Ocular epidemiology involves the application of descriptive epidemiological metrics to outline the frequency and distribution of eye diseases, as well as to elucidate their occurrence and patterns of spread. It makes use of analytical metrics to explore the reasons behind the distribution of eye diseases, the factors that influence their prevalence, and the effectiveness of prevention, treatment, or management measures. Within the framework of the modern bio-psycho-social medical model, ophthalmologists are expected to conduct clinical research alongside disease prevention and treatment efforts. Such research builds upon clinical medicine and integrates relevant disciplines such as epidemiology, biostatistics, social medicine, and health economics. The scope of study extends from individual patient cases to groups of affected populations, and from the clinical management of individual patients within hospitals to the prevention and treatment of diseases within community populations. This approach facilitates more comprehensive and in-depth investigations into the occurrence, progression, prognosis, and prevention of ocular diseases.
Common Research Methods in Ocular Epidemiology
Research methods in ocular epidemiology are broadly categorized into descriptive and analytical studies.
Descriptive Research
Descriptive research focuses on describing the number and distribution of diseases within a specific population rather than testing a specific causal hypothesis. It primarily addresses "who" is susceptible, "where" the disease occurs, and "when" it is most likely to appear, laying the groundwork for further analytical research. This approach employs qualitative or quantitative techniques, such as surveys, interviews, and observations, to collect data. The main types of descriptive studies are outlined as follows:
Case Reports
Case reports focus on individual or specific series of cases without the use of a control group. They primarily describe the occurrence and distribution of studied cases and are not intended for estimating the risk of the disease in question. The strengths of this approach include ease of data collection, fewer resources (time, manpower, and material requirements), and the opportunity for patients to receive appropriate treatment during the study. However, case reports are limited by weak evidentiary strength and less reliability. The absence of control groups may lead to potentially erroneous conclusions.
Epidemiological Description of Disease Occurrence
This type of research describes the occurrence and distribution of diseases within populations based on individual characteristics such as age, gender, ethnicity, education level, occupation, marital status, socioeconomic status, and personality traits. It also considers geographical variations (e.g., by country, city, or rural area) and temporal factors (e.g., seasonal trends). The primary goal is to identify "who" is more likely to develop the disease.
Descriptive Cross-Sectional Studies
Cross-sectional studies collect data on the prevalence of diseases or the health status of a defined population during a specific time period. These are also referred to as prevalence studies. Cross-sectional studies include two main approaches: sample surveys and censuses. Sample surveys involve randomly selecting a subset of the population (i.e., a sample) for observation. The results of the sample are then used to estimate the characteristics of the overall population it represents. To ensure representativeness, randomization principles must be followed during sampling. Common sampling methods include simple random sampling, systematic sampling, stratified sampling, cluster sampling, and multistage sampling. Sample surveys are cost-effective in terms of manpower, resources, and time, but they are more complex in terms of study design, implementation, and data analysis. Errors due to oversight or repetition are harder to identify, and this approach may not be suitable for research subjects with significant variability.
In contrast, censuses involve investigating or examining every member of a defined population within a specified time frame to obtain data on a particular disease or health condition.
Analytical Research
Analytical research refers to methods used to test specific causal hypotheses by examining the relationship between the exposure to a particular risk factor and the occurrence of a disease. It can be categorized into two main types: observational studies and experimental studies.
Observational Studies
Observational studies do not involve direct control over the level of exposure to the risk factor being studied. Instead, they rely on observation and analysis to achieve the research objectives. Common types of observational studies include:
Analytical Cross-Sectional Studies
Analytical cross-sectional studies measure both disease status and exposure to risk factors within a sample population at a specific point in time to identify potential associations. The benefits of this approach include lower costs and easier implementation. It does not require follow-up time and allows for the examination of relationships between multiple diseases and various exposure factors. Such studies can provide useful data for health planning in specific populations and are less disruptive to participants’ work and lives, ensuring better cooperation. However, limitations include the inability to determine the temporal sequence between disease occurrence and exposure factors, which prevents establishing a causal relationship. This method is unsuitable for studying diseases with very low prevalence, may face difficulties in randomly selecting large population samples, and can only measure disease prevalence rather than incidence or relative risk of disease development.
