BBH 440 exam 3 Epidemioliogy

Determination of exposure status

- After the selection of cases and controls, researchers must then ascertain the exposure status of the individuals in each group. This information can be obtained through various ways, including questionnaires, review of medical records, interviews, etc. In situations where the identified case has died in the process of organizing the study, sometimes reports from individuals close to them (e.g. family members) are used in place of an interview with the case herself. Of course, this is not ideal as it creates more possibilities for errors in recall.

disadvantages of ecological studies

- No cause-effect interpretation possible - Possibility of ecological fallacy - still a descriptive observational design that doesn't provide us with the ability to make a determination of causality

other threats to validity (ecological studies)

Because of how prone the interpretation of ecological studies is to the ecological fallacy, we have to be very careful in interpreting the results of ecological studies. There are many potentially confounding variables in an ecological study (e.g. demographic characteristics like age, other changes taking place in society) that may better explain the relationship between our variables of interest. For this reason, ecological studies on their own are also not effective at demonstrating causality. Ecological studies are generally inexpensive and easy to conduct. They allow us to justify further investigations or maybe to generate new hypotheses, but they are still a descriptive observational design that doesn't provide us with the ability to make a determination of causality.

A research group conducts a one-time survey of over 5,000 individuals that includes questions on the number of alcoholic beverages consumed per day and number of cigarettes smoked per day. The group analyzes the data and finds a positive association between the two variables. This is an example of a/an:

Cross-sectional study - The data for this study was collected at one point in time- with individuals. It would not be passive surveillance because there is an active solicitation of data going on. A case series is only a description of a small group of individuals. An ecological study would involve aggregate data, not data on the individual level.

the "gold standard" for research

Due to the amount of control researchers have over an experimental study, this study design is considered the "gold standard" of research. A good experimental study can allow us to state with a significant degree of confidence that an exposure causes an outcome, if that's what the results of the study suggest. Thus, an experimental design has the most explanatory power of any design we have discussed so far.

A research group obtains school district data on the number of students with intellectual disabilities and also obtains city data on the level of certain pollutants in the water supply. They find that the level of water pollutants is positively correlated with the number of students with intellectual disabilities. This is what type of study?

Ecological study

An investigator obtains data on mortality rates due to home accidents in 20 states as well as data indicating the total liquor sales per year in those 20 states. The investigator finds a correlation between deaths due to home accidents and total liquor sales. This is what type of study?

Ecological study - the investigator in this case is using group data, rather than describing a single case or providing individuals an opportunity to report some health events. Thus, the answer is ecological study

"correlation does not imply causation"

Just because two variables are CORRELATED, that does not mean one causes the other. It simply means that they are associated in some way. - commonly quoted phrase in science - we are inclined to think one event causes another simply because the two events are correlated, or they change in a predictable way in relation to one another. However, we cannot say thing A causes thing B simply because they are correlated with one another because there may be a multitude of other explanations for the perceived relationship between A and B.

Humans as pattern seekers

Looking for patterns in our environment (even though they may not really be there). - We see things that aren't really there because our brains are so focused on identifying patterns. This can be problematic in many areas of life, but it also creates problems in trying to understand the causes of disease. We have a tendency to believe that when two things occur close to one another in time, or when they change together (e.g. as one increases, so does the other), that one of those things is causing a change in the other. However, this is not always the case.

A group of researchers conduct a cohort study that examines the relationship between blood cholesterol levels and diabetes. At the end of the study they decide to conduct an additional study to explore another hypothesis that a particular genetic background increases the risk of diabetes. They take a sample of participants who developed diabetes during their cohort study and a sample who did not, then they compare the likelihood of having the particular genetic background between both groups. This additional study is what type of study?

Nested case-control study

A research group conducts a study to examine the relationship between unemployment and drug use. They get unemployment numbers from state governments for 6 states and they obtain numbers regarding number of drug arrests in those same 6 states. They find a positive correlation between unemployment and drug arrests. They conclude that being unemployed makes it more likely someone will use and/or be arrested for drug use. Is the logic they used accurate?

No

A researcher publishes a case series that describes 18 cases of a previously unknown disease. Based on the information in the case series, the author concludes that the disease is caused by excessive sun exposure. Disregarding any of the potential details of the case series, is the researcher able to make this conclusion?

No

A study is conducted that involves surveying participants about their diet, exercise level, and weight early in the year and then following up with them and again asking about their diet, exercise level, and weight in December of the same year. Is this a cross-sectional study?

No

A research group is examining the relationship between long-term smoking (5+ years) and blood pressure. They recruit a group of long-term smokers and a group of non-smokers, then measure their blood pressure. What general experimental approach is this considered?

Observational

disadvantages of case studies/case series

One thing case studies and case series are not useful for, however, is making inferences about what caused a disorder. Because in a case study or series we are only dealing with one patient or one group of patients, we don't have multiple groups to compare to one another to be able to make determinations about what makes the affected group different.

