Inferential Methods
Which of the following are true of cohort studies? Effect-cause is a major problem for most cohort studies. Effect-cause can never occur in a cohort study. While specification (aka restriction) can minimize the influence of important confounders, it does so at the expense of generalizability. An important weakness common to all cohort studies is that blinding of the outcome assessment is not possible. The likelihood of chance associations can be minimized by maximizing the sample size.
While specification (aka restriction) can minimize the influence of important confounders, it does so at the expense of generalizability. and The likelihood of chance associations can be minimized by maximizing the sample size.
Which of the following are true of decision analyses? While typically designed using data from studies involving many patients, they can be tailored to an individual patient's unique probabilities. They can provide quantitative answers to difficult clinical decisions. A main weakness is that they do not allow incorporation of quality of life measures. They can convert diverse benefits into a single currency (i.e., QALYs) allowing comparison across decisions.
While typically designed using data from studies involving many patients, they can be tailored to an individual patient's unique probabilities. They can provide quantitative answers to difficult clinical decisions. and They can convert diverse benefits into a single currency (i.e., QALYs) allowing comparison across decisions.
Which of the following are true of qualitative research studies? (Choose all statements that are true) "Thick description" can be used to provide the reader with contextual information to understand what it was like for subjects to have been in the study. A common sampling goal is to assure that the proportions of various characteristics (age, gender, race, viewpoints, etc.) in the study sample match their distribution in the target population. "Focus groups" can provide a qualitative assessment of how subjects in the target population are thinking about the topic of interest "Member checking" is a common sampling strategy.
"Thick description" can be used to provide the reader with contextual information to understand what it was like for subjects to have been in the study. and "Focus groups" can provide a qualitative assessment of how subjects in the target population are thinking about the topic of interest
Which of the following are true of the use of blinding (using methods aimed at preventing subjects and/or investigators/study staff from knowing the study arm to which each subject was assigned)? A major purpose is to protect against performance bias A major purpose is to protect against detection/ascertainment bias A major purpose is to assure that the outcome of interest is not present at baseline A major purpose is to minimize attrition bias
A major purpose is to protect against performance bias and A major purpose is to protect against detection/ascertainment bias
All of the following are common examples of "co-interventions" EXCEPT. (Choose ONE) A psychiatrist involved in a RCT studying a new antidepressant for depression learns of the study assignments and provides more therapeutic support for subjects in the treatment arm. A participant in a trial of a weight loss remedy does not know he/she is receiving placebo and adheres to the treatment protocol. A caregiver of a participant in a study to improve quality of life in patients with brain cancer is aware the participant is in the control group and spends more time socializing with the participant. A participant enrolled in a trial evaluating a new sleep medication knows she is receiving placebo and takes melatonin (an over the counter sleep aid).
A participant in a trial of a weight loss remedy does not know he/she is receiving placebo and adheres to the treatment protocol.
Kacie received a grant to conduct a pilot study (a small and inexpensive preliminary study) comparing two types of flea collars among German Shepherds. Ideally, her study design should ________________________. (Choose ALL that are True) include clinically relevant outcomes such as quality of life. evaluate more than one outcome. recruit study participants who are at low risk for developing the outcome to assure adequate sample size. specify a single primary outcome to test the main hypothesis.
include clinically relevant outcomes such as quality of life, evaluate more than one outcome and specify a single primary outcome to test the main hypothesis.
Which of the following are true of a well-designed randomized controlled trial? (Choose ALL that are True) When selecting the intervention, the investigators consider the dosage, duration, and frequency of the intervention which best balances safety and efficacy. When selecting the intervention, complicated interventions are preferable to simple interventions. When setting up the randomization process, rigorous precautions are taken to prevent tampering with the randomization sequence such as use of numbered, sealed, opaque envelopes. If properly carried out, randomized trials provide the most definitive evidence of causality.
All except, "When selecting the intervention, complicated interventions are preferable to simple interventions."
All of the following are steps that an investigator should take when designing a sample EXCEPT? (Choose ONE) Conceptualize the target population using a specific set of inclusion criteria Select an appropriate accessible population Design an approach to sampling the population Design and implement strategies for recruiting a sample of subjects that is small enough to control random sources of error
All of the following describe important steps in designing a study sample except for option #4. As stated on page 30, it is important that the investigator select a sample of subjects that is sufficiently representative of the target population to control systematic sources of error and LARGE enough to control random sources of error.
For each item of the following methods of dealing with confounding, please indicate whether they are used during the design phase or analysis phase. Statistical adjustment is a _______phase strategy for dealing with confounding.
Analysis-Phase
For each item of the following methods of dealing with confounding, please indicate whether they are used during the design phase or analysis phase. Stratification is a _______phase strategy for dealing with confounding.
Analysis-Phase
When reviewing the results of a decision analysis, the reader should consider the following. (Choose ALL that are True) Are each of the components contributing to the model valid? Do the probability estimates generally agree with what is observed in clinical practice? Is the magnitude of expected benefit from the favored clinical strategy clinically important? Are the results of the decision analysis stable?
Are each of the components contributing to the model valid? Do the probability estimates generally agree with what is observed in clinical practice? Is the magnitude of expected benefit from the favored clinical strategy clinically important? Are the results of the decision analysis stable? (All)
Which of the following are true of cross-sectional studies? Effect-cause is unlikely with this design. Chance is not a concern for this design. Attrition bias is not a concern for this design. Detection bias is not a concern for this design. Confounding is unlikely with this design.
Attrition bias is not a concern for this design.
Which of the following are true of before-after studies? (Choose all that are true) Before-after studies can be prospective or retrospective in design. Using different people in the before and after groups is a good way to minimize the problem of confounding. Using the same people for both the before and after groups is a good way to avoid the problem of learning effects. Including an external control group can help to assess whether or not the findings may have been influenced by secular trends. Lengthening the interval between the before and after measurements can minimize the problem of secular trends
Before-after studies can be prospective or retrospective in design. and Including an external control group can help to assess whether or not the findings may have been influenced by secular trends.
