Health Policy and Technology xxx (xxxx) xxx

Contents lists available at ScienceDirect

Health Policy and Technology

journal homepage: www.elsevier.com/locate/hlpt

Digital health inclusion towards achieving universal health coverage for

Bangladesh utilizing general practitioner model
Moinul H. Chowdhury a, Rony Chowdhury Ripan a, A.K.M. Nazmul Islam a, Rubaiyat
Alim Hridhee a, Farhana Sarker a, b, Sheikh Mohammed Shariful Islam c, Khondaker
A. Mamun a, d, *
a
CMED Health (A Digital Health Inclusion Initiatives for UHC), Dhaka, Bangladesh
b
Department of CSE, University of Liberal Arts Bangladesh, Dhaka, Bangladesh
c
Institute for Physical Activity and Nutrition, Faculty of Health, Deakin University, Melbourne, Australia
d
AIMS Lab, Department of CSE, United International University, Dhaka, Bangladesh

A R T I C L E I N F O A B S T R A C T

Keywords: Objective: Bangladesh’s health care system, particularly in rural areas, experiences enormous obstacles in
Digital GP model providing complete preventive and primary healthcare services due to the lack of adequate healthcare facilities,
Digital health system resource constraints, and a non-functional referral system. To alleviate these problems, in this study, we intro­
Digital health referral system
duce the digital general practitioner (GP) model for rural Bangladesh, digital platforms and present a statistical
Rural health care system
Universal health coverage
analysis of the data that was gathered from the pilot project.
Digital health inclusion Methods: A total of 12,746 people were provided regular health services during the pilot project, from all genders
and age groups, and provided their socio-demographic and healthcare-related data. We analyzed healthcare-
related data by carrying out both descriptive and inferential statistics.
Results: By utilizing this digital GP model, rural residents can receive routine health screenings at their homes,
identify health risks early, receive consultation and health education, and be referred to GP and upper-level
health facilities as needed. We found that hypertension was more prevalent (4.84% of the served population),
and cancer was the least prevalent of all the NCDs in the studied population (0.05% of the served population).
The population for stroke, hypertension, diabetes increased until the 50–59 age range as age increased, following
which the population proportion declined as age increased. Additionally, 3.96% of young females were severely
malnourished, comparably higher proportion than young males (2.34%).
Conclusion: NCDs such as hypertension, diabetes was prevalent among rural people. Necessary steps should be
taken to raise preventive and primary healthcare awareness among rural people.
Public interest summary: The absence of proper healthcare facilities, resource constraints, and a non-functional
referral system hamper Bangladesh’s health care system’s ability to provide comprehensive preventive and
primary healthcare services in rural area. As a result, patients develop advanced ailments, including non-
communicable diseases (NCDs), and must seek treatment at an expensive specialty hospital. To resolve this
issue, we introduce a digital GP model for rural Bangladesh, then show digital platforms that use the concept, and
lastly summarize significant findings from the piloted digital GP model. By utilizing this digital GP model, rural
residents can receive routine health screenings at their homes, identify health risks early, receive consultation
and health education, and be referred to GP and upper-level health facilities as need. From our data analysis, we
discovered high burden of NCDs such as hypertension and diabetes in the piloted area. Necessary steps should be
taken to raise preventive and primary healthcare awareness among rural people.

* Corresponding author at: Department of Computer Science & Engineering, Advanced Intelligent Multidisciplinary Systems Lab, Institute of Advanced Research,
United International University, United City, Madani Ave, Dhaka 1212, Bangladesh.
E-mail address: mamun@cse.uiu.ac.bd (K.A. Mamun).

https://doi.org/10.1016/j.hlpt.2023.100731

2211-8837/© 2023 Fellowship of Postgraduate Medicine. Published by Elsevier Ltd. All rights reserved.

Please cite this article as: Moinul H. Chowdhury et al., Health Policy and Technology, https://doi.org/10.1016/j.hlpt.2023.100731
M.H. Chowdhury et al. Health Policy and Technology xxx (xxxx) xxx

