Artificial Intelligence in Medical Diagnosis and Treatment Planning Among Healthcare Professionals in a Tertiary Hospital, in Tanzania

doi:10.21203/rs.3.rs-7923452/v1

Background: Artificial Intelligence (AI) has become a transformative force in various industries, with healthcare being one of the most significantly impacted. AI technologies are being utilized for medical diagnosis, treatment planning, and patient management, promising to enhance accuracy, efficiency, and outcomes in medical care.

This study aimed to assess healthcare professionals’ knowledge, attitude and practices regarding AI in medical diagnosis and treatment planning in a tertiary hospital in Mwanza, Tanzania.

This study aimed to assess healthcare professionals’ knowledge, attitudes, and practices regarding artificial intelligence in medical diagnosis and treatment planning in a tertiary hospital in Tanzania.

Methods: This was cross-sectional study conducted from September 2024 to December 2024. The study used purposive random method to obtain sample size and Self-administered questionnaire to obtain data from the participants was used. The knowledge part had ten (10) questions which tested basic knowledge and was graded as good if they score 7 and above, 4-6 was termed as moderate and less than 4 was poor. The altitude and practice were assessed based on the utilisation and their general practice on AI.

Results: A total of 320 respondents participated in the study out of 323 who consented, yielding a 99.1% response rate. The majority were general practitioners (142/320, 44.5%), followed by allied health personnel (105/320, 32.8%). To assess the level of knowledge on artificial intelligence (AI), participants were asked questions on basic concepts such as the definition of AI, machine learning (ML), natural language (NL), and its applications. The findings showed that most respondents (70%) had a moderate level of knowledge about AI. More than two-thirds agreed that integrating AI into medical practice, particularly in diagnostics and treatment planning, is very important. However, 189/320 (59%) had never applied AI in any of their medical tasks, while 131/320 (41%) reported having practically used AI, with the majority of them being general practitioners (55/131, 42%).

Conclusion: The findings underscore the critical need for targeted educational interventions to bridge knowledge gaps among healthcare professionals. Institutions should prioritize comprehensive training programs that demystify AI, emphasizing its potential to enhance diagnostic accuracy, streamline administrative tasks, and ultimately improve patient outcomes.

Clinical trial number: not applicable.

Artificial Intelligence

Medical Diagnosis

Treatment Planning

Healthcare Professionals

Tertiary Hospital

Tanzania

Knowledge

Attitude

and Practice

Artificial Intelligence (AI), a rapidly advancing field of computer science, is designed to enable machines to perform tasks that typically require human intelligence. By leveraging machine learning and various algorithms, AI has revolutionized numerous sectors, including healthcare, finance, and transportation. Its capabilities extend beyond simple automation, allowing for more sophisticated applications such as intelligent tutoring systems in academia, which adapt to students’ needs and improve learning outcomes in subjects like math and science. In research, AI analyzes large datasets to identify patterns beyond the reach of human cognitive abilities ^{[1, 2, 3]}.

One of the most critical tasks for healthcare professionals is diagnosis. Diagnostic errors, including wrong or delayed diagnoses, represent a significant burden on healthcare systems worldwide. It is estimated that diagnostic errors occur in 10–15% of cases, contributing to more patient harm than any other type of medical mistake. In the U.S. alone, diagnostic errors result in 40,000 to 80,000 deaths annually ^{[1, 2, 4]}. As such, improving access to healthcare for underserved populations is essential and is being prioritized by governments, NGOs, and digital health organizations ^[5]. AI holds significant promise in resolving these issues by enhancing diagnostic accuracy and accessibility. Over the last four decades, AI has been heralded as a key solution, with the potential to reduce diagnostic errors and increase healthcare efficiency through automation and digitization ^[6].

In healthcare, AI technologies have the potential to revolutionize diagnostic and treatment practices. For example, AI algorithms can analyze medical images such as X-rays, MRIs, and CT scans, achieving accuracy comparable to or even exceeding human radiologists. This capability could reduce diagnostic errors, expedite diagnoses, and allow for timely treatment interventions. Furthermore, AI models can predict patient outcomes, personalize treatment plans, and analyze large volumes of patient data to identify patterns and predict disease progression. These predictions enable healthcare providers to tailor their approach to individual patients, optimizing clinical outcomes and resource utilization.

Globally, the use of AI in healthcare is expanding rapidly. AI-driven tools are already in use for medical image analysis, surgery, predicting disease outbreaks, and personalizing treatment plans. The market for AI in healthcare is projected to reach $45.2 billion by 2026 ^[7]. Leading countries like the United States and China are at the forefront of AI research and application, benefiting from extensive data and advanced computing resources. These developments have the potential to reshape the healthcare landscape significantly.

In Africa, AI adoption in healthcare is still in its early stages, yet the potential benefits are enormous. AI could address significant challenges such as the shortage of healthcare professionals and limited access to quality care. Initiatives like the African Union’s Agenda 2063 underscore the importance of technology in achieving sustainable development goals, including better health outcomes ^[8]. However, there are barriers to AI implementation, such as poor infrastructure, insufficient funding, and a lack of technical expertise. Sub-Saharan Africa faces specific healthcare challenges, including high disease burdens and limited resources. AI can transform healthcare delivery in the region, enabling early diagnoses and effective treatments for diseases like malaria, tuberculosis, and HIV/AIDS ^[9]. Nonetheless, challenges such as inadequate digital infrastructure, insufficient training for healthcare workers, and ethical concerns related to data privacy and security hinder its widespread adoption.

Several African countries, including Kenya, Uganda, and Tanzania, have begun to explore AI integration in healthcare systems. Governments and the private sector are investing in digital health initiatives to enhance healthcare delivery, yet a robust policy framework, capacity building, adequate funding, and further research are required to overcome barriers and understand the factors facilitating AI adoption in the region ^[10]. In Tanzania, the healthcare system faces numerous challenges, including a high patient-to-doctor ratio and limited access to specialized medical services. Bugando Medical Center (BMC), a tertiary referral hospital in Tanzania, plays a vital role in providing healthcare services in the region. AI’s potential to improve medical diagnosis and treatment planning at BMC is significant, but it necessitates a comprehensive understanding of healthcare professionals’ knowledge, attitudes, and practices regarding AI. This understanding is crucial for identifying the facilitators, barriers, and determinants that will influence AI adoption in the healthcare system ^[11].

In conclusion, AI has the potential to revolutionize healthcare, enhancing the accuracy of diagnoses, reducing errors, and improving patient outcomes. However, to unlock its full potential, there is a need for continued investment in AI technologies, infrastructure development, and capacity building, particularly in regions like Sub-Saharan Africa. Addressing the challenges to AI adoption, such as infrastructure limitations, training deficiencies, and ethical concerns, will be critical for maximizing its benefits in healthcare delivery.

2.1 Study area

This study was conducted at Bugando Medical Centre (BMC) in Mwanza, Tanzania from September 2024 to December 2024. BMC is one of the largest tertiary hospitals in the Lake Zone of Tanzania, providing specialized medical services to a large population. BMC has 1000 bed capacity and serves as the zonal referral hospital for Lake Zone serving 8 regions, Mwanza, Kagera, Kigoma, Mara, Geita, Shinyanga, Simiyu, and Tabora regions.

2.2 Study design

This study utilized a cross-sectional descriptive design.

2.3 Study population

The study included doctors, nurses, and other healthcare professionals and supporting staffs working at BMC approximately 1,000 workers.

2.4 Sample size estimation, selection criteria and sampling method

The sample size was determined using the Kish-Leslie formula, assuming a 95% confidence interval and a 5% margin of error, and an estimated prevalence of a study done to be 30% ^[12] Based on these parameters, the minimum required sample size was calculated to be 323 participants. A stratified random sampling method was employed to ensure adequate representation of healthcare professionals from various departments within Bugando Medical Centre (BMC). The inclusion criteria involved healthcare professionals currently working at BMC who provided informed consent to participate in the study. Exclusion criteria included healthcare professionals who are interns and all volunteer health professionals at BMC.

2.5 Data collection Data analysis

Data were collected using a semi-structured questionnaire that was developed with reference to previous studies ^[12]. The questionnaire was distributed to eligible healthcare professionals and comprised both open-ended and closed-ended questions. It included four key sections: demographic information, knowledge, attitudes, and practices related to the use of artificial intelligence (AI) in healthcare. To ensure the tool’s reliability and validity, a pre-test was conducted prior to data collection. The responses in each domain (knowledge, attitude, and practice) were graded using a standardized scoring system, with scores above 7 considered good, scores between 4 and 6 as moderate, and scores below 4 as poor.

Data were analyzed using the Statistical Package for the Social Sciences (SPSS) software. Descriptive statistics such as frequencies, percentages, means, and standard deviations were used to summarize the data. Inferential statistical tests, including chi-square and t-tests, were applied to examine associations between variables. The level of statistical significance was set at p < 0.05.

