Cross-sectional Study

Presented by
Mohanto Nijaya
PhD Candidate
Student ID: 20217774
Drug Delivery Research Lab
College of Pharmacy
Chosun University
 Learning Objectives

 Definition of the cross-sectional studies

 Define the design of cross-sectional studies
 Describe types of cross-sectional studies
 Identify steps for conducting cross-sectional studies
 Describe the analysis of data in cross-sectional studies
 Common applications of the cross-sectional study
 Explain the biases of cross-sectional studies
 Describe the advantages and disadvantages of cross-sectional
studies
 What is cross-sectional studies

Other names: Transverse study/ Prevalence study/ Cross-sectional

analysis

 A cross-sectional study is a type of observational study, or descriptive research,

that involves analyzing information about a population at a specific point in
time

 More specifically, Cross-sectional studies measure simultaneously the exposure

and health outcome in a given population and in a given geographical area at a
certain time

 It is described as a “snapshot” of a population at a certain point in time because

exposure and outcome are determined simultaneously for each subject

 The aim of this study is to obtain reliable data that make it possible to generate,
robust conclusions, and create new hypotheses that can be investigated with
new research
Diagrammatic representation of cross-
sectional studies

Observational study

Study Population

Time
Past Present Future

Figure 1: Diagrammatic representation of cross-sectional studies

Cross-sectional studies

 Cross-sectional studies can be helpful in determining how many

people are affected by a condition and whether the frequency of
the occurrence varies across groups or population characteristics

 It involves-
 Identifying a defined population at a particular point in time
 Measuring a range of variables on an individual basis
 Measurement of exposure of interest and outcome of interest
is carried out at the same time
 There is no in-built directionality as both exposure and
outcome are present in the study subject for quite same time
Cross-sectional studies don’t have
a direction

Cohort

Exposure Cross-sectional Outcome

Case-control
Design of cross-sectional studies

 Cross-sectional studies measure existing disease and

current exposure levels

 They provide some indication of the relationship between

the disease and exposure or non-exposure
Design of cross-sectional studies

Exposure &
Outcome
Sampling

Exposure &
No outcome
Reference
Sample
population

No exposure
& Outcome
Measurement

No exposure
& No
outcome
 Design of cross-sectional studies

Measure disease and exposure

status simultaneously among
individuals in a well-defined
population at a point in time

Collect data on
exposure and
outcome (e.g.
disease)

Identify Subjects
from population
 Types of Cross-Sectional Studies

Cross-Sectional Studies

Descriptive Analytical

 It provides information about

It yields information about a
the presence and strength of
single variable or about each of
associations between
the number of separate
variables, permitting the
variables in a study population
testing of hypothesis
 Compare the prevalence of
disease in persons with and
without exposure of interest
When to Conduct a Cross-Sectional Study

 To estimate the prevalence of a health condition or

prevalence of a behavior or risk factor

 To learn about characteristics such as knowledge, attitude,

and practices of individuals in a population

 To monitor trends over time with serial cross-sectional

studies
 For example: In Korea, cross-sectional studies done by
the Korean National Health Insurance Service (NHIS)

 Another National example of cross-sectional studies of

great importance is the Korea National Health and
Nutrition Examination Survey (KNHANES)
 Steps in conducting a cross-sectional study

State the criteria for the disease/ condition clearly

Define co-variables to be measured

Examine ethical issues

Define a population of interest (reference or source

population)

Defining inclusion/ exclusion criteria

 Steps in conducting a cross-sectional study

Determine minimum number of sample required

Select study subjects through appropriate sampling

procedure

Define measurement or analysis procedure

Carry out data collection

Clinical examination (Laboratory investigations)

Steps in conducting a cross-sectional study

Clinical records and other documents

Interviews and Questionnaires

Summarize data

Analyze and interpret finding

Report
Analysis of data in cross-sectional
studies

A. Descriptive cross-sectional studies (Measurement of variables)

 For Continuous variables (Measurement Data)

Mean, Standard Deviation, Median & Percentiles

 Nominal Data (Count Data)

Prevalence rates and Proportions

95% confidence intervals (CI) can also be calculated

Analysis of data in cross-sectional
studies

B. Analytical cross-sectional studies (Measurement of association

between variables)

 For Continuous variables (Measurement Data)

Correlation and regression coefficient

 Nominal Data (Count Data)

Odds ratio, Rate ratio (Prevalence ratios) and Exposure ratio
Analysis of data in cross-sectional studies

