This document discusses different types of research designs, including observational and intervention designs. It focuses on non-intervention designs like case reports, case series, and cross-sectional studies. Case reports describe the occurrence, diagnosis, treatment and follow-up of an individual patient, especially unusual cases. Case series describe aspects of a disease or treatment by following a group of patients with common characteristics. Both case reports and case series are useful for generating hypotheses but have limitations due to lack of a control group.
An epidemiological experiment in which subjects in a population are randomly allocated into groups, usually called study and control groups to receive and not receive an experimental preventive or therapetuic procedure, maneuver, or intervention .
Study designs, Epidemiological study design, Types of studiesDr Lipilekha Patnaik
Study design, Epidemiological study designA study design is a specific plan or protocol
for conducting the study, which allows the investigator to translate the conceptual hypothesis into an operational one.
Observational analytical study: Cross-sectional, Case-control and Cohort stu...Prabesh Ghimire
This presentation provides overview of three observational analytical studies: cross-sectional study design, case-control study design and cohort study design
A meta-analysis is the use of statistical methods to summaries the results of the studies. Meta-analyses are conducted to assess the strength of evidence present on a disease and treatment. The results of a meta-analysis can improve precision of estimates of effect, answer questions not posed by the individual studies, settle controversies arising from apparently conflicting studies, and generate new hypotheses. In particular, the examination of heterogeneity is vital to the development of new hypotheses.
An epidemiological experiment in which subjects in a population are randomly allocated into groups, usually called study and control groups to receive and not receive an experimental preventive or therapetuic procedure, maneuver, or intervention .
Study designs, Epidemiological study design, Types of studiesDr Lipilekha Patnaik
Study design, Epidemiological study designA study design is a specific plan or protocol
for conducting the study, which allows the investigator to translate the conceptual hypothesis into an operational one.
Observational analytical study: Cross-sectional, Case-control and Cohort stu...Prabesh Ghimire
This presentation provides overview of three observational analytical studies: cross-sectional study design, case-control study design and cohort study design
A meta-analysis is the use of statistical methods to summaries the results of the studies. Meta-analyses are conducted to assess the strength of evidence present on a disease and treatment. The results of a meta-analysis can improve precision of estimates of effect, answer questions not posed by the individual studies, settle controversies arising from apparently conflicting studies, and generate new hypotheses. In particular, the examination of heterogeneity is vital to the development of new hypotheses.
Factors Associated with Anemia among Pregnant Women of Underprivileged Ethnic...Prabesh Ghimire
Abstract
Background. This study aims at determining the factors associated with anemia among pregnant women of underprivileged ethnic groups attending antenatal care at the provincial level hospital of Province 2. Methods. A hospital-based cross-sectional study was carried out in Janakpur Provincial Hospital of Province 2, Southern Nepal. 287 pregnant women from underprivileged ethnic groups attending antenatal care were selected and interviewed. Face-to-face interviews using a structured questionnaire were undertaken. Anemia status was assessed based on hemoglobin levels determined at the hospital’s laboratory. Bivariate and multiple logistic regression analyses were used to identify the factors associated with anemia. Analyses were performed using IBM SPSS version 23 software. Results. The overall anemia prevalence in the study population was 66.9% (95% CI, 61.1–72.3). The women from most underprivileged ethnic groups (Terai Dalit, Terai Janajati, and Muslims) were twice more likely to be anemic than Madhesi women. Similarly, women having education lower than secondary level were about 3 times more likely to be anemic compared to those with secondary level or higher education. Women who had not completed four antenatal visits were twice more likely to be anemic than those completing all four visits. The odds of anemia were three times higher among pregnant women who had not taken deworming medication compared to their counterparts. Furthermore, women with inadequate dietary diversity were four times more likely to be anemic compared to women having adequate dietary diversity. Conclusions. The prevalence of anemia is a severe public health problem among pregnant women of underprivileged ethnic groups in Province 2. Being Dalit, Janajati, and Muslim, having lower education, less frequent antenatal visits, not receiving deworming medication, and having inadequate dietary diversity are found to be the significant factors. The present study highlights the need of improving the frequency of antenatal visits and coverage of deworming program in ethnic populations. Furthermore, promoting a dietary diversity at the household level would help lower the prevalence of anemia. The study findings also imply that the nutrition interventions to control anemia must target and reach pregnant women from the most-marginalized ethnic groups and those with lower education
Factors Associated with Enrolment of Households in Nepal’s National Health In...Prabesh Ghimire
Abstract
Background: Nepal has made remarkable efforts towards social health protection over the past several years. In 2016, the Government of Nepal introduced a National Health Insurance Program (NHIP) with an aim to ensure equitable and universal access to healthcare by all Nepalese citizens. Following the first year of operation, the scheme has covered 5 percent of its target population. There are wider concerns regarding the capacity of NHIP to achieve adequate population coverage and remain viable. In this context, this study aimed to identify the factors associated with enrolment of households in the NHIP.