An organization provides a phone number and web site that individuals can use to voluntarily report any allergic reactions they experience after using cosmetic products. This is what type of surveillance?

Passive

Use of the odds ratio

The statistic often used for testing the hypothesis and for quantifying the association between the exposure and outcome in a case-control study is the odds ratio. Remember from our previous discussion of the odds ratio that an odds ratio of 1.0 indicates there is no relationship between the outcome and the exposure. In a case-control study, the null hypothesis would then be that the odds ratio is equal to 1.0. Therefore, if we are to determine that a significant relationship exists between our exposure and outcome, the odds ratio must be statistically significantly above or below 1.0.

Researchers conduct a cohort study of the relationship between daily alcohol consumption and motor vehicle accidents and find the relative risk of getting in an accident for the alcohol drinkers compared to non-drinkers is 4.5 with a confidence interval of 3.2-6.9. This suggests that:

daily alcohol consumption is associated with an increased risk of motor vehicle accidents

Misclassification (case-control study)

occurs when participants are misclassified as to their exposure or outcome status. Imagine, for example, that researchers are attempting to study the relationship between depression and obesity (they hypothesize that depression is an exposure that will increase the likelihood of obesity as an outcome). To identify patients as "depressed," the researchers use the results from a survey that contains the question, "Have you experienced extreme feelings of sadness in the past 30 days?" They classify participants who answer yes as having the exposure and those who answer no as not having the exposure. Do you see any problem with this approach?

disadvantages of cross-sectional studies

only tells us about things we can detenct now; Due mostly to this feature, cross-sectional studies are not effective at demonstrating causality. Due to their one-point-in-time design, they don't do a good job at helping us to show that one event preceded (and thus caused) another

The Vaccine Adverse Event Reporting System (VAERS) is a program set up by the CDC. People who experience adverse events associated with vaccines can call VAERS or go to their website and report their experience if they choose to. This is what type of surveillance?

passive - in active surveillance, researchers actively solicit information from individuals in the population. In passive surveillance they provide people with a way to report information if they desire to

Recall bias (case-control studies)

refers to the tendency for recollections of past information to be incomplete or inaccurate. Because many case-control studies rely on participants' reports of their exposure status, recall bias is often a problem. For example, when asking a participant about her dietary practices ten years ago, it is likely she will not be able to remember some details accurately. She may, in that case, either not provide the information or attempt to provide information and do so inaccurately. Using inaccurate information regarding exposure status can affect the ability of the researchers to see a relationship between the exposure and outcome.

clinical trials

represents the type of experiment you would probably think of if I asked you to imagine what a typical experiment looks like. It involves 1) selecting a sample of human participants, 2) exposing about half of them to the independent variable, and then 3) comparing the two groups to see if the exposure resulted in a greater change in the exposed group than was seen in the non-exposed group. In a clinical trial, this is all done with the purpose of determining the effect the independent variable will have on human health.

Selection of cases (case-control study)

requires that investigators first decide on some criteria they can use to define a case and then that they identify cases in the population using those criteria. - A case is a participant who has the outcome (e.g. a disease of interest), a control does not.

Ecological studies

type of cross-sectional study- has one feature that distinguishes it from other cross-sectional studies (like surveys). Instead of making use of data from individuals, like you would with a survey, an ecological study uses data that describes groups of individuals. ecological studies use group data, not individual data. This will allow you to tell them apart from something like a cross-sectional survey. - Sometimes the statistics for the individuals in the group are not available, or sometimes they may be much more difficult to obtain than the group or aggregate data. Thus, an ecological study often makes use of aggregate data that is readily available; for this reason, ecological studies can be relatively inexpensive ways of conducting a preliminary investigation into a relationship of interest.

advantages of surveillance studies

used solely as descriptive epidemiological studies—to tell epidemiologists what is going on the population so that information can be used to monitor the population's health. Because no variables are being manipulated, surveillance studies are also observational.

advantages of case studies/case series

useful in identifying new disorders or alerting people's attention to some new risk factor for a disorder. The information in them can be used to generate hypotheses or ideas for future studies.

Cross-section definition

what we obtain when we make a slice through an object. it allows us to look at the inner workings of the object we sliced through, but only at the point of which we made the slice. example: - look at this cross-section made by slicing through an orange. You can see the plane at which the slice was made, but nothing below it; if you looked at the other end of the orange you would see a similar thing: only the plane where the slice was made would be visible.

