Part 2

A case series describes a group of X who have the same health X or who have undergone the same X.

A case report is a report that describes # patient. 

A case series is a report that describes a X of individuals who have the same X or X or who have undergone the same X. A case series can be written and disseminated only when a researcher has access to an appropriate source of cases and there is a compelling reason to write about those cases. This study approach can be useful for a variety of purposes, including:

Describing the characteristics of and similarities among a group of individuals with the same signs and/or symptoms of disease

Identifying new syndromes and refining case definitions

Clarifying typical progression of a disease or disorder

Describing atypical presentations of a disease or disorder or unusual complications from a treatment

Developing hypotheses for future research

A case series describes a group of patients who have the same health condition or who have undergone the same procedure.

A case report is a report that describes one patient. 

A case series is a report that describes a group of individuals who have the same disease or disorder or who have undergone the same procedure. A case series can be written and disseminated only when a researcher has access to an appropriate source of cases and there is a compelling reason to write about those cases. This study approach can be useful for a variety of purposes, including:

Describing the characteristics of and similarities among a group of individuals with the same signs and/or symptoms of disease

Identifying new syndromes and refining case definitions

Clarifying typical progression of a disease or disorder

Describing atypical presentations of a disease or disorder or unusual complications from a treatment

Developing hypotheses for future research

Key Characteristics of a Case Series

• Objective: Describe a X of individuals with a X.

• Primary study question: What are the key characteristics of the X included in the study?

• Population: All individuals included in the study must have the same X or X or have undergone the same X.

• When to use this approach: A source of X is available, and no X group is required or available.

• Requirement: An appropriate source of X is available.

First steps

• Specify what new and important information the X will provide.

• Identify a source of X.

• Assign a X definition.

• Select the characteristics of the study X that will be described.

• What to watch out for: A lack of X.

• Key statistical measure: Only X statistics are required.

Key Characteristics of a Case Series

• Objective: Describe a group of individuals with a disease.

• Primary study question: What are the key characteristics of the cases included in the study?

• Population: All individuals included in the study must have the same disease or disorder or have undergone the same procedure.

• When to use this approach: A source of cases is available, and no comparison group is required or available.

• Requirement: An appropriate source of cases is available.

First steps

• Specify what new and important information the analysis will provide.

• Identify a source of cases.

• Assign a case definition.

• Select the characteristics of the study population that will be described.

• What to watch out for: A lack of generalizability.

• Key statistical measure: Only descriptive statistics are required.

A researcher conducting a case series must select # disease, disorder, or procedure of interest, determine what will be new and interesting about the study, and identify an appropriate and available source of X.

The next step is to establish a clear case X. 

A case definition is a list of the X and X criteria that must be met in order for an individual to be classified as a person with the X of interest in a case series, a case–control study, or another type of study. Case definitions are also essential for any X investigation, no matter which study approach is used to investigate the event.

The first step in writing a case definition is clarifying what constitutes the X or disorder of X. 

A researcher conducting a case series must select one disease, disorder, or procedure of interest, determine what will be new and interesting about the study, and identify an appropriate and available source of cases.

The next step is to establish a clear case definition. 

A case definition is a list of the inclusion and exclusion criteria that must be met in order for an individual to be classified as a person with the disease of interest in a case series, a case–control study, or another type of study. Case definitions are also essential for any outbreak investigation, no matter which study approach is used to investigate the event.

The first step in writing a case definition is clarifying what constitutes the disease or disorder of interest. 

….objective indication of disease that can be clinically observed, such as a rash, cough, fever, or elevated blood pressure. Sometimes a case definition can be based solely on these types of clinical observations. However, for many diseases and disorders, clinical observations and laboratory test results are not sufficient on their own to yield a valid diagnosis. 

…subjective indication of illness that is experienced by an individual but cannot be directly observed by others. For example, when a patient rates his or her pain on a scale from 0 to 10, no one else can verify that the patient is accurately reporting his or her pain level. A clinician might observe behaviors consistent with severe pain but cannot measure pain. Illness and sickness describe other subjective components of a disease.

….describes how a person perceives his or her own experience of having an adverse health condition. 

…describes how a person with an adverse health condition relates to and is regarded by his or her community. 

… is a collection of signs and symptoms that occur together.

A sign is an objective indication of disease that can be clinically observed, such as a rash, cough, fever, or elevated blood pressure. Sometimes a case definition can be based solely on these types of clinical observations. However, for many diseases and disorders, clinical observations and laboratory test results are not sufficient on their own to yield a valid diagnosis. 

A symptom is a subjective indication of illness that is experienced by an individual but cannot be directly observed by others. For example, when a patient rates his or her pain on a scale from 0 to 10, no one else can verify that the patient is accurately reporting his or her pain level. A clinician might observe behaviors consistent with severe pain but cannot measure pain. Illness and sickness describe other subjective components of a disease.

 Illness describes how a person perceives his or her own experience of having an adverse health condition. 