Methods: A cross-sectional household survey using face to face interview was carried out in 2 Palikas (municipalities) of Ilam district. 570 households were studied by recruiting equal number of NHIP enrolled and non-enrolled households. We used Pearson’s chi-square test and binary logistic regression to identify the factors associated with household’s enrolment in NHIP. All statistical analyses were performed using IBM SPSS version 23 software.
Results: Enrolment of households in NHIP was found to be associated with ethnicity, socio-economic status, past experience of acute illness in family and presence of chronic illness. The households that belonged to higher socio-economic status were about 4 times more likely to enrol in the scheme. It was also observed that households from privileged ethnic groups such as Brahmin, Chhetri, Gurung, and Newar were 1.7 times more likely to enrol in NHIP compared to those from underprivileged ethnic groups such as janajatis (indigenous people) and dalits (the oppressed). The households with illness experience in 3 months preceding the survey were about 1.5 times more likely to enrol in NHIP compared to households that did not have such experience. Similarly, households in which at least one of the members was chronically ill were 1.8 times more likely to enrol compared to households with no chronic illness.
Conclusion: Belonging to the privileged ethnic group, having a higher socio-economic status, experiencing an acute illness and presence of chronically ill member in the family are the factors associated with enrolment of households in NHIP. This study revealed gaps in enrolment between rich-poor households and privileged-underprivileged ethnic groups. Extension of health insurance coverage to poor and marginalized households is therefore needed to increase equity and accelerate the pace towards achieving universal health coverage.
Recent Advances in Evidence Based Public Health PracticePrabesh Ghimire
This product is the result of compilation from various sources. I acknowledge all direct and indirect sources although they have not been mentioned explicitly in the document.
Development of test instruments
Includes information about:
Methods of collecting information
Interview techniques and tools
Observation: concept and observation checklist
This is the product of compilation from various sources. I would like to acknowledge all direct and indirect sources although they have not been mentioned explicitly within the document.
This product is the result of compilation from various sources. I would like to acknowledge all direct and indirect sources, although they have not been explicitly mentioned within the document.
This product is the result of compilation from various sources. I acknowledge all direct and indirect sources although they have not been mentioned explicitly in the document.
New Organogram of Nepalese Health System (Please check the updated slides on ...Prabesh Ghimire
This slide has been updated to accommodate the recent changes. Please check the following link for the updated presentation:
https://www.slideshare.net/PrabeshGhimire/organogram-organization-structure-of-nepalese-health-system-updated-nov-2021
Bilateral and Multilateral Organizations in NepalPrabesh Ghimire
Declaration: The materials incorporated in this document have come from variety of sources and compiler bears no responsibilities for any information contained herein. The compiler acknowledges all the sources although references have not been explicitly cited for all the contents in this document.
