Epidemiology quiz

Synthesis and Practical Application: Study Design

David D. Celentano, ScD, MHS Johns Hopkins University

2

Module 2: Study Designs

3

Objectives

► Review and compare epidemiologic study designs

► Contrast different study designs by designing an epidemiologic study

► Name epidemiologic study designs based on information provided in abstracts of published papers

4

Study Designs

5

Summary of Study Designs: Main Types of Epidemiologic Studies

Study type Characteristics

Experimental Studies prevention and treatment of diseaseInvestigator actively manipulates which groups receive the study agent

Observational Studies causes, prevention, and treatment for diseasesInvestigator watches as nature takes its course

Cohort Examines multiple health effects of an exposureSubjects defined by exposure levels and followed for disease occurrence

Case-control Typically examines multiple exposures in relation to a diseaseSubjects are defined as cases and controls and exposure histories compared

Cross-sectional Examines relationship between exposure and disease prevalence in a defined population at one point in time

Ecological Examines relationships between exposure and disease with population-level data rather than individual data

The material in this video is subject to the copyright of the owners of the material and is being provided for educational purposes under rules of fair use for registered students in this course only. No additional copies of the copyrighted work may be made or distributed.

Randomized Control Trial

7

Experimental Study Designs

Study type Characteristics

Experimental

Studies prevention and treatment of disease Investigator actively manipulates which groups receive the study agent

= Randomized Trial

8

Design of a Randomized Trial—1

Start with:

9

Design of a Randomized Trial—2

Start with:

Then randomize to treatment:

10

Design of a Randomized Trial—3

Start with:

Then randomize to treatment:

Then follow up:

11

What Is the Primary Purpose of Randomization?

To prevent bias in the choice of treatment

12

Observational Study Designs

Study type Characteristics

Observational Studies causes, prevention, and treatment for diseasesInvestigator watches as nature takes its course

1. Cohort Examines multiple health effects of an exposureSubjects defined by exposure levels and followed for disease occurrence

2. Case-control Typically examines multiple exposures in relation to a diseaseSubjects are defined as cases and controls and exposure histories compared

3. Cross- sectional

Examines relationship between exposure and disease prevalence in a defined population at one point in time

4. Ecological Examines relationships between exposure and disease with population-level data rather than individual data

The material in this video is subject to the copyright of the owners of the material and is being provided for educational purposes under rules of fair use for registered students in this course only. No additional copies of the copyrighted work may be made or distributed.

Cohort Study

2

Observational Study Designs Observational study design Data level

1. Cohort Individual-level data

2. Case-control Individual-level data

3. Cross-sectional Individual-level data

4. Ecological Population-level data

3

Observational Study Designs—Cohort Studies

► Examines multiple health effects of an exposure

► Subjects defined by exposure levels and followed for disease occurrence

► Individual-level data

4

Design of a Cohort Study—1

Start with:

5

Design of a Cohort Study—2

Start with:

Then follow up:

6

Alternative Design of a Cohort Study—1

Start with:

7

Alternative Design of a Cohort Study—2

Start with:

Then measure:

8

Alternative Design of a Cohort Study—3

Start with:

Then measure:

Then follow up:

9

Design of a Cohort Study

The material in this video is subject to the copyright of the owners of the material and is being provided for educational purposes under rules of fair use for registered students in this course only. No additional copies of the copyrighted work may be made or distributed.

The Two Types of Cohort Studies

2

Observational Study Designs: The Two Types of Cohort Studies

► Two types of cohort studies: 1. Prospective 2. Retrospective

3

This Design Holds for Both Prospective and Retrospective Cohort Studies

4

So What Is the Difference between Prospective and Retrospective Cohort Studies?—1

5

So What Is the Difference between Prospective and Retrospective Cohort Studies?—2

6

Design a Prospective Cohort Study—1

Today (September 2018):

Today (September 2018):

Future (for example, September 2021):

7

Design a Prospective Cohort Study—2

8

Design a Prospective Cohort Study—3

9

Design a Prospective Cohort Study—4

10

Design a Prospective Cohort Study—5

11

Design a Retrospective Cohort Study—1

Past (September 2011):

Past (September 2011):

Anytime between September 2011 and the Present

(September 2018):

12

Design a Retrospective Cohort Study— 2

13

Design a Retrospective Cohort Study— 3

14

Design a Retrospective Cohort Study— 4

15

Retrospective vs. Prospective Cohort Studies—1

16

Retrospective vs. Prospective Cohort Studies—2

17

Retrospective vs. Prospective Cohort Studies—3

18

Is This a Prospective or Retrospective Cohort Study?—1

19

Is This a Prospective or Retrospective Cohort Study?—2

20

Is This a Prospective or Retrospective Cohort Study?—3

21

Retrospective vs. Prospective Cohort Studies—4

The material in this video is subject to the copyright of the owners of the material and is being provided for educational purposes under rules of fair use for registered students in this course only. No additional copies of the copyrighted work may be made or distributed.

