), and only a small proportion held a diploma (12.3%) or Ph.D. (15.9%). Slightly less than half had attended AI-related training (48.5%), and the vast majority of those reported CME/CPD accreditation (80.7%). Similarly, half (50.1%) attended EBP training, and more than four-fifths (80.6%) reported CME accreditation.

Table 1.

Descriptive frequency distribution with percentage of demographic variables (N = 503)

Demographic Variables	Subcategories	Frequency (N)	Percentage (%)
Age Group	24 Years	28	5.6%
	25–34 Years	181	36.0%
	35–44 Years	163	32.4%
	45–54 Years	116	23.1%
	55 + Years	15	3.0%
Gender	Male	312	62.0%
	Female	191	38.0%
Nationality	Saudi	376	74.8%
	Non-Saudi	127	25.2%
Primary location	Rural area	29	5.8%
	Town	167	33.2%
	Urban City	307	61.0%
Current Role	Faculty	56	11.1%
	Nurse	104	20.7%
	Physician	106	21.1%
	Public Health Specialist	144	28.6%
	Other allied health professions	93	18.5%
Institution Types	Govt Hospital	231	45.9%
	PHC	128	25.4%
	Private Hospital	89	17.7%
	University Teacher	55	10.9%
Years of Experience	2–6	162	32.2%
	7–10	148	29.4%
	10+	193	38.4%
Highest Academic Qualification	Diploma	62	12.3%
	Bachelors	226	44.9%
	Masters	135	26.8%
	Ph.D.	80	15.9%
Have you attended any formal courses, workshops, or training on AI in healthcare?	Yes	244	48.5%
	No	259	51.5%
If you answered yes, were the training sessions accredited for CME/CPD hours?	Yes	197	80.7%
	No	47	19.3%
Have you attended any formal courses, workshops, or training on EBP?	Yes	252	50.1%
	No	251	49.9%
If you answered yes, were the training sessions accredited for CME/CPD hours?	Yes	203	80.6%
	No	49	19.4%

Table 2 summarizes the descriptive results for the Knowledge, Attitude, and Practice (KAP) components. The average knowledge score among participants was 2.45 (SD = 0.72), and the domain’s reliability was acceptable (α = 0.763). Knowledge scores indicated that respondents were most aware of the role of artificial intelligence in supporting decision-making in their field (mean = 2.58). In contrast, they were least aware of the distinction between traditional statistical models and AI-based predictive models (mean = 2.29). Attitude scores generally reflected positive views, with an overall mean of 2.46 (SD = 0.69) and a reliability coefficient of 0.815. Items indicating strong attitudes included agreement that evidence-based practice should be part of medical and health curricula (mean 2.53) and acknowledgment that AI can improve healthcare accuracy and efficiency (mean 2.48). Conversely, lower attitude scores were observed for confidence in interpreting AI-generated outputs (mean 2.30) and for concerns about AI replacing healthcare roles (mean 2.40). Practical use of AI was the weakest among the three domains, with an average score of 2.22 (SD 0.75) despite a high reliability coefficient of 0.938. Items related to the everyday use of AI tools, such as supporting clinical decision-making (mean 2.05) and predicting health outcomes (mean 2.02), scored lower than other items. This indicates that although participants have moderate knowledge and mostly positive attitudes, the actual use of AI in professional tasks remains limited.

Table 2.

Descriptive analysis of KAP model with mean, S. D., and reliability (N = 503)

