Hippocrates - "Air, Water and Place", a systematic statement on the relationship between the natural environment and health and disease

①Venice, Italy—Seaport Quarantine

②Sui Dynasty—Quarantine

John Graunt, London, UK - Life table proposed to establish a comparison group

Created research methods for chronic non-communicable diseases, including risk estimation methods.

for example

This was a period of considerable development in methods of epidemiological analysis, including the distinction between confounding and bias, interactions, and the practical development of case-control study designs.

for example

It is a period when epidemiology intersects with other disciplines, updates concepts and models, continuously introduces new sub-disciplines, and expands the application fields of epidemiology.

For example, in 1993, Schulte proposed ecological epidemiology

(Including communicable diseases, non-communicable diseases, parasitic diseases, etc.)

(Including accidents, disabilities, mental retardation, physical and mental injuries, etc.)

Including various functional states of body physiology and biochemistry, pre-disease state and longevity, etc.

“That is, a state of complete physical, mental and social adaptability in all aspects, not just the absence of disease or infirmity.”

i.e. reveal prevalence (mainly infectious diseases) or distribution (other diseases, injuries and health)

This can be achieved through descriptive epidemiological methods.

That is, starting from the analysis object to find out the rules and causes of prevalence and distribution.

Proposed etiological hypotheses can be tested or validated with the help of analytical epidemiological methods.

That is, rationally using the results of the first two stages to find prevention or control strategies and measures.

It can be verified by experimental epidemiological methods.

molecular epidemiology

infectious disease epidemiology

Special features️

Current situation research

monitor

ecological research

generate hypotheses

Case-control study (can both generate and test hypotheses)

array research

test hypothesis

individual test

community trial

(1) Definition

(2) Calculation formula

(3) Factors to be considered when calculating

(4) Application

(1) Definition

(2) Calculation formula

(3) Advantages

(4) Application

(1) Definition

(2) Subsequent cases

(3) Calculation formula

(4) Application

(1) Definition

(2) Type

(3) Reasons affecting the prevalence rate

(4) Relationship between prevalence, incidence, and course of disease

(5)Application

(6) Comparison with incidence rate

(1) Definition

(2) Application

(3) Calculation formula

(1) Related concepts

(2) Application

(3) Calculation formula

(1) Definition

(2) Application

(3) Calculation formula

(1) Definition

(2) Application

(3) Calculation formula

(1) Definition

(2) Application

(3) Calculation formula

(1) Definition

(2) Content

(1) Definition

(2) Content

Refers to the general level of incidence rate over the past years.

There is no obvious connection between the cases in terms of time and place of onset, and they appear to occur sporadicly.

Generally speaking, for a larger area, it is usually compared with the local incidence rate of the disease in the past three years.

① Diseases in which immunity is long-lasting after illness, or diseases in which the population maintains a certain level of immunity due to vaccination, often appear sporadic, such as measles.

② Some diseases that are mainly invisible infections often exist in sporadic forms, such as poliomyelitis and Japanese encephalitis.

③Some infectious diseases whose transmission mechanism is not easy to realize can also appear sporadic, such as typhus and anthrax

④Some long-term latent infectious diseases also exist in sporadic form, such as leprosy

Refers to local areas or collective units,

A sudden occurrence of many patients with the same symptoms in a short period of time.

①These people often have the same source of infection or transmission route.

②Most patients often appear at the same time between the shortest and longest incubation period of the disease. Such as outbreaks of measles, hand, foot and mouth disease, mumps, hepatitis A and other diseases in day care institutions.

It means that the incidence rate of a certain disease in a certain area significantly exceeds the annual incidence rate of the disease.

① Compared with sporadic cases, when an epidemic occurs, there are obvious temporal and spatial connections between cases.

②The prevalence of a certain disease indicates that there may be common transmission factors

The incidence rate of a certain disease significantly exceeds the annual incidence rate of the disease,

The disease spreads rapidly and covers a wide area.

In a short period of time, it has become a worldwide epidemic across provincial, national and even state boundaries.

Such as influenza and cholera.

