SPH 696 Epidemiology Methods II Final Exam Assignment Solution

This document provides a comprehensive solution to the SPH 696 Epidemiology Methods II final exam. The solution addresses a public health emergency scenario, proposing a retrospective cohort study to identify risk factors for coronavirus-related deaths, detailing study design, data analysis, and potential biases. It then analyzes an article on political affiliation and mortality, discussing longitudinal study characteristics and immortal time bias. Finally, the solution examines an article on social fragmentation and sleep among adolescents, defining reliability, discussing intra-subject, intra-observer, and inter-observer variability, assessing the validity of sleep measurement, and providing calculations for sensitivity and specificity. The solution also explains the difference between percent agreement and kappa statistic and discusses the acceptability of the measurement tool, providing a well-rounded response to the exam questions.

Final Exam Name: ____________________
SPH 696 Epidemiology Methods II
Dr. Roman Pabayo
Part One: 40 marks
1. On January 28, 2020, Canada's first case of a novel coronavirus was identified and confirmed.
As of April 3, 2020, there are around 20,000 cases and 177 deaths. The Public Health Agency of
Canada has hired you to identify potential risk factors for deaths due to coronavirus infection.
How would you investigate this pressing public health emergency? You can make reasonable
assumptions.
Retrospective cohort study
- cohort studies allows for the investigator or to identify subjects at a point in time when
they have negative interests outcome and makes the comparison of incidence of the
outcome of interests among the exposed and non exposed groups. Cohorts can be
identified retrospective or prospectively, however in either case they need to be
established at two times. There is need for a consideration that cohort did have Covid -19
outcome at the beginning of the observation and need to observed if the outcome disease
developed later through assessment of incidence in each groups of study.
- Retrospective study compares the exposure status and makes a comparison based on the
disease incidence levels, in this case the status of the exposure and outcome are
ascertained in a retrospective manner.
- In this process, identification of a useful cohort which is free from a disease and is at
risks is selected. The records applicable in this will least’s say travel history of the cohort
individuals, assessment of each subject exposure status at the beginning of observation
period and then assessment of the subjects encounters in the two groups.
2. Data analysis
- In this study design, essential data tools which will be employed d entailed incidence
assessment of the disease, relative risks or risk ratios of developing the disease, risks
1

Final Exam Name: ____________________
SPH 696 Epidemiology Methods II
Dr. Roman Pabayo
difference between the two groups and attributable proportion leading to deaths of the
disease in both the groups and disease occurrence.
- When the risks of the two groups are different, then we affirm that there is no association
between the disease and exposure state, while if the risks of the disease are higher or
lower in the exposed group then we make a conclusion that exposure is associated with
the covid -19 occurrences among the subjects.
3. Sources of bias, distortion or any other limitations.
- Cohort studies have inherent advantages; they provide clarity on temporal sequence.
They clearly indicate the temporal sequences occurring between the exposure and
outomce, due to the fact that cohort are known to be disease free at start of the
observation when the status of exposure is being established.
- Further cohort studies allows for calculation of incidences of diseases between the
groups. In this case fundamental measures entail absolute risks-incidence, risks ratio,
risks differences and attributable risks % of the disease.
- Cohort studies are essential in allowing multiple effects selection and examination of a
single exposure.
- Selection bias can be observed in this case since the otuomces already might have been at
different times.
- The usage of record in this design is crucial. Usage of record not for the design of the
study makes the data obtained to be of poor quality. Further, there is often the absence of
potential confounding factors if the data was obtained in the past. Further it has been
observed that it can be a challenge to identify the correct exposed cohort and appropriate
comparison group, however in this case the Covid 19 victims are available and easily
traceable. In large cohorts loss to follow up can occur such as death occurrences as
observed din this case there high mortality rates thus can hinder this process leading to
bias retrospective assessment.
(Supino & Borer,2012; Gallin & Ognibene, 2012)
2

