Chapter One Introduction

1.1 Introduction.
Universal Health Coverage (UHC) ensures that everyone can access the necessary
health services without facing financial difficulties. This comprehensive system
encompasses essential health services, spanning from preventive measures to palliative
care (Organization, Tracking universal health coverage: First Global Monitoring
Report., 2015)

Target 3.8 in Goal 3 of the Sustainable Development Goals (SDGs) is about making
sure everyone gets good healthcare. This includes protecting people from money
problems when they need health services and making sure they can access high-quality
healthcare, medicines, and vaccines that are safe, effective, and affordable. This target
is really important because it helps not only with Goal 3 but also with health-related
goals in other Sustainable Development Goals (SDGs). (WB report (Nations, Sustain.
Dev. Goals 17 Goals to Transform ourWorld, 2015)

The countries in the United Nations (UN) made a promise to achieve Universal Health
Coverage (UHC) by the year 2030. This commitment is part of the Sustainable
Development Goal (SDG) number 3, which they recently agreed upon (Nations,
Transforming our world: The 2030 Agenda for Sustainable Development, 2015).This
promise has two important goals. First, it wants to make sure at least 80% of necessary
health services are available to everyone, no matter how much money they have.
Second, it aims to protect people from paying for healthcare expenses out of their own
pockets (OOP) by reaching 100% financial risk protection. Many countries have made
Universal Health Coverage (UHC) a top priority because they believe it's crucial for
their national health systems. They want to reduce poverty and make sure everyone has
equal access to healthcare, improving health for everyone (Nations, Sustainable
Development Goals [Internet]., 2015).

Many countries have chosen Universal Health Coverage (UHC) because it's an
important way to make sure healthcare is affordable and fair for everyone. It helps
create a strong health system and is a big step towards achieving better health for all
(Organization, Tracking universal health coverage: First Global Monitoring Report,
2015).

1|Page
Chapter One Introduction

The World Health Organization (WHO) and World Bank (WB) worked together to
create a plan for checking if Universal Health Coverage (UHC) is working well. They
look at three things: how many people are covered, what health services are available,
and how much of the money for health comes from a system where everyone helps
each other out. (World Health Organization, 2014)

Bangladesh's health system is facing two big problems: there are many diseases, not
enough services, and the money protection for health is not working well. The
government heavily subsidizes public health services, and primary care at health clinics
is provided at nearly no cost. Each patient is charged a small fee of Bangladesh Taka 13
(equivalent to US$ 0.17 in 2011) for each outpatient visit (Watch, 2012).

Bangladesh has a diverse healthcare system that lacks strong regulation, primarily
involving four main players: the government, the for-profit private sector, the not-for-
profit private sector (mostly non-governmental organizations), and international
development organizations. (Ahmed, 2015).In countries with fragile health systems,
such as Bangladesh, inequality is a significant concern. Disadvantaged populations
often face limitations in their financial access to healthcare services.

People in Bangladesh often have to pay a significant portion of their healthcare costs
from their own pockets. Around 63% of the total health costs are covered directly by
households. Unfortunately, this high out-of-pocket spending puts families at risk of
financial difficulties and impoverishment (MOHFW 2015) (Organization., 2011).

The higher burden of financial hardship found health financing in Bangladesh relies
heavily on OOP payments for both public and private health services. Health financing
is inadequately funded, with only 2.63 percent of the gross domestic product (GDP)
allocated to health, marking the lowest expenditure in the South Asia region (Current
health expenditure (% of GDP) - Bangladesh, Nepal, Bhutan, India, Maldives, Sri
Lanka, Pakistan, Afghanistan, 2020).

From 1997 to 2020, the overall pattern of health expenditure as a percentage of GDP
showed a consistent increase, averaging around 3%. In contrast, Over the period from
1997 to 2020, the proportion of government spending in relation to the GDP has
consistently decreased, consistently staying below one percent (<1%). This decline
became more pronounced in 2020, plummeting to 0.66%, according to BNHA (1997-
2020).

2|Page
Chapter One Introduction

Bangladesh's initial 20-year healthcare financing strategy, formulated in 2012, aimed to

reduce the proportion of out-of-pocket payments in total health expenditure by half.
The vision also included the implementation of a social health protection scheme by
2032, achieved through increased tax revenue and mandatory contributions to social
health. The government of Bangladesh has set a target to raise the budget allocation for
health to 15% by 2032, a substantial increase from the current level of approximately
5%. Achieving this goal may pose significant challenges.

This incentivizes individuals to turn to private sector healthcare, which tends to be

more costly. Almost all health expenditure in private health facilities comes directly
from out-of-pocket payments (93%) (World Development Indicators:, 2014).

Health financial coverage is limited, with nine percent of households experiencing

catastrophic health payments, and seven percent resorting to distress financing
(borrowing or selling household assets to cover healthcare costs) (Islam, 2017).

Two recent studies have examined the effects of out-of-pocket (OOP) payments for
healthcare on economic impoverishment and poverty in Bangladesh (Rahman et al.,
2013; Hamid et al., 2014).

Rahman et al. (2013) focused on a metropolitan city (Rajshahi) in Bangladesh,

exploring the factors influencing high healthcare expenditure and financial catastrophes
related to healthcare. (Rahman MM, 2013) (Hamid SA, 2014). It was discovered that
households, on average, allocated 11% of their budget to health expenses. Half of the
residents dedicated 7% of their monthly per capita consumption expenditure to cover
the costs of one illness, and 9% of all households experienced financial catastrophe.
Rahman et al. (2013) conducted their investigation in a metropolitan city, concentrating
on catastrophes linked to out-of-pocket (OOP) spending. As a result, this study
overlooked catastrophes in rural areas and the impact on poverty in both urban and
rural areas separately. (Rahman MM, 2013).

Hamid et al. (2014) focused on the impact of Catastrophic Health Expenditure (CHE)
in the context of low-income populations. Key findings include: Nationwide, 14.2% of
households experienced Catastrophic Health Expenditure (CHE), Catastrophic Health
Expenditure (CHE) was more common in rural areas (16.3%) than in urban areas
(8.6%), especially affecting households with lower socioeconomic status, Annually,
3.5% of the total population fell into poverty due to out-of-pocket (OOP) spending on

3|Page
Chapter One Introduction

healthcare, leading to economic impoverishment for 5 million people in Bangladesh.

(Hamid SA, 2014). Private healthcare providers, working alongside government
medical services, primarily concentrate on curative services, which include general
practitioner clinics, medical centers, and private hospitals. Currently, Bangladesh does
not have a national health insurance scheme, and the private insurance market is not
well-established. (Organization., 2011). Several small-scale community insurance
schemes operated by NGOs, often in collaboration with micro-financing programs,
exist. However, these schemes cover less than 1% of the total population and primarily
focus on serving impoverished communities (Organization., 2011).

The government introduced a health insurance pilot project, "Shasthyo Shurokhsha

Karmasuchi (SSK)," providing free treatment for 50 diseases to households below the
poverty line. While it aims to prevent financial hardship, the current plan may not fully
protect low-income families. Addressing rising healthcare costs requires a wide
national health insurance scheme, particularly prioritizing the needs of the poor.

Implementing a nationwide health scheme with a more comprehensive benefit package,

similar to the Universal Coverage Scheme of Thailand or Seguro Popular of Mexico,
has the potential to protect the entire population from the heavy burden of out-of-
pocket payments. [38, 39].

By implementing risk pooling mechanisms, many countries with varying income

levels—low, middle, and high—have effectively decreased user fees at the point of
care. This strategy has proven successful in alleviating the economic risks associated
with out-of-pocket payments for families. [33, 40].

The Bangladeshi Constitution makes a commitment to tackle inequalities in accessing

healthcare in rural areas. Furthermore, the country has joined the global community in
pledging to achieve Universal Health Coverage (UHC) by 2030 as part of the
Sustainable Development Goals (SDGs)[5(13)].

It is important to study the impact of OOP healthcare spending on economic condition

reflecting the rural scenarios in Rangpur Division The authors of this study explored
how OOP payments influence Catastrophic Health Expenditure (CHE) and distress
financing in rural areas of Rangpur Division. The study aims to evaluate the occurrence
of catastrophic health expenditure and distress financing linked to out-of-pocket (OOP)
payments.

4|Page
Chapter One Introduction

1.2 Health care system in Bangladesh

The Ministry of Health and Family Welfare (MOHFW) and the Ministry of Local
Government, Rural Development, and Cooperatives (MOLGRDC) hold the primary
responsibility for public health services in Bangladesh.

Public sector: In Bangladesh, the public sector plays a vital role in providing
healthcare services to citizens, especially in underserved areas. To achieve this, the
public sector has established primary healthcare facilities such as Union Health and
Family Welfare Centers (UHFWCs), Community Clinics (CCs), and Upazila Health
Complexes (UHCs) that offer preventive, promotive, and curative care. The public
sector's focus is on maternal and child health, immunization, and infectious disease
control, leading to a decrease in mortality rates and an improvement in public health. In
addition, the public sector creates healthcare policies, regulations, and guidelines that
ensure quality standards and patient safety. It invests in healthcare professionals'
development by providing training and capacity-building programs. Moreover, the
public sector provides timely healthcare support during natural disasters and
humanitarian crises, playing a critical role in emergency response and disaster
management. Overall, the public sector's efforts are essential in advancing healthcare
accessibility, equity, and quality across Bangladesh

Private sector: It’s facilities such as hospitals, clinics, and diagnostic centers play a
significant role in providing healthcare services to the population. They offer
specialized care, advanced medical treatments, and surgical procedures, often with a
focus on efficiency and innovation. Private healthcare providers invest in modern
equipment and technologies, enhancing the quality and range of services available to
patients. Patient choice between public and private healthcare providers encourages
both sectors to strive for excellence. Collaborations between the public and private
sectors are also emerging to address healthcare challenges more effectively. Overall,
the private sector contributes to improved health outcomes and the overall well-being
of the population.

In Bangladesh, the healthcare system is divided into three levels: primary, secondary,
and tertiary. The primary level is mainly focused on providing basic healthcare services
and preventive measures to the general population. Secondary healthcare offers
specialized treatments and diagnostic services to patients with specific health

5|Page
Chapter One Introduction

conditions. Finally, tertiary healthcare facilities provide advanced medical services and
treatments for complex diseases and conditions. These facilities typically consist of
medical colleges, teaching hospitals, and specialized centers.

Figure 1.2.1: Health care levels in Bangladesh. (source: Internet)

Primary Healthcare: Primary healthcare which is the initial level of healthcare

services and usually the first-place people go for medical help, is delivered through
community clinics, union health centers, and rural health centers. These locations
provide fundamental services like vaccinations, maternal and child health support, basic
medical consultations, and preventive care.

Table 1.2.1: Healthcare services at primary level Facilities.

Facilities Services
Family planning, immunization, control of
Community Clinics (CCs) communicable diseases, treatment of common
ailments, and referral services are provided.

Outpatient services include family planning,

Health & Family Welfare Centre communicable disease control, clinical care, normal
delivery, and adolescent health care.

Outpatient services, as well as inpatient services,

Upazila Health Complexes including diagnostic and operative treatments, are
(UHCs) offered.

Source: Ministry of Health and Family Welfare (MOHFW)

6|Page
Chapter One Introduction

Secondary Healthcare: Secondary-level healthcare refers to medical services that are

more specialized than primary care but less specialized than those offered at tertiary
care facilities. These services typically include diagnosis, treatment, and management
of a wide range of health conditions and diseases that require more advanced medical
expertise and resources. Secondary

level healthcare facilities often include district hospitals, specialized clinics, and
regional medical centers.

Table:1.2.2 Healthcare services at secondary level

Facilities Services

Outpatient and inpatient services are

District Hospitals (DHs) available, along with laboratory,
radiographic, and ambulance services.

Outpatient, inpatient, and emergency

General Hospitals
services are provided.

Family planning, preconception, prenatal,

and postnatal care, along with outpatient
Maternal and Child Welfare Centers
and inpatient services, are offered,
(MCWCs)
including child and adolescent mental
health services.

Source: Ministry of Health and Family Welfare (MOHFW)

Tertiary-level: healthcare facilities have expanded in Dhaka and various divisional

cities nationwide. Dhaka hosts numerous specialized medical institutes, while Medical
College Hospitals and General Hospitals are typically located in divisional and major
district headquarters. Bangladesh boasts thirty-six medical colleges under the
Directorate General of Health Services (DGHS), along with six Armed Forces and
Army medical colleges, and nine dental colleges with dental units within medical
colleges. Additionally, Dhaka houses several super-specialized hospitals focusing on
teaching and research. Tertiary-level health facilities offer a wide range of medical
services, including outpatient, inpatient, and emergency obstetric care (EMOC).