Which of the following are true of before-after studies? (Choose all that are true) Before-after studies can be prospective or retrospective in design. Using different people in the before and after groups is a good way to minimize the problem of confounding. Using the same people for both the before and after groups is a good way to avoid the problem of learning effects. Including an external control group can help to assess whether or not the findings may have been influenced by secular trends. Lengthening the interval between the before and after measurements can minimize the problem of secular trends.
Before-after studies can be prospective or retrospective in design. and Including an external control group can help to assess whether or not the findings may have been influenced by secular trends.
Which of the following are true with regard to outcomes in randomized trials? If a study is adequately powered to detect differences in efficacy (it has enough participants to demonstrate a statistically significant difference if one exists) , it is automatically powered to detect differences in important adverse outcomes Blinding of assessors is less important when evaluating hard outcomes. If the participants and providers can not be blinded, there is no point in trying to blind the outcome assessor. Surrogate markers are often used when the clinical outcome of interest is uncommon or takes a long time to develop.
Blinding of assessors is less important when evaluating hard outcomes. and Surrogate markers are often used when the clinical outcome of interest is uncommon or takes a long time to develop.
The following are potential challenges a researcher may face when conducting a clinical trial. (Choose ONE) It is generally expensive to conduct a clinical trial. It can be quite time-consuming to run a clinical trial. Both A and B Neither A nor B
Both A and B
Now let's remind ourselves of a few of the major intentions behind clinical trial registries like ClinicalTrials.gov. Which of the following are intended advantages of requiring investigators to register key trial information in a publicly available registry prior to enrolling the first participant in their study? By pre-specifiying primary and secondary outcomes before enrollment of subjects, there will be a lower likelihood of selective reporting of outcomes. Negative trials will be guaranteed publication in peer reviewed journals The existence of small and/or negative trials will be documented and publicly accessible whether or not they are ever published. By pre-specifying inclusion and exclusion criteria, there will be a lower likelihood of detection bias
By pre-specifiying primary and secondary outcomes before enrollment of subjects, there will be a lower likelihood of selective reporting of outcomes. and The existence of small and/or negative trials will be documented and publicly accessible whether or not they are ever published.
Which of the following could enhance your ability to draw causal inferences from the findings of a before-after study? (Choose all that are true) When the before and after subjects are different people, design the study so that all eligible subjects who are available during their respective time periods are enrolled. Carefully collect data on potential confounders from each group so that the analysis can be appropriately adjusted. Collect the same information, during the same before and after periods, from a population that was not exposed to the intervention or exposure of interest. Use objective measures for key outcomes of interest. Be sure that the time periods used for the before and after periods are comparable in terms of possible seasonal variations.
Carefully collect data on potential confounders from each group so that the analysis can be appropriately adjusted. Collect the same information, during the same before and after periods, from a population that was not exposed to the intervention or exposure of interest. Use objective measures for key outcomes of interest. Be sure that the time periods used for the before and after periods are comparable in terms of possible seasonal variations. (All)
Which of the following are true of case-control studies? Case-control studies are similar to multiple cohorts in that the comparison groups are sampled separately in both designs. Case-controls studies are similar to retrospective cohorts in that they are both highly susceptible to differential recall bias. Case control studies are similar to randomized trials in that they both evaluate several outcomes of interest.
Case-control studies are similar to multiple cohorts in that the comparison groups are sampled separately in both designs. While "Number 1" is a true statement, "Number 2" is false because retrospective cohorts are not susceptible to differential recall bias. This is because the cohort is defined at a time 0 when the outcome is not yet present so assessments of the predictor variable could not be influenced by knowledge of the outcome. "Number 3" is false because while RCTs can look at many outcomes, case-control studies are structured to evaluate potential predictors of a specific outcome (the one that defines the cases).
Kacie and her boyfriend Otto are conducting a systematic review of optimal treatments for canine skin allergies. As part of their review process, they should make sure to do all of the following EXCEPT (Choose One). Provide clear definitions of inclusion and exclusion criteria. Abstract data from each study in a uniform and unbiased fashion. Create data abstraction forms, which include any data that will subsequently appear in the text, tables, or figures. Clearly allocate the responsibility of data abstraction to one reviewer.
Clearly allocate the responsibility of data abstraction to one reviewer.
Which of the following are true of before-after studies? (Choose all that are true) Because the comparison groups are separated by time, differential performance bias is not possible. Because the comparison groups are separated by time, differential detection bias is not possible. The phenomenon of "regression to the mean" is most important to consider when the before and after groups are different people. Compared to using different subjects for the before and after groups, using the same subjects can reduce the problem of differential selection bias. Compared to using different subjects for the before and after groups, using the same subjects can increase the problem of differential attrition bias.
Compared to using different subjects for the before and after groups, using the same subjects can reduce the problem of differential selection bias. and Compared to using different subjects for the before and after groups, using the same subjects can increase the problem of differential attrition bias.
All of the following are ethical issues that must be considered when designing a randomized clinical trial EXCEPT. (Choose ONE) The intervention and control arms must be in equipoise (there must be uncertainty or controversy over which is superior). Control groups must receive appropriate standard of care (eg. routine interventions that are already known to be effective). Confidentiality is only relevant if the investigator personally knows the subjects. Trial must be discontinued once it has been shown that one arm is more effective or harmful.
Confidentiality is only relevant if the investigator personally knows the subjects. (Not true!)
Which of the following are ways in which validity can be assessed? (Select all that are true) Content Validity Face Validity Construct Validity Systematic Validity Criterion-related Validity
Content Validity, Face Validity, Construct Validity, and Criterion-related Validity (systematic validity does not exist)
This design combines within and between group designs to enhance control over confounding. Non-inferiority Crossover Time-series Factorial
Crossover
For each item of the following methods of dealing with confounding, please indicate whether they are used during the design phase or analysis phase. Matching is a _______phase strategy for dealing with confounding.