Sickness describes how a person with an adverse health condition relates to and is regarded by his or her community. 

A syndrome is a collection of signs and symptoms that occur together.

After a clinician makes a diagnosis based on X and X, details about a patient’s condition are added to the patient’s medical record. Diagnostic and procedural codes used as part of this charting process can be a valuable part of a case definition, especially since the codes allow electronic databases of clinical records to be searched for potentially eligible patients.

 X codes are diagnostic categorizations based on the International Classification of Diseases (ICD), more formally called the International Statistical Classification of Diseases and Related Health Problems. ICD codes can be used to search clinical X for patients with the same X. 

CPT codes are the Current Procedural Terminology codes published by the American Medical Association. CPT codes can be used to search clinical or insurance records for patients who have undergone the same X or received the same X.

After a clinician makes a diagnosis based on signs and symptoms, details about a patient’s condition are added to the patient’s medical record. Diagnostic and procedural codes used as part of this charting process can be a valuable part of a case definition, especially since the codes allow electronic databases of clinical records to be searched for potentially eligible patients.

 ICD codes are diagnostic categorizations based on the International Classification of Diseases (ICD), more formally called the International Statistical Classification of Diseases and Related Health Problems. ICD codes can be used to search clinical records for patients with the same diagnosis. 

The other important step in establishing a case definition is selecting the relevant PPTs, which stands for the X, X, and X characteristics that set the context for a case series or for other types of descriptive epidemiology studies. 

An ICD code or CPT code alone is not sufficient as a case definition, because it does not provide information about the X of cases and other eligibility criteria. Will demographic characteristics, such as age or sex, be used to eliminate some patients with relevant diagnoses from the study population? Will only patients of particular clinics or hospitals, or residents of particular cities or counties, be eligible for inclusion? Will the study include 1 yearof files, or will it include 5 or more years? 

A comprehensive case definition must include a X description plus an appropriate set of X that specify additional X and X criteria.

A case series might be constructed from X data acquired by X cases about their experiences using a X and/or X techniques. These data might be supplemented or confirmed with a review of the participants’ medical X. Alternatively, a case series can be—and often is—based solely on X data, usually acquired from a review of patient X.

The other important step in establishing a case definition is selecting the relevant PPTs, which stands for the person, place, and time characteristics that set the context for a case series or for other types of descriptive epidemiology studies. 

An ICD code or CPT code alone is not sufficient as a case definition, because it does not provide information about the sources of cases and other eligibility criteria. Will demographic characteristics, such as age or sex, be used to eliminate some patients with relevant diagnoses from the study population? Will only patients of particular clinics or hospitals, or residents of particular cities or counties, be eligible for inclusion? Will the study include 1 yearof files, or will it include 5 or more years? 

A comprehensive case definition must include a disease description plus an appropriate set of PPTs that specify additional inclusion and exclusion criteria.

A case series might be constructed from primary data acquired by interviewing cases about their experiences using a questionnaire and/or qualitative techniques. These data might be supplemented or confirmed with a review of the participants’ medical records. Alternatively, a case series can be—and often is—based solely on secondary data, usually acquired from a review of patient charts.

Using Medical Records in a Case Series (Simplified Notes)

• Create a data extraction form/questionnaire to collect information from patient charts consistently.

• Limitation: Medical records are written for X care, not X, so important details may be missing.

• Signs, symptoms, patient comments, and clinician observations are not always X.

• If something is not mentioned in the chart, it does not mean it was X—it may simply not have been documented.

• For each item on the data collection form, include three response options:

  • X = symptom/exposure was present

  • X = symptom/exposure was absent

  • No X = chart does not indicate whether it was present or absent

• During analysis, pay close attention to the amount and type of missing information, as it can affect interpretation of the results.

Key idea: "Not recorded" does not mean "not present."

Using Medical Records in a Case Series (Simplified Notes)

• Create a data extraction form/questionnaire to collect information from patient charts consistently.

• Limitation: Medical records are written for clinical care, not research, so important details may be missing.

• Signs, symptoms, patient comments, and clinician observations are not always recorded.

• If something is not mentioned in the chart, it does not mean it was absent—it may simply not have been documented.

• For each item on the data collection form, include three response options:

  • Yes = symptom/exposure was present

  • No = symptom/exposure was absent

  • No information = chart does not indicate whether it was present or absent

• During analysis, pay close attention to the amount and type of missing information, as it can affect interpretation of the results.

Key idea: "Not recorded" does not mean "not present."

Ethical Considerations in Case Series Studies (Simplified Notes)Protecting Patient Privacy

• Researchers must protect patient confidentiality and privacy.

• Follow all laws and regulations regarding the use of medical records.

• All case series studies require approval from at least # research ethics committee/IRB.

Informed Consent

• Researchers usually need X that patients consented to the use of their health data.

• Some hospitals obtain X X from patients to use X records for research.