7|Page
Chapter One Introduction

Chronic diseases: It represent a substantial health concern in Bangladesh, adding to

the nation's disease burden and healthcare complexities. Common chronic conditions
include cardiovascular diseases, diabetes, chronic respiratory diseases, cancer, and
mental health disorders. Factors such as rapid urbanization, lifestyle changes, unhealthy
dietary habits, tobacco use, and air pollution contribute to the increasing prevalence of
chronic illnesses in Bangladesh. Additionally, limited access to healthcare services,
inadequate preventive measures, and socioeconomic disparities exacerbates the burden
of chronic diseases, particularly among vulnerable populations. Addressing chronic
illnesses requires a comprehensive approach, including health promotion, disease
prevention, early detection, access to affordable treatment, and ongoing management
and support for patients. Collaborative efforts involving government agencies,
healthcare providers, NGOs, and community organizations are essential to effectively
address the challenges posed by chronic illnesses and improve the health outcomes of
the population in Bangladesh.

Acute diseases: The people of Bangladesh are facing several acute diseases due to
environmental conditions, sanitation, and population density. Respiratory infections
like the common cold, influenza, and pneumonia are widespread, while gastrointestinal
illnesses such as diarrhea and food poisoning are also common. Mosquito-borne
diseases like dengue fever and malaria pose significant health risks, especially in urban
areas. Skin infections, urinary tract infections, and acute injuries resulting from
accidents or trauma are also frequently encountered. Healthcare facilities can be
overwhelmed during outbreaks. Public health initiatives focusing on sanitation, hygiene
education, vector control, and vaccination campaigns, along with access to healthcare
services, are essential to manage these illnesses effectively.

There are several means of travel to service/treatment providing personnel. The most
common modes of transportation include buses, trains, taxis, and rickshaws. However,
the quality and availability of these transportation options can vary depending on the
location and the time of day. In some areas, it may be difficult to find reliable
transportation, while in others, there may be a variety of options to choose from.

Care-seeking behavior: In Bangladesh, care seeking behavior varies depending on

several factors such as socio-economic status, education level, cultural beliefs, and
accessibility to healthcare services. People living in rural areas tend to rely on

8|Page
Chapter One Introduction

traditional healers and home remedies, while those in urban areas are more likely to
seek care from modern healthcare providers. However, there is a growing trend towards
using digital health services, especially among younger generations. Despite these
differences, common barriers to seeking care in Bangladesh include cost, lack of trust
in healthcare providers, and transportation issues.

1.3 Objectives of the study:

[Link] access healthcare facilities available to rural population of Rangpur
division.

[Link] find the nature of financial hardship in rural population.

[Link] investigate the cause of distress financing in rural Rangpur division

1.4 Organization of the study:

Chapter 1: The opening chapter contains an introduction with Bangladeshi
and worldwide literature of the study, objective of the study and layout of the
study.
Chapter 2 This chapter, we discussed the background characteristics of the
sample, sampling technique, statistical methods of finding the association
between financial hardship and the socio-demographic, health related
variables.
Chapter 3: In this chapter contains the results of basic characteristics, health
care facilities, health care cost and financial hardship and the relationship of
financial hardship with the independent variables.
Chapter 4: This chapter summarizes the association and relationship
between distress financing, catastrophic payment and health care facilities.
Chapter:5 Finally, the chapter presents the conclusion and policy
implications

9|Page
Chapter Two Methods and Methods

2.1 Materials

2.1.1 Study area

Bangladesh has eight administrative divisions: Barisal, Chattogram, Dhaka,

Mymensingh, Khulna, Rajshahi, Rangpur, and Sylhet. Rangpur is one of them and it is
the northernmost division of Bangladesh. There are 17.30 million people living in the
Rangpur division, and average households’ size is approximately four members and the
population density 1,088 per square kilometer. The literacy rate is 74.66% at the
national level and the literacy rate is 68.95% for male and female, 72.13%, 65.88% for
males and female respectively (planning., november,2023). It has 8 districts:
Nilphamari, Gaibandha, Kurigram, Lalmonirhat, Panchagarh, Rangpur and Thakurgaon

Figure 2.1.1: Districts of Rangpur Division (Source: Internet)

10 | P a g e
Chapter Two Methods and Methods

2.2 Methods

2.2.1 Sample size determination:

A primary sampling unit (PSU) in this survey is individual households of the rural
Rangpur division. For this survey, we identified financial distress for seeking health
care as the key target variable. It is found that the prevalence of financial prevalence of
theoretical distress in Bangladesh is 7%. (Islam, 2017) We, therefore, calculated the
minimum number of households needed for this survey as:

where,

P=Prevalence of financial distress 7%

d2 =Margin of error.
The formula provided that 100 household samples would be required for my study.

2.2.2 Sample selection procedures:

In this study, I used multistage cluster sampling to collect primary data on health
service coverage and financial hardship in rural areas in the Rangpur division. At first, I
selected 5 districts by simple random sampling from 8 districts. These districts are
Kurigram, Lalmonirhat, Panchagarh, Rangpur, and Thakurgaon.

Then we selected a upazila from these districts by using simple random sampling. After
that we randomly selected 5 union from the 5 upazila each. Finally, I chose one village
from each union randomly, and the entire households of those selected villages are my
sample Data on health service coverage and financial hardship has been collected from
these villages. The total number of family members is 463 (118 households).

11 | P a g e
Chapter Two Methods and Methods

2.2.3 Data collection:

I interviewed household heads using a structured questionnaire, collecting data on short

term illnesses symptoms (last 30 days), chronic illnesses, duration of suffering,
consultation locations, transportation methods, consulting from (qualified doctor or
salesman of a pharmacy) and socio-economic status.

Interviewers, with consent, collected data from household heads on socio-demographic

details and the consumption/expenditure of household members in the past 30 days or
12 months, using a structured questionnaire. The study used a 30-day recall period for
all illnesses and a minimum of 12 months' duration for chronic diseases in the year
before the interview.

A health problem was classified as chronic if it continued or was predicted to exist for
more than a year. Deta was collected on the duration of illness, treatment cost,
consulting for the treatment (qualified doctor/sales man of a pharmacy), travel to the
consulting provider, distance to the provider, diagnosis, and treatment response.
Interviewees were asked about their primary

12 | P a g e
Chapter Two Methods and Methods

2.2.4 Conceptual Framework:

Chronic illness

Hospitilized

Consulted For Illness

Financial Catastropic
Get the medicine
Hardship(Distress Payment(health cost
form
Financing) over 25%)

Pay for the medicine.

Travel to the provider

Old family members

Figure 2.2.1: Conceptual framework

13 | P a g e
Chapter Two Methods and Methods

2.5 Variables

2.5.1 Outcome variables

There were two outcome variables in this study,

(a)Distress financing, Distress financing is about getting money in difficult times,

especially for sudden expenses like medical treatments. This can include borrowing,
selling items, or making urgent financial arrangements. In healthcare, it's used when
people have significant medical costs that exceed their regular income or savings. It
was measured by household head’s self-reported. Our data collector asked question to
the household head’s:

Financing sources for medical treatment?

(a)Regular income, (b) Household saving, (c) Sold Agricultural product/Tree (d)
Mortgage of Assets/Land Sold livestock/ permanent assets/belonging, (e) Borrowed
from Friends/ Relatives/Office

Among them Mortgage of Assets/Land, sold livestock, permanent assets/belonging,

borrowed from Friends/ Relatives/Office are considered as distress financing. And
Regular income, Household saving, Sold Agricultural product/Tree are conceded as
normal financing. The normal financing is coded as 0, and distress financing is coded
as 1.

(b) Catastrophic payment, In Universal Health Coverage (UHC), catastrophic

payments are high healthcare costs that can seriously affect individuals or households
financially. Various studies use different methods to define these expenses as
catastrophic. Some studies consider out-of-pocket health expenditures as catastrophic if
they surpass a specified percentage (e.g., 10% or 25%) of income or consumption. In
our study, we define out-of-pocket health expenditures as catastrophic when they
exceed a specific percentage (more than 25%) of consumption.

Interviewers recorded information on household member’s household consumption or

expenditure (sector wise: food, education, treatment, medicine, household goods,
others) in the past 30 days or past 12 months using a structured questionnaire from
household heads after obtaining informed consent. People were questioned about their
primary symptoms and eventual diagnoses, including whether professional medical

14 | P a g e
Chapter Two Methods and Methods

doctors, specifically MBBS doctors, made the diagnosis. Data were collected on the
onset or duration of illness, diagnosis, treatment response. When the overall cost of
treatment (including both treatment and medication) goes beyond 25, it is termed a
catastrophic health payment. Total treatment cost (more than 25% of expenditure) is
considered as catastrophic health payment and coded as (1) otherwise (0)

Table 2.5.1: List of selected dependent variables with groups and

codes.

Variables Group Code No

Normal financing 0
Distress Financing
Distress financing 1
No 0
Catastrophic payment
Yes 1

2.5.2 Independent variables

There were some socio-economic, demographic and health related factors which were
assumed to be associated factors of our outcome variables.

Socio-economic and demographic factors: Household head educational level, Mother's

occupation of household members, Family income (monthly BDT), Number of old
(aged over 65), Number of child (aged under 5), House hold expenditure.

Health factors: Presence of chronic illness, Duration of chronic illness, treatment cost,
consulting for the treatment (qualified doctor/sales man of a pharmacy), travel to the
consulting provider, distance to the provider, Get the medicine form, Pay for the
medicine, Short-term illness symptoms (last 30 days), socio-economic status.

15 | P a g e
Chapter Two Methods and Methods

Table 2.5.2: List of selected independent variables with groups and

codes.

Variables Group Code No

No 0
Chronic illness exists?
Yes 1

Inpatient 0
Inpatient or Outpatient?
Outpatient 1

Qualified Doctor 0
Consulted for illness?
Salesman of a Pharmacy 1

Health cost more than 25 No 0

percent? Yes 1

Govt health center 0

Get the medicine form?
Pharmacy/Dispensary 1

Did you pay for the No 0

medicines? Yes 1

Walking 1

Ambulance 2
How did you travel to the
Bus 3
provider?
Auto Rickshaw 4

Motorcycle 5

No 0
Old members?
Yes 1

16 | P a g e
Chapter Two Methods and Methods

2.6 Statistical analysis

Frequency distribution (percentage) was used for both dependent and independent
variables to examine financial hardship (distress financing, chaotropic payment) among
the rural population in Rangpur Division.

We computed descriptive statistics for quantitative variables. In this study, we

employed the χ2-test to determine the association between dependent and independent
variables. Significantly associated factors, identified through the χ2-test, were utilized
as independent variables in logistic models.

Statistical analysis was performed by using SPSS software (IBM version 20). The
binary logistic regression model considered financial hardship (distress financing,
chaotropic payment) as the outcome variable, with various socio-economic,
demographic, and health-related characteristics as independent variables.

The dependent variables, categorized into two groups, were analyzed using a binary
logistic regression model. This aimed to assess the impact of socio-economic,
demographic, and health-related characteristics on financial hardship (distress
financing, chaotropic payment) within the rural population in Rangpur Division.

The fitness of the multiple binary logistic regression model was tested through the
Hosmer and Lemeshow test and Nagelkerke R². Model accuracy was evaluated using
sensitivity, specificity, positive predictive values, and negative predictive values.

2.6.1. Frequency distribution

A frequency distribution shows us a summarized grouping of data divided into

mutually exclusive classes and the number of occurrences in a class. It is a way of
showing unorganized data notably to show results of different socio-economic and
demographic, anthropometric factors of a study. Some of the graphs that can be used
with frequency distributions are histograms, line charts, bar charts and pie charts.
Frequency distributions are used for both qualitative and quantitative data for
constructing frequency distribution:

 Decide the number of classes. Too many classes or too few classes might not reveal
the basic shape of the data set also it will be difficult to interpret such frequency

17 | P a g e
Chapter Two Methods and Methods

distribution The ideal number of classes may be determined or estimated by

formula:

Number of classes=C=1+3.3 log n (log base 10) or by the square-root choice formula
C= √n where n is the total number of observations in the data

 Calculate the range of the data (Range=Max-Min) by finding the minimum and
maximum data values. Range will be used to determine the class interval or class
width.
 Decide the width of the classes, denoted by hand obtained by, h=

 generally, the class interval or class width is the same for all classes. The classes all
taken together must cover at least the distance from the lowest value (minimum) in
the data to the highest (maximum) value. Equal class intervals are preferred in
frequency distribution, while unequal class intervals (for example logarithmic
intervals) may be necessary in certain situations to produce a good spread of
observations between the classes and avoid a large number of empty or almost
empty classes.
 Decide the individual class limits and select a suitable starting point of the first
class which is arbitrary, it may be less than or equal to the minimum value. Usually,
it is started before the minimum value in such a way that the midpoint (the average
of lower- and upper-class limits of the first class) is properly placed.
 Take an observation and mark a vertical bar (|) for a class it belongs. A running
tally is kept till the last observation.
 Find the frequencies, relative frequency, cumulative frequency etc. as required.