Cross-sectional studies

where we take a cross-section of the population at one point in time. we gather data about our sample once, in the present, but we dont do any follow-up studies to find out more about what they look or what what happens to them in the future; often take the form of surveys. Any time you've completed a survey that didn't involve some other follow-up survey at some point in the future, you've taken part in a cross-sectional study.

experimental studies

you are actually changing the exposure status of the participants. In other words, you (the researcher) is creating the difference between the treatment and control groups by giving one group the drug and the other a placebo. You are manipulating the independent variable to see if that manipulation creates a difference in the dependent variable. example: - you want to conduct a test to see if a particular medication is effective at reducing high blood pressure; the hope is that the medication can one day be used as an anti-hypertensive drug. It probably isn't that difficult for you to think of a way to test the medication. A common approach would be to: 1) take a group of people who have high blood pressure, 2) measure their blood pressure, 3) give half of them the medication while the other half receives a placebo, and 4) measure their blood pressure again. If the drug worked, you would expect to see a greater reduction in blood pressure in the group who received the drug than you would in the group who received the placebo.

Response bias (cross-sectional studies)

you might simply run into the situation where responses over a survey may not be the most accurate representations of reality - might be due to participants not remembering the correct answers, or perhaps due to them being reluctant to disclose information they deem sensitive.

cohort study procedure

1. Researchers form at least two groups based on whether or not individuals have had exposure to the independent variable 2. Researchers follow-up with the participants in both groups for an extended period of time to see if the condition they are interested in develops 3. Researchers compare the incidence of the health condition over the course of the study in both groups to see if the exposed group was more or less likely to develop the condition.`

Causation

A cause and effect relationship in which one variable controls the changes in another variable.

Correlation

A measure of the extent to which two factors vary together, and thus of how well either factor predicts the other.

Case-control study done with cohort study data

A nested case-control study is conducted just like any other case-control study, with the exception that it is done as part of a cohort study. Sometimes investigators in cohort studies plan to conduct one or more nested case-control studies over the course of the study, and sometimes (like in the example above) they are planned in response to some interesting data that emerges over the course of the cohort study. Despite being conducted using data from a cohort study, nested case-control studies still suffer from the same biases as case-control studies and thus have the same strengths and weaknesses.

retrospective study

A study that monitors people who have been exposed to an environmental hazard at some time in the past.

Causality or causation

a relationship between two variables where one is understood to effect change in the other - almost every study in epidemiology is concerned in some way (indirectly or directly) with causality example: - investigators may set out to determine the exposure that led to an outbreak of food-borne illness. In this situation, they are interested in discovering the cause of the illness.

cohort study design

a study in which one or more study groups (cohorts) are selected and their outcomes analyzed throughout a follow-up period (can be prospective or retrospective).

Analytic studies

allows us to ask why the population is the way it is; they involve the use of multiple groups that we can compare to determine what makes one group different from the other. - the purpose of analytic studies is to find out why something is occurring, so an analytic study would allow us to begin to explore the causes underlying this increase in obesity shown above. example: - we could use an analytic study design to test the hypothesis that increased availability of high-fat foods increases the risk of obesity and we could use the results of this study to help to explain the changes seen in the figure above. Analytic studies can be either observational or experimental.

disadvantages of cohort studies

- a cohort study on its won cannot determine that an exposure causes an outcome - biases; researchers typically like to see that study repeated with a different sample or if possible, using an experimental study - loss to follow up; Due to the long-term nature of cohort studies, it is almost inevitable that you will not end the study with the same number of participants you began with. Some participants may move out of town, others may die, and others may simply lose interest in the study and stop responding. When participants are unavailable for follow-up for any reason, and researchers are unable to obtain data regarding their health status, it is known as loss to follow-up. - misclassification and confounds - not well suited for studying very rare diseases - prospective cohort studies are very time-consuming and expensive; can last decades

Sampling bias (cross-sectional studies)

- a survey may be less likely to reach some participants than others; for example, a survey administered over the internet will only reach those who have internet access. - those with internet access might be different in other ways from those without it (e.g. of higher socioeconomic status).

Common causal variable

- a variable that is not part of the research hypothesis but that causes both the predictor and the outcome variable and thus produces the observed correlation between them example: - as ice cream sales have increased at the largest ice cream retailer in town, drowning deaths have also increased. You immediately begin to suspect that this ice cream retailer's ice cream is so heavy with milk fat that it is making it impossible for people to swim after they eat it! Of course, this isn't true. Can you think of a third variable that might affect both ice cream sales and the frequency of drowning deaths? Temperature. Ice cream sales go up in the summer and so do drowning deaths (because people are swimming more). In this situation, instead of A causing B, there is a C—a third variable outside of our relationship of interest that is having an effect on that relationship.

advantages of cohort studies

- best designs in terms of their ability to provide convincing evidence of a relationship between an exposure and an outcome - allow researchers to obtain current data on exposure status and identify the occurrence of the outcome fairly close to when it actually happens. - can be used to support statements about a causal relationship between an exposure and an outcome

two common types of experimental studies

- clinical trials - community trials

advantages of case-control studies

- inexpensive and relatively easy to conduct; especially true when case and or exposure information is readily available through public databases - also work well for diseases that are not very common (rare diseases); odds ratio is a better approximation of risk when the prevalence for the disease is low and because the other major observational analytic design like cohort study is difficult to conduct when diseases are not commonly seen in the population.