• Researchers may access records only after identifying information (e.g., names, addresses) has been removed.

• Patients who opt out of research are excluded from the data provided to researchers.

• Additional study-specific consent may be required if:

  • Researchers contact patients directly.

  • Researchers access identifiable information.

Use of Photographs

• Patients must give X consent before photographs are taken.

• Researchers must follow medical center policies and privacy laws.

• Consent is usually required before photos are:

  • Presented X.

  • Published in X.

• Consent may be needed even if the image appears non-X.

Preventing Identification of Participants

• Extra caution is needed when:

  • The disease/procedure is X.

  • The X and X details are very specific.

• People familiar with the community may still identify participants.

• Remove all potentially identifiable information before publication.

Key Idea

The main ethical concern in a case series is protecting patient privacy through consent, confidentiality, and deidentification of data.

Most case study reports do not require any numbers beyond simple counts and percentages. When the sample size is sufficiently large, statistical tests may be used to compare subpopulations of cases or to compare before-and-after measures of included patients.

Ethical Considerations in Case Series Studies (Simplified Notes)Protecting Patient Privacy

• Researchers must protect patient confidentiality and privacy.

• Follow all laws and regulations regarding the use of medical records.

• All case series studies require approval from at least 1 research ethics committee/IRB.

Informed Consent

• Researchers usually need proof that patients consented to the use of their health data.

• Some hospitals obtain general consent from patients to use deidentified records for research.

• Researchers may access records only after identifying information (e.g., names, addresses) has been removed.

• Patients who opt out of research are excluded from the data provided to researchers.

• Additional study-specific consent may be required if:

  • Researchers contact patients directly.

  • Researchers access identifiable information.

Use of Photographs

• Patients must give separate consent before photographs are taken.

• Researchers must follow medical center policies and privacy laws.

• Consent is usually required before photos are:

  • Presented publicly.

  • Published in journals.

• Consent may be needed even if the image appears non-identifiable.

Preventing Identification of Participants

• Extra caution is needed when:

  • The disease/procedure is rare.

  • The location and time details are very specific.

• People familiar with the community may still identify participants.

• Remove all potentially identifiable information before publication.

Key Idea

The main ethical concern in a case series is protecting patient privacy through consent, confidentiality, and deidentification of data.

Some case series may benefit from the use of various measures of morbidity and mortality. 

The case fatality rate (CFR) is the proportion of people with a particular X who X as a result of that X. The CFR is different from the mortality rate, which is the percentage of members of a population who die from any X (an all-cause mortality rate) or from a particular condition (a cause-specific mortality rate) during a specified time period.

 Mortality rates are typically expressed in units such as “per #.”

Some case series may benefit from the use of various measures of morbidity and mortality. 

The case fatality rate (CFR) is the proportion of people with a particular disease who die as a result of that condition. The CFR is different from the mortality rate, which is the percentage of members of a population who die from any condition (an all-cause mortality rate) or from a particular condition (a cause-specific mortality rate) during a specified time period.

 Mortality rates are typically expressed in units such as “per 100,000.”

The CFR is also different from the proportionate mortality rate (PMR), which is the proportion of X in a population during a particular X period that were attributable to a particular X. Both the CFR and the PMR are often expressed as X.

X X: all deaths/all populations

X: death from disease X/All deaths

X: Death from disease X/People with disease X

The CFR is also different from the proportionate mortality rate (PMR), which is the proportion of deaths in a population during a particular time period that were attributable to a particular cause. Both the CFR and the PMR are often expressed as percentages.

Mortality rate: all deaths/all populations

PMR: death from disease X/All deaths

CFR: Death from disease X/People with disease X

A cross-sectional study provides a snapshot of the X status of a X at # point in time

A cross-sectional study, also called a X study, measures the proportion of members of a X who have a particular X or X at a particular point in time. Typically made over a X duration of time and must be based on a X sample of the source population. Cross-sectional studies are used to:

Describe communities.

Assess population needs.

Support program planning.

Monitor and evaluate programs.

Establish baseline data prior to the initiation of longitudinal studies.

A cross-sectional study provides a snapshot of the health status of a population at one point in time

A cross-sectional study, also called aprevalence study, measures the proportion of members of a population who have a particular exposure or disease at a particular point in time. Typically made over a short duration of time and must be based on a representative sample of the source population. Cross-sectional studies are used to:

Describe communities.

Assess population needs.

Support program planning.

Monitor and evaluate programs.

Establish baseline data prior to the initiation of longitudinal studies.

Key Characteristics of Cross-Sectional Studies

• Objective: Describe the X and/or X status of a X.

• Primary study question: What is the X of the X and/or X in the X?

• Population: The study participants must be X of the source population from which they were drawn.

• When to use this approach: Time is X and/or the budget is X.

• Requirement: The exposures and outcomes are relatively X, and the researchers expect to be able to recruit several # participants.

First steps

• Define a source X.

• Develop a strategy for recruiting a X sample.