KAP Questions	Mean	S. D.
KNOWLEDGE
I am familiar with the concept of Evidence-Based Practice (EBP).	2.50	0.711
I am familiar with the concept of AI in healthcare.	2.43	0.748
AI can support decision-making in my field.	2.58	0.663
I understand how AI supports evidence-based decision-making.	2.52	0.670
I can differentiate between traditional statistical models and AI-based predictive models	2.29	0.774
I am familiar with examples of AI applications in healthcare.	2.38	0.766
Overall Mean and S. D. (Standard Deviation)	2.45	0.722
Cronbach’s Alpha	0.763
ATTITUDE
EBP should be a core component of all medical and health education programs.	2.53	0.679
AI should be integrated into medical and health education curricula	2.52	0.679
I feel confident using AI tools relevant to my field of practice.	2.42	0.690
Ethical concerns limit the use and adoption of AI in healthcare.	2.44	0.659
AI can improve the accuracy and efficiency of healthcare interventions.	2.48	0.694
AI may replace certain healthcare roles in the future.	2.40	0.759
I am confident in interpreting AI-generated outputs for medical decisions.	2.30	0.719
EBP is not sufficiently utilized in current healthcare practice.	2.34	0.731
I trust AI-generated recommendations when they are supported by scientific evidence.	2.47	0.685
I am concerned about data privacy and security when using AI in healthcare.	2.48	0.670
I am interested in receiving further training in EBP and AI.	2.65	0.613
Overall Mean and S. D. (Standard Deviation)	2.46	0.689
Cronbach’s Alpha	0.815
PRACTICES
I use AI tools to search for relevant literature.	2.38	0.657
I rely on AI to summarize or extract findings from research articles.	2.32	0.691
I use AI to identify recent or high-quality studies for decision-making.	2.33	0.717
I use AI tools to support systematic reviews and evidence synthesis.	2.28	0.732
I use AI to track new publications or clinical guidelines.	2.33	0.715
I use AI to evaluate the credibility of research evidence	2.26	0.752
I use AI to detect data patterns or trends that inform evidence appraisal.	2.21	0.744
I use AI to compare multiple evidence sources before making decisions.	2.25	0.730
I apply AI to assess potential biases or limitations in research.	2.23	0.747
I use AI to support the critical appraisal of evidence for interventions.	2.26	0.755
I use AI tools to support clinical decision-making in my daily work.	2.05	0.789
I rely on AI to predict health outcomes or disease trends.	2.02	0.817
I use AI tools to tailor evidence-based recommendations to meet the needs of specific patients or groups.	2.13	0.806
I use AI to integrate patient or community preferences with scientific evidence in decision-making	2.08	0.784
I use AI to evaluate outcomes of implemented interventions.	2.18	0.772
Overall Mean and S. D. (Standard Deviation)	2.22	0.747
Cronbach’s Alpha	0.938

Table 3 presents the statistical relationships between demographic factors and levels of knowledge, attitudes, and practices regarding artificial intelligence and evidence-based practice. Most demographic variables were not significantly associated with attitudes or practice scores. However, knowledge scores were significantly related to current professional role (p = 0.009), years of experience (p = 0.014), and academic qualification (p = 0.000), with higher levels of knowledge seen among faculty members, those with 7–10 years of experience, and participants holding postgraduate degrees. Institutional type was also associated with notable differences: respondents working in government hospitals and universities reported higher knowledge levels than those in primary care settings. Attitude scores generally did not differ significantly across groups, except for age, in which younger participants aged 24 or younger showed a slightly higher percentage of negative attitudes (p = 0.012). Practice scores varied significantly by experience (p = 0.001) and academic qualifications (p = 0.027), with higher practice levels reported among respondents with master’s and doctoral degrees and those with fewer years of work experience. However, these findings were built on bivariate analysis and did not conder for potential confounding variables such as academic qualification, professional role, and institutional type. Therefore, these relationships should be interpreted cautiously, as they may change after adjustment for multiple demographic factors. Overall, the findings suggest that although demographic factors influence knowledge and practice, attitudes toward AI and EBP remained relatively consistent across most groups.

Table 3.

Statistical association between KAP model categories with demographic variables (N = 503)