Prospective cohort study

- most common cohort study A research study that follows over time groups of individuals who are alike in many ways but differ by a certain characteristic (for example, female nurses who smoke and those who do not smoke) and compares them for a particular outcome (such as lung cancer). - occurs over a period of time, cohort studies are sometimes also called longitudinal studies

Hormone replacement therapy (HRT) - as an example of how correlation may not always indicate causation

- previous thought that HRT would lower risk of CHD, stroke and breast cancer until a 9 year study was done after the WHI study was done years before, which proved that HRT did not lower the risk of CHD, stroke and breast cancer in these women and it actually increased their risks. (could be due to other factors) The quagmire surrounding the reversal of opinion about HRT's potential health benefits, however, exemplifies the difficulties with inferring causality prematurely. Because in most studies there is always the possibility that confounding variables or biases may be influencing the results, it becomes difficult to make statements about causality until a subject has been thoroughly researched using high-quality experimental designs. Thus, although making a determination of causality is the goal of any researcher, that goal is especially elusive.

Observational study

- the investigator is not actually changing the exposure status of his or her participants. Instead, at the start of the study some individuals have the exposure and some do not. Then, the occurrence of the outcome is compared between these two groups. example: - one way to study smoking during pregnancy and behavioral problems in offspring would be to recruit a sample of pregnant women---some who smoke and some who don't. Then, you would compare the frequency of behavioral problems in the children of the smokers to the frequency of behavioral problems in the children of the non-smokers. In this example, you are simply observing to see if there are differences between the offspring of smokers and non-smokers.

descriptive studies

- the researcher is reporting data that describes a particular situation of interest; useful in providing information about populations but they don't generally allow us to ask why the population is the way it is (thats what analytic studies do) example: - take a look at this image that shows obesity prevalence in the United States in 1990, 2000, and 2010. This image doesn't help us to understand why obesity prevalence has increased significantly over this 30-year period; instead, it simply describes the situation for us. It tells us what percentage of adults in each state are obese at each of these time points.

Case-control study processes

- the selection of cases - the selection of controls - the determination of exposure status

Coincidence

- two or more things occurring at the same time by chance example: - there is a statistical oddity called the Redskins Rule, named after the Washington Redskins, a National Football League team. For every election that occurred in the United States from the year 1940 through 2000, the outcome of the Washington Redskins' last home football game prior to the election correctly predicted the winner of the election. If the Redskins won, the incumbent party retained control of the White House. If the Redskins lost, the opposition party won - There is clearly a correlation here between the outcome of the Redskins' last home game before the election and the outcome of the election, but do we really think one is causing the other? No, this is a complete coincidence. Sometimes correlations don't represent a real relationship between two variables at all—it just appears that they do.

Reverse causation

- when variable 1 is assumed to cause variable 2, yet the opposite direction of causation may be the case - relationship in which the correlation between two things is due not to the first thing causing the second, but to the second causing the first

advantages of case-crossover study

- works well with exposure that is transient and also ones that you were expecting to immediately precede the outcome. - cases and controls are all the same people, thus, confounding effects in a typical case-control study that might arise from differences between the types of individuals found in the case and control groups are less likely to occur in a case-crossover studies. - case-crossover studies are a popular choice of study design when exposures are transient or intermittent and the onset of the health condition being studied is abrupt.

disadvantages of case-control studies

-measurement of exposure may be inaccurate -representativeness of cases and controls may be unknown -provide indirect estimates of risk -the temporal relationship between exposure factor and outcome cannot always be ascertained - biases

case-crossover studies

-Only cases -Compare exposures for "Case time" vs. "control time" - each individual in the study serves as both a case and a control. This may sound a bit confusing; it might help if we looked at an example to help illustrate the study design. example: - Anger and Heart attacks -- To conduct the case-crossover study, you would select a group of individuals who have had a heart attack (note: everyone in this study has experienced the outcome). Then, you would identify a period of time immediately before the outcome as the "case" time. In this example, let's say you identified the case time as the 2 hours immediately before the person had a heart attack. Then, you would identify a "control" time. Let's say in this case it was a 2-hour period at the same time of the day on the day before the participant had a heart attack. - Next, you ask the participants about whether they experienced an episode of anger in the 2 hours before their heart attack and whether they experienced an episode of anger in the 2 hours on the day prior. What you would expect to see, if your hypothesis was correct, would be more episodes of anger (i.e. higher levels of exposure) during the case time than the control time. An odds ratio could be calculated to express this relationship.