Declaration: The materials incorporated in this document have come from variety of sources and compiler bears no responsibilities for any information contained herein. The compiler acknowledges all the sources although references have not been explicitly cited for all the contents in this document.
International Non Government Organizations (INGOs) in NepalPrabesh Ghimire
Declaration: The materials incorporated in this document have come from variety of sources and compiler bears no responsibilities for any information contained herein. The compiler acknowledges all the sources although references have not been explicitly cited for all the contents in this document.
Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...Subhajit Sahu
Abstract — Levelwise PageRank is an alternative method of PageRank computation which decomposes the input graph into a directed acyclic block-graph of strongly connected components, and processes them in topological order, one level at a time. This enables calculation for ranks in a distributed fashion without per-iteration communication, unlike the standard method where all vertices are processed in each iteration. It however comes with a precondition of the absence of dead ends in the input graph. Here, the native non-distributed performance of Levelwise PageRank was compared against Monolithic PageRank on a CPU as well as a GPU. To ensure a fair comparison, Monolithic PageRank was also performed on a graph where vertices were split by components. Results indicate that Levelwise PageRank is about as fast as Monolithic PageRank on the CPU, but quite a bit slower on the GPU. Slowdown on the GPU is likely caused by a large submission of small workloads, and expected to be non-issue when the computation is performed on massive graphs.
Techniques to optimize the pagerank algorithm usually fall in two categories. One is to try reducing the work per iteration, and the other is to try reducing the number of iterations. These goals are often at odds with one another. Skipping computation on vertices which have already converged has the potential to save iteration time. Skipping in-identical vertices, with the same in-links, helps reduce duplicate computations and thus could help reduce iteration time. Road networks often have chains which can be short-circuited before pagerank computation to improve performance. Final ranks of chain nodes can be easily calculated. This could reduce both the iteration time, and the number of iterations. If a graph has no dangling nodes, pagerank of each strongly connected component can be computed in topological order. This could help reduce the iteration time, no. of iterations, and also enable multi-iteration concurrency in pagerank computation. The combination of all of the above methods is the STICD algorithm. [sticd] For dynamic graphs, unchanged components whose ranks are unaffected can be skipped altogether.
Adjusting primitives for graph : SHORT REPORT / NOTESSubhajit Sahu
Graph algorithms, like PageRank Compressed Sparse Row (CSR) is an adjacency-list based graph representation that is
Multiply with different modes (map)
1. Performance of sequential execution based vs OpenMP based vector multiply.
2. Comparing various launch configs for CUDA based vector multiply.
Sum with different storage types (reduce)
1. Performance of vector element sum using float vs bfloat16 as the storage type.
Sum with different modes (reduce)
1. Performance of sequential execution based vs OpenMP based vector element sum.
2. Performance of memcpy vs in-place based CUDA based vector element sum.
3. Comparing various launch configs for CUDA based vector element sum (memcpy).
4. Comparing various launch configs for CUDA based vector element sum (in-place).
Sum with in-place strategies of CUDA mode (reduce)
1. Comparing various launch configs for CUDA based vector element sum (in-place).
Chatty Kathy - UNC Bootcamp Final Project Presentation - Final Version - 5.23...John Andrews
SlideShare Description for "Chatty Kathy - UNC Bootcamp Final Project Presentation"
Title: Chatty Kathy: Enhancing Physical Activity Among Older Adults
Description:
Discover how Chatty Kathy, an innovative project developed at the UNC Bootcamp, aims to tackle the challenge of low physical activity among older adults. Our AI-driven solution uses peer interaction to boost and sustain exercise levels, significantly improving health outcomes. This presentation covers our problem statement, the rationale behind Chatty Kathy, synthetic data and persona creation, model performance metrics, a visual demonstration of the project, and potential future developments. Join us for an insightful Q&A session to explore the potential of this groundbreaking project.
Project Team: Jay Requarth, Jana Avery, John Andrews, Dr. Dick Davis II, Nee Buntoum, Nam Yeongjin & Mat Nicholas
3. Types of Research Design
Non-Intervention
Observational
Design
Intervention Design
Population Based Individual Based
Descriptive
(Health
Survey)
Ecological
Study
Descriptive
Case Report/
Case Series
Analytical
Cross-sectional or
Prevalence Study
Case- Control Study Cohort Study
Randomized
Control Trial or
Clinical Trial
Non-randomized
Quasi Experimental
Field Trial
Cross-over
Design
Parallel
Design
Research
Designs
Prabesh Ghimire, MPH 3
4. Types of Research Design
• Observational/ Non-Intervention Design
• Observe both exposures and outcomes
• Intervention Design
• Assign exposures
• Observe outcomes
Prabesh Ghimire, MPH 4
6. Observational Design
• Allows nature to take its course
• The investigator observes/ measures but does not intervene.
• Types
• Descriptive (Case report, case series, ecological, cross-sectional)
• Analytical (Cross-sectional, case-control, cohort)
Prabesh Ghimire, MPH 6
8. Case Report
• A case report is a detailed description of disease occurrence,
diagnosis, treatment, response to treatment, and follow-up after
treatment of an individual person.
• Case reports usually describe an unusual or novel occurrence
and as such, remain one of the cornerstones of public health
progress and provide many new ideas in public health.
• Unusual features of the case may suggest a new hypothesis
about the causes or mechanisms of disease.
Prabesh Ghimire, MPH 8
9. Case Report
Case reports often describe:
• Unique cases that cannot be explained by known diseases or
syndromes
• Cases that show an important variation of a disease or
condition
• Cases that show unexpected events that may yield new or
useful information
• Cases in which one patient has two or more unexpected
diseases or disorders
Prabesh Ghimire, MPH 9
10. Reasons for preparing case report
Case reports are prepared to keep record of
• an unexpected association between diseases or symptoms;
• an unexpected event in the course observing or treating a
patient;
• findings that shed new light on the possible pathogenesis of a
disease or an adverse effect;
• unique or rare features of a disease;
• unique therapeutic approaches; variation of anatomical
structures.
Prabesh Ghimire, MPH 10
11. Case Report
• Case reports are considered the lowest level of evidence, but
they are also the first line of evidence, because they are where
new issues and ideas emerge.
• If multiple case reports show something similar, the next step
might be a case-control study to determine if there is a
relationship between the relevant variables.
Prabesh Ghimire, MPH 11
12. Case Report
Case report should provide the following case details
• Case description (socio-demographic information)
• Case history
• Physical examination results
• Results of pathological tests and other investigations
• Treatment plan
• Expected outcome of the treatment plan
• Actual outcome
Prabesh Ghimire, MPH 12
13. Case Report
Strengths
• Can help in the identification of new trends or diseases
• Can help detect new drug side effects and potential uses
(adverse or beneficial)
• Educational -a way of sharing lessons learned
• Identifies rare manifestations of a disease (for example in covid-
19)
Prabesh Ghimire, MPH 13
14. Case Report
Limitations
• Cases may not be generalizable
• Not based on systematic studies
• Causes or associations may have other explanations
• Can be seen as emphasizing the bizarre or focusing on
misleading elements
Prabesh Ghimire, MPH 14
16. Case Series
• A case series is a descriptive study that follows a group of
patients with common characteristics used to describe some
clinical, pathophysiological or operational aspects of a disease,
treatment or diagnostic procedures.
• The primary purpose of a case series is generation of
hypotheses that subsequently can be tested in studies of
greater methodological rigor.
Prabesh Ghimire, MPH 16
17. Case Series
• It is most useful for describing the potential effectiveness of
new interventions, for describing the effectiveness of
interventions on unusual diagnoses, and for describing
unusual responses (either good or bad) to interventions.
• Case series can be conducted retrospectively or prospectively.
Prabesh Ghimire, MPH 17
18. When to consider a Case Series
• When a more cautious description of interventions in several
settings in required.
• To report on novel diagnostic or therapeutic strategies,
particularly when the option of waiting for comparative evidence
is considered unacceptable.
Prabesh Ghimire, MPH 18
19. Importance of Case Series
Clinical case-‐ series are of value in public health field for:
• Studying predictive symptoms, signs, and tests.
• Creating case definitions
• Clinical education, audit, and research
• Health services research
• Establishing safety profiles
Prabesh Ghimire, MPH 19
20. Types of Case Series
On the basis of recruitment
Consecutive case series:
• Includes all eligible patients identified by the researchers during the
study period.
• The patients are treated in the order in which they are identified.
• Consecutiveness increases the quality of the case series.
Non-consecutive case series:
• Includes some, but not all, of the eligible patients identified by the
researchers during the study period.
Prabesh Ghimire, MPH 20
21. Types of Case Series
On the basis of sampling
Exposure-based sampling
• Include all patients treated and have specific outcomes or adverse events.
• Sampling is based on both a specific outcome and presence of a specific
exposure.
Outcome-based sampling
• Includes patients with the specific outcome regardless of exposure.
• Thus neither absolute risk nor relative risk can be calculated.
• Selection is based only on a specific outcome, and data are collected on
previous exposures.
Prabesh Ghimire, MPH 21
22. Designing a good case series
Research Question
• The study question should list its study population, the intervention,
and the primary outcome.
Setting
• Select a suitable observation period and identify cases with events in
this period.
• It may be tempting to include patients seen over a large period of
time to increase sample size.
• However, the use of a short inclusion period minimizes known and
unknown changes over time in co-interventions, prognosis, and even
in the intervention under study
Prabesh Ghimire, MPH 22
23. Designing a good case series
Number of Cases
• The general number of cases reported in a case series range from
20 to 50.
• But may vary from as few as 2 or 3 to as many as more than 100 or
even thousands.
Data collection
• Reports of case series usually contain detailed information about the
individual patients.
• This includes demographic information (for example, age, gender,
ethnic origin) and information on diagnosis, treatment, response to
treatment, and follow-up after treatment
Prabesh Ghimire, MPH 23
24. Designing a good case series
What
• The diagnosis or case definition should be clear and applied
equally to all individuals in the series.
• The case definition should mention the inclusion and exclusion
criteria, which should be based on widely used validated
definitions.
• When: The date when the disease or death occurred (time).
• Where: The place where the person lived, worked etc (place).
Prabesh Ghimire, MPH 24
25. Designing a good case series
Who
• The characteristics of the population (person).
• Noting the socio-demographic characteristics of a series of
cases, as well as the temporal and spatial distributions can
sometimes provide a clue to risk factors and hence help
generate a hypothesis.
• This can be tested subsequently with more elaborate analytic
studies.
Prabesh Ghimire, MPH 25
26. Designing a good case series
• A detailed description of the intervention and the co-intervention
should be stated. This will ensure repeatability of the study by
other investigators.
• It is very important to thoroughly describe co-interventions (for
example, physical therapy)
• The most important outcomes in care are those that measure
patient satisfaction, relief of symptoms, and a feeling of well-
being.
• An example is the Short Form-36 questionnaire, which not only
measures physical function but also mental well-being.
Prabesh Ghimire, MPH 26
27. Designing a good case series
Methods of data collection
• The method of data acquisition (telephone interview, clinical
measurement, or chart review) should be addressed in the study
report
Analysis
• Only descriptive statistics should be used.
• Findings can be presented as proportions (%) of the study
populations with the outcome, confidence intervals; means,
standard deviations for continuous variables
• No comparative tests yielding p values should be done.
Prabesh Ghimire, MPH 27
28. Designing a good case series
Reporting
• A statement of the external validity of the obtained data should be
given. This includes patient characteristics and completeness of
follow-up.