The following are some commonly used methods of depicting frequency:

Bar graph

A bar chart is like a picture made of rectangles. The length of each rectangle shows
how much of something there is. The longer the rectangle, the more there is of that
thing. You can make these rectangles stand up or lie down. If they stand up, it's called a
vertical bar chart. Some people also call it a column bar chart.

18 | P a g e
Chapter Two Methods and Methods

Frequency distribution table

A frequency distribution table is like a list that shows how many times different things
happen in a group. It helps organize and summarize the information about how often
each value appears. So, it's like making a neat list to understand which values occur
more or less in a sample.

Joint frequency distributions

A joint frequency distribution table is like a special list that shows how often two things
happen together. It helps us see how different pairs of things occur. Each entry in the
table tells us how many times both of these things happen at the same time. It's a way
of organizing information to understand the relationships between two variables.

2.6.2 Descriptive statistics

Descriptive statistics involves using summary statistics to quantitatively describe

features in a set of data. It is the process of analyzing and utilizing these statistics. This
type of statistic is distinct from inferential statistics, which aims to learn about a
population based on a sample. Descriptive statistics focus on summarizing the sample
itself. Chapter two delves into theoretical approaches, emphasizing that descriptive
statistics offer straightforward summaries of both the sample and the observations
within it. These summaries can take the form of quantitative summary statistics or easy-
to-understand visual graphs. Whether these summaries serve as the initial description in
a broader statistical analysis or prove sufficient for a specific investigation, descriptive
statistics play a crucial role in simplifying and conveying information about the data at
hand. [Wikipedia]

2.6.3 Chi-square test

A chi-square test, often denoted as chi-square (χ²) test, is a statistical method used to
test hypotheses when the distribution of the test statistic follows a chi-squared
distribution under the assumption of a true null hypothesis. When not otherwise
specified, the term 'chi-squared test' is commonly shorthand for Pearson's chi-squared
test. This test evaluates if there's a significant difference between the expected and
observed frequencies within one or more categories (Pearson, 1900).

19 | P a g e
Chapter Two Methods and Methods

In practical terms, a chi-squared test helps in deciding whether to reject the null
hypothesis, indicating that the data may not be independent. When applied to
qualitative data, the test statistic is based on counts representing items in each category.
It measures the agreement between actual and expected counts, assuming
independence.

An important note is that the chi-squared distribution approximation becomes more

accurate as the sample size increases. This asymptotic property means that, under the
assumption of a true null hypothesis, the sampling distribution can closely approximate
a chi-squared distribution with increasing sample size.

Examples of situations where the chi-squared distribution is only approximately valid

in chi-square tests include:
➤ Pearson's chi-squared test determines if paired data on two variables, as shown in a
contingency table, show independence. It is sometimes referred to as the chi-squared
goodness-of-fit test or the chi-squared test for independence. One example of its use
would be to examine survey data from people of different nationalities to see if
responses are influenced by nationality (Cochran and William, 1952).

➤ A modification called Yates's adjustment for continuity, or Yates' chi-squared test, is

applied to improve the precision of the chi-squared test findings.

➤ The precision of the chi-squared test results is increased by utilising a modification

known as Yates's correction for continuity, or Yates' chi-squared test.

➤ The linear-by-linear association chi-squared test analyzes the trend or pattern in

the relationship between two categorical variables.

➤ Likelihood-ratio tests are statistical tools that compare the fit of two models by
evaluating the likelihood of observed data under each. The tests help researchers
determine whether a more complex model significantly improves the explanation of the
data compared to a simpler model.

Test of independence

The Chi-Square test of Independence helps us find out if two groups of things are
linked. It checks how often one group matches with different options of the other
group. Think of it like organizing this information in a table with rows and columns. If

20 | P a g e
Chapter Two Methods and Methods

there are rows (r) and columns (c), the Chi-Square test helps us see if the two groups
are independent or if there's a connection between them. In a test of independence, we
have the null hypothesis and the alternative hypothesis:

Ho: There is no relationship between the categorical variables.

VS
H1: There is a between the categorical variables.
In this situation, an "observation" means noting down the results of two things that
happen. The null hypothesis, or the idea we're testing, suggests that these results don't
have a statistical connection—they occur independently. We organize each observation
into a specific spot in a table with rows and columns, and this table is like a grid with
different sections, each representing a combination of the two [Link] the
statistics used to test the hypothesis are,

∑∑

which is approximately distributed as χ2 with (r-1) (k-1) df.

where
χ²=Pearson's cumulative test statistic.
Oi,j= an observed frequency;
Ei,j=an expected (theoretical) frequency, asserted by the null hypothesis;
n=the number of cells in the table.
The fitting of the "independence" model leads to a reduction in the number of degrees
of freedom, specifically by the amount given by the equation: p=r+c-1.

The approximation to the chi-squared distribution becomes unreliable when expected

frequencies are excessively low. Generally, it is deemed acceptable as long as less than
20% of the events exhibit expected frequencies below 5. However, when dealing with a
single degree of freedom, the approximation lacks reliability if expected frequencies
fall below 10. In such instances, a more accurate approximation can be achieved by
applying Yates's correction for continuity. This involves reducing the absolute value of
each difference between observed and expected frequencies by 0.5 before squaring; this
is called is called Yates’s correction for continuity. In cases where the expected value,
E, is found to be small (indicating either a small underlying population probability, or a
small number of observations), the normal approximation of the multinomial

21 | P a g e
Chapter Two Methods and Methods

distribution can fail, and in such cases, it is found to be more appropriate to use the G-
test, a likelihood ratio-based test statistic. Where the total sample size is small, it is
necessary to use an appropriate exact test, typically, either the binomial test or (for
contingency tables) Fishers exact test; but note that this test assumes fixed and known
marginal totals.

In statistics, G-tests are likelihood-ratio or maximum likelihood statistical significance

tests that are increasingly being in situations where chi-squared tests were previously
recommended. The general formula for G is,

G=2 ∑[Link](Oi /Ei).

where,

Oi is the observed frequency in a cell,

Ei is the expected frequency on the null hypothesis,

and the sum is taken over is taken over all cells, and where in denotes the natural
logarithm (log to the base e) and the sum is taken over all non-empty cells.

Fisher's exact test is a statistical significance test used in the analysis of contingency
tables. Although in practice it is employed when sample sizes are small, it is valid for
all samples sizes. It is named after its inventor, R.A. Fisher, and is one of a class of
exact tests, so called because the significance of the deviation from a null hypothesis
can be calculated exactly, rather than relying on an approximation that becomes exact
in the limit as the sample size grows to infinity, as with many statistical tests.

In statistics, Yates's correction for continuity (or Yates's chi-squared test) is used in
certain situations when testing for independence in a contingency table. In some cases,
Yates's correction may adjust too far, and so its current use is limited. Using the chi-
squared distribution to interpret Pearson's chai-squared statistic requires one to assume
that the discrete probability of observed binomial frequencies in the table can be
approximated by the continuous chi-squared distribution. The assumption is not quite
correct, and introduces some error. To reduce the error in approximation, Frank Yates,
an English-statistician, suggested a correction for continuity that adjusts the formula for

22 | P a g e
Chapter Two Methods and Methods

Pearson's chi-squared test by subtracting 0.5 from the difference between each
observed value and its expected value in a 2×2 contingency table (Yates, 1934). This
reduces the chi-squared value obtained and thus increases its p-value. The effect of
Yates's correction is to prevent overestimation of statistical significance for small data.
This formula is chiefly used when at least one cell of the table has an expected count
smaller than 5. Unfortunately, Yates's

correction may tend to overcorrect. This can result in an overly conservative result that
fails to reject the null hypothesis when it should (a type II error).
So it is suggested that Yates's correction is unnecessary even with quite low sample
sizes (Sokaland Rohlf, 1981), such as:∑
The following is Yates's corrected version of Pearson's chi-squared statistic:
| |
χ²Yates=∑ ,

where
Oi =An observed frequency
Ei = An expected (theoretical) frequency, asserted by the null hypothesis.
N= Number of distinct events.

2.6.4 Logistic regression analysis approach

In statistics, logistic regression or logit regression or logit mode is a regression model

where the dependent variable (DV) is categorical. This article covers the case of binary
dependent variables-that is, where it can take only two values, such as pass/fail,
win/lose, alive/dead or healthy/sick. Cases with more than two categories are referred
to as multinomial logistic regression or if the multiple categories are ordered, as ordinal
logistic regression. Logistic regression was developed by statistician David Cox in
1958. The binary logistic model is used to estimate the probability of a binary response
based on one or more predictor (or independent) variables (features). As such it is not a
classification method. It could be called a qualitative response/discrete choice model in
the terminology of economics. Logistic regression measures the relationship between
the categorical dependent variable and one or more independent variables by estimating
probabilities using a logistic function, which is the cumulative logistic distribution.
Thus, it treats the same set of problems as probity regression using similar techniques,

23 | P a g e
Chapter Two Methods and Methods

with the latter using a cumulative normal distribution curve instead. Equivalently, in the
latent variable interpretations of these two methods, the first assumes a standard logistic
distribution of errors and the second a standard normal distribution of errors. Logistic
regression can be seen as a special case of the generalized linear model and thus
analogous to linear regression. The model of logistic regression, however, is based on
quite different assumptions (about the relationship between dependent and independent
variables) from those of linear regression. Logistic regression is an alternative to
Fisher's 1936

method, linear discriminate analysis. If the assumptions of linear discriminate analysis

hold, the conditioning can be reversed to produce logistic regression. The converse is
not true, however, because logistic regression does not require the multivariate normal
assumption of discriminate analysis.

2.6.5 Fields and example applications

Logistic regression is used in various fields, including the medical and social sciences.
For example, the Trauma and Injury Severity Score (TRISS), which is widely used to
predict mortality in injured patients, was originally developed by Boyd et al. using
logistic regression. Many other medical scales used to assess severity of a patient have
been developed using logistic regression. Logistic regression may be used to predict
whether a patient has a given disease (e.g. diabetes; coronary heart disease), based on
observed characteristics of the patient (age, sex, body mass index, results of various
blood tests, etc). Another example might be to predict whether an American voter will
vote Democratic or Republican, based on age, income, sex, race, state of residence,
votes in previous elections, etc. The technique can also be used in engineering,
especially for predicting the probability of failure of a given process, system or product.
It is also used in marketing applications such as prediction of a customer's propensity to
purchase a product or halt a subscription, etc. In economics it can be used to predict the
likelihood of a person's choosing to be in the labor force, and a business application
would be to predict the likelihood of a homeowner defaulting on a mortgage.
Conditional random fields, an extension of logistic regression to sequential data, are
used in natural language processing.

24 | P a g e
Chapter Two Methods and Methods

Logistic regression can be binomial, ordinal or multinomial. Binomial or binary logistic

regression deals with situations in which the observed outcome for a dependent variable
can have only two possible types (for example, "dead" vs. "alive" or "win" vs. "loss").
Multinomial logistic regression deals with situations where the outcome can have three
or more possible types (e.g., "disease A" vs. "disease B" vs. "disease C") that are not
ordered. Ordinal logistic regression deals with dependent variables that are ordered. In
binary logistic regression, the outcome is usually coded as "0" or "1", as this leads to
the most straightforward interpretation. If a particular observed outcomefor the
dependent variable is the noteworthy possible outcome (referred to as a "success" or a
"case") it is usually coded as "1" and the contrary outcome (referred to as a "failure" or
a "non-case") as "0". Logistic regression is used to predict the odds of being a case
based on the values of the independent variables (predictors). The odds are defined as
the probability that a particular outcome is a case divided by the probability that it is a
non-case. Like other forms of regression analysis, logistic regression makes use of one
or more predictor variables that may be either continuous or categorical. Unlike
ordinary linear regression, however, logistic regression is used for predicting binary
dependent variables (treating the dependent variable as the outcome of a Bernoulli trial)
rather than a continuous outcome. Given this difference, the assumptions of linear
regression are violated. In particular, the residuals cannot be normally distributed. In
addition, linear regression may make nonsensical predictions for a binary dependent
variable. What is needed is a way to convert a binary variable into a continuous one
that can take on any real value (negative or positive). To do that logistic regression first
takes the odds of the event happening for different levels of each independent variable,
then takes the ratio of those odds (which is continuous but cannot be negative) and then
takes the logarithm of that ratio. This is referred (to as logit or log-odds) to create a
continuous criterion as a transformed version of the dependent variable. The logit of
success is then fitted to the predictors using linear regression analysis. The predicted
value of the logit is converted back into predicted odds via the inverse of the natural
logarithm, namely the exponential function. Thus, although the observed dependent
variable in logistic regression is a zero-or-one variable, the logistic regression estimates
the odds, as a continuous variable, that the dependent variable is a success (a case). In
some applications the odds are all that is needed. In others, a specific yes-or-no
prediction is needed for whether the dependent variable is or is not a case; this

25 | P a g e
Chapter Two Methods and Methods

categorical prediction can be based on the computed odds of a success, with predicted
odds above some chosen cutoff value being translated into a prediction of a success.

