📄 medicine

Dataset of survey responses on Artificial Intelligence adoption in the healthcare sector of Bangladesh: Stakeholder perspectives from patients, providers, and administrators

This study presents a dataset of survey responses from Bangladeshi healthcare stakeholders, revealing moderate to positive perceptions and readiness for AI adoption driven by technological awareness, personal innovativeness, and social media influence, thereby offering evidence to support policy formulation and implementation of AI-driven healthcare services in Bangladesh.

Original authors: Md. Hasan Tarek, Mohammad Rakibul Islam Bhuiyan, Saiful Islam

Published 2026-07-03

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Md. Hasan Tarek, Mohammad Rakibul Islam Bhuiyan, Saiful Islam

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine Bangladesh's healthcare system as a massive, bustling train station. It's a place where millions of people (patients) try to get medical help, and a dedicated team of workers (doctors, nurses, and managers) tries to keep the trains running on time. Right now, this station is facing some growing pains: it's crowded, resources are tight, and the old ways of doing things are getting a bit slow.

Enter Artificial Intelligence (AI). Think of AI not as a robot taking over the station, but as a super-smart, high-tech control panel that could help run the trains more efficiently, predict delays before they happen, and guide passengers to the right platform faster.

This paper is essentially a survey report asking everyone at the station: "How do you feel about installing this new control panel?"

Here is the breakdown of what the researchers found, using simple analogies:

1. The Survey Crew and the Crowd

The researchers (from Begum Rokeya University) didn't just ask the station managers; they asked everyone. They gathered 515 people from different roles:

The Passengers (Patients): About 39% of the group.
The Station Managers (Administrators): About 27%.
The Tech Crew (Technical Assistants): About 21%.
The Train Conductors (Doctors and Nurses): A smaller group (about 13%).

They asked these people to fill out a checklist (a survey) where they could rate their feelings from "Strongly Disagree" to "Strongly Agree."

2. The General Vibe: "Cautiously Optimistic"

If the station were a room full of people, the mood isn't "We hate this idea!" and it's not "We need this yesterday!" It's more like a moderate, positive hum.

Most people think AI is a good idea.
They are generally ready to try it out.
The data shows that people aren't terrified; they are curious and willing to learn.

3. What Makes People Say "Yes"?

The researchers found that three specific things act like fuel for people's excitement about AI:

Knowing How It Works (Technological Awareness): The more people understand what AI actually is (not just magic, but math and data), the more they like it. It's like understanding how a GPS works makes you trust it more than just blindly following a stranger's directions.
Being a "Early Adopter" (Personal Innovativeness): People who naturally like trying new gadgets or apps are the first to say, "Let's do this!" They are the ones who buy the latest phone before anyone else.
Hearing About It Online (Social Media Influence): If people see positive stories about AI on social media, they feel more ready to accept it. It's like hearing a friend say, "This new coffee shop is amazing," makes you want to try it.

4. The "But..." (Concerns)

Even though the vibe is positive, there are some speed bumps on the road:

Job Worries: About half of the healthcare workers are worried that AI might take their jobs, like a robot replacing a ticket seller.
Privacy Fears: People are worried about their medical secrets being leaked, similar to worrying about someone hacking your personal diary.
The Infrastructure Gap: In some parts of the country (especially rural areas), the "power lines" (internet) and "roads" (computers) aren't strong enough yet to support this high-tech control panel.

5. Did the Survey Make Sense? (The Quality Check)

The researchers didn't just collect random answers; they checked their work to make sure the survey was solid.

Reliability: They checked if the questions were consistent. Imagine asking someone the same question in three different ways; if they give the same answer every time, the survey is reliable. The data passed this test.
Validity: They made sure they were actually measuring what they said they were measuring. It's like making sure a thermometer is actually measuring temperature and not humidity. The data passed this test too.
No Single Bias: They checked to make sure one single answer wasn't dominating all the others (like if everyone just said "Yes" to everything). The answers were varied and honest.

6. What the Paper Doesn't Say

It is important to stick to what this specific report claims:

This paper does not say that AI is currently being used in every hospital in Bangladesh.
It does not claim that AI has already cured diseases or saved lives in this specific study.
It does not provide a step-by-step guide on how to build the AI systems.

The Bottom Line:
This paper is a snapshot of the mood. It tells us that the people in Bangladesh's healthcare system are ready and willing to embrace AI, provided they are educated about it, their job security is respected, and the internet infrastructure is strong enough to support it. The "train station" is ready for the upgrade, but the actual installation of the control panel needs to happen carefully and with the right support.

Technical Summary: Dataset of Survey Responses on Artificial Intelligence Adoption in the Healthcare Sector of Bangladesh

Problem Statement
Bangladesh's healthcare sector faces persistent challenges, including bureaucratic inefficiencies, resource limitations, and unequal access to services. While the national "Digital Bangladesh" initiative has established a technological foundation, the integration of Artificial Intelligence (AI) into healthcare remains hindered by significant gaps in awareness and understanding among stakeholders. Existing literature indicates that while general attitudes toward AI are positive, a substantial knowledge gap exists, particularly among older professionals and those with limited digital exposure. Furthermore, concerns regarding job displacement, data privacy, and algorithmic bias persist. There is a critical need for empirical data to understand the specific factors influencing AI readiness and adoption among diverse stakeholders (patients, providers, and administrators) to inform policy and implementation strategies.