Case-control study

A type of epidemiologic study where a group of individuals with the diseases, referred to as cases, are compared to individuals without the disease, referred to as controls; Remember that these studies start out by grouping participants into cases and controls. This will help you to tell them apart from cohort studies. - One approach to testing this hypothesis would be a case-control study. To conduct a case-control study, you recruit a sample of individuals—some of whom already have the outcome you're interested in (i.e. glioma), and some of whom do not. Those individuals who have the outcome are considered cases, while those who do not are considered controls. Ideally, your sample would be about 50% cases and 50% controls, although it rarely works out to be split this perfectly. After you have formed a sample that consists of both cases and controls, then you must in some way determine the level of exposure seen in each group

An organization solicits statistics on the number of diagnosed cases of prostate cancer by contacting hospitals throughout the country and asking for their data on prostate cancer cases. This is what type of surveillance?

Active

Researchers are exploring the relationship between trans fat in the diet and obesity. They recruit a sample of individuals; about half of them are obese and the other half are a healthy weight. Then they ask participants questions about their diet to ascertain their level of trans fat intake. They find that individuals who are obese are more likely to also consume high levels of trans fats. This is what type of study?

Case-control - this study involves taking a sample of people, half of whom already have the condition being studied as the outcome. thus, half of the sample are cases while the other half are controls. remember at the beginning of a cohort study, nobody should have the condition being studied as the outcome

A group of researchers are investigating the relationship between shoe size and reading ability in a sample of students that range in age from 6-14. They find a positive correlation between shoe size and reading ability and conclude that shoe size is causally related to reading ability. Which error of causation best describes this hypothesized relationship

Common causal variable

Observational Analytic Studies

Compares groups, but researcher has no control over assignment. There are instances where randomized controlled trial cannot be done due to ethical reasons, and therefore observational study is the best option (e.g., smoking and lung disease) - ecologic, case-control, cohort

Confounds (case-control studies)

Every study has the potential to be affected by confounds, or external factors that are associated with the variables of interest in the study. Because of both their observational and their retrospective natures, case-control studies are highly susceptible to confounds. In an observational study, what a person does in their life during the experiment is outside of your control. Thus, there may be many other factors that could contribute to the relationship between your exposure and outcome—many of which you may not be aware of. There is even less control over such things in a case-control study, because it is retrospective, meaning any exposure the participant had has already happened. Thus, you have to rely purely on their accurate reporting to you of any exposures they've had, which as we mentioned above is not always valid. - Due to the observational nature, retrospective design, and susceptibility to bias that case-control studies have, they are not adequate studies for demonstrating a causal relationship. Case-control studies can be very useful for identifying potential relationships and supporting the investigation of further research into such a relationship, but they are not used on their own to make statements of causality.

Cross-sectional studies can be used to make and cause & effect interpretation about an exposure and an outcome

False

Advantages of cross-sectional studies

Faster and cheaper than Longitudinal Studies; provides valuable descriptive information about the population or sample of interest; can be used to obtain estimates of prevalence; researchers can also use them to identify potential associations between exposures and outcomes- these potential associations can be followed up later with more advanced study designs

Selection bias in cohort studies (healthy user bias)

Healthy user bias occurs because those who are most likely to agree to take part in a study—and perhaps especially a long-term study like a cohort study—are probably healthier than the general population. They are probably more health-conscious and thus more likely to have healthier behaviors, follow medical advice, etc. Therefore, the effect of an exposure may be underestimated in these individuals since they will overall be less likely to develop a negative health outcome.

difference between prospective and retrospective cohort studies

In a retrospective cohort study, the group of interest already has the disease/outcome. In a prospective cohort study, the group does not have the disease/outcome, although some participants usually have high risk factors

Case-control study example: "effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): a case-control study.

Research question: studying the relationship between a number of risk factors and the occurrence of myocardial infarction cases: 15,000 from hospitals in 52 different countries; all cases had been admitted to the hosptial recently with symptoms from their first heart attack controls: 15,000 controls from hospitals and communities in the same countries How the exposure status was determined: some of the information they acquired via questionnaires (smoking status, fruit and vegetable intake), interviews with the participant (height, weight), or by collecting blood that was later tested (cholesterol levels) Odds ratio: found significant associations with the outcome of a heart attack for all of them; risk factors like current smoking and diabetes have odds ratios above 1.0, indicating an increased risk, while exposures like exercise and vegetables and fruits daily have odds ratios below 1.0, suggesting that these exposures are associated with a decreased risk of myocardial infarction

Cohort study example: "rotating night shifts and risk of breast cancer in women participating in the nurses' health study"

