Clinical Epidemiology and Biostatistics

Abstract

Epidemiology is the study of the distribution and determinants of disease frequency1. In the fifth century BC, Hippocrates suggested that the development of human disease might be related to the external and internal environment of an individual1. In the 1600s and 1800s in England, John Graunt and William Farr quantified vital statistics on the basis of birth and death records1. In the 1850s, John Snow associated cholera with water contamination in London by observing higher cholera rates in homes supplied by certain water sources1. Epidemiological methods gradually evolved with use of the case-control study to demonstrate an association between smoking and lung cancer, use of the prospective cohort study to determine risk factors for cardiovascular disease in the Framingham Heart Study, and use of the randomized clinical trial for the poliomyelitis vaccine1. The evidence-based medicine and patient-derived outcomes assessment movements burst onto the scene of clinical medicine in the 1980s and 1990s as a result of contemporaneous medical, societal, and economic influences. Pioneers such as Sackett and Feinstein emphasized levels of evidence and patient-centered outcomes assessment2-10. Work by Wennberg and colleagues revealed large small-area variations in clinical practice, with some patients being thirty times more likely to undergo an operative procedure than other patients with identical symptoms merely because of their geographic location11-16. Additional critical research suggested that up to 40% of some surgical procedures might be inappropriate and up to 85% of common medical treatments were not rigorously validated17-19. Meanwhile, the costs of health care were rapidly rising to over two billion dollars per day, increasing from 5.2% of the gross domestic product in 1960 to 16.2% in 199720. Health maintenance organizations and managed care emerged. In addition, increasing federal, state, and consumer oversight was brought to bear on the practice of clinical medicine. These forces have led to an increased focus on the effectiveness of clinical care. Clinical epidemiology provides the methodology with which to assess this effectiveness. This article presents an overview of the concepts of study design, hypothesis testing, measures of treatment effect, diagnostic performance, evidence-based medicine, outcomes assessment, data, and statistical analysis. Examples from the orthopaedic literature and a glossary of terminology (terms italicized throughout the text) are provided. Study Design In observational studies researchers observe patient groups without allocation of the intervention, whereas in experimental studies researchers allocate the treatment. Experimental studies involving humans are called trials. Research studies may be retrospective, meaning that the direction of inquiry is backward from the cases and that the events of interest transpired before the onset of the study. Alternatively, studies may be prospective, meaning that the direction of inquiry is forward from the cohort inception and that the events of interest transpire after the onset of the study (Fig. 1). Cross-sectional studies are used to survey one point in time. Longitudinal studies follow the same patients over multiple points in time.Fig. 1: Comparison of prospective and retrospective study designs on the basis of the direction of inquiry and the onset of the study.All research studies are susceptible to invalid conclusions due to bias, confounding, and chance. Bias is the non-random systematic error in the design or conduct of a study. Bias usually is not intentional; however, it is pervasive and insidious. Forms of bias can corrupt a study at any phase, including patient selection (selection and membership bias), study performance (performance and information bias), patient follow-up (nonresponder and transfer bias), and outcome determination (detection, recall, acceptability, and interviewer bias). Frequent biases in the orthopaedic literature include selection bias, when dissimilar groups are compared; nonresponder bias, when the follow-up rate is low; and interviewer bias, when the investigator determines the outcome. A confounder is a variable that has independent associations with both the independent (predictor) and dependent (outcome) variables, thus potentially distorting their relationship. For example, an association between knee laxity and anterior cruciate ligament injury may be confounded by female sex since women may have greater knee laxity and a higher risk of anterior cruciate ligament injury. Frequent confounders in clinical research include gender, age, socioeconomic status, and co-morbidities. As discussed below in the section on hypothesis testing, chance may lead to invalid conclusions based on the probability of type-I and type-II errors, which are related to p values and power. The adverse effects of bias, confounding, and chance can be minimized by study design and statistical analysis. Prospective studies minimize bias associated with patient selection, quality of information, attempts to recall preoperative status, and nonresponders. Randomization minimizes selection bias and equally distributes confounders. Blinding can further decrease bias, and matching can decrease confounding. Confounders can sometimes be controlled post hoc with the use of stratified analysis or multivariate methods. The effects of chance can be minimized by an adequate sample size based on power calculations and use of appropriate levels of significance in hypothesis testing. The ability of study design to optimize validity while minimizing bias, confounding, and chance is recognized by the adoption of hierarchical levels of evidence on the basis of study design (see Table [Levels of Evidence for Primary Research Question] in Instructions to Authors of this issue of The Journal). Furthermore, the standard to prove cause-effect is set higher than the standard