Socio-
demographic
characteristics
MortalityPolitical party
affiliation and
political ideology
Moderates
Conservatives
Republicans
Final Exam Name: ____________________
SPH 696 Epidemiology Methods II
Dr. Roman Pabayo
Part Two: 20 marks
The following questions 1-5 are based on information from the article “Political Party
Affiliation, Political Ideology and mortality.”
4. Main exposure(s), outcome and a conceptual model
Exposures = political party affiliations and self-reported happiness and self-rated health
Outcomes = time to death and political beliefs and time to death
Conceptual framework
5. The authors argue that this article is a longitudinal study. Do you agree? Because ;
- The study is a longitudinal study bearing in mind the subjects were followed from the
year 1976 until the year 2008, this enabled a prospective assessment of political party
affiliations and risks of mortality rates among the subjects over these years.
- In essence, a longitudinal study is a research design study that observed represented
variables especially people over a long period to assess the underlying n measures being
of the study.
3

Final Exam Name: ____________________
SPH 696 Epidemiology Methods II
Dr. Roman Pabayo
- Longitudinal studies as observed by Caruana, Roman, Hernández-Sánchez & Solli (2015)
employ a continuous process of repeated measures among individuals for a longer
duration of time. This study, the main measures of interest were political party beliefs
assessed through political party affiliations and ideologies followed for the duration.
6. Some epidemiologists argue that immortal time bias was committed by the authors. No, because
;
- Immortal time bias refers to the period in which death cannot occur during a cohort
follow up. Immortal bias according to Rotham et al (2008) as the occurrence of the period
in which the cohort entry and first expire date of assessment in which the exposure of
interest has not occurred. In this case, it is either misclassified or not excluded from the
analysis.
- Issues with immortal time bias in cohort studies have been shown how it can affect or
lead to bias results in favour of treatment groups and how it is often challenging to avoid
it. Often is recommended that cohort studies need to assess the presence of immortal time
bias based on validity criteria. In this case, accounting for immortal time bias could be an
issue in this study.
- Based on the measures of the outcome being investigated which entailed the association
of political afflation’s and ideologies on mortality tend to lead to non-avoidance of
immortal time bias. The reasoning underlying this aspect is mortality was the result
outcome result in this case hence cannot be justified for the occurrence of immortal time
bias. Focusing on this bias concerning this study means that the end outcome of the study
could not be realized since death occurrence is the primary outcome of interests.
(Melamed, A., & Robinson, 2019)
4

Final Exam Name: ____________________
SPH 696 Epidemiology Methods II
Dr. Roman Pabayo
Part 3: 40 marks
The following questions are based on information in the article: “The relationship between social
fragmentation and sleep among adolescents living in Boston, Massachusetts.”
The outcome of this study is sleep behaviour:
"To measure usual sleep duration, students were asked, "How many hours and minutes of sleep
do you usually get on a school night?" Based on the CDC's recommendation (Http: Based on the
CDC's recommendation (https://
www.cdc.gov/sleep/about_sleep/how_much_sleep), responses were dichotomized to ≥8.5 and
<8.5 h per night. Also, the number of hours of sleep was converted to the number of minutes to
obtain a continuous measure of sleep duration.
7. Define reliability. Describe and discuss intra-subject variability, intra-observer variability, and
inter-observer variability of sleep measurement used in this study and likely variation observed;
- Reliability refers to the degree to which the research method being used is producing
stable and consistent results. Having a consistent measure is essential to consider reliable
results.
- A critical aspect of variation in cross over design s entails intra-subject variability which
is exhibited through a single subject when the exposure of interests is given with repeated
exposures. intra-observer variability can be affected by confounds that influence the
dependent and independent variables leading to a spurious association. It is often difficult
to quantify this base don experimental studies, more often it is neglected in analysis in
psychophysical sampling avenues, thus in most cases, these experiments are often hard to
be realistic and results obtained are not representative of the true picture (Rinderknecht,
Lambercy, & Gassert, 2018a).
- Intra-observer variability refers to the amount of variation between the results obtained
on two or more examination observations on the same material. This illustrates the
5