Introduction and out-of-pocket expenses account for 72 percent [12]. Whereas Global
Health Spending Per Capita (HSPC) is 1467$, Government Health
A general practitioner (GP) is a doctor who provides primary health Spending Per Capita (GHSP) is 865$, and out-of-pocket costs are 18%
care (PHC) services for several chronic illnesses, provides preventive [12]. Due to this suffering, 16.4% people avoid treatment and 8.6
and primary treatments within a catchment area, and refers patients to a million people pushed into poverty due to out-of-pocket expenses [12].
hospital or specialist after risk assessment [1]. Many developed coun­ Even though these lacking’s, numerous telemedicine services have
tries have implemented GP models that make their health care systems recently appeared in Bangladesh. In Bangladesh, for example, "Doc­
more effective and reduce the burden on hospitals for primary level Time" provides 24 h telemedicine services [13]. It is, nevertheless, un­
health issues. In the early twentieth century, primary care was intro­ popular in rural areas due to a lack of understanding and access to the
duced in the UK, emphasizing the concept of referral [2]. The USA first internet. Similarly, despite the fact that “Digital Healthcare Solutions”
introduced the assistant physician role to serve primary care in 1960 due provides medical consultation, micro health insurance, and health pro­
to a shortage of doctors [3]. Norway has a strong primary care service grams for diabetes, communicable diseases, and maternal and child care
and its people have been relying on general practitioners since 2001 as [14], its reach to rural populations is not thoroughly measured. Addi­
their regular doctors [3]. In Australia, general practitioners provide tionally, these organizations lack IoT devices that allow patients to
treatment for common illnesses, chronic diseases, and diabetes, and also monitor their vital signs [15].
provide vaccinations. To address all these issues, after the exploration of the health care
Additionally, developed countries began digitizing their health care system, digital health system, and the rural health service situation in
systems in the 1990s and developed the Health Level Seven (HL7) Bangladesh, CMED Health [16] designed and implemented an inte­
standard, which ISO adopted as a reference for international standard­ grated digital GP platform for rural areas to provide comprehensive
ization, by compiling several frameworks and related standards for the preventive and primary healthcare service, which is named the "Rural
exchange, integration, sharing, and retrieval of electronic health records General Practitioner" (RGP) model. People can get primary care on their
(EHRs) [4]. Previously, all documents were written by hand, and pa­ doorstep through CMED’s digital health kits and mobile applications,
tients did not have access to them. However, with the advent of public which are maintained by trained health workers. After proper risk
EHR systems, various benefits for a public healthcare system have been assessment through the clinical decision support system (CDSS), health
recognized, including reduced and more efficient management of ex­ workers refer the patients to an integrated GP center or facilitate tele­
penses, more effective management of vast volumes of patient data, and medicine services on their phone, where doctors provide further inter­
centralized medical patient records [5]. The digital healthcare system vention and prescribe for the patient digitally. All the members of a
has been enhanced a lot after this revolutionary inclusion. For example, household can have all the benefits of the digital GP model by spending
"Babylon Health" launched a "GP at hand" service that includes an 100 Bangladeshi Taka (US $1.20) monthly. The primary objectives of
instant symptom checker, face-to-face appointments, telephone ap­ the study were the following: a) To introduce a digital GP model for
pointments, video call appointments with a GP, and so on [6]. In 2017, Rural Bangladesh. b) To demonstrate digital platforms that incorporate
England’s National Health Service (NHS) tested a system in which the digital GP model. c) To outline key findings based on data collected
smartphone-based applications were utilized to monitor chronic dis­ from the piloted digital GP model.
eases such as chronic obstructive pulmonary disease (COPD) and The remainder of the paper is structured according to the following:
gestational diabetes. These apps enabled clinicians to remotely access Section 2 provides a comprehensive overview of the digital GP model for
patient data via a smart device and prescribe them [7]. Gelogo et al. [4] rural Bangladesh. In Section 3, we present our methodology for
proposed a system for ubiquitous health monitoring that consists of analyzing data that was gathered from the piloted digital GP model. We
numerous sensors (embedded in a wearable belt) and android mobile analyze all the data and outline the findings in Section 4. In Section 5, we
application. When users wear the belt, it transmits vital physiological discuss our findings, compare them with existing literature, outline our
data to their phone, which the users are able to view. Additionally, the strengths and limitations, and finally conclude this paper with future
app includes an alert system in case of an emergency. work in Section 6.
However, low- and middle-income countries (LMICs) like
Bangladesh, where rate of population growth is high, are still facing Design and development of digital GP model for rural areas
several issues in digitizing and structuring their health care systems.
Bangladesh has a population of 167,885,689 and is ranked number 8 System overview
among the most populated countries in the world [8]. Besides, areas
comprising around 61.82% of Bangladesh are rural [9]. Some of the The Digital GP Model is implemented by CMED Health and United
main challenges of Bangladesh’s health sector are due to a lack of Trust in Nayangar union, a small rural area in Jamalpur district,
healthcare infrastructure at the rural level, as well as a scarcity of skilled Bangladesh. The model was developed to provide comprehensive pre­
general practitioners and health workers. According to a study from ventive and primary health care services to rural populations (with a
2014, there are 18.2 physicians, 5.8 nurses, and 0.8 dentists per 10,000 focus on NCDs and maternal care) via doorstep service delivery, a
people in urban areas, while the corresponding figures are 1.1, 0.8, and structured and functioning referral system that adheres to WHO guide­
0.08 respectively in rural areas [10]. In addition, due to a lack of edu­ lines, and telemedicine—all facilitated by integrated digital solutions.
cation, rural people are not aware of their basic rights and do not address The digital GP model works as follows: First, trained health visitors visit
their health-related issues. Unfortunately, every day, rural people face households and during this visit they provide basic primary health care
challenges accessing health services, which can lead to avoidable health services, including symptomatic health checks, screening, and coun­
complications, including NCDs. Nujhat et al. reported that in 2018 seling, as well as antenatal and postnatal care. Each served individual
prevalence of hypertension was 41.6% and diabetes was 4.3% among has their unique health account where their health data is stored. In
rural people [11]. Even if they face problems with these NCDs, due to addition, the health workers use mobile app to collect socio-
lack of money, they often visit the local pharmacy and get medication demographic data and health-related data. The health workers app is
from quack doctors, which is even more alarming for their health. integrated with an A.I.-driven clinical decision support system that no­
However, sometimes they go to secondary and tertiary health clinics for tifies health workers, based on physical and biochemical measurements,
some primary and preventive health care, costing them a fortune, which whether or not to refer the patient to a doctor. If the patient consents to
can be treated by merely seeing a GP doctor. Additionally, the World see a doctor based on the outcome of the primary health evaluation, they
Bank reports that Bangladesh’s Health Spending Per Capita (HSPC) is are referred to the Sushatho Digital Healthcare Platform (SDHP) or GP
123 dollars, the Government’s Health Expense Per Capita is 22.9 dollars, center. The comparison between SDHP and GP is that SDHP is a virtual

2
M.H. Chowdhury et al. Health Policy and Technology xxx (xxxx) xxx

Fig. 1. A comprehensive graphical overview of the digital GP model for rural Bangladesh.

platform integrated with telemedicine services. On the other hand, GP is features such as: 1) doctors can see the total served patient number till
a physical center integrated with comprehensive primary healthcare date, the total served patient number on a particular day, how many
services and diagnostic tests. After inspecting the patient, if necessary, patients are waiting, and so on. 2) they can store patients’ complaints,
doctors may choose to refer the patient to a secondary or tertiary comorbidity, etc. and also see previous drug history, gynecological
healthcare facility where they can get treatment from a specialist. A history, and can give diagnosis lab tests. 3) Besides, they can prescribe
comprehensive graphical overview of the digital GP model is presented all the medicine, medicine dosage, and instructions for each drug, and
in Fig. 1. also provide advice and suggestions. A comprehensive overview of these
features is depicted in Fig. 4.
Admin’s Dashboard: The admin’s dashboard is created to control,
Digital platforms
monitor, and provide quality assurance of the digital GP model. When
the admin logs in to the dashboard, at first glance, he can see all the
This digital GP model is entirely integrated by 4 applications: 2
organizations they are partnered with, how many unions (Union is
android applications and 2 web applications. All these applications were
Bangladesh’s smallest rural administrative and local government unit
designed based on a Cloud based medical system framework shortly
[18].) they are serving, how many doctors they have, how many health
called CMED [17].
officers they have, total number and percentage of households surveyed,
Health Worker Application: The health worker application is one of
total number and percentage of members served, total number of health
the primary applications of this GP model, as it collects the majority of
cards the members bought, etc. Also, by scrolling down, they can see
sociodemographic and health data. When a health worker uses CMED
several visualization charts of socio-demographic data and
smart health kits to collect health-related data such as SpO2, Blood
health-related data. A simple overview of the admin dashboard can be
Pressure, and so on, this application allows the health worker to view all
observed in Fig. 5.
of the data and measurement results. Additionally, this application in­
corporates a system for offline synchronization. Due to the lack of a
Methodology
reliable internet connection in rural areas, all data collected by health
workers is initially stored on the device’s local system. After that, all the
Study design, setting, and population
data are automatically uploaded to the cloud once the device establishes
a reliable internet connection. Apart from that, another primary feature
CMED health piloted the digital GP model in the Nayanagar Union of
of this application is its data driven clinical decision support system
Melandaha Upazila in the Jamalpur district, a rural area of Mymensingh
(CDSS) that makes complex referral decisions based on a patient’s health
division. This study was done with data that was collected from March
vital measurements and medical history. This can be used by health
26th, 2021 to March 31st, 2022. During this one-year period, the digital
visitors and health officers to refer patients to doctors. Fig. 2 depicts
GP model served 12,746 rural people from 5643 households with the
several critical aspects of a health worker’s application.
help of 4 GP doctors, 3 registered health officers. Total targeted mem­
User Application: This user application includes several critical fea­
bers for this study were 21049 people. Among them, 12746 people
tures, including the following: 1) Patients can constantly monitor and
(response rate 60.55%) paid 100 Bangladesh Taka (US $1.20) monthly
track their health records. 2) Additionally, they can educate themselves
to receive the digital GP model services.
on primary and preventive healthcare by reading several articles. 3)
They can use "Search" to locate hospitals, ATMs, blood banks, and
pharmacies in their immediate vicinity. and so forth. A comprehensive Data collection procedure
overview of these features is depicted in Fig. 3.
Doctor’s Dashboard: The GP doctor dashboard has several important A group of well-trained health workers (both male and female)

