Table of Contents

Acknowledgment ............................................................................................3
List of Table ....................................................................................................4
List of Figure ..................................................................................................4
Summary.........................................................................................................4
Acronyms ........................................................................................................5
In order to have a homogenous cohort and reduce the risk of confounding
bias, the analysis will not consider certain groups. Of which, Patients with
secondary hypertension and pregnant women suffering from hypertensive
disorders of pregnancy are excluded, because the etiology, treatment
response, and clinical course of these conditions are quite different from
those of primary, essential hypertension. Moreover, to maintain the
methodological strongest, patients with incomplete critical baseline data
such as; missing initial BP or starting drug regimen will be left out, as
these gaps would hinder the correctness of the survival analysis. In
addition, patients whose medical records are lost, inaccessible, or have
fewer than the minimum required BP measurements will be excluded in
order to assure the reliability and completeness of the longitudinal data
necessary for this study. ............................................................................24
Where, ....................................................................................................30
APPENDIX.....................................................................................................51
 Image

BAHIR DAR UNIVERSITY

COLLEGE OF MEDICINE AND HEALTH
SCIENCES
DEPARTMENT OF EPIDEMIOLOGY AND
BIOSTATISTICS
TIME TO CONTROL HYPERTENSION AND
ITS ASSOCIATED FACTORS AMONG
HYPERTENSIVE PATIENTS WHO HAVE
FOLLOWED AT DEBRE TABOUR REFERAL
HOSPITAL, NORTHWESTERN ETHIOPIA
A RESEARCH PROPOSAL SUBMITTED TO THE
DEPARTMENT OF EPIDEMIOLOGY AND
BIOSTATISTICS, SCHOOL OF PUBLIC
HEALTH, COLLEGE OF MEDICINE AND
HEALTH SCIENCES, IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR
THE DEGREE OF MASTERS IN PUBLIC
HEALTH (BIOSTATISITCS)
BY: BIRARA ALEMU
ADVISOR: PROFESSOR ESSEY KEBEDE
(PROFESSOR OF BIOSTATISTICS)
CO-ADVISOR: MS. TAYE ABUHAY (ASSISTANT
PROFESSOR OF BIOSTATISTICS)
JANURAY 2025
BAHIR DAR, ETHIOPI

Acknowledgment

I would like to express my deepest gratitude to my

advisor, Professor Essey K. (Professor of Biostatistics)
and my co-advisor, Taye A. (Assistant Professor of
Biostatistics) for their unwavering guidance, insightful
feedback, and unreserved support throughout the
development of this research proposal. Their expertise,
patience, and encouragement have been invaluable in
shaping the direction and rigor of this work.

I also extend my sincere appreciation to the faculty and

administration of Bahir Dar University for providing the
academic environment and opportunities. My thanks go
to the management and staff of Debre Tabor Referral
Hospital for their willingness to facilitate this study.

List of Table

List of Figure

Summary

Background: Hypertension remains a leading

modifiable risk factor for cardiovascular diseases
globally, with disproportionately low control rates in
Sub-Saharan Africa. This study aimed to identify the
predictors of hypertension among patients at Debre
Tabor Referral Hospital in Northwestern Ethiopia. Key
determinants, including demographic variables,
comorbid conditions, medication adherence, lifestyle
factors, treatment protocols, type of health insurance,
among others, will be examined.

Method: A retrospective cohort study will be conducted

using records of 402 hypertensive patients sampled from
Debre Tabor Hospital (2021-2025). Descriptive statistics
will summarize patient characteristics. Kaplan-Meier
curves will estimate the median time to control. The Log-
Rank test will compare this time between key patient
groups (e.g., by drug regimen or residence). Accelerated
Failure Time (AFT) models will identify predictors of
control time, with a shared frailty component tested for
unobserved clustering. The data will be recorded using
Epi data and the analysis will use Stata and R software.

Expected Results: The analysis will yield three

principal findings corresponding to the analytical plan.
First, the median survival time to hypertension control
will be determined for the overall cohort using the
Kaplan-Meier estimator, establishing a crucial local
benchmark for management efficiency. Second, the
survival status of different patients will be compared
using the log-rank test to identify significant disparities
in time to control. Finally, the risk factors (predictors)
associated with time to control will be identified through
Accelerated Failure Time (AFT) modeling.

Work Plan & Budget: The study will be completed over

five months, encompassing data extraction, analysis, and
thesis writing. The total required budget is estimated at
34,233 ETB, covering personnel, data access, materials,
and contingency.

Keywords: Hypertension control; Time-to-event;

Survival analysis; Accelerated Failure Time model.

Acronyms
AFT Accelerated Failure Time

AIC Akaike

BIC Bayesian Information Criterion (BIC

AOR Adjusted Odd Ratio

BP Blood Pressure

CVD Cardiovascular Diseases

DBP Diastolic Blood Pressure

DASH Dietary Approaches To Stop Hypertension

FMOH Federal Ministry Of Health

IBR Institutional Board Review

LMICS Low- And Middle-Income Countries

MMM May Measurement Month

MLE Maximum Likelihood Estimation

NCD Non-Communicable Diseases

OR Odd Ratio

SBP Systolic Blood Pressure

SSA Sub-Saharan African

WHO World Health Organization

Hypertension, characterized by a persistent elevation of

systolic blood pressure (SBP) ≥140 mmHg and/or
diastolic blood pressure (DBP) ≥90 mmHg, is a chronic,
non-communicable disease (NCD) and a leading
modifiable risk factor for global morbidity and mortality
(1). Often termed the "silent killer," it frequently
presents with no overt signs or symptoms until
significant damage has occurred to target organs such
as the heart, brain, and kidneys. When symptoms
manifest, they can include headaches, shortness of
breath, dizziness, and chest pain, although these are
non-specific. The primary pathophysiological mechanism
involves increased systemic vascular resistance and
cardiac workload, leading to endothelial dysfunction,
arterial stiffening, and remodeling over time (2).

The global burden of hypertension has reached alarming

levels. Current estimates indicate that approximately 1.4
billion adults worldwide live with hypertension, with the
prevalence steadily rising due to aging populations and
lifestyle transitions(3). This burden is disproportionately
borne by low- and middle-income countries (LMICs),
where over two-thirds of affected individuals reside, and
health systems are often least equipped to manage this
chronic condition effectively(4). In the World Health
Organization (WHO) African Region, the age-
standardized prevalence is among the highest globally,
driven by rapid urbanization, dietary shifts, and
increasing rates of obesity and physical inactivity (5).
Within this context, Sub-Saharan Africa faces a
particularly complex challenge, facing a double burden
of persistent infectious diseases and a sharply rising
epidemic of non-communicable diseases, such as
hypertension(6). In Ethiopia, hypertension is now
recognized as a leading non-communicable disease
(NCD), contributing to the nation's double burden of
disease alongside persistent infectious challenges, and
national estimates suggest a hypertension prevalence
exceeding 20% among adults, with significant variations
observed across different regions and study
populations(7). In response, the Ethiopian Federal
Ministry of Health has recognized hypertension as a
priority and has developed strategic plans and treatment
guidelines (8). However, a persistent gap exists between
guideline recommendations and real-world clinical
outcomes. Evidence from hospital-based studies across
Ethiopia consistently reveals suboptimal control rates,
often reported between 26% and 53%(9, 10). These
cross-sectional studies cannot express the dynamic
process of achieving control.

In this study, Accelerated Failure Time (AFT) models will

be utilized, which are a class of parametric survival
analysis techniques particularly suited for this research.
Unlike methods that only assess proportional hazards,
AFT models are ideal for analyzing the time required to
reach a desired clinical endpoint (such as BP control)
because they directly model how predictors accelerate or
decelerate the time to an event. This provides clinically
intuitive results, showing which factors shorten or
prolong the journey to control the disease. Given the
potential for unmeasured heterogeneity in patient
responses, the analysis will also explore shared frailty
extensions of the AFT model to account for latent
clustering, thereby ensuring robust and nuanced
insights.

Therefore, this study aimed to determine the time to

hypertension control and its predictors among patients
at Debre Tabor Comprehensive Specialized Hospital
using these advanced time-to-event analytical
approaches, generating evidence to inform more timely
and effective hypertension management strategies.

Hypertension continues to be the leading modifiable risk

factor for cardiovascular disease globally, affecting an
estimated 1.4 billion adults, two-thirds of whom live in
low- and middle-income countries(3). The global medical
community has shown that blood pressure control must
be achieved within 1–3 months of treatment initiation to
minimize cardiovascular risk (2). High-income countries
consistently meet this standard; Spain achieved control
in 49 days, the US in 3.3 months, and South Korea in just
2.1 months (4, 11, 12). However, as healthcare systems
transition from high-income to low-resource settings,
this timeline expands dramatically, creating what can
only be described as a global timeline equity crisis.