Demographic Variable	Knowledge Categories					Attitude Categories			Practice Categories
	Total No. (%)	Low (< 60%) No. (%)	Moderate (60–75%) No. (%)	High (≥ 75%) No. (%)	p- value	Negative (< 60%) No. (%)	Positive (≥ 60%) No. (%)	p- value	Poor (< 60%) No. (%)	Moderate (60–75%) No. (%)	High (≥ 75%) No. (%)	p-value
Age Group
≤ 24 Years	28(5.6%)	0(0.0%)	5(17.9%)	23(82.1%)	0.382	4(14.3%)	24(85.7%)	0.012	5(17.9%)	2(7.1%)	21(75.0%)	0.097
25–34 Years	181(36.0%)	18(9.9%)	51(28.2%)	112(61.9%)		4(2.2%)	177(97.8%)		37(20.4%)	58(32.0%)	86(47.5%)
35–44 Years	163(32.4%)	15(9.2%)	42(25.8%)	106(65.0%)		4(2.5%)	159(97.5%)		34(20.9%)	46(28.2%)	83(50.9%)
45–54 Years	116(23.1%)	11(9.5%)	25(21.6%)	80(69.0%)		5(4.3%)	111(95.7%)		23(19.8%)	26(22.4%)	67(57.8%)
55 + Years	15(3.0%)	3(20.0%)	3(20.0%)	9(60.0%)		2(13.3%)	13(86.7%)		5(33.3%)	2(13.3%)	8(53.3%)
Gender
Male	312(62.0%)	28(9.0%)	72(23.1%)	212(67.9%)	0.353	14(4.5%)	298(95.5%)	0.286	67(21.5%)	85(27.2%)	160(51.3%)	0.715
Female	191(38.0%)	19(9.9%)	54(28.3%)	118(61.8%)		5(2.6%)	186(97.4%)		37(19.4%)	49(25.7%)	105(55.0%)
Nationality
Saudi	376(74.8%)	38(10.1%)	97(25.8%)	241(64.1%)	0.411	16(4.3%)	360(95.7%)	0.428	77(20.5%)	94(25.0%)	205(54.5%)	0.289
Non-Saudi	127(25.2%)	9(7.1%)	29(22.8%)	89(70.1%)		3(2.4%)	124(97.6%)		27(21.3%)	40(31.5%)	60(47.2%)
Primary location of your current workplace
Rural Village	29(5.8%)	3(10.3%)	5(17.2%)	21(72.4%)	0.901	2(6.9%)	27(93.1%)	0.276	6(20.7%)	5(17.2%)	18(62.1%)	0.373
Small City	167(33.2%)	15(9.0%)	42(25.1%)	110(65.9%)		4(2.4%)	163(97.6%)		36(21.6%)	52(31.1%)	79(47.3%)
Urban City	307(61.0%)	29(9.4%)	79(25.7%)	199(64.8%)		13(4.2%)	294(95.8%)		62(20.2%)	77(25.1%)	168(54.7%)
Current Role
Faculty	56(11.1%)	3(5.4%)	5(8.9%)	48(85.7%)	0.009*	3(5.4%)	53(94.6%)	0.734	10(17.9%)	16(28.6%)	30(53.6%)	0.279
Nurse	104(20.7%)	11(10.6%)	28(26.9%)	65(62.5%)		2(1.9%)	102(98.1%)		17(16.3%)	21(20.2%)	66(63.5%)
Other allied health professions	93(18.5%)	12(12.9%)	31(33.3%)	50(53.8%)		4(4.3%)	89(95.7%)		24(25.8%)	25(26.9%)	44(47.3%)
Physician	106(21.1%)	13(12.3%)	27(25.5%)	66(62.3%)		5(4.7%)	101(95.3%)		27(25.5%)	28(26.4%)	51(48.1%)
Public Health Specialist	144(28.6%)	8(5.6%)	35(24.3%)	101(70.1%)		5(3.5%)	139(96.5%)		26(18.1%)	44(30.6%)	74(51.4%)
Types of Institution
Govt Hospital	231(45.9%)	23(10.0%)	52(22.5%)	156(67.5%)	0.003*	11(4.8%)	220(95.2%)	0.747	48(20.8%)	60(26.0%)	123(53.2%)	0.323
PHC	128(25.4%)	17(13.3%)	45(35.2%)	66(51.6%)		3(2.3%)	125(97.7%)		32(25.0%)	37(28.9%)	59(46.1%)
Private Hospital	89(17.7%)	3(3.4%)	21(23.6%)	65(73.0%)		3(3.4%)	86(96.6%)		13(14.6%)	20(22.5%)	56(62.9%)
University Teacher	55(10.9%)	4(7.3%)	8(14.5%)	43(78.2%)		2(3.6%)	53(96.4%)		11(20.0%)	17(30.9%)	27(49.1%)
Years of Experience
2–6 Years	162(32.2%)	13(8.0%)	34(21.0%)	115(71.0%)	0.014*	8(4.9%)	154(95.1%)	0.160	38(23.5%)	32(19.8%)	92(56.8%)	0.001*
7–10 Years	148(29.4%)	8(5.4%)	35(23.6%)	105(70.9%)		2(1.4%)	146(98.6%)		16(10.8%)	49(33.1%)	83(56.1%)
10 + Years	193(38.4%)	26(13.5%)	57(29.5%)	110(57.0%)		9(4.7%)	184(95.3%)		50(25.9%)	53(27.5%)	90(46.6%)
Highest Academic Qualification
Diploma	62(12.3%)	13(21.0%)	22(35.5%)	27(43.5%)	0.000*	1(1.6%)	61(98.4%)	0.418	14(22.6%)	23(37.1%)	25(40.3%)	0.027*
Bachelors	226(44.9%)	26(11.5%)	74(32.7%)	126(55.8%)		10(4.4%)	216(95.6%)		55(24.3%)	61(27.0%)	110(48.7%)
Masters	135(26.8%)	6(4.4%)	21(15.6%)	108(80.0%)		7(5.2%)	128(94.8%)		20(14.8%)	36(26.7%)	79(58.5%)
Ph.D.	80(15.9%)	2(2.5%)	9(11.3%)	69(86.3%)		1(1.3%)	79(98.8%)		15(18.8%)	14(17.5%)	51(63.8%)