Design-Phase
For each item of the following methods of dealing with confounding, please indicate whether they are used during the design phase or analysis phase. Specification is a _______phase strategy for dealing with confounding.
Design-Phase
An investigator is exploring the association between alcohol consumption and pancreatitis. In order to help minimize the risk of bias, she should _____. (Choose all that are TRUE) Enroll a small sample of subjects. Ensure that the sample of study subjects represents the population of interest. Make sure to inform the outcome assessors of the predictor variables. Ensure that the measurements of the outcome variables represent the outcomes of interest.
Ensure that the sample of study subjects represents the population of interest. and Ensure that the measurements of the outcome variables represent the outcomes of interest.
Which of the following are considered characteristics of a good research question? (Select all that are True) Feasible Interesting Necessary Ethical Repeatable
FINER! Feasible Interesting, and Ethical (N= novel and R= relevant)
This design allows two or more independent trials to be carried out for the price of one. Non-inferiority Crossover Time-series Factorial
Factorial
(True/false) Marginal (or incremental) cost effectiveness is calculated by comparing the new strategy being considered to doing nothing, as opposed to comparing it to the current practice.
False
(True/false) When a new strategy being considered costs less and is more effective than the current strategy, the new strategy is "dominated".
False
The Jadad Score includes assessment of whether the allocation to study arm was concealed. (True/False)
False
True or False: "Regression to the mean" refers to the concept that people with extreme values are likely to have more extreme values on repeated measures.
False
True or False: A cost-effectiveness analysis would be relevant if the new strategy being considered was both more costly and less effective than the existing strategy.
False
True or False: Effect-cause is a common problem in cohort studies of disease causation because risk factor measurements are made among subjects who are known to have the disease.
False
True or False: Random error is a wrong result due to bias.
False
Which of the following are true of cohort studies? For some research questions, they are the highest quality study design that can be used. They are good for evaluating rare exposures. A weakness is that they are limited to a single outcome of interest. A weakness is their inherent susceptibility to differential recall-bias.
For some research questions, they are the highest quality study design that can be used. and They are good for evaluating rare exposures.
Which of the following are true of "surrogate markers"? Ideally, they are intermediate in the main pathway to the clinical outcome of interest. They often require a longer duration of follow-up than clinical outcomes They are always subjective. An example is HIV viral load.
Ideally, they are intermediate in the main pathway to the clinical outcome of interest. and An example is HIV viral load.
Which of the following are true of Funnel Plots? They provide a visual display of the effect size and confidence interval of each study as well as a summary estimate showing the pooled effect size and variance. They are used for visually assessing for the presence of attrition bias. If no bias is present, they should look like an inverted funnel with the apex near the summary effect estimate. A lack of correlation between study outcomes (effect size) and sample size suggests the presence of publication bias.
If no bias is present, they should look like an inverted funnel with the apex near the summary effect estimate.
Which of the following are true of the prospective cohort design. (Choose all that are TRUE) The investigator only measures the level of the predictors after the outcome has occurred. It is cheap and efficient for studying rare outcomes. It allows the investigator to measure variables more completely and accurately than is possible retrospectively. It enables the investigator to calculate incidence, the number of new cases of a condition occurring over time.
It allows the investigator to measure variables more completely and accurately than is possible retrospectively. and It enables the investigator to calculate incidence, the number of new cases of a condition occurring over time.
This design compares a new intervention to existing standards of care and often requires a large sample size. Non-inferiority Crossover Time-series Factorial
non-inferiority
Which of the following are true of the Jadad Score? It incorporates seven items related to randomization, blinding and withdrawals. Possible scores range from 0 (weakest methodological quality) to 7 (strongest methodological quality). The Jadad Score works especially well because if an element is not reported in an article, you can be confident it was not done. Possible scores range from 0 (weakest methodological quality) to 5 (strongest methodological quality).
It incorporates seven items related to randomization, blinding and withdrawals. and Possible scores range from 0 (weakest methodological quality) to 5 (strongest methodological quality).
Which of the following are true of a decision analysis? (Choose ALL that are True) It is a systematic approach to structuring a decision. It includes a decision model that outlines the strategies being compared and important potential outcomes of each strategy. The investigator collects relevant information about the probability of each outcome being considered. By definition, a decision analysis should only include two clinical outcomes.
It is a systematic approach to structuring a decision. It includes a decision model that outlines the strategies being compared and important potential outcomes of each strategy. and The investigator collects relevant information about the probability of each outcome being considered.
According to the primer, which of the following are true of cost-effectiveness analysis. (Choose ALL that are True) It is a technique for comparing the relative value of various clinical strategies. The critical reader should examine the information used for effectiveness. The funding source typically has no bearing on a CEA analysis. The critical reader should look at the utilization counts and have some confidence about the face validity of the dollars attached to them.
It is a technique for comparing the relative value of various clinical strategies. The critical reader should examine the information used for effectiveness. and The critical reader should look at the utilization counts and have some confidence about the face validity of the dollars attached to them.
Which of the following are true of a "threshold value" in decision analysis? It is the value at which the strategies being compared have an equal expected value. If it occurs in the plausible range for the variable being varied, it means the model is stable with regard to that variable. It is the point at which the "answer" to the model shifts from one strategy to the other It is the value beyond which neither strategy is favored.
It is the value at which the strategies being compared have an equal expected value. and It is the point at which the "answer" to the model shifts from one strategy to the other
All of the following are features of case-crossover studies EXCEPT (Choose ONE) It is a variant of the case-control design. Exposures of the cases at the time (or right before) the outcome occurred are compared with exposures of those same cases at one or more other points in time. It is useful for studying the long-term effects of intermittent exposures. As with ordinary care-control studies, case-crossover studies begin with a group of cases (people who have had the outcome of interest).