Final Exam Name: ____________________
SPH 696 Epidemiology Methods II
Dr. Roman Pabayo
number of variations in which observer’s experiences when observing the same outcomes
more than n occurrence.
- Intra-observer variability is Essential Avenue for information on the quality of
observational data, the objectivity of the observers based on the same behavioural
methodology is essential in the quality of the observation. This determination is essential
in managing inter-observer agreement in developing observer measures (Rinderknecht et
al., 2018b). In this case, the levels of assessment for sleep behaviour as the determinant is
essential in aspects of intra-observer variability.
- Interobserver variation entails the variation between the results yielded by two or more
observers in assessing outcome levels. The coefficient of observation agreement is a key
aspect of observational data quality. Determining inter-observer variability is key in
developing observational measures.
- In this study, the Inter-observer variation is likely to be higher due to the many
obsevr4res assessing sleep behaviour as the study responses are being sought by the
students through question response avenue. Thus in this usage of more observes could
lead to a variety of results due to variation of examination of one key outcome for a long
period.
8. In your opinion, do you think that this measurement is valid? Yes, and its measurement entails:
- Assessment of validity is essential in research and requires underlying evidence to build a
comprehensive case to measure what is supposed to measure. Validity determination can
be viewed as an avenue for constructing an evidenced-based argument with regards to the
tool that can measure what is supposed to measure.
- The instrument in the research thus needs to be reliable and valid for the results of the
study to be credible. Validity aspects need to be always be reported in each assessment
tool. At times applying a tool with high reliability is not sufficient, combining assessment
of validity measures is essential in determining the credibility of the tool.
- Assessment of validity calls for measures such as content, response process and variables
relationships to be considered in the study. Various tools for validity emphasis on
6

Final Exam Name: ____________________
SPH 696 Epidemiology Methods II
Dr. Roman Pabayo
different approaches. In utilizing observer ratings such as in this study, inter rate
agreement will be essential (Sullivan, 2011).
- In this study employing a validated tool by CDC is an essential element that allows the
validity of the results obtained. Based on the credibility of the CDC is an established
organization, applying and utilizing their recommendations of sleep assessment t assures
the reliability based on the underpinning evidence and basis for the development of the
key categories thus offering more validity ratings in this study.
9. In the space below, draw two 2 x 2 tables. One shows that the measurement in the study is highly
sensitive and specific, and another 2 x 2 table indicating the instrument is lowly sensitive and
specific. Use the cut-off of 8.5 hours as sufficient vs. insufficient sleep. Show any calculations. 5
marks
Highly sensitive and specificity
Testing the tool in a group of 113 students, 79 have sufficient sleep of 8.5 hours while in another
group of 217 students with insufficient less than 8.5 hours, 10 of them have sufficient sleep
Sufficient Insufficient
Tests positive 79 39
Tests negative 10 207
Hence with this
Sensitivity will be = TP/(TP+FN) = 79/(79+39) =67%
Specificity will be =TN/(TN+FP) =207/(207+10) =95%
Low sensitive and specificity
Testing the tool in a group of 113 students, 10 have sufficient sleep of 8.5hour while in another
group of 217 students with insufficient less than 8.5 hours, 200 of them have sufficient sleep
Sufficient Insufficient
Tests positive 10 103
Tests negative 200 17
7