Chi-Square and Fisher’s Exact test applied: p < 0.05 is considered as statistically significant and p < 0.01 is considered as statistically highly significant

Table 4 presents the correlation matrix for the Knowledge, Attitude, and Practice (KAP) components and the overall KAP score. The findings indicate that each domain is positively and significantly correlated with the others, demonstrating that respondents with higher knowledge scores also tended to exhibit more favorable attitudes and better practice behaviors. The association between practice and the overall score was the strongest of the observed correlations, suggesting that the practical application of AI and evidence-based practice contribute most substantially to the overall performance measure. In comparison, knowledge and attitude showed more moderate correlations, indicating that awareness and perceptions alone may not always translate directly into practice. Overall, the results suggest interconnectedness among the KAP domains, with practice appearing to play a central role within the model.

Table 4.

Correlation analysis among the KAP Model (N = 503)

Correlation		Knowledge Score	Attitude Score	Practice Score	Overall Score
Knowledge Score	r	1	0.384**	0.452**	0.671**
	p value		0.000	0.000	0.000
Attitude Score	r	0.384**	1	0.374**	0.696**
	p value	0.000		0.000	0.000
Practice Score	r	0.452**	0.374**	1	0.899**
	p value	0.000	0.000		0.000
Overall Score	r	0.671**	0.696**	0.899**	1
	p value	0.000	0.000	0.000

**. Karl Pearson’s correlation coefficient (r) is significant, p < 0.01

Table 5 revealed the frequency of evidence sources used by participants in their professional activities. National and international guidelines, as well as continuing education or training workshops, were the most used sources, each reported by 47.3% of respondents. Peer-reviewed journal articles were also frequently consulted, cited by 43.5% of the sample. Institutional protocols and expert opinion were used by 42.1% of participants, indicating that organizational policies and professional judgement continue to play a substantial role in practice. Health information systems or surveillance data were accessed by 39.2% of respondents, while 38.4% relied on colleagues or professional networks. 35.4% of participants reported using journal clubs or academic discussions, and 32.8% reported using AI-based tools or decision-support systems. Grey literature, such as reports and conference papers, was the least commonly used evidence source, accounting for 30.2%.

Table 5.

Frequency distribution with percentage of given options (N = 503)

Which of the following sources of evidence do you regularly use in your professional practice?	Frequency (No.)	Percentage (%)
Peer-reviewed journal articles	219	43.5%
National / International guidelines (e.g., WHO, CDC, MOH)	238	47.3%
Institutional protocols or policy document	212	42.1%
Professional experience or an expert’s opinion	212	42.1%
Health information systems/surveillance data	197	39.2%
AI-based tools or decision-support system	165	32.8%
Continuing education or training workshop	238	47.3%
Colleagues / professional networks	193	38.4%
Grey literature (reports, theses, conference papers)	152	30.2%
Journal clubs or academic discussions	178	35.4%

Table 6 presents the logistic regression analysis examining demographic predictors of attitudes toward AI and evidence-based practice. In this model, the attitude category functions as the dependent variable, while demographic characteristics are included as independent variables. Most demographic variables showed no significant predictive power, as indicated by non-significant p-values across gender, nationality, workplace location, role, institution type, and most age groups. However, individuals aged 35–44 years and 45–54 years had statistically significantly higher odds of reporting a positive attitude compared to those aged 24 years or younger, with odds ratios of 16.13 (95% CI: 1.10–237.51) and 10.81 (95% CI: 1.19–98.66), respectively. Other demographic factors did not reach statistical significance, although some trends were observed, including higher odds among Saudi nationals and among those with more than 10 years of experience. Still, confidence intervals were wide and crossed the null value. Overall, the results suggest that age is the primary demographic factor associated with more favorable attitudes, whereas most other characteristics do not independently account for differences in attitude levels.

Table 6.