It is useful for studying the long-term effects of intermittent exposures.
Which of the following are true of randomization? Its purpose is to balance potentially confounding variables present at baseline between the study groups It increases the likelihood that a positive finding is due to chance If done in a tamperproof, concealed way, it can protect against performance bias. If done in a tamperproof, concealed way, it can protect against selection bias.
Its purpose is to balance potentially confounding variables present at baseline between the study groups And If done in a tamperproof, concealed way, it can protect against selection bias.
True or False: Inferences about cause and effect are strengthened by measuring and adjusting for potential confounding variables.
true
True or False: The biggest drawback to a systematic review is that it can produce a reliable appearing summary estimate based on the results of individual studies that are of poor quality.
true
Which of the following are true of matching in case-control studies? Matching is a good method for ensuring that cases and controls are comparable with respect to important factors that are related to the disease. Matching eliminates the option of studying the matched variables as predictors. An ideal variable for matching is one that is intermediate in the causal pathway to the outcome of interest. Overmatching is a method for increasing the power in case-control studies.
Matching is a good method for ensuring that cases and controls are comparable with respect to important factors that are related to the disease. and Matching eliminates the option of studying the matched variables as predictors.
A Medline search is a key component of any systematic review search strategy. Which of the following statements are true with regard to Medline? Medline is a database that can be searched using a variety of interfaces including PubMed and OVID. When searching Medline, MeSH terms and keywords are the same thing. The controlled vocabulary for searching Medline is conveniently identical to that used for searching other similar electronic search databases including EMBASE. The "NOT" operator is the preferred method for limiting sets when designing a search strategy for a systematic review.
Medline is a database that can be searched using a variety of interfaces including PubMed and OVID.
All of the following are true of the multiple cohort design EXCEPT. (Choose ONE). Begins with two or more separate samples of subjects; one group with exposure to a potential risk factor and one or more groups with no exposure or a lower level of exposure. May be the only feasible approach for studying rare environmental or occupational exposures. Minimizes the problem of confounding. May use a population based registry data as the external control group.
Minimizes the problem of confounding.
Which of the following are true of measuring and analyzing outcomes? Most rare outcomes are not seen until an intervention is in widespread use. Blinded outcome assessment is important for reducing the role of chance. While there are a variety of methods for asking subjects about adverse outcomes, they all tend to give the same result. Many studies are underpowered for evaluating important adverse outcomes.
Most rare outcomes are not seen until an intervention is in widespread use. and Many studies are underpowered for evaluating important adverse outcomes.
Which of the following are true of variations on the classic case-control design? Case-crossover studies sample cases from one population and controls from another. Nested case-control studies can eliminate the problem of differential recall-bias. Nested case-controls and nested case-cohorts can be especially useful for research questions that require expensive testing to determine predictor variables. In a nested case-cohort study, the control group is a random sample of the entire cohort in which the study is nested.
Nested case-control studies can eliminate the problem of differential recall-bias. Nested case-controls and nested case-cohorts can be especially useful for research questions that require expensive testing to determine predictor variables. and In a nested case-cohort study, the control group is a random sample of the entire cohort in which the study is nested.
Which of the following are true of case-control studies? Case-control studies are similar to cohort studies in that the comparison groups are defined by the presence or absence of the outcome of interest. Case-controls studies and cohort studies share the weakness that effect-cause is a major concern. Nested-case control studies and cohort studies share the advantage that differential recall bias is unlikely.
Nested-case control studies and cohort studies share the advantage that differential recall bias is unlikely. "Number 1" is false because while in a case-control study the comparison groups are defined by the presence or absence of the outcome of interest, in a cohort study the comparison groups are defined by the presence or absence of the exposure/predictorof interest. "Number 2" is false because effect-cause is not a major concern in cohort studies. Because cohorts start at a time 0 when the subjects do not have the outcome, and follow them forward in time to assess for the development of the outcome, it is a good design for establishing the temporal relationship between the exposure/predictor and the outcome. While effect-cause is possible in a cohort (if the outcome has a long pre-clinical phase in which there are no symptoms or obvious findings making it difficult to exclude at baseline), it is not a common problem.
In Chapter 8 of the Cochrane Handbook the authors recommend against the use of the term "methodological quality" to describe the risk for bias in a study. They point out that the results of a study can be unbiased despite a methodological flaw, and conversely a study can be conducted at the highest possible standards, yet be at risk for bias. The Cochrane authors instead recommend the term "Risk of Bias" and have created a tool for assessing each included trial. Please assess which of the following statements related to Cochrane's Risk of Bias Tool are True. The Risk of Bias tool is designed to provide an overall numeric score that can be used in classifying each study as having high, moderate or low risk for bias. The Risk of Bias tool has seven domains, each of which is assessed as Yes or No. Rather than presenting an overall "Risk of Bias" score, the findings of each domain should be presented for each study. One way to assess the "Reporting Bias" domain would be to compare the outcomes reported in the published study to those listed in the protocol on ClinicalTrials.gov.
Rather than presenting an overall "Risk of Bias" score, the findings of each domain should be presented for each study. and One way to assess the "Reporting Bias" domain would be to compare the outcomes reported in the published study to those listed in the protocol on ClinicalTrials.gov.
Which of the following are true of case-series? The goal of most case-series is broad generalizability. Referral bias occurs when the most difficult or unusual cases are referred to specialty centers so the clusters (or "series") of unusual cases seen there are often the most extreme examples. Confounding is typically not an issue when using a case-series to evaluate the effect of a novel intervention. Some case-series compare their findings to "historical controls". Case-series are always observational.
Referral bias occurs when the most difficult or unusual cases are referred to specialty centers so the clusters (or "series") of unusual cases seen there are often the most extreme examples. and Case-series are always observational.