Methodology
This study employs a quantitative research approach utilizing a structured survey to gather data from 515 valid respondents across Bangladesh between January 2024 and April 2026. The participants represent a diverse cross-section of the healthcare ecosystem, including patients (38.64%), management bodies (27.38%), technical assistants (20.97%), doctors (7.77%), and nurses (5.24%). The sample is predominantly male (63.30%) and highly educated, with the majority holding Honor's/BBA/BSC (46.80%) or Master's degrees (33.40%).

Data collection was conducted via an online questionnaire distributed through Google Forms and social media platforms. The instrument utilized a five-point Likert scale (1–5) to measure seven key latent constructs:

Perceived Susceptibility (PS)
Technological Awareness (TA)
Perceived Risk (PR)
Personal Innovativeness (PI)
Social Media Influence (SMI)
Readiness of AI (RAI)
Adaptation of AI (AAI)

Each construct was measured by three specific items. The dataset underwent rigorous preprocessing, including cleaning to remove inconsistencies. Statistical analysis was performed using Python (3.8.15) with libraries including NumPy, SciPy, and Pandas. The analytical framework included:

Descriptive Statistics: To assess central tendency and distribution.
Reliability and Validity Testing: Cronbach's Alpha, Composite Reliability (CR), and Average Variance Extracted (AVE) were calculated to ensure internal consistency and convergent validity.
Factor Analysis: Kaiser-Meyer-Olkin (KMO) and Bartlett's Test of Sphericity were used to confirm data suitability.
Bias Assessment: Harman's one-factor test was conducted to check for common method bias.
Discriminant Validity: Assessed using the Fornell-Larcker criterion and the Heterotrait-Monotrait (HTMT) ratio.

Key Contributions
The primary contribution of this work is the publication of a comprehensive, publicly available dataset (hosted on Mendeley Data) containing survey responses from 515 stakeholders in the Bangladeshi healthcare sector. The study provides:

Empirical Evidence: A detailed profile of the perceptions, awareness levels, and readiness of patients, providers, and administrators regarding AI adoption.
Validated Measurement Model: The study demonstrates that the survey instrument possesses satisfactory reliability and validity, with all constructs meeting standard thresholds for Cronbach's Alpha (>0.70), CR, and AVE (>0.50).
Correlation Insights: The analysis reveals significant positive correlations between constructs, notably between Personal Innovativeness and Adaptation of AI ( $r=0.715$ ), and between Readiness and Adaptation ( $r=0.709$ ).
Demographic Context: The dataset offers a granular view of how education levels and professional roles correlate with AI perceptions, highlighting that the sample is skewed toward educated demographics.

Results
The analysis indicates that stakeholders in Bangladesh generally exhibit moderate to positive perceptions and readiness toward AI adoption.

Descriptive Findings: Mean values for most constructs ranged between 2.90 and 3.34, suggesting opinions are moderately positive rather than polarized. Skewness values were slightly negative, indicating a trend toward higher scale responses.
Reliability: All constructs demonstrated strong internal consistency, with Cronbach's Alpha scores exceeding 0.70. Composite Reliability (CR) scores were high, and AVE scores exceeded 0.50, confirming convergent validity.
Factor Analysis: The KMO value was 0.952 (excellent), and Bartlett's Test was significant ( $p < 0.001$ ), confirming the dataset's suitability for factor analysis.
Discriminant Validity: Both the Fornell-Larcker criterion and HTMT ratios (mostly below 0.85) confirmed that the constructs are distinct from one another.
Key Drivers: Factors such as technological awareness, personal innovativeness, and social media influence were found to be positively associated with AI readiness and adoption. Specifically, individuals with higher personal innovativeness showed a stronger likelihood of adopting AI-based healthcare services.

Significance and Claims
The authors claim that this dataset provides a foundational evidence base for policymakers and healthcare administrators in Bangladesh. The study suggests that respondents are receptive to AI-driven healthcare services, highlighting the potential for expanding AI applications within the national system. The findings support the notion that targeted educational interventions to enhance AI literacy, particularly among older professionals and those with limited digital exposure, are necessary. Furthermore, the data underscores the importance of addressing infrastructure limitations and developing robust regulatory frameworks for data governance to facilitate effective implementation.

The paper modestly acknowledges its limitations, noting that the sample size (515) may not fully represent the entire Bangladeshi healthcare sector, and that the study focuses on perceptions rather than actual implementation outcomes. Consequently, the authors position the work as a preliminary step toward understanding AI adoption dynamics, offering a resource for further research into the real-world effectiveness and long-term trends of AI in the region's healthcare system.