Generally refers to various factors that can affect outcomes (such as diseases)

Characteristics or status of the research subjects related to the outcome (such as age, gender, occupation, genetics, behavior, lifestyle, etc.) or exposure to certain factors related to the outcome (such as X-ray exposure, heavy metals, environmental factors, etc.)

A characteristic, state, or factor is an exposure factor

fixed queue

dynamic queue

Generally refers to factors that can cause a specific adverse outcome or increase the probability of its occurrence, including personal behavior, lifestyle, environment, genetics and other factors.

According to whether the research subjects are exposed to a certain research factor or its different levels, the research subjects are divided into the exposed group (E) and the non-exposed group (Ē)

Follow up for a certain period of time and compare the difference in the incidence of the studied outcome (outcome) between the two groups to analyze the relationship between exposure factors and the study outcome.

forward-looking in time

is an observational study

Establish controls and group according to exposure or non-exposure

It is a study from cause to effect, consistent with chronological order, and has strong causal connection ability.

Determination of study cohort is now

Subjects are grouped according to their current exposure

Follow-up required*

The ending will occur at some point in the future

Chronological order enhances the credibility of etiological inferences

Direct access to exposure and outcome data, making the results credible

incidence rate

First-hand data has less bias

Requires a large sample size, costs a lot, and takes a long time

Research subjects are easily lost to follow-up, affecting feasibility

① Have clear test hypotheses

② The incidence of the disease under study is relatively high, generally not less than 5%

③ Clearly define exposure factors and outcome variables

④ Have reliable measurement methods

⑤ Sufficient observation groups and their exposure conditions can be obtained

⑥ Most people can complete the follow-up

⑦ Have sufficient human, financial and material resources

Grouping based on historical information available to the researcher at the beginning of the study regarding the subject's exposure at some point in the past

No prospective observation is required, the outcome has already occurred when the study begins

However, its nature is still forward-looking (from past exposure to current outcome), and it is a study from "cause" to "effect"

Save time, effort, and get results quickly

Complete data collection and analysis in a short period of time

The data is not under the control of the researcher when it is accumulated, and the content may not meet the requirements.

Sufficiently complete and reliable historical records or archival materials about the exposure and outcome of the research subjects during a certain period in the past are required

prone to selection bias and information bias

Difficulty controlling the interference of confounding factors

The integrity and authenticity of historical data will directly affect the feasibility of the research and the authenticity of the research results.

Continuing prospective cohort studies based on historical cohort studies

Suitable for evaluating the effects of exposure factors that have both short-term and long-term effects on human health

Determination of study cohort is past

Grouping research subjects based on exposure at a certain time in the past

Need follow-up

Some endings may have already occurred

It is a design model that combines prospective cohort studies with historical cohort studies, so it has the advantages of the above two categories.

professional group

Specially exposed groups

general population

organized group of people

internal control

external control

total population control

Multiple control

If the outcome incidence rate p1 of the exposed group cannot be obtained, you can try to obtain the relative risk (RR) value. p1 can be obtained by the formula p1=RR×p0

RR is the ratio of morbidity or mortality in the exposed population to the unexposed population.

To ensure the reliability of the study, the power level should be at least 0.80

P¯: the average of two incidence rates, q=1- P¯ P1: expected incidence rate in the exposed group, q1=1- P1 P0: expected incidence rate in the control group, q0=1- P0 Zα=1.96, Zβ=1.282 (critical values ​​of standard normal distribution corresponding to α and β)

① Demographic information (age, gender, occupation, education level, marital status, etc.) and possible confounding factor information

②Exposure factor information

③Outcome indicator information

Develop a uniform and detailed questionnaire

Check hospital, factory, unit and personal health records or files

Conduct relevant physical examinations, laboratory examinations and special examinations on research subjects

Conduct regular monitoring of environmental factors

Generally consistent with the content of the baseline data

The focus is on the outcome variable

All eligible subjects who completed the baseline survey (either exposed group or control group)

Research subjects experienced expected results

Normally: illness or death

Some indicators change

lost to follow-up

The deadline for the entire research work

Influencing factors

The interval is too short, which wastes manpower and material resources, and also brings unnecessary trouble and even harm to the research subjects.

If the interval is too long, it is easy to lose to follow-up, and no intermediate changes can be observed.

Generally, the follow-up interval for chronic diseases can be set at 1 to 3 years. The follow-up period for Framingham cardiovascular disease is every two years.

Then the follow-up person should be the baseline investigator to increase the comparability of follow-up and baseline surveys.

The study population is relatively large and the population is relatively stable (i.e., a fixed cohort). Regardless of the intensity of the disease and the length of observation time, the cumulative incidence of the disease under study can be calculated.

Divide the number of cases during the entire observation period by the population at the beginning of the observation period

Indicates the cumulative incidence of

The length of accumulation time must be stated, otherwise, its epidemiological significance will not be clear

Applicable conditions

When the observed population is unstable, the observation objects enter the study at different times, and are lost to follow-up for various reasons. Each observation object has a different follow-up time. Such a queue is a dynamic queue, and the total number of people is used as the unit. The calculation rate is unreasonable. The incidence rate is calculated in person-hour units, which is called incidence density.

The most commonly used unit of person-time is person-year. For example, if 10 research subjects are observed for 1 year or 1 research subject is observed for 10 years, it is 10 person-years.

Applicable conditions

When the number of research subjects is small and the incidence of outcome events is relatively low, it is not appropriate to directly calculate the rate regardless of the length of observation time. At this time, the incidence (or death) rate of the entire population can be used as a standard to calculate the expected number of incidences (or deaths) in the observation population; then the actual number of incidences (or deaths) in the observation population and the expected number of incidences (or deaths) can be obtained Ratio of people, that is, standardized incidence (or death) ratio

Applicable conditions

The meaning of SMR

Applicable conditions

Ie: incidence (or death) rate in the exposed group

Io: incidence (or mortality) rate in the non-exposed group

a: Number of cases in the exposed group

n1: sample size of exposure group

c: Number of cases in the control group

n0: sample size of control group

1) RR=1 means there is no difference in morbidity or mortality between the two groups;

2) RR>1 means that the morbidity or mortality rate of the exposed group is higher than that of the non-exposed group, exposure can increase the risk of morbidity or death, and exposure factors are risk factors for diseases;

3) RR<1 means that the morbidity or mortality rate of the exposed group is lower than that of the non-exposed group. Exposure can reduce the risk of morbidity or death, and exposure factors are protective factors.

The farther the RR value is from 1, the greater the effect of exposure and the greater the strength of the association between exposure and outcome.

From the perspective of RR, compared with non-smokers, smokers have a much greater risk of dying from lung cancer than from cardiovascular disease, indicating a strong link between smoking and lung cancer; From the perspective of AR, smoking has a greater effect on cardiovascular disease. If the smoking factor is eliminated, the mortality rate from cardiovascular disease can be reduced more significantly, that is, the social effect of prevention is greater.

Select a population that is convenient for follow-up

Expand calculated study sample by 10%*

Strengthen the management of follow-up staff

Develop follow-up plans and monitoring measures

Interim analysis

Conduct supplementary checks on objects with missing or omitted items

Limit research objects, match

Stratified analysis, standardized or multivariate analysis