Which of the following are accurate descriptions of publication bias or ways of dealing with it? (Choose ALL that are True) Reviewers can attempt to identify unpublished results by querying investigators and reviewing abstracts, meeting presentations, and doctoral theses. The presence of publication bias can be assessed by creating a funnel plot of included studies. Even if there is adequate information to determine study eligibility, reviewers should never include the results of unpublished studies in the overall summary estimate. In the presence of publication bias, a plot of study sample size versus outcome should have a bell or funnel shape.
Reviewers can attempt to identify unpublished results by querying investigators and reviewing abstracts, meeting presentations, and doctoral theses. and The presence of publication bias can be assessed by creating a funnel plot of included studies.
Strategies for minimizing the risk of sampling bias in a case-control study include________. (Choose all that are TRUE) Sampling controls and cases in the same way. Performing a population-based study. Using only one control group. Matching cases with controls.
Sampling controls and cases in the same way. Performing a population-based study. and Matching cases with controls.
According to the Hulley text, which of the following are potential approaches for dealing with or assessing the problem of sampling bias in case-control studies? Sampling the cases and controls in the same admittedly unrepresentative way. Blinding the observers and subjects to the case/control status. Using several control groups sampled in different ways. Matching the cases and controls on certain predictor variables.
Sampling the cases and controls in the same admittedly unrepresentative way. Using several control groups sampled in different ways. and Matching the cases and controls on certain predictor variables.
Which of the following search methods are primarily aimed at reducing the chance of missing unpublished studies? Searching both OVID and PubMed. Searching the proceedings of scientific conferences for abstracts presented at their annual meetings. Searching the reference lists of relevant studies for citations of published articles your electronic search may have missed. Searching trial registries like ClinicalTrials.gov.
Searching the proceedings of scientific conferences for abstracts presented at their annual meetings. and Searching trial registries like ClinicalTrials.gov.
Which of the following are true of systematic review methodology? Inclusion criteria should be flexible enough to accommodate changes once the findings of the review are known. Efforts to contact authors for missing information are usually not successful and thus are not considered to be worth the effort. Sensitivity analyses can be used to evaluate how various decisions in the review process (i.e., to allow unblinded studies) might have influenced the findings of the review. Data abstraction should be performed by at least two independent reviewers so that they can compare their findings.
Sensitivity analyses can be used to evaluate how various decisions in the review process (i.e., to allow unblinded studies) might have influenced the findings of the review. and Data abstraction should be performed by at least two independent reviewers so that they can compare their findings.
Which of the following are common weaknesses of case-control studies? The need for a much larger number of subjects than cohorts and RCTs. Separate sampling of the cases and controls. Lack of a control group. Recall bias related to retrospective measurement of predictor variables.
Separate sampling of the cases and controls. and Recall bias related to retrospective measurement of predictor variables.
Which of the following are true? Serial surveys are often used to investigate patterns over time. The term "serial survey" is synonymous with "cohort study". Serial surveys follow a single group of subjects over time.
Serial surveys are often used to investigate patterns over time.
Strategies for minimizing the risk of differential measurement bias in a case-control study include__________. (Choose ONE) Asking subjects recently diagnosed with a serious disease (e.g. cancer) to recall past exposures. Standardizing the operational definitions of variables. Both A and B Neither A nor B
Standardizing the operational definitions of variables.
Which of the following are true regarding cohort studies? Prospective cohorts are ideal for evaluating the effect of an intervention because the investigator can determine who gets the intervention and who gets the control. Retrospective cohorts are preferred over prospective cohorts because the investigator can more easily control how the variables were measured. Specification (aka restriction) is a reversible method for controlling for potential confounders during the analysis phase of a study. Stratification involves analyzing the results of a study separately for different groups (or strata), for example subjects with or without diabetes. In cohort studies, sources of confounding and bias always make an observed association appear stronger than it is.
Stratification involves analyzing the results of a study separately for different groups (or strata), for example subjects with or without diabetes.
Which of the following are true of randomization techniques? Strong randomization techniques are critical to assuring that confounding variables remain balanced throughout the course of the trial Stratified blocked randomization can be used to ensure that an important potential confounder is evenly distributed between study groups. In a study with two arms, fixed block randomization using blocks of 6 would assure that the study arms differ in size by no more than 6. In a non-blinded study, fixed block randomization can expose a study to selection bias
Stratified blocked randomization can be used to ensure that an important potential confounder is evenly distributed between study groups. and In a non-blinded study, fixed block randomization can expose a study to selection bias
Which of the following are true with regard to case-control studies? Effect-cause is not possible with this design. Including "dummy variables" is a good method for reducing the likelihood of differential measurement bias. While matching helps balance confounders, it greatly limits generalizability. Survivor bias should be considered when subjects need to be alive to be included in the study.
Survivor bias should be considered when subjects need to be alive to be included in the study.
Which of the following are true of heterogeneity? The "test for heterogeneity" evaluates whether the results of the individual studies contributing to your summary estimate vary beyond what would be expected due to chance. In order to determine that the heterogeneity of the summary estimate is acceptable (not significant enough to be concerned), the p-value should be <0.10 A large degree of heterogeneity is highly desirable as it assures the findings are more broadly generalizable. If a summary estimate demonstrates significant heterogeneity, one option is to not combine the studies.
The "test for heterogeneity" evaluates whether the results of the individual studies contributing to your summary estimate vary beyond what would be expected due to chance. and If a summary estimate demonstrates significant heterogeneity, one option is to not combine the studies.
The Cochrane Library is another important database that should be included in the search strategy of a high quality systematic review. This question is aimed at helping you get acquainted with what is there and how to do a basic search. We will be doing more formal "Advanced" searches next term but for this introductory question, we will stick with a basic search. . Go to The Cochrane Library website and use the search box in the upper left corner to see how many and what types of articles the library contains related to public health. (Type "public health" without the quotes in the box and click on "GO"). The Cochrane Library contains over 9000 reviews indexed as being related to public health. The Cochrane Library contains more than 200 Cochrane Reviews and more than 100 "Other" Reviews related to public health There are at least three Review Protocols posted in the Cochrane Library as being related to public health. There is not yet a Cochrane Group for public health topics.
