Chapter 1: A Concept of Clinical Research

The ultimate purpose of a profession is to develop a knowledge base that will maximize the effectiveness of practice. To that end, health professionals have recognized the necessity for documenting and testing elements of clinical practice through rigorous and objective analysis and scientific inquiry. The concept of evidence-based practice represents the fundamental principle that the provision of quality care will depend on our ability to make choices that have been confirmed by sound scientific data, and that our decisions are based on the best evidence currently available. If we look at the foundations of clinical practice, however, we are faced with the reality that often compels practitioners to make intelligent, logical, best-guess decisions when scientific evidence is either incomplete or unavailable.

This situation is even more of an issue because of the economic challenges that continue to confront health care. Clinical research has, therefore, become an imperative, driving clinical judgments, the organization of practice, and reimbursement. The task of addressing the needs of the present and future is one that falls on the shoulders of all clinicians—whether we function as consumers of professional literature or scientific investigators—to collect meaningful data, to analyze outcomes, and to critically apply research findings to promote optimal clinical care. Through collaborative and interdisciplinary efforts, researchers and clinicians share a responsibility to explore the broadest implications of their work, to contribute to balanced scientific thought. The purpose of this text is to provide a frame of reference that will bring together the comprehensive skills needed to promote critical inquiry as part of the clinical decision making process.

In this chapter we develop a concept of research that can be applied to clinical practice, as a method of generating new knowledge and providing evidence to justify treatment choices. We will explore an historic perspective of clinical research, the framework of evidence-based practice, the different types of research that can be applied to clinical questions, and the process of clinical research.

The concept of research in health professions has evolved along with the development of techniques of practice and changes in the health care system. Traditionally, research has connoted controlled laboratory experiments, run by scientists in white lab coats using complex instrumentation; however, the maturation of a clinical profession brings with it the realization that research has a broader meaning as it is applied to the patients and situations encountered in practice. Clinical research is a structured process of investigating facts and theories and exploring connections. It proceeds in a systematic way to examine clinical conditions and outcomes, to establish relationships among clinical phenomena, to generate evidence for decision making and to provide the impetus for improving methods of practice.

Clinical research must be empirical and critical; that is, results must be observable, documented and examined for their validity.¹ This objective process is, however, also a dynamic and creative activity, performed in many different settings, using a variety of quantitative and qualitative measurement tools and focusing on the application of clinical theory and interventions. It is a way of satisfying one's curiosity about clinical phenomena, stimulating the intellectual pursuit of truth to understand or explain clinical events, and generating new or different ways of viewing clinical problems.

The context of clinical research is often seen within a prevailing paradigm. Scientific paradigms have been described as ways of looking at the world that define both the problems that can be addressed and the range of legitimate evidence that contributes to solutions.² We can appreciate changes in research standards and priorities in terms of three paradigm shifts that have emerged in rehabilitation and medicine through the latter half of the 20th century: the focus on outcomes research to document effectiveness, the application of models of health and disability and most recently an attention to evidence-based practice.

The concept of looking at outcomes as the validation of quality care is not a new one. Historically, the triad of structure, process and outcomes has been used as the barometer of healthcare quality.³ Structure was assessed through organizational standards, and process through quality assurance programs examining details such as charges and record keeping. Outcomes of care were typically assessed in terms of morbidity, mortality, length of stay and readmissions.

In rehabilitation, outcomes were often related to improvements in impairments or pathologies, with the assumption that such changes would be linked to the ultimate outcomes of interest. Today, the concept of outcomes has been expanded to fit with the World Health Organization's definition of health, which includes physical, social and psychological well-being.⁴ Looking at the effects of intervention now includes consideration of patient satisfaction, patient preferences, self-assessment of functional capacity and quality of life. Clinicians and especially patients have always considered functional outcome as the ultimate measure of the success of intervention. At this time, however, consumers and reimbursement policies have obligated health care practitioners to define and document outcomes, and to substantiate the efficiency and effectiveness of treatment.

To be meaningful the outcomes agenda must influence public policy, routine monitoring of medical care and standardized assessment of patient outcomes.⁵ Clinical practice databases must be developed to include functional outcome measures and other relevant information to contribute to the evaluation of outcomes.⁶ Clinical managers now use these data to support practice and organizational structure. The objective of outcomes management has generated a renewed understanding of the link between clinical management decisions, treatment decisions and measured documentation of effectiveness.^7,8,9,10,11 Outcomes management has emerged as an interdisciplinary process aimed at determining best practices and identifying opportunities for improvement of clinical quality through intermediate and long-term outcome analysis.¹²

Outcomes research refers to the study of success of interventions in clinical practice, with a focus on the end results of patient care in terms of disability and survival.¹³ Such studies often use large administrative databases that include information about insurance coverage and utilization of services in addition to functional outcomes. Patients are frequently followed over time after discharge.