• Decide on the methods to be used for data collection.

• What to watch out for: Non-X of the study population.

• Key statistical measure: PX.

Key Characteristics of Cross-Sectional Studies

• Objective: Describe the exposure and/or disease status of a population.

• Primary study question: What is the prevalence of the exposure and/or disease in the population?

• Population: The study participants must be representative of the source population from which they were drawn.

• When to use this approach: Time is limited and/or the budget is small.

• Requirement: The exposures and outcomes are relatively common, and the researchers expect to be able to recruit several hundred participants.

First steps

• Define a source population.

• Develop a strategy for recruiting a representative sample.

• Decide on the methods to be used for data collection.

• What to watch out for: Non-representativeness of the study population.

• Key statistical measure: Prevalence.

Because cross-sectional studies are X- and X-effective, they are the most popular approach used for X epidemiology.

Cross-sectional studies use the X epidemiological study design. The researcher asks a number of people—usually a few hundred—to complete a short X, and then those data are analyzed. However, there is one very important requirement: The participants must be reasonably X of one well-defined population. 

….is the degree to which the participants in a study are similar to the source population from which they were drawn. If the results are intended to reflect the profile of a particular population group, then the study’s sampling strategy must recruit a population that is as diverse as the source population. 

Because cross-sectional studies are time- and cost-effective, they are the most popular approach used for descriptive epidemiology.

Cross-sectional studies use the simplest epidemiological study design. The researcher asks a number of people—usually a few hundred—to complete a short questionnaire, and then those data are analyzed. However, there is one very important requirement: The participants must be reasonably representative of one well-defined population. 

Representativeness is the degree to which the participants in a study are similar to the source population from which they were drawn. If the results are intended to reflect the profile of a particular population group, then the study’s sampling strategy must recruit a population that is as diverse as the source population. 

For research conducted with human populations, a X is the gathering of data from individuals using a list of questions. 

One commonly used tool for collecting data during a cross-sectional study is a KAP survey, a survey instrument that asks participants about their X, X (or X or X), and X (or X). Helpful for identifying gaps between what people X and how they X on that knowledge. For example, the adults who complete a KAP survey form might demonstrate high knowledge about the benefits of exercise on cardiovascular health but at the same time indicate that they exercise rarely because a variety of perceived barriers prevent them from being as physically active as they know they ought to be for maximum fitness.

A repeated cross-sectional study is a series of cross-sectional studies that resample and resurvey representatives from the X source population at two or more different X points. A repeated cross-sectional study design does not track the X individuals forward in time. Instead, a X set of participants is sampled from the source populationfor each round of data collection. Some people may happen by chance to be selected for more than one round of surveying, but their answers to the different surveys are not linked. This is different from a longitudinal study that follows the X people forward in time. Repeated cross-sectional surveys can reveal trends in X-level status over time, but they do not allow for the examination of X-level changes.Repeated cross-sectional surveys are often conducted annually or every few years as part of national health surveillance programs. 

For research conducted with human populations, a survey is the gathering of data from individuals using a list of questions. 

One commonly used tool for collecting data during a cross-sectional study is a KAP survey, a survey instrument that asks participants about their knowledge, attitudes (or beliefs or perceptions), and practices (or behaviors). Helpful for identifying gaps between what people know and how they act on that knowledge. For example, the adults who complete a KAP survey form might demonstrate high knowledge about the benefits of exercise on cardiovascular health but at the same time indicate that they exercise rarely because a variety of perceived barriers prevent them from being as physically active as they know they ought to be for maximum fitness.

A repeated cross-sectional study is a series of cross-sectional studies that resample and resurvey representatives from the same source population at two or more different time points. A repeated cross-sectional study design does not track the same individuals forward in time. Instead, a new set of participants is sampled from the source populationfor each round of data collection. Some people may happen by chance to be selected for more than one round of surveying, but their answers to the different surveys are not linked. This is different from a longitudinal study that follows the same people forward in time. Repeated cross-sectional surveys can reveal trends in population-level status over time, but they do not allow for the examination of individual-level changes.Repeated cross-sectional surveys are often conducted annually or every few years as part of national health surveillance programs. 

…is the process of continually monitoring health events in a population so that emerging public health threats can be detected and appropriate control measures can be implemented quickly. Health officials can use surveys of randomly sampled population members as part of monitoring long-term trends in public health. Population-based surveys are used for many of the largest studies conducted by the U.S. Centers for Disease Control and Prevention (CDC), including the Behavioral Risk Factor Surveillance System (BRFSS), the National Health and Nutrition Examination Survey (NHANES), and the U.S. National Health Interview Survey (NHIS).

X surveillance is the process of public health officials contacting healthcare X in their jurisdictions to ask how often the clinicians are X particular types of X. 

X surveillance is the compilation of X of notifiable disease diagnoses submitted by medical X.