3
M.H. Chowdhury et al. Health Policy and Technology xxx (xxxx) xxx

Fig. 2. Some of the main features from “Health Workers” application.

usually went to all the households and collected socio-demographic and for 30 s. Before taking oxygen saturation, and pulse rate, participants’
health-related data through a mobile application called the "Health fingers were warmed by some hand exercises, and while measuring, all
Workers App," as stated before. At this stage, health workers used IoT- participants were instructed to sit still on a chair and remove any nail
enabled smart health devices to measure diabetes, blood pressure, polish. In order to assess blood pressure, health workers gave each
pulse rate, weight, and SpO2 (blood oxygen saturation) as shown in participant five minutes to relax in a comfortable position with a straight
Fig. 6. Health workers also document height, temperature, and MUAC back and uncrossed legs. Then, a health worker placed a smart blood
(Mid-Upper Arm Circumference) measurements (only for children), and pressure monitor on the participant’s left arm and tightened the arm
input them into the mobile application manually. Participants were cuff. Concurrently, the health worker ensured that the participant’s arm
asked to take off their shoes and any headgear they were wearing before was put on a table such that it was parallel to their heart. Using a smart
having their height measured. Afterwards, they were instructed to stand glucometer, blood sugar is measured. Initially, health professionals
with their feet together, their heels on the floor, their knees straight, and ensured the cleanliness of the smart glucometer. The participants were
their eyes level with their ears. A measuring tape was then used to then asked to wash their hands with warm water and hand soap. After
measure height (cm). Following the removal of their shoes and light thoroughly drying the participant’s hand, health care professionals
casual attire, participants were to stand still on a weighing scale on a punctured the participant’s finger to get a little drop of blood for the
firm and flat platform to measure their weight. In kilograms (kg), the smart glucometer. Ultimately, the smart glucometer examined the blood
portable weighing scale was used to measure the weight of the subject. sample. All of the above-mentioned data were gathered with the consent
After manually entering height and weight, BMI was automatically of participants. Besides, all of the above-mentioned data were gathered
calculated by "Health Workers APP." Arm circumference was measured maintaining adequate privacy. The ethical review board of United In­
by a plastic tape placed horizontally around the arm. SpO2 and pulse rate ternational University approved to conduct the study.
were measured by placing the index finger into the smart pulse oximeter

4
M.H. Chowdhury et al. Health Policy and Technology xxx (xxxx) xxx

Fig. 3. Some of the main features from “User” application.

Fig. 4. Some of the main features from “Doctor” dashboard.

Quality assurance the data collection through health officials; (3) comprehensive system
monitoring via a separate admin dashboard; (4) to ensure safety, all four
To maintain the quality control of the study, CMED took several applications were tested by experts in software quality assurance before
measures: (1) pre-piloted training of the team members, including deploying in the field for data synchronization and integration; (5) Using
admins, health officials, and health workers, to outline the procedures long-lasting and intelligent measurement equipment for physical and
and potential difficulties associated with data collection and taking biochemical measurements.
measurements; (2) strict monitoring at the field level to closely monitor

5
M.H. Chowdhury et al. Health Policy and Technology xxx (xxxx) xxx

Fig. 5. A overview of “Admin” dashboard.

Fig 6. The digital GP model services and smart health devices.

Statistical analysis Python libraries like pandas, matplotlib, NumPy, SciPy, stats model and
seaborn.
The data that were collected through health workers app were kept
in Amazon Web Service (AWS). Then we collected this data from AWS as Results
a dump file and stored them in MySQL server. After that we applied
MySQL (a structured query language) to extract necessary data. Sociodemographic information of the study population
Following that, we examined the data for inconsistency, missing data,
coding errors, and outliers [19]. Descriptive statistics were used to As stated before, there were 12,746 people in total, with 11,491
evaluate the distribution of the studied population. To evaluate the adults, and 1255 young people. From Table 1, it can be observed that
distributions of all measurements, all categorical variables were pre­ overall, the female population was higher (total: 7651, in%: 60.03) than
sented using frequency, percentages, and 95% CI (confidence interval), the male population (total: 5095, in%: 39.97). It is because health
while continuous variables were presented using mean and standard workers usually visited during the day when most of the male members
deviation. In addition, the association of NCDs and its risk predictors usually go to work. However, the mean age of the males (41.12 ± 18.95
were calculated using binomial logistic regression and presented in years) was higher than the females’ mean age (37.95 ± 17.01 years).
crude odds ratio along with 95% CI. Before using binary logistic The average age of the study population was 39.22 ± 17.88 (Standard
regression, assumption check was done for normality, multicollinearity, Deviation) years. Then we analyzed the comparison between male and
and outliers. The bar graph and line graph were used to illustrate the female proportions among different age groups. The female population
major findings. All these analyses were done with the help of powerful was higher in all age groups except the ">=70′′ age groups, as shown in

6
M.H. Chowdhury et al. Health Policy and Technology xxx (xxxx) xxx

Table 1
Sociodemographic status of the study population.
Characteristics Overall N = 12,746 95% CI MALE N = 5095 (39.97%) 95% CI FEMALE N = 7651 (60.03%) 95% CI
Age (Mean ± S.D) 39.22 ± 17.88 41.12 ± 18.95 37.95 ± 17.01

