observational studies

what type of study design provides the greatest strength of evidence? and the lowest?

• greatest = RCT

  • experimental study designs > observational study designs

• lowest = case series and case reports

SATA: experimental study designs

a) Cohort study

b) RCT

c) cross-sectional

d) quasi-experimental

b) RCT

d) quasi-experimental

SATA: which of the following are analytical studies?

a) RCT

b) Case Report

c) Cohort Study

d) Case-Control

e) case series

f) cross-sectional

c) Cohort Study

d) Case-Control

f) cross-sectional

SATA: which of the following are descriptive studies?

a) RCT

b) Case Report

c) Cohort Study

d) Case-Control

e) case series

f) cross-sectional

b) Case Report

e) case series

f) cross-sectional

descriptive observational study

• distribution of one or more variables, without regard to an causal or other hypothesis

  • not one clinical/epidemiological study can answer a research question definitively

  • info = routinely collected/available → less expensive and time consuming

  • essential for formulating hypothesis/research questions

which one is case report and a case series?

• single patient = ??

• series of patients = ??

• single = case report

• series = case series

case report and case series role

• first alert for possible link between clinical medicine + epidemiology of disease

• chance observation may open new line of investigation

cross-sectional studies

• estimates distribution of outcomes and/or exposures within a population at a given time point

• helpful assessing disease burden + healthcare needs

• can estimate prevalence of exposure and/or diseases

analytical study

• perform hypothesis testing to establish association between exposure and outcome

  • comparison group(s)

  • timing sequence

what are the two types of analytical studies?

1. observational (case-control studies, cohort studies)

2. interventional (experimental) (RCT)

analytical observational studies

• non-experiment studies that attempt to establish an association between exposure(s) and outcome(s)

  • exposure = naturally determined

analytical observational study sampling design

• did they define with regards to exposure status or disease status?

if an analytical observational study was defined based on exposure, what type of study is it?

a) cohort study

b) case-control study

c) prospective

d) retrospective

a) cohort study

if an analytical observational study was defined based on outcome, what type of study is it?

a) cohort study

b) case-control study

c) prospective

d) retrospective

b) case-control study

analytical observational study timing

• whether the investigator started the study before or after the outcome

• did they start collecting data before or after the disease occurred?

if an analytical observational study starts the study before the outcome/disease occurrence, what type of study is it?

a) cohort study

b) case-control study

c) prospective

d) retrospective

c) prospective

if an analytical observational study starts the study after the outcome/disease occurrence, what type of study is it?

a) cohort study

b) case-control study

c) prospective

d) retrospective

d) retrospective

study conducted in 2023

• 2 groups

  • one group exposed to antipsychotics

  • another group without exposure to antipsychotics

• follow patients for 5 years + recorded the incidence of diabetes in both groups

what type of analytical observational study?

• cohort study

• prospective

study conducted in 2023

• sampled patients who attended primary care clinics from 2015-2018

• 2 groups

  • one group exposed to antipsychotics

  • another group without exposure to antipsychotics

• used data from EMR to record the incidence of diabetes in both groups after 5-years of exposure to antipsychotics

what type of analytical observational study?

• cohort

• retrospective

study conducted in 2023

• sampled patients who attended primary care clinics from 2015-2018

• 2 groups

  • one group with diabetes

  • another group without diabetes

• used data from EMR to record exposure to antipsychotics in both groups after 5-years of exposure to antipsychotics

what type of analytical observational study?

• case-control

• retrospective

all case studies are (prospective/retrospective)

• retrospective

the selection of cohorts depends on what factors?

1. type of exposure being investigated

2. frequency of exposure in population

3. accessibility of patients

how do cohorts establish a temporal relationship?

• be free of the outcome of interest at the start of the study

counterfactual ideal for selecting cohorts

• all groups should be equally susceptible to developing the outcome at the start of the study

T/F: it is okay if the degree of surveillance varies between groups in cohort studies

false

• degree of surveillance should be similar between groups

SATA: what are some measures of associations are used in cohort studies?

a) risk ratio

b) absolute risk reduction/increase

c) odds ratio

d) incidence rate ratio

a) risk ratio

b) absolute risk reduction/increase

d) incidence rate ratio

how to calculate risk ratio?

• CI_exp+ / CI_Exp-

  • (A/ A+B) / (C / C+D)

what are the pros for cohort studies

• clearer temporal relationship

• able to study multiple outcomes

• efficient for studying uncommon or rare exposures

• incidence data available

what are the cons for cohort studies

• expensive + time consuming (compared to case-control)

• often requires large sample

• loss to follow-up may be problematic

• inefficient for rare outcomes

• not easy to reproduce

a cohort study was conducted to investigate the risk of diabetes with antipsychotics. The investigators identified 255 patients with antipsychotic exposure and 284 patients without antipsychotic exposure and followed them for 5-years. At the end of the study, 78 exposed patients developed diabetes and 64 unexposed patients developed diabetes.

calculate the RR for the risk of diabetes with antipsychotic exposure. Calculate relative risk increase

RR= 1.35 (1.35 times the risk)

relative risk increase = 1.35 -1 = 0.35 (35% increase)

participants are selected based on what for case-control studies?

• case status

what are cases?

• patients with the disease (outcome)

  • based on case definition (eligibility criteria for inclusion/exclusion in the study, sensitive and specific)

what are controls in case-control studies?

• patients without the disease (outcome)

what are some considerations in selecting controls?

