Overview of Study Design University of Nairobi Winter 2013 Ann Vander Stoep, PhD Psychiatric Epidemiologist

What is research design A research design is the arrangement of conditions for collection analysis of data in a manner that aims to combine relevance to the research purpose with economy in procedure It is a strategy specifying which approach will be used for gathering and analyzing the data

Types of Studies Descriptive vs. Analytic Descriptive Studies: Descriptive studies present general characteristics of the distribution of a disease (person/place/time) Exploratory: No specific hypotheses Analytic Studies: Analytic studies test a specific hypothesis to determine whether a factor influences the risk of a disease Usually compare two groups (diseased vs. non- diseased or exposed vs. non-exposed)

Approach to research Quantitative vs. Qualitative Qualitative is based on measurement of quantity or amount Qualitative research is concerned with quality or kind e.g why do the people think and do certain things(concerned with assessment of attitudes, opinion and behavior

Research approach Prospective: follow-up of subjects to determine outcome status occurs after study begins Retrospective: follow-up time between exposure and outcome has already occurred when study begins

Characteristics of Studies Unit of study Individuals Groups (Ecological Study) –Don’t know whether diseased individuals actually had the exposure or not

Characteristics of Studies Time Single point in time (Cross-sectional) Multiple points in time (Cohort, Case-Control)

Descriptive Studies Purpose “Provide information on patterns of disease occurrence within populations according to characteristics such as age, gender, race,…” Used when: little is known about causes of a disease objective is to generate hypotheses about etiology objective is to follow disease trends over time early stages of disease investigation

Types of Descriptive Studies 1.Case report No denominator 2.Case series 3.Population-based studies Prevalence IncidenceDenominator Mortality

Incidence Incidence and prevalence are both measures of the extent of disease in a population. Incidence tells us about a change in status from non-disease to disease, thus being limited to new cases.

Prevalence Prevalence includes both new cases and those who contracted the disease in the past and are still surviving.

The Numerator Both prevalence and incidence consist of a numerator and a denominator. The numerator of an incidence measure is a count of the newly developed cases that arise in a population over a specified time interval such as one year.

Numerator The numerator of a prevalence measure consists of a count of the total number of individuals in a population suffering from the disease at a point in time.

The Denominator The denominator of both measures consists of the number of persons in the population from which cases of disease arise, for example, the population of a city, county,state or country.

Descriptive Studies Descriptive studies provide the data and information that become the foundation for the development of hypotheses…. …leading to… …Analytic studies

Analytic Studies Purpose To investigate the relationship between an exposure and a disease outcome

Types of Analytic Studies Observational Studies Researcher systematically “observes” the association between “exposure” and “outcome” Cross-sectional studies, Cohort studies, Case-control studies Document naturally occurring associations Experimental Studies Investigator assigns “exposure” of interest Clinical trials, Community trials Randomized, non-randomized (quasi-experimental)

Observational Analytic Studies Cross-Sectional Studies Purpose: To assess relationships between “exposure” and “outcome” (risk factor and disease) at one point in time.

Observational Analytic Studies Cross-Sectional Studies Procedure: 1.Measure “disease” (outcome), exposure, and other factors simultaneously in defined population 2.Identify individuals with/without condition and assess presence/absence of “exposures” 3.Compare frequency of exposure among “diseased” and “non-diseased”

Observational Analytic Studies Cross-Sectional Studies In cross-sectional studies, it is not possible to determine the temporal sequence (which came first). Did the exposure precede the disease??? Or did the disease precede the exposure???

Observational Analytic Studies Cohort Studies Purpose: “A cohort study involves comparing disease incidence over time between groups that differ on their exposure to a factor of interest.” (Koepsell and Weiss, 2003)

Observational Analytic Studies Cohort Studies Procedure: 1.Subjects without disease are classified on the basis of exposure to potential risk factor: “Exposed” / “Unexposed” 2.Observed over time to determine who develops disease 3.Disease incidence compared in Exposed and Unexposed 4.Yields “Relative Risk”

Observational Analytic Studies Cohort Studies Used when: Exposure is uncommon Outcome is not too uncommon Exposure may be difficult to recall accurately if assessed long afterwards Objective is to investigate many outcomes of one exposure

Observational Analytic Studies Case-Control Studies Purpose: To investigate hypotheses about disease etiology (causes, risk factors, etc.) “At first, a case-control study may appear to be attacking the problem backwards, by proceeding from effect to cause.” (Koepsell and Weiss, 2003)

Observational Analytic Studies Case-Control Studies Procedure: 1.Subjects selected based on disease status: Cases (with disease); Controls (without disease) 2.Exposure history ascertained 3.Frequency of exposure compared in Cases and Controls 4. Yields “Odds Ratio” (an estimate of the Relative Risk)