Logistic Regression between dependent variables (Attitude categories) and independent variables(N = 503)

Demographic Variables	B	S.E.	Wald	Sig.	Exp(B)	95% C.I. for EXP(B)
						Lower	Upper
Constant	3.278	1.682	3.796	0.051	26.516
Age Group
24 Years	Reference
25–34 Years	0.703	1.502	0.219	0.640	2.020	0.106	38.341
35–44 Years	2.781	1.372	4.106	0.043	16.130	1.095	237.513
45–54 Years	2.381	1.128	4.453	0.035	10.811	1.185	98.659
55 + Years	1.659	1.056	2.467	0.116	5.252	0.663	41.602
Gender
Female	Reference
Male	0.241	0.581	0.171	0.679	1.272	0.407	3.977
Nationality
Non-Saudi	Reference
Saudi	1.384	0.816	2.873	0.090	3.989	0.806	19.751
Primary location
Rural Village	Reference
Small City	− 0.953	0.913	1.089	0.297	0.386	0.064	2.309
Urban City	0.416	0.627	0.441	0.507	1.517	0.443	5.186
Current Role
Faculty	Reference
Nurse	-1.417	0.975	2.112	0.146	0.242	0.036	1.639
Other allied health professions	0.988	0.983	1.009	0.315	2.686	0.391	18.456
Physician	0.062	0.772	0.006	0.936	1.064	0.234	4.828
Public Health Specialist	− 0.667	0.732	0.830	0.362	0.513	0.122	2.156
Types of Institution
Govt Hospital	Reference
PHC	-1.179	1.053	1.254	0.263	0.308	0.039	2.421
Private Hospital	− 0.596	1.208	0.243	0.622	0.551	0.052	5.878
University Teacher	− 0.680	1.131	0.361	0.548	0.507	0.055	4.650
Years of Experience
2–6 Year	Reference
7–10 Year	0.419	0.883	0.226	0.635	1.521	0.270	8.579
10 + Year	1.548	0.992	2.435	0.119	4.700	0.673	32.836
Highest Academic Qualification
Diploma	Reference
Bachelors	− 0.706	1.569	0.202	0.653	0.494	0.023	10.697
Masters	-2.061	1.246	2.736	0.098	0.127	0.011	1.464
Ph.D.	-2.256	1.189	3.600	0.058	0.105	0.010	1.077

Table 7 shows the ordinal regression analysis assessing the relationship between demographic characteristics and the ordered categories of knowledge and practice, with the highest category serving as the reference. The results indicate that years of professional experience and academic qualifications were statistically significant determinants of knowledge level. Participants with 2–6 and 7–10 years of experience had significantly higher odds of being classified in lower knowledge categories than those with more than 10 years of experience. This finding differs from the bivariate results presented in Table 3, where participants with fewer years of experience appeared to explain higher knowledge levels. This discrepancy likely reflects the influence of confounding variables, specifically academic qualification and professional role, which were controlled for in the ordinal regression model. Similarly, individuals holding diploma- or bachelor-level qualifications were substantially more likely to be classified in lower knowledge groups than those with doctoral or master-level degrees. In contrast, most demographic characteristics were not associated with knowledge at a statistically meaningful level, including age, gender, nationality, workplace location, or institution type. Regarding practice, the only variable that showed a significant association was professional role: nurses had markedly higher odds of reporting lower practice levels than public health specialists. No other demographic variables were significantly associated with practice scores.

Table 7.

Ordinal regression analysis in demographic variables with knowledge and practice categories with high categories as reference categories (N = 503)