Age Groups
<18 1255 (9.85%) 9.33–10.36 599 (11.76%) 11.2–12.32 656 (8.57%) 8.09–9.06
18–29 2611 (20.48%) 19.78–21.19 724 (14.21%) 13.6–14.82 1887 (24.66%) 23.92–25.41
30–39 2783 (21.83%) 21.12–22.55 1029 (20.2%) 19.5–20.89 1754 (22.93%) 22.2–23.65
40–49 2081 (16.33%) 15.69–16.97 894 (17.55%) 16.89–18.21 1187 (15.51%) 14.89–16.14
50–59 1917 (15.04%) 14.42–15.66 789 (15.49%) 14.86–16.11 1128 (14.74%) 14.13–15.36
60–69 1349 (10.58%) 10.05–11.12 653 (12.82%) 12.24–13.4 696 (9.1%) 8.6–9.6
>=70 750 (5.88%) 5.48–6.29 407 (7.99%) 7.52–8.46 343 (4.48%) 4.12–4.84
Marital Status
Married 10,791 (84.66%) 84.04–85.29 4327 (84.93%) 84.31–85.55 6464 (84.49%) 83.86–85.11
Unmarried 1328 (10.42%) 9.89–10.95 702 (13.78%) 13.18–14.38 626 (8.18%) 7.71–8.66
Divorced 23 (0.18%) 0.11–0.25 4 (0.08%) – 19 (0.25%) 0.16–0.33
Widower 35 (0.27%) 0.18–0.37 35 (0.69%) 0.54–0.83 – –
Others 59 (0.46%) 0.35–0.58 27 (0.53%) 0.4–0.66 32 (0.42%) 0.31–0.53
Widow 510 (4.0%) 3.66–4.34 – – 510 (6.67%) 6.23–7.1
Education
Illiterate 5488 (43.06%) 42.2 – 43.92 2026 (39.76%) 38.42 – 41.1 3462 (45.25%) 44.13 – 46.37
Primary 2707 (21.24%) 20.53 – 21.95 1183 (23.22%) 22.06 – 24.38 1524 (19.92%) 19.03 – 20.81
Secondary 1116 (8.76%) 8.27 – 9.25 428 (8.4%) 7.64 – 9.16 688 (8.99%) 8.35 – 9.63
College or higher 786 (6.17%) 5.75 – 6.59 408 (8.01%) 7.26 – 8.76 378 (4.94%) 4.45 – 5.43
Literacy 920 (7.22%) 6.77 – 7.67 258 (5.06%) 4.46 – 5.66 662 (8.65%) 8.02 – 9.28
Others 1729 (13.57%) 12.98 – 14.16 792 (15.54%) 14.55 − 16.53 937 (12.25%) 11.52 – 12.98

Fig. 7. In “>=70′′ age groups male proportion was higher. Most of the
studied population was married (total: 10,791, in%: 84.66) and illiterate
(total: 5488, in%: 43.06). In addition, among 5643 households, 5374
households had safe water, 4473 households had sanitary latrine.

Analysis of NCD risk factors

According to Fig. 8(a), hypertension was more prevalent than all

other NCDs in the studied population. There were 617 people in the
Nayanagar union who had hypertension, which was 4.84% of the overall
population. Following that, diabetes was more prevalent. A total of 283
diabetes cases were observed, representing 2.22% of the total popula­
tion. Cancer was the least prevalent among all NCDs. There were just 6
people (0.05%) with cancer and 11 people (0.09 percent) with kidney
illness. The Fig. 8(b) depicts all of the NCDs classified by gender. It is
seen from Fig. 8(b) that the male population numbers with strokes
(male = 21, female = 14), COPD (male = = 41, female = 30) and CVD
(male = 41, female = 32) were higher than the female population. On
the other hand, kidney disease (male = 4, female = 7), diabetes (male =
Fig. 7. A bar chart representation of gender proportions according to
age groups.
126, female = 157), and hypertension (male = = 235, female = 382),
were more dominant in the female population.
Additionally, Table 2 categorizes all NCDs by gender and age.
Additionally, the most vulnerable age group was “>=40′′ , as all the
NCDs were found to be more prevalent in this age group. For stroke,

Fig. 8. (a) A bar chart representation of population number according to NCD’s. (b) A bar chart representation all of the NCDs segregated into gender.

7
M.H. Chowdhury et al. Health Policy and Technology xxx (xxxx) xxx

Table 2 In addition, Table 3 represents crude odds ratio association of

NCD’s segregated as gender and age. different risk factors with different NCDs. Hypertension was associated
Age Groups with higher odds of stroke (OR: 59.24, P < 0.01, 95% C.I.: 27.64
NCDs <40 >=40 − 126.9). Also, diabetes had strong association with hypertension (OR:
Stroke Overall 1 34 21.24, P<0.01, 95% C.I.: 16.53 – 27.29). Moreover, CVD was associated
0.01% 0.27% with higher odds of hypertension (OR: 20.25, P < 0.01, 95% C.I.: 12.70 –
Male 1 20 32.28).
0.02% 0.39%
Female 0 14
0.0% 0.18%
Measurement analysis
Cancer Overall 1 5
0.01% 0.04% There were seven distinct units of measurement, as listed in Table 4.
Male 1 2 Among them, blood pressure was taken from the highest number of
0.02% 0.04%
people (11,328). The GP model’s measurement severity status is
Female 0 3
0.0% 0.04% designed for adult (age >= 18) and young (age < 18) people following
CVD Overall 13 60 the WHO PEN Protocol [20]. So, all the measurements were categorized
0.1% 0.47% for adult people (Table 5) and young people (Table 6) since most of the
Male 5 36 severity statuses were different among these two groups [21].
0.1% 0.71%
Female 8 24
Adult People: The Table 5 depicts the distribution of the overall
0.1% 0.31% measured adult population (over the age of 17) by BP, BMI, Pulse Rate,
COPD Overall 17 54 TEMP, SpO2, and Blood Sugar. As can be seen from this table, the ma­
0.13% 0.42% jority of adult people in Nayanagar union were healthy, as the "normal"
Male 7 34
status number for all measurements is greater than the rest of the
0.14% 0.67%
Female 10 20 severity status numbers. However, in terms of Blood Sugar, among the
0.13% 0.26% adult population, 466 (4.06%) were suspected for diabetes, 29 (0.25%)
Kidney Disease Overall 4 7 as High (Borderline), 40 (0.35%) as Low (Hypoglycemia), and 62
0.03% 0.05% (0.54%) as pre-diabetic. From “BMI” measurement, it can be observed
Male 1 3
0.02% 0.06%
that 164 (1.22%) adult people were obese and 1094 (9.52%) adult
Female 3 4 people were underweight. In addition, female adults (744, in%: 10.65)
0.04% 0.05% were underweight at a greater number than male adults (350, in%:
Hypertension Overall 78 539 7.78). In terms of blood pressure, 3390 (29.5%) adult people were in
0.61% 4.23%
prehypertension stage and 2147 (18.7%) adult people were suspected
Male 24 211
0.47% 4.14% for hypertension.
Female 54 328 Young People: The following Table 6 summarizes the distribution of
0.71% 4.29% the whole measured population (under the age of 18) by BP, BMI, TEMP,
Diabetes Overall 51 232 MUAC, Pulse Rate, SpO2, and Blood Sugar. In terms of "Blood Sugar,"
0.4% 1.82%
Male 13 113
two individuals had "High" blood sugar levels and one individual had
0.26% 2.22% "Low" blood sugar levels. Even though the female population was larger,
Female 38 119 44 (7.35%) young males were obese, significantly more than their fe­
0.5% 1.56% male counterparts (23, or 3.51%). However, females were underweight
CVD = Cardiovascular disease; COPD = Chronic Obstructive Pulmonary Disease. at a higher rate (75, or 11.43%) than males (53, or 8.85%). According to
MUAC measurements, 40 young people (3.19%) were suffering from
cancer, CVD, COPD, diabetes, hypertension, and kidney disease, the severe malnutrition. Females were more prevalent than males among
“>=40′′ age group had a significantly higher prevalence of 34 (0.27%), them. 26 (3.96%) young females were severely malnourished, whereas
5 (0.04%), 60 (0.47%), 54 (0.42%), 232 (1.82%), 539 (4.23%), and 7 14 (2.34%) young males were severely malnourished. Having said that,
(0.05%), respectively. For stroke, hypertension, diabetes, overall pop­ the majority of the young population in Nayanagar union was in good
ulation number increased with age until the 50–59 age range, after health.
which overall population number decreased as age increased. However,
for CVD, overall population number increased as age increased. Fig. 9 Discussion
provides a more detailed look of these patterns.
The digital GP model for rural Bangladesh as a whole is comprised of

Fig. 9. A population (in%) line graph according to their age groups.