 In a cross-sectional study we can calculate the prevalence of disease and

the prevalence of exposure, using the 2 X 2 table

Disease No Disease Total

Exposed a b a+b

Not Exposed c d b+d

Total a+c b+d

 Prevalence of disease in exposed compared to non-exposed: a/a+b vs

c/c+d
 Prevalence of exposure in diseased compared to non-diseased: a/a+c vs
b/b+d
 Analysis of data in cross-sectional studies

 Odds ratio:  t-test:

 chi-square test :
Common applications of cross-sectional study

 In community health care- They are usually conducted to

estimate the prevalence of the outcome of interest for a given
population, commonly for the purposes of public health
planning

 In clinical practice and patient care

 In program evaluation- Used to understand the prevalence

of various conditions, treatments, services or other outcomes
and the factors associated with such outcomes

 In acquiring new knowledge- Appropriate for screening

hypotheses because they require relatively shorter time and
fewer resources
Biases in cross-sectional study

 Choice of the sampling frame- when the study participants are

systemically different in their characteristics compared with eligible
participants who were not selected for the study

 Non-response bias

 Information bias

 Observer bias

 Prevalence bias in hospital studies

Advantages of cross-sectional study

 Relatively quick and easy to conduct

 Data on all variables is only collected once

 Able to measure prevalence for all factors under investigation

 Less costly

 Recall bias

 Multiple outcomes and exposures can be studied

 Good for descriptive analyses and for generating hypotheses

Disadvantages of cross-sectional study

 Does not provide an estimate of disease occurrence (incidence)

 No direct estimate of risk possible

 Rare diseases, short duration, high case fatality not detected

 Natural history of disease information minimal

 Prone for biases from selective survival

 Not possible to establish temporality

 Therefore, it is a weak design for proving casualty

References

 Carlson, Melissa DA, and R. Sean Morrison. "Study design, precision, and
validity in observational studies." Journal of palliative medicine 12.1 (2009):
77-82.
 Bland M. An Introduction to Medical Statistics. 3rd Edn. Oxford: Oxford
University Press; 2001.
 Celentano, David D., and Scd Mhs. Gordis Epidemiology. Elsevier, 2018.
 Hulley, Stephen B., ed. Designing clinical research. Lippincott Williams &
Wilkins, 2007.
 Haynes, R. Brian. Clinical epidemiology: how to do clinical practice research.
Lippincott williams & wilkins, 2012.
 The Centre for Evidence-Based Medicine develops, promotes and disseminates
better evidence for healthcare. Study Design. NA. Accessed September 13,
2019: https://www.cebm.net/2014/04/study-designs/
 Pine CM, Pitts NB, Nugent ZJ. British Association for the Study of Community
Dentistry (BASCD) guidance on sampling for surveys of child dental health. A
BASCD coordinated dental epidemiology programme quality standard.
Commun Dent Health 1997; 14(Suppl 1):S10–S17
Research article about a
cross-sectional study
A cross-sectional study of distress: A cancer
response

Hanna Ekman, Alexandra Pettersson, Liselotte Jakobsson, Pernilla Garmy

Introduction

 All people with cancer experience some degree of distress as a

direct result of the disease, regardless of which phase of the
disease the patient is in

 Cancer-related distress can include feelings about-

 Vulnerability
 Grief and fear
 Some problems that can be disabling, such as depression,
anxiety, panic, social isolation and existential and spiritual
crises
 Background

 There are large variations in the degree of distress

experienced by patients and the areas that affect distress
including gender, age and marital status

 Research shows that more women than men experience a

Distress high degree of distress
Thermometer
 In general, younger patients have a higher degree of
distress regardless of gender, whereas younger women
experience higher distress compared with older women

 There is also a variation between those who are married

and young and middle-aged who experience higher
distress than those who are older

The Distress Thermometer is an assessment tool

recommended for the assessment of cancer
rehabilitation needs internationally
Design
Method This cross-sectional study aimed to investigate the
cancer rehabilitation experiences of working-age
cancer survivors

This cross-sectional survey was approved by the Regional Ethical

Review Board in Lund, Sweden prior to the recruitment of participants

 The study was performed at the Cancer Rehabilitation Clinic on a

medium size hospital in southern Sweden which is cooperated with the
Swedish Social Insurance Agency in five neighboring municipalities
with between 7000–85,000 inhabitants per municipality

 A questionnaire and information were distributed in October 2016

via mail to all patients over the age of 18 years who had a cancer
diagnosis and were enrolled in the Swedish Social Insurance Agency in
the included municipalities from January 2013– April 2016 (N= 384)
Design
Method

 One reminder was sent after 4 weeks

 Written informed consent was requested and was completed

by hand and returned together with the questionnaire in a
provided stamped reply envelope

 A total of 168 patients answered the questionnaire (response

rate of 44%)
 Design
Survey questionnaire

Background data
The questionnaire consisted of questions about background data:
 Gender
 Age
 Civil state
 Children
 Birth Country
 Other diseases
 Ongoing treatment
 Long-term treatment
 Psychiatric disease
 Addictions
 Design
The distress thermometer

 The Distress Thermometer is an assessment tool that measures

distress and is provided by the National Comprehensive Cancer
Network (NCCN)Gender

 The form contains two parts:

 The thermometer- where the patient estimates the degree
of distress in the last week on a visual analog scale from
0–10 and a problem list with 39 common problem areas
related to cancer and its treatment

 By understanding the underlying causes of distress, the

health professional can better apply targeted
interventions and thus reduce the distress of people with
cancer
Analysis

 Data were analyzed by descriptive and analytical statistics with means,

standard deviations, and percentages

 A t-test was used to investigate if there were age or gender differences

between those who had ongoing treatment versus those who had completed
treatment

 Bivariate analyses were used with a chi-squared test between those who had
ongoing and completed treatment in relation to the issues of the Distress
Thermometer

 Bivariate analysis and multiple logistic regression were used to compare the
0–3 and 4–10 distress groups and identify risk factors for distress

 The variables for the risk factors associated with distress agreed with practical
experience

 The Hosmer Lemeshow and the Nagelkerke R2 tests were reported for the
logistic regression analysis. IBM SPSS version 22 was used for the analysis
 Result
Prevalence of different sample
(Descriptive Cross-Sectional analysis)

Description of the sample (N = 161)

 A total of 168 persons reported on having distress or not
 161 persons reported on ongoing or completed cancer treatment

Total of 161 people reported on ongoing or completed treatment, among

them, 105 people reported on completed treatment. So, What is the
prevalence of a completed treatment people?
=Number of people who reported on completed treatment/ Total
population X 100
= 105 /161 X 100
= 65%

Cont’d
 Prevalence of different sample
(Descriptive Cross-Sectional analysis)

Another calculation
Total of 161 people reported on ongoing or completed treatment, among
them, 56 people reported on ongoing treatment. So, What is the prevalence
of a ongoing treatment people?
=Number of people who reported on ongoing treatment/ Total
population X 100
= 56 /161 X 100
= 35%

Cont’d
 Table 1: Description of the sample (N = 161)
Sample description Ongoing Completed treatment p-value
treatment (N=56, (N= 105, 65%)
35%)
Age in years (means, SD) 56.18 (8.6 SD) 56.90 (9.2 SD) .634a
Sex
Women, N (%) 38 (67.8%) 72 (68.5%)
Men, N (%) 18 (32.2%) 33 (31.5%)
Country of birth .165b
Sweden, N (%) 50 (91%) 101 (96%)
Nordic countries, N (%) 2 (3.6%) 0 (0%)
Europe, N (%) 3 (5.4%) 3 (3%)
Outside Europe, N (%) 0 (0%) 1 (1%)
Living with a partner, N (%) 48 (89%) 78 (76%) .061b
Parent of children, N (%) 50 (91%) 94 (91%) .719b
Children under the age of 18, N (%) 15 (30%) 19 (20%) .400b

 Note: Participants were divided based on whether they had ongoing treatment or had
completed cancer treatment.
 Missing cases: 0%–7%. Abbreviation: SD, standard deviation.
 a
t test. bChi-squared test.
 Prevalence of various problem areas that affected distress
(Descriptive Cross-Sectional analysis)

 A high degree of distress was defined as >3 according to the Distress

Thermometer (Distress “0-3” and “4-10”)
 Table 2 shows the prevalence of various problem areas that
affected distress and the stated degree among those with ongoing
or completed treatment

 Of those who had completed their treatment, 29% still had a high degree of
distress-

**Total of 105 people completed their treatment, among them, 30 people

had a high degree of distress (4-10). So, What is the prevalence of a high
degree of distress (4-10)?
=Number of people who had a high degree of distress / Total population
X 100
= 30 /105 X 100
= 29%

Cont’d
 Prevalence of various problem areas that affected distress
(Descriptive Cross-Sectional analysis)