2.6.6 Logistic regression model

The dependent variable in logistic regressions is usually dichotomous, that is, the
dependent variable can take the value 1 with a probability of success  or the value 0
with probability of failure 1-. The relationship between the predictor and response
variable is not a linear function in logistic regression; here logit transformation of  is
given by:

̂ ,

where

 the constant of the equation

 the coefficient of the predictor variables.

An alternative form of the logistic regression equation is:

[ ] * += Where, Y is the

dependent variable assigning 1 if women body mass index is greater than 21 and 0 if
less than or equal to that value.

2.6.7 Estimation technique

The most common method used to estimate unknown parameters in linear regression is
least squares. Under usual assumptions, least squares equators have some desirable
properties. But when least square method is applied to estimate a model with
dichotomous outcome the estimators on longer have these same properties. In such
situations, the general method for estimating the parameters of logistic regression
model is the method of maximum likelihood.

In logistic regression the likelihood equation is non-linear and explicit function of

unknown parameters. Therefore, we use a very effective and well-known iterative
method, Newton-Raphson method.

Now let us consider a single regression model as

26 | P a g e
Chapter Two Methods and Methods

where

Yi is a dichotomous variable, it takes two values 0 and 1.

So, Yi is a Bernoulli random variable. So, the p.d.f of Yi is given by

Where, Pi is a probability that define as

Implies, ( ) ( )

Since those are assumed to be independent, the joint probability density function is:

Since the logarithm is a monotonic function, so, taking logarithm on (iv), we get

∑[ ]

∑[ { }]

∑[ ( ) ]

27 | P a g e
Chapter Two Methods and Methods

∑[ ( )]

∑[ { }]

∑ ∑ { }

Now differentiating equation (v) with respect to 0and 1 respectively,

∑ ∑* +

∑ ∑

∑ ∑* +

∑ ∑

∑ ∑

∑ ∑
( ) ( )

Where,

1 X1   Y1   p1 
  Y  p 
X 
X2 
Y   , P   2 
 2
1
,
      
  Y  p 
1 Xn   n  n

Now we put , then we get

28 | P a g e
Chapter Two Methods and Methods

XT (Y - p) = 0.................................... (vii)

or XTY = XTp

or ̂ ̂

The solution to equation (vii) will satisfy

( ̂)

Equation (viii) is generally solved by using Newton-Raphson method. This entails first
determining.

{ } [ ]

( )
But we have, ( )

{ }
{ }

{ } { }

{ }
{ }

where, ( )

Iterative estimates of are then obtained as

With playing the role of Yi in this iteratively reweighted least square approach.

29 | P a g e
Chapter Two Methods and Methods

̂
̂
̂( ̂)

̂
With ( ) Notice that i plays an important role of Yi and ̂( ̂)
is the

residual corresponding to divided by the estimated variance of Yi. If we wish to write

equation (ix) in an equivalent form that shown the updating of ̂ , we may write Z as:

̂ ̂ , then we obtain

̂ ( ̂ )

̂ The updating formula given by

equation (x) is used until the estimate converges. The first is to obtain initial
estimates; (0). Various approaches are used to obtain these. As originally derived by
Lachebrunch (1975) displayed by Hosmer and Lemeshow (1989), the initial estimates
obtained using the discriminant functions are given by:

̂
̂ ( ) ̂ ̂ ̂
̂ ̂
( )
̂ ̂ ̂
( ̂ )

Here, ̂ ̅ and ̂ ̅ ,

Where ̅ and ̅ are the average of the n- values

When Y = 0 and Y = 1 respectively. And ̂ ̅ and ̂ ̂

And ̂

Where, and are the usual sample variance computed using Y = 0 and Y = 1
respectively, and n0 and are the corresponding sample size.

2.6.8 Testing the significance of the coefficients

To assess the effect of independent variables on i’th dependent variable, we have to

follow some procedures incorporated with logistic regression model, such as:

(1) Likelihood Ratio Test, (2) Wald Test, (3) Score Test

30 | P a g e
Chapter Two Methods and Methods

We have used Wald test to test our hypothesis

H0: The contribution of the covariates in the model is equal to zero

Vs

H1 = At least one of them is non-zero

2.6.9 Interpretation of parameters

Interpretation of parameters of logistic regression model is not a straightforward as in

linear regression model. So, it is relevant to present a little discussion about it. Since the
logit transformation of g(x) is linear in parameters, we can interpret the parameters
using the arguments of linear regression. Thus, the interpretation may be described as
follows: We know that the logit transformation of a logistic regression model is

( )

( )
is linear in parameters.

That is, * +

So, arguing analogously as in the case of linear regression model, we can say that

represents the rate of change in * + for one unite

change in Xj (other variable remaining constant).

The Interpretation of parameters of logistic regression has another interesting aspect. In

fact, this is the proper interpretation. To describe this, we first consider that the
independent variable (Xj) is dichotomous. This case is not only simplest but also it
gives the conceptual foundation for all other saturations. The description is as
following:

We begin our consideration of the interpretation of logistic regression coefficients with

the situation where the independent variable nominal scale and dichotomous (that is
measured at two levels). This case provides the conceptual foundation for all the other
situations.

We have

( )

31 | P a g e
Chapter Two Methods and Methods

Now if Xj is a dichotomous variable taking values 0 and 1, then the odds ratio, denoted
R, is defined as the ratio of odds (say) Xj = 1 and Xj= 0 is (keeping all other Xs fixed).

pi ( Yi  1 X , X j  1 ) / 1  pi (Yi  1 X , X j  1 )
O
pi ( Yi  0 X , X j  0 ) /1  pi (Yi  0 X , X j  0 )

So, we can directly estimate the coefficients of a logistic regression model as ̂ and
hence can interpret. The simple relationship between the coefficient and the odds ratio
is the fundamental reason why logistic regression has proven to be such a powerful
analytical research tool. The odds ratio is a measure of association which has found
wide use, especially in epidemiology as it approximates how much more likely (or
unlikely) is for the outcome to be present among those with x = l (or 2, or 3) then
among those with x = 0. For example if Y denotes presence or absence of lung cancer
and x denotes whether the person is a smoker, then 0 =2,estimate that lung cancer is
twice more likely to occur among smokers than non-smokers in the study population
.As an example , suppose denotes presence or absence of heart disease and denotes
whether or not the person engaged in regular strenuous physical exercise, If the
estimated odds ratio is 0 = 0.5,then occurrence of heart disease is one half as likely to
occur among those who exercise than among those who do not in the study population.
The interpretation gives for the odds ratio is based on the fact that in many instances it
approximates a quantity called relative risk. The odds ratio approximates the relative
risk if { } { } .This holds when p(x) is small for both 1 and 0.

2.6.10 Computation of probability

We can compute the probability form Pi the estimated odds ratio. This calculation is
very simplest, given a data set of x variables in following equations.

( ) ∑

Where of course s are estimated from the fitted model, then we have

( ) ( )

From this equation (i) can be computed easily.

32 | P a g e
Chapter Two Methods and Methods

In case of continuous independent variable, the interpretation of the estimated

coefficient is similar to that of nominal scaled variables, an estimated log odds ratio. A
meaningful change must be defined for the continuous variable, which is the primary
difference (Hosmer and Lemeshow, 1989).

2.6.11 Binary logistic regression analysis

Binary logistic regression analysis is a popular and widely used analysis that is similar
to linear regression analysis except that the outcome is dichotomous (e.g.,
success/failure or yes/no or died/lived). The epidemiology module on regression
analysis provides a brief explanation of the rationale for logistic regression and how it
is an extension of multiple linear regressions. In essence (see page 5 of that module). In
essence, we examine the odds of an outcome occurring (or not), and by using the
natural log of the odds of the outcome as the dependent variable the relationships can
be linearized and treated much like multiple linear regression. Simple logistic
regression analysis refers to the regression application with one dichotomous outcome
and one independent variable; multiple logistic regression analysis applies when there
is a single dichotomous outcome and more than one independent variable. Here again
we will present the general concept. Hosmer and Lemeshow provide a very detailed
description of logistic regression analysis and its applications. The outcome in logistic
regression analysis is often coded as 0 or 1, where 1 indicates that the outcome of
interest is present, and 0 indicates that the outcome of interest is absent. If we define p
as the probability that the outcome is 1, the multiple logistic regression model can be
written as follows:

exp b0  b1 X 1    bp X p 
pˆ 
1  exp b0  b1 X 1    bp X p 

p̂ is the expected probability that the outcome is present; X1 through Xp are distinct
independent variables; and b0 through bp are the regression coefficients. The multiple
logistic regression model is sometimes written differently. In the following form, the
outcome is the expected log of the odds that the outcome is present,

 pˆ   pˆ 
ln   , ln    b0  b1 X 1  b2 X 2    bp X p
 1  pˆ    1  pˆ  

33 | P a g e
Chapter Two Methods and Methods

Logit transformation
e  o  1 X
The logistic regression model is given by P (Y | X ) 
1  e  o  1 X

 P(Y | X ) 
Which is equivalent to ln    o  1 X
 1  P (Y | X ) 

This is called the logit transformation.

Dichotomous predictor

Consider a dichotomous predictor (X) which represents the presence of risk (1 =

present)

 P(Y  1 X  1)
Odds for Disease with Risk Present   e 0  1
p  1  P(Y  1 X  1)
 e 0 1 X 
1 p Odds for Disease with Risk Absent  P(Y  1 X  0)  e 0
 1  P(Y  1 X  0)


Therefore, the odds ratio (OR)

Odds for Disease with Risk Present e 0 1

Dichotomous predictor   0  e 1
Odds for Disease with Risk Absent e

Therefore, for the odds ratio associated with risk presence we have OR  e 1

Taking the natural logarithm, we have

ln(OR)  1

Thus, the estimated regression coefficient associated with a 0-1 coded dichotomous
predictor is the natural log of the OR associated with risk presence!

Logit is directly related to odds

The logistic model can be written as,

34 | P a g e
Chapter Two Methods and Methods

 PY X  
ln    ln  P    0  1 X
 1  PY X  
 1 P 

P   X
This implies that the odds for success can be expressed as e 0 1
1 P

The multiple binary logistic regression model corresponding to the distress

financing among the rural people in Rangpur Division was:

Log( =β1X1+β2X2+ β3X3+ β4X4+ β5X5+ β6X6+ β7X7 + β8X8

where, P the probability of normal financing, 1-P the probability of distress financing;
X1 Chronic illness exists? X2= Inpatient or Outpatient? X3= Consulted for illness? X4=Health
cost more than 25 percent? X5=Get the medicine form? X6= Did you pay for the medicines?
X7 =How did you travel to the provider? X8 =Old members? The parameter βi refers to the
effect of Xi on the log odds such that Y = 1, controlling the other Xi.

The multiple binary logistic regression model corresponding to the catastrophic

payment among the rural people in Rangpur Division was:

Log( =β1X1+β2X2+ β3X3+ β4X4+ β5X5+ β6X6+ β7X7 + β8X8

where, P the probability of catastrophic payment(no), 1-P the probability of

catastrophic payment(yes); X1 Chronic illness exists? X2= Inpatient or Outpatient? X3=
Consulted for illness? X4=Health cost more than 25 percent? X5=Get the medicine
form? X6= Did you pay for the medicines? X7 =How did you travel to the provider? X8
=Old members? The parameter βi refers to the effect of Xi on the log odds such that Y
= 1, controlling the other Xi.

There is an important assumption in multiple regressions either linear or logistic that

there is no multicollinearity problem (dependent each to other) among the independent
variables. However, there is no exact method of detecting multicollinearity in multiple
logistic regression. In this study, multicollinearity was evaluate using the standard error
(SE). If the SE magnitude falls between 0.001 and 0.5, it is considered indicative of no
multicollinearity (Chan, 2004). Statistical significance was considered at (p<0.05). The
analysis was conducted using SPSS software (IBM version 26) and Microsoft Excel
(version 19).

35 | P a g e
Chapter Two Methods and Methods

Hosmer and Lemeshow test

The Hosmer–Lemeshow test is a statistical test for goodness of fit for logistic
regression models. It is used frequently in risk prediction models. The test assesses
whether or not the observed event rates match expected event rates in subgroups of the
model population. The Hosmer-Lemeshow test specifically identifies subgroups as the
deciles of fitted risk values. Models for which expected and observed event rates in
subgroups are similar are called well calibrated.

2.6.12 Hosmer and Lemeshow goodness of fit test

The Hosmer-Lemeshow goodness of fit test is based on dividing the sample up

according to their predicted probabilities, or risks. Specifically, based on the estimated
parameter values ̂ 0, ̂ 1…, ̂ pfor each observation in the sample the probability that Y=1
is calculated, based on each observation's covariate values:

ˆ 

exp ˆ0  ˆ1 X 1    ˆ p X p

1  exp ˆ  ˆ X    ˆ X
0 1 1 p p 
The observations in the sample are then split into g groups (we come back to choose of
g later) according to their predicted probabilities. Suppose (as is commonly done) that
g=10. Then the first group consists of the observations with the lowest 10% predicted
probabilities. The second group consists of the 10% of the sample whose predicted
probabilities are next smallest, etc.