Demographic Variables	Knowledge						Practice
	Estimate	Std. Error	Wald	95% Confidence Interval		p value	Estimate	Std. Error	Wald	95% Confidence Interval		p value
				Upper Bound	Lower Bound					Upper Bound	Lower Bound
Threshold
Low	-2.486	0.788	9.950	-4.031	− 0.941	0.002	-3.182	0.910	12.240	-4.965	-1.399	0.000
Moderate	− 0.632	0.779	0.659	-2.159	0.894	0.417	0.059	0.876	0.005	-1.658	1.776	0.946
Age Group
24 Years	1.419	0.831	2.913	− 0.210	3.048	0.088	1.541	0.963	2.565	− 0.345	3.428	0.109
25–34 Years	0.272	0.646	0.177	− 0.993	1.537	0.674	1.000	0.784	1.627	− 0.536	2.536	0.202
35–44 Years	0.598	0.596	1.006	− 0.571	1.767	0.316	0.543	0.726	0.561	− 0.879	1.966	0.454
45–54 Years	0.649	0.590	1.209	− 0.508	1.806	0.271	0.984	0.728	1.827	− 0.443	2.411	0.176
55 + Years	0a						0a
Gender
Male	0.275	0.210	1.716	− 0.137	0.687	0.190	− 0.220	0.242	0.830	− 0.695	0.254	0.362
Female	0a						0a
Nationality
Saudi	− 0.032	0.280	0.013	− 0.580	0.517	0.910	0.215	0.302	0.507	− 0.376	0.806	0.476
Non-Saudi	0a						0a
Primary location
Rural Village	− 0.119	0.460	0.067	-1.020	0.783	0.796	− 0.158	0.521	0.092	-1.179	0.863	0.762
Small City	− 0.070	0.218	0.102	− 0.497	0.358	0.750	− 0.515	0.247	4.367	− 0.999	− 0.032	0.037
Urban City	0a						0a
Current Role
Faculty	0.386	0.519	0.554	− 0.631	1.403	0.457	0.478	0.496	0.928	− 0.494	1.451	0.335
Nurse	0.436	0.330	1.744	− 0.211	1.082	0.187	1.394	0.401	12.077	0.608	2.181	0.001
Other allied health professions	− 0.369	0.308	1.438	− 0.971	0.234	0.230	0.648	0.368	3.099	− 0.074	1.370	0.078
Physician	− 0.315	0.300	1.106	− 0.902	0.272	0.293	0.425	0.348	1.493	− 0.257	1.106	0.222
Public Health Specialist	0a						0a
Institution Types
Govt Hospital	0.155	0.433	0.128	− 0.693	1.002	0.721	0.427	0.463	0.848	− 0.481	1.335	0.357
PHC	− 0.259	0.447	0.336	-1.136	0.617	0.562	0.222	0.493	0.202	− 0.744	1.188	0.653
Private Hospital	0.396	0.480	0.680	− 0.545	1.336	0.410	0.930	0.521	3.189	− 0.091	1.951	0.074
University Teacher	0a						0a
Years of Experience
2-6 Years	0.938	0.343	7.504	0.267	1.610	0.006	− 0.223	0.368	0.368	− 0.945	0.499	0.544
7–10 Years	0.939	0.303	9.580	0.344	1.534	0.002	0.366	0.327	1.249	− 0.276	1.007	0.264
10+ Years	0a						0a
Highest Academic Qualification
Diploma	-2.074	0.482	18.507	-3.020	-1.129	0.000	− 0.727	0.523	1.933	-1.751	0.298	0.164
Bachelors	-1.715	0.421	16.602	-2.540	− 0.890	0.000	− 0.217	0.419	0.268	-1.039	0.605	0.605
Masters	− 0.353	0.440	0.645	-1.215	0.509	0.422	− 0.164	0.396	0.171	− 0.941	0.613	0.679
Ph.D.	0a						0a

a: This parameter is set to zero because it is redundant, which means reference categories

Overall, the model suggests that knowledge disparities are primarily shaped by educational attainment and length of experience. In contrast, practice behavior appears to be influenced mainly by the nature of the professional role.

Hence, the adjusted regression results indicate a more relevant interpretation of the determinants of knowledge and practice compared with the unadjusted bivariate analysis shown in Table 3.

Discussion

This study examined healthcare providers’ knowledge, attitudes, and practices regarding the integration of EBP with AI in Saudi Arabia. Overall, participants acknowledged the potential value of AI-supported EBP and expressed willingness to enhance their competencies; however, routine integration into daily professional practice remains limited. In line with these findings, moderate levels of knowledge and generally positive attitudes were observed, whereas practice levels were comparatively lower. This gap is important, as EBP is closely associated with improved patient outcomes and healthcare system performance, and AI is increasingly viewed as a key tool for enhancing access to, synthesis of, and application of evidence in clinical care [1, 5].

A key finding within the knowledge domain is that many respondents demonstrated familiarity with the general concept of AI supporting healthcare decision-making yet reported lower confidence in the technical and interpretive skills required for its safe application. Challenges were noted in evaluating AI-generated outputs within real clinical contexts and appropriately integrating them with professional judgment. This distinction is important, as effective and safe use of AI depends on understanding uncertainty, recognizing system limitations, and discerning when AI support is beneficial versus potentially misleading. Similar patterns have been documented in both Saudi and international studies, where awareness of AI continues to increase, while practical readiness and confidence vary across professional roles and healthcare settings [5, 12, 15].