Final Exam Name: ____________________
SPH 696 Epidemiology Methods II
Dr. Roman Pabayo
Hence with this
Sensitivity will be = TP/(TP+FN) = 10/(10+103) =8%
Specificity will be =TN/(TN+FP) =17/(17+200) =7%
10. Researchers tested the validity and reliability of the measurement tool among Japanese adults
aged 50-65. The percent agreement is 90.0% and the kappa statistic is 0.75. Describe the
difference between the percent agreement and kappa. Pabayo et. al. argue that this measurement
tool is acceptable for use in their study. Do you agree? Yes, Explain. 5 marks
- Kappa statistics are a measure essential for measuring inter-rater reliability. It has been
argued that is a robust tool compared to a simple percent agreement. Concern has often
been the challenge in interpreting the results. The measure of agreement between
examiners is thus an essential component in inter rate analysis, the Kappa tool is essential
in assessing this agreement rates. The perfect agreement is κ=1. A zero value signifies
agreement by chance, greater than agreement then it is κ≥0, less agreement obtained
entails κ≤0. In percent agreement, the total number of times in which the examiners agree
and divided by the total number of classification made.
- In this case, a percent agreement of 90% was found which is positive. A value that is
higher than 0.75 often signifies excellent agreement. Thus for this reason I agree with the
Pabayo et al on the application of the tool in this study.
11. Among adolescents, researchers have observed that current smokers were more likely to report
poor sleep quality in comparison to non-smokers. Which public health strategy would you use to
improve population health? In other words, how would you decrease smoking behaviour among
adolescents? Would you use a high-risk strategy or a population strategy (Briefly explain why
and give an example of an intervention)? Describe any strengths and limitations of your strategy.
(10 marks)
- In this setting, a high-risk strategy will be essential in this context. Targeting smokers as
they are the population of the risks is essential in addressing public health concerns. High
risks strategy is essential in targeting the population of interest in this case the smokers
who are experiencing and likely to encounter poor sleep patterns.
8

Final Exam Name: ____________________
SPH 696 Epidemiology Methods II
Dr. Roman Pabayo
- An example of a strategy to apply to this health promotion entails an education and media
campaign targeting the smokers and forming groups for behaviour change. This will be
crucial in ensuring the target group can realize their inherent behaviour risks associated
and engage in a behaviour change approach.
- Engaging in education and awareness campaign strategy to target the smokers on the
effects of smoking will be essential. This will employ behavior change models especially
health belief model in its application. This will engage the behaviors of the smokers on
being aware that they are facing sleep related problems, being linked to smoking. Upon
this acceptance, they will be engaged on a venue of improving the smoking behavior
through a multi stage cessation program. Support services will thus be initiated to support
the recovery process of the smokers in the general population.
- The underlying strengths of this strategy entail a targeted behavior approach to smoking
behavior while an underlying weakness entails failure or low rates of acceptance by the
clients that they have a problem. Behavior change starts with self-acceptance, lack of this
tend to jeopardize the whole process.
- Further, stigmatization issues could affect the behaviour change acceptance among the
smokers. Stigma from within themselves and peer pressure related issues could have a
significant effect on the manner in which they engage on the behavioiur change avenue.
Thus critical concerns need to be considered in this avenue.
9

Final Exam Name: ____________________
SPH 696 Epidemiology Methods II
Dr. Roman Pabayo
References
Caruana, E. J., Roman, M., Hernández-Sánchez, J., & Solli, P. (2015). Longitudinal studies.
Journal of thoracic disease, 7(11), E537–E540. https://doi.org/10.3978/j.issn.2072-
1439.2015.10.63
Gallin, J. I., & Ognibene, F. P. (Eds.). (2012). Principles and practice of clinical research (3rd
ed.).
Melamed, A., & Robinson, J. N. (2019). Case–control studies can be useful but have many
limitations: Study design: case‐control studies. BJOG: An International Journal of
Obstetrics & Gynaecology, 126(1), 23-23.
Rinderknecht, M. D., Lambercy, O., Raible, V., Büsching, I., Sehle, A., Liepert, J., & Gassert, R.
(2018b). Reliability, validity, and clinical feasibility of a rapid and objective assessment of
post-stroke deficits in hand proprioception. Journal of neuroengineering and
rehabilitation, 15(1), 47.
Rinderknecht, M. D., Lambert, O., & Gassert, R. (2018a). Enhancing simulations with intra-
subject variability for improved psychophysical assessments. PloS one, 13(12).
Rothman, K. J., Greenland, S., & Lash, T. L. (2008). Case-control studies. Encyclopedia of
Quantitative Risk Analysis and Assessment, 1.
Sullivan G. M. (2011). A primer on the validity of assessment instruments. Journal of graduate
medical education, 3(2), 119–120. https://doi.org/10.4300/JGME-D-11-00075.1
Supino, P. G., & Borer, J. S. (Eds.). (2012). Principles of research methodology: A guide for
clinical investigators.
10