Case-Control Study

2

Observational Study Designs—Case-Control Studies

► Individual-level data

3

Design of a Case-Control Study—1

4

Design of a Case-Control Study—2

5

Design of a Case-Control Study—3

6

Observational Study Designs: Types of Case-Control Studies

► Types of case-control studies 1. Traditional 2. Within a defined cohort

a) Nested case-control b) Case cohort

7

Nested Case-Control Studies

► Types of case-control studies 1. Traditional 2. Within a defined cohort

a) Nested case-control b) Case cohort

8

Design of a Case-Control Study Nested in a Cohort

9

Another Way to Think about Case-Control Studies within a Defined Cohort

► Types of case-control studies 1. Traditional 2. Within a defined cohort (These are efficient types of cohort studies!)

a) Nested case-control b) Case cohort

The material in this video is subject to the copyright of the owners of the material and is being provided for educational purposes under rules of fair use for registered students in this course only. No additional copies of the copyrighted work may be made or distributed.

Cross-Sectional Study

2

Observational Study Designs—Cross-Sectional Studies

► Individual-level data

3

Design of a Cross-Sectional Study

4

Let’s contrast different study designs by designing an epidemiologic study

5

Types of Outcome Data from Different Study Designs

Study design Design Outcome data & measures of association

Randomized trial Randomize participants to treatment and then follow up to see who develops the outcome

Incident cases of disease; incidence (risk); Incidence rate

Cohort Exposure measured and then participants followed up to see who develops the outcome

Incident cases of disease; incidence (risk); Incidence rate

Case-control Participants selected into study based on outcome; then go back and assess previous exposure

Mix of prevalent and incident cases (can design to limit to incident cases); odds

Cross-sectional Exposure and outcome measured at the same time Prevalence

6

Research Question

► Is regular physical activity at matriculation into a Master’s degree program associated with successful completion of that degree? ► Exposure = Physical activity (yes vs. no) ► Outcome = Graduation (yes vs. no)

7

Study Population

► Scientific considerations ► What is the research question? ► Who is at risk of the outcome? ► Person, place, time

► Practical considerations ► Participation ► Feasibility and financial considerations

8

Study Population—Students in 340.721

► Limit to students enrolled in a master’s program

9

Study Population—Students in 340.721 (cont.)

► Limit to students enrolled in a master’s program

10

Prospective Cohort Study

► Enroll students in 2018

► Assess physical activity

► Follow up students to determine who graduates

11

Retrospective Cohort Study

► Go back to school records to determine who was enrolled back in 2016

► Using school records, such as ID swipes at the Cooley Recreation Center, determine who was physically active in 2016

► Follow students to determine who has graduated on time

12

What Is a Problem with This Retrospective Cohort Study Design?

► Assumes we can actually figure out who is physically active and who is not ► We’ve limited determining physical

activity to tracking ID swipes at the recreation center (but people can be physical active in other ways)

The material in this video is subject to the copyright of the owners of the material and is being provided for educational purposes under rules of fair use for registered students in this course only. No additional copies of the copyrighted work may be made or distributed.

Case-Control Examples

2

Case-Control Study—1

► Identify graduates and non- graduates in 2018

► Ask participants if they were physically active at the time they took 340.721

3

Case-Control Study—2

► Identify graduates and non- graduates in 2017

► Ask participants if they were physically active at the time they took 340.721

4

What Is a Problem with This Case-Control Study Design?

► Identify graduates and non- graduates in 2017

► Ask participants if they were physically active at the time they took 340.721

5

Case-Control Study Nested Within a Cohort Study

► Enroll students in 2018

► Assess physical activity (using accelerometry collected 24 hours a day for 7 days)

► Follow up students to determine who graduates

► Select cases (graduates) and a sample of controls (non-graduates) for the study

► Go back and analyze accelerometry data for cases and controls

6

Cross-Sectional Study

7

What Is a Problem with This Cross-Sectional Study Design?

► Don’t know what came first

The material in this video is subject to the copyright of the owners of the material and is being provided for educational purposes under rules of fair use for registered students in this course only. No additional copies of the copyrighted work may be made or distributed.

Practical Applications— Identifying Study Designs

2

Let’s Practice!