Attitudes toward AI and EBP were generally positive, with many participants emphasizing the importance of integrating AI and EBP into healthcare education and expressing interest in further training. At the same time, concerns related to ethics, privacy, and governance were commonly reported. These concerns should be viewed as appropriate caution rather than resistance, particularly as health systems work to establish clearer standards for transparency, accountability, data protection, and the equitable use of AI technologies. Consistent with this perspective, the WHO guidance highlights that responsible AI adoption requires robust ethical and governance frameworks, including safeguards to ensure privacy, fairness, and accountability [30, 31].

Despite positive attitudes, practice remained comparatively limited, suggesting that AI-assisted EBP is not yet embedded into routine workflows for many providers. As shown in Table 2, the lowest-scoring practical behavior reflected limited use of AI for higher-stakes tasks such as predicting health outcomes or disease trends (mean 2.02), which may indicate cautious engagement with applications that require stronger validation, clearer accountability, and smoother integration into clinical processes. This staged adoption aligns with broader evidence that AI delivers value when validated, embedded into workflow, and used to complement, not replace, clinical judgment and contextual decision-making [10, 32, 33].

The differences found across professional and workplace characteristics support the interpretation that uptake is shaped by more than individual motivation alone. Bivariate analyses demonstrated significant associations between knowledge and variables such as professional role, years of experience, institution type, and academic qualification. Although bivariate results showed a higher knowledge among those with 2–6 years of experience, this pattern Inverted after adjustment. The regression indicated that participants with 2–6 and 7–10 years of experience were more likely to have lower knowledge than those more than 10 years. This shift is principally described by academic qualifications, a key confounder, since higher experienced participants were more likely to hold advanced degrees linked with advanced knowledge. while practice varied significantly by experience and qualification. Knowledge and practice were not uniform across the workforce, reflecting uneven readiness. Organizational conditions such as protected time, access to resources, leadership support, and practical workflow fit are well-established drivers of EBP adoption and likely shape AI use as well. Evidence from implementation research and nursing practice environments similarly highlights that real-world uptake improves when practical barriers (time, access, workflow integration) are addressed alongside training [3, 6, 7, 20, 23].

Findings from adjusted regression analyses further suggest that “who you are in the system” can influence practice more than general enthusiasm. In the ordinal regression model, professional role emerged as a significant predictor of practice level, with nurses having significantly higher odds of falling into lower practice categories compared with the reference group. This highlights the need for role-sensitive implementation approaches that reflect differences in decision authority, access to digital tools, workload pressures, and opportunities for evidence work. Rather than interpreting this as a deficiency, it should be viewed as a signal to tailor support through practical training, structured access to approved tools, mentorship, and supportive workplace routines to the realities of each professional group [6, 7, 20, 22].

Evidence-source patterns also provide context for how EBP is currently enacted. Participants reported relying primarily on clinical guidelines and continuing professional education, alongside peer-reviewed literature and institutional protocols, while routine use of AI-based decision-support tools was less common. This is reassuring because guideline-based care remains central to EBP, but it also indicates that AI pathways for routine clinical use are still developing and may require clearer institutional endorsement, governance, and integration into everyday systems. These findings align with ongoing national digital health transformation priorities, where progress is substantial, but readiness varies across settings [1, 11, 17].

Finally, the relationships among KAP domains suggest that knowledge, attitudes, and practice move together, but not perfectly. Higher knowledge tended to align with more favorable attitudes and better practice, yet practice appears to be the key step that determines whether readiness translates into real impact. Correlation analysis demonstrated significant positive associations among all domains, with the strongest relationship observed between practice and the overall KAP score. These findings suggest that while knowledge and attitudes provide an essential foundation, practice represents the critical step through which readiness is converted into meaningful change in healthcare delivery. Taken together, the findings indicate that Saudi healthcare providers are receptive to AI-integrated EBP, but routine practice remains constrained by practical, organizational, and governance conditions. A realistic way ahead is a combined strategy: build interpretive competence and responsible-use skills through targeted training, and simultaneously strengthen enabling environments through approved tools, workflow integration, protected time, mentorship, and clear ethical governance, so that positive attitudes can translate into safe, consistent, and sustainable practice [3, 6, 7, 31, 32].