In Africa, especially sub-Saharan Africa, the prevalence

of hypertension has reached 27%, which is the highest of
any world region, yet the rates of control are less than
16%. The timeline to reach control stretches from
months to years: 19.6 months in Ghana, 22.4 months in
Kenya, and astoundingly, 48 months in Ethiopia,
according to a study based at Felege-Hiwot Hospital,
Bahir Dar, Ethiopia. (13-15). This represents a 35-fold
increase compared with Spain and a 15-fold increase
compared with the global standard. Each additional
month of delay compounds cardiovascular risk by 2-3%,
meaning that African patients face dramatically higher
probabilities of stroke, heart failure, and kidney disease
due to extended periods of uncontrolled
hypertension(16).
In Ethiopia, the picture of hypertension management is both fragmented
and contradictory. Despite the country experiencing an alarmingly
increasing hypertension burden, estimated at 20–30% of adults (or roughly
15–20 million people), evidence on the effectiveness of its management
remains low and incoherent. Two studies available on time-to-control paint
a contrast: a study in Bahir Dar presents 48 months, while another study
conducted in the Afar Region shows 13 months (13, 17). The 35-month
difference represents not only a regional difference but also an
unmistakable example of a medical system that is at risk and in crisis, as
indicated by the 48-month delay compared to the 13-month delay. Both
numbers represent a significant variation from the globally accepted norm
of 1-3 months. Patients suffer prolonged exposure to avoidable
cardiovascular risks, regardless of whether they experience a delay of one
or four years. The most urgent question is no longer how these two periods
differ. Rather, why do both times differ from the absolute target by so far?
The gap between the standard of care that can be realized and what is
actually provided to patients in Ethiopia indicates that there is a serious
need to fully investigate the combined effects of systemic, clinical, and
patient-level issues that are preventing patients from receiving timely
medical care and are leading to numerous patient deaths each month.

The methodological limitations of existing Ethiopian hypertension research

add to this knowledge gap. A systematic review indicated that 87% of
studies employ cross-sectional designs to assess prevalence and control
rates at single time points, whereas only 6% investigate time-to-event
outcomes (18). Studies have confirmed the scarcity of comparable time-to-
control blood pressure studies in Ethiopia (17). This builds an evidence base
that can tell us how many patients achieve control but not how long it takes
them to get there, which is a critical distinction for clinical management
and health system planning. Moreover, existing time-to-control studies
examine limited predictors, mainly basic demographics, while ignoring
health insurance type and medication adherence predictors, even though
they are critical factors in hypertension control. This study will employ a
survival analysis approach to determine the median time to blood
pressure control and identify the key predictors of blood pressure
control among hypertensive patients at Debre Tabor Hospital.
Specifically, a parametric Accelerated Failure Time (AFT) model will
be used for its direct interpretability of predictors on survival time.
To account for potential unobserved heterogeneity or clustering
within the data (e.g., by physician, clinic unit, or residence), a
shared frailty component will be incorporated. The results will provide
evidence for designing interventions that accelerate control and ultimately
lower cardiovascular risk in this vulnerable population, establish a critical
local benchmark, and identify modifiable care bottlenecks.

Accordingly, the following research questions will be answered:

• What is the median time to control blood

pressure among hypertensive patients ?
• What are the potential predictors that affect the time
to control blood pressure among hypertensive
patients?
• Which parametric models fit the time to control blood
pressure among hypertensive patients?
•
This study aims to provide the first local evidence on
how long it takes for people with high blood pressure to
control it at Debre Tabor Hospital. This study addresses
a significant gap in Ethiopia's research on high blood
pressure. Instead of just looking at control rates at one
point in time, this research will check how long it takes
things to happen. It will use special models to set a key
standard, find things we can change that cause delays,
and spot problems in how people get care.

The study findings will help health professionals provide

care based on risk and help health facilities make better
follow-up plans, medicine rules, and use resources well.
In the end, this means faster, better and more personal
high blood pressure care for people in the area.

The results will back wider health programs and rules by

backing Ethiopia’s plan to cut heart disease and early
deaths from other diseases. By getting useful info on
treatment times and things that matter, this study can
help adapt the country's high blood pressure rules to fit
places with few resources, like Debre Tabor. This work
aims not just to count how long control takes but to
speed it up. It hopes to close the gap between starting
treatment and getting safe blood pressure, lower
problems that can be stopped, and boost the life quality
for lots of high blood pressure patients in the area.

Cardiovascular disease (CVD) has become one of the

leading causes of death around the globe(19). From the
risk factors for CVD, high blood pressure (BP) is
accounts with the strongest evidence for causation and it
has a high prevalence of exposure. Many cohort studies
showed that a high blood pressure the main risk factor
for chronic heart disease, heart failure, heart valve
disease, atrial fibrillation, aortic syndromes and
dementia, as far as to coronary heart disease and
stroke(20). Beyond personal health, hypertension has a
significant economic impact globally. Hospital stays,
long-term medication, managing complications, and lost
productivity are all ways that uncontrolled hypertension
that increased healthcare costs(21). As a result of this,
controlling hypertension is not only a medical priority
but also a financial necessity for health systems around
the world, especially areas with scarce resources where
the burden is highest and the ability to manage its
complications and prevalence is most limited and
difficult (22).

Current estimates indicate that approximately 1.4 billion

adults worldwide live with hypertension(3). High blood
pressure is the silent killer that is responsible for more
than 10 million deaths every year (5). WHO African
region has reported to have the highest prevalence of
hypertension in the world at 46%, and it is expected to
continue increasing(23). Nearly one out of three Africans
has hypertension, with prevalence varying across the
regions. Africa is the magnetic pole of global
hypertension, and its prevalence is expected to be
rising(24).

Studies suggest the prevalence of hypertension is

increasing in Africa, and most of hypertensive individuals
are not aware of their own status(25). A household
survey analysis of 17 African countries, revealed that the
prevalence of hypertension is high but low rate
diagnosis and treatment in SSA (26).
A systematic review and meta-analysis conducted in SSA
showed that the burden of hypertension has increased.
However, a large proportion of individuals with
hypertension remains undiagnosed and untreated and
this is leading to the rising burden of cardiovascular
disease in the areas(27). In SSA, there is a high
prevalence hypertension, because of contrastingly low
awareness, diagnosis, and treatment and control
rates(28).

Another systematic review and meta-analysis using

PubMed, Google Scholar, and Web of Science databases
to get the relevant studies and twenty-six studies with a
sample size of 11,600 participants revealed that the
pooled and the mean prevalence of uncontrolled
hypertension were 50.29% and 49.55% respectively.
Moreover, it indicated that the overall pooled prevalence
of uncontrolled hypertension was considerably high(29).

A systematic review and meta-analysis conducted in

Ethiopia showed that, hypertension is becoming a major
public health problem in Ethiopia and approximately two
out of ten individuals of age greater than 18 years in
Ethiopia are living with hypertension(30).

A random effect of meta-analysis conducted using nine

studies revealed that prevalence of hypertension among
Ethiopian population was estimated to be 19.6 % and
result of the evidence provided that, proper attention
has to be given to primary prevention and control of
hypertension in the Ethiopian population especially for
adult hoods(31).

A meta-analysis and systematic review showed that the

pooled prevalence of hypertension was 21%, with a
significant difference in the prevalence of hypertension
across the regions of Ethiopia. That is, the highest
prevalence of hypertension was in the Oromia region
(27%), while the lowest was in Addis Ababa (17%)(32).

A community-based cross-sectional study conducted at

Debre Berhan town with a total sample size of 680
showed that the prevalence of hypertension at the
community level was 27.5%(33).

An institutional based cross sectional study conducted at

Felege-Hiwot Comprehensive Referral Hospital on the
prevalence and associated factors of hypertension with a
total of 308-sample size; showed that the prevalence of
hypertension in this study was 27.3. Moreover, this study
suggests that, almost one-third of the study population
was hypertensive that means there was a hidden
epidemic in hypertension in this population(34).

A hospital-based cross-sectional study done from January

24 to February 25, 2023 in governmental hospitals in
Wolaita zone, South Ethiopia, with a total sample size of
422 showed that the prevalence of
prehypertension/hypertension was 42.8%(35).