8
M.H. Chowdhury et al. Health Policy and Technology xxx (xxxx) xxx

Table 3 (NDHM) in 2020 with the goal of providing a digital health id, a digital
NCD’s associated with risk factors. patient health record system, an electronic medical record web appli­
Disease Total Factor Crude P- 95% CI (for cation, a digi doctor platform, and a health facility registry [22]. They
Cases Odds value Odds Ratio) will accomplish this by providing all citizens with a single, secure health
(N) Ratio ID, electronic prescriptions, digital referrals and consultations, clinical
(OR)
decision support, and the interchange of health information between
Stroke 35 Age public and private health care facilities, among other things. While
>=40 37.28 <0.01 5.10–272.42 CMED’s digital primary care model incorporates the majority of these
services, it is still unable to transmit health information between public
<40
Hypertension
Present 59.24 <0.01 27.64–126.9 and private health care facilities. However, we believe that the digital
Absent GP model for rural areas will have this capability if scaled up with
Diabetes government assistance. In addition, we were unable to locate any
Present 13.39 6.03 – 29.74
<0.01
research that discusses the digital GP model in Pakistan. However, we
Absent
CVD discovered a study [23] that highlighted some telemedicine programs
Present 16.93 <0.01 5.07 – 56.58 such as "eDoctor" and "Sehat Kahini," as well as electronic health systems
Absent such as "Dengue activity tracking system" and mobile health immuni­
COPD zation record systems such as "Teeku."
Present 5.31 0.10 0.72 – 39.34
Absent
From our result analysis we found that, hypertension was more
Hypertension 617 Age prevalent, and cancer was the least prevalent of all other NCDs in the
>=40 10.45 <0.01 8.03 – 13.59 studied population. In this study, hypertension was prevalent among
<40 4.84% of the studied population, which is very smaller than 25.9%, that
Diabetes
was found by this study [24]. We argue that, it is because we only took
Present 21.24 <0.01 16.53 – 27.29
Absent data from one union and their life style could be healthier than the study
Diabetes 283 Age mentioned above. However, according to this study [25], the prevalence
>=40 5.96 <0.01 4.33 – 8.21 of hypertension ranged from 1.10% to 75.0% in the investigated pop­
< 40 ulation, which supports the findings of our study. In this study, the male
Hypertension
population numbers with stroke (male = 21, female = 14), COPD (male
Present 21.24 <0.01 16.53 – 27.29
Absent = = 41, female = 30) and CVD (male = 41, female = 32) were higher
CVD 73 Age than the female population, which is supported by these studies [26,27],
>=40 6.20 <0.01 3.26 – 11.78 and [28]. Besides, people with hypertension were more likely to have
<40
stroke which is also shown in this study [28]. On the other hand, for
Diabetes
Present 16.54 <0.01 9.67 – 28.29 kidney disease (male = 4, female = 7), diabetes (male = 126, female =
Absent 157), and hypertension (male = 235, female = 382), the female popu­
Hypertension lation was more extensive than the male population. Similar results for
Present 20.25 <0.01 12.70 – 32.28 kidney and hypertension were also found in these studies [29,24],
Absent
However, for diabetics, the male population was more prevalent than
Kidney 11 Hypertension
Disease the female population, as found by these studies [30,31],. Also, we
Present 16.51 <0.01 5.02–54.24 found that NCDs were more prevalent among the “>= 40′′ years age
Absent group. It is hard to compare with other studies since different studies
Diabetes
represented the age range differently, and most of the studies were only
Present 25.51 <0.01 7.43–87.65
Absent
for the adult population. With the exception of CVD, the population for
CVD stroke, hypertension, diabetes increased until the 50–59 age range,
Present 67.85 <0.01 17.63–261.08 following which the population proportion declined as age increased.
Absent From “BMI” measurement, it can be observed that 164 (1.22%) adult
CVD = Cardiovascular disease; COPD = Chronic Obstructive Pulmonary Disease. people were obese, and 1094 (9.52%) adult people were underweight. In
addition, female adults (744, in%: 10.65) were underweight at a greater
number than male adults (350, in%: 7.78). Similar results also observed
Table 4 for young people. Even though young female population was larger, 44
Measurement value count. (7.35%) young males were obese, significantly more than their female
Measurement Name No. of People
counterparts (23, or 3.51%). However, females were underweight at a
higher rate (75, or 11.43%) than males (53, or 8.85%). According to
BP (Blood Pressure) 11,328
MUAC measurements, 40 young people (3.19%) were suffering from
Pulse Rate 11,178
BMI (Body Mass Index) 7065 severe malnutrition. Females were more prevalent than males among
Blood Sugar 6444 them. 26 (3.96%) young females were severely malnourished, whereas
SpO2 (Oxygen Saturation) 3453 14 (2.34%) young males were severely malnourished. We argue that this
TEMP (Temperature) 3398 discrimination is due to rural people’s tendency to give more food to the
MUAC (Mid-Upper Arm Circumference) 192
male members than the female members [32].
Our study has several strengths. To the best of our knowledge this is
four digital platforms and GP centers. Two of these four digital platforms the first digital GP model introduced for rural Bangladesh. In addition,
("Health workers application" and "User application") are Android ap­ we analyzed health related data from Nayanagar union and outlined
plications, while the other two are web applications ("Doctor dashboard" some key findings. Having said that, our study has several limitations.
and "Admin dashboard"). Besides, patients can visit with doctors via To begin with, we analyzed only one union, Nayanagar. As a result, our
telemedicine or in person at a GP center. Additionally, the GP center study’s findings cannot be inferred to the entire rural population in
provides diagnostic services such as "blood tests" and "urine tests." Our Bangladesh. Finally, because the data is skewed toward females, it may
neighbor country, India began the National Digital Health Mission underestimate the true prevalence of non-communicable diseases and
health vital measurements in the study group.