Case-Control Studies
Case-control studies compare a group of individuals with a disease (case group) to one or more groups of individuals without the disease (control group) to analyze past or present exposure to risk factors. The goal is to study the association between risk factors and disease occurrence and to determine the degree of this association. This is a "retrospective" observational study moving from "effect to cause," which can generate new hypotheses. The advantages of case-control studies include suitability for investigating rare diseases or diseases with long latency periods, the requirement of a smaller sample size, high efficiency, and relatively lower costs and time demands. However, drawbacks include low efficiency when researching rare exposure factors, lower reliability of retrospectively collected data, and difficulties in establishing temporal sequences between disease onset and exposure factors, making it challenging to confirm causality. There is also a high likelihood of bias when selecting case and control groups. These studies usually focus on a single exposure factor and do not provide data on disease prevalence, incidence, or relative risk of disease development.
Cohort Studies
Cohort studies involve comparing a group exposed to a particular risk factor (exposed group) with another group not exposed to the same risk factor (control group) to observe the occurrence of a specific disease over a certain period of time. Researchers in such studies do not randomly assign or actively control the exposure factors or treatments. Cohort studies are generally conducted using a prospective approach, making them a "cause-to-effect" type of research. These studies are applicable for describing the incidence of a particular disease over a defined time period and for analyzing the relationship between exposure factors and disease occurrence.
The advantages of cohort studies include the ability to establish causal relationships between exposure factors and disease because the effects of exposure are clearly assessed before the disease develops. Since disease onset occurs after the exposure, the status of the disease itself does not influence the selection of study participants or the measurement of exposure factors. Cohort studies are effective tools for identifying disease incidence and exploring potential causes and are useful for investigating multiple diseases that result from a single exposure factor and understanding their relationships with that factor.
However, cohort studies have limitations. They can be expensive and time-consuming, and efficiency is typically low when studying rare diseases. These studies often require a large sample size. Participant loss to follow-up reduces the effective sample size, and if a significant number of individuals exposed to a particular factor develop the disease before the study's follow-up period concludes, serious ethical concerns in medical research may arise.
Experimental Research or Clinical Trials
In experimental research, researchers observe the effects of exposure to certain factors on disease progression while maintaining control over the level of exposure to these factors. This method is often used in animal studies and is referred to as experimental research. Direct application of this method to humans may violate medical ethics. However, the effectiveness and safety of new drugs or treatments must be proven through clinical research before they can be widely applied in practice. Under conditions that do not violate medical ethics, the use of experimental research methods in human studies is referred to as clinical trials.
The advantages of clinical trials include better control over treatment and other confounding factors affecting disease outcomes, clear establishment of the temporal sequence between exposure and disease, and a higher likelihood of reproducible results. However, clinical trials also have disadvantages. Participants are often highly selected and may not be representative of the general population, making it difficult to generalize the findings to broader groups. Handling independent variables can sometimes be challenging, and ethical issues may arise in certain situations.
The basic principles of clinical trials include the use of controls, randomization of groups, and blinding methods.
Establishing Control Groups Under Balanced and Identical Conditions
Control groups serve as the basis for comparison in clinical trials, and proper design of the control group is a core element of trial design. The purpose of establishing control groups is to study the effects of treatment factors through comparative methods, exclude the influence of non-study factors on treatment effectiveness, and minimize or prevent the impact of bias and chance-related errors on trial results. Greater balance and comparability between the experimental and control groups enhance the ability to detect the effects of the study factors. The principle of balanced comparability requires that, other than the study factors, all other conditions should be as similar as possible between the experimental and control groups. Control design methods include:
- Paired Comparison Design: Subjects are paired based on certain characteristics or conditions, with each pair receiving different treatments.
- Within-Subject Control Design: The same patient serves as their own control by comparing treatment outcomes before and after treatment.
- Between-Group Comparison Design: Cases are divided into experimental and control groups, with control group design principles ensuring proper setup before the trial begins. Comparisons within the same time period are preferred over comparisons across different periods, and comparisons within the same institution are preferred over those involving external institutions. Both experimental and control groups should follow the principle of random group assignment.
Randomization of Groups
Clinical trials adhere to the principle of randomization when forming groups. Due to limitations in time, manpower, and resources, not all patients can be included in a trial, requiring a subset of the population to be selected as a sample to represent the whole. Randomization ensures that the results of a study can be generalized to the larger population. Random sampling differs from arbitrary selection, as the assignment of patients to experimental or control groups must exclude subjective biases from researchers or participants. Randomization requires specific techniques for implementation. Sampling involves selecting a study sample from the total population in a way that gives each unit an equal chance of being chosen. Random sampling techniques include simple random sampling, stratified random sampling, systematic sampling, and cluster sampling, which are often used in combination. Randomization methods vary, including drawing lots, rolling dice, or more scientific and convenient approaches such as using random number tables. Simpler methods also utilize calculators or computers with random number generators to produce a series of random numbers directly.