• sampled form the same underlying cohort as the cases or form the same source population that gave rise to the cases

• sampled independent of the exposure

matching

• common approach to ensure cases and controls are similar with respect to important risk factors (counterfactual ideal)

a higher ratio in matching controls to a case (1:1 vs 4:1) also increases ____

• precision

  • less variability

what measure of association is used in case-control studies

a) risk ratio

b) NNT/NNH

c) hazard ratio

d) odds ratio

d) odds ratio

how to calculate odds ratio?

• (A*D)/(C*B)

T/F: in case-control studies, investigators can measure incidence

• false

  • cannot measure incidence because they start with the outcome

  • the presence/absence of the disease is fixed by design

pros of case-control studies

• short duration + inexpensive

• can study multiple exposures

• efficient for studying uncommon or rare diseases

• efficient for studying diseases with long latency

• requires smaller sample size

cons of case-control studies

• temporality can be an issue

• difficult to identify appropriate controls

• unable to estimate prevalence or incidence (cannot calculate RR or risk differences)

• inefficient for rare exposures

a cohort study was conducted to investigate the risk of diabetes with antipsychotics. The investigators identified 200 patients with diabetes, then matched 400 patients without diabetes and collected whether they had exposure to antipsychotics over the past for 5-years. At the end of the study, 55 patients with diabetes had exposure to antipsychotics and 87 without diabetes had exposure to antipsychotics

calculate the OR for the risk of diabetes with antipsychotic exposure.

OR = 1.36 (1.36 times the odds)

nested case-control studies

• case-control study conducted within a well-defined cohort (or cohort study)

pros for nested case-control studies

• ensures inclusion of incident cases (limits issue of temporality)

• ensures controls are selected from the same cohort/source population that gave rise to the cases

• could be more efficient than prospective cohort

  • less costing, no need to study all patients in cohort

cons for nested case-control studies

• takes longer than traditional case-control studies

• less strength of evidence than cohort study

Which of the following study designs has the highest strength of evidence? 

1. Cross-sectional study

2. Case-control study

3. Randomized controlled trial

4. Cohort study

3. Randomized controlled trial

An observational study is being conducted in 2023 to investigate whether vasopresssin increases the risk of venous thromboembolism (VTE) among critically ill patients. The investigators sampled patients who were hospitalized in the ICU from 2020-2022 and obtained two groups: one group who were treated with vasopressin and another group who were not treated with vasopressin. The investigators then recorded whether patients in each group developed VTE during their hospitalization.

Which of the following best describes this study?

1. Retrospective cohort study

2. Nested case-control study

3. Prospective cohort study

4. Traditional case-control study

1. Retrospective cohort study

   • defined the study sample by exposure ststaus (vasopressin use vs no vasopressin use)

n observational study is being conducted in 2023 to investigate whether bisphosphonate use increases the risk of fractures in older women. The investigators sampled patients from a population-based cohort study of women who started the study between 2010 and 2012. The cohort was followed up for 10 years. The investigators obtained two groups: one group who were hospitalized with a fracture during the 10-year period and another group who did not have a hospitalization for a fracture during the 10-year period. The investigators then recorded exposure to bisphosphonates in both groups.

Which of the following best describes this study?

1. Retrospective cohort study

2. Nested case-control study

3. Prospective cohort study

4. Traditional case-control study

2. Nested case-control study

   • case study because study sample is defined by disease status (fractures vs. no fractures)

   • nested case-control study because case-control study was conducted within a population-based cohort. 

Which of the following factors influence the selection of the cohort in a cohort study?

1. Accessibility of patients

2. All of the above

3. Frequency of exposure in the population

4. Type of exposure being investigated

2. All of the above

In cohort studies, the cohort selected should be outcome-free at the start of the study. Which of the following is the most appriopriate reason for this procedure?

1. To establish a temporal relationship between the exposure and outcome

2. To increase accessibility of patients to be included in the study

3. To ensure all patients with the exposure are included 

4. To exclude patients not at risk for the outcome

1. To establish a temporal relationship between the exposure and outcome

A cohort study was conducted to investigate whether vasopressin use increases the risk of venous thromboembolism among critically ill patients. The investigators identified 175 patients with vasopressin exposure and 175 patients without vasopressin exposure. At the end of the study, 27 exposed patients developed VTE and 24 unexposed patients developed VTE.

Calculate the appropriate relative measure of association for the risk of VTE with vasopressin.

1. Risk ratio = 0.89

2. Odds ratio = 1.15

3. Odds ratio = 0.87

4. Risk ratio = 1.13

4. Risk ratio = 1.13

Which of the following is the first step in conducting a case-control study?

1. Matching controls to cases

2. Identifying cases using a case definition

3. Selecting controls from the source population

4. Recording prior exposure to the drug of interest

2. Identifying cases using a case definition

Which of the following is TRUE regarding selecting controls? 

1. Controls should be sampled from the same source population as the cases and dependent of the exposure

2. Controls should be sampled from a different source population as the cases and dependent of the exposure

3. Controls should be sampled from the same source population as the cases and independent of the exposure

4. Controls should be sampled from a different source population as the cases and independent of the exposure

3. Controls should be sampled from the same source population as the cases and independent of the exposure

A nested case-control study was conducted to investigate whether bisphosphonate use increases the risk of fractures among older women. The investigators sampled 200 women with fractures and 200 women without fractures. At the end of the study, 42 women with fractures were using bisphosphonates and 34 women without fractures were using bisphosphonates. 

Calculate the appropriate relative measure of association for the risk of fractures with bisphosphonates.

1. Odds ratio = 1.25

2. Risk ratio = 0.81

3. Risk ratio = 1.24

4. Odds ratio = 0.80

Odds ratio = 1.25