 Of those with ongoing treatment, 48% had a high degree of distress-

** Total of 56 people had ongoing treatment, among them, 27 people had

a high degree of distress (4-10). So, What is the prevalence of a high
degree of distress (4-10)?
=Number of people who had high degree of distress / Total
population X 100
= 27 /56 X 100
= 48%

Cont’d
 Table 2: Description of the Distress Thermometer (N = 161); Prevalence of various
problem areas that affected distress (Descriptive Cross-Sectional analysis)

Distress thermometer Ongoing treatment Completed p-valuea

(N=56) treatment (N=105)
Distress 0–3, N (%) 29 (52%) 75 (71%) .013
Distress 4–10, N (%) 27 (48%) 30 (29%)
Practical problems
Child care, N (%) 0 (0%) 1 (1%) .464
Housing, N (%) 3 (5%) 3 (3%) .425
Insurance/financial, N (%) 9 (16%) 9 (9%) .150
Transportation, N (%) 4 (7%) 4 (4%) .354
Work/School, N (%) 10 (18%) 13 (12%) .344
Treatment decisions, N (%) 8 (14%) 5 (5%) .035
Family problems
Dealing with children, N (%) 3 (5%) 4 (4%) .647
Dealing with partner N (%) 5 (9%) 6 (6%) .441
Ability to have children, N (%) 2 (4%) 1 (1%) .242
Family health issues, N (%) 18 (32%) 21 (20%) .087

Note: p-values < .05 were considered significant and are marked in bold
Cont’d
 Table 2: Description of the Distress Thermometer (N = 161); Prevalence of various
problem areas that affected distress (Descriptive Cross-Sectional analysis)

Distress thermometer Ongoing treatment Completed p-valuea

(N=56) treatment (N=105)
Emotional problems
Depression, N (%) 9 (16%) 8 (8%) .096
Fear, N (%) 18 (32%) 11 (10.5%) .001
Nervousness, N (%) 13 (23%) 16 (15%) .210
Sadness, N (%) 24 (43%) 23 (22%) .005
Worry, N (%) 35 (62.5%) 33 (31%) <.001
Loss of interest in usual activities, 13 (23%) 15 (14%) .155
N (%)
Spiritual/ religious concerns, N (%) 4 (7%) 2 (2%) .095
Physical problems
Appearance, N (%) 8 (14%) 9 (9%) .261
Bathing dressing, N (%) 4 (7%) 1 (1%) .031
Breathing, N (%) 11 (20%) 2 (2%) .000
Changes in urination, N (%) 7 (12.5%) 5 (5%) .075
Constipation, N (%) 8 (14%) 8 (8%) .178

Note: p-values < .05 were considered significant and are marked in bold
Cont’d
 Table 2: Description of the Distress Thermometer (N = 161); Prevalence of various
problem areas that affected distress (Descriptive Cross-Sectional analysis)

Distress thermometer Ongoing treatment Completed p-valuea

(N=56) treatment (N=105)
Diarrhoea, N (%) 10 (18%) 6 (6%) .014
Eating, N (%) 5 (9%) 8 (8%) .771
Fatigue, N (%) 38 (68%) 46 (44%) .004
Feeling swollen, N (%) 19 (34%) 14 (13%) .002
Fevers, N (%) 2 (4%) 4 (4%) .939
Getting around, N (%) 20 (36%) 23 (22%) .059
Indigestion, N (%) 7 (12.5%) 4 (4%) .037
Memory/concentration, N (%) 25 (45%) 31 (30%) .055
Mouth sores N (%) 5 (9%) 6 (6%) .441
Nausea, N (%) 12 (21%) 5 (5%) .001
Skin dry/itchy, N (%) 23 (41%) 21 (20%) .004
Sleep, N (%) 20 (36%) 36 (34%) .856
Substance use, N (%) 0 (0%) 0 (0%)
Tingling in hands and feet, N (%) 25 (45%) 32 (30.5%) .073

Note: p-values < .05 were considered significant and are marked in bold
Prevalence of various problem areas that affected distress
(Descriptive Cross-Sectional analysis)

 Patients experienced several problems after treatment had

ended such as fatigue (44%), sleep problems (34%), worries
(31%), pain (31%), tingling in hands and feet (31%), and
problems with memory/concentration (30%)

 Patients with financial/insurance problems had significantly

higher distress than those who did not have these problems
 Bivariate analysis of the risk factors linked to the high
degree of distress ≥ 4 (Descriptive Cross-Sectional)

