4. All studies should start with a research
question that addresses what the investigator
would like to know
Goal is to find an important research question
that can be developed into a feasible and
valid study plan
5. Usual format (analytic studies):
“In a population of [study population], is
[predictor variable] associated with [outcome
variable]?”
6. In a population of injection drug users with
HIV infection, is HAART use associated with
slower progression of chronic hepatitis B ?
In a population of children with falciparum
malaria, is HIV associated with higher
mortality ?
In a population of truck drivers, is
unprotected sex with sex workers
associated with HIV infection ?
7. Format for descriptive study
Most studies will have more than one
research question
8. In descriptive studies:
“In a population of [study population], what is
the prevalence (or mean, median, etc.) of
[outcome variable]?”
9. Acute respiratory tract infections caused by
Streptococcus pneumoniae are a leading cause
of morbidity and mortality in young children. We
evaluated the efficacy of a 9-valent
pneumococcal conjugate vaccine in a
randomized, double-blind study in Soweto,
South Africa. At 6, 10 and 14 weeks of age,
19,922 children received the 9-valent
pneumococcal polysaccharide vaccine
conjugated to a noncatalytic cross-reacting
mutant of diphtheria toxin (CRM197), and
19,914 received placebo…
Klugman KP, Madhi SA, Huebner RE, et al. A trial of 9-valent
pneumococcal conjugate vaccine in children with and without HIV
infection.
10. In a population of South African infants is
immunization with a 9-valent
pneumococcal conjugate vaccine
associated with invasive pneumococcal
disease ?
In a population of [study population] is [predictor
variable] associated with [outcome variable]?
11. Mastering the literature
◦ Scholarship -- be a scholar and know the literature
◦ Get a mentor
Be alert to new ideas and techniques
◦ Importance of teaching
Be imaginative
17. A population is a complete set of people or
objects (e.g., clinics, neighborhoods) with a
specified set of characteristics
◦ Clinical
◦ Demographic
◦ Geographic
◦ Period of time
A sample is a subset of the population
18. The target population is the group of all people
in the world to which results will apply (e.g.,
inpatients Ahmadu Bello University Teaching
Hospital)
The accessible population is the subset of the
target population that is potentially available to
the investigator for study (e.g., O&G inpatients
with postpartum hemorrhage at Ahmadu Bello
University Teaching Hospital)
The study population or study sample is the
subset of the accessible population that is asked
to participate in the study (e.g., a random 20%
sample of all O&G inpatients with postpartum
hemorrhage at Ahmadu Bello University Teaching
Hospital)
19. In the research question, the population
refers to the target population
◦ Defined by clinical, demographic and geographic
characteristics
The research plan specifies the accessible
population and how to sample it to get the
study population
20. Findings in the study sample are inferred to
be present in the accessible population and,
by extension, in the target population
Study findings are generalized from the study
sample to the accessible population to the
target population
21. Rather than selecting participants from the
entire accessible population, we often
define a subset of participants for the
study using inclusion and exclusion criteria
and then choose participants
22. Inclusion criteria are the main
characteristics of the target population -
demographic, clinical, geographic and
temporal
Exclusion criteria are the subset of
individuals in the accessible population
that should not be included (other causes
for outcomes, poor data quality)
◦ The fewer the exclusion criteria, the more
generalizable the results
23. Demographic
characteristics
Clinical
characteristics
Geographic
characteristics
Temporal
characteristics
Male injection drug
users 20-40 years
old
HIV infection
Attending
investigator’s clinic
Seen between
January 1 and
December 31, 2006
24. A high likelihood of
being lost to follow up
An inability to provide
good data
Being at high risk for
side effects
Characteristics that
make it unethical to
withhold study
treatment
Not a permanent
resident of the study
area
History of
encephalopathy or
meningitis
≥2 failed HAART
regimens
Applies primarily to
experiments
25. Study sample or study population, further defined by inclusion and
exclusion criteria
Accessible
population
27. External validity?
◦ How different are those invited to participate
from those eligible to participate in the accessible
population but not invited, that is, not selected
for the sample (demographics, illness status,
geography, risk profile)?
Internal validity?