This study offers several notable strengths alongside important limitations to consider. Including a nationwide sample from a wide range of professions across the government and private sectors enhances the generalizability of the findings. Employing a structured KAP framework with acceptable to good internal consistency supports the reliability of the measurement tools, and presenting both bivariate and adjusted regression analyses strengthens the interpretation of independent predictors. Additionally, examining sources of evidence and practical AI-related behaviors adds applied relevance beyond attitudes alone. Finally, the questionnaire which is used in this study was developed for the purpose of this study, as no validated KAP instrument combining artificial intelligence and evidence-based practice currently exists during our study. Although internal consistency was assessed using Cronbach’s alpha, factor analysis was not measured to establish construct validity, which considered as limitation in the psychometric robustness of the instrument.

However, first, this cross-sectional design precludes causal inference regarding the relationship between participants’ characteristics and their KAP toward integrating EBP and AI. Longitudinal or experimental designs are required to determine temporality and causality. The sampling approach, based on non-probability snowball sampling, facilitates rapid, wide-reaching recruitment across multiple professions and sectors, but introduces potential selection bias and constrains external validity. Participants may share characteristics such as high engagement with technology or interest in AI/EBP-related topics, which may not reflect the broader student or professional populations. Although the study achieved nationwide breadth, the lack of random sampling limits generalizability, and future work could incorporate probability-based sampling or apply post hoc weighting if population benchmarks are available. All measures relied on self-report, which is vulnerable to social desirability, recall, and misreporting biases; even with structured KAP instruments, respondents might overstate favorable AI- and EBP-related practices, potentially inflating effect estimates. Construct validity may be challenged if AI-related practices and EBP adoption are multidimensional or context-dependent, and measurement error could attenuate associations or obscure nuanced relationships. Reporting evidence of validity (e.g., convergent validity, factor structure) and conducting sensitivity analyses with alternate operationalizations would strengthen interpretation. Additionally, the questionnaire was newly developed, and while internal consistency and inter-domain correlations were acceptable, confirmatory factor analysis (CFA) was not conducted to formally establish construct validity. Future studies should validate the instrument across diverse populations and settings. Furthermore, the snowball sampling methodology prevents the determination of the total denominator, and therefore, a true response rate cannot be calculated. This limits the assessment of non-response bias and introduces selection bias, thereby limiting generalizability. Finally, the very low number of participants with negative attitudes may have affected the stability of the logistic regression model, as reflected in the wide confidence intervals for some estimates.

Conclusion

Healthcare providers in Saudi Arabia showed moderate knowledge and positive attitudes toward integrating AI with EBP, but routine practice remains limited. This reflects some KAP gaps, where willingness exists but daily use is restricted by confidence, access, workflow fit, and governance. Knowledge and practice varied across professional roles, experience, institution types, and qualifications, suggesting uneven readiness across the workforce. Role-sensitive training, plus institutional enablement (time, resources, mentorship, governance), is needed to support safe, consistent integration into routine care.

Abbreviations

EBP

Evidence-based practice

AI

Artificial Intelligence

KAP

Knowledge, attitudes, and practices

WHO

World Health Organization

SDAIA

The Saudi Data and Artificial Intelligence Authority

SCFHS

Saudi Commission for Health Specialties

ORs

Odds ratios

CIs

95% confidence intervals

Author contributions

A.A., visualization, investigation, writing – review & editing, funding acquisition, data curation; A.M., methodology, data curation, and writing-orginal draft; K.M., visualization, investigation, writing – original draft, data curation; H.A., formal analysis and writing – original draft; M.S.K., investigation, data curation, and writing – original draft; M.A., methodology, data curation, and writing-original draft; M.M.H., investigation and writing – original draft; R.M.A., data curation, and writing-original draft; A.D., visualization, investigation, methodology, writing – review & editing., data curation. All authors reviewed and approved the final manuscript prior to submission.

Funding

This study was funded by the Deanship of Scientific Research at King Khalid University through the Small Group Research Project under Grant Number RGP1/150/46.

Data availability

Data are available upon reasonable request. All data relevant to the study are available upon request for academic research.

Declarations

Ethics approval and consent to participate

Ethical approval for this study was obtained from the Research Ethics Committee at King Khalid University (Approval No. KKU-88-2025-25). The research was carried out in accordance with the principles outlined in the Declaration of Helsinki. Informed consent was obtained through the first page of the online questionnaire, which provided participants with clear information about the study’s objectives, the voluntary nature of participation, and data confidentiality. Proceeding with the survey was regarded as implied consent to participate.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.