to suggest an association. Inference of causation requires supporting data from non-observational studies such as a randomized clinical trial, a biologically plausible explanation, a relatively large effect size, reproducibility of findings, a temporal relationship between cause and effect, and a biological gradient demonstrated by a dose-response relationship. Observational study designs include case series, case-control studies, cross-sectional surveys, and cohort studies. A case series is a retrospective, descriptive account of a group of patients with interesting characteristics or a series of patients who have undergone an intervention. A case series that includes one patient is a case report. Case series are easy to construct and can provide a forum for the presentation of interesting or unusual observations. However, case series are often anecdotal, are subject to many possible biases, lack a hypothesis, and are difficult to compare with other series. Thus, case series are usually viewed as a means of generating hypotheses for additional studies but not as conclusive. A case-control study is a study in which the investigator identifies patients with an outcome of interest (cases) and patients without the outcome (controls) and then compares the two groups in terms of possible risk factors. The effects in a case-control study are frequently reported with use of the odds ratio. Case-control studies are efficient (particularly for the evaluation of unusual conditions or outcomes) and are relatively easy to perform. However, an appropriate control group may be difficult to identify, and preexisting high-quality medical records are essential. Moreover, case-control studies are susceptible to multiple biases, particularly selection and detection biases based on the identification of cases and controls. Cross-sectional surveys are often used to determine the prevalence of disease or to identify coexisting associations in patients with a particular condition at one particular point in time. The prevalence of a condition is the number of individuals with the condition divided by the total number of individuals at one point in time. Incidence, in contradistinction, refers to the number of individuals with the condition divided by the total number of individuals over a defined time period. Thus, prevalence data are usually obtained from a cross-sectional survey creating a proportion, whereas incidence data are usually obtained from a prospective cohort study and a time value is contained in the denominator. Surveys are also frequently performed to determine preferences and treatment patterns. Because cross-sectional studies represent a snapshot in time, they may be misleading the research the disease over time. Surveys also in terms of adequate and A cohort study is one in which a of interest is and is in to determine outcomes and associations with risk factors. or cohort studies can also be in studies, cohort are on the basis of and the follow-up is or in the studies are for the and risk factors of a disease because they are meaning that a group of is over time. The effects in a cohort study are frequently reported in terms of risk Because cohort studies are prospective, they can optimize follow-up and data quality and can minimize bias associated with selection, information, and In addition, they have the to provide evidence However, studies are are often a for and are for the assessment of unusual outcomes or Experimental study designs may the use of or controls. The randomized clinical trial with is the standard of clinical evidence as it provides the conclusions by minimizing the effects of bias and confounding. with patients is the means of confounding. The performance of a randomized control trial the of a that sample size, for the trial, of assessment of and data analysis. Because allocation is selection bias is minimized and confounders and are equally between Blinding minimizes performance, and Blinding may be at the intervention, outcome and analysis minimizes nonresponder and transfer bias, while determination adequate power. The that patients be the treatment group to which they were randomized in to the of the randomized clinical trial is the of clinical research the of such include their and time to of patients and by may be rapidly a may an randomized clinical trial or a randomized clinical trial difficult to randomized clinical clinical of treatment in the for to and to adverse and while randomized clinical have internal some have their because the practice and the of patients in a randomized clinical trial may be and are to the design and conduct of clinical research studies. is of and it is the focus of of the of be with the and the of as they to of and of and for The of hypothesis is to from a sample to the from which it or the that the not by chance but because of a association between the hypothesis of a study that is association between whereas the hypothesis that is a association. the of a study are not the hypothesis, whereas the are can the hypothesis and the Thus, research studies that are based on a sample an the in the a of the possible outcomes of a study can that the of a study is a association is not when is association or a association is when is a association. However, a study can have two of A type-I or error when a association is when is association in a study that a A type-II or error that is association in a study that a association error error value probability of type-I probability of type-II The refers to the probability of a type-I the of significance is set at which means that the of a association is than a one in chance that the association was due to chance Thus, the p which is with a statistical is a of the of the evidence by the data in of the the p value is than the then the evidence the hypothesis is for to it and that the result is values frequently are used in clinical research and are by and however, is a in to p values because a significance