Outcome Measures

Outcomes can be documented in many ways. Economic indicators are traditional outcomes, interpreted within the context of cost effectiveness or cost-benefit ratio; that is, what is the relative cost in terms of success of outcomes? For example, Harp¹⁴ demonstrated how revenue, patient outcomes, staff productivity, costs and patient satisfaction could reflect the success of a rehabilitation program for patients with back and neck problems. Fakhry and associates¹⁵ looked at the effectiveness of using evidence-based guidelines for treating patients with traumatic brain injury (TBI), a condition they described as costing billions of dollars annually. They used a protocol based on Brain Trauma Foundation guidelines, and demonstrated that adherence to this protocol resulted in a reduction of mortality, length of stay and disability as well as financial resources.

The development of questionnaires to measure outcomes in terms of function and health status has become a major thrust of health research and has provided a mechanism for understanding how functional outcomes relate to specific elements of health care. Generic instruments that assess quality of life, and more specifically health-related quality of life (HRQOL), have provided an overarching perspective for understanding the outcomes of health care in terms of physical, psychological and social function.

Many health status scales have been developed to assess these constructs. Two of the more widely used instruments, the Medical Outcomes Study Short-Form 36 (SF-36) and the Sickness Impact Profile (SIP), have been validated in many languages^{16,17,18,19,20,21,22} and for many different patient populations.^{23,24,25,26,27,28,29,30,31,32,33,34} These scales have also been tested in abbreviated forms that demonstrate an efficiency of validity. These instruments allow for calculation of a summary score or subscale scores that are theoretically related to different dimensions of function and health status. For example, the SF-36 provides eight subscale scores that reflect physical function, physical role limitations, mental function, social function, vitality, general health, bodily pain, and emotional role limitations.³⁵

Those who study HRQOL have debated the usefulness of generic measures over disease-specific (or condition or region-specific) instruments that include items focused on issues relevant to a particular disorder. For example, the Western Ontario and McMaster Osteoarthritis Index (WOMAC) is specific to arthritis.³⁶ The Minnesota Living with Heart Failure Questionnaire targets individuals with heart disease.³⁷ Studies that compare generic and specific tools generally lead to the conclusion that specific tools are more powerful for understanding impairments and function, but both are needed to get a full picture of the individual's quality of life.^38,39,40

Issues of validity for outcome measures remain paramount, as researchers and clinicians must understand the conditions and situations for which these tests are appropriate. The interpretation of outcomes based on these tests must also be made with consideration of the constructs that are being measured. Clinical decisions based on such outcomes must account for the context of the scale used and its measurement properties. Chapter 6 will focus on these issues in greater detail.

A second concept in understanding the evolution of medical research is related to the overriding framework for the delivery of health care. This was historically based on the biomedical model, which focuses on a linear relationship between pathology and resulting impairments. Within this model, health is viewed as the absence of disease and the assumption is made that disease and injury can be treated and cured. The biomedical model confines attention to physical aspects of health, without consideration of how the patient is affected by illness.⁴¹ The primary outcomes of interest under this model are the traditional endpoints of cure, disease or death.^42,43 However, as health care advances and people live longer, practitioners appreciate the inadequacies of the biomedical model for dealing with the common problems of aging, chronic disease and disability, which do not fall within the rubric of "treat and cure," and the consequent need to look differently at the assessment of "successful" interventions.

An expansion of this model has been applied to a broader perspective in rehabilitation. The disablement model (see Figure 1.1A) has provided a framework for assessing the effect of acute and chronic conditions by emphasizing functional consequences and social role. This model demonstrates the relationships among pathology, impairments, functional limitations and disability.^44,45 Although variants of this model have been proposed with different terminology, they have all included the basic elements of pathology, organ system dysfunction, restrictions in activities of daily living (ADL) and limitations of role performance as a member of society.⁴⁶ Accordingly, this model provides a conceptual basis for looking at outcomes within the broader context of health, including psychological and social domains, general health status and quality of life.^42,47,48,49

The International Classification of Functioning, Disability and Health

In 2001 the World Health Organization (WHO) published a revised model of the International Classification of Functioning, Disability and Health (ICF) (see Figure 1.1B).⁵⁰ The ICF is the result of an international and multidisciplinary effort to provide a common language for the classification and consequences of health conditions. Rather than focusing on disability, the intent of the ICF is to describe how people live with their health condition. It is a comprehensive representation of health and health-related domains based on the relationships among health conditions, body functions and structures, activities and participation. The domains are classified from body, individual and societal perspectives. Since an individual's functioning and disability occur in a context, the ICF also includes specific reference to environmental and personal factors that can affect function.⁵¹

The ICF holds a parallel to the disablement model. Health conditions correspond to pathology; body functions/structures correspond to impairments; activity corresponds to functional limitation; and participation corresponds to disability. The ICF may be more useful, however, to understand and measure health outcomes, looking beyond mortality and disease. It shifts the focus to "life," and describes how people live with their health condition. This context makes the model useful for health promotion as well as illness and disability. The ICF has been evaluated in a variety of patient populations, cultures, age groups and diagnoses.^{52,53,54,55,56,57,58}

Each of the components of the ICF can be described in positive or negative terms, as shown in Figure 1.1B.⁵⁹ For example, body systems or anatomical structures will be intact or impaired. Individuals will be able to perform specific activities or they will demonstrate difficulties in executing specific tasks, described as activity limitations. They will either be able to participate in life roles or they may experience participation restrictions. Environmental and personal factors will either facilitate function or create barriers.