Suppose for the moment, artificially, that all of the observations in the first group had a
predicted probability of 0.1. Then, if our model is correctly specified, we would expect
the proportion of these observations who have Y=1 to be 10%. Of course, even if the
model is correctly specified, the observed proportion will deviate to some extent from
10%, but not by too much. If the proportion of observations with Y=1 in the group
were instead 90%, this is suggestive that our model is not accurately predicting
probability (risk), i.e., an indication that our model is not fitting the data well.

In practice, as soon as some of our model covariates are continuous, each observation
will have a different predicted probability, and so the predicted probabilities will vary
in each of the groups we have formed. To calculate how many Y=1 observations we
36 | P a g e
Chapter Two Methods and Methods

would expect, the Hosmer-Lemeshow test takes the average of the predicted
probabilities in the group, and multiplies this by the number of observations in the
group. The test also performs the same calculation for Y=0, and then calculates a
1 g
Okl  ekl 2
Pearson goodness of fit statistic is, 
k  0 l 1 ekl
, where

O01denotes the number of observed Y=0 observations in the lth group, O1l denote the
number of observed Y=1 observation in the lth group, and e0l and e1l similarly denote
the expected number of zeros. It should be emphasized that a large p-value does not
mean the model fits well, since lack of evidence against a null hypothesis is not
equivalent to evidence in favor of the alternative hypothesis. In particular, if our sample
size is small, a high p-value from the test may simply be a consequence of the test
having lower power to detect mis-specification, rather than being indicative of good fit.

2.6.13 Classification table

In Logistic Regression, case classification relies on the predicted probability that the
case will be an event (having a higher value on the dependent variable, DV), calculated
using the model equation. By default, a case is projected to be in the event class (e.g., I
in DV coded as 0 and 1) if its predicted probability is at least 0.5. If the event is rare in
the sample, then the predicted probability may be less than 0.5 for all cases. If the
predicted probability of the event falls within the range of 0.01 to 0.49 for cases that
genuinely experienced the event, all these cases would still be predicted as nonevents.
Indeed, the cut-off value for predicted probability in classification can be adjusted. For
instance, setting the cut-off at 0.2 would classify cases as events if the predicted
probability is equal to or exceeds 0.2. In the scenario where the predicted probability
ranged from 0.01 to 0.49 for true event cases, adjusting the cut-off to 0.2 would
correctly classify some of those true events (with predicted probability between 0.2 and
0.49) as events. Adjusting the cut-off introduces a tradeoff; some true nonevents may
be misclassified as events. While the overall correct prediction rate might not improve,
the likelihood of detecting a true event increase. The choice of the cutoff depends on
the relative importance of sensitivity (detecting true event cases) and the false positive
rate (misclassifying nonevents as events).

37 | P a g e
Chapter :Three Results

3.1 Basic Characteristics of the Respondents

In this study, data has been collected from 118 households. Information about the family
members of selected households is given in Table 1. The total number of family members
is 463 with an average household size around four. Of them, 228 are male and 236 are
female. Additionally, there are 32 members who are aged 65 years or older. There are also
36 members who are under 5 years old.

Table 3.1.1: Descriptive statistics of Family Information.

Family Information Minimum Maximum Total Mean

Male 0 5 228 1.93

Female 1 5 236 2.00
Old (65+) 0 2 32 .27
Child (<5) 0 1 27 .23
Total Family Member 1 8 463 3.92

It represents the proportion of male and female participated in this study. It is observed that
51% of the household members are female and 49% of them are male (Figure:3.1.1).

PROPORTION OF MALE AND FEMALE

Female Male
51% 49%

Figure 3.1.1: Proportion of male and female.

38 | P a g e
Chapter :Three Results

In table 3.1.2 it shows the educational status of household heads. We observe that the
majority of the household heads (27.1%) have a Secondary level education followed by
higher education (26.3%), No Education (25.4%), and Primary Education (21.1%).

Table 3.1.2: Educational status of household head:

Educational level Frequency Percentage

No Education 30 25.4

Primary Education 25 21.1

Secondary Education 32 27.1

Higher Education 31 26.3

Total 118

It table 3.1.3 represents the frequency of occupation status according to the household
head. Most of them, about 40%, are farmers. Some are teachers, about 9%. There are also
businessmen, shopkeepers, and service workers, around 7-14%. A few people have other
jobs, making up the last 7.6%.

Table 3.1.3: Occupation status of household head.

Occupation Frequency Percentage

Farmer 47 39.8

Teacher 11 9.3

Day Labor 9 7.6

Business Man 17 14.4

Shop Keeper 9 7.6

Service 16 13.6

Other 9 7.6

Total 118 100

39 | P a g e
Chapter :Three Results

According to (Table 3.1.4), 33.1% of households came from lower middle-class families
(monthly income: 10,000 to 20000 BDT), while 30.5% came from middle-class families
(monthly income: 20,000 to 30,000 BDT). Only 5% of households were from the poorest
families (monthly income less than 10,000 BDT), and just 4% came from rich families
(monthly income: 50,000 BDT and above).

Table 3.1.4: Income status of household head

Monthly family income (BDT) group Frequency(N%)

Less than 10,000 (Poor) 5 (4.3)

10000 to 20,000 39 (33.1)

20000 to 30,000 (Middle) 36 (30.5)

30000 to 40,000 25 (11.1)

40000 to 50,000 9 (7.6)

50000 and above (Rich) 4 (3.4)

Total 118

In table 3.1.5 is like a picture showing if family members got sick in the last 30 days. Out
of a total of 463 family members, 212 of them (about 46%) got sick, and the rest, 251
(about 54%), did not get sick. So, it shows us that a bit more than half of the family
members did not have any illness recently.

Table 3.1.5: Presence of illness symptoms in the last 30 days.

Presence of illness in the
Frequency Percentage
last 30 days
Yes 212 45.8

No 251 54.2

Total 463 100

40 | P a g e
Chapter :Three Results

Table:3.1.6 displays details about family members' illness symptoms over the past 30 days,
providing insight into their health. Notably, fever is the most prevalent symptom at 42.4%,
followed by pain (18.8%), weakness (6.8%), and other symptoms. Diarrhea affects 9.4%,
5.6% have breathing trouble issues and 9.2% have various health problems under others.

Table 3.1.6: Symptoms of illness in members of households

Illness Symptoms
Illness Symptoms.
Frequency Percentage

Fever 106 42.4

Pain 47 18.8

Weakness 17 6.8

Diarrhea 19 7.6

Breathing trouble 14 5.6

Blood Pressure 10 4

Female diseases 8 3.2

Pregnancy related 6 2.4

Other 23 9.2

Table 3.1.7 shows information about family members having a chronic illness. Out of a
total of 463 family members, 117 (about 25%) have a chronic illness, while the majority,
346 (about 75%), do not have any chronic illness.

41 | P a g e
Chapter :Three Results

Table 3.1.7: Symptoms of chronic illness in members of households

Family member with

Frequency Percentage
chronic illness.
Yes 117 25.3
No 346 74.7
Total 463 100

Table 3.1.8 shows the prevalence of chronic health conditions among family members. The
data indicates that some of the most commonly reported conditions include blood pressure
issues (24.13%), asthma/breathing trouble/diabetes (19.79%), chronic heart disease
(11.72%), and gastric/ulcer issues (9.65%). Other conditions, such as chronic fever
(6.22%), paralysis (6.21%), and arthritis/rheumatism, eczema, injuries, etc. (15.86%) were
also reported

Table 3.1.8: Chronic diseases among family members within the past year

42 | P a g e
Chapter :Three Results

Table 3.1.9 provides about how family members get their medicine. Additionally, it shows
that around 6% of the 240 family members get medicine from government and NGO
health centers, while the majority, about 82%, get their medicine from pharmacies or
dispensaries. Some people, around 5%, get medicine from other shops, and about 7% get it
from both (Govt. health centers and Pharmacies)

Table 3.1.9 Family members get their medicine from.

Get the medicine Form Frequency Percentage

Govt. and NGO health center 15 6.2

Pharmacy/ dispensary 196 81.7

Other shop 12 5

GHC+ Pharmacy 17 7.1

Total 240 100

Table 3.1.10 gives information about where family members go when they need health
advice. Out of 240 family members, about 13% consult government health workers, while
5% go to Ayurved, Kabiraji, or Hekim. Only a few, about 2%, consult doctors from NGO
facilities. A large group, around 46%, prefer doctors from private facilities, and about 31%
talk to pharmacists. A small number, around 3%, choose self-treatment

43 | P a g e
Chapter :Three Results

Table 3.1.10: Family members sought consultation from

Consulted from Frequency Percentage

[Link] 30 12.5

Ayurved /Kabiraji / Hekim 12 5

Doctor from NGO Facility 4 1.7

Doctor from Private 111 46.3

Facility

Salesman of a pharmacy 75 31.3

Self-treatment 8 3.3

Total 240 100

Table 3.1.11 provides insights into the ways families seek care when someone is ill. The table
shows that out of 240 family members, a significant portion (10%) opt for inpatient care, while
around 8% choose outpatient care from public hospitals. A larger group of about 35% prefer
private hospitals for their outpatient care needs. Additionally, some families (7%) go for both
public and private outpatient care. Quite a few (around 35%) consult pharmacists

Table 3.1.11: Family members Care-seeking behavior:

Care-seeking behavior Frequency Percentage
Inpatient 25 10.4
Outpatient Public 19 7.9
Outpatient Private 85 35.4
Out Patient (Public and Private) 16 6.7
Self-treatment and Salesman of Pharmacy 83 34.6
Ayurved /Kabiraji / Hekim 12 5
240 100

44 | P a g e
Chapter :Three Results

The figure (3.1.2) and the table (3.1.12) indicate that an average food-related expenses
accounted for almost half of a household's expenditures. Education-related expenses made
up 16% of household expenses, while healthcare expenses were around 15.5% of total
household expenditure. In addition, roughly 16% of total household expenditure is spent
on other sectors, including household goods and maintenance.

Figure 3.1.2: Monthly Household expenditure in different sectors.

Table 3.1.12: Household expenditure in different sectors.

Percentage of Percentage of Percentage of Percentage

food cost Education cost family monthly of other
monthly Monthly treatment cost cost

Mean 51.67 16.94 15.51 16.09

Minimum 20.00 .00 1.32 .00
Q1 40.00 0.00 8.17 7.02
Median 50.00 15.00 13.10 13.54
Q3 62.87 27.59 21.42 24.16
Maximum 90.00 69.23 60.61 52.63

45 | P a g e
Chapter :Three Results

In table 3.1.12 showed that the monthly treatment cost of the households participated in the
survey was as low as 1.32% of their total consumption and as high as 21.43% with a median
13% indicating that 50% of the households had treatment cost more than 13% of the total
monthly consumption (Table 3.1.12).

The monthly food cost in the survey was as low as 20 % and as high as 90% of their total
consumption. And median 50% indicating that 50% of the households had food cost more
than 50% of the total monthly consumption.

The minimum education cost in the households was 0.00 percent. The maximum education
cost in the households was 69.23%, and he median percentage of households cost was 15%
indicating that 50% of the households had education cost more than 15% of their monthly
consumption.

The maximum households expenditure in others sector was 52.63% and median households
cost was 13.54%.

46 | P a g e
Chapter :Three Results

3.2 Coverage of essential health services

Outpatient, inpatient, and preventive care are primarily provided by the public sector,
whereas the private sector is predominantly utilized for outpatient and inpatient curative
care. The Ministry of Health and Family Welfare (MOHFW) is the primary organization
responsible for public health services. There are three main levels: primary health care,
district, and divisional or tertiary levels.

In table 3.2.1 provides an overview of the healthcare facilities available in the study area,
located in Binoypur village in Panchagarh district

At the primary level, healthcare facilities include the Community Clinic (2 km), Union
Health and Family Welfare (4 km), Upazila Health Care (13 km, 50 beds), and
Ayurved/Koniraji/Hekim (4/3 km,). Moving to the secondary level, there is a District
Hospital (36 km, 100 beds).

At the tertiary level, notable healthcare facilities include Rangpur Medical College
Hospital (120 km, 1500 beds), M Abdur Rahim Medical College Hospital (63 km, 500
beds), and Rangpur Mother and Children Hospital (117 km).