Surveillance is the process of continually monitoring health events in a population so that emerging public health threats can be detected and appropriate control measures can be implemented quickly. Health officials can use surveys of randomly sampled population members as part of monitoring long-term trends in public health. Population-based surveys are used for many of the largest studies conducted by the U.S. Centers for Disease Control and Prevention (CDC), including the Behavioral Risk Factor Surveillance System (BRFSS), the National Health and Nutrition Examination Survey (NHANES), and the U.S. National Health Interview Survey (NHIS).

Active surveillance is the process of public health officials contacting healthcare providers in their jurisdictions to ask how often the clinicians are diagnosing particular types of disease. 

Passive surveillance is the compilation of reports of notifiable disease diagnoses submitted by medical laboratories.

… is the proportion of people with a disease who are diagnosed as having that disease. The CDR will be low if few patients are tested for suspected infectious diseases or other conditions that can be verified through various types of medical tests. 

The case detection rate (CDR) is the proportion of people with a disease who are diagnosed as having that disease. The CDR will be low if few patients are tested for suspected infectious diseases or other conditions that can be verified through various types of medical tests. 

… is the process of tracking potential outbreaks or other disease events based on reports of symptoms rather than relying solely on counts oflaboratory-confirmed diagnoses. Crowd-sourced data culled from social media sites and other platforms can assist with the recognition of growing population health concerns.

Syndromic surveillance is the process of tracking potential outbreaks or other disease events based on reports of symptoms rather than relying solely on counts oflaboratory-confirmed diagnoses. Crowd-sourced data culled from social media sites and other platforms can assist with the recognition of growing population health concerns.

…surveillance is the continuouscollection and analysis of high-quality data from a limited number of clinics or hospitals so that public health officials will be able to detect changes in health status occurring in the larger population from which the sentinel sites were sampled.

 Sentinel surveillance is the continuouscollection and analysis of high-quality data from a limited number of clinics or hospitals so that public health officials will be able to detect changes in health status occurring in the larger population from which the sentinel sites were sampled.

Cross-sectional studies measure the X of various exposures or exposure histories, diseases and disorders, and demographic characteristics in # well-defined X at one point in time or over a short duration of time, with all data collected within a few days, weeks, or months. 

The most common result reported for a cross-sectional survey is the X, the percentage of members of a population who have a given trait at the time of a study.

… measures the proportion of a population with a particular characteristic at one point in time…. is a “snapshot” of population status, such as the percentage of 18- to 64-year-olds in a city who were current smokers as of July 1, 2020. 

…. measures the proportion of a population with a particular characteristic during a defined time period, such as several weeks or several months. A…might ask what percentage of eighth-grade students in a school district have ever been told by a doctor that they have asthma or what percentage of those students have had a dental checkup during the past 6 months.

Cross-sectional studies measure the prevalence of various exposures or exposure histories, diseases and disorders, and demographic characteristics in one well-defined population at one point in time or over a short duration of time, with all data collected within a few days, weeks, or months. 

The most common result reported for a cross-sectional survey is the prevalence, the percentage of members of a population who have a given trait at the time of a study.

A point prevalence measures the proportion of a population with a particular characteristic at one point in time. A point prevalence is a “snapshot” of population status, such as the percentage of 18- to 64-year-olds in a city who were current smokers as of July 1, 2020. 

A period prevalence measures the proportion of a population with a particular characteristic during a defined time period, such as several weeks or several months. A period prevalence might ask what percentage of eighth-grade students in a school district have ever been told by a doctor that they have asthma or what percentage of those students have had a dental checkup during the past 6 months.

X measures can also be used as part of the analysis of data from cross-sectional studies. For example, …compares the prevalence of a characteristic in two independent populations (or independentsubpopulations of study participants) by taking a ratio of their prevalence rates. Populations are independent when no individual study participants are members of more than one of the populations being compared. A PR could compare the prevalence rates for males and females or compare the prevalence rates for people whoidentify as current smokers and those who identify as never having used tobacco products.

Because a cross-sectional study has no X dimension, it cannot be used to assess X. An exposure can be said to be “X with” or “X to” a disease, but a cross-sectional study cannot show that an exposure X a disease.

Comparative measures can also be used as part of the analysis of data from cross-sectional studies. For example, a prevalence ratio (PR) compares the prevalence of a characteristic in two independent populations (or independentsubpopulations of study participants) by taking a ratio of their prevalence rates. Populations are independent when no individual study participants are members of more than one of the populations being compared. A PR could compare the prevalence rates for males and females or compare the prevalence rates for people whoidentify as current smokers and those who identify as never having used tobacco products.

Because a cross-sectional study has no time dimension, it cannot be used to assess causality. An exposure can be said to be “associated with” or “related to” a disease, but a cross-sectional study cannot show that an exposure caused a disease.

A case–control study X the X histories of people with and without a particular disease in order to identify likely X factors for the disease.

A case–control study is a study that X the X histories of people with disease (X) and people without disease (X). 

