![© 2017 American Health Information Management Association
Experimental Research
• Experimental research is conducted to establish cause-and-effect
(causal) relationships
– Strict procedures, random assignment of subjects to groups,
manipulation of subjects’ experience, and measurement of
resulting physical, behavioral, or other changes
– Strictly controlled situations and environments
– Four features
• Randomization (random sampling, experimental [study] group [arm])
• Observation before and after treatment (dependent variable)
• Presence of control group(s) (arm)
• Treatment manipulating independent variable
– Test hypotheses following protocols
– Example: RCT evaluating effectiveness of HIT in medication
safety](https://image.slidesharecdn.com/ab120916ch01-210127140703/85/HM404-Ab120916-ch01-26-320.jpg)
HM404 Ab120916 ch01
- 1. © 2017 American Health Information Management Association © 2017 American Health Information Management Association Health Informatics Research Methods: Principles and Practice, Second Edition Chapter 1: Research Frame and Designs
- 2. © 2017 American Health Information Management Association Learning Objectives • Use and explain the terms research, research frame, theory, model, and research methodology. • Designate the appropriate placement of a research project on the continuum of research from basic to applied. • Differentiate among research designs. • Provide appropriate rationales that support the selection of a research design. • Use key terms associated with research frames and designs appropriately.
- 3. © 2017 American Health Information Management Association Definition of Research • Research is a systematic process of inquiry aimed at discovering or creating new knowledge about a topic, confirming or evaluating existing knowledge, or revising outdated knowledge
- 4. © 2017 American Health Information Management Association What Are Health Informatics and HIM Research? • Health informatics research is the investigation of the process, application, and impact of computer science, information systems, and communication technologies to health services – Example: Identification of commonly recurring safety issues related to EHRs • HIM research is investigation into the practice of acquiring, analyzing, storing, disclosing, retaining, and protecting information vital to the delivery, provision, and management of health services – Example: Meanings of the term “medical record” in research forms
- 5. © 2017 American Health Information Management Association Purposes of Health Informatics and HIM Research • Formulate theories and principles of health informatics and HIM • Test existing theories, models, and assumptions about the principles of health informatics and HIM • Build a set of theories about what works, when, how, and for whom • Advance practice by contributing evidence that decision makers can use • Train future practitioners and researchers • Develop tools and methods for the process of health informatics research and HIM research
- 6. © 2017 American Health Information Management Association Research Frame • Research frame: Overarching structure of a research project, including – Theory or theories underpinning the study – Models illustrating the factors and relationships of the study – Assumptions of the field and the researcher – Methods – Analytical tools
- 7. © 2017 American Health Information Management Association Theories and Models • Theory: Systematic organization of knowledge that explains or predicts phenomena – Interrelating concepts in a logical, testable way – Providing definitions, relationships, and boundaries – Parsimonious • Model: Idealized representation that abstracts and simplifies a real-world situation so the situation can be studied, analyzed, or both – Visual depiction of theory – Portrayal with objects, smaller-scaled version, or graphic representation
- 8. © 2017 American Health Information Management Association Examples of Theories and Models Used in Health Informatics and HIM Research: • Theories – Adult learning theories – Change theories – Diffusion of innovations – General systems theory – Information behavior theories – Information processing or cognitive learning theories – System of systems theory – User acceptance theories – And others (table 1.1) • Models – AHIMA Data Quality Management Model – Dominant design (A-U model) – Information systems success model – Sociotechnical model – Swiss cheese model – Systems development life cycle model – Technology acceptance model – And others (table 1.1)
- 9. © 2017 American Health Information Management Association Research Methodology and Research Method • Research methodology: Study and analysis of research methods and theories – Example: Which method of data collection results in the greatest response rate? • Research method: Set of specific procedures used to gather and analyze data – Example: Surveys (chapter 2) and focus groups (chapter 3)
- 10. © 2017 American Health Information Management Association Continuum of Basic and Applied Research • Basic research – Answer question “Why?” – Focuses on development of theories and their refinement – “Bench science” • Applied research – Answers questions “What?”, “How?”, “When?”, or “For whom?” – Focuses on implementation of theories and models into practice – Most health informatics and HIM researchers are applied researchers