of p is a point can be misleading is between p and p but the is the p value information the of the and the p value may be without the being to the on p values include the use of variable levels of significance based on the of the type-I error and the of p values without the of in of p values has as information the significance of not they are the of and the of by the of the the p value is often as being or the provides a of values that the to the of the In addition, while p values have are in the of the variable of which the to the For example, the of a study of the of the for with of the managed to a clinical practice may that was a for patients to the with the of p values are used or to for patients to the and to for patients not to the are The statistical significance whereas the statistical significance not clinical significance of the and of the is the probability of a association one and is defined as the probability of a type-II power is set at which means that is chance that the study demonstrate association when is a association. In practice, when a study a the error of is the type-I error as by the p However, when a study the error of is the type-II error as by in a study that effect, may be effect or may be a effect but the study was because the sample size was or the were Thus, when a study effect, the power of the study be The calculations for power on the statistical methods that are for the however, are in a power effect size, and sample size size is the that to be to with the and is based on a clinical large a be are often defined in on the basis of a in values divided by the standard for a of two sample effect and large decrease the power of a study. of power is in clinical to minimize the use of when a study and to the validity of a study. calculations are performed when a study is being power is set at is set at the effect size and are from data or the and the is for the sample of power after the study has power and is A diagnostic can result in possible the is and the disease is the is and the disease is the is and the disease is and the is and the disease is The of a is the of patients with the disease who are as a result of the A with of patients with the have a have a A result of a disease The of a is the of patients without the disease who are as a result of the A with of patients without the have a have a A result of a disease in and can be a the which is the probability of a divided by the probability of a and can be in studies in which the of a diagnostic are with of the of in the same example, by the of with a rate rate value and value and are of diagnostic performance and have for and clinical practice however, they are in the clinical because the not the patient has the The are the probability of the patient the disease when the result is value and the probability of the patient not the disease when the result is value The and values are that an of the prevalence of the disease in the and they can be with use of that is an between and Because is some between the and groups with to a the investigator can a with a rate optimize or a rate optimize (Fig. In practice, are on the basis of the of a or a the of a the of a of a condition as of the in a more is This relationship between the and of a diagnostic can be with use of a A the relationship between the rate on the and the rate on the at possible (Fig. diagnostic performance can be on the basis of the the of Because is between the and the over a of diagnostic values is an between and the are as to the of point is because patients with the disease are as a However, is because some of the patients without the disease are as a result the are as to the of point is because some patients with the disease are as a result However, is because patients without the disease are as a for a clinical for from of the in The rate is on the and is on the The the the diagnostic performance of a or a diagnostic For a the the is For the the is of of include probability and is a between and that likely an is to on the basis of the number of events per the number of trials. The probability of on a is is the of the probability of an to the probability of the not The odds of up on a is Because probability and odds are they can be odds risk can be in a prospective cohort risk the incidence of disease in the cohort divided by the incidence of disease in the cohort For example, a prospective cohort study of with of the anterior cruciate ligament a higher of knee in who are not with a than in who are with a the risk is This can be as a times higher risk of knee injury in a with anterior cruciate ligament who is not with a than in such a who is with a A in a retrospective case-control study which incidence be is the odds which is the of the odds of the disease in the study group to the odds of the disease in the control group Experimental group a group rate experimental rate control odds experimental odds risk odds risk risk and number to that are likely to the or of a disease are called risk factors. The effect of a that the probability of an adverse outcome can be quantified by the risk the risk and the number to The effect of a that the probability of an adverse outcome can be quantified by the risk the risk and the number to refers to the medical care that a patient whereas outcome refers to the result of that medical care. The of the outcomes assessment has patient-derived outcomes measures include and measures of patient such as the are used to assess health or quality of as based on the of such as the knee or the are used to assess of a condition or of patient are used to assess of care and have including the evaluation of quality of patient-centered of and quality The of an outcomes the and creating A large number of outcomes have and used without assessment of their and to refers to the of an and to the of the when used by and by the same at can be by the to the same patient on two without an in the medical These are usually reported