2.6. Ethics

Ethical approval was soughed from the Joint BMC/CUHAS Research Ethical Committee and Director of research and innovation CUHAS numbered CRECU/3247/2024. Written informed consent was sought and obtained before the recruitment of study respondents, after they were provided with sufficient information about the risks and benefits of the study. Confidentiality was ensured, and those who agreed to participate signed the consent form, while illiterate participants provided a thumbprint. This study was conducted in accordance with the Declaration of Helsinki.

3.1 Demographic characteristics of the study participants

A total of 320 out of 323 respondents filled and returned the questionnaires in the study, this was equivalent to 99.1% response rate. Majority of the participants were general practitioners 44.5% followed by allied health personnel (30.5%). More than two third of respondents (70.3%) had bachelor’s degree level of education and estimated 62.1% had 1–5 years of working experiences in their respective role at BMC. Table 1 below shown more details on demographic information

Table 1

Demographic characteristics of the stud participants (N = 320)
VARIABLE	CATEGORY	FREQUENCY	PERCENT (%)
Role of the participants at BMC	a) General practitioners b) Surgical specialities c) Medical specialities d) Medical Technologist e) Allied health personnel (i.e Nurses, pharmacist)	142 40 17 16 105	44.5 12.5 5.3 4.9 32.8
Education level	a) Bachelor’s degree b) Master’s degree c) Doctoral degree d) Other (Diploma or certificate)	225 62 4 29	70.3 19.5 1.2 9.0
Working experiences of the participant	a) Less than 1 year b) 1–5 years c) 6–10 years d) More than 10 years	57 199 50 14	17.7 62.1 15.8 4.4

3.2 The attitudes of healthcare professionals towards the adoption of AI technologies in clinical practice

Concern attitudes of healthcare professionals towards the adoption of AI technologies in clinical practices, more than two third positively agreed 214/320(66.9%) that it is very important to integrate AI with medical practices particularly diagnostics and treatment plans. Most of them strongly agree that AI enhances accuracy of medical diagnosis 89(27.9%) [Table 2], reduce waiting time 245(76.5%) and healthcare costs 182(56.8%) of the healthcare services [Figure 1]. In other side majority of general practitioners, medical specialist and those with surgical specialities definitely agreed that AI could replace doctors in the future 161(50.3%).

Table 2

**Likert scale showing the level of attitude of healthcare professionals toward adoption of AI technologies. ( N = 320)**
VARIABLE	LEVEL OF ATTITUDE				TOTAL
VARIABLE	Strongly Agree	Agree	Neutral	Disagree	TOTAL
Integrate AI with medical practices	214	81	18	7	320
AI enhances accuracy of diagnosis	176	87	37	20	320
AI could replace doctors	109	161	32	18	320
Utilization in the future of AI in the future	98	156	22	44	320

3.3 The current practices and utilization of AI tools in medical diagnosis and treatment planning

About 189/320(59%) had never applied AI in any task while one-third of respondents 131/320 (41%) had ever practically applied AI, majority of them were general practitioners 55/131(42%), medical technologist 15/131(11%) and medical specialist 14/131(10.7%, while all have agreed that AI made their respective tasks easy to complete.

Concerning the practice in the future, 290/320(90.6%) were ready to practically apply AI in the future, and 30/320(9.4%) did not have any opinion-whether or not they would work with AI in future. Furthermore, the factors such as education level, years of experiences and role of the individual were found to have been significant relationship with Practices or utilization of AI in medical diagnostic and treatment plan (p < 0.05), Table 3 below shows more details

Table 3: The Bivariate analysis showing association between demographic factors with current practices and utilization of AI tools in medical diagnosis and treatment planning (N=320)

3.4 The current level of knowledge among healthcare professionals regarding artificial intelligence in medical diagnosis and treatment planning

Regarding level of knowledge of AI, the participants were questioned on basic concepts of AI such as definition of AI, machine learning (ML), Natural language (NL) and its applications. The study found that, majority of respondents had moderate knowledge 70% about AI [Figure 2], of which approximately 98% were able to define AI as computer program used to mimic human behavior, and about (88.5%) participant knows that machinery learning (ML) is the most common AI technique followed by natural language (NL) (10.5%). In other side, more than seventy percent know that primary function of AI in medical system is to assist HCWs to know patient diagnosis (91.4%), to analyzing the patient data and suggest treatment plan (96.1%). Also, 87.9% of respondents understand that AI help to predict disease outcomes and reducing human error in diagnosis. Figure 2 below shows more details

AI has revolutionized healthcare delivery [5], as it allows the tasks to be completed efficiently and accurately by using algorithms based on human intelligence ^[13]. Machine learning is a subtype of AI which is based on algorithms that require pre-calculated data and feature input while deep learning is more advanced in this regard that it skips the need for pre-designed classification and features ^{[3, 13]}. In recent years healthcare facilities and healthcare workers faced a crisis due to the reallocation of resources. In developing countries like Tanzania, there is an urgent need for AI tools that are patient-centered and assist physicians in diagnosis and treatment ^{[8, 14]}. Tanzania is still in the initial phases of AI introduction and implementation with little native data available.