①Definition

②Research objects

③Research content

④Research conclusion

⑤Research on diseases

⑥Application

①Definition

②Research objects

③Research content

④Research conclusion

⑤Research on diseases

⑥Research significance

It is the working environment, socioeconomic status, health and education level, and physical labor intensity of the workers.

and mental stress and other factors.

Related to exposure to pathogenic factors in the working environment.

Relevant to working conditions and protective facilities.

Different occupational groups have different types of diseases and different prevention and treatment priorities.

influence on disease occurrence.

1. City

2. Rural areas

1. City

2. Rural areas

1. City

2. Rural areas

1. City

2. Rural areas

1. City

2. Rural areas

The frequency of disease incidence and illness in a certain area is higher than that in surrounding areas,

The frequency of disease in the area exceeds random probability, which is called regional clustering of disease.

It is suggested that specific pathogenic factors in this region have an impact on population health.

(1) Definition of locality

(2) Endemic diseases

(3)Classification

Also known as exotic diseases,

Any infectious disease that does not exist or has been eliminated in the country or region,

When introduced from abroad or other regions, it is called an imported infectious disease.

Such as AIDS.

It is generally a disease epidemic or outbreak that lasts for several days, weeks or months.

It is the special way of existence of the disease

Outbreak→smaller range

Short-term fluctuation → larger range

Food poisoning in collective canteens

Outbreaks of typhoid, dysentery, and measles

chemical poisoning

The incidence of diseases increases in certain seasons

①Strictly seasonal

②Seasonal increase

It means that the frequency of disease fluctuates regularly at certain time intervals.

There is a peak of popularity every few years.

Commonly seen in respiratory infectious diseases, including influenza, epidemic cerebrospinal meningitis, whooping cough, chickenpox, diphtheria, etc.

(1) Factors such as dense population, heavy traffic and poor sanitary conditions are conducive to the spread of diseases.

(2) The communication mechanism is easy to implement

(3) Diseases that can form stable immunity after illness

(4) Pathogen mutation and the speed of mutation

Also called long-term variation or long-term change,

refers to a long period of time, usually several years or decades,

Changes in clinical characteristics, distribution status, epidemic intensity, etc. of the disease

Such as gastric cancer and scarlet fever.

①Changes in the cause or causative factors,

②Mutation of pathogens,

③Changes in the body’s immune status,

④The improvement of medical treatment and prevention level,

⑤The degree of improvement of the reporting and registration system, etc.

If the difference in morbidity or mortality of a disease is mainly the result of environmental factors,

Then the incidence or mortality rate of the disease in the immigrant population is different from that of the population in the country of origin (region).

And the morbidity or mortality rate is close to that of the local population in the country of immigration.

If the difference in incidence or mortality of a disease is mainly related to genetic factors,

Unlike the morbidity or mortality rates in the native population of the country of immigration,

The incidence or mortality rate of the disease among immigrant populations is similar to that of the population in the country of origin (region).

Also called case investigation,

Refers to going to the scene of the disease, the contact history of new cases, the disease or health status of family members and surrounding people

and investigate environmental factors that may be related to the disease

① Find out the causes and conditions of the cases under study,

②The purpose of controlling the spread of the epidemic and eliminating the source of the epidemic to prevent the recurrence of similar diseases.

Generally, they are patients with infectious diseases or non-infectious diseases, or cases with unknown causes.

By focusing on a specific point in time or period,

and collection and description of the distribution of diseases or health conditions and related factors among a specific range of people.

Provide clues to the cause for further research

①Time angle️

②From the perspective of observation and analysis indicators️

However, during data processing and analysis,

Groupings can be based on exposure characteristics or disease status.

Current situation research focuses on a specific point in time

or the status and association of exposure and disease in a group over a specific period of time.

Such as the time to measure high blood pressure and diagnose whether it is high blood pressure.

Reason one

Reason two

Factors such as gender, race, blood type, genotype, etc.

It exists before the onset of the disease and does not change depending on whether you have the disease or not.

Prerequisite 1

Prerequisite 2

Prerequisite three

It is required that the time interval between current situation studies should not be too long,

There was little change in incidence over this time frame, and the course of the disease was stable.

Incidence rate = the difference between two occurrence rates divided by the time interval.