The study described in the abstract is part of a large effort known as the Nurse's Health Study. This study began in 1976 and is now recruiting a third group of participants. Each cohort so far has consisted of over 100,000 participants—all female nurses; the study has looked at a number of exposures like smoking, oral contraceptive use, alcohol intake, and diet, and their relationship with outcomes like cancer and heart disease. - second wave of the nurses health study; hypothesizing a hormone called melatonin that is produced at its highest levels in the middle of the night might have cancer-preventing properties, Melatonin's primary physiological role seems to be to help your body regulate its circadian rhythms (i.e. biological rhythms that occur on a 24-hours cycle), such as your sleep-wake cycle. Production of this hormone, however, is disrupted when an individual is exposed to bright light during normal nighttime hours. - the researchers were interested to see if women who worked rotating night shift were more likely to develop cancer. If they worked rotating night shift it meant that their work shifts were not always at night, but occurred throughout the night occasionally—in this study "occasionally" was defined as at least three nights per month. It is believed that having unpredictable disruptions to your circadian rhythms (e.g. several times a month) is more of a health threat than simply working night shift all the time, as in the latter scenario your body simply gets used to living on a "flipped" 24-hour schedule. The researchers used a sample of 78,562 women who were part of the Nurse's Health Study and had answered questions regarding rotational shift work but did not have cancer at the start of the study. The women were followed-up with for 10 years. Over that time, 2441 cases of breast cancer occurred among the women in the sample. The researchers found that women who worked 1-14 or 15-29 years on rotating night shift did not have a significantly increased risk of cancer. The relative risk of cancer for women who worked 30 or more years on rotating night shift, however, was 1.36 (1.04-1.78). Thus, women who worked for 3 decades or more on a rotating night shift schedule had a 36% increased risk of developing breast cancer. This example shows all of the classic features of a cohort study. The groups were formed based on the presence or absence of the exposure (i.e. working rotating night shift), there was a follow-up period, and incidence of the outcome over the follow-up period was compared between the exposed and non-exposed groups to formulate an estimate of risk. The group's hypothesis in this experiment was supported in the women with the highest levels of exposure.

ecological fallacy

assumes that a generalized cultural value applies equally well to all members of the culture - we can never draw conclusions about individuals within a group based on the summarized characteristics of a group. When we do, we make a mistake known as - ecological fallacy (threat to validity)

A researcher is interested in studying the relationship between alcohol intake and type 2 diabetes. Out of a sample of 500 people, she forms one group consisting of people who have type 2 diabetes and another consisting of people who don't. Then she gathers information about their alcohol intake. This is what type of study?

case-control study

Researchers want to investigate the link between cotards delusion, an extremely rare mental disorder where a person comes to believe they are dead, and previous psychiatric disturbances. which study would be better suited to investigate this?

case-control study

A research group is interested in studying the relationship between stress in social relationships and depressive episodes. They find a group of people who have had a depressive episode and ask them whether they experienced significant social stress in the week before their episode began. They also ask them if they experienced significant social stress in another 1-week period sometime the month before (when they didn't have a depressive episode). This is what type of study?

case-crossover study

public health surveillance

closely monitor the health of populations to watch for unusual events that might indicate a public health problem; a major responsibility of large epidemiological organizations like the WHO or CDC. Surveillance studies are organized efforts at public health surveillance, the goals of which are to monitor a population to gather information that can be used to prevent or control threats to health. Surveillance studies come in two broad forms: active or passive.

Selection of controls (case-control study)

controls used in a case-control study should be as similar as possible to the case subjects in a case-control study with one exception: they do not have the condition being studied. Ideally then, the controls are drawn from the same population the cases are drawn from but are representative of the general population in terms of the likelihood they have been exposed to the independent variable. They should be normal, healthy individuals that are similar in demographic characteristics (e.g. age, sex) to the cases, and they should be selected randomly from the population.

cross- sectional studies and causality

cross-sectional studies can suffer from various types of bias and they do not allow researchers to demonstrate with confidence that there is a causal relationship between two variables. They are, however, relatively simple to conduct (depending on the size of the sample and method of data collection), and can provide valuable information regarding statistics like prevalence. Thus, they are a widely used descriptive study design in epidemiology. The data collected via cross-sectional studies can also be used in more advanced study designs that do allow for the formal testing of hypotheses, which we will discuss in the next lesson.

A researcher conducts a survey that is used to find the prevalence of smoking in the United States. Based on this information alone, this is what type of research

desriptive

true or false: A cohort study is an experimental study, not an observational study.

false

disadvantages of passive surveillance

if you are waiting for people to report information to you then you are more likely to miss some cases of interest simply because people did not care to (or forgot to) report them. Thus, passive surveillance can cause researchers to under-estimate the true incidence of a particular problem in the population. Additionally, other factors may influence the number of reports that are made. example: - if the media reports on a potential adverse effect of a drug, then suddenly reports to MedWatch about those adverse effects will increase significantly. Some of those cases, however, may be due simply to individuals who take the drug having anxious reactions and believing they are experiencing negative side effects when really they may not be (a reaction known as the nocebo effect).