…is a study participant with the infectious or parasitic disease, noncommunicable disease, neuropsychiatric condition, injury, or other disease, disorder, disability, or health condition of interest to the researcher. … in a case–control study is a participant who does not have the disease being examined. (The term “control” has a different meaning in experimental studies. In an experiment, a control is a participant who is assigned not to receive the active X.)

Individual participants in a case–control study are selected for inclusion in the study based on their X status, then both cases and controls are asked the same set of questions about past X. A case–control study is often the best study approach for identifying possible X factors for a disease. This is especially true when the disease is X, so a study of the general population would be unlikely to yield more than a few cases. A special type of statistic—an X ratio—is used to identify likely risk factors.

A case–control study compares the exposure histories of people with and without a particular disease in order to identify likely risk factors for the disease.

A case–control study is a study that compares the exposure histories of people with disease (cases) and people without disease (controls). 

A case is a study participant with the infectious or parasitic disease, noncommunicable disease, neuropsychiatric condition, injury, or other disease, disorder, disability, or health condition of interest to the researcher. A control in a case–control study is a participant who does not have the disease being examined. (The term “control” has a different meaning in experimental studies. In an experiment, a control is a participant who is assigned not to receive the active intervention.)

Individual participants in a case–control study are selected for inclusion in the study based on their disease status, then both cases and controls are asked the same set of questions about past exposures. A case–control study is often the best study approach for identifying possible risk factors for a disease. This is especially true when the disease is uncommon, so a study of the general population would be unlikely to yield more than a few cases. A special type of statistic—an odds ratio—is used to identify likely risk factors.

Key Characteristics of Case–Control Studies

• Objective: X exposure histories of people with a disease (cases) and people without that disease (controls).

• Primary study question: Do X and X have different X histories?

• Population: Cases and controls must be X except for their x status.

• When to use this approach: The disease is relatively X, but a source of X is available.

• Requirement: A source of X is available.

First steps

• Identify a source of X.

• Assign a case X.

• Decide what type of X population will be appropriate for the study.

• Decide whether cases and controls will be X.

• What to watch out for: X x

• Key statistical measure: X X (x).

Key Characteristics of Case–Control Studies

• Objective: Compare exposure histories of people with a disease (cases) and people without that disease (controls).

• Primary study question: Do cases and controls have different exposure histories?

• Population: Cases and controls must be similar except for their disease status.

• When to use this approach: The disease is relatively uncommon, but a source of cases is available.

• Requirement: A source of cases is available.

First steps

• Identify a source of cases.

• Assign a case definition.

• Decide what type of control population will be appropriate for the study.

• Decide whether cases and controls will be matched.

• What to watch out for: Recall bias.

• Key statistical measure: Odds ratio (OR).

Case–Control Study: How Participants Are Classified

Step 1: Determine Current X Status

• Disease (X)

• No disease (X)

Step 2: Determine Past X Status

Ask: Did the individual have the exposure at some point in the past?

Key Idea

• Participants are selected based on X status first (case or control).

• Researchers then look backward to determine whether they had the X.

• The exposure histories of cases and controls are compared to identify possible X factors.

• The main statistical measure used is the X X (OR).

Case–Control Study: How Participants Are ClassifiedStep 1: Determine Current Disease Status

• Disease (case)

• No disease (control)

Step 2: Determine Past Exposure Status

Ask: Did the individual have the exposure at some point in the past?

Key Idea

• Participants are selected based on disease status first (case or control).

• Researchers then look backward to determine whether they had the exposure.

• The exposure histories of cases and controls are compared to identify possible risk factors.

• The main statistical measure used is the Odds Ratio (OR).

Because case–control studies require an X number of cases in order to be valid, the first step in designing a case–control study is to identify an appropriate and accessible source of individuals who have the disease or disorder of interest. Hospitals, specialty clinics, physicians’ offices, public health agencies, disease registries, and disease support groups may be able to assist researchers in identifying individuals who are likely to meet the study’s case definition.

Most organizations will not release any information about patients or members until after a research protocol has been approved by the appropriate X X committee. Researchers accessing health data from cases and controls must exercise extreme care to protect the privacy of potential participants and the confidentiality of their personal information.

All cases must have the X disease, disorder, disability, or other health-related condition, and the study’s case definition must specify exactly what characteristics must be present or absent for a person to be deemed a case. Clinical manuals and reports about previous studies of the disease can be helpful references for drafting and refining the inclusion and exclusion criteria. The case definition should include X, X, and X (X) characteristics.

Because case–control studies require an adequate number of cases in order to be valid, the first step in designing a case–control study is to identify an appropriate and accessible source of individuals who have the disease or disorder of interest. Hospitals, specialty clinics, physicians’ offices, public health agencies, disease registries, and disease support groups may be able to assist researchers in identifying individuals who are likely to meet the study’s case definition.

Most organizations will not release any information about patients or members until after a research protocol has been approved by the appropriate ethics oversight committee. Researchers accessing health data from cases and controls must exercise extreme care to protect the privacy of potential participants and the confidentiality of their personal information.