Another systematic review and meta-analysis using

electronic bibliographic data base search from PubMud,
Google Scholar and other that contains 26 studies with
9046 hypertension patients reveled that ;the estimated
prevalence of uncontrolled HTN in Ethiopia is 51%(9).

A study that uses Demographic and Health Survey (DHS)

in India; in which the data collected from 636,699
households across all states and it includes the blood
pressure measurements for 17,08,241 individuals aged
15 and above and using binary logistic regression to
identify key risk factors of hypertension. Moreover, their
study showed that age is statistically significant and
when the age of individuals increased (when the age of
individuals 60 years or over is ten times more [OR: 8.97
(CI: 8.83–9.12)] likely to have hypertension compared to
a person of age 15 to 29 years)(36).

A cross sectional study conducted to identify the

predictors of blood pressure control among patients with
hypertension treated at the University of Teaching
Hospital, Calabar, Nigeria, with a total sample size of
441 hypertensive patients. And using bivariate and
multivariate logistic regression methods showed that
individuals whose aged greater than 54 years were
less likely to attain adequate BP control than young
adults, b=−1.19, p<0.001, AOR=0.30 (95% CI 0.18,
0.52)(37).

Another a hospital-based cross-sectional study was

conducted with 369 participants on blood pressure
control among adults with hypertension at a tertiary
hospital in Ethiopia using multivariate binary logistic
regression showed when the age of individual is
increased there is poor blood pressure control (AOR
(95% CI) : 1.05 (1.00-1.11)(38) .

A systematic review and meta-analysis that conducted to

estimate the prevalence of multi-morbidity among adults
in community settings using 126 peer-reviewed studies
that included nearly 15.4 million people using the
searched PubMed, Science Direct, Embase and Google
Scholar databases, revealed that having conditions of
comorbid highly influences the management and
outcomes of hypertensive patients. Globally,
hypertension rarely exists in itself; from 70-80% of
hypertensive populations have at least one additional
chronic condition, and 30-40% has multiple
comorbidities(39).

A systematic review and meta-analysis that conducted in

SSA that containing 20 studies involving 11,400
participants was included and their finding showed that
the existence of diabetes is a strong significant predictor
of delayed hypertension control(40).

A facility-based retrospective cohort study conducted

among adult hypertensive patients aged 18 years and
older in five public hospitals, using a total of 443-
hypertensive individuals by applying a Cox regression
analysis showed that individuals without underlying
comorbidities had a 1.9 times higher hazard of achieving
BP control compared to those with comorbidities(17).
Other an institution-based retrospective follow-up study
that conducted from September 2015 to April 2016 on
Blood pressure control status and associated factors
among adult hypertensive patients using a sample of 395
participants revealed that individuals that have
overweight and obese a less likely to controlled BP.
Thus, the likelihood of BP control decreased by 50% for
overweight and 44% for obese patients as compared to
normal/underweight patients(41).

Globally, the medication adherence to antihypertensive

therapy is a primary risk factor for poor control rates,
with non-adherence contributing to an estimated 45% of
uncontrolled hypertension cases. For time-to-event
analysis, poor adherence does not merely correlate with
worse outcomes; it prolongs the period to achieve the
target blood pressure, as treatment efficacy has
consistently interrupted(42). A cross sectional study with
a total of 249 samples and using regression analysis
showed that most patients scored high on imperfect
adherence (59.5%) to antihypertensive medication and
this leads to prolongs the period to achieve the target
blood pressure(43).

Another a hospital-based cross-sectional study conducted

to assess medication adherence and associated factors
among 197 hypertensive patients on treatments at
Shashemene Referral Hospital and using a chi square
measure of association. From the study subjects with
uncontrolled blood pressure, 50 (25.5%) respondents
had low adherence to antihypertensive medications, and
the remaining subjects 31 (15.7%), 13 (6.7%) had
medium adherence, and high adherence respectively and
this suggests that most of study participants has low
medication adherence(44).

Lifestyle modification is one of and basic

hypertension management and control and it is
recommended as a primary intervention mechanism in
all major clinical guidelines, including those from the
WHO, International Society of Hypertension, and
Ethiopian National Guidelines(45).

A clinical trial that includes 459 adults with SBP of less

than 160 mm hg and DBP of 80 to 95 mm hg on the
effects of dietary patterns on blood pressure. Their study
revealed that Dietary Approaches to Stop Hypertension
(DASH) diet rich in fruits, vegetables, whole grains, and
low-fat dairy reduce systolic BP by 5.5-11.4 mmHg in
as little as two weeks, a reduction comparable to single-
drug therapy and it is more effective when combined
with sodium restriction(46). Other studies showed that
high sodium intake (>5g/day) is linked to a 15-20%
longer median survival time to achieve blood
pressure control, as it has more efficacy that of many
antihypertensive drugs(47).

A systematic review and meta-analysis that used a data

of 36 articles including 2865 participants, their study
showed that excess drinking (>3 drinks/day) has a
significant effect. Decreasing high intake to moderate
levels can lower systolic BP by up to 4 mmHg. In
clinical cohorts, high drinkers have shown to take 3-4
months longer to reach BP targets than non-drinkers or
moderate drinker(48). Hospital-based cross sectional
study conducted from June 4 to August 31, 2019 on
Blood pressure control and its associated factors among
hypertensive patients in federal. Using 320 participants,
their finding showed that, there was a statistically
significant association between blood pressure control
and smoking (p<0.047), participants who smoke highly
being control BP is prolonged (49).
A facility-based retrospective cohort study was conducted among adult
hypertensive patients aged 18years and older in five public hospitals
between September 7, 2019 and January 8, 2023 on time to achieve blood
pressure control. The study showed that hypertensive patients with
creatinine level ≥ 1.5 mg/dL had a 62.9% lower hazard of BP control as
compared to those with creatinine < 1.5 mg/dL. This could be explained by
the fact that impaired kidney function, characterized by decreased water
and salt excretion and reduced production of BP-regulating hormones, leads
to delayed time to BP control(17).

A systematic review and meta-analysis that containing

42 randomized trials and the study concluded
that initial combination therapy (typically a CCB +
ACEI/ARB or ACEI/ARB + diuretic) achieved BP
control 2.8 times faster (Hazard Ratio: 2.8, 95% CI:
2.3–3.5) than monotherapy. The median time to control
decreased from 16.1 weeks with monotherapy to 5.8
weeks with combination therapy (50).

Most of previous Ethiopian studies conducted a cross-

sectional study design, which use data at a single point
in time. This approach limited for studying time to
control hypertension. Cross sectional study identifies
predictors associated with being controlled or
uncontrolled hypertension on a specific time, but is
unable to establish a temporal sequence. This study
design exposed for survivor bias, it systematically
excludes patients who censored before the survey
period, it highly underestimating the true difficulty to
achieving BP control (51). Moreover, the previous
research highly contradicted with studies conducted in
single hospitals or specific urban/rural settings. Results
from specialized hospital in Addis Ababa may not apply
to a primary hospital in the Afar region or a rural health
center. This creates uneven of local evidence without
a coherent national picture. The contradiction
between the studies result of the 48-month time to
control hypertension in Bahir Dar and the 13-month time
to control in Afar a showcase of that local factors
including clinician training, patient demographics, and
follow-up systems dominate outcomes. Yet, most studies
do not deeply analyze these system-level predictors,
focusing instead on patient-level factors(52).

In general, Ethiopia faces a major public health crisis

due to its high rates of hypertension. Unfortunately,
there are also low levels of awareness, treatment, and
ongoing control for this condition among those affected.
This literature will look at the many patient-specific risk
factors associated with hypertension, such as age,
comorbidities, and compliance with taking medications
and making healthy lifestyle choices that studied using
cross-sectional survey data.

However, there is an important element that remains

unexplored: How long does it take for people in Ethiopia
to achieve stable blood pressure control? Very little data
is available, and what exists is inconsistent. Times
reported for achieving stable blood pressure control
range from a median of 13 to 48 months, meaning
unmeasured contextual factors may have a significant
impact on rates of control. Additionally, the majority of
studies conducted to examine this topic employed
inappropriate statistical models (most often basic logistic
regression) that do not adequately represent the
complexity of the clinical care process and, therefore,
created a substantial gap in knowledge

This study will address the critical gap in understanding

the dynamics of hypertension management by employing
Accelerated Failure Time (AFT) models as the primary
analytical framework. The AFT model will be selected
over the conventional Cox proportional hazards model
due to its direct clinical interpretability for the research
question at hand. Unlike hazard ratios, the Time Ratios
(TR) generated by AFT models explicitly quantify how
patient characteristics and treatment factors accelerate
or decelerate the time to achieving blood pressure
control. This is more intuitive for clinicians and planners
who need to know not just if a factor increases risk, but
how much sooner or later control is achieved because of
it. Furthermore, the AFT framework does not require the
proportional hazards assumption, which is often violated
in real-world clinical data where treatment effects may
change over time. By applying this robust parametric
approach and incorporating a shared frailty term to
account for unobserved heterogeneity, the study will
generate a methodologically sound and locally relevant
evidence base. The findings will provide actionable, time-
oriented insights such as identifying which modifiable
factors cause the greatest delays enabling Debre Tabor
Hospital to design targeted interventions that shorten
the path to control and reduce long-term cardiovascular
risk for its hypertensive population.