Table 3.2.1: Health care facilities and Distance (Binoypur)

Study Area Level Health Care Distance Bed

Facilities (Km) Capacity
1. Community Clinic 2 -
[Link] health and family -
Primary 4
welfare
Level
[Link] Health Care 13 50
[Link]/Koniraji/hekim 4/3 -
Vill: Binoypur
Secondary 100
[Link] Hospital 36
Level
Post: Kaligonj
Rangpur Medical college 1500
Upozila:Debigonj 120
Hospital
District:Panchagarh
M Abdur Rahim Medical 500
Tertiary
College 63
Level
Hospital
Rangpur Mother and -
117
Children Hospital

47 | P a g e
Chapter :Three Results

In the studied region (specifically Daikotbari village in Dinajpur district), the Primary
Level of healthcare includes the Community Clinic, situated 2 kilometers away, the Union
Health and Family Welfare Center, also 2 kilometers away, Upazila Health Care facilities
located 10 kilometers away with a bed capacity of 50, and Ayurved/Koniraji/hakim,
situated 4 kilometers away (Table 3.2.2).

Secondary Level facilities encompass the District Hospital, positioned 20 kilometers away
with a bed capacity of 250, and Maternal and Child Welfare Centers, also 20 kilometers
away from the village. Additionally, NGO facilities (BRAC/World Vision) are available in
this area, located 12 kilometers away from the village.

At the Tertiary Level, the prominent healthcare institution is Rangpur Medical College
Hospital, located 120 kilometers away with a bed capacity of 1500.

Table 3.2.2: Health care facilities and Distance (Daikotbari)

Distance Bed
Study Area Level Health Care Facilities
(Km) Capacity
1. Community Clinic 2 -
[Link] health and family
Primary 2 -
welfare
Level
[Link] Health Care 10 50
Vill: Daikotbari
[Link]/Koniraji/hekim 4 -
[Link] Hospital 20 250
Post: Balandor Secondary
[Link] and child welfare
Upazila: Birol Level 20 -
Centers
District:
NGO BRAC/World Vision 12 -
Dinajpur
[Link] Medical college
120 1500
Tertiary Hospital
Level M Abdur Rahim Medical
24 500
College Hospital

48 | P a g e
 Chapter :Three Results

Table 3.2.3 shows the rural area of Gandhamarua, Lalmonirhat district, at the Primary
Level, there is a Community Clinic, Union Health and Family Welfare facilities, and
Upazila Health Care located 13 kilometers away, with a combined bed capacity of 50.
Additionally, Ayurved/Koniraji/hekim is situated five kilometers away from
Gandhamarua.

At the Secondary Level, there is a District Hospital positioned 50 kilometers away, with a
combined bed capacity of 250.

At the Tertiary Level, there is a Medical College Hospital located 47 kilometers away, with
a bed capacity of 1500, a Specialized Hospital 49 kilometers away, and M Abdur Rahim
Medical College Hospital positioned 112 kilometers away, with a bed capacity of 500.

Table 3.2.3: Health care facilities and Distance (Gandhamarua)

Level Health Care Facilities Distance Bed

Study Area
(Km) Capacity
1. Community Clinic 1 -

[Link] health and family

4 -
Primary Level welfare

Vill: Gandhamarua [Link] Health Care 13 50

[Link]/Koniraji/hekim 5 -

Post: Durgapur Secondary

[Link] Hospital 21 250
Upazila: Aditmari. Level

District:Lalmonirhat [Link] college Hospital 47 1500

Tertiary Level [Link] Hospital 49 -

3.M Abdur Rahim Medical

112 500
College Hospital

49 | P a g e
Chapter :Three Results

In the specified study area (Notun Onontopur) in Kurigram district, at the Primary Level,
there is a Community Clinic located 2 kilometers away. Additionally, Union Health and
Family Welfare facilities are situated 4 kilometers away, while Upazila Health Care
centers are positioned 6 kilometers away, offering a combined bed capacity of 50.
Ayurved/Koniraji/hekim facilities are also available, 8 kilometers away from the village.
(Table 3.2.4)

At the Secondary Level, a District Hospital is located 20 kilometers away, with a bed
capacity of 250, while Maternal and Child Welfare Centers are situated 22 kilometers
away.

At the Tertiary Level, Rangpur Medical College Hospital is positioned 65 kilometers

away, providing a substantial bed capacity of 1500. Additionally, there is a Specialized
Hospital located 67 kilometers away.

Table 3.2.4: Health care facilities and Distance (Notun Onontopur)

Distance Bed
Study Area Level Health Care Facilities
(Km) Capacity
1. Community Clinic 2 -
[Link] health and family
4 -
Primary Level welfare
Vill: Notun
[Link] Health Care 6 50
Onontopur
[Link]/Koniraji/hekim 8 -
Post: Ulipur
[Link] Hospital 20 250
Upazila: Ulipur. Secondary
Level [Link] and child
District: 22 -
welfare Centers
Kurigram
[Link] Medical
65 1500
Tertiary Level college Hospital
[Link] Hospital 67 -

50 | P a g e
Chapter :Three Results

It detailing healthcare facilities and distances in the study area covering Vill: Parul, Post:
Deuty, Upazila: Pirgacha, District: Rangpur (Table 3.2.5).

At the Primary Level, there is one Community Clinic, one Union Health and Family
Welfare facility, and an Upazila Health Care center located 12 kilometers away, with a
combined bed capacity of 50. Additionally, Ayurved/Koniraji/hekim facilities are
available.

At the Secondary and Tertiary Levels, Rangpur Medical College Hospital is situated 25
kilometers away, boasting a significant bed capacity of 1500. Furthermore, there is a
Specialized Hospital located 27 kilometers away

Table 3.2.5: Health care facilities and Distance (Parul)

Distance Bed
Study Area Level Health Care Facilities
(Km) Capacity

1. Community Clinic 1 -

[Link] health and family

3 -
Vill: Parul welfare
Primary Level
Post: Deuty [Link] Health Care 12 50
Upozila:
[Link]/Koniraji/hekim 5 -
Pirgacha
District: Secondary [Link] Medical
25 1500
Rangpur Level college Hospital
[Link] Medical
25 1500
college Hospital
Tertiary Level
[Link] Hospital 27 -

51 | P a g e
Chapter :Three Results

It displays information on whether families had to bear the cost of medicines or not. Out of
240 family members, around 86% of them had to pay for their medicines. Only a small
percentage, approximately 6%, did not have to pay for their medicines. However, some
families, about 7%, had to pay for some medicines while not paying for others (Table
3.2.6).

Table 3.2.6: Did Family members pay for the medicines or not?
Pay for the medicines or Frequency Percentage
not.
YES 207 86.3
NO 15 6.3
Partial 18 7.5
Total 240 100

Table 3.2.7 illustrates that auto rickshaws are the most favored means of transportation for
individuals who travel to health providers, as chosen by 37.9% of respondents. Walking is
also a significant means of travel at 28.9%, which suggests that health facilities are located
relatively close by. Other methods of transportation include public transport (16.7%),
motorcycles (9.6%), and ambulances (5%). Interestingly, 2.1% of respondents prefer
calling a doctor

Table 3.2.7: Transportation to the healthcare provider.

Transportation the healthcare Provider Frequency Percentage

Walking 69 28.9
Bus 40 16.7
Auto rickshaw 91 37.9
Motorcycle 23 9.6
Ambulance 12 5
Calling doctor 5 2.1
Total 240 100

52 | P a g e
Chapter :Three Results

3.3 Health care cost and financial protection.

Health Care Spending:

Figure 3.3.1 shows how much households spend on health care. Out of 118 households,
65.3% of them spend more than 10% of their income on health care, and 44.1% spend
more than 15%. Unfortunately, 8.7% of households face financial difficulties and have to
resort to Catastrophic payment methods.

Health care expenditure.

70 65.3

60

50 44.1

40

30

20
8.7
10
1.7

0
More than 10% More than 15% More than 25% More than 40%

Figure:3.3.1 Health care spending.

53 | P a g e
Chapter :Three Results

3.3.(a) Catastrophic health spending:

There is no right or wrong approach to measuring catastrophic health expenditures.
Different studies adopt different approaches. Some studies define out-of-pocket health
expenditures as catastrophic when they exceed a given percentage (for example, 10% or
25%) of income or consumption (1). In our study we define out-of-pocket health
expenditures as catastrophic when they exceed a given percentage (more than 25%) of
consumption.

In our study we found that 8.7% of households who participated in the survey had
treatment cost over 25% of their total consumption. (Figure 3.3.2)

Financing for the treatment

8.7%
Chatastropic Payment (treatment
cost over 25%)
Treatment total cost less than 25%

91.3%

Figure:3.3.2 Financing for the treatment (Catastrophic or not).

54 | P a g e
Chapter :Three Results

3.3.1Relation of Catastrophic health payments with health-related variable:

We use the Chi-square test to find the link between Catastrophic health payments (health
costs exceeding 25% of expenditure) and health-related factors among rural residents in
Rangpur Division. Table 3.3.1 shows that catastrophic health spending is linked including
chronic illness, hospitalization (inpatient-outpatient), consultation, and travel to the
healthcare provider. In our sample, it is observed that obtaining medicine from (a
government health center or pharmacy/dispensary), paying for the medicine, and the age
group of elderly individuals (old members) did not show a significant association with
distress financing.
The study also finds that the presence of chronic illness is significantly linked to distress
financing (p-value = 0.00). When it comes to catastrophic health expenses (health costs
exceeding 25%), individuals with chronic illnesses face a higher risk than those without.
our study finds that 29.1% of people with chronic illness experience catastrophic health
payments, while only 4.1% of those without chronic illness do. In other words, those with
chronic illness are more likely to face financial difficulties due to their health expenses
than those without.
It is found that there is a statistically significant association between inpatient care and
catastrophic health payments (p-value=0.000<0.05). For individuals who receive inpatient
care, 87.5% encounter catastrophic health spending, whereas 12.5% do not. Conversely,
among those receiving outpatient care, 5.0% experience catastrophic health payments,
while 95.9% do not
Significant association between consulting a qualified doctor and catastrophic health
payments (p-value=0.001<0.05). In our sample, we notice that for catastrophic health
payments, 24.5% consult a qualified doctor, while 1.2% consult a salesperson from a
pharmacy. In the case of fairly normal health spending, 75.5% consult a qualified doctor,
and 98.8% consult a salesperson from a pharmacy The association between transportation
mode and catastrophic health payments is found to be statistically significant (p-value =
0.000). Among those who walk, none (0.00%) experience catastrophic health payments.
Ambulance users have a high rate of catastrophic health spending (75.0%), while bus users
face a rate of 42.1%. Auto rickshaw users encounter a 9.9% rate of catastrophic health
spending, and motorcycle users have a 20.0% rate.

55 | P a g e
Chapter :Three Results

Table: 3.3.1 Association between catastrophic payment and different

socioeconomic and health related characteristics among rural people in
Rangpur Division.

Catastrophic Payment Chi-

Characteristics p-value
square
No, N (%) Yes, N (%)

Chronic illness exists?

No 118 (95.9) 5 (4.1)
27.526 0.00
Yes 69 (59.0) 34 (29.1)
Inpatient or Outpatient?
Inpatient 3 (12.5) 21 (87.5)
99.472 0.00
Outpatient 198 (91.7) 18 (8.3)
Consulted for illness?
Qualified Doctor 117 (75.5) 38 (24.5) 21.97 0.00
Salesman of a Pharmacy 84(98.8) 1 (1.2)
Get the medicine form?
Govt health centre 30 (93.8) 2 (6.3)
2.713 0.100
Pharmacy/Dispensary 171 (82.2) 17 (17.8)
Did you pay for the medicines?
No 12 (75.0) 4 (25.0)
0.177 0.674
Yes 187 (83.5) 37 (16.5)
How did you travel to the provider?
Walking 74 (100) 0 (0.00)
Ambulance 3 (25.0) 9 (75.0)
Bus 22 (57.9) 16 (42.1) 66.42 0.000
Auto Rickshaw 82 (90.1) 9 (9.9)
Motorcycle 20 (80.0) 5 (20.0)
Aged members (>65)?
No 55 (72.5) 14 (20.3)
0.205 0.651
Yes 15 (75.0) 5 (25.0)

56 | P a g e
Chapter :Three Results

Table 3.3.2 shows how people fund medical treatment. The majority (66.9 %) rely on
regular income, while household savings contribute significantly (21.04%). Selling
livestock/permanent assets (3.2%), borrowing (from Friends/ Relatives/Office) (6.8%), and
other methods collectively form diverse financial strategies. %)

Table 3.3.2: Financing sources for the treatment.

Finance from the treatment

Financing sources for the treatment
Frequency (Percentage)

Regular income 70 (66.9)

Household saving 25 (21.4)

Sold livestock/ permanent assets/belonging 4 (3.2)

Mortgage of Assets/Land 2 (1.7)

Borrowed from Friends/ Relatives/Office 7 (6.8)

3.3. (b) Distress financing on health

Distress financing means getting money during tough times, especially for sudden costs
like treatment bills. It involves like borrowing money, selling things, or making urgent
financial arrangements. In healthcare, people might use distress financing when they have
big medical expenses that are hard to pay with their usual income or savings.