• The follow-up rates and reasons for loss to follow-up should be
stated.
• No absolute conclusions on the studied treatment should be stated.
As mentioned before, the lack of a comparison group prohibits any
hypothesis from being tested.
• Valid conclusion: “Patients treated by treatment X showed good outcome Y
after Z months of follow-up.”
• Stating that “treatment X is better than treatment Y” or even that “treatment X
is effective” would be invalid.
Prabesh Ghimire, MPH 28
29. Strengths and Limitations
Strengths
• High external validity: the study results are closer to those obtained in
routine clinical practice and may, therefore, be considered more relevant.
• It could be useful when a randomized controlled trial is not appropriate or
possible.
• No interference in the treatment decision process
• Wide range of patients can be studied
• In-expensive
• Conduction of study take little time
• Useful for hypothesis generation
• Informative for very rare disease with few established risk factors.
Prabesh Ghimire, MPH 29
30. Strengths and Limitations
Limitations
• Lack of a control (or comparison) group
• Lack of a denominator to calculate rates of disease.
• Causal inferences cannot be made
• Data collection often incomplete
• Susceptible to bias (selection bias, measurement bias)
Prabesh Ghimire, MPH 30
31. For further reading
• https://asploro.com/what-is-case-
series/#:~:text=Non%2DConsecutive%20Case%20Series%3A%20%5B,quality%20of%20
the%20case%20series.
• Mathes, T., & Pieper, D. (2017). Clarifying the distinction between case series and cohort
studies in systematic reviews of comparative studies: potential impact on body of
evidence and workload. BMC medical research methodology, 17(1), 107.
https://doi.org/10.1186/s12874-017-0391-8
• Abu-Zidan, F. M., Abbas, A. K., & Hefny, A. F. (2012). Clinical "case series": a concept
analysis. African health sciences, 12(4), 557–562.
• https://www.researchgate.net/publication/327449197_What_is_case_series
Prabesh Ghimire, MPH 31
33. Ecological Study
• Observational study in which data are analyzed at the
population or community level rather than individual level.
• Disease rates and exposures are measured in each of a series
of populations and their relations is examined.
• Often the information about disease and exposure is abstracted
from published statistics and therefore does not require
expensive or time consuming data collection.
• In ecological studies health outcomes are aggregates of
individual health data. E.g. prevalence, incidence, rate of
diseases.
Prabesh Ghimire, MPH 33
34. Years of education Teenage pregnancy
(Yes/No)
Prevalence/ Rate of
Teenage Pregnancy
Association
Avg. no. of years of
education
Prabesh Ghimire, MPH 34
35. Purpose of ecological study
• To monitor population health so that public health strategies may be
developed and directed.
• To make large scale comparisons, e.g. comparisons between
countries;
• To study the relationship between population-level exposure to risk
factors and diseases or in-order to look at the contextual effect of
risk factors on the population
• When disease under investigation is rare, requiring aggregation of
data for any analysis to be carried out.
• When measurement at individual level are not available. E.g.
confidentiality might require that individuals are anonymized by
aggregation of data to small area level.
Prabesh Ghimire, MPH 35
36. Types of Ecological Study
Geographical
• One common approach is to look for geographical correlations
between disease incidence or mortality and the prevalence of
risk factors.
• For example, mortality from coronary heart disease in local
authority areas of England and Wales has been correlated with
neonatal mortality in the same places 70 and more years
earlier.
• This observation generated the hypothesis that coronary heart
disease may result from the impaired development of blood
vessels and other tissues in fetal life and infancy.
Prabesh Ghimire, MPH 36
37. Types of Ecological Study
Longitudinal/Time trends
• Many diseases show remarkable fluctuations in incidence over time.
• Epidemics of chronic disorders such as lung cancer and coronary
heart disease evolve over decades.