Figure 1:
Conceptual Framework

This framework is based on Andersen's Behavioral Model

of Health Services Utilization, adapted for hypertension
control outcomes(53). It proposes that time to
hypertension control (the outcome) is influenced by
multiple interacting factors. Demographic factors (age,
sex) and enabling factors (insurance, residence) create
the initial conditions. These influence healthcare system
factors (treatment decisions) and behavioral
factors (patient actions), while need factors (clinical
severity) directly affect both treatment decisions and
outcomes. The framework also accounts for potential
interactions between key variables.

The main purpose of this study will be to determine the

time to control of blood pressure and to identify its
predictors among hypertensive patients attending at the
chronic follow-up clinic at Debre Tabor referral hospital,
Northwestern Ethiopia.

• Todetermine the median time to control hypertension

starting from treatment initiation.
• To identify predictors of time to hypertension
control among hypertensive patients .
• To compare the different parametric survival
models that fit time to control of blood pressure
among hypertensive patients.
•
A hospital based retrospective cohort study design will
be conducted at Debre Tabor comprehensive specialized
hospital from January 1, 2021 to December 30, 2025.

The study will be conducted at Debre Tabor

Comprehensive Specialized Hospital, located in the town
of Debre Tabor, northwestern Ethiopia. Debre Tabor
serves as the administrative capital of the South Gondar
Zone, within the Amhara National Regional State in
northwestern Ethiopia.

The Amhara region is the second most populous region

in Ethiopia, with an estimated population exceeding 30
million people(54). The South Gondar Zone is one of the
most populous zones within the region. According to the
latest national statistics, the zone has an estimated total
population of approximately 2.7 million people(55).
Debre Tabor town itself, the zonal capital where the
hospital is located, has a growing urban population
estimated at over 120,000 inhabitants (55). Debre Tabor
is approximately 670 kilometers north of Addis Ababa,
the national capital, and about 102 kilometers southeast
of Bahir Dar, the capital city of the Amhara region.
Debre Tabor Comprehensive Specialized Hospital is a
major public health institution with a substantial
catchment area that extends beyond the town to include
much of the South Gondar Zone. According to the
hospital's administration information, it provides
healthcare services to an estimated population of 3.5
million people from its catchment area. The hospital is a
significant employer and healthcare provider, with a
workforce of approximately 550 healthcare professionals
in addition to its administrative and support staff.

Furthermore, the hospital has an important academic

role. It functions as a primary teaching hospital for the
College of Medicine and Health Sciences at Debre Tabor
University, providing practical training for students
across various medical and health science disciplines.
This dual role as both a specialized service delivery
center and an academic institution underscores its
importance in the region's health.

The source population for this study will all diagnosed

adult hypertensive patients (≥18 years) who have been
placed on antihypertensive drug treatment, followed at
least three times by the Chronic Illness Follow-Up Clinic
of Debre Tabor Comprehensive Specialized Hospital
between 1/1/2021 and 12/31/2024

The study population will consist of a random systematic

selection of eligible patients, based on eligibility criteria
2021-2025.

The study will be include all the adult population of

patients aged 18 years and above who have been
diagnostically confirmed to suffer from primary
hypertension, thus identifying the relevant clinical
population. Patients should have started
antihypertensive drug therapy at or before their first
clinic visit within the study window (January 2021 to
December 2025) and have a recorded baseline blood
pressure (BP) to be considered a valid cohort for time to
event analysis. Additionally, a follow, up period of at
least six months from the start of therapy is necessary to
be able to observe significant clinical trajectories. Each
participant should have at least three BP measurements
from different clinic visits; this condition is not an
arbitrary one, but it is methodologically necessary to
evaluate the changing state of BP control over time. The
first measurement serving as a baseline, and subsequent
measurements allowing for the identification of the event
(e.g., loss of control after a period of control) in survival
analysis.

In order to have a homogenous cohort and reduce the risk

of confounding bias, the analysis will not consider certain
groups. Of which, Patients with secondary hypertension
and pregnant women suffering from hypertensive
disorders of pregnancy are excluded, because the
etiology, treatment response, and clinical course of these
conditions are quite different from those of primary,
essential hypertension. Moreover, to maintain the
methodological strongest, patients with incomplete
critical baseline data such as; missing initial BP or starting
drug regimen will be left out, as these gaps would hinder
the correctness of the survival analysis. In addition,
patients whose medical records are lost, inaccessible, or
have fewer than the minimum required BP measurements
will be excluded in order to assure the reliability and
completeness of the longitudinal data necessary for this
study.

Outcome variable of this study will time to control of

hypertension, measured in completed months from the
initiation of antihypertensive agent until the patient's BP
control according to national clinical guidelines.

The potential predictor variables will be Age, Sex,

residence type, Baseline SPB and DBP, Frequency of
appointment, Comorbidity(Eg, DM and other ), Lifestyle
(smoking, alcohol use, chat use), Medication adherence,
creatinine (mg/DL), blood urea nitrogen ,Type of
antihypertensive drugs Number of medication and
Health insurance Type

• Time to Hypertension Control: The duration,

measured in completed in day or months, from the
date of initiation of the first antihypertensive
medication to the date of the first clinic visit at which
the patient's blood pressure is recorded as controlled (
<140 mmHg and DBP <90 mmHg) according to
national guidelines (5) .
• Controlled Hypertension: Defined as an average systolic blood pressure
(SBP) of less than 140 mmHg and an average diastolic blood pressure
(DBP) of less than 90 mmHg, based on at least two consecutive readings
during a single clinic visit, following the initiation of antihypertensive
therapy (56).
• Uncontrolled Hypertension: Defined as an average SBP ≥140 mmHg
and/or DBP ≥90 mmHg at a follow-up visit after treatment initiation(56).
• Monotherapy : Treatment of hypertension using a single
antihypertensive drug class (e.g., ACE inhibitor, Calcium Channel Blocker,
Diuretic alone) (50).
• Combination Therapy: Treatment of hypertension
using two or more antihypertensive drugs from
different classes, either as separate pills or a fixed-dose
combination, initiated concurrently or sequentially (50)
.
• Comorbidity: The presence of one or more additional
chronic conditions (e.g., Diabetes Mellitus, Chronic
Kidney Disease, Dyslipidemia) documented in the
patient's medical record at or before hypertension
diagnosis (39)
•
The sample size for this time-to-event (survival) study is
determined using a formal power calculation for the Cox
proportional hazards regression model. This is the
gold-standard method for studies where the primary
objective is to compare the time until an event (blood
pressure control) between groups while adjusting for
other variables and the use of parameters from the
recent Afar region study in Ethiopia(17). The calculation
is based on the work of for the Cox model. The required
number of events (E) (i.e., patients achieving BP control)
is the fundamental unit for power in survival
analysis(57). The formula is.
Where:

• = Total number of events (BP control) required.

• = The standard normal deviate for a two-sided Type I error (α). For
α=0.05, Z = 1.96.
• ) = The standard normal deviate for power (1-β). For 80% power, β=0.20,
Z = 0.84.
• = The proportion of participants in the group of primary interest (e.g., on
combination therapy).
• = The proportion of participants in the comparison group (e.g., on
monotherapy) = (1 - p₁).
• = The natural logarithm of the hazard ratio (HR) the study is powered to
detect
•
The calculation was based on a hazard ratio of 1.66 (derived from the prior
study), with an assumed event probability of 57%, a proportion of 51% in
the combination therapy group, 80% power, and a 5% significance level. A
design effect of 1.5 is applied, and the sample will incomplete by 20% to
account for records with insufficient or missing data.

There for,

=122

And next calculate the initial sample size ()

=
214Next calucaltethe final sample size by applying the
design effect (DE) and Contingency for Ineligible
Records (W) of 1.5 and 20% respectively.

= 214*1.5 = 321

Result: Rounding up, the minimum required sample size

is 402 hypertensive patients.