An experience of distress financing occur mainly due to excessive out-of-pocket (OPP)

spending are those made by people at the time of getting any type of service (preventive,
curative, long-term care) provided by any type of provider. They include cost-sharing (the
part not covered by a third party like an insurer) but they exclude insurance premiums.
Out-of-pocket payments could be financed out of a household’s income, its savings, or by

57 | P a g e
Chapter :Three Results

borrowing. They do not include any payment made by a government agency, a health
insurance fund, or a private insurance provider.

It indicates the financing condition for medical treatment. The majority, 88.3%, managed
their treatment cost with household consumption, a smaller yet notable percentage,
constituting 11.7%, experiences financial distress when seeking funds for their medical
needs (Figure.3.3.3).

Figure 3.3.3: Family member’s health financing condition

Financing For the treatment

11.7%

88.3%

Managed from household consumption Distress Financing

58 | P a g e
Chapter :Three Results

The Chi-square test shows a significant association (p-value = 0.00), between catastrophic
health spending and distress financing. Majority (79.5%) of the health seekers had
experiencing of distress financing who experience catastrophic payment.

Table: 3.3.3 Association between Catastrophic health payments and

distress financing.

3.3.2 Relationship of distress financing with health-related variable:

The Chi-square test is being used to determine the association between distress financing
and different socioeconomic and health-related characteristics among rural people in
Rangpur Division. It demonstrates that different socioeconomic and health-related
characteristics such as chronic illness, hospitalization (inpatient-outpatient), consultancy,
catastrophic payment (health cost more than 25%), and travel to the provider are associated
with distress financing. In our samples, we observe that getting the medicine from a Govt
health center, Pharmacy/Dispensary, paying for the medicine, and aged people (old
members) were not significant related with distress financing (Table 3.3.4).

It is found that there is significant association between chronic illness and distress
financing (p-value=0.001<0.05) Among our sample, we observe that 41% of chronic
patients had distress financing which only about 5% of non-chronic patient had distress
financing.

Significant association between inpatient care and distress financing is also found (p-
value=0.000<0.05) Among those who receive inpatient care, 75.0% experience distress

59 | P a g e
Chapter :Three Results

financing, while 25.0% do not. For individuals receiving outpatient care, 16.7% experience
distress financing, while 83.3% do not.

Consulting qualified doctors is significantly associated with distress financing (p-value =

0.00). There is a noticeable difference in the prevalence of distress financing between those
who consult a qualified doctor and those who consult a salesman at a pharmacy. Among
those who consult a qualified doctor, 33.5% experience distress financing, while 66.5% do
not. Of individuals who consult a salesman at a pharmacy, only 2.4% experience distress
financing, while 97.6% do not.

The association between transportation and distress financing is statistically significant (p-
value = 0.000). These who travelled by ambulance for health seeking face a high distress
financing rate (83.3%), while bus users experience 71.1%. Auto rickshaw users have
13.2% distress financing, and motorcycle users have 20.0%.

60 | P a g e
Chapter :Three Results

Table: 3.3.4 Association between distress financing and different

socioeconomic and health-related characteristics among rural people in
Rangpur Division.

Distress Financing Chi-

Characteristics p-value
No, N (%) Yes, N (%) square
Chronic illness exists?
No 11 (95.3) 6 (4.9)
44.93 0.00
Yes 69 (59.0) 48 (41.0)
Inpatient or Outpatient?
Inpatient 6 (25.0) 18 (75.0)
42.15 0.00
Outpatient 180 (83.3) 36 (16.7)
Consulted for illness?
Qualified Doctor 103 (66.5) 52 (33.5)
30.63 0.00
Salesman of a Pharmacy 83(97.6) 2 (2.4)
Health cost more than 25 percent?
No 178 (88.6) 23 (11.4)
86.726 0.00
Yes 8 (20.5) 31 (79.5)
Get the medicine form?
Govt health center 27 (84.4) 5 (15.6)
1.001 0.317
Pharmacy/Dispensary 159 (76.4) 49 (23.6)
Did you pay for the medicines?
No 12 (75.0) 4 (25.0)
0.061 0.804
Yes 174 (77.7) 50 (22.3)
How did you travel to the provider?
Walking 74 (100) 0 (0.00)
Ambulance 2 (16.7) 10 (83.3)
Bus 11 (28.9) 27 (71.1) 102.9 0.000
Auto Rickshaw 79 (86.8) 12 (13.2)
Motorcycle 20 (80.0) 5 (20.0)
Aged members (>65)?
No 50(72.5) 19(27.5)
1.140 0.286
Yes 12(60.0) 8(40.0)

61 | P a g e
Chapter :Three Results

3.3.3 Determinants of distress financing

Binary logistic regression estimates the effects of different socioeconomic and health-
related characteristics factors on distress financing among rural people in Rangpur
Division. We consider the independent variables in the logistic regression model that are
significantly associated with distress financing. Financing health costs are categorized into
two categories: distress and not distress financing, which is considered the dependent
variable.

Standard error shows that there is no evidence of multicollinearity problems among

independent variables. Hosmer and Lemeshow test [χ2 =10.910, p-value=0.03] showed
that our selected model was good model for exploring the dependent variables. Showed
that about 43.2% of the variation in the model dependent variable (distress financing) is
explained by this logistic model. The results of binary logistic regression analysis of
different socioeconomic and health-related characteristics factors on distress financing are
displayed.

People who are affected by chronic disease are 3.584 times more likely to had distress
financing than those who are not affected by chronic disease [AOR=3.584, 95%CI:1.24-
10.355; P<0.05].

People who reported they get their treatment from a qualified doctor are 4.709 times more
likely to had distress financing than those who get their treatment from a salesman at a
pharmacy disease at a 10% level of significance [AOR=4.709, 95%CI:0.938-23.267;
P<0.10].

Rural people in Rangpur are likely to experienced distress financing 2.043 times more if
they receive inpatient care than outpatient care [AOR=5.848, 95%CI:2.043-16.745;
P<0.05].

If the distance (consulting qualified doctors) increases by one unit, the distress financing
scale will also increase by one unit [AOR=1.001, 95%CI:1.001-1.006; P<0.03].

62 | P a g e
Chapter :Three Results

Table: 3.3.5 Determinants of distress financing

95% C.I. for

p- Exp(β EXP(β)
Variables β SE Wald DF
value )
Lower Upper

1.276 0.5 5.56 1 .018 3.584 1.240 10.355

Chronic Illness (Yes/No)
41

Distance to the .008 0.0 4.63 1 .031 1.008 1.001 1.016

provider? (Km) 04

1.766 0.5 10.82 1 .001 5.848 2.043 16.745

Inpatient vs outpatient
37
Qualified Doctor vs 1.549 0.8 3.54 1 0.060 4.709 0.938 23.267
Salesman of a Pharmacy 23
Hosmer and Lemeshow Chi-Value=10.910 Nagelkerke R Square=0.431
test

Classification Table:

The overall accuracy of this model to predict residents having distress financing (with a
predicted probability of 0.5 or greater) is 83.3% (see Table 3.3.5). The sensitivity is
calculated as 186/180=96.8%, and the specificity is 21/54=38.9%. The Positive Predictive
Value (PPV) is 180/213=84.50%, and the Negative Predictive Value (NPV) is 77.77%.

Predicted
Observed Distress financing
Percentage correct
No Yes
No 180 6 96.8
Distress financing
Yes 33 21 38.9
Overall percentage 83.8

Table 3.3.6 Classification table of binary logistic regression for distress

financing

63 | P a g e
Chapter :Three Results

3.3. (c) Financial protection:

Insufficient healthcare coverage, particularly in rural areas, remains a significant obstacle.

With just 2.5% of the population having health insurance, the majority faces financial
burdens associated with healthcare expenses (Dhaka Tribune). In our study, we find only 2
families with health insurance among the 118 households surveyed.

In response to these challenges, the government of Bangladesh sets an ambitious target of

achieving Universal Health Coverage (UHC) by 2032. Recently, the government
implements a pilot project of health insurance called “Shasthyo Shurokhsha Karmasuchi
(SSK).” Developed by the Health Economics Unit (HEU) under the Ministry of Health and
Family Welfare (MoHFW), SSK (Social Security for the Poor) aims to provide
complimentary healthcare services to households below the poverty line (BPL).

The benefit package includes issuing one health card per household and offering free
treatment services for 50 specified diseases. This initiative is designed to safeguard these
households from facing financial hardship in the event of catastrophic illness. However,
the existing benefit package may not be sufficient to protect low-income households from
the financial burden of out-of-pocket payments, especially given the ongoing rise in
healthcare costs.

Achieving a significant decrease in the burden of out-of-pocket payments appears

challenging unless there is implementation of a national health insurance scheme covering
all citizens, with a specific focus on prioritizing the needs of the poor population

64 | P a g e
Chapter: Four Discussion

DISCUSSION

Residents in the rural areas of Rangpur Division faced significant challenges in

healthcare financing, as revealed by this study. This is one of the first study at rural area
in Rangpur (to include evidence of financial hardship and financial coverage) regarding
UHC.

Households in the rural areas of Rangpur Division have to spent about 15 % of their
consumption for seeking healthcare. We find among the households, 65.3% spend over
10% on healthcare, 44.1% spend over 15%, and 8.7% face financial challenges
resorting to Catastrophic payment (approximately one in nine households experience a
financial catastrophe.).

Around 13% of people consult government health workers for treatment. A significant
number, approximately 46%, prefer to seek medical advice from private doctors. Due
to the lack of quality of treatment in government hospital, people generally prefer
private health care.

13% consult government health workers. A large group, around 46%, prefer doctors
from private facilities, and about 31% talk to pharmacists. A small number, around 3%,
choose self-treatment.

As the lack of quality of treatment in rural area people generally prefer private health
care.

Our study also finds that around 11.7% of households face distress financing
(borrowing or selling household assets) to pay for healthcare costs. There is a
relationship between catastrophic health spending and distress financing. Among those
who experience Catastrophic Payment, only 8 (20.5%) do not face distress financing,
while the majority, 31 (79.5%), do. Thus, households facing catastrophic health
payments are at a higher risk of resorting to distress financing.

There is a relationship between catastrophic health spending and distress financing.

Among those who experience Catastrophic Payment, only 20.5% do not face Distress
Financing, while the majority, (79.5%), do. The Chi-square test shows a significant
association (p-value = 0.00), indicating an association between catastrophic health

65 | P a g e
Chapter: Four Discussion

spending and distress financing. This suggests that those households experiencing
catastrophic health payments are more likely to have distress financing.

Similar to findings in studies conducted in developing nations (Steinhardt LC, 2009),

the risk of using distress financing was strongly associated with financial hardship
including catastrophic payment.

Catastrophic health payments and distress financing are associated with health-related
traits, including chronic illness, hospitalization, consultation (qualified doctor or
salesman in a pharmacy), distress financing (borrowing or selling assets), and
transportation to the healthcare provider.

Healthcare costs are significantly higher for individuals suffering from chronic illnesses
compared to those with illness symptoms. Having a member with a chronic illness in a
household nearly 3.584 times increases the risk of resorting to distress financing.

Financial difficulties are strongly connected to the use of health services in Bangladesh.
This study reveals that the occurrence of financial hardship varies significantly between
those seeking inpatient care and those receiving care in outpatient facilities. People who
receive inpatient care suffer 2.043 times more than those who receive outpatient care.
These findings are consistent with studies conducted in developing countries (Laura C
Steinhardt, 2008).

Many people face distress financing when consulting a qualified doctor, compared to
only a few who face it when consulting a pharmacy salesman for their illness. Seeking
treatment from a qualified doctor increases the chance of distress financing by 4.709
times, compared to obtaining treatment from a pharmacy salesman.

Many people face financial difficulties when it comes to transportation. Furthermore,

the survey reveals that the farther people have to travel to visit a qualified doctor, the
more likely they are to experience financial difficulties. For every one unit increase of
distance, the chance of experiencing financial distress increases by one percent.

While other studies have shown that elderly individuals (>65) are significantly
associated with care-seeking decisions (Pannarunothai and Mills, 1997), our study
found no significant association after adjusting for other covariates.

66 | P a g e
Chapter: Four Discussion

The obstacle of limited healthcare coverage, particularly in rural areas, is significant.

With only 2.5% of the population having health insurance, the majority faces
vulnerability to financial strains associated with healthcare expenditures (Dhaka
Tribune). Furthermore, our study finds that out of 118 households surveyed, only two
report having health insurance.

According to Xu and colleagues (Ke Xu, July 12, 2003), the main factors leading to
financial risk include catastrophic payments, distress financing, and the absence of
health insurance. Our study also identifies that all these conditions are present in the
area we examined. Therefore, it is clear that public health services fail to serve as an
effective social safety net.