(1) Concept

(2) Purpose

(1) Concept

(2)Purpose

(3)Attention

(1) Census

(2) Sampling survey

(1) Factors affecting sample size

(2)Sample size calculation formula

1. Non-random sampling

2. Random sampling

3. Comparison of sampling errors

①Basic personal situation.

②Occupational situation.

③Living habits and health care conditions.

④ Women’s fertility status.

⑤Environmental information.

⑥Demographic information.

Personnel participating in the investigation should be uniformly trained according to standard methods.

Let them master the investigation methods and ensure the consistency of data collection methods and standards.

(1) Obtained through experimental measurement or inspection.

(2) Prepare a questionnaire and survey the research subjects to obtain exposure and disease information.

(3) Utilize relevant information from routine data, routine registrations and reports, special inquiries and correspondence surveys, clinical examinations and other special examinations

①Select research objects subjectively.

② Randomly transform the sampling method.

③No response bias

④Survivor bias

①Reporting bias

②Recall bias

③Survey bias

④Measurement bias

In data analysis, the existence of confounding factors leads to discrepancies between the results and the actual situation.

It is widely used in the study of the etiology of chronic diseases or the study of the relationship between environmental variables and population disease states.

Provide a basis for the establishment of research hypotheses.

In disease surveillance, ecological studies can be applied to estimate and monitor disease trends,

Provide basis for formulating strategies and measures for disease prevention and control

Both are descriptive studies,

It is a cross-sectional study of a certain disease or health condition among people within a specific range at a specific time.

①Current situation research

②Ecological research

By observing the distribution of a certain disease among different groups of people or regions,

Etiological hypotheses are then proposed based on differences in disease distribution.

Comparative ecological studies do not require exposure data or complex data analysis methods.

Gastric cancer mortality in coastal areas is higher than in other areas️

The environment or diet in coastal areas may be one of the risk factors for gastric cancer.

Continuously observe changes in the average exposure level of a factor in a population

its relationship with changes in the incidence and mortality of a certain disease, and understand its changing trends;

By comparing the changes in disease frequency before and after exposure levels, the connection between a certain factor and a certain disease can be judged.

Often used interchangeably with comparative ecological research.

Smoking rate among the cardiovascular disease MONICA program population, etc.;

Changes in cardiovascular disease morbidity and mortality.

①Definition

②Cause of occurrence

In particular, some variables related to sociodemographic and environmental aspects are easy to correlate with each other (multicollinearity problem)

A group of people within a specific range is called a queue

Ⅰ. According to the nature of the queue definition

Ⅱ. According to the time when people enter the queue, they are divided into

The exposures in the study were not artificially administered or randomly assigned,

It exists objectively before research and is not determined by the will of the researcher.

A control group is usually established during the research design stage.

Control groups can also be established as needed during the data analysis phase.

The control group can be from the same population as the exposed group, or it can be from a different population.

In a cohort study, the subject's exposure is established at the outset (before the onset of disease).

Then explore the relationship between exposure factors and diseases, that is, first determine the causes, and then observe the effects longitudinally.

The researchers grasped the exposure status of the study subjects and followed up on the occurrence of outcomes.

and the outcome occurs in an exact number of exposed people,

Able to accurately calculate the incidence of outcomes,

The ability to estimate the risk of a certain outcome among the exposed population and determine the causal relationship is strong.

(1) Definition

(2) Advantages

(3) Disadvantages

(1) Definition

(2) Advantages

(3) Disadvantages

Also called a mixed cohort study,

That is, based on historical cohort studies, continue prospective observation for a period of time,

It is a design model that combines prospective cohort studies with historical cohort studies.

Qian has the above two types of advantages and makes up for their respective shortcomings.

① There should be clear test hypotheses, and the exposure factors of the test must be accurately identified

②The morbidity or mortality rate of the disease under study should be relatively high, such as not less than 5%.

③ The measurement of exposure factors should be clearly defined, and the exposure data of the observed population should be obtained with certainty.

④Clearly specify the outcome variables.

⑤There are enough people to observe.

⑥Most of the observation population should be followed up to the end of the study, and complete and reliable data should be obtained.