The Cochrane Library contains more than 200 Cochrane Reviews and more than 100 "Other" Reviews related to public health and There are at least three Review Protocols posted in the Cochrane Library as being related to public health.
Which of the following are true of case-control studies? The comparison groups are determined by the exposure of interest. The comparison groups are sampled from separate populations. Case-control studies are ideal for calculating the incidence of disease. Cases always have a disease.
The comparison groups are sampled from separate populations.
Which of the following are true of before-after studies? (Choose all that are true) The groups being compared can consist of different people. The groups being compared can consist of the same people. The groups being compared are determined by the outcome of interest. Time separates the assessment of outcomes for the groups being compared. Before-after studies often provide insights using a real world setting.
The groups being compared can consist of different people. The groups being compared can consist of the same people. Time separates the assessment of outcomes for the groups being compared. Before-after studies often provide insights using a real world setting.
Now please refer to the PRISMA Statement Checklist. You could either use the published article that is in your reading or go to prisma statement website and look at the linke for the "PRISMA Checklist". The names of all databases searched should be provided, including dates covered and last date searched. Contact with authors should be described. Sources that prove unfruitful (you don't find any articles) do not need to be reported. A full electronic search strategy for at least one database should be included.
The names of all databases searched should be provided, including dates covered and last date searched. Contact with authors should be described. and A full electronic search strategy for at least one database should be included.
With regard to analyzing the results of a randomized controlled trial, which of the following are true? The preferred analysis method for efficacy outcomes is usually by "intention to treat". The preferred analysis method for adverse outcomes is usually by "intention to treat". "Per protocol" analyses are ideal because they preserve the benefits of randomization. As treated" analyses are ideal because they preserve the benefits of randomization.
The preferred analysis method for efficacy outcomes is usually by "intention to treat".
Which of the following are considered tangible elements that make up the study plan? (Select all that are True) The research question and its significance The study plan Measurement approaches Study subjects
The research question and its significance, The study plan, Measurement approaches, and Study subjects (all true!)
For each of the four main differential biases (selection, performance, attrition and detection), we considered four discrete questions--what it is, when it can occur, how to minimize it, and how to assess the likelihood that it occurred (how effective were the methods used to minimize it). Which of the following would allow you to assess the likelihood that differential attrition bias compromised the results of a completed randomized trial? The study's overall loss to follow-up was 20, split evenly among the two arms The study provided a CONSORT Flow Diagram including the numbers and reasons for attrition by study arm The study had nurses call patients weekly, sent holiday and birthday cards, and paid participants based on the number of study visits they attended. The study was double-blinded and blinding was found to be very effective.
The study's overall loss to follow-up was 20, split evenly among the two arms and The study provided a CONSORT Flow Diagram including the numbers and reasons for attrition by study arm
Which of the following are true of case-series? The subjects often have a shared unique, novel, or unusual presentation. They often contain detailed descriptions of the individual subjects. They are good for rare outcomes. They can sometimes be converted to a case-control study by adding one or more appropriate control groups.
The subjects often have a shared unique, novel, or unusual presentation. They often contain detailed descriptions of the individual subjects. They are good for rare outcomes. and They can sometimes be converted to a case-control study by adding one or more appropriate control groups.
According to the before-after primer, the validity of the inference is threatened if which of the following is true? (Choose ONE) All participants in the before group are eligible for the after group. There is evidence of a prevailing temporal trend. The observed change after the intervention is implemented is substantial The observed change occurs quickly after the intervention.
There is evidence of a prevailing temporal trend.
Which of the following are true of cross-sectional studies? There is no follow-up period. They are ideal for measuring incidence. "Relative prevalence" could be considered the cross-sectional equivalent of "relative risk". They can be used to characterize the participants in a cohort study or randomized trial at a specific point in time.
There is no follow-up period. "Relative prevalence" could be considered the cross-sectional equivalent of "relative risk". and They can be used to characterize the participants in a cohort study or randomized trial at a specific point in time.
Which of the following are true of cost-effectiveness ratios? They are calculated by dividing the net cost of the intervention by its net benefit. Interventions with high C/E ratios are the most cost-effective (high priority). Interventions with low C/E ratios should be abandoned. The Net Cost used for calculating a C/E ratio is determined by how much it costs to implement the intervention of interest minus how much income it generates.
They are calculated by dividing the net cost of the intervention by its net benefit.
Which of the following are true of QALYs? They are calculated by multiplying the life-expectancy in years by a quality adjustment factor such as a utility. QALY stands for quality-adjusted life year. They are determined by dividing the net cost of the intervention by its net benefit. If QOL after stroke is valued at 0.5 utilities, then living for 5 years in this outcome state would be worth 5.5 QALYs.
They are calculated by multiplying the life-expectancy in years by a quality adjustment factor such as a utility. QALY stands for quality-adjusted life year.
Which of the following are strengths of case-control studies? They are efficient for studying rare outcomes. Sampling from two populations reduces the likelihood of bias. The design allows you to look at many outcomes at a time. They are good for hypothesis generating.
They are efficient for studying rare outcomes. and They are good for hypothesis generating.
Which of the following are true of cohort studies? They are ideal for measuring the incidence of a condition. A major weakness is their susceptibility to confounding. Only prospective cohorts can establish a time sequence between exposure and outcome variables; retrospecitve cohorts are unable to do so. They are ideal for evaluating possible causes of rare outcomes.
They are ideal for measuring the incidence of a condition. and A major weakness is their susceptibility to confounding.