Observational Analytic Studies Case-Control Studies Used when: Disease is uncommon in general population Long latency period between exposure and disease Want to investigate many different exposures for one outcome Accurate recall is possible or records of exposure are available

Experimental Analytic Studies Purpose: Test hypothesis regarding etiology of disease Evaluate method of intervention or prevention Eliminate bias by individual subject characteristics

Experimental Analytic Studies Procedure: 1.Form study groups from same underlying population 2.Researcher controls conditions under which subjects are exposed 3. Researcher imposes the exposure (e.g. intervention or level of risk factor)

Experimental Analytic Studies Used when: much is known about epidemiology of disease want to evaluate clinical treatment options want to evaluate prevention approaches

Experimental Analytic Studies Randomized Trials Randomized [clinical or controlled] trials: Subjects randomly assigned to Experimental or Control groups “Gold standard” of epidemiological studies Eliminate selection bias

Experimental Analytic Studies “Quasi-Experimental” Studies Quasi-experimental studies: Subjects (or subject groups) not randomized Researcher imposes (or does not impose) exposure (intervention) on groups

Experimental Analytic Studies Community Trials Community trials: Entire population of community does or does not receive intervention Communities may be randomized or non- randomized to intervention or control condition Same approach might be applied to classrooms, schools, neighborhoods, regions (groups, not individuals)

Ann takes over the exercise

Toxic Shock Syndrome (TSS) Clinical description: “Typically begins suddenly with high fever, vomiting, and profuse watery diarrhea, sometimes accompanied by sore throat, headache, and myalgias. The disease progresses to hypotensive shock within 48 hours, and the patient develops a diffuse, macular, erythematous rash with non-purulent conjunctivitis. Urine output is often decreased, and patients may be disoriented or combative…” MMR; May, 1980

Toxic Shock Syndrome (TSS) Laboratory: “elevated blood urea nitrogen…white blood cell counts…; in 33 of 45 (73%) patients cultured, Staphylococcus aureus was isolated from the throat, cervix, vagina, or rectum”

Descriptive Epidemiology TSS: Descriptive Epidemiology Todd et al. described toxic shock syndrome (TSS) in 7 children 2. May 1980 – “Since October 1, 1979, 55 cases have been reported to CDC. Fifty-two of these (95%) have been in women. Of 40 patients in whom a menstrual history was obtained, 95% had onset of illness within the 5-day period following onset of menses…. Seven deaths have occurred, for a case fatality ratio of 13%.”

TSS: Case-Control Study 1980 – “Because of the striking association with menses, …the CDC administered a telephone questionnaire to 52 women who had illness meeting a clinical case definition of TSS and 52 age- and sex-matched controls. …No significant differences were found between the patients and controls in marital status, contraceptive methods, brand of tampon or sanitary napkin used, absorbency…However, 50 of 50 cases with onset during menstruation (100%) used tampons as compared to 43 of 50 controls (86%) (p=.02). Among case-control pairs who used tampons, more cases used tampons at all times (day and night) while menstruating (p <.05).”

TSS: Case-Control Study “Eighty women who had onset of TSS between October 1, 1979, and September, 1980, and 160 age- and sex- matched neighborhood controls participated in this study to evaluate risk factors associated with TSS in women. Of the 80 women, 76 had onset of illness during their menstrual periods. The odds ratio for developing menses-associated TSS with any use of tampons compared with no use of Tampons was (p < 0.001). When exclusive use of a particular tampon brand was compared with exclusive use of all other brands, Rely was the only brand with a significantly increased odds ratio (2.49; p=0.005)”

TSS: Cohort Study A group of women (n = 50) who had used Rely Tampons during the previous year were compared with another group of women (n=50) who had used other brands of tampons during the previous year. Of the Rely Tampon users, 5 had developed TSS; of the other tampon users, only 2 developed TSS. Women who use Rely tampons are at increased risk of TSS compared to women who use other types of tampons (Relative Risk = 2.5)

TSS: Randomized trial 100 women, aged 20-40, who were not pregnant and had regular menses, were recruited for this experimental study. After being informed of the potential risks, they consented to participate. The women were randomly divided into two groups. One group was provided Rely tampons and told to use them during the next year; the other group was provided with another brand of tampons. At the end of one year, 10 of the women in the Experimental group had dropped out of the study. Of those remaining, 4 had developed TSS. Of the women in the Control group, 2 had developed TSS.

STUDY DESIGNS DESCRIPTIVEANALYTIC * Case Report OBSERVATIONAL EXPERIMENTAL * Case Series * Cross-sectional *Comparative trial * Survey * Cohort * RTC * Case-control Ecological or Individual unit of study