 The degree of distress was dichotomized into groups of low (0–3)

and high (4–10) distress

 The bivariate analysis (it involves the analysis of two variables)

explored whether there was any significant correlation between the
degree of distress and age, sex, living with partner or alone, having
children under 18 years old, living in the home, other bodily
diseases, mental illness, completed treatment, ongoing treatment or
some form of long-term treatment (Table 3)

Cont’d
 Table 3: Bivariate analysis of the risk factors linked to the high
degree of distress ≥ 4 (N = 168)

Variables Distress 0–3 Distress 4–10 p-valuea

(N = 110) (N = 58)
Age under: 40 4 (4%) 3 (5.4%) .085
Age: 40–60 53 (50%) 28 (50%)
Age over: 60 49 (46.2%) 25 (45%)
Sex: men 38 (35%) 17 (30%) .470
Sex: women 71 (65%) 41 (70%)
Living with partner 84 (80%) 45 (82%) .784
Living alone 21 (20%) 10 (18%)
Children under 18 living home 18 (19%) 16 (29%) .168
Insurance/finance 6 (6%) 13 (22%) .001
Have or have had other physical 40 (38%) 25 (45%) .423
diseases
Have or have had mental illness 5 (5%) 4 (6.9%) .532
Have completed the cancer treatment 75 (71%) 30 (29%) .013

Have ongoing cancer treatment 29 (52%) 27 (48%)

Have any long-term treatment 37 (35%) 32 (57%) .006

Note: p-values < .05 were considered significant and are marked in bold
Missing cases: 0%–9.5%. aChi-squared test. Cont’d
 Bivariate analysis of the risk factors linked to the high
degree of distress ≥ 4 (Descriptive Cross-Sectional)

 The analysis shows that there was a significant relationship (p

< .05) between insurance/finance problems and a high degree of
distress. There was also a significant relationship (p < .05) between
the degree of distress and cancer treatment (Table 3)
 Multiple logistic regression analysis of the risk factors
linked to the high degree of distress (Analytical Cross-Sectional)

 The multiple regression analysis describes whether there was any

significant relationship between distress ≥ 4 and factors such as age, sex,
living with partner or alone, having children under 18 years old,
insurance/finance, other bodily diseases, mental illness, completed
treatment and whether they underwent any form of long-term treatment

Cont’d
 Table 4: Multiple logistic regression analysis of the risk factors
linked to the high degree of distress ≥ 4 (N = 144)

95% confidence
Variables Odds ratio (OR) p-value
intervals (CIs) for OR

Age 1.007 0.953–1.064 .811

Sex 1.227 0.520–2.893 .641
Living with a partner or 1.297 0.474–3.547 .613
living alone
Children under 18 living 1.225 0.651–2.306 .530
home
Insurance/ finance 4.953 1.312–18.694 .018
Have or have had other 1.116 0.498–2.502 .789
bodily diseases
Have or have had a 0.648 0.143–2.937 .573
mental illness
Have completed the 1.517 0.630–3.653 .353
cancer treatment
Have any long-term 0.612 0.265–1.411 .249
treatment

Note: p-values < .05 were considered significant and are marked in bold
 Multiple logistic regression analysis of the risk factors
linked to the high degree of distress (Analytical Cross-Sectional)

 The multiple logistic regression analysis showed that there

was a significant relationship (p < .05) between the degree
of distress and insurance/finance problems

 Ongoing or completed treatment was not a significant factor

in this analysis

Cont’d
 Study limitations

 The study was cross-sectional, which means it captures a snapshot of

prevailing conditions at a specific time. Thus, this study cannot address
how the conditions change over a longer period

 The response rate was 44% after the reminder. The group actively
decided not to distribute a second reminder because the participants had
cancer and might be preoccupied or they might also have not received the
questionnaire because they may have also moved or have been hospitalized

 The low participation might also affect the generalizability of the

result. This is a sensitive subject, and it is unclear if the dropout needs to
be high for cancer studies
 Conclusion

 This study showed that distress is a common problem regardless of whether

people are undergoing cancer treatment or have completed treatment

 This distress can have an impact on their work life, which is increasingly
relevant due to older retirement ages and should be studied from a social
perspective

 From the oncology perspective, these results increase understanding of

which patients are at risk of experiencing a higher degree of distress and
thus have poorer conditions for compliance with oncological treatment

 In addition to the patients’ benefits, the data can also have a socio-
economic benefit because many known consequences of cancer treatment
can be prevented if they are identified early. This will reduce the burden on
the healthcare system.
Thank You