Cohort study example

imagine a group of researchers are interested in exploring the relationship between tanning bed use and the occurrence of malignant melanoma (a type of skin cancer). They hypothesize that tanning beds will increase the likelihood of cancer. - imagine a group of researchers are interested in exploring the relationship between tanning bed use and the occurrence of malignant melanoma (a type of skin cancer). They hypothesize that tanning beds will increase the likelihood of cancer. - At the end of the study, they find that the incidence of melanoma was much higher in the group who used tanning beds than the group who did not. They can express the relationship between tanning bed usage and melanoma using a measure like relative risk, which is a common measure of association used for cohort studies. Note that relative risk is used in cohort studies instead of an odds ratio because incidence is a statistic that can be obtained from a cohort design. When the measure obtained from a cohort study is incidence-density, then the rate ratio is used

advantages of ecological studies

inexpensive and easy to conduct - allow us to justify further investigations or maybe to generate new hypotheses

Passive surveillance

instead of the researcher(s) actively soliciting information regarding health, they provide the population of interest with a way of reporting such information voluntarily. example: - the Food & Drug Administration (FDA) has a system for reporting adverse events people experience after using over-the-counter or prescription drugs (as well as a variety of other products). If someone experiences, for example, an unusual side effect after taking a prescription drug, they can go to the FDA site MedWatch (Links to an external site.) and report the problem they experienced to the FDA. The FDA keeps track of these reported events and uses the information as an early indication system to notify them of potential problems with medications and other substances individuals are using. The reporting of adverse events through MedWatch has led to the FDA issuing warnings about drugs and even pulling drugs from the market.

Active surveillance

involves the regular collection of data from organizations or individuals; the individuals organizing the surveillance effort actively solicit the information. In other words, they will ask the organizations or individuals to provide them with the information. - Active surveillance efforts allow researchers to potentially reach large segments of the population, and thus can provide accurate representations of the factors affecting such populations. Because of this, they are useful sources of information about the health status of the population. They are, however, also relatively expensive and require a significant commitment of resources. Because active surveillance involves the active solicitation of information from a large number of individuals, it can be difficult to accomplish; for this reason, many active surveillance efforts are conducted by large organizations that have ample resources to do so, like the CDC. example: - the CDC every year conducts the largest telephone survey in the world, known as the Behavioral Risk Factor Surveillance System (BRFSS). To do so, they contact and conduct interviews with more than 400,000 individuals in all 50 states in order to monitor the prevalence of risk factors like smoking and conditions like obesity. The information from this large surveillance effort is used to identify particularly concerning trends in the United States and to organize prevention efforts to address the effects of such trends.

prevalence incidence bias in case-control studies (type of selection bias)

occurs because, if investigating a potentially fatal disease, then the individuals who are available to study are inevitably those who have survived long enough to be studied. There may be, however, other cases who have died recently from the disease because their disease progression was much quicker and more severe. These more severe cases would be those most likely to have had high levels of exposure to the independent variable (based on the researcher's hypothesis). Thus, the fact that these individuals are no longer alive can cause the relationship between the exposure and outcome to be underestimated.

Nested case-control studies

is not a variation on a cohort study, but a variation on a case-control study that is done in conjunction with a cohort study. This variation simply refers to a case-control study that is conducted as part of a cohort study.

Advantages of passive surveillance

much less expensive than active surveillance, and the commitment of resources is lower since the solicitation of information is not active.

Non-respondant bias (cross-sectional studies)

not all people who receive a survey may decide to fill it out and return it. Sometimes the individuals who don't return the survey are different in some way from those who do; perhaps, for example, they are less healthy.; - these can significantly affect cross-sectional surveys. Additionally, a survey may be less likely to reach some participants than others; for example, a survey administered over the internet will only reach those who have internet access.

disadvantages of surveillance

not capable of demonstrating causality, but that is not generally their purpose.

Selection bias (case-control studies)

refers to bias involving the selection of participants to participate in the experiment; in case-control studies this generally involves selecting participants for the study who don't accurately represent the relationship between exposure and disease that would be seen in the general population. one type is called prevalence-incidence bias

A group of researchers are interested in the relationship between binge drinking and metabolic syndrome. They recruit a sample of participants; about half of them binge drink once or twice a week while the other half do not binge drink at all. Then, the researchers follow-up with those participants over a 10 year period to see how many of them develop metabolic syndrome. This is what type of study? Group of answer choices

prospective cohort - at the start of the study, none of the participants have the outcome. thus you can exclude case-control designs. the study is not retrospective; researchers are following up with participants over a 10-year period and waiting to see who develops metabolic syndrome. therefore, it is a prospective study