with use of the or refers to the measures it to validity an is of the being to validity an relationship to an validity an hypotheses and with to an values over the disease and treatment. medicine the and use of evidence in the care of medicine the research evidence with clinical and patient The of evidence-based medicine the for information an the evidence to that the evidence with to and and the critical with clinical and the values and The of in evidence-based medicine are to patients who have a particular Evidence is on the basis of study design (see Table [Levels of Evidence for Primary Research Question] in Instructions to Authors of this issue of The with an on randomized clinical and can be in evidence-based from the of Clinical and and evidence-based and A systematic is a of the medical literature in which methods are used to a literature and critical of studies. A of systematic is in which methods are used to the of independent studies randomized clinical to For example, a study that the literature for and of for internal and for the treatment of and then the outcomes and is a systematic the other a study in which the the literature for and of and then the patient data to a statistical analysis is a Clinical or clinical practice are that are on the basis of the to and optimize may also potentially of and variations in practice and decrease analysis is a that evaluation of conditions of The analysis is that a be often of and that The of analysis the of a to the determination of outcome and analysis to the value of to determine the (Fig. and analysis to determine the effect of outcome and on (Fig. analysis can identify the and this with variations in outcome or patient This used or with the of analysis for operative of are by chance are by and are by outcome are to the of the to are the or 1). treatment is because it has a higher value with analysis for operative of The probability of a from operative treatment is on the The represent the value for the operative and the value of from operative treatment treatment is study designs in medicine include studies, and In studies, the costs of the treatment are In the costs and clinical outcome are and reported as per clinical outcome. In both costs and are in In and are and are reported as per The on which a is has for the in which information is and can be or data or and can be of as The not represent an Examples include and a outcome. data are also called data are in terms of or and are reported in or is an then the data are The represent an but are not to Examples include and injury data are also in terms of or and are reported in or data are on a for which the between have meaning on a Examples include age, and a can have the of is called data are in terms of and standard and can be reported in or can be in terms of measures of such as and and in terms of measures of such as standard and can be by such as the and (Fig. or analysis the relationship of a independent and a dependent for means of that are include the for two independent groups and the for For or that are not statistical for include the as the to compare two independent groups and the to compare of is used to compare means of or more independent groups in which the data are The result is the which a p value is an determine are between are used to multiple between the include the and The is used to compare of or more independent groups in the data not follow a The is a to analysis of analysis of is used for when the study has The to compare or more groups is called the for and of of of analysis of used to compare for or include the for two or more independent groups and the when are or For the is used for two and the is used for or more that are used to determine associations include the for variables, for variables, and for analysis is used to data when the outcome of interest is the time an A group of patients is to determine they the of The point in analysis can be or a clinical point such as of a total A patient is when the of interest not to that the study period. is the of time from a the study the of interest or the time of The time for patient is when a is the of time that the total has or the of time that it before it to be is patients have not a at the of the study but at risk for in the These are in which information is because time is data are with use of the in which the from the is time that an but not at analysis is used when the of the point is points that have not are as at the of the follow-up for the analysis. analysis a the number of time and the number of patients the A can be to the of patients from on the and the follow-up time since the on the (Fig. The can be the at with use of for groups can be by the for the of the rates between patients who with a and who without a The rates were for patients with a p The error the represent by The of patients on the were based are in analysis between multiple is a of a relationship between an outcome variable and an variable or a set of independent is used when the outcome variable is with the of the that from data to a that as a of two or more or is used when the outcome variable is or and it has the common of multivariate analysis for methods include data and data is used to or to independent associations for and or For example, can be used to determine of of the in on the basis of an of and can be used to determine independent determinants of patient outcome with use of a outcome Because many usually a particular it is to use multivariate analysis. The of multivariate analysis is to identify, from the many patient and surgical and related to the outcome. multivariate a of information, and requires is an to have a in statistical methodology in the multivariate analysis. Epidemiology and are the of clinical of study design, hypothesis testing, diagnostic performance, measures of effect, outcomes assessment, evidence-based medicine, and is both for the investigator clinical research and for the clinical research