67 | P a g e
Chapter: Five Conclusion and Recommendation

5.1 Conclusion

This study investigated the health service coverage for rural population of Rangpur
division and the financial hardship for seeking health care services; factors of financial
hardship were also determined. We found that only a small percentage, approximately
13%, of people choose to consult government health workers for treatment. This is due
to the lack of quality treatment available in government healthcare facilities, which
pushes people to seek private healthcare options instead. It is found that 8.7% of rural
Rangpur households experienced catastrophic expenses, indicating that approximately
one in nine households faces financial catastrophe related to healthcare expenditures.
Furthermore, approximately 11.7% of households resort to distress financing, which
involves borrowing money or selling household assets to meet healthcare expenses. On
the other hand, limited insurance coverage exists in rural areas; we found only two
households having health insurance, while the rest have no insurance coverage. The
financial hardship, including catastrophic payments and distress financing, is associated
with the presence of chronic illness, consulting a qualified doctor, inpatient care, and
distance to the treatment provider.

In Bangladesh, health seekers usually have to pay from their own pocket. When people
suffer from some chronic illness, treatment cost increases due to consultation fee,
hospital fee, medicine and hospitalization cost. This cost increase at unbearable level, if
someone prefer specialized doctors in some private set up. Since the insurance
coverage is very low in rural areas, catastrophic payment and distress financing are
obvious outcome.

Implementing a risk-pooling mechanism, increasing GDP and spending on health, and

ensuring effective monitoring of subsidized programs in public health facilities can
protect households from healthcare-related financial burdens.

68 | P a g e
Chapter: Five Conclusion and Recommendation

5.2 Recommendation

 It is essential to prioritize health as a top sector in the national budget, as there

has been a noticeable decrease in the budget allocation for the health ministry
over time.
 Enhance access to quality healthcare services at the district and upazila level
government hospitals.
 Optimize the use of current resources and ensure the maximum utilization of the
annual allocation in the government's health budget for enhanced efficiency.
 Promote health insurance as another way to pay for healthcare in both the
government and private sectors in Bangladesh. Set up a health insurance
marketplace with the same features for each insurance company to make it easy
for people to choose.
 Increase investment in the production of Human Resources for Health (HRH)
and necessary infrastructure, focusing on generating skilled professionals,
particularly doctors and trained nurses.
 Reforms should involve increasing government spending on health by
reallocating budgetary resources, ensure effective monitoring of subsidized
programs, ensuring standardized costs for both official and unofficial fees
across all public facilities and committing to providing health insurance
coverage for the whole population.
 There is an urgent need to implement mandatory pre-payment schemes for
formal sector employees and subsidized schemes for the impoverished
population, as outlined in the health financing strategy.

69 | P a g e
 Reference

Reference

World Development Indicators:. (2014). (The World Bank) Retrieved from

[Link]

Current health expenditure (% of GDP) - Bangladesh, Nepal, Bhutan, India, Maldives,

Sri Lanka, Pakistan, Afghanistan. (2020). Retrieved from

[Link]

BT-IN-MV-LK-PK-AF

Ahmed, S. M. (2015). Bangladesh Health System Revie. S. M. Ahmed, B. B. Alam, I.

Anwar, T. Begum, R. Huque, J. A. M. Khan et al., Bangladesh Health System

Review, A. Naheed and K. Hort, Eds., World Health Organization, Dhaka,

Bangladesh, 2015.

Bank, W. ( 2016). UHC in Africa: A framework for action. Washington, DC: World

Bank Group.

Hamid SA, A. S. (2014). Applied Health Economics and Health Policy, 12, 421-33.

Islam, M. R. (2017). “Inequalities in financial risk protection in Bangladesh: an

assessment of universal health coverage,”. International Journal for Equity in

Health(BMC).

Ke Xu, D. B. ( July 12, 2003). Household catastrophic health expenditure: a

multicountry. THE LANCET •, Vol 362 •.

Laura C Steinhardt, 1. H. (2008). The effect of wealth status on care seeking and health

expenditures in Afghanistan. Oxford University Press in association with The

London School of Hygiene and Tropical Medicine, 24:1–17.

70 | P a g e
 Reference

Nations, U. (2015). Sustainable Development Goals [Internet]. (Sustain. Dev. Goals 17

Goals to Transform ourWorld) Retrieved from

[Link]

Nations, U. (2015). Sustain. Dev. Goals 17 Goals to Transform ourWorld. ( United

Nations) Retrieved from [Link]

development-goals/

Nations, U. (2015). Transforming our world: The 2030 Agenda for Sustainable

Development. New York: United Nations.

Organization, W. H. (2015). Tracking universal health coverage: First Global Monitoring

Report. Geneva, Switzerland: WHO Press.

Organization, W. H. (2015). Tracking universal health coverage: First Global Monitoring

Report. Geneva, Switzerland: WHO Press.

Organization., W. H. (2011). World health statistics 2011. Geneva: World Health

Organization.

planning., B. b. (november,2023). Population and housing census 2022. Bangladesh

bureau of statistics and informatics division ministry of planning.

Rahman MM, G. S. (2013). Health related inancial catastrophe, inequality and chronic

illness in Bangladesh. PLoS One, e56873, 8.

Steinhardt LC, W. H. (2009). The effect of wealth status on care seeking and health

expenditures in Afghanistan. Health Policy Plan, 24:1–17.

Watch, B. H. (2012). Bangladesh health watch 2011 – Moving towards universal health

coverage. Dhaka: James P Grant School of Public Health.

71 | P a g e
 Reference

Welfare., M. o. (2010). Health Bulletin. Dhaka: Government of the People’s Republic of

Bangladesh.

World Health Organization, W. B. (2014). Monitoring progress towards universal health

coverage at country and global levels, Framework measures and targets. Geneva:

World Bank.

Hamid SA, A. S. (2014). Applied Health Economics and Health Policy, 12, 421-33.

Islam, M. R. (2017). “Inequalities in financial risk protection in Bangladesh: an

assessment of universal health coverage,”. International Journal for Equity in

Health(BMC).

Nations, U. (2015). Sustain. Dev. Goals 17 Goals to Transform ourWorld. ( United

Nations) Retrieved from [Link]

development-goals/

[Link]

Nations, U. (2015). Sustain. Dev. Goals 17 Goals to Transform ourWorld. ( United

Nations) Retrieved from [Link]

development-goals/

72 | P a g e
 Glossary

Universal Health Coverage (UHC): It is a healthcare system that ensures all

individuals and communities receive the quality health services they need without
facing financial hardship. It encompasses access to essential healthcare services,
including preventive, promotive, curative, rehabilitative, and palliative care, as well as
access to safe, effective, quality, and affordable essential medicines and vaccines for
all. UHC aims to improve health outcomes, reduce health inequalities, and protect
individuals from the financial risks associated with paying for healthcare services out
of pocket. It is a cornerstone of sustainable development and a fundamental human
right, promoting health and well-being for all.
Out-of-pocket payments (OOP): These are among the most important financing
mechanisms in many health systems, especially in developing countries, adversely
affecting equality and leading vulnerable groups to poverty. Therefore, this scoping
review study was conducted to identify the strategies involving OOP reduction in
health systems.
Catastrophic OOP payments occur when spending on health care is above a certain
threshold that results in financial distress. However, the threshold levels for estimating
catastrophic payments differ in the literature, ranging from 10% to 25% of total
consumption expenditure
Distress financing: It refers to the practice of seeking funds during financial crises or
emergencies, often at high costs or through unfavorable terms. This type of financing
typically arises when individuals or organizations face urgent cash needs due to
unforeseen circumstances such as medical emergencies, business downturns, or
unexpected expenses. Distress financing options may include high-interest loans,
payday loans, credit card cash advances, or borrowing against assets. While providing
immediate relief, distress financing can exacerbate financial problems in the long run
due to the high costs and risks involved, potentially leading to a cycle of debt and
further financial distress if not managed carefully.
Financial hardship: Financial hardship in healthcare can have significant implications
for individuals and families, including delayed or forgone medical care, increased stress
and anxiety, and even bankruptcy in extreme cases. Addressing this issue requires a
comprehensive approach, including policies aimed at reducing healthcare costs,
expanding access to affordable insurance coverage, and providing assistance programs
for those facing financial difficulties related to medical expenses.

73 | P a g e
 Glossary

Coefficient: A coefficient is a multiplicative factor in a polynomial, series, or any

expression, it is typically a number, although it can be any expression. The variables
that appear in the coefficients are generally referred to as parameters in this scenario,
and they must be differentiated from the other variables.
Confidence Interval (CI): A confidence interval (CT) is a sort of interval estimate in
statistic that is generated from the statistics of observed data and may contain the true
value of an unknown population parameter. The interval has a confidence level
associated with it that, in a nutshell, quantifies the amount of certainty that the
parameter is within the interval.
Constant: A constant is a value that remains constant during any mathematical
operation of research.
Degrees of freedom: The number of parameters of a system that can vary
independently is known as degrees of freedom in several scientific domains. A point on
the plane, for example, has two degrees of freedom for translation: its two coordinates;
a non-infinitesimal object on the plane, on the other hand, may have additional degrees
of freedom for orientation. This concept is represented in mathematics as the dimension
of a manifold or an algebraic variety. When degrees of freedom are used instead of
dimensions, it usually signifies that the system's manifold or variety is only implicitly
defined.
Dependent variable: What you measure in the experiment and what is impacted
throughout the experiment is referred to as a dependent variable. A dependent variable
is a variable that takes on its value as a result of a second variable or a system of
variables taking on any set of arbitrary values provided to them.
Healthcare: The prevention, diagnosis and treatment of disease, illness, injury and
other physical and mental disabilities in humans is what health care or healthcare is all
about. Aesthetic health care is provided by allied health professionals (providers or
practitioners). These health-care organizations include and physicians physician
associates professionals Dentistry, midwifery, nursing, medicine, optometry, audiology,
pharmacy, psychology, occupatio nal therapy, therapy and other health professions are
all part of health care. It includes work done in providing primary care, secondary care,
and tertiary, as well as in public health.
Independent variable: An independent variable is a variable that can have any
arbitrary value assigned to it.

74 | P a g e
 Glossary

Healthcare: The prevention, diagnosis, and treatment of disease, illness, injury, and
other physical and mental disabilities in humans is what health care or healthcare is all
about. Aesthetic health care is provided by allied health professionals (providers or
practitioners). These health-care organizations include physicians and physician
[Link] Dentistry, midwifery, nursing, medicine, optometry, audiology,
pharmacy, psychology, occupational therapy, therapy and other health professions are
all part of health care. It includes work done in providing primary care, secondary care,
and tertiary. as well as in public health.
Independent variable: An independent variable is a variable that can have any
arbitrary value assigned to it.
Intercept: In mathematics, the meaning of intercept is to separate or include, as
between two points or lines. The constant term in linear regression analysis, on the
other hand, is known as intercept. It's simply the point on the y-axis where the fitted
line crosses.
Logistic regression: Logistic regression, often known as logit regression or logit mode
in statistics, is a regression model in which the dependent variable (DV) is categorical.
This article focuses on binary dependent variables, which have just two possible values,
such as pass/fail, win/lose, alive/dead, or healthy/sick. Multinomial logistic regression
is used when there are more than two categories, or ordinal logistic regression when the
numerous categories are ordered.
Multiple logistic regressions: When one nominal variable is combined with two or
more measurement variables. The dependent (Y) variable is the nominal variable, and
you're looking at how the independent (X) variables affect the likelihood of getting a
specific value for the dependent variable.
Sensitivity: In various domains, sensitivity (also known as true positive rate, recall, or
likelihood of detection) refers to the percentage of true positives that are accurately
identified as soch (eg. the percentage of sick people who are correctly identified as
having the condition).
Specificity: The proportion of true negatives correctly identified as such is measured
by specificity (also known as the true negative rate) (e.g., the percentage of healthy
people who are correctly identified as not having the condition).
Standard error (SE): The standard deviation of a statistic's sampling distribution or an
estimate of that standard deviation is the standard error (SE) of that statistic (typically
an estimate of a parameter). The standard error of the mean is the term used when the
75 | P a g e
 Glossary

parameter or statistic is the mean (SEM). The sampling distribution of a population

mean is created by taking repeated samples and keeping track of the results. This
results in a distribution of many means, each with its own mean and variance. The
variance of the sampling distribution obtained is equal to the variance of the population
divided by the sample size, as calculated mathematically. Because sample means
cluster more closely around the population mean as sample size grows, this is the case.
As a result, the relationship between the standard error and the standard deviation is
that the standard error equals the standard deviation divided by the square root of the
sample size for every given sample size. To put it another way, the standard error of the
mean is a measure of how far sample means differ from the population mean
Variable: A variable is a quantity (or symbol), which can take various values.
Mathematical model: Mathematical model is a system of mathematical equations
allowing all the known experimental data to be taken into account.
X-axis: Usually the x-axis is drawn horizontal (abscissa), with x increasing to the right.
Y-axis: Usually the y-axis is drawn vertical (ordinate), with y increasing going up.

76 | P a g e