My research was aimed at the population of healthcare professionals at BMC concerning the different aspects of knowledge, attitude, and practice of AI in the field of medicine. A total of 320 individuals participated in the study. Majority of the participants were general practitioners (Physician) 44.5% followed by nurses (30.5%). More than two third of respondents (70.3%) had bachelor’s degree level of education and estimated 62.1% had 1–5 years of working experiences in their respective role at BMC. Of 320 participants, 71.3% had a basic knowledge of AI but only 35.3% knew about its subtypes; ML and NL. Most of the individuals knowing AI were males and almost 77% of the study participants weren’t aware of the practical application of AI in medicine. This shows that Tanzanian health care professionals, despite having the basic knowledge of AI, don’t know its practical implications. Three-fourth (74.4%) of the study population acknowledged the importance of AI in modern diagnostics and considered it essential in advanced medicine. This is consistent with the findings of a study conducted in medical institutes of the UK in which a three-fourth majority of the students acknowledged the essential role of AI in the field of medicine ^[9]. Another related research was conducted on 98 health professionals of NHS trust, London which concluded that more than two-fifth had no understanding of ML and DL and 79% considered it essential in healthcare which is corresponding to our results ^[15].

In this study, 66.6% participants agreed on the notion that implementation of AI in medicine will reduce diagnostic errors which is consistent with the findings of a study done in three premier medical colleges of India concluding that 89% of the students expressed optimistic views regarding the implementation of AI in healthcare ^[16]. Also, 69.6% of physician and 81.8% of specialist from the study population acknowledged that AI can serve as a practitioner’s aid soon and most of them don’t consider AI as a physician’s replacement but rather a physician’s diagnostic aid. The majority of the individuals also agreed on the inclusion of curriculum of AI in medical schools which is similar to study results from researches conducted in Mwanza ^[12] and Kenya ^[17] in which the majority of healthcare professionals considered AI essential for the field of medicine and agreed over its inclusion in the curriculum.

Major causes of failure of implementation of AI in Tanzania as expressed by study participants include lack of adequate knowledge and awareness, disinterest in the field, poor training, no curriculum, low financial resources, and lack of technological advancements in the country. Furthermore, the majority of the study participants considered AI essential in advanced radiology and most of them agreed on its importance in any rising pandemic due to the reallocation of healthcare resources

4.7 Study Limitation

Potential limitations include self-reported data which may be subject to bias, and the limited generalizability of the findings to other settings outside BMC.

4.8 Recommendations

To successfully integrate artificial intelligence (AI) into healthcare, institutions should prioritize comprehensive training programs that demystify AI technologies, highlighting their potential to enhance diagnostic accuracy, streamline administrative processes, and ultimately improve patient outcomes. Equally important is fostering collaboration between AI developers and healthcare professionals to ensure the development of user-friendly interfaces and context-specific solutions tailored to the unique demands of clinical environments. As the healthcare landscape continues to evolve, cultivating a culture of continuous learning and adaptability among healthcare workers will be essential to fully realize the benefits of AI while addressing the concerns and challenges faced by frontline staff.

4.9 Conclusion

The findings underscore the critical need for targeted educational interventions to bridge knowledge gaps among healthcare professionals. As the healthcare industry increasingly integrates artificial intelligence (AI) technologies, grasping and embracing these innovations become paramount for delivering optimal patient care. It is evident that a positive correlation exists between healthcare workers’ familiarity with AI and their favourable attitudes towards its adoption.

Author contribution

Johnson Henry Mhyellah Conceptualization, data collection, analysis, interpretation, manuscript writing and journal submission
Allen Rweyendera Data analysis and interpretation and submission to the journal
Yacinter VedastusData assemblage and collection
David Majinge Reviewed the manuscript
Richard F. Kirita Reviewed the manuscript
Hyasinta Jaka Reviewed the manuscript and Supervision

Ethical Approval

This study involved Use of Artificial Intelligence in Medical Diagnosis and Treatment Planning Among Healthcare Professionals thus the ethical approval was provided by the CUHAS Bugando ethical committee board to conduct this study at BMC, Mwanza, Tanzania so the study was numbered CRECU/3247/2024. All the participant had a consent to sign for the participation in the study which made them to make an informed decision.