Since the total population of hypertensive patients who

have followed at Debre Tabor Hospital from January 1,
2021 to December 30, 2025 is 1,577, and applying a
finite population correction would reduce the sample to
355, but i have chosen to retain the larger sample of
402. This conservative approach ensures sufficient
power even if the actual hypertension control rate is
lower than the assumed 57%, and it provides stable
parameter estimation for the AFT model when examining
multiple predictors.

A systematic random sampling technique will be

employed to select a representative sample of
hypertensive patients from the 1,577 individuals in the
chronic follow-up clinic registry between January 2021
and December 2025. This method is chosen for its
efficiency and ability to ensure each patient has an equal
probability of selection while providing a practical
approach to sampling from a sequentially ordered
medical record system. The sampling interval, k, will be
determined precisely by dividing the total population
(N=1,577) by the required sample size (n); in this case,
n=402, then k would be approximately four, meaning
every fourth patient from a randomly ordered list based
on identification number will be selected. To ensure
robust and consistent data collection, a structured and
pre-tested data abstraction checklist will be used to
extract all relevant clinical and demographic variables,
with the process conducted by trained personnel under
supervision and subject to random quality audits to
maintain data integrity and reliability throughout the
study.

Data will be extracted using a structured, pre-

tested data abstraction form developed specifically for
this study. The tool designed based on the study
objectives, a comprehensive literature review, and
consultation with clinical experts in hypertension
management and epidemiology.

Survival analysis, is a statistical method focused on

analyzing the time until the occurrence of a specific
event of interest while accounting for censored data, will
serve as the principal methodological framework in this
study. It will be employed to assess the durability of
blood pressure (BP) control in hypertensive patients by
modeling the time from treatment initiation to until the
first achieved control or event of loss of control (BP
≥140/90 mmHg). Survival analysis has been extensively
applied, in hypertension research to study time-to-event
outcomes critical to understanding disease progression
and treatment effectiveness. Previous studies have
predominantly used this method to model the time from
diagnosis or treatment initiation to major adverse
cardiovascular events (e.g., stroke, myocardial
infarction, or heart failure), providing key evidence for
treatment guidelines by quantifying the long-term
benefits of blood pressure control(58).

The survival function, the probability that an individual

survives beyond t(59). Let T be a random variable
associated with the time until some specified event (the
survival time) and let 𝑡 be the value of the variable 𝑇.
This event may be death, the appearance of a tumor or in
this case the first time to achieve blood pressure and so
forth. 𝑇 is a nonnegative random variable from a
homogeneous population and f (t) be the underlying
probability density function of the survival time T. The
cumulative distribution function F (t), which represents
the probability that a subject selected at random will
have a survival time less than some stated value t, is
given

……………………………………..(1)

And represents the probability that the survival time is

less than some value 𝑡 or the probability

of observing a survival time of up some value 𝑡 time units

And using the above cumulative distribution function,

…………………………………..(2)

From equations (1) and (2) the relationship between f (t)

and S (t) can be derived as

………………………………..(3)

And

Hazard function is also known as the instantaneous

death rate or conditional failure rate in reliability, the
force of mortality in demography, the intensity function
in stochastic processes, the age-specific failure rate in
epidemiology, the inverse of the Mill’s ratio in
economics, or simply as the hazard rate. We can also
think of the hazard function, ℎ(𝑡), as the rate at which
failures occur at time 𝑡(59).

The hazard function ℎ(𝑡), is then the limiting value of this

quantity as 𝛿𝑡 → 0, so that

…………………………………….(4)

Where,

………………………………….(5)

By applying the theory of conditional probability and the

relationship in equation,

…………………………………..(6)

Where,

•: Hazard function (instantaneous risk of the event at

time )
• : Probability density function of survival time
• : Survival function ,
• : Natural logarithm of the survival function
•
The corresponding cumulative hazard function is
defined as,

and then,

And

Understand if Specifying one of the four functions f(t),

S(t), h(t) or H(t) specifies the other three

Functions. Under the parametric approach, the baseline

hazard function is defined as a parametric function and
the vector of its parameters, say Ψ, is estimated together
with the regression coefficients and the frailty
parameter(s).

The standard estimator of the survival function,

proposed by Kaplan and Meier (1958), is called the
Product-Limit estimator. This estimator incorporates
information from all the observations available, both
uncensored (event times) and censored, by considering
survival to any point in time as a series of steps defined
at the observed survival and censored times.

We assume that we have a sample of independent

observations denoted ( , 𝛿𝑖 ), 𝑖 = 1, 2, 3, … , 𝑛 of the
underlying survival time variable 𝑇 and censoring
indicator variable 𝛿 (𝛿𝑖 = 0 if censored) Assume also that
𝑟 ≤ 𝑛 of the observations are recorded death times. The
rank-ordered survival times are denoted by (1) < (2) <. .
. < 𝑡(𝑟).

Let 𝑛𝑖 =the number at risk of dying at (𝑖) and 𝑑𝑖 =the

observed number of deaths at (𝑖). Then the KM
estimator of the survivor function at time t is

…………………………………………..(7)

This estimator is a step function that changes values only

at the time of each birth interval. The Cumulative hazard
function of the KM estimator can be estimated as:

Where is KM estimator

Comparison of survivorship (survival) functions is used

in survival analysis to determine whether two or more
groups have different survival experiences over time (for
example, male vs female patients, treatment A vs
treatment B). This comparison is usually done using
statistical tests that compare the Kaplan–Meier survival
curves of the groups.

The general form of Weighted Log Rank test is given by;

………………………………………………..(8)

From the expression ,

and

is the number at risk at observed survival time t(i) in

group 1

is the number at risk at observed survival time t(i)in the

group 2

is the number of observed deaths in group 1

is the total number of individuals or risk before time t(i)

is the total number of deaths at t(i)

is the total number of deaths at time t (i)

The test statistic depends on which of the various tests is

used, but each maybe expressed in the form of a ratio of
weighted sums over the observed survival times. Q
follows a chi-square distribution with one degree of
freedom. We can also use the above test to compare k
groups. In this study, we used the log rank test, which is
a special case of Q.
Log Rank test to test, whether survival functions of two or more groups
are equal and most widely used method and it gives equal weight to all
event times.

The hypothesis;
 H: Survival functions are the same for all groups

H1: At least one survival function is different

The formula, log rank test statistic is

…………………………………………………….(9)

For this study, modeling the time from the starting of a

treatment or intervention until the successful control of
hypertension (defined as achieving sustained blood
pressure below 140/90 mmHg). The Accelerated Failure
Time (AFT) model will be chosen for its direct,
interpretable relationship between covariates and
survival time. Unlike proportional hazards models, the
AFT model assumes that the effect of a covariate is
to accelerate or decelerate the time-to-event by a
constant factor(60). This characteristic allows for an
easier interpretation of the results because the
parameters measure the effect of the correspondent
covariate on the mean survival time. The types of the
AFT model class include the exponential AFT model,
Weibull AFT model, log logistic AFT model, log-normal
AFT model, and gamma AFT model
The Weibull Accelerated Failure Time (AFT) model is a parametric
survival regression model where covariates multiplicatively accelerate or
decelerate the time-to-event. The model assumes survival times follow
a Weibull distribution, and the effect of covariates is to "accelerate" or
"decelerate" the time until an event occurs.

In the AFT framework:

• If a covariate has an acceleration factor > 1, it lengthens survival time

(protective effect).
• If a covariate has an acceleration factor < 1, it shortens survival time
(harmful effect).
•
In terms of the log–linear representation of the model, if
Ti has a Weibull distribution, then ε has a type of the
extreme value distribution known as Gumbel
distribution. This is an asymmetric distribution with
survival function , −∞ < ε < ∞

Then the survivor function of ) is

………………………………….(10)

In addition, the Hazard Function is

…………………………………(11)

The Log-Normal Accelerated Failure Time (AFT) model is

a parametric survival regression model where the
natural logarithm of survival time follows a normal
distribution. This implies that the actual survival times
follow a log-normal distribution. The baseline survival
function and hazard function are given by,

Where μ and σ are parameters, (t) is the probability

density function and (t) is the cumulative density
function. The survival function for the ith individual is

The log-logistic accelerated failure time (AFT)

model is a parametric survival analysis model that
assumes survival times follow a log-logistic distribution,
with covariates acting multiplicatively on survival time
Unlike Weibul hazard, Log-logistic distribution is not a
monotonic function of time. However, situations in which
the hazard function changes direction can arise. In cases
where one comes across to censored data, using log-
logistic distribution is mathematically more
advantageous than other distributions.
The probability density function of Log-logistic
distribution given by

, and are the scale and shape parameter respectively.