9
M.H. Chowdhury et al. Health Policy and Technology xxx (xxxx) xxx

Table 5
Distribution of overall measured population over the age of 18 years based on BP, BMI, Pulse Rate, SpO2, TEMP and Blood Sugar.
Overall 95% CI Male 95% CI Female 95% CI
Total Population (age >=18) 11,491 4496 (39.13%) 6995 (60.87%)

BP 11,105 (96.64%) 96.31–96.97 4258 (94.71%) 94.05–95.36 6847 (97.88%) 97.55–98.22

Low 1191 (10.36%) 9.81–10.92 358 (7.96%) 7.17–8.75 833 (11.91%) 11.15–12.67
Normal 4377 (38.09%) 37.2–38.98 1502 (33.41%) 32.03–34.79 2875 (41.1%) 39.95–42.25
Prehypertension 3390 (29.5%) 28.67–30.34 1534 (34.12%) 32.73–35.51 1856 (26.53%) 25.5–27.57
Mild High 1399 (12.17%) 11.58–12.77 562 (12.5%) 11.53–13.47 837 (11.97%) 11.21–12.73
Moderate High 569 (4.95%) 4.56–5.35 227 (5.05%) 4.41–5.69 342 (4.89%) 4.38––5.39
High 5 (0.04%) – 2 (0.04%) – 3 (0.04%) –
Severe High 174 (1.51%) 1.29–1.74 73 (1.62%) 1.25–1.99 101 (1.44%) 1.16–1.72
Pulse Rate 10,959 (95.37%) 94.99–95.75 4253 (94.6%) 93.93–95.26 6706 (95.87%) 95.4–96.33
Low 291 (2.53%) 2.25–2.82 203 (4.52%) 3.91–5.12 88 (1.26%) 1.0–1.52
Normal 9677 (84.21%) 83.55–84.88 3804 (84.61%) 83.55–85.66 5873 (83.96%) 83.1–84.82
High 991 (8.62%) 8.11–9.14 246 (5.47%) 4.81–6.14 745 (10.65%) 9.93–11.37
BMI 6477 (56.37%) 55.46–57.27 1982 (44.08%) 42.63–45.53 4495 (64.26%) 63.14–65.38
Underweight 1094 (9.52%) 8.98–10.06 350 (7.78%) 7.0–8.57 744 (10.64%) 9.91–11.36
Normal 4384 (38.15%) 37.26–39.04 1378 (30.65%) 29.3–32.0 3006 (42.97%) 41.81–44.13
Overweight 834 (7.26%) 6.78–––7.73 230 (5.12%) 4.47–5.76 604 (8.63%) 7.98–9.29
Obesity 140 (1.22%) 1.02–1.42 21 (0.47%) 0.27–0.67 119 (1.7%) 1.4–2.0
Highly Obesity 19 (0.17%) 0.09–0.24 2 (0.04%) – 17 (0.24%) 0.13–0.36
Morbid Obesity 6 (0.05%) – 1 (0.02%) – 5 (0.07%) –
TEMP 2701 (23.51%) 22.73–24.28 922 (20.51%) 19.33–21.69 1779 (25.43%) 24.41–26.45
Very low 272 (2.37%) 2.09–2.65 72 (1.6%) 1.23–1.97 200 (2.86%) 2.47–3.25
Low 1200 (10.44%) 9.88–11.0 431 (9.59%) 8.73––10.45 769 (10.99%) 10.26–11.73
Normal 1200 (10.44%) 9.88–11.0 401 (8.92%) 8.09–9.75 799 (11.42%) 10.68–12.17
High 29 (0.25%) 0.16–0.34 18 (0.4%) 0.22–0.58 11 (0.16%) 0.06–0.25
Very high – – – – – –
SpO2 2832 (24.65%) 23.86–25.43 1060 (23.58%) 22.34–24.82 1772 (25.33%) 24.31–26.35
Very low 54 (0.47%) 0.34–0.59 26 (0.58%) 0.36–0.8 28 (0.4%) 0.25–0.55
Low 144 (1.25%) 1.05–1.46 81 (1.8%) 1.41–2.19 63 (0.9%) 0.68–1.12
Normal 2634 (22.92%) 22.15––23.69 953 (21.2%) 20.0–22.39 1681 (24.03%) 23.03–25.03
Blood Sugar 6421 (55.88%) 54.97–56.79 2100 (46.71%) 45.25–48.17 4321 (61.77%) 60.63–62.91
Low (Hypoglycemia) 40 (0.35%) 0.24 – 0.46 11 (0.24%) 0.1 – 0.38 29 (0.41%) 0.26 – 0.57
Normal 5660 (49.26%) 48.34–50.17 1826 (40.61%) 39.18–42.05 3834 (54.81%) 53.64–55.98
Pre- Diabetic 62 (0.54%) 0.41–0.67 28 (0.62%) 0.39–0.85 34 (0.49%) 0.32–0.65
Diabetic (need confirmation) 466 (4.06%) 3.69–4.42 153 (3.4%) 2.87–3.93 313 (4.47%) 3.99–4.96
High (Borderline) 29 (0.25%) 0.16–0.34 17 (0.38%) 0.2–0.56 12 (0.17%) 0.07–0.27
High 164 (1.43%) 1.21–1.64 65 (1.45%) 1.1–1.79 99 (1.42%) 1.14–1.69

BP = Blood Pressure; BMI = Body Mass Index; TEMP = Temperature; SpO2 = Blood Oxygen Saturation.

Conclusion & future work Competing interests

Overall, the digital GP model for rural Bangladesh presents digital None.
health accounts, health records, healthcare at doorsteps, telemedicine at
doorsteps, data driven operational decision making, digital prescription Ethical approval
and promote health awareness among rural people. We believe, by
adapting this digital GP model, rural people may be able to escape The ethical review board of United International University
higher-level health facilities for primary healthcare services and reduce approved to conduct the study. All of the above-mentioned data were
out-of-pocket expenditure. By scaling up this digital GP model across the gathered with the consent of participants.
Bangladesh may achieve sustainable development goals (SDG), and
universal health coverage. Moreover, from our data analysis, we see that
CRediT authorship contribution statement
NCDs such as hypertension, diabetes were prevalent among rural peo­
ple. Necessary steps should be taken to raise awareness among rural
Moinul H. Chowdhury: Formal analysis, Writing – original draft,
people. In the future, we aim to show impact of digital GP model by
Writing – review & editing. Rony Chowdhury Ripan: Conceptualiza­
collecting data from rural areas across Bangladesh. Also, future re­
tion, Investigation, Methodology, Writing – review & editing. A.K.M.
searchers can analyze NCD behavioral risk factors by accessing socio-
Nazmul Islam: Conceptualization, Investigation, Methodology, Formal
demographic data [33] and can find risk factors and predictors of BP,
analysis, Writing – review & editing. Rubaiyat Alim Hridhee:
diabetes, etc. [24]. In addition, the existing GP model can be improved
Conceptualization, Investigation, Methodology, Formal analysis,
by implementing an A.I. based symptom checker. Also, CMED health is
Writing – review & editing. Farhana Sarker: Conceptualization,
working towards making this digital GP model more sustainable, cost
Investigation, Methodology, Writing – review & editing. Sheikh
effective. This digital GP model is only implemented in one union and
Mohammed Shariful Islam: Investigation, Writing – review & editing,
now CMED health is trying to scale it so that all the rural people of
Supervision. Khondaker A. Mamun: Conceptualization, Investigation,
Bangladesh can get healthcare benefits from this model.
Methodology, Formal analysis, Writing – original draft, Writing – review
& editing, Supervision.
Funding

None.
Declaration of Competing Interest

The authors have no conflict of interest.