A research group is interested in studying the relationship between pesticide exposure and Alzheimer's disease. They identify a group of people who have had high levels of pesticide exposure and another who have had no pesticide exposure. They then follow-up with them every year for the next 25 years to see how many in each group develop cancer. This is what type of study?

prospective cohort study

What statistic compares the incidence density rate in the exposed to the incidence density rate in the unexposed?

rate ratio

When conducting a case-control study on the relationship between high carbohydrate levels in the diet and blood pressure, researchers ask participants to respond to a survey about the average daily carbohydrate levels in their diet over the past year. What type of bias is this question likely to be affected by?

recall bias

Case series

researchers will describe a group of patients with similar characteristics. example: - A famous case series was published by the Centers for Disease Control (CDC) in June of 1981. It described five cases of a disease called Pneumocystis carinii pneumonia in middle-aged men in the Los Angeles area. P. carinii pneumonia is a type of lung disease that is usually only seen in individuals with compromised immune systems. None of the cases included in the case series were known to have a compromised immune system; this was the reason the CDC published the case series—it seemed worth bringing to people's attention that these unusual cases of P. carinii pneumonia were occurring in one area of the United States. Immediately after publication, many more reports of similar cases of P. carinii pneumonia emerged. Soon after the publication of this case series, epidemiologists realized these middle-aged men were suffering from a disease that compromised the immune system. This is a famous example of a case series because it brought what would come to be one of the most devastating new diseases of the 20th century to the attention of clinicians: Acquired Immune Deficiency Syndrome (AIDS). - but when it comes to determining causal relationships, one must use a more advanced design

In 2017, a research group is interested in studying the relationship between exposure to a particular chemical and cancer. They identify a group of people who were exposed to the chemical during an industrial accident in 1985 and another group who were never exposed. Then they examine medical records to see if those in the exposed group were more likely to develop cancer over the 20 years following their exposure. This is what type of study?

retrospective cohort study

A group of researchers sets out to investigate the relationship between toxic waste exposure and cancer. In the year 2017, they obtain medical records for individuals who lived near toxic waste spills between the years 1990-1995 as well as medical records for individuals with no toxic waste exposure during that time frame. They examine the medical records to determine how many people in each group developed cancer over a 10-year period from 1995 to 2005. This is what type of study?

retrospective cohort study - this study is retrospective, since it involves looking backwards in time. it is not however a case control study because at the start of the period of time being analyzed no one had the outcome. instead some participants had the exposure and some did not. and medical records were analyzed to see who developed the outcome. thus it is a retrospective cohort

Because a case-control study begins after the outcome has already occurred, we call it a

retrospective study

Descriptive designs

study designs in which the researcher defines a problem and variable of interest but makes no prediction and does not control or manipulate anything

Case study definition (case report)

study of a single case, a single individual that involves a description of the clinical features of that case, including any and all details that might help the reader understand the important details of the case that might help the reader to understand the important details of the case. - often a case study is done when an unusual case is presented to a clinician; An unusual case warrants a case study because a case study may be a way of bringing this case to the attention of other clinicians or researchers so they are aware of the case type or factors that might have been involved in bringing it about.

misclassification and confounds for cohort studies

susceptible to errors in classification of exposure or the outcome and can affect the outcome of the study - researchers may have a little more control over confounds than in a case-control study however the effect of confounds cannot be completely removed from any study regardless of design

Webs of causation

the more it becomes apparent that very few health conditions have just one cause. This makes the determination of causality even more complex, as we must be aware that most diseases have multiple causal factors and many other contributing factors, and which of these are most important depends to a large degree on the individual and factors like his or her genetic makeup, early life environment, experiences, and current environment. - stress--> diet and hormones--> smoking, obestiy, physical ativity, genetic background, insulin resistance--> type 2 diabetes - This type of design, with multiple precipitating conditions and events, is a better representation of causality than a more simplistic indication that shows only one—or even just a few—factors that influence the likelihood of a health condition. It is helpful to keep in mind as you learn about health that the true relationship between a health condition and the factors that lead to it looks more like this than a simple straight line from one causal factor to the disease.

retrospective cohort study

the researchers would examine the medical records on the individuals exposed to the chemstuff spill, looking for cases of cancer that occurred for a period of 20 years (the exact time frame would depend on the study) after the exposure. They would compare the rate of cancer occurrence in this exposed group to another group of factory workers who were not exposed. If their hypothesis was correct, they would see more cases of cancer in the workers exposed to chemstuff than in those who were not. - Thus, this study is set up in exactly the same way as a prospective cohort study (e.g. groups based on exposure, post-exposure period examined to look for occurrence of outcome), but the exposure and the outcome have all occurred in the past. A retrospective cohort study can have more explanatory power than a case-control study, but it is generally considered inferior to a prospective cohort study for several reasons.