All cases must have the same disease, disorder, disability, or other health-related condition, and the study’s case definition must specify exactly what characteristics must be present or absent for a person to be deemed a case. Clinical manuals and reports about previous studies of the disease can be helpful references for drafting and refining the inclusion and exclusion criteria. The case definition should include person, place, and time (PPT) characteristics.

Next, an appropriate source of controls must be selected. Depending on the goals of the study, controls may be recruited from, among other sources:

Friends and relatives of cases

Hospital or clinic patients without the disease of interest

The general population

A control definition is a list of all of the eligibility criteria for X in a comparison population. Controls must be X to cases except for their X status, so the inclusion and exclusion criteria for cases that do not specifically relate to the disease must also apply to controls. For example, if cases must be males between 25 and 39 years of age, controls must also be men in this age group.

Individuals who do not meet the case definition or the control definition must be X from the study. Excluded individuals may not meet one of the PPT criteria for inclusion, or they may have an intermediate or indeterminate disease status that prevents them from meeting either the case or the control definition. 

Next, an appropriate source of controls must be selected. Depending on the goals of the study, controls may be recruited from, among other sources:

Friends and relatives of cases

Hospital or clinic patients without the disease of interest

The general population

A control definition is a list of all of the eligibility criteria for inclusion in a comparison population. Controls must be similar to cases except for their disease status, so the inclusion and exclusion criteria for cases that do not specifically relate to the disease must also apply to controls. For example, if cases must be males between 25 and 39 years of age, controls must also be men in this age group.

… the process of recruiting one or more controls who are demographically similar to each case. (For a cohort study, matching describes the process of recruiting one or more X individuals who are demographically similar to each X person participating in the study.) There are three often-used options for matching: no matching, frequency (group) matching, and matched-pairs (individual) matching.

Many case–control studies use X matching. They simply assume that similar inclusion and exclusion criteria for cases and controls will result in case and control populations that have similar distributions according to sex, age group, socioeconomic status, and other characteristics that may be confounders of the association between the key exposure and the disease.

Matching in a case–control study describes the process of recruiting one or more controls who are demographically similar to each case. (For a cohort study, matching describes the process of recruiting one or more unexposed individuals who are demographically similar to each exposed person participating in the study.) There are three often-used options for matching: no matching, frequency (group) matching, and matched-pairs (individual) matching.

Many case–control studies use no matching. They simply assume that similar inclusion and exclusion criteria for cases and controls will result in case and control populations that have similar distributions according to sex, age group, socioeconomic status, and other characteristics that may be confounders of the association between the key exposure and the disease.

…(also called X matching) is a sampling design that ensures that cases and controls in a case–control study have similar group-level demographic characteristics. (In cohort studies, frequency matching ensures that X and X participants have similar X-level profiles.) For example, suppose a study is using hospitalized cases and controls. For each case, the researcher may select one control from the hospital registration files who was admitted the same week as the case, is the same sex as the case, and is within ±3 years of the age of the case, but does not have the disease of interest. Frequency matching can be used to identify one, two, or several controls for each case. For group matching, the goal is to recruit a X population that is similar to the X population. Individual cases are not linked to individual controls during analysis, so the analysis uses the same approaches that are used for unmatched case–control studies.

Frequency matching (also called group matching) is a sampling design that ensures that cases and controls in a case–control study have similar group-level demographic characteristics. (In cohort studies, frequency matching ensures that exposed and unexposed participants have similar group-level profiles.) For example, suppose a study is using hospitalized cases and controls. For each case, the researcher may select one control from the hospital registration files who was admitted the same week as the case, is the same sex as the case, and is within ±3 years of the age of the case, but does not have the disease of interest. Frequency matching can be used to identify one, two, or several controls for each case. For group matching, the goal is to recruit a control population that is similar to the case population. Individual cases are not linked to individual controls during analysis, so the analysis uses the same approaches that are used for unmatched case–control studies.

… matching (also called X matching) is a sampling design that links X case in a case–control study to one or more controls with similar characteristics, such as genetic siblings or community members with the same date of birth. (In cohort studies, matched-pairs matching links each X individual to one or more X individuals.) This approach is common in X studies that link each case to a close genetic relative. When this type of matching is used for a study, the pairs are X for analysis. The unit of analysis is the X, not the X. This approach requires the use of special statistical tests.

Matched-pairs matching (also called individual matching) is a sampling design that links each case in a case–control study to one or more controls with similar characteristics, such as genetic siblings or community members with the same date of birth. (In cohort studies, matched-pairs matching links each exposed individual to one or more unexposed individuals.) This approach is common in genetic studies that link each case to a close genetic relative. When this type of matching is used for a study, the pairs are linked for analysis. The unit of analysis is the pair, not the individuals. This approach requires the use of special statistical tests.