10
M.H. Chowdhury et al. Health Policy and Technology xxx (xxxx) xxx

Table 6
Distribution of overall measured population under age 18 based on BP, BMI, Pulse Rate, MUAC, SpO2, TEMP and Blood Sugar.
Overall 95% CI Male 95% CI Female 95% CI
Total Population (age < 18) 1255 599 (47.73%) 656 (52.27%)

BP 223 (17.77%) 15.65–19.88 80 (13.36%) 10.63–16.08 143 (21.8%) 18.64–24.96

Low 110 (8.76%) 7.2–10.33 30 (5.01%) 3.26–6.76 80 (12.2%) 9.69–14.7
Normal 85 (6.77%) 5.38–8.16 36 (6.01%) 4.11–7.91 49 (7.47%) 5.46–9.48
High 28 (2.23%) 1.41–3.05 14 (2.34%) 1.13–3.55 14 (2.13%) 1.03–3.24
Pulse Rate 219 (17.45%) 15.35–19.55 79 (13.19%) 10.48–15.9 140 (21.34%) 18.21–24.48
Low 6 (0.48%) – 3 (0.5%) – 3 (0.46%) –
Normal 195 (15.54%) 13.53–17.54 68 (11.35%) 8.81–13.89 127 (19.36%) 16.34–22.38
High 18 (1.43%) 0.78–2.09 8 (1.34%) – 10 (1.52%) –
BMI 588 (46.85%) 44.09–49.61 258 (43.07%) 39.11–47.04 330 (50.3%) 46.48–54.13
Underweight 128 (10.2%) 8.52–11.87 53 (8.85%) 6.57–11.12 75 (11.43%) 9.0–13.87
Normal 354 (28.21%) 25.72–30.7 149 (24.87%) 21.41–28.34 205 (31.25%) 27.7–34.8
Overweight 39 (3.11%) 2.15–4.07 12 (2.0%) 0.88–3.13 27 (4.12%) 2.6–5.64
Obesity 67 (5.34%) 4.09–6.58 44 (7.35%) 5.26–9.43 23 (3.51%) 2.1–4.91
TEMP 697 (55.54%) 52.79–58.29 327 (54.59%) 50.6–58.58 370 (56.4%) 52.61–60.2
Very low 42 (3.35%) 2.35–4.34 14 (2.34%) 1.13–3.55 28 (4.27%) 2.72–5.82
Low 248 (19.76%) 17.56–21.96 106 (17.7%) 14.64–20.75 142 (21.65%) 18.49–24.8
Normal 391 (31.16%) 28.59–33.72 196 (32.72%) 28.96–36.48 195 (29.73%) 26.23–33.22
High 15 (1.2%) 0.59–1.8 11 (1.84%) 0.76–2.91 4 (0.61%) –
Very high 1 (0.08%) – – – 1 (0.15%) –
SpO2 621 (49.48%) 46.72–52.25 306 (51.09%) 47.08–55.09 315 (48.02%) 44.2–51.84
Very low 36 (2.87%) 1.95–3.79 16 (2.67%) 1.38–3.96 20 (3.05%) 1.73–4.36
Low 23 (1.83%) 1.09–2.57 11 (1.84%) 0.76––2.91 12 (1.83%) 0.8–2.85
Normal 562 (44.78%) 42.03–47.53 279 (46.58%) 42.58–50.57 283 (43.14%) 39.35–46.93
Blood Sugar 23 (1.83%) 1.09–2.57 6 (1.0%) – 17 (2.59%) 1.38–3.81
Low 1 (0.08%) – 1 (0.17%) – – –
Normal 20 (1.59%) 0.9–2.29 4 (0.67%) – 16 (2.44%) 1.26–3.62
High 2 (0.16%) – 1 (0.17%) – 1 (0.15%) –
MUAC 192 (15.3%) 13.31–17.29 99 (16.53%) 13.55–19.5 93 (14.18%) 11.51–16.85
Normal 151 (12.03%) 10.23 – 13.83 84 (14.02%) 11.24–16.8 67 (10.21%) 7.89–12.53
Moderate 1 (0.08%) – 1 (0.17%) – – –
Severe 40 (3.19%) 2.22–4.16 14 (2.34%) 1.13–3.55 26 (3.96%) 2.47–5.46

BP = Blood Pressure; BMI = Body Mass Index; TEMP = Temperature; SpO2 = Blood Oxygen Saturation; MUAC = Mid-Upper Arm Circumference.