The activity and participation domains can be further conceptualized in terms of an individual's capacity to act versus actual performance; that is, what one is able to do versus what one actually does. These elements are defined with reference to the environment in which assessment is taking place. Performance relates to the "current" or actual environment in which the individual participates. Capacity relates to a "standardized" optimum environment, which may be real or assumed. According to the ICF, the gap between capacity and performance reflects the impact of the environment on an individual's ability to perform.⁵⁹ Therefore, an individual with rheumatoid arthritis may have joint deformities (impairments), have difficulty walking (activity limitation) and be unable to perform work activities because of inappropriate height of furniture (participation restriction and environmental barrier). The individual may have the capacity to do the work, but cannot perform the activities in the current environment. However, if the office environment is modified, this individual may be able to perform activities without pain, and therefore, her performance and participation level will improve.

Both the ICF and the disablement model provide a framework for identifying which outcome measures are relevant for specific clients or patients. Health status and functional questionnaires present one avenue for examining outcomes. We must also continue to look at changes in impairments and basic functional activities, such as gait or the performance of a particular functional or occupational task, to provide a complete picture of improvement. Data on psychological and social aspects of health must also be collected, including the impact of the environment. As we become more involved in the documentation of outcomes it is imperative that we understand the measurement properties of the tools we use so they can be applied and interpreted properly (see Chapters 4,5,6).

Stressing the importance of objective documentation in clinical research does not mean that practice can be reduced to a finite science. There is no pure "scientific method" that can account for the influence of experience, intuition and creativity in clinical judgment. Making clinical decisions in the face of uncertainty and variability is part of the "art" of clinical practice. We cannot, however, dissociate the art from the science that supports it. The framework of evidence-based practice (EBP) helps to put this in perspective.

Sackett and colleagues⁶⁰ have provided a popular definition of evidence-based practice as the "conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients." EBP is also described as the "integration of best research evidence with our clinical expertise and our patient's unique values and circumstances."⁶¹ Embedded in this definition is an important concept, that evidence is applied in a process of clinical decision making within the context of a patient or clinical scenario. It emphasizes that research literature provides one, but not the only, source of information for decision making. Perhaps more aptly called evidence-based decision making, this process requires considering all relevant information and then making choices that provide the best opportunity for a successful outcome given the patient-care environment and available resources.

The process of EBP starts by asking a relevant clinical question. The question provides a direction for decision making related to a patient's diagnosis, prognosis or intervention. Questions may also relate to etiology of the patient's problem, the validity of clinical guidelines, safety or cost-effectiveness of care. It should not be surprising that the ability to formulate a good question is essential to finding a relevant answer! It is not a general question, but one that focuses specifically on the characteristics of the patient and issues related to his or her management. The acronym PICO has been used to represent the components of a good clinical question: Patients, Intervention, Comparison, Outcome (see Box 1.1).

The question leads to a search for the best evidence that can contribute to a decision about the patient's care. The terms defined within the PICO format can be used as keywords in a search for literature. The concept of "best evidence" is important, as it refers to the availability of valid and relevant research information. Clinicians must be able to search and access literature (see Chapter 31), critically appraise studies to determine if they meet validity standards, and then determine if and how research results apply to a given clinical situation. A working knowledge of research design and statistics is important for clinicians to use this information wisely. For instance, in describing the Hypothesis-Oriented Algorithm for Clinicians II (HOAC II), Rothstein et al.⁶² have made the assessment of evidence a clearly identifiable part of the decision making process.

This assessment, however, must be made by a clinician who then integrates his or her own clinical judgment and experience with the patient's needs and unique characteristics to make a decision about the patient's care (see Figure 1.2). This decision will also take into account the current circumstances of care, including available equipment, space, time, the patient's comorbidities and the clinical setting. Even Sackett acknowledges that

There are many useful journal articles,* books⁶¹ and websites^{64,65,66,67,68} that describe the concepts related to EBP, including interpretation of statistical outcomes. We will include many of these concepts throughout this text as we discuss research designs, statistical analyses and the use of published material for clinical decision making. The success of evidence-based practice will continue to be assessed as we examine outcomes based on use of published guidelines and treatment effects.

The information that is used to make clinical decisions and to support clinical research can be acquired in many different ways. As one participates in the pursuit of knowledge, it is interesting to reflect on the sources of information that guide our thinking and decision making. How do we decide which test to perform, which intervention will be most successful, which patients have the best chance of responding positively to a given treatment? Oftentimes clinical problems can be solved on the basis of scientific evidence, but in many situations such evidence does not exist or is not directly applicable. It is important, then, to consider how we come to "know" things, and how we can appropriately integrate what we know with available evidence as we are faced with clinical problems (see Figure 1.3).

Tradition

As members of an organized culture, we accept certain truths as givens. Something is thought to be true simply because people have always known it to be true. Within such a belief system, we inherit knowledge and accept precedent, without need for external validation. Rehabilitation science is steeped in tradition as a guide to practice and as a foundation for treatment. We have all been faced with clinical, administrative or educational practices that are continued just because "that is the way they have always been done."

Tradition is useful in that it offers a common foundation for communication and interaction within a society or profession. Therefore, each generation is not responsible for reformulating an understanding of the world through the development of new concepts. Nevertheless, tradition as a source of knowledge poses a serious problem in clinical science because many traditions have not been evaluated for their validity, nor have they been tested against potentially better alternatives. Sole reliance on precedent as a reason for making clinical choices generally stifles the search for new information, and may perpetuate an idea even when contrary evidence is available.

Authority

We frequently find ourselves turning to specialized sources of authority for answers to questions. If we have a problem with finances, we seek the services of an accountant. If we need legal advice for purchasing a home, we hire a real estate lawyer. In the medical profession we regularly pursue the expertise of specialists for specific medical problems. Given the rapid accumulation of knowledge and technical advances and the need to make decisions in situations where we are not expert, it is most reasonable and natural to place our trust in those who are authoritative on an issue by virtue of specialized training or experience.

Authorities often become known as expert sources of information based on their success, experience or reputation. When an authority states that something is true, we accept it. As new techniques are developed, we often jump to use them without demanding evidence of their scientific merit, ignoring potential limitations, even when the underlying theoretical rationale is unclear.^69,70 Too often we find ourselves committed to one approach over others, perhaps based on what we were taught, because the technique is empirically useful. This is a necessary approach in situations where scientific evidence is unavailable; however, we jeopardize our professional responsibility if these techniques are not critically analyzed and if their effects are not scientifically documented.

The danger of uncritical reliance on authoritative canon is well illustrated by the unyielding belief in the medical tenets of Galen (a.d. 138–201), whose teachings were accepted without challenge in the Western world for 16 centuries. When physicians in the 16th and 17th centuries began dissecting human organs, they were not always able to validate Galen's statements. His defenders, in strict loyalty and unwilling to doubt the authority, wrote that if the new findings did not agree with Galen's teachings, the discrepancy should be attributed to the fact that nature had changed!⁷¹

Trial and Error

The trial and error method of data gathering was probably the earliest approach to solving a problem. The individual faced with a problem attempts one solution and evaluates its effects. If the effects are reasonably satisfactory, the solution is generally adopted. If not, another solution is tried. We use this method when we have no other basis for making a decision. We have all used trial and error at one time or another in our personal lives and in professional practice. Trial and error incorporates the use of intuition and creativity in selecting alternatives when one approach does not work.

The major disadvantage of trial and error is its haphazard and unsystematic nature and the fact that knowledge obtained in this way is usually not shared, making it inaccessible to others facing similar problems. In situations where a good response is not obtained, a continuous stream of different solutions may be tried, with no basis for sorting out why they are not working.

Trial and error is by nature extremely time consuming and limiting in scope, for although several possible solutions may be proposed for a single problem, the process generally ends once a "satisfactory" response is obtained. Experience is often based on these solutions, and when similar situations arise, a better solution, as yet untried, may never be tested. Therefore, a clinician using this method should never conclude that the "best" solution has been found.

Logical Reasoning

Many clinical problems are solved through the use of logical thought processes. Logical reasoning as a method of knowing combines personal experience, intellectual faculties, and formal systems of thought. It is a systematic process that has been used throughout history as a way of answering questions and acquiring new knowledge. Two distinctive types of reasoning are used as a means of understanding and organizing phenomena: deductive and inductive reasoning (see Figure 1.4).

Deductive Reasoning

Deductive reasoning is characterized by the acceptance of a general proposition, or premise, and the subsequent inferences that can be drawn in specific cases. The ancient Greek philosophers introduced this systematic method for drawing conclusions by using a series of three interrelated statements, called a syllogism, containing (1) a major premise, (2) a minor premise and (3) a conclusion. A classic syllogism will serve as an example:

1. All living things must die.   [major premise]

2. Man is a living thing.     [minor premise]

3. Therefore, all men must die.  [conclusion]

In deductive reasoning, if the premises are true, then it follows that the conclusion must be true. Scientists use deductive logic by beginning with known scientific principles or generalizations, and deducing specific assertions that are relevant to a specific question. The observed facts will cause the scientist to confirm, to reject or to modify the conclusion. The greater the accuracy of the premise, the greater the accuracy of the conclusion.

For example, we might reason that exercise will be an effective intervention to prevent falls in the elderly in the following way:

1. Impaired postural stability results in falls.

2. Exercise improves postural stability.

3. Therefore, exercise will decrease the risk of falls.

This system of deductive reasoning produces a testable hypothesis: If we develop an exercise program for individuals who have impaired stability, we should see a decrease in the number of falls. This has been the basis for a number of studies. For example, Wolf and colleagues⁷² used this logic as the theoretical premise for their study comparing balance training and tai chi exercise to improve postural stability in a sample of older, inactive adults. Carter and coworkers⁷³ designed an exercise program aimed at modifying risk factors for falls in elderly women with osteoporosis. Similarly, Barnett et al.⁷⁴ studied the effect of participation in a weekly group exercise program over one year on the rate of falling in community dwelling older people. All three studies found that the exercise groups either had a lower incidence of falls or delayed onset of falls, supporting the premise from which the treatment was deduced.

Of course, deductive reasoning does have limitations. Its usefulness is totally dependent on the truth of its premises. In many situations, the theoretical assumptions on which a study is based may be faulty or unsubstantiated, so that the study and its conclusions have questionable validity. In addition, we must recognize that deductive conclusions are only elaborations on previously existing knowledge. Deductive reasoning can organize what is already known and can suggest new relationships, but it cannot be a source of new knowledge. Scientific inquiry cannot be conducted on the basis of deductive reasoning alone because of the difficulty involved in establishing the universal truth of many statements dealing with scientific phenomena.

Inductive Reasoning

Inductive reasoning reflects the reverse type of logic, developing generalizations from specific observations. It begins with experience and results in conclusions or generalizations that are probably true. This approach to knowing was advanced in the late 16th century by Francis Bacon, who called for an end to reliance on authority as absolute truth. He proposed that the discovery of new knowledge required direct observation of nature, without prejudice or preconceived notions.⁷⁵ Facts gathered on a sample of events could lead to inferences about the whole. This reasoning gave birth to the scientific approach to problem solving, and often acts as the basis for common sense. For example, we might observe that those patients who exercise do not fall, and that those who do not exercise fall more often. We might then conclude, through induction, that exercise will improve postural stability.

Inductive reasoning has its limitations as well. The quality of the knowledge derived from inductive reasoning is dependent on the representativeness of the specific observations used as the basis for generalizations. To be absolutely certain of an inductive conclusion, the researcher would have to observe all possible examples of the event. This is feasible only in the rare situations where the set of events in question is very small, and we therefore find ourselves relying mostly on imperfect induction based on incomplete observations. In the preceding example, if we observe the effects of exercise on a sample of elderly persons, and if balance and exercise responses are related to aging, our conclusion may not be a valid one for younger individuals.

Even with these limitations, the process of logical reasoning, both deductive and inductive, is an essential component of scientific inquiry and clinical problem solving. Both forms of reasoning are used to design research studies and interpret research data. Introductory statements in research articles often illustrate deductive logic, as the author explains how a research hypothesis was developed from an existing theory of general body of knowledge. Inductive reasoning is used in the discussion section of a research report, where generalizations or conclusions are proposed from the data obtained in the study. Even though imperfect induction does not allow us to reach infallible conclusions, it is the clinical scientist's responsibility to evaluate critically the validity of the information and to draw reasonable conclusions (see Box 1.2). These conclusions must then be verified through further empirical testing.

The following statement, attributed to Galen, illustrates the potential for the abuse of logic:

The Scientific Method

The scientific method is the most rigorous process for acquiring new knowledge, incorporating elements of deduction and induction in a systematic and controlled analysis of phenomena. The scientific approach to inquiry is based on two assumptions related to the nature of reality. First, we assume that nature is orderly and regular and that events are, to some extent, consistent and predictable. Second, we assume that events or conditions are not random or accidental and, therefore, have one or more causes that can be discovered. These assumptions allow us to direct clinical thinking toward establishing cause-and-effect relationships so that we can develop rational solutions to clinical problems.

The scientific approach has been defined as a systematic, empirical, controlled and critical examination of hypothetical propositions about the associations among natural phenomena.¹ The systematic nature of research implies a sense of order and discipline that will ensure an acceptable level of reliability. It suggests a logical sequence that leads from identification of a problem, through the organized collection and objective analysis of data, to the interpretation of findings. The empirical component of scientific research refers to the necessity for documenting objective data through direct observation. Findings are thereby grounded in reality rather than in personal bias or subjective belief of the researcher.

The element of control, however, is the most important characteristic that sets the scientific method apart from the other sources of knowledge. To understand how one phenomenon relates to another, the scientist practitioner must attempt to control factors that are not directly related to the variables in question. Clinical problems such as pain, functional disability, cognitive dysfunction, deformity, cardiopulmonary insufficiency or motor control concern highly complex phenomena and often involve the effects of many interacting factors. Investigators must be able to control extraneous influences to have critical confidence in research outcomes. This important concept is explored in greater detail in Chapter 9.

A commitment to critical examination means that the researcher must subject findings to empirical testing and to the scrutiny of other scientists. Scientific investigation is thereby characterized by a capacity for self-correction based on objective validation of data from primary sources of information. This minimizes the influence of bias, and makes the researcher responsible for logical and defensible interpretation of outcomes.

Limitations of the Scientific Method

Although scientific research is considered the highest form of acquiring knowledge, it is by no means perfect, especially when it is applied to the study of human behavior and performance. The complexity and variability within nature and the environment and the unique psychosocial and physiological capacities of individuals will always introduce some uncertainty into the interpretation and generalization of data. These issues differentiate clinical research from laboratory research in physical and biological sciences, where environment and even heredity are often under complete control. This does not mean that the scientific method cannot be applied to human studies, but it does mean that clinical researchers must be acutely aware of extraneous influences to interpret findings in a meaningful way. Some clinical findings may actually be strengthened by the knowledge that patients generally improve with certain treatments despite physiological and environmental differences.

The research process delineates a general strategy for gathering, analyzing, and interpreting data to answer a question. A variety of schema have been used to classify research strategies according to their purpose and objectives.

Quantitative and Qualitative Research

In categorizing clinical research, researchers often describe studies by distinguishing between quantitative and qualitative methods. Quantitative methods may be used all along the continuum of research approaches, whereas qualitative data are generally applied to descriptive or exploratory research. Quantitative research involves measurement of outcomes using numerical data under standardized conditions. The advantage of the quantitative approach is the ability to summarize scales and to subject data to statistical analysis. Quantitative information may be obtained using formal instruments which address physical or physiological parameters, or by putting subjective information into an objective numerical scale.

Qualitative research is more concerned with a deep understanding of a phenomenon through narrative description, which typically is obtained under less structured conditions. In qualitative methodology, "measurement" is based on open-ended questions, interviews and observations, as the researcher attempts to capture the context of the data, to better understand how phenomena are experienced by individuals. The purpose of the research may be to simply describe the state of conditions, or it may be to explore associations, formulate theory, or generate hypotheses.

Basic and Applied Research

One system of classification is based on the objective of the research, or the degree of utility of the findings. Basic research is done to obtain empirical data that can be used to develop, refine, or test theory. Basic research is directed toward the acquisition of new knowledge for its own sake, motivated by intellectual curiosity, without reference to the potential practical use of results. Typically done in a laboratory, basic research is often called "bench research." Researchers who study how blood cells function or who examine the structure and function of parts of the brain are doing basic research. Of course, basic studies may eventually lead to numerous practical applications, such as developing a treatment for leukemia or grafting brain cells to treat Parkinson's disease. But these are not the direct goals of the basic scientist.

In contrast, applied research is directed toward solving immediate practical problems with functional applications and testing the theories that direct practice. It is usually carried out under actual practice conditions on subjects who represent the group to which the results will be applied. Most clinical research falls into this category. When therapists study the effect of electrical stimulation for reducing muscle spasm or compare the effectiveness of eccentric and concentric exercises for increasing strength, they are doing applied research.

Although the distinction between basic and applied research appears to create a dichotomy, in reality a continuum exists between the two extremes. We recognize that rehabilitation and health care are applied sciences, but that many of the theories that guide practice are founded on basic science principles. Today, clinical research is often a hybrid, combining elements of both basic and applied science. Many studies provide clinical application as well as new knowledge that contributes to a theoretical understanding of behavior.

Translational Research

The term translational research refers to the application of basic scientific findings to clinically relevant issues, and simultaneously, the generation of scientific questions based on clinical dilemmas.⁷⁶ It is often described as taking knowledge from "bench to bedside," or more practically from "bedside to bench and back to bedside."⁷⁷ Although certainly not a new concept, the medical community has experienced a renewed emphasis on the application of laboratory-based findings to clinically important problems. The NIH Roadmap, proposed in 2002, has called for a new paradigm of research to assure that "basic research discoveries are quickly transformed into drugs, treatments, or methods for prevention."⁷⁸ Questions related to understanding the mechanisms of disease or therapies, molecular changes or different responses of normal or abnormal tissues are examples of how fundamental work at the bench can eventually benefit patients directly.⁷⁹

All too often, the successes of scientific breakthroughs in the laboratory or in animal models have not translated into major changes in medical care for humans. The success of the Human Genome Project is an example of where important scientific discoveries have not yet realized their full potential.⁸⁰ Although markers for specific genetic defects can be identified, these do not exist in isolation from other physical and physiological conditions, and so the complexity of the human organism creates a challenge to apply these discoveries to patient outcomes. Other examples of promising translational research include the study of regeneration in spinal cord injury,⁸¹ and new interventions to optimize treatment of diabetes.⁸² Research aimed at developing therapies for inhibiting angiogenesis in tumors has also sparked questions related to their use in nononcological diseases, such as rheumatoid arthritis, psoriasis and diabetic retinopathy.⁸³ As these examples illustrate, the success of translational research will lie in the close collaboration among laboratory researchers who understand the basic science, clinicians who understand human behavior and response to disease, and patient communities.^78,84 It will include reflection on challenging clinical problems, rigorous investigation with basic science techniques, insights into clinical innovations, and consideration of new directions for future research.⁷⁷

Experimental and Nonexperimental Research

Another common classification defines research as either experimental or nonexperimental. Experimental research refers to investigations where the researcher manipulates and controls one or more variables and observes the resultant variation in other variables. The major purpose of an experiment is to compare conditions or intervention groups, to suggest cause-and-effect relationships. Nonexperimental research refers to investigations that are generally more descriptive or exploratory in nature and that do not exhibit direct control over the studied variables. This latter type of research is often referred to as observational research, to reflect the idea that phenomena are observed rather than manipulated.

A Continuum of Research

In a more practical scheme, research can be viewed along a continuum that reflects the type of question the research is intended to answer. Within this continuum, illustrated in Figure 1.5, research methods can be classified as descriptive, exploratory, or experimental. These classifications reflect different purposes of research, and within each one various types of research can be used. As a continuum suggests, however, different types of research can overlap in their purpose and may incorporate elements of more than one classification.

While many view this continuum as a hierarchy, with experimental designs at the top (suggesting a relative value for these research approaches), each type of research fulfills a particular purpose and need. Each brings specific strengths to an investigation of clinical phenomena. The appropriate use of various designs will depend on the research question and the available data, with questions related to intervention, diagnosis and prognosis requiring different approaches.

Experimental Research

Experimental designs provide a basis for comparing two or more conditions for the purpose of determining cause and effect relationships. They control or account for the effects of extraneous factors, providing the greatest degree of confidence in the validity of outcomes, and allowing the researcher to draw meaningful conclusions about observed differences. The randomized controlled trial (RCT) is considered the "gold standard" of experimental designs, typically involving the controlled comparison of an experimental intervention and a placebo. There are, however, many alternative models, some simple and others more complex, that provide opportunities to examine the cause of outcomes, including the systematic study of one or several individuals using single-subject designs within the clinical environment.

In quasi-experimental studies the degree of control is limited by a variety of factors, but interpretable results can still be obtained. When true experimental conditions cannot be achieved, these designs permit comparisons, but they also acknowledge the limitations placed on conclusions.

Exploratory Research

In exploratory research a researcher examines a phenomenon of interest and explores its dimensions, including how it relates to other factors. In epidemiology health researchers examine associations to describe and predict risks for certain conditions using cohort and case-control studies. Using correlational methods, the researcher is able to search for these relationships and may generate predictions that these relationships suggest. Predictive models can then be used as a basis for decision making, setting expectations and prognosis. By establishing associations, researchers can also test or model theoretical propositions. Many efforts in outcomes research use this approach to study relationships among pathologies, impairments, functional limitations and disability.

Methodological studies will use correlational methods to demonstrate reliability and validity of measuring instruments. Historical research reconstructs the past, on the basis of archives or other records, to generate questions or suggest relationships of historical interest to a discipline.

Descriptive Research

In descriptive research the researcher attempts to describe a group of individuals on a set of variables, to document their characteristics. Descriptive research may involve the use of questionnaires, interviews or direct observation. Descriptive data allow researchers to classify and understand the scope of clinical phenomena, often providing the basis for further investigation. Several designs can be used within this approach.

Developmental research is intended to investigate patterns of growth and change over time within selected segments of a population, or it may chronicle the natural history of a disease or disability. Normative studies focus on establishing normal values for specific variables, to serve as guidelines for diagnosis and treatment planning. Qualitative research involves collection of data through interview and observation, in an effort to characterize human experience as it occurs naturally, and to generate hypotheses about human behavior. A case study or case series may consist of a description of one or several patients, to document unusual conditions or the effect of innovative interventions.

Sources of Data

In designing research studies, investigators will describe the methods used for collecting data. Most research involves direct data collection based on the performance of subjects, according to the investigator's defined protocol. Surveys or questionnaires are often used to collect data on subject characteristics or opinions, as part of descriptive, exploratory or experimental studies. As large databases begin to develop, researchers often use secondary analysis as a mechanism for exploring relationships. This approach typically involves the use of data that were collected for another purpose, or it may be based on data from ongoing surveys.

Synthesis of Literature

As bodies of evidence continue to grow through publication of research, clinicians face the challenge of aggregating information to adequately answer a clinical question. Systematic reviews present a comprehensive analysis of the full range of literature on a particular topic, typically an intervention, diagnostic test or prognostic factors. Meta-analysis is a process of statistically combining the findings from several studies to obtain a summary analysis. These forms of review, when done well, provide the clinician with a critical analysis of current research that can be used for clinical decision making. They also allow the clinician to recognize the scope of research and knowledge in a particular content area, and to appreciate a balance in the interpretation of information.

Clinical research involves a systematic process of sequential steps that guide thinking, planning and analysis. Whether one is collecting quantitative or qualitative data, the research process assures that there is a reasonable and logical framework for a study's design and conclusions. We conceptualize research as a series of nine sequential steps shown in Figure 1.6, recognizing that the order may vary and the steps may overlap in different research models. These steps can be grouped into five major categories.

Step 1: Identify the Research Question

The first step of the research process involves delimiting the area of research and formulating a specific research question that provides an opportunity for scientific testing (see Chapter 7). During this stage, the researcher must define the type of individual to whom the results will be generalized. Through a review of scientific literature, the researcher should be able to provide a rationale for the study, a justification of the need to investigate the problem, and a theoretical framework for interpreting results. Research hypotheses are proposed to predict how response variables and treatment variables will be related and to predict clinically relevant outcomes. In descriptive or qualitative studies, guiding questions may be proposed that form the framework for the study.

Step 2: Design the Study

In step 2, the researcher designs the study and plans methods of subject selection, testing, and measurement so that all procedures are clearly mapped out (see Chapters 5,6,7,8,9,10,11,12,13,14,15,16). The choice of research method reflects how the researcher conceptualizes the research question. Many alternative approaches are available, depending on the nature of the data and the type of subjects. The researcher must carefully define all measurements and interventions so that the methods for data analysis are clear. The completion of the first two steps of planning results in the formulation of a research proposal (see Chapter 32).

Step 3. Methods

During the third step of the research process, the researcher implements the plans designed in steps 1 and 2. Data collection is typically the most time consuming part of the research process. After data are collected and recorded, the researcher must reduce and collate the information into a useful form for analysis. Forms or tables are created for compiling the "raw data." Just as much attention to precision must be given during data reduction as during data collection (see Chapter 30).

Step 4: Data Analysis

The fourth step of the research process involves analyzing, interpreting, and drawing valid conclusions about the obtained data. It is the pulling together of all the materials relevant to the study, to apply them to a generalized or theoretical framework. Statistical procedures are applied to summarize quantitative information in a meaningful way, usually with the assistance of a computer (see Chapters 17,18,19,20,21,22,23,24,25,26,27,28,29). It is at this stage that the research hypothesis will be either supported or rejected. In qualitative studies, the researcher will look for themes that characterize the data. Through the analysis of results, the study should also lead to new questions that will stimulate further study.

Step 5: Communication

Research done in a vacuum is of little use to anyone. Researchers have a responsibility to share their findings with the appropriate audience so that others can apply the information either to clinical practice or to further research. Research reports can take many forms including journal articles, abstracts, oral presentations, and poster presentations. Students may be required to report their work in the lengthier form of a thesis or dissertation (see Chapter 33).

Finally, no research project is a dead end. Results of one study always lead to new questions. Researchers contribute to the advancement of their own work by offering suggestions for further study and recommending what kinds of additional studies would be useful for contributing to the theoretical foundations addressed in the current study.

The focus of a text such as this one is naturally on the methods and procedures of conducting research, on the mechanisms of how research is done: how phenomena are observed and measured; how different types of research fit varied research questions; how to design conditions so that relationships can be examined; and how to control and manipulate variables to demonstrate cause-and-effect relationships. By understanding the processes, definitions and analytic procedures of research, the clinician has the building blocks to structure an investigation or interpret the work of others.

Methodology is only part of research, however. Research designs and statistical techniques cannot lead us to a research question, nor can they specify the technical procedures needed for studying that question. Designs cannot tell us what to investigate, nor do they assign meaning to the way clinical phenomena behave. Two other aspects are equally important to the concept of research: knowledge of the subject matter that will be studied and the research philosophy of the clinical discipline.

A thorough knowledge of content related to a research question is necessary to determine relevant applications of the methods for answering the question. The researcher must be able both to determine which instruments are appropriate for measuring different variables and to apply measurement tools properly. The scientific bases for observed responses must be thoroughly understood to design a study and interpret results. Without a complete background in the relevant content, the researcher may make serious errors in data collection and analysis.

The philosophy of a discipline concerns itself with the way the subject matter is conceptualized, the overall significance of the knowledge generated by research, and what scientific approaches will contribute to an understanding of practice. How one conceives of a discipline's objectives and the scope of practice will influence the kinds of questions one will ask. We must recognize the influence of professional values on these applications. These values reflect the researcher's inclinations to consider treatment alternatives, to search for new knowledge to substantiate certain types of clinical decisions, or to investigate particular types of questions with particular methods. For instance, different paradigms will direct some clinical investigators to study behavior at the level of impairments versus outcomes, or to use qualitative versus quantitative methods.

There is no right or wrong in these contrasts. As we explore the variety of research approaches and the context of evidence-based practice, we urge the reader to continually apply his or her own framework for applying these methods. Our emphasis on clinical examples throughout the book is a limited attempt to demonstrate these connections. It is also relevant to consider the interdisciplinary clinical associations inherent in health care, the team approach to patient care, and the shared research agendas that might emerge from such associations. The framework that supports a research question will likely be broader than any one discipline's objectives and might be well served by a team of professionals.