• If time or secular trends in disease incidence correlate with changes
in a community’s environment or way of life then the trends may
provide important clues to aetiology.
• Example: In Britain, successive rises and falls in mortality from
cervical cancer have been related to varying levels of sexual
promiscuity, as evidenced by notification rates for gonorrhoea.
Prabesh Ghimire, MPH 37
38. Types of Ecological Study
Migrant studies
• In migrant studies, the disease rate among persons who have
migrated from one location to another is compared with the
disease rate in persons who did not migrate.
• Second generation Japanese migrants to the USA have
substantially lower rates of stomach cancer than Japanese
people in Japan, indicating that the high incidence of the
disease in Japan is environmental in origin.
Prabesh Ghimire, MPH 38
39. Ecological Study
• Advantage
• Inexpensive and easy to carry-out using routinely collected data
• Useful for performing international comparisons and studying group-
level effects (correlation between rates from CVD and cigarette sales
per capita)
• Disadvantage
• Prone to bias and confounding
• Caution is needed when applying grouped results to the individual level
Prabesh Ghimire, MPH 39
40. Ecological Study Examples
• Assessment of various dietary factors and cancer mortality and
incidence by country.
• Incidence rates for 27 cancers in 23 countries and mortality rates for 14
cancers in 32 countries have been correlated with a wide range of
dietary and other variables.
• Source: https://onlinelibrary.wiley.com/doi/pdf/10.1002/ijc.2910150411
Prabesh Ghimire, MPH 40
41. Ecological Fallacy
• Type of confounding specific to ecological studies.
• Occurs when relationships which exists for groups are assumed to
also be true for groups.
• It is an error in the interpretation of the results of an ecological study,
where conclusions are inappropriately inferred about individuals from
the results of aggregate data.
• The fallacy assumes that individual members of a group all have the
average characteristics of the group as whole, when in fact any
association observed between variables at the group level does not
necessarily mean that the same association exists for any given
individual selected from the group.
Prabesh Ghimire, MPH 41
42. Ecological Fallacy
• For example, it has been observed that the number of
televisions per capita is negatively associated with the rate of
deaths from heart disease.
• However, it would be an ecological fallacy to infer that people
who don’t own televisions die from heart disease.
• Indeed, in this scenario there are other potentially causative
factors that could be common to both, such as reduced physical
activity or a poorer diet associated with less affluent societies.
Prabesh Ghimire, MPH 42
43. Ecological Fallacy
• In ecologic studies, only information on aggregate measures,
such as the average exposure in City A and the death rate in
City A can be known.
• At the individual level, however, we can, for example, determine
the proportion of people who died within each of the categories
of exposure (low or high).
Prabesh Ghimire, MPH 43
44. Example of ecological fallacy
• Suppose indoor air pollution is higher in Bajura than in Achham, but
mortality from COPD is lower in Bajura than in Achham.
• It would be fallacious to conclude that indoor air pollution protects against
COPD deaths.
• It is possible that persons dying of COPD in Achham may have moved
from cities with high indoor air pollution or that another risk factor for
COPD – such as smoking – is more prevalent in Achham than Bajura.
• We do not know the cumulative exposures of cases and non-cases in
either district.
• The heterogeneity of lifetime air pollution exposure among individuals in
each district makes the average exposure unrepresentative of the
distribution of exposure among individuals in the population.
Prabesh Ghimire, MPH 44
45. Criteria for ecological fallacy
Ecological fallacy exists if it meets all of these three criteria
• Results must be obtained with ecological data
• Data must be inferred to individuals.
• Results obtained with individual data are contradictory
Prabesh Ghimire, MPH 45
46. Reasons for ecological fallacy
• It is not possible to link exposure with disease in individuals -
those with disease may not be the same people in the
population who are exposed.
• The data used may have originally been collected for other
purposes.
• Inability to control for confounding.
Prabesh Ghimire, MPH 46