⑦ Have sufficient human, material and financial resources.

①～⑤, ⑥ It requires sufficiently complete and reliable historical records or archival data on the exposure and outcome of the research subjects during a certain period in the past.

Historical cohort study Prospective study, that is, historical The time from exposure to present observation cannot meet the research requirements

If the ending event does not occur or does not occur completely.

a. The outcome is not limited to morbidity and death, but also changes in health status and quality of life;

b. It can be either the ultimate outcome (such as morbidity or death) or an intermediate outcome (such as changes in molecules or serum);

c. Outcome variables can be either qualitative or quantitative, such as serum antibody titers, urine glucose, and blood lipids;

d. The outcome variable can be either negative (such as disease occurrence) or positive (such as disease recovery).

1. Selection of exposed groups

2. Selection of control groups

Generally, the sample size of the control group should not be smaller than that of the exposure group, usually the same amount.

It is necessary to estimate the loss to follow-up rate in advance and appropriately expand the sample size.

Assuming that the loss to follow-up rate is 10%, 10% can be added to the calculated sample size as the actual sample size.

The larger the d value, the smaller the sample size required.

If the incidence rate p₁ of the exposed group cannot be obtained, an estimate of its relative risk (RR) can be obtained.

p₁ can be obtained from P1=RR×p₀.

That is, the Type I error (false positive error) α value when testing a hypothesis.

The smaller the probability of false positive errors, the larger the sample size required.

Usually α=0.05 or 0.01 is chosen. When 0.01 is chosen, the sample size required is larger than when 0.05 is chosen.

Also called certainty (1-β),

β is the probability of Type II error (false negative error) when testing the hypothesis.

And 1-β is the ability to avoid false negatives when testing a hypothesis, that is, efficacy.

If the required power (1-β) is larger, that is, the β value is smaller, the larger the sample size required.

Usually β is 0.10, sometimes 0.20 is used.

① The selected research subjects should be followed until the observation termination period.

② Follow-up interviews are required for those who are lost to follow-up. If the follow-up cannot be traced, try to understand the reasons and analyze the reasons for the loss to follow-up.

① Direct face-to-face interviews, telephone interviews, environmental and disease monitoring, hospital medical records, etc.

② The same follow-up method should be adopted for the exposed group and the control group, and should remain unchanged throughout the follow-up process.

Generally, the content is consistent with the baseline data, but the focus of follow-up collection is on outcome variables.

Refers to the research object showing the expected results and reaching the observation endpoint.

This study subject will no longer be followed up.

The emphasis is on the expected results

If the expected outcome of an observation is coronary heart disease but a subject develops hypertension, the observation endpoint should not be considered to have been reached.

①Definition

②Two factors need to be considered

If the observation time is short, data can be collected once at the end of the observation.

If it is longer, multiple follow-up visits are needed. Generally, the follow-up interval for chronic diseases can be set at 1 to 2 years.

① Depending on the content of the follow-up, the investigator can be an ordinary interviewer, a laboratory technician, a clinician, etc., but the follow-up investigator must be carefully trained.

② The researcher can participate in the follow-up, but it is best not to participate in person, because the researcher's follow-up is prone to subjective bias, and uninformed outsiders can obtain more objective information.

① Scope of application

②Formula

Scope of application

formula

① Scope of application

②Formula

③Meaning

④Essence

⑤Epidemiological significance

① Scope of application

②Formula

When the study sample size is large and both p and 1-p are not too small, for example, when np and n(1-p) are both greater than 5, the frequency distribution of the sample rate approximates a normal distribution. At this time, the principle of normal distribution can be applied To test whether the difference in rates is significant, use the U test method to test the difference in rates between the exposed group and the control group.

If the rate is relatively low and the sample is small, the direct probability method, binomial distribution test or Poisson distribution test can be used instead.

When the rate is slightly larger and the sample is slightly larger, the significance test of the rate can be performed using the chi-square test of the four-grid table data.

The test of SMR or SPMR is actually a test of the deviation of the result value from 1. The test method can be x² test or scoring test.