Acknowledgments [11] Nujhat S, Alam W, Parajuli A, Mohsen WAM, Banyira L, Gupta RD, et al.
Prevalence of risk factors for non-communicable diseases in a rural population of
Bangladesh: a cross-sectional study. Lancet Glob Health 2020;8(S21). https://doi.
The authors are grateful to the United Trust and United International org/10.1016/S2214-109X(20)30162-5 [Internet]Available from.
University, Bangladesh for their support and cooperation to successfully [12] Current health expenditure (% of GDP) - Bangladesh | Data [Internet]. [cited 2022
implement the pilot program. Apr 11]. Available from: https://data.worldbank.org/indicator/SH.XPD.CHEX.GD.
ZS?locations=BD.
[13] Online Doctor and Telemedicine Services in BD. [cited 2022 Mar 23]. Available
References from: https://www.doctime.com.bd/.
[14] Our Services | Digital Healthcare Solutions [Internet]. [cited 2022 Mar 22].
[1] Bulletin JH-TS for the SH of M, 1983 undefined. The origin of the royal college of Available from: https://dh.health/OurServices.
general practitioners. europepmc.org [Internet]. [cited 2022 Mar 22]; Available [15] Rahman M., Chowdhury M.H., Hridhee R.A., Islam T. Implementation of a digital
from: https://europepmc.org/article/med/11611312. healthcare service model for ensuring preventive and primary healthcare in rural.:
[2] Malone R. The role of the physician associate: an overview. Ir J Med Sci 2021;191: 1–14.
1277–83. [16] CMED - Connecting People, Saving Lives [Internet]. [cited 2022 Mar 22]. Available
[3] Inga Marthe Gronseth, Kirsti Malterud & Stein Nilsen, Why do doctors in Norway from: https://cmed.com.bd/.
choose general practice and remain there? A qualitative study about motivational [17] Sailunaz K, Alhussein M, Shahiduzzaman M, Anowar F, Mamun KA. CMED: cloud
experiences. Scandinavian Journal of Primary Health Care, 38 (2) 184-191, based medical system framework for rural health monitoring in developing
10.1080/02813432.2020.1753348, 12 may 2020. countries. Comput Electr Eng 2016;53:469–81. Jul 1.
[4] Silva BMC, Rodrigues JJPC, de la Torre Díez I, López-Coronado M, Saleem K. [18] Union councils of Bangladesh - Wikipedia [Internet]. [cited 2023 Jan 14].
Mobile-health: a review of current state in 2015. J Biomed Inform 2015;56:265–72. Available from: https://en.wikipedia.org/wiki/Union_councils_of_Bangladesh.
Aug 1. [19] Ripan RC, Sarker IH, Hossain SMDM, Anwar MM, Nowrozy R, Hoque MM, et al.
[5] Martínez-Pérez B, de La Torre-Díez I, López-Coronado M, Herreros-González J. A data-driven heart disease prediction model through k-means clustering-based
Mobile apps in cardiology: review. JMIR Mhealth Uhealth 2013;1(2):e15. anomaly detection. SN Comput Sci 2021;2(2):1–12. 2021 2:2 [Internet]Feb 24
[6] Chris Salisbury for NHS Hammersmith and Fulham CCG and NHS England. [cited 2022 Apr 23]Available from, https://link.springer.com/article/10.1007/s42
Evaluation of babylon GP at hand final evaluation report prepared by Ipsos MORI 979-021-00518-7.
and York health economics consortium with prof. Chris Salisbury for NHS [20] World Health Organization. Package of essential noncommunicable (PEN) disease
Hammersmith and Fulham CCG and NHS England; 2019 [Internet]Available from, interventions for primary health care in low-resource settings WHO PEN protocol 1
http://www.ipsos-mori.com/terms. prevention of heart attacks, strokes and kidney disease through integrated
[7] Matt Reynolds, NHS to start prescribing health apps that help manage conditions | management of diabetes and hypertension. World Health Organization; 2012.
New Scientist [Internet]. [13 February 2017]. Available from: https://www.news p. 14–7. PEN.
cientist.com/article/2121164-nhs-to-start-prescribing-health-apps-that-help [21] Managing hypertension in primary care in the caribbean [Internet]. [cited 2022
-manage-conditions/. Mar 28]. Available from:https://www.researchgate.net/publication/291957607_
[8] Bangladesh Population 2023 (Demographics, Maps, Graphs) [Internet]. [cited Managing_Hypertension_in_Primary_Care_in_the_Caribbean/figures?lo=1.
2023 Mar 22]. Available from: https://worldpopulationreview.com/countries/ba [22] Bajpai N., Wadhwa M. India’s national digital health mission. 2020 [cited 2022
ngladesh-population. Apr 5]; Available from: https://academiccommons.columbia.edu/doi/10.7916/d
[9] Bangladesh: share of rural population in Bangladesh from 2012 to 2021 | Statista 8-wq8f-9094.
[Internet]. [cited 2023 Mar 22]. Available from: https://www.statista.com/statis [23] Kazi AM, Qazi SA, Ahsan N, Khawaja S, Sameen F, Saqib M, et al. Current
tics/760934/bangladesh-share-of-rural-population/. challenges of digital health interventions in Pakistan: mixed methods analysis.
[10] Md. Shariful Alam Khandakar, Rural Health Care System and Patients... J Med Internet Res 2020;22(9):e21691. https://www.jmir.org/2020/9/e21691
Satisfaction towards Medical Care in Bangladesh: An Empirical Study; Journal of [Internet]. 2020 Sep 3 [cited 2022 Apr 5];22(9):e21691.
Business Studies, XXXV, (2) 83-102, August 2014.. Available from: https://www. [24] Barua L, Faruque M, Rakhshanda S, Banik PC, Shawon RA, Mashreky SR. Baseline
fbs-du.com/news_event/14664850464.%20Md.pdf. prevalence of high blood pressure and its predictors in a rural adult population of

11
M.H. Chowdhury et al. Health Policy and Technology xxx (xxxx) xxx

Bangladesh: outcome from the application of WHO PEN interventions. J Clin [29] Banik S, Ghosh A. Prevalence of chronic kidney disease in Bangladesh: a systematic
Hypertens 2021;23(12):2042–52. Dec 1. review and meta-analysis. Int Urol Nephrol 2020;53(4):713–8. 2020 53:4
[25] Ziaul M, Chowdhury I, Rahman M, Akter T, Akhter T, Ahmed A, et al. Hypertension [Internet]Aug 12 [cited 2022 Mar 23]Available from, https://link.springer.com/ar
prevalence and its trend in Bangladesh: evidence from a systematic review and ticle/10.1007/s11255-020-02597-6.
meta-analysis. Clin Hypertens 2020;26(1):1–19. 2020 26:1 [Internet]Jun 1 [cited [30] Mørkrid K, Ali L, Hussain A. Risk factors and prevalence of diabetic peripheral
2022 Mar 23]Available from, https://clinicalhypertension.biomedcentral.com/arti neuropathy: a study of type 2 diabetic outpatients in Bangladesh. Int J Diabetes
cles/10.1186/s40885-020-00143-1. Dev Ctries 2010;30(1):11 [Internet]Jan 1 [cited 2022 Mar 23]Available from, htt
[26] Saha UK, Alam MB, Rahman AKMF, Hussain AHME, Mashreky SR, Mandal G, et al. ps://pmc/articles/PMC2859278.
Epidemiology of stroke: findings from a community-based survey in rural [31] Biswas T, Islam A, Rawal LB, Islam SMS. Increasing prevalence of diabetes in
Bangladesh. Public Health 2018;160:26–32. Jul 1. Bangladesh: a scoping review. Public Health 2016;138:4–11. Sep 1.
[27] Sutradhar I, Gupta RD, Hasan M, Wazib A, Sarker M. Prevalence and risk factors of [32] Gender inequality and severe malnutrition among children in a remote rural area
chronic obstructive pulmonary disease in Bangladesh: a systematic review. Cureus of Bangladesh on JSTOR [Internet]. [cited 2022 Mar 23]. Available from: https://
2019;11(1) [Internet]Jan 28 [cited 2022 Mar 23]Available from, https://www.cu www.jstor.org/stable/23498867?casa_token=ZssPTIYaDdQAAAAA%3AJYvQJF
reus.com/articles/15943-prevalence-and-risk-factors-of-chronic-obstructive-pu e64Zu0T9x3alT3-W15_2K3bFwN81SDj2sfccVIzqp4c7LNGTaVIY_Rme7y44GJGaR
lmonary-disease-in-bangladesh-a-systematic-review. 1trSsq5e00TqvVoC9xRAhw6jccMEPI0NmzKRxuB6rCks&seq=1.
[28] Monwarul Islam A., Mohibullah A., Paul T., Professor A. Cardiovascular disease in [33] Negi PC, Chauhan R, Rana V, Vidyasagar LK. Epidemiological study of non-
bangladesh: a review. communicable diseases (NCD) risk factors in tribal district of Kinnaur, HP: a cross-
sectional study. Indian Heart J 2016;68(5):655–62. Sep 1.

12