Which of the following are true of utilities? They are numerical expressions of patient preferences for a particular state of health. They are typically assessed on a scale from -1 (death or the worst health imaginable) to 1 (perfect health). They can be used to calculate QALYs. One method to determine them is the "standard gamble."
They are numerical expressions of patient preferences for a particular state of health. They can be used to calculate QALYs. and One method to determine them is the "standard gamble."
Which of the following are true of "clinical" outcomes? They are outcomes that can be directly experienced by patients. They are typically considered the least meaningful outcome type. An example of a "clinical" outcome is change in LDL cholesterol level. They are always objective.
They are outcomes that can be directly experienced by patients.
Which of the following are true of before-after studies. (Choose ALL that are True) They are sometimes referred to as pre-post studies. They always include a concurrent control group. They can be used to evaluate the effects of a clinical management or quality improvement intervention. They have minimal risks for confounding and bias.
They are sometimes referred to as pre-post studies. and They can be used to evaluate the effects of a clinical management or quality improvement intervention.
Which of the following are strengths of cross-sectional studies? They are typically faster and less expensive than prospective studies. They are ideal for studying the natural history of a disease. They are ideal for studying rare diseases in samples of the general population. They can provide the prevalence of both the predictors and outcomes. They can be converted to a cohort by adding a time component.
They are typically faster and less expensive than prospective studies. They can provide the prevalence of both the predictors and outcomes. and They can be converted to a cohort by adding a time component.
Which of the following are true of cost-effectiveness analyses? They can convert diverse benefits into a single currency (i.e., $/QALY). They can provide a quantitative method for selecting among competing wants where resources are limited. Low cost-effectiveness ratios help identify the lowest cost choice. They are used in the approval process for new drugs or technologies in some countries.
They can convert diverse benefits into a single currency (i.e., $/QALY). They can provide a quantitative method for selecting among competing wants where resources are limited. and They are used in the approval process for new drugs or technologies in some countries.
Which of the following are true of randomized controlled trials? They can establish a temporal relationship between the intervention and outcome. They are ideal for rare outcomes. Neither the intervention nor the outcome is present at baseline. They do not suffer from attrition bias because the investigator controls the intervention.
They can establish a temporal relationship between the intervention and outcome. and Neither the intervention nor the outcome is present at baseline.
Which of the following are true of qualitative research studies? (Choose each statement that is true) A major goal is to determine whether the way one group thinks about a topic is statistically significantly different from another group. They don't involve much time investment. They can explore questions that would be difficult or impossible to explore any other way. Sampling is often considered complete when the researcher feels he/she has reached "saturation" with regard to his/her understanding of a phenomenon of interest.
They can explore questions that would be difficult or impossible to explore any other way. and Sampling is often considered complete when the researcher feels he/she has reached "saturation" with regard to his/her understanding of a phenomenon of interest.
Which of the following are unique advantages of including a meta-analysis as part of a systematic review? (Choose One) They can increase the power to answer a research question. They always include individual-level data that can be used to adjust for potential confounding. Both A and B Neither A nor B
They can increase the power to answer a research question.
Which of the following are true of meta-analyses? They enhance the validity of the included studies by combining them together. They are often limited by the reduced power that results from combining multiple smaller studies. They can result in a statistically significant summary estimate even when none of the studies contributing to the summary estimate were statistically significant. Sub-group analyses can reveal populations that are more likely to benefit from the intervention being evaluated.
They can result in a statistically significant summary estimate even when none of the studies contributing to the summary estimate were statistically significant. and Sub-group analyses can reveal populations that are more likely to benefit from the intervention being evaluated.
Systematic reviews generally include which of the following types of information. (Choose ALL that are True) They display the findings of the individual studies. They describe important characteristics of each study such as the characteristics of the study population. In the absence of significant heterogeneity, they include sensitivity and subgroup analyses. In the absence of significant heterogeneity, they present summary estimates and confidence intervals based on the findings of all the included studies.
They display the findings of the individual studies. They describe important characteristics of each study such as the characteristics of the study population. In the absence of significant heterogeneity, they include sensitivity and subgroup analyses. In the absence of significant heterogeneity, they present summary estimates and confidence intervals based on the findings of all the included studies. (ALL)
Which of the following are true of qualitative research studies? (Choose each statement that is true) A major strength is that the findings are considered "exhaustive" so when a theme does not emerge you can be sure it does not exist. They generally strive for a large sample sizes in order to maximize generalizability. They generally do not start with a priori theories to be tested. They often use a single method of data collection to try to answer 3 questions at the same time (also known as "triangulation"). They often alter their ongoing data collection goals and methods based on looking at interim results.
They generally do not start with a priori theories to be tested. and They often alter their ongoing data collection goals and methods based on looking at interim results.
Features of secondary data analysis include ________. (Choose ALL that are True) They greatly reduce the amount of time necessary to complete a research study. They acquire data from sources such as existing research projects, electronic medical records and administrative datasets. They enable the investigator to have greater control over the selection of the population to study compared to primary studies. They are a less expensive way to conduct research.
They greatly reduce the amount of time necessary to complete a research study. They enable the investigator to have greater control over the selection of the population to study compared to primary studies. and They are a less expensive way to conduct research.
Advantages of the case-control design include________. (Choose all that are TRUE) They have a relatively short duration. They require relatively few subjects. They enable researchers to study multiple outcomes. They yield information about the incidence of disease.
They have a relatively short duration. and They require relatively few subjects.
Which of the following are true of multiple-cohorts? A population is sampled and divided into multiple comparison groups based on the level of the predictor variable. They may be the only feasible approach for studying rare exposures and exposures to potential occupational and environmental hazards. By sampling exposed and unexposed subjects from separate populations, confounding is minimized. Subjects are selected from different populations for the purpose of enhancing generalizability.
They may be the only feasible approach for studying rare exposures and exposures to potential occupational and environmental hazards.