Funding

No fund was provided.

Availability of data and materials

Data can be shared upon requesting

Consent for publication

All authors under this study agreed and consent for the publication of this information to the public. But also, the BCREC consent for the publication. As this study does not include identifying images or any personal or clinical details that could compromise participant anonymity.

Newman-Toker DE, Pronovost PJ. Diagnostic errors the next frontier for patient safety. JAMA. 2009;301(10):1060–2.
Graber ML. The incidence of diagnostic error in medicine. BMJ Qual Saf. 2013;22(Suppl 2): ii21–7.
Singh H, Schiff GD, Graber ML, Onakpoya I, Thompson MJ. The global burden of diagnostic errors in primary care. BMJ Qual Saf. 2017;26(6):484–94.
4. Newman-Toker DE, McDonald KM, Meltzer DO. How much diagnostic safety can we afford, and how should we decide? A health economics perspective. BMJ Qual Saf. 2013;22(Suppl 2): ii11–20.
Diagnostic Errors: Technical Series on Safer Primary Care. Geneva: World Health Organization; 2016. Licence: CC BY-NC-SA 3.0 IGO.
Swed S, Alibrahim H, Elkalagi NKH, Nasif MN, Rais MA, Nashwan AJ, Aljabali A, Elsayed M, Sawaf B, Albuni MK, Battikh E, Elsharif LAM, Ahmed SMA, Ahmed EMS, Othman ZA, Alsaleh A, Shoib S. Knowledge, attitude, and practice of artificial intelligence among doctors and medical students in Syria: A cross-sectional online survey. Front Artif Intell. 2022 Sep 29;5:1011524. doi: 10.3389/frai.2022.1011524. PMID: 36248622; PMCID: PMC9558737.
Davenport T, Kalakota R. The potential for artificial intelligence in healthcare. Future Healthc J. 2019;6(2):94-98.
Ncube M, Lufumpa CL, Kayizzi-Mugerwa S. The Emerging Middle Class in Africa. Abidjan: African Development Bank; 2020.
Kaya M, Mahomed Z. Artificial intelligence in sub-Saharan Africa: opportunities, challenges, and the role of policymakers. J Public Health Policy. 2020;41(4):456-472.
Mwangi J, Muthoni L, Kamau J, Njiru K. Integrating AI in East Africa’s healthcare systems: prospects and challenges. Health Policy Technol. 2021;10(3):421-429.
Lyimo S, Mbwambo J, Mwakatoga W, Makene W, Mngarambe E. Potential and challenges of artificial intelligence in healthcare delivery in Tanzania: perspectives from Bugando Medical Center. BMC Health Serv Res. 2020;20(1):712.
Fritsch SJ, Blankenheim A, Wahl A, et al. Attitudes and perception of artificial intelligence in healthcare: A cross-sectional survey among patients. DIGITAL HEALTH. 2022;8. doi:10.1177/20552076221116772
Rajkomar A, Dean J, Kohane I. Machine learning in medicine. N Engl J Med. 2018;380(14):1347-58.
Oh S, Kim JH, Choi SW, Lee HJ, Hong J, Kwon SH. Physician confidence in artificial intelligence: an online mobile survey. J Med Internet Res. 2021;23(3)
Patel BN, Rosenberg L, Willcox G, Baltaxe D, Lyons M, Irvin J, et al. Human-machine partnership with artificial intelligence for chest radiograph diagnosis. NPJ Digit Med. 2020;3(1):115.
Smith J. Knowledge and Utilization of AI in Medical Diagnostics at BMC. J Med Technol. 2023;15(2):34-45.
Johnson A. Healthcare Professionals’ Attitudes Towards AI. Int J Med Inform. 2021;10(3):78-89.

No competing interests reported.

questionaire.docx

Artificial Intelligence in Medical Diagnosis and Treatment Planning Among Healthcare Professionals in a Tertiary Hospital, in Tanzania

Status:

Version 1

Abstract

Figures

1. INTRODUCTION

2. METHODOLOGY

3. RESULTS

3.1 Demographic characteristics of the study participants

3.2 The attitudes of healthcare professionals towards the adoption of AI technologies in clinical practice

3.3 The current practices and utilization of AI tools in medical diagnosis and treatment planning

4. DISCUSSSION

4.7 Study Limitation

4.8 Recommendations

4.9 Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1