For both frequency matching and matched-pairs matching, it is important not to X. ...describes the recruiting challenges and possible statistical bias that can result from matching too many characteristics of the cases and controls (or exposed and unexposed participants). The demographic and exposure variables used as matching criteria cannot be evaluated as possible X factors for the disease. For example, suppose that cases and controls are frequency matched based on the date of hospital admission, sex, and age. The case and control populations will, by design, have the same proportion of admissions in April, the exact same percentages of males and females, and nearly identical mean ages. As a result of this forced similarity, the study will not be able to examine whether cases are more or less likely than controls to require hospitalization in a certain month, to be males, or to be octogenarians. Additionally, when there are more matching characteristics, it can be difficult to find controls who meet all of the matching criteria. The study population may end up being quite different from the general population because of the strict eligibility requirements, and this may limit the usefulness of the study. Overmatching may also result in a statistical X that obscures the relationship between an exposure and the disease.

For both frequency matching and matched-pairs matching, it is important not to overmatch. Overmatching describes the recruiting challenges and possible statistical bias that can result from matching too many characteristics of the cases and controls (or exposed and unexposed participants). The demographic and exposure variables used as matching criteria cannot be evaluated as possible risk factors for the disease. For example, suppose that cases and controls are frequency matched based on the date of hospital admission, sex, and age. The case and control populations will, by design, have the same proportion of admissions in April, the exact same percentages of males and females, and nearly identical mean ages. As a result of this forced similarity, the study will not be able to examine whether cases are more or less likely than controls to require hospitalization in a certain month, to be males, or to be octogenarians. Additionally, when there are more matching characteristics, it can be difficult to find controls who meet all of the matching criteria. The study population may end up being quite different from the general population because of the strict eligibility requirements, and this may limit the usefulness of the study. Overmatching may also result in a statistical bias that obscures the relationship between an exposure and the disease.

As the protocol is developed, researchers should seek to minimize the likelihood of various types of X occurring. 

…is a systematic flaw in the design, conduct, or analysis of a study that can cause the results of the study not to accurately reflect the truth about the source population. A key word in the definition of bias is “X,” which means methodical, orderly, and routine. By contrast, bias is not something that happens X or by X. A flawed research protocol will introduce bias into any sample drawn from the same source population, because bias is a X error.

A strong survey instrument for a case–control study will ask each participant questions that confirm whether the respondent is a X, a X, or X. The disease status for each participant may need to be confirmed by clinical or laboratory testing or other types of secondary verification. The researcher must ensure that only confirmed cases and confirmed controls are included in the analysis. 

Adhering to X definitions for what constitutes a case and what constitutes a control minimizes the risk of X bias, which occurs when participants are not correctly categorized, such as when some controls in a case–control study are incorrectly classified as cases or some cases are incorrectly classified as controls due to a systematic problem with the case definition or the control definition.

The protocol for a case–control study must also seek to minimize X bias, which occurs when cases and controls systematically have different memories of the past. Participants are often asked to remember events from the distant past that cannot be confirmed by documents created around the time when the exposure would have occurred. Cases may be searching for answers to questions about why they have become ill. As a result, they may have more vivid memories of participation or lack of participation in activities perceived to be risky or beneficial. Although there is no way to prove that recall bias is occurring because of systematically different memories among cases and controls, the results of case–control studies must be interpreted cautiously in light of the possibility that differential recall may have influenced the findings.

As the protocol is developed, researchers should seek to minimize the likelihood of various types of bias occurring. 

Bias is a systematic flaw in the design, conduct, or analysis of a study that can cause the results of the study not to accurately reflect the truth about the source population. A key word in the definition of bias is “systematic,” which means methodical, orderly, and routine. By contrast, bias is not something that happens randomly or by chance. A flawed research protocol will introduce bias into any sample drawn from the same source population, because bias is a systematic error.

A strong survey instrument for a case–control study will ask each participant questions that confirm whether the respondent is a case, a control, or neither. The disease status for each participant may need to be confirmed by clinical or laboratory testing or other types of secondary verification. The researcher must ensure that only confirmed cases and confirmed controls are included in the analysis. 

Adhering to strict definitions for what constitutes a case and what constitutes a control minimizes the risk of misclassification bias, which occurs when participants are not correctly categorized, such as when some controls in a case–control study are incorrectly classified as cases or some cases are incorrectly classified as controls due to a systematic problem with the case definition or the control definition.

The protocol for a case–control study must also seek to minimize recall bias, which occurs when cases and controls systematically have different memories of the past. Participants are often asked to remember events from the distant past that cannot be confirmed by documents created around the time when the exposure would have occurred. Cases may be searching for answers to questions about why they have become ill. As a result, they may have more vivid memories of participation or lack of participation in activities perceived to be risky or beneficial. Although there is no way to prove that recall bias is occurring because of systematically different memories among cases and controls, the results of case–control studies must be interpreted cautiously in light of the possibility that differential recall may have influenced the findings.