①Risk ratio (OR)

②Rate Ratio (RR)

① Both risk ratio and rate ratio are the most useful indicators that reflect the strength of the association between exposure and morbidity (death). However, the values ​​of risk ratio and rate ratio in the same study are different because the cumulative incidence rate and incidence density are not equal.

②RR indicates how many times the risk of illness or death in the exposed group is that of the control group.

③The larger the RR value, the greater the effect of exposure and the greater the strength of the association between exposure and outcome.

④It has etiological significance.

Formula①

Formula②

Also called specific risk, risk difference (RD), excess risk

is the absolute value of the difference between the incidence rate in the exposed group and the incidence rate in the control group

Indicates the extent to which a hazard is specifically attributable to an exposure factor.

① Refers to the increased number of disease occurrences in the exposed population compared with the non-exposed population. If the exposure factors are eliminated, this number of disease occurrences can be reduced.

②It has disease prevention and public health significance

RR and AR are calculated by comparing the rates in the exposed group with those in the control group,

Describe the biological effect of the exposure, i.e., how causative the exposure is

Also known as etiology score (EF)

Refers to the morbidity or mortality in the exposed population, the proportion attributable to the exposure as a percentage of the total morbidity or death

The proportion of the hazard that is specifically attributable to the exposure

The portion of total population incidence that is attributable to exposure.

The increased incidence of disease among the exposed population compared with the general population

PAR=It-I0

PAR and PAR% are calculated by comparing the rates of the entire population and the control group.

and the extent to which morbidity or mortality in this population might be reduced by eliminating this factor.

They are related both to RR and AR and to the proportion of exposed persons in the population.

Also called population etiology score (PEF), the percentage of all morbidities (or deaths) in the total population

Describe the proportion of the population that is attributable to exposure

If there is a dose-response relationship for an exposure,

That is, the greater the dose of exposure, the greater the effect.

The more likely that exposure is the cause.

First list the incidence rates at different exposure levels,

Then the lowest exposure level group was used as the control,

Calculate the relative risk and attributable risk for each other exposure level group.

When necessary, a trend test should be conducted on changes in risk (or rate).

(1) Definition

(2) Situation

(3) Control

If misclassification bias occurs to the same extent in all observed groups,

The results will only affect the accuracy of diagnosis but not the relative relationship between two or more groups.

Their relative risk will generally be closer to 1 than it actually is.

If misclassification bias occurs in one group but not in the other group, or if the degree of misclassification is different between the two groups,

The result may be higher or lower than the actual relative risk.

① Certain restrictions can be placed on the research subjects (such as a certain age group, a certain gender) in order to obtain homogeneous research samples

② Matching method can be used in comparison selection to ensure the comparability of the two groups on some important variables.

First, the possibility of confounding should be judged based on the criteria for confounding.

Use stratified analysis, standardization or multi-factor analysis methods.

At the beginning of the study, subjects are not selected based on a certain exposure factor, but all people within a certain range who meet certain conditions are selected to form a comprehensive cohort (without grouping), and the exposure status of the cohort population to multiple suspicious factors is collected and their health is checked. situation;

Prospective observation, including changes in exposure to various suspected causes and the occurrence of various health outcomes;

After the study ends, the cohort members are divided into different exposure groups and control groups based on their exposure to a certain suspicious factor at the beginning of the study, and then the relationship between the exposure and possible related outcomes is analyzed according to the cohort study method.

(1) At the beginning of the study, there were no specific research factors and no grouping, but a large amount of information on suspicious factors was collected;

(2) In the final analysis, the researcher selects one research factor at a time according to his own interests, divides the cohort into the exposure group and the control group according to the presence or absence of the research factor in the baseline survey, and compares the outcome incidence rates of each group.

(3) It is essentially a cohort study. In the later stage of the study, the analysis report for each research factor is divided into the exposure group and the control group according to the presence or absence of the factor at the beginning of the study. It is prospective and based on the cause. For those who seek results, exposure also happens naturally.

(4) It can be understood as a complex of multiple cohort studies, which is very suitable for comprehensive etiological research on a certain type of disease, especially when the cause is not very clear and various diseases may have some common causes or complex relationships. Be applicable;

(5) This design is very suitable for studying the relationship between multiple causes and diseases, can reduce the cost of a single study, improve research efficiency, and effectively overcome some limitations of traditional cohort studies. For example, the Framingham Cardiovascular Disease Study in the United States and the Prospective Chronic Disease Study in China are typical studies of this type.