◦ How different are those who choose to participate
in comparison to those who are invited but
decline?
28. Acute respiratory tract infections caused by
Streptococcus pneumoniae are a leading cause of
morbidity and mortality in young children. We
evaluated the efficacy of a 9-valent pneumococcal
conjugate vaccine in a randomized, double-blind study
in Soweto, South Africa. At 6, 10 and 14 weeks of age,
19,922 children received the 9-valent pneumococcal
polysaccharide vaccine conjugated to a noncatalytic
cross-reacting mutant of diphtheria toxin (CRM197),
and 19,914 received placebo…
Klugman KP, Madhi SA, Huebner RE, et al. A trial of 9-valent
pneumococcal conjugate vaccine in children with and without HIV
infection. N Engl J Med 2003; 349:1341-8.
29. “The 9-valent pneumococcal conjugate
vaccine is currently under development
for licensure in both developed and
developing countries but has not yet
been licensed for use. Our study
provides evidence to support the wider
development and use of this vaccine to
prevent invasive pneumococcal disease,
reduce antibiotic resistance among
pneumococcal strains, and diminish the
incidence of pneumonia in children.”
30. Successful recruitment generally means…
◦ response, generalizable sample, adequate size
For database only studies (Not usually a big
problem)
For hands-on studies (e.g., surveys, cohorts, trials)
◦ Expect that it will be harder than you think!
◦ Use reasonable inclusion/exclusion criteria
◦ Acceptable burden to subjects with potential
benefits
◦ Try to minimize subject non-response (non-
response bias)
31. Advantage of sampling (as opposed to
including everyone in the target
population)
◦ Efficiency
◦ Allows investigator to draw inferences about a
large population by examining a sample at
relatively small cost in time and effort
Disadvantages of sampling
◦ Introduces error
◦ If sample is insufficiently representative of
target population, findings may not
generalize to target population
33. Convenience sample
◦ Population accessible to investigator
◦ No need to randomly select participants
◦ Representativeness? Selection bias?
Consecutive sample
◦ All members of target population accessible to
investigator recruited as they present for care, etc.
◦ Minimizes selection bias
Chain-referral or snowball sample
Quasi-probability sampling
◦ Respondent-driven sampling
◦ Time-location sampling
34. Probability sampling
◦ Each unit of population has equal chance of being
selected to participate
Simple random sample
Stratified random sample
Cluster sample
Systematic sample
35. Simple random sample
Stratified random sample
Cluster sample
Systematic sample
36. RQ: Among patients admitted with PPH is
there an association between induction of
labor/labor augmentation and excessive
bleeding (blood loss >1000ml)?
◦ What is the target population?
?
◦ What is the accessible population?
?
37. Sampling frame: …….(including dead
patients)?
Inclusion criteria: all vs. subset?
◦ Demography:?
◦ Geography/administrative:
◦ Clinical:?
◦ Time period: one year, one month?
38. Previously diagnosed with…. (presence of
chronic diseases)
Any other ???
Remember: exclusion criteria are applied to a subset of individuals in
the accessible population that should not be included (other causes
for outcomes, poor data quality), and the fewer the exclusion criteria,
the more generalizable the results to the accessible and target
populations.
39. You decide you need to include 500
patients and approximately 1500 patients
passed by the O&G services and meet the
major inclusion criteria.
Name three sampling strategies and their
advantages and disadvantages:
◦ …
◦ …
◦ …
40.
41. ◦ assess household composition and characteristics;
◦ assess characteristics of respondents and women’s
status;
◦ measure fertility levels and preferences;
◦ assess knowledge, attitudes and use of family planning
methods;
◦ assess family planning and maternal health services
accessibility and availability;
◦ assess maternal mortality, morbidity and women’s
reproductive health status; and
◦ produce a report that characterizes the reproductive
health status of the study population that will serve as
baseline for reproductive health program priority
setting and development of public health interventions
42. Is the problem a major public health concern?
Is there a high demand for a solution?
Is there a readily available (or can be developed)
solution?
Will the solution have a significant impact on
correcting the problem
Is the solution feasible to implement (i.e. will it
reach the most vulnerable population)
Are the limitations of the solution acceptable?
Deciding on Interventions: