Diagnosis
Frequency
Fueled by a spirit of exploration, research is an ongoing quest for greater understanding. The most important purpose of chiropractic research is to provide chiropractors with accurate information so that they can be more helpful to their patients. Researchers have a responsibility to ask the questions that matter most. Students and practitioners also have a responsibility: to keep current with newly emerging information by being avid consumers of research.
Ideally this means subscribing to one or more scholarly journals and reading them on a regular basis. At the very least, it means reading detailed summaries and discussions of current research and searching MEDLINE and other databases when necessary for evidence-based answers to questions that arise in the course of daily practice. Every profession that holds science in high esteem expects all of its practitioners to stay well informed throughout their careers. For chiropractic to occupy a respected place among the health professions, this is absolutely essential.
Aside from its practical value in improving the quality of patient care, research also has a broader sociologic purpose; this purpose is related to the credibility and cultural authority that chiropractic derives from generating new knowledge. Without exception, the professions that thrive are those that are perceived as being the experts in a particular field. Although expertise has various components, none is more essential than generating new knowledge. If the chiropractic profession fails to do so in its areas of specialty, others will fill the gap.
Chiropractic research has three main divisions: (1) clinical, (2) basic science, and (3) health services research. Clinical research addresses issues directly related to patient care, as in a study by Boline and colleagues, 1 which found that chiropractic outperformed the medication amitriptyline for tension headache patients. Basic science research explores underlying structures and mechanisms, as in a study by Dishman and Bulbulian, 2 which investigated reflexes in the gastrocnemius muscle after spinal adjustment/manipulation and offered support for the theory that manual spinal procedures may lead to short-term inhibitory effects on the human motor system. Health services research studies chiropractic’s role in society, as in a study by Hawk and Long3 that evaluated use of chiropractic in several Midwestern states and reported that over 40% of workers with low back pain use chiropractic, that use increases with age, and that it is more common for rural compared with nonrural residents.
For chiropractic as for other health professions, all three types of research are important. The focus of this chapter is clinical research.
Active control—A procedure that may have therapeutic effects, performed on the control group in a clinical trial.
Abnormality—A condition that is not normal. Abnormality may be defined either statistically or in terms of adverse health consequences.
Anecdote—A story about a clinical event or case.
Basic science research—Research that explores structures and mechanisms.
Bias—A process tending to produce results that depart systematically from true values.
Blinding—Shielding the subject, the person administering the treatment, or the person assessing the outcome from knowledge that might compromise the integrity of a study.
Case series—A written report on the details of a series of related cases.
Case study—A written report on the details of a single case.
Clinical epidemiology—The application of epidemiologic principles to clinical case management (see definition of epidemiology).
Clinical research—Research that addresses issues directly related to patient care.
Clinical trial—A prospective longitudinal experiment designed to assess the comparative efficacy or effectiveness of a treatment, often labeled a randomized clinical trial (RCT) if random assignment of subjects is made to each of the comparison treatment groups (see definition of randomized clinical trial).
Cohort—A defined group of people observed over a period of time.
Confounding bias—Bias that results from the interaction of two or more factors in a cause-and-effect relationship such that the presence of one factor makes it difficult to evaluate the true effect of the other factors.
Control group—A comparison group assignment in a clinical trial that receives no treatment, a placebo treatment, or an alternative treatment (see definitions of placebo and active control).
Correlation—A consistent statistical relationship between two variables such that one variable tends to predict the other. It may suggest but not prove a causal relationship.
Cost-effectiveness—The relative health value of an intervention compared with its financial cost.
Cross-sectional studies—Measurements taken at one moment in time.
Diagnosis—The identification and measurement of an abnormality in a particular patient that has clinical ramifications.
Evidence-based health care—A health care system in which, to the greatest extent possible, procedures used by health care providers have been subjected to rigorous standards of scientific observation, experimentation, and documentation.
Experimental group—The group in a clinical trial that receives the intervention being tested.
External validity—The degree to which the results of a study can be expected to hold true in other settings.
False positive—(1) A test that is positive despite the absence of the disease or condition being tested; (2) a situation in which Intervention A is no better than Intervention B, but errors in the data or its interpretation lead one to reasonably conclude that Intervention A actually is better than Intervention B. Also called Type I error.
False negative—(1) A test that is negative despite the presence of the disease or condition being tested; (2) a situation in which Intervention A is actually superior to Intervention B, but errors in the data or its interpretation lead one to reasonably conclude that it is not superior to Intervention B. Also called Type II error.
Frequency—A general or statistical expression of how often a condition or disease occurs. Statistical expressions of frequency take two forms: (1) prevalence and (2) incidence (see definitions of prevalence and incidence).
Gold standard—A measure of agreed upon accuracy and validity.
Hawthorne effect—The tendency of people in a research study to change their behavior, compromising the validity of the data.
Health services research—Research that studies the structure, process, and outcomes of the health delivery system and its role in society.
Incidence—The proportion of a clearly defined group (i.e., population) initially free of a condition, that develops it over a given period of time.
Internal validity—The degree to which the results of a study are correct for the methods and sample of patients included in the study (see definitions of validity and external validity).
Interobserver reliability—The consistency of measured results between different practitioners evaluating the same thing (see definition of intraobserver reliability).
Intervention—A procedure used for treatment or prevention of illness.
Intraobserver reliability—The consistency with which one practitioner can consistently arrive at the same result (see definition of interobserver reliability).
Longitudinal—Involving serial measurements taken over time.
Magnitude—The size of an effect.
Measurement bias—Systematic error in data resulting from poor methods of observation.
Meta-analysis—A systematic review that usually includes a ranking of the quality of each study, plus statistical pooling of the data from all studies to determine the average effect of treatment (see definition of systematic review).
Natural history—The usual progression of an illness in the absence of intervention.
Observational study—A study in which the researcher observes events as they occur naturally or in the course of normal practice, without attempting to influence them.
Operational definition—A description of the methods, tools, and procedures required to make an observation (i.e., a definition that is specific and allows objective measurement).
Outcome measures—Parameters that indicate a change in health status, used in studies of treatment effectiveness. These must be specifically defined at the beginning of the study.
P-values—A statistical statement generally representing the probability of a false-positive conclusion, or Type I error. They are used to account for the possibility of random error and to measure the statistical significance of a finding.
Placebo—An intervention designed to mimic a “real” treatment and performed on a control group in a clinical trial. Its purpose is to control for the nonspecific therapeutic effects of care so that the specific effects of an actual treatment can be measured.
Pretest posttest design—A study design in which baseline (i.e., pretreatment) and outcome (i.e., posttreatment) measurements are performed.
Prevalence—The proportion of a population having a particular condition or outcome at a given moment.
Prognosis—The predicted course of an illness.
Prospective study—A study that generates new data based on events that occur after the study begins (see definition of retrospective).
Random error—The influence of chance variation.
Randomized controlled trial—A prospective, longitudinal study in which patients are divided into two or more groups on a randomized basis (see definition of clinical trial).
Reliability—The consistency of a measurement when repeated.
Retrospective study—A study that reviews events that have already occurred (see definition of prospective).
Risk—The likelihood of an adverse event or outcome. Risk is determined by measuring the relationship between the presence of possible risk factors and the subsequent incidence of particular conditions.
Sample—A subset of a population.
Selection bias—Bias that occurs when the group selected for study differs in significant ways from the true population.
Self-care—Methods used by individuals to enhance their own health.
Sensitivity—The proportion of times a diagnostic procedure is correct in patients without a specific diagnosis (see definition of specificity).
Sham adjustment/manipulation—A manual intervention used for research purposes that must be convincing enough for research subjects to believe it is a real adjustment/manipulation, although it has none of the specific physiologic or therapeutic effects of a real adjustment/manipulation. Also called placebo adjustment/manipulation.
Single-blind study—A study in which the patients are blinded as to whether they are in the experimental group or the comparison group.
Specificity—The proportion of times a diagnostic procedure is correct in patients without a specific diagnosis (see definition of sensitivity).
Statistical significance—A statement based on inferential statistics indicating that a good probability exists that a conclusion is not wrong because of random error. The magnitude of a therapeutic effect and the size of the sample are two key factors influencing statistical significance.
Systematic error—Error reflecting bias (see definition of bias).
Systematic review—A summary of scientific knowledge in an area accomplished by a review of published research, in which explicit objective methods are used to evaluate the methodological quality and the results.
Triple-blind study—A study in which patients, doctors, and outcome assessors are blinded.
Validity—The degree to which an observation or measurement provides an indication of the true state of the phenomena being measured. Also called accuracy.
CLINICAL EPIDEMIOLOGY
To be an intelligent reader of research papers, one must be familiar with the basic concepts and terminology of clinical epidemiology, because these form the basic structure of clinical research in the health sciences. Epidemiology is the study of the incidence and progression of diseases in populations. Clinical epidemiology, which is the application of epidemiological principles to patient case management, represents an attempt to bring order and quantitative methodology to the process of answering specific clinical questions that arise during patient care.
The questions addressed in clinical epidemiology are mostly the same questions that practitioners consciously or unconsciously ask themselves when evaluating patients and determining the most appropriate courses of action. As described in Fletcher, Fletcher, and Wagner’s classic text, Clinical Epidemiology: The Essentials,4 these questions can be grouped into the following categories: abnormality, diagnosis, frequency, risk, prognosis, treatment, prevention, cause, and cost. In applying research to case management, it helps to break down the process into these various components, to learn how to use the scientific literature to help answer the questions that arise each step along the way.
Abnormality
The concept of abnormality is a foundation for all health care practice, including chiropractic, because if no abnormality exists there may be no need for the intervention of a health care practitioner. To know whether something is abnormal presumes that one knows what is normal, but defining normal is not as simple as it first appears. A purely statistical approach defines normal as any condition that occurs most of the time in most people. However, this assumes that “usual” is the equivalent of “good” (i.e., that no disease or dysfunction occurs in the majority of a population). This is an unwise assumption, as demonstrated by the fact that most American adults have plaque deposits in their arteries, which are related to heart attacks and early deaths. In addition, some phenomena that are unusual are known to be positive health indicators. For example, very low cholesterol levels are rare in the population of the United States, but they tend to confer greater protection against heart disease than do statistically average (i.e., normal) cholesterol levels.
Far better than the statistical definition of abnormality is a model that defines abnormality in terms of a relation to negative health consequences. Using this model, normal means good and abnormal means bad, which is consistent with the way most people use these terms. To demonstrate whether a particular parameter is normal or abnormal (i.e., whether it is associated with adverse health consequences) can require extensive epidemiologic research. High blood pressure is a good example. Research has demonstrated that very high blood pressure in combination with other factors is closely related to stroke and other problems. Even mild hypertension has been shown to increase the risk of future cardiovascular disease. The key point here is that clinicians consider a certain range of blood pressure to be normal, not because most people have it, but because a threshold level exists beyond which negative health consequences begin.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