Which of the following are true of systematic reviews? They use a well-defined and systematic approach to identifying all completed studies that address a specific study question. They present the characteristics and results of each included study. They can reveal peculiarities of individual studies that might not have been apparent when viewing each study individually. To be considered for publication, many journals require that they be accompanied by a PRISMA Checklist and Flow Diagram.
They use a well-defined and systematic approach to identifying all completed studies that address a specific study question. They present the characteristics and results of each included study. They can reveal peculiarities of individual studies that might not have been apparent when viewing each study individually. To be considered for publication, many journals require that they be accompanied by a PRISMA Checklist and Flow Diagram. (all)
In decision analysis, which of the following are true with regard to "repeating the analysis while allowing one variable in the model to vary over a range of values. This is a method for determining how "stable" the model is. This is called a sensitivity analysis. The range should cover all possible values for the variable. The range should cover all plausible values for the variable.
This is a method for determining how "stable" the model is. This is called a sensitivity analysis. and The range should cover all plausible values for the variable.
This design has a single group with outcomes compared within each subject during periods on and off the intervention. Non-inferiority Crossover Time-series Factorial
Time-series
Which of the following are benefits of pre-specifying primary and secondary outcomes? To clarify your hypothesis. To discourage data dredging To prevent attrition bias. To prevent reporting bias.
To clarify your hypothesis. To discourage data dredging and To prevent reporting bias.
(True/False) In general, CEAs give a single answer aimed at maximizing the benefits for a population and thus may not be relevant to individual patients.
True
True or False: A before-after study design is often used in situations where an RCT or cohort study would not be feasible due to the discrete onset and universal coverage of the intervention or exposure (e.g., a nationwide law).
True
True or False: Blinding subjects and investigators to the study arm assignment minimizes the likelihood of unbalanced co-interventions and the risk for biased outcome ascertainment and adjudication.
True
True or False: Cost-effectiveness analysis (CEA) is a technique for selecting among competing wants wherever resources are limited.
True
True or False: In a decision analysis, a threshold value is the value at which two strategies have equal expected value.
True
True or False: Precision is reduced by random error (chance) from three sources of variability: the observer, the subject, and the instrument
True
True or False: The nested case control is a study in which cases and controls are drawn from a larger prospective study (i.e. a cohort or randomized trial).
True
True or False: The retrospective cohort study design is only possible if adequate data about the predictors are available on a cohort of subjects that has been assembled for other purposes.
True
True or False: The variability in the findings of individual studies included in a systematic review is called heterogeneity.
True
True/false: A very desirable case-control design (also known as a "population-based case-control study") is to sample cases from a disease registry and controls from the population served by the registry.
True
Which of the following are strategies for minimizing or assessing differential measurement bias in case-control studies? Using data collected before the outcome occurred. Blinding the observers and subjects to the risk factor being studied Nesting the case-control study in a prospective study. Blinding the observers and subjects to the case-control status of each subject.
Using data collected before the outcome occurred. Blinding the observers and subjects to the risk factor being studied Nesting the case-control study in a prospective study. Blinding the observers and subjects to the case-control status of each subject. (All)
Approaches to minimizing or assessing for bias during the assessment of predictor variables in case-control studies include________(Choose all that are TRUE) Using data recorded before the outcome occurred. Including "dummy" questions about plausible risk factors not associated with the disease. Blinding the observer to the case-control status of the study subjects. Blinding subjects and the interviewer to the specific risk factors being studied.
Using data recorded before the outcome occurred. Including "dummy" questions about plausible risk factors not associated with the disease. Blinding the observer to the case-control status of the study subjects. Blinding subjects and the interviewer to the specific risk factors being studied.
According to the primer, all of the following are true of utilities EXCEPT (Choose ONE) They are numerical expressions of patient preferences for a particular state of health. Utilities are typically assessed on a scale from 1 (death or worst health imaginable) to 100 (best health). A familiar application of utilities is the quality-adjusted life year (QALY). One simple approach to measuring utilities is using a visual analogue scale
Utilities are typically assessed on a scale from 1 (death or worst health imaginable) to 100 (best health).
According to the Kuper et al paper, which of the following are key questions that the reader should ask themselves when reading qualitative research studies? (Choose ALL that are True) Was the sample used in the study appropriate to its research question? Does the study adequately address potential ethical issues? Were the data collected appropriately? Is what the researchers did clear?
Was the sample used in the study appropriate to its research question? Does the study adequately address potential ethical issues? Were the data collected appropriately? Is what the researchers did clear? (All)
Which of the following are elements of a good systematic review? (Choose ALL that are True) Well-formulated, clear research question. Only display the findings of studies with the most robust outcomes. Comprehensive and unbiased identification of completed studies. Clear presentation of inclusion and exclusion criteria.
Well-formulated, clear research question. and Comprehensive and unbiased identification of completed studies. Clear presentation of inclusion and exclusion criteria.
Which of the following describe main weaknesses of the before-after study design? (Choose ALL that are True (i.e. main weakness)) The before-after design provides NO insight about the use of a widespread intervention in a real world setting. When the before and after groups are made up of the SAME people, there can be regression to the mean. Since the before and after measurements are made at different time points, differences could be due to secular trends. When the before and after groups are made up of the SAME people, there can be learning effects.
When the before and after groups are made up of the SAME people, there can be regression to the mean. Since the before and after measurements are made at different time points, differences could be due to secular trends. and When the before and after groups are made up of the SAME people, there can be learning effects.
Which of the following are important factors to consider when critically evaluating the findings of a before-after study? (Choose all that are true) Whether the participants were selected because they were outliers. Whether the randomization was well-concealed. Whether everyone who is included in the before group would be eligible to be in the after group. Whether or not there was a prevailing temporal trend during the time period of the study.
Whether the participants were selected because they were outliers. Whether everyone who is included in the before group would be eligible to be in the after group. Whether or not there was a prevailing temporal trend during the time period of the study.