► Name epidemiologic study designs based on information provided in abstracts of published papers

3

The Role of Contraceptive Use in Cervical Cancer: The Maryland Cervical Cancer … Study—1

“Abstract: Recent evidence on the importance of sexual history and sexually transmissible agents in cervical cancer has been reported. Case-control studies have frequently demonstrated increased risk of cervical cancer for women using oral contraceptives, while laboratory results have shown that vaginal spermicides inactivate various sexually transmissible agents. To determine the role of contraceptive use in cervical cancer, 153 cases of Maryland women with invasive cervical cancer and age, race, and residence-matched controls were interviewed in 1985, focusing on sexual history, health care utilization patterns, screening history, contraceptive use, and smoking. Overall, lifetime use of contraceptives was protective of cervical cancer (odds ratio (OR) = 0.38, 95% confidence interval (CI) = 0.2-0.7). Use of oral contraceptives (OR = 0.48), diaphragm (OR = 0.29), and vaginal spermicides (OR = 0.28) were more frequent in controls than cases. After adjustment for behavioral factors (age at first intercourse, smoking, gaps in Papanicolaou smear testing, and obstetrician-gynecologist visits), use of vaginal spermicides remained significant (OR = 0.30), although use of oral contraceptives and barrier methods of contraception failed to remain significant. The effectiveness of vaginal spermicides in preventing cervical cancer may be due to their antiviral action.”

► Which study design was used? ► Cross-sectional ► Case-control ► Retrospective cohort ► Prospective cohort

Source: Celentano, Klassen, Weisman, & Rosenshein. (1987). Am J Epidemiol, 126(4), 592–604.

4

The Role of Contraceptive Use in Cervical Cancer: The Maryland Cervical Cancer … Study—2

► Which study design was used? ► Case-control

Source: Celentano, Klassen, Weisman, & Rosenshein. (1987). Am J Epidemiol, 126(4), 592–604.

5

Cancer Preventive Screening Behavior among Elderly Women—1

“Screening procedures and recommendations for early detection of cervical and breast cancer are reviewed with special reference to suggested guidelines for screening among the elderly. Data on preventive practices related to cancer detection were obtained from 675 women in Maryland by telephone interviews. The elderly (65 years of age and older) reported fewer detection tests of any type; specifically, 23% report never having had a Pap test and an additional 28% have not had one within 5 years, compared with women under the age of 65, where 98% report having had a Pap test and 79% have had one within the past two years. Elderly women were also less likely than younger women to report receiving routine breast examinations by their physicians, to perform BSE, or to have been taught BSE by a health professional. Type of provider utilized, having a personal physician, and going to a gynecologist account for a significant proportion of the variance in screening practices.”

► Which study design was used? ► Cross-sectional ► Case-control ► Prospective cohort ► Randomized trial

Source: Celentano, Shapiro, & Weisman. (1982). Prev Med, 11(4), 454–63.

6

Cancer Preventive Screening Behavior among Elderly Women—2

► Which study design was used? ► Cross-sectional study

Source: Celentano, Shapiro, & Weisman. (1982). Prev Med, 11(4), 454–63.

7

Risk Factors for HIV-1 Seroconversion among Young Men in Northern Thailand—1

“Objective: To identify behavioral and sociodemographic risk factors for incident human immunodeficiency virus-1 (HIV-1) infection among healthy young men in northern Thailand. Design: Men inducted into military service in northern Thailand in May and November 1991 were followed at 6-month intervals until discharge 2 years later. Trained nonmilitary interviewers identified risk factors for HIV- 1 infection through interviews with the men. Setting: Thirteen military bases in northern Thailand. Participants: A total of 1932 seronegative men, aged 19 to 23 years (average age, 21 years) at enrollment, conscripted into the Royal Thai Army and Air Force from six upper-northern Thai provinces. Main outcome measures: Human immunodeficiency virus-1 seroincidence as determined through enzyme-linked immunosorbent assay and verified by Western blot, and univariate and multivariate analyses of risk factors related to HIV-1 incidence.”

► Which study design was used? ► Cross-sectional ► Case-control ► Prospective cohort ► Randomized trial

Source: Celentano et al. (1996). JAMA, 275(2), 122–7.

8

Risk Factors for HIV-1 Seroconversion among Young Men in Northern Thailand—2

► Which study design was used? ► Prospective cohort study

9

Hearing Impairment and … Dementia and Cognitive Decline in Older Adults: The Health ABC Study—1

“Background: Age-related peripheral hearing impairment (HI) is prevalent, treatable, and may be a risk factor for dementia in older adults. In … analysis, we quantified the association of HI with … dementia and with domain-specific cognitive decline in memory, perceptual speed, and processing speed.

Methods: Data were from the Health, Aging and Body Composition (Health ABC) study, a biracial cohort of well-functioning adults aged 70- 79 years. Dementia was defined using a prespecified algorithm incorporating medication use, hospital records, and neurocognitive test scores. A pure-tone average in decibels hearing level (dBHL) was calculated in the better hearing ear using thresholds from 0.5 to 4kHz, and HI was defined as normal hearing (≤25 dBHL), mild (26-40 dBHL), and moderate/severe (>40 dBHL). Associations between HI and … dementia and between HI and cognitive change were modeled using Cox proportional hazards models and linear mixed models, respectively.”