Blinding Method
Clinical trials aim to obtain unbiased results, although bias may occur at any stage, from trial design to data analysis. Bias can originate from participating healthcare providers or the patients themselves. Implementing blinding methods effectively prevents such biases, and there are three levels of blinding—single-blind, double-blind, and triple-blind methods.
Single-Blind Method
The researcher is aware of the treatment each subject receives, but the subject remains entirely unaware. This method is simple and easy to implement, eliminates psychological biases from the participants, and allows physicians to address any issues during treatment promptly. However, it may result in researcher-influenced bias when collecting and evaluating data. Healthcare providers may inconsistently apply efficacy standards between the experimental and control groups or feel uneasy about control group patients not receiving treatment, potentially providing "compensatory" treatment that affects the trial results. Control groups can involve the use of placebos, which are substances resembling the experimental drug in appearance but lack specific active ingredients. The placebo’s color, smell, solubility, and packaging should closely match the experimental drug. When placebos adversely affect patient outcomes, standard treatments may be used, though they must similarly match the experimental drug in form and appearance.
Double-Blind Method
Neither the participants nor the observers know the group assignment or treatment received by participants. This method reduces informational bias caused by subjective factors of both parties. Double-blind trials require a comprehensive coding and confidentiality system, along with measures to ensure participant safety. This method may not be suitable for treating critically ill patients.
Triple-Blind Method
Neither the participants, observers, nor data analysts are aware of group assignments. This method prevents bias during data analysis but can present practical challenges during implementation.
Double-blind, randomized controlled trials are commonly used in clinical studies.
Common Indicators in Ophthalmic Epidemiological Research
The primary tool for measuring diseases in epidemiological research is rates, which clearly illustrate the likelihood and risk of disease occurrence in a specific population during a given time period. Rates used to describe the frequency of diseases generally fall into two main categories:
Prevalence
Prevalence measures the likelihood of a specific disease having already occurred within a population at a particular point in time or over a specific time period. For calculation, the numerator represents the total number of existing cases of the disease, and the denominator represents the total population surveyed. Prevalence is not suitable for studies analyzing causative factors, but it can be a useful tool when planning health facilities and workforce needs. When necessary data for calculating incidence are unavailable, prevalence can also be used to estimate the significance of a disease within a population.
Incidence
Incidence determines the likelihood of healthy individuals, who are exposed to certain risk factors, developing a specific disease within a designated time frame. For calculation, the numerator represents the total number of new cases during the specified time period, while the denominator represents the total number of individuals at risk for developing the disease during the same period. Incidence serves as a fundamental tool for etiological studies as it directly measures the frequency of occurrence for acute or chronic diseases.
There is a clear relationship between incidence and prevalence, which can be expressed as P ≈ I x D, where P stands for prevalence, I for incidence, and D for the duration of the disease. If patients recover or die, they are no longer considered prevalent cases. This relationship indicates that prevalence changes directly with incidence and the duration of the disease. If incidence remains stable, the disease has a long duration, and the mortality rate among the diseased population is the same as for others in the population, then P = I x D. In such cases, knowing two of the variables enables calculation of the third.
The statistical intensity of the association between disease occurrence and exposure factors can be expressed using the ratio of incidence or prevalence between two groups.
In cohort studies, the ratio of incidence between two groups is referred to as the Relative Risk (RR), which indicates the statistical strength of the relationship between disease occurrence and exposure factors. When RR=1, it signifies that the incidence rate in the exposed population is the same as that in the unexposed population, indicating no association between the exposure factor and the disease; the factor is unlikely to be a cause. When RR >1, it indicates that the disease incidence in the exposed population is significantly higher than in the unexposed population, suggesting that the factor may be a cause. Conversely, when RR<1, it indicates that the factor is not a cause and may even have a protective effect, reducing the risk of disease in the population.
In case-control studies, where incidence cannot be calculated, the statistical strength of the relationship between disease and exposure factors can be expressed using the ratio of prevalence between two groups, referred to as the Odds Ratio (OR). When OR=1, there is no association between the exposure factor and the disease. When OR>1, it indicates an increased risk of disease due to the exposure factor. When OR<1, it implies a decreased risk of disease due to the exposure factor, meaning the factor may have a protective effect.