This kind of research is essentially a historical cohort study based on big data.

Convenient and fast, it can overcome the limitations of manpower, material resources, financial resources and time.

① First raise the research question (such as the relationship between obesity and hypertension);

② Find the exposure group and control group from big data (for example, those with normal blood pressure and obesity 10 years ago are the exposure group, and those with normal blood pressure and weight during the same period are the control group);

③ Check the current outcome (blood pressure) information of the two groups;

④Analyze the relationship between exposure and outcome

Also called a grouped case-control study,

That is, among the case and control populations specified in the design,

A certain number of research subjects were selected for comparison between groups.

There are no other restrictions or regulations on the selection of controls.

Generally, the number of controls should be equal to or greater than the number of cases.

However, the number of cases and controls does not need to be strictly proportional.

This method is easier to implement than the matching method.

However, the ability of the method itself to control confounding factors is weak.

This should be compensated for in the statistical analysis.

That is, the selected controls are required to be consistent with the cases in certain factors or characteristics.

The goal is to balance matching factors between cases and controls,

This eliminates the interference of these factors on the results.

This method can increase the statistical testing capabilities during analysis and improve research efficiency.

However, it also increases the difficulty of selecting controls, and data collection and statistical analysis are more troublesome.

(1) Definition

(2)Basic design methods

Case group: all cases

Control group: randomly selected from all cohort members included at baseline

(1) Definition

(2)Basic design methods

(3) Differences from nested case-control study

(1) Definition

(2) Application

transient effect

(1) Definition

(2) Application

(3) Requirements

(1) Cases should meet unified and clear disease diagnostic criteria.

(2) Try to use internationally accepted or domestically unified diagnostic standards, and use gold standards as much as possible, and apply pathological diagnosis as much as possible when necessary.

(3) For diseases for which there are no clear diagnostic criteria, the criteria can be customized according to the needs of research, but attention should be paid to balancing the false positive rate and false negative rate of the diagnostic criteria to make the diagnostic criteria moderately lenient and stringent.

(4) For a special research purpose, researchers can stipulate or restrict certain characteristics of research subjects, such as elderly cases, female cases, severe cases, cases in a certain city, etc.

(1) New cases

(2) Current cases

(3) Death cases

(1) Select cases from hospitals

(2) Select cases from the community population

① Must be a person who is confirmed not to suffer from the disease under study based on the same diagnostic criteria as the case.

② The control should be representative of the source population where the cases occurred, that is, the exposure distribution of the control should be consistent with the exposure distribution of the case source population.

(1) Patients with other diseases diagnosed in the same or multiple medical institutions

(2) Non-disease cases or healthy people in the community or group population

(3) Neighbors of the case or healthy people in the same residential area or non-cases of the disease

(4) The case’s spouse, compatriots, relatives, classmates or colleagues, etc.

(1) Non-matching

(2) Match

The number of controls is rXn.

The number of cases (n) required for a case-control study when the number of cases and controls is not equal

That is, the distribution frequency of the general characteristics of the research subjects, such as age, gender, occupation, place of residence, etc., and the clinical classification of the cases is described.

If it is a random sample of cases from a certain population, it needs to be compared with the characteristics of all cases in the corresponding time and region.

That is to compare whether certain basic characteristics of the case group and the control group are similar or identical.

The purpose is to test the comparability of the two groups. If the difference between two groups in some basic characteristics is statistically significant,

Then its possible impact on the research results should be considered and controlled during inferential analysis.

1. Analysis of association between exposure and disease→X² test

2. Association strength analysis→OR (odds ratio)

3. Calculate the 95% confidence interval of OR

4. Estimated attributable risk percentage (AR%) and population attributable risk percentage (PAR%)

1. Analysis of the correlation between exposure and disease

1. Analysis of the correlation between exposure and disease

3. Calculate the 95% confidence interval of OR → Miettinen method

1. Definition

2. Form

4. Calculation steps