► Which study design was used? ► Cross-sectional ► Nested case-control ► Retrospective cohort ► Prospective cohort

Source: Deal et al. (2017). J Gerontol A Biol Sci Med Sci, 72(5), 703–9. http://dx.doi.org/10.1093/gerona/glw069

10

Hearing Impairment and … Dementia and Cognitive Decline in Older Adults: The Health ABC Study—2

► Which study design was used? ► Prospective cohort study

Source: Deal et al. (2017). J Gerontol A Biol Sci Med Sci, 72(5), 703–9. http://dx.doi.org/10.1093/gerona/glw069

11

A … Feasibility Pilot … of Hearing Treatment for Reducing Cognitive Decline: Results from the Aging and Cognitive Health Evaluation … —1

“Introduction: Hearing loss (HL) is prevalent and independently related to cognitive decline and dementia. There has never been a … to test if HL treatment could reduce cognitive decline in older adults. Methods: A 40-person (aged 70-84 years) pilot study in Washington County, MD, was conducted. Participants were randomized 1:1 to a best practices hearing or successful aging intervention and followed for 6 months. … Results: The Aging and Cognitive Health Evaluation in Elders Pilot (ACHIEVE-P) Study demonstrated feasibility in recruitment, retention, and implementation of interventions with no treatment-related adverse events. A clear efficacy signal of the hearing intervention was observed in perceived hearing handicap (mean of 0.11 to -1.29 standard deviation [SD] units; lower scores better) and memory (mean of -0.10 SD to 0.38 SD). Discussion: ACHIEVE-P sets the stage for the full-scale ACHIEVE … (N = 850, recruitment beginning November 2017), the first … to determine efficacy of a best practices hearing (vs. successful aging) intervention on reducing cognitive decline in older adults with HL.”

► Which study design was used? ► Cross-sectional ► Case-control ► Prospective cohort ► Randomized trial

Source: Deal et al. (2017). Alzheimers Dement (N Y), 3(3), 410–5. http://dx.doi.org/10.1016/j.trci.2017.06.003

12

A … Feasibility Pilot … of Hearing Treatment for Reducing Cognitive Decline: Results from the Aging and Cognitive Health Evaluation … —2

► Which study design was used? ► Randomized trial

Source: Deal et al. (2017). Alzheimers Dement (N Y), 3(3), 410–5. http://dx.doi.org/10.1016/j.trci.2017.06.003

13

Association between Plasma Cholesterol and Prostate Cancer in the PSA Era—1

“Abstract: We previously found that statin users had a lower risk of advanced and possibly high-grade prostate cancer compared with nonusers. We hypothesize that statins' effects on cholesterol synthesis may explain those findings because prostate cancer cells exhibit cholesterol dysregulation. Thus, we investigated whether low plasma cholesterol is associated with prostate cancer overall and by stage and grade. Participants were drawn from the 18,018 members of the Health Professionals Follow-Up Study who provided blood in 1993-1995. We ascertained 698 incident cases through January 2000. Controls were 698 men who had a PSA test and were matched to cases. Plasma cholesterol was measured enzymatically. Conditional logistic regression was used to estimate multivariable ORs and 95% CIs of total, clinically organ-confined (n = 518), advanced (T3b or worse; n = 61), low- grade (Gleason sum < 7; n = 386) and high-grade (Gleason sum > or = 7, n = 247) disease. Low cholesterol (<25th percentile vs. > or =25th percentile) was not associated with total (OR = 0.93, 95% CI: 0.72-1.20), organ-confined (OR = 0.87, 95% CI: 0.64-1.18) or low-grade (OR = 1.06, 95% CI: 0.75-1.51) disease. However, men with low cholesterol had a lower risk of high-grade disease (OR = 0.61, 95% CI: 0.39-0.98), especially if organ-confined (OR = 0.54, 95% CI: 0.29-0.99). The association for advanced disease appeared inverse, but number of cases was small (OR = 0.42, 95% CI: 0.13-1.36). Associations remained after excluding cholesterol-lowering drug users. These results coupled with prior statin findings suggest that mechanistic studies on cholesterol metabolism should be pursued to understand a possible target for preventing poorly differentiated prostate cancers.”

► Which study design was used? ► Cross-sectional ► Case-control ► Nested case-control ► Prospective cohort

Source: Platz, Clinton, & Giovannucci. (2008). Int J Cancer, 123(7), 1693–8. http://dx.doi.org/10.1002/ijc.23715

14

Association between Plasma Cholesterol and Prostate Cancer in the PSA Era—2

► Which study design was used? ► Nested case-control study

Source: Platz, Clinton, & Giovannucci. (2008). Int J Cancer, 123(7), 1693–8. http://dx.doi.org/10.1002/ijc.23715

15

You Should Be Able to…

► Distinguish between and identify epidemiologic study designs