And survivor function and the hazard functions of Log-

logistic

Distributions are;

and

If the variable, in the log-linear formulation has a log-

logistic distribution with mean

zero and variance, and the survival function of is

The survival function of 𝑇𝑖 is then,

……………………………….(12)

And the hazard function of 𝑇𝑖 for the 𝑖𝑡ℎ individual is

then

………………………………..(13)

In survival analysis, parametric AFT estimation

determines model parameters (coefficients and scale
factors) by assuming survival times follow a specific
probability distribution, such as the Weibull, Lognormal,
or Gamma. Unlike the semi-parametric Cox model, which
leaves the baseline hazard unspecified, this approach
uses Maximum Likelihood Estimation (MLE) to account
for both observed events and censored data (61). By
defining the underlying distribution, it allows for the
direct estimation of how covariates accelerate or
decelerate survival time.
In this study, an Accelerated Failure Time (AFT) model
will be serve as the primary analytical method to identify
predictors of time to hypertension control, providing
directly interpretable time ratios that indicate whether
factors accelerate or delay treatment success. To ensure
the robustness of these findings, a shared frailty term
will be incorporated into the AFT framework as a
secondary sensitivity analysis. This frailty-augmented
AFT model will account for potential unmeasured
clustering; such as non-random patient assignment to
specific clinical units or consulting teams within Debre
Tabor Comprehensive Specialized Hospital, their
residence type which introducing a random effect that
captures latent within-group correlations. The frailty
term, typically modeled using a gamma or inverse
Gaussian distribution, will estimate the variance of these
unobserved cluster-level influences. Should this variance
prove statistically significant, it will indicate the
presence of unmeasured heterogeneity, and the adjusted
time ratios from the frailty model will be reported
alongside the primary AFT estimates to validate their
stability. Conversely, a negligible frailty variance will
reinforce the assumption of observation independence
and confirm the reliability of the standard AFT
results(62).

Suppose we have k observations and 𝑖 subgroups. Each

subgroup consists of 𝑛𝑖 observations and , where n is
the total sample size. The hazard rate for the kth
individual in the ith

subgroup is given by:

 i= 1,2,3…………………….r and k = 1.,2,3………..n
,………(16)

Where the baseline hazard function and β is is a vector of

fixed effect parameters to be estimate. If the
proportional hazards assumption does not hold, the
accelerated failure time frailty model which assumes

………………………………………..(17)

The shared frailty model addresses individual

heterogeneity by incorporating an unobserved random
effect, or frailty (Zi), shared among all members of a
cluster to account for within-group dependence(62).
While a univariate model is obtained when each group
contains only one subject(63). To ensure mathematical
tractability and analytical convenience, researchers
frequently utilize parametric distributions for these
frailties, with the Gamma distribution being the most
widely adopted due to its computational simplicity(64).
Other viable alternatives include the inverse Gaussian
and positive stable distributions, which also effectively
model the underlying heterogeneity within populations.

In survival analysis, the Gamma Shared Frailty

Distribution is the most widely used mathematical
framework for modeling "hidden" correlations between
individuals in a group (like members of the same family
or patients in the same hospital)(64).

The density of a gamma-distributed random variable

with parameter is;

Where,
And the gamma function, it corresponds to a Gamma
distribution Gam(μ,ϴ) with μ fixed to 1 for identifiability.
Its variance is then θ, with Laplace transform.

And the conditional survival function and hazard

Function is

Model selection for Accelerated Failure Time (AFT)

models typically involves comparing different parametric
distributions—such as Weibull, log-logistic,
lognormal, to identify the one that best fits the
observed survival data. This is will be done
using information criteria like the Akaike Information
Criterion (AIC) or Bayesian Information Criterion (BIC),
alongside graphical checks such as residual plots and
hazard shape comparisons. Covariate selection may use
stepwise procedures or penalized likelihood methods,
while overall model fit is assessed via goodness-of-fit
tests and validation of the distributional assumptions
against the data(61).
To ensure the validity and reliability of the findings, strong assessment of
model assumptions will be an integral component of this study's analysis.
The Accelerated Failure Time (AFT) and shared frailty models each rest on
specific statistical assumptions that must be verified for the results to be
valid. The approach to model adequacy checking will therefore be two-fold,
encompassing diagnostic procedures for both the primary AFT model and
the sensitivity frailty- AFT model.

The standard AFT model assumes a specific parametric distribution for the
survival times (e.g., Weibull, log-logistic, log-normal) and a linear
relationship between the log-transformed survival time and the covariates.
To verify these assumptions, the following steps will be conducted:

• Distribution Selection: The appropriateness of different parametric

distributions (Weibull, log-logistic, log-normal) will be evaluated using the
( AIC) and (BIC) . The distribution yielding the lowest AIC/BIC will be
selected for the final model.
• Residual Analysis: The adequacy of the chosen distribution and the
linearity assumption will be assessed graphically using Cox-Snell and
deviance residuals . For a well-fitted model, a plot of the cumulative
hazard function against the Cox-Snell residuals should approximate a
straight line with a slope of 1. Deviations from linearity will suggest misfit
and prompt consideration of alternative distributions or model
transformations.
• Goodness-of-Fit Test: The Kolmogorov-Smirnov test will be applied
to the standardized residuals to statistically evaluate whether they follow
the expected extreme-value distribution under the chosen AFT model.
•
The frailty-augmented AFT model introduces additional assumptions
regarding the unobserved random effect:

• Frailty Distribution: The model assumes the frailty term follows a

specific distribution, typically the gamma distribution . The
appropriateness of this assumption will be checked by comparing the
model fit (via AIC/BIC) with alternative frailty distributions (e.g., inverse
Gaussian).
• Independence of Frailty: The frailty is assumed independent of the
observed covariates. This will be assessed by examining the correlation
between the estimated frailties and key patient characteristics.
• Proportionality of Frailty Effect: The frailty term multiplicatively
affects the hazard, implying a shared effect over time. This will be
evaluated by testing for time-varying effects in the frailty component.
•
• Data Management
•
Data will be extracted from patient medical records
using a standardized, pre-tested data abstraction form.
The form will be designed and structured using Epi
Data software (version 4.6). Both Stata version 17.0
and R 4.5.1 software version will be used for the analysis
because together they give us a powerful and flexible
toolkit for handling the complex “time-to-event” statistics
this study requires. Their built-in functions and packages
will help us fit the specialized survival models that
planned, check our assumptions with clear graphs, and
ensure our results are both robust and reproducible,
giving us confidence in our final findings.
To ensure data integrity, a multi-stage quality control
process will be implemented. This includes pre-training
data abstractors using a standardized tool, pilot testing
the extraction form on a sample of records, and
performing daily spot-checks for consistency and
completeness by the principal investigator. Double data
entry will be employed for a random 10% sample to
calculate and minimize inter-rater reliability
discrepancies, and all collected data will undergo range
and consistency checks before the final analysis to
identify and manage outliers or implausible values.

Ethical approval for this study will be sought from the

Institutional Review Board (IRB) of Bahir Dar University.
As a retrospective study using anonymzed secondary
data from medical records, the requirement for
individual informed consent will be waived by the IRB.
Confidentiality will be strictly maintained by using a
coded identification number instead of patient names,
and all data will be stored on a password-protected
computer accessible only to the research team, in
compliance with national data protection guidelines.
Activities Responsible Time
body frame

Nov 14 – Dec 28 Dec 28 Dec 29 Feb Feb Apr 05 - May

2025. 2025. 2025. 20 May 02 10,
2026 2026. 2026

Proposal and Investigator

tool
development

Proposal >>
presentation

Data collection >>

and Data
cleaning

Data processing >>

and analysis

Result writing >>

Final Defense >>

Table 1: Work Plan

Personal cost
and
transportation

s.n Activities Person No Working Daily Total

involved day wage in
EBR

1 Data Investigator 5 8 500 4000

and Data
collector

Sub- total 4,000

Costs of
stationary

[Link] Item Unit Quantity Unit Total

price price

1 Notebook Number 3 25 75

2 Pen Number 5 12 60

5 Paper Pack 2 300 600

6 Internet and GB 20GB 150 3000

communication

Sub total 3735

Costs of
accessing the
data and

[Link] Activities Quantity Unit price Total

price

1 Payment for 402 Data 50 20,100

accessing Data

2 Printing 2 3birr/page 6
questionnaire

3 Copying 402 5 201

questionnaire

4 Printing 45 10 450
proposal

5 Printing Final 82 82
Thesis

Sub total 23,386

Budget Total 31121

summary

Contingency 3112
(10%)

Grand 34,2
total

Table 2: Budget Plan

1Al-Makki A, DiPette D, Whelton PK, Murad MH,

Mustafa RA, Acharya S, et al. Hypertension
pharmacological treatment in adults: a World Health
Organization guideline executive summary.
Hypertension. 2022;79(1):293–301.

2Volpe M, Gallo G, Tocci G. Is early and fast blood

pressure control important in hypertension
management? International Journal of Cardiology.
2018;254:328–32.

3WHO. [Link]
content/uploads/2025/05/Ethiopia-National-NCD-
Protocols_2021-.pdf. 2025.

4Martin-Fernandez M, Vinyoles E, Real J, Soldevila N,

Munoz M-A, Del-Val JL, et al. The prognostic value of
blood pressure control delay in newly diagnosed
hypertensive patients. Journal of Hypertension.
2019;37(2):426–31.

5Organization WH. Global report on hypertension: the

race against a silent killer: World Health Organization;
2023.

6Misganaw A, Mariam DH, Araya T. The double

mortality burden among adults in Addis Ababa, Ethiopia,
2006-2009. Preventing chronic disease. 2012;9:E84.

7Awedew AF, Berheto TM, Dheresa M, Tadesse S,

Hailmariam A, Tollera G, et al. The Burden of Non-
Communicable Diseases and Its Implications for
Sustainable Development Goals Across Regions in
Ethiopia. Ethiopian Journal of Health Development.
2023;37(2).

8FMOH. [Link]
content/uploads/2025/05/Ethiopia-National-NCD-
Protocols_2021-.pdf. 2021.

9Gobezie MY, Hassen M, Tesfaye NA, Solomon T,

Demessie MB, Fentie Wendie T, et al. Prevalence of
uncontrolled hypertension and contributing factors in
Ethiopia: a systematic review and meta-analysis.
Frontiers in cardiovascular medicine. 2024;11:1335823.

10Kebede B, Chelkeba L, Dessie B. Rate of blood

pressure control and its determinants among adult
hypertensive patients at Jimma University Medical
Center, Ethiopia: Prospective cohort study. SAGE open
medicine. 2021;9:20503121211006000.

11Hong SH, Wang J, Tak S. A Patient‐Centric Goal in

Time to Blood Pressure Control from Drug Therapy
Initiation. Clinical and Translational Science.
2013;6(1):7–12.

12Kim H-L, Lee EM, Ahn SY, Kim K-i, Kim HC, Kim JH, et
al. The 2022 focused update of the 2018 Korean
Hypertension Society Guidelines for the management of
hypertension. Clinical Hypertension. 2023;29(1):11.

13Sendek EM, Hebo SH. Modeling time-to-good control

of hypertension using Cox proportional hazard and
frailty models at Bahir-Dar Felege Hiwot Referral
Hospital. Open Access Medical Statistics. 2017:27–36.

14Bosu WK, Bosu DK. Prevalence, awareness and control

of hypertension in Ghana: A systematic review and meta-
analysis. PloS one. 2021;16(3):e0248137.

15Munyombwe T, & Mayige, M. Time to blood pressure

control and its predictors among hypertensive patients in
Kenya: A retrospective cohort study. . East African
Medical Journal,. 2021; 98*(2), :87–95.

16Huang T. The effect of biological variables and

autonomic measures on acute blood pressure responses
to slow breathing: McGill University (Canada); 2021.
17Gashe AD, Worku FA, Mahamud KA, Woldemichael
DZ, Eshete T, Endries AY. Time to achieve blood
pressure control and its predictors among hypertensive
patients treated at public hospitals in Afar region, North-
Eastern Ethiopia: A retrospective cohort study. 2024.

18Legese N, Tadiwos Y. Epidemiology of Hypertension in

Ethiopia: A Systematic Review. Integr Blood Press
Control. 2020;13:135–43.

19Kaptoge S, Pennells L, De Bacquer D, Cooney MT,

Kavousi M, Stevens G, et al. World Health Organization
cardiovascular disease risk charts: revised models to
estimate risk in 21 global regions. The Lancet global
health. 2019;7(10):e1332–e45.

20Fuchs FD, Whelton PK. High blood pressure and

cardiovascular disease. Hypertension. 2020;75(2):285–
92.

21Kjeldsen SE. Hypertension and cardiovascular risk:

General aspects. Pharmacological research.
2018;129:95–9.

22Khatib R, McKee M, Shannon H, Chow C, Rangarajan

S, Teo K, et al. Availability and affordability of
cardiovascular disease medicines and their effect on use
in high-income, middle-income, and low-income
countries: an analysis of the PURE study data. The
Lancet. 2016;387(10013):61–9.

23Sani RN, Connelly PJ, Toft M, Rowa-Dewar N, Delles

C, Gasevic D, et al. Rural-urban difference in the
prevalence of hypertension in West Africa: a systematic
review and meta-analysis. Journal of human
hypertension. 2024;38(4):352–64.

24Olowoyo P, Okekunle AP, Asowata OJ, Atolani S,

Morsy MI, Caiazzo E, et al. Prevalence of hypertension in
Africa in the last two decades: systematic review and
meta-analysis. Cardiovascular Research.
2025;121(12):1815–29.

25Adeloye D, Basquill C. Estimating the prevalence and

awareness rates of hypertension in Africa: a systematic
analysis. PloS one. 2014;9(8):e104300.

26Shakil SS, Ojji D, Longenecker CT, Roth GA. Early

stage and established Hypertension in Sub-saharan
Africa: results from Population Health surveys in 17
countries, 2010–2017. Circulation: Cardiovascular
Quality and Outcomes. 2022;15(12):e009046.

27Ataklte F, Erqou S, Kaptoge S, Taye B, Echouffo-

Tcheugui JB, Kengne AP. Burden of undiagnosed
hypertension in sub-saharan Africa: a systematic review
and meta-analysis. Hypertension. 2015;65(2):291–8.

28Aminde LN, Agbor VN, Fongwen NT, Ngwasiri C,

Nkoke C, Nji M, et al. High burden of non-adherence to
blood pressure-lowering medications: meta-analysis of
data from over 34,000 adults with hypertension in Sub-
Saharan Africa. medRxiv. 2024:2024.05. 28.24308082.

29Aytenew TM, Kassaw A, Simegn A, Nibret Mihretie G,

Asnakew S, Tesfahun Kassie Y, et al. Uncontrolled
hypertension among hypertensive patients in Sub-
Saharan Africa: A systematic review and meta-analysis.
Plos one. 2024;19(6):e0301547.
30Tiruneh SA, Bukayaw YA, Yigizaw ST, Angaw DA.
Prevalence of hypertension and its determinants in
Ethiopia: a systematic review and meta-analysis. PloS
one. 2020;15(12):e0244642.

31Kibret KT, Mesfin YM. Prevalence of hypertension in

Ethiopia: a systematic meta-analysis. Public health
reviews. 2015;36(1):14.

32Tesfay FH, Backholer K, Zorbas C, Bowe SJ, Alston L,

Bennett CM. The magnitude of NCD risk factors in
Ethiopia: Meta-Analysis and systematic review of
evidence. International Journal of Environmental
Research and Public Health. 2022;19(9):5316.

33Geset Haile D, Sharew NT, Mekuria AD, Abebe AM,

Mezemir Y. Prevalence of hypertension and associated
factors among adults in Debre Berhan Town, North Shoa
Zone, Ethiopia, 2020. Vascular health and risk
management. 2021:203–10.

34Belachew A, Tewabe T, Miskir Y, Melese E, Wubet E,

Alemu S, et al. Prevalence and associated factors of
hypertension among adult patients in Felege-Hiwot
Comprehensive Referral Hospitals, northwest, Ethiopia:
a cross-sectional study. BMC research notes.
2018;11(1):876.

35Endrias EE, Tesfaye Mamito T, Geta Hardido T, Ataro

BA. Prevalence of Pre-Hypertension/Hypertension and its
Associated factors among adults in the Wolaita Zone of
Southern Ethiopia: a cross-sectional study. INQUIRY:
The Journal of Health Care Organization, Provision, and
Financing. 2024;61:00469580241246968.
36Mohammad R, Bansod DW. Hypertension in India: a
gender-based study of prevalence and associated risk
factors. BMC public health. 2024;24(1):2681.

37Ukpeh IE, Akpan MI, Ugbe UM, Enyievi PB,

Osuchukwu NC, Bassey EE, et al. Predictors of blood
pressure control among patients with hypertension: a
cross-sectional study in a Nigerian tertiary health
facility. BMJ Open. 2025;15(3):e091010.

38Barega B, Seifu L, Melkie A, Abebe S, Taye M. Blood

pressure control among adults with hypertension at a
tertiary hospital in Ethiopia. Ethiopian Journal of Health
Sciences. 2023;33(4).

39Chowdhury SR, Das DC, Sunna TC, Beyene J, Hossain

A. Global and regional prevalence of multimorbidity in
the adult population in community settings: a systematic
review and meta-analysis. EClinicalMedicine. 2023;57.

40Kibirige D, Olum R, Turyamureeba W, Morgan B,

Kyazze AP, Nsubuga Y, et al. Prevalence and predictors
of microvascular and macrovascular diabetes
complications in adult Ugandans: a systematic review
and meta-analysis. medRxiv. 2024:2024.10.
15.24315542.

41Animut Y, Assefa AT, Lemma DG. Blood pressure

control status and associated factors among adult
hypertensive patients on outpatient follow-up at
University of Gondar Referral Hospital, northwest
Ethiopia: a retrospective follow-up study. Integrated
blood pressure control. 2018:37–46.

42Burnier M. Drug adherence in hypertension.

Pharmacological research. 2017;125:142–9.
43Abeasi DA, Abugri D, Akumiah PO. Predictors of
medication adherence among adults with hypertension in
Ghana. 32–23:(1)8;2022 .‫ﻧﺸﺮﯾﻪ ﻣﺮﺍﻗﺒﺖ ﭘﺮﺳﺘﺎﺭﯼ ﻣﺪﺩﺟﻮ ﻣﺤﻮﺭ‬.

44Hasen G, Abdo A. Medication adherence and

associated factors in management of hypertension in
Shashemene referral Hospital, Ethiopia. Journal of
Hypertension and Management. 2020;3.

45WHO Guidelines Approved by the Guidelines Review

Committee. Guideline for the pharmacological treatment
of hypertension in adults. Geneva: World Health
Organization

World Health Organization 2021.; 2021.

46Appel LJ, Moore TJ, Obarzanek E, Vollmer WM,

Svetkey LP, Sacks FM, et al. A clinical trial of the effects
of dietary patterns on blood pressure. New England
journal of medicine. 1997;336(16):1117–24.

47He FJ, Tan M, Ma Y, MacGregor GA. Salt reduction to

prevent hypertension and cardiovascular disease: JACC
state-of-the-art review. Journal of the American College
of Cardiology. 2020;75(6):632–47.

48Roerecke M, Kaczorowski J, Tobe SW, Gmel G, Hasan

OS, Rehm J. The effect of a reduction in alcohol
consumption on blood pressure: a systematic review and
meta-analysis. The Lancet Public Health.
2017;2(2):e108–e20.

49Oyemomi IA, Anthony AT, Muftau SO, Emmanuel OO,

Olalekan IW, Olamide D-DO. Blood pressure control and
its associated factors among hypertensive patients in
federal teaching hospital, ido-ekiti, Southwest Nigeria.
Archives of Community Medicine and Public Health.
2020:167–73.

50Fuchs FD, Fuchs SC, Berwanger O, Whelton PK.

Clinical Trials in Hypertension: A Mathematical
Endorsement for Diagnosis and Treatment.
Hypertension. 2025;82(3):411–8.

51Wang X, Cheng Z. Cross-sectional studies: strengths,

weaknesses, and recommendations. Chest.
2020;158(1):S65–S71.

52Tiruneh GT, Demissie M, Worku A, Berhane Y.

Predictors of maternal and newborn health service
utilization across the continuum of care in Ethiopia: A
multilevel analysis. PloS one. 2022;17(2):e0264612.

53Andersen RM. Revisiting the behavioral model and

access to medical care: does it matter? Journal of health
and social behavior. 1995:1–10.

54Csa E. Population projection of Ethiopia for all regions

at wereda level from 2014–2017. Central Statistical
Agency of Ethiopia. 2013;1:167–76.

55Population size by woreda Tawn [Internet]. 2022.

Available from: [Link]
size-by-sex-zone-and-wereda-july-2022/.

56Unger T, Borghi C, Charchar F, Khan NA, Poulter NR,

Prabhakaran D, et al. 2020 International Society of
Hypertension global hypertension practice guidelines.
Hypertension. 2020;75(6):1334–57.
57Schoenfeld DA. Sample-size formula for the
proportional-hazards regression model. Biometrics.
1983:499–503.

58Whelton PK, Carey RM, Mancia G, Kreutz R, Bundy

JD, Williams B. Harmonization of the American College
of Cardiology/American Heart Association and European
Society of Cardiology/European Society of Hypertension
blood pressure/hypertension guidelines: comparisons,
reflections, and recommendations. European heart
journal. 2022;43(35):3302–11.

59Rodrıguez G. Parametric survival models. Princeton

University, Rapport technique, Princeton. 2010.

60Emmert-Streib F DM. Introduction to survival analysis

in practice. Machine Learning and Knowledge
Extraction. 2019;Sep 8;1(3):.1013–38.

61Collett D. Modelling survival data in medical research:

Chapman and Hall/CRC; 2023.

62Hougaard P. Analysis of multivariate survival data:

Springer; 2000.

63Wienke A. Frailty models in survival analysis:

Chapman and Hall/CRC; 2010.

64Vaupel JW, Manton KG, Stallard E. The impact of

heterogeneity in individual frailty on the dynamics of
mortality. Demography. 1979;16(3):439–54.

APPENDIX

Information extraction form/questionnaire

Title: time to control hypertension and its associated
factors hypertensive patients who follow up at Debre
Tabour Referal Hospital, Northern Ethiopia

Date of data
collection________________________________________________
___________

Data
collector___________________________Signature____________
____________________

Supervisor______________________________Signature_______
________________________
1. Skip
demographic
Characteristics

1 Patient ID
(MRN)

2 Sex Male
Female

3 Age in years _______________________

4 Residency Urban
Rural

5 Type of Health Community-Based Health Insurance

Insurance (CBHI)
Social Health Insurance (SHI)
Private Insurance
Out-of-Pocket (No Insurance)
Other (Specify): _________________

2. Clinical
factors

6 Date of __________________________________
hypertension
 diagnosis (

7 Date of _____________________________________
antihypertensive
drug initiation (

8 Follow-up BP controlled If censored,

outcome Censored (uncontrolled, TO, death) skip to Q10

9 If controlled Date:_________
(BP) SBP:__________
DBP:____________

10 If censored Date: ____________________________

(TO, Death, LTF
and study end)

11 Base line blood Mild HTN BP:_____________

pressure Severe HTN BP:_____________
(hypertension
stage)

12 Does patient has Yes If no, skip to

the comorbid No Q14
disease/

13 If yes to DM
question Asthma
number 12, Other chronic respiratory diseases
state the Malignancies
comorbid HIV/AIDS
disease/s Liver disease
Others (specify)________________

14 Does patient has Yes If no, skip to

hypertensive No Q16
complication/

15 If yes to Heart diseases

question stroke
number 14, renal disease
state Ocular disease
complication /s Others (specify)_______________

3. Drug and
therapy
related factors

16 Drug Therapy Monotherapy If

Combination therapy Combination
therapy
skip to Q18

17 If CCBs
Diuretics______________
ACE inhibitors ___________
Beta-blockers__________
ARB_________________

18 If combination CCBs with diuretics

therapy CCBs with ARB-blocker:__________
ACE inhibitor with
diuretics:________
CCBs with ACE Inhibitor:_________
ARBs with diuretics:_____________
Others (specify):__________________

19 Treatment A. yes
intensification B. No

20 Medication High Adherence (>80% of doses

Adherence taken)
(Based on Medium Adherence (50-80%)
clinician Low Adherence (<50%)
notes/refill Not documented
records)

21 Prescription Once per day

pattern Twice per day
Three times per day

22 Other Yes
concomitant No
therapy

23 Frequency of Monthly
Scheduled Every two month
Follow-up Every three month
Appointments: Other (Specify)_________________
23 Statin Therapy Yes
No

Behavioral
characteristics

24 Smoking Status Yes

No

25 Alcohol use Yes

No

26 Chat Intake High

Moderate
Low/Restricted
Other

Laboratory

27 Serum ___________________________________

28 Fasting ___________________________________

29 Blood Urea __________________________________

Nitrogen (BUN)
(mg/