- 11. © 2017 American Health Information Management Association Approaches:Quantitative, Qualitative, and Mixed Methods Quantitative • Explanation of phenomena by making predictions, collecting and analyzing evidence, testing alternative theories, and choosing the best theory • Numeric data • Generalizability and positivism • Qualitative • Investigation to describe, interpret, and understand processes, events, and relationships as perceived by individuals or groups • Nonnumeric data • Context Mixed methods • Combination of quantitative and qualitative theoretical perspectives, methods, sampling strategies, data collection techniques, data sets, analytical procedures, representational modes, or any combination of these aspects • Suited to studies of healthcare
- 12. © 2017 American Health Information Management Association Scientific Inquiry • Scientific inquiry: Way of generating knowledge by systematically gathering data about phenomena, critically analyzing the data, proposing explanations based on evidence, and developing understanding and knowledge – Empiricism – Inductive reasoning (induction): From specific to the general; drawing conclusions based on a limited number of observations (“bottom up”) – Deductive reasoning (deduction): From general to the specific; drawing conclusions based on generalizations, rules, or principles (“top down”)
- 13. © 2017 American Health Information Management Association Rigor • Rigor establishes the validity and reliability of a research study’s results and conclusions – Quantitative definition: Strict application of the scientific method to ensure unbiased and well- controlled experimental design, methodology, analysis, interpretation and report of results and includes transparency in reporting full experimental details so that others may reproduce and extend the findings – Qualitative definition: Trustworthiness of the interpretation of the study’s findings
- 14. © 2017 American Health Information Management Association Research Designs • Research design: Plan to achieve the researchers’ purpose— answering a question, solving a problem, or generating new information – Infrastructure of the study – Seven common research designs • Historical research • Descriptive research • Correlational research • Observational research • Evaluation research • Experimental research • Quasi-experimental research – Selecting the appropriate research design increases the likelihood that the evidence collected are relevant, high quality, and directly related to the research question or problem
- 15. © 2017 American Health Information Management Association Historical Research • Historical research: Examines historical materials to explain, interpret, and provide a factual account of events – Purposes • Discovering new knowledge • Identifying trends that could provide insights into current questions or problems • Relating the past to contemporary events or conditions • Creating official records – Process of historical research is to systematically collect, critically evaluate, and analyze and interpret evidence from historical materials, known as primary and secondary sources – Example: US National Library of Medicine’s Conversations with Medical Informatics Pioneers
- 16. © 2017 American Health Information Management Association Primary and Secondary Sources • Primary sources – First-hand sources also called primary data – Examples • Original documents • Artifacts • Oral histories – Created or collected for a specific purpose – Preferable to secondary sources • Secondary sources – Second-hand sources also called secondary data – Created by people uninvolved with the event – Aggregate, summarize, critique, analyze, or manipulate the primary sources and, thus, are derived from primary sources
- 17. © 2017 American Health Information Management Association Descriptive Research • Descriptive research determines and reports the current status of topics and subjects – Seeks to accurately capture or portray dimensions or characteristics of people, organizations, situations, technology, or other phenomena – Should be chosen to answer questions, such as “what is,” “what was,” or “how much” – Is best way to collect information that will demonstrate relationships and describe the world as it exists – Some descriptive research studies are also correlational— detecting relationships – Example: Study of consumers’ perspectives on the healthcare environment and use of services
- 18. © 2017 American Health Information Management Association Correlational Research • Correlational research detects the existence, direction, and strength (or degree) of associations among characteristics – Characteristics can be phenomena, factors, attitudes, organizational features, properties, traits, indicators, performance measures, or any other attribute of interest – Quantitative, exploratory, and indicative of existing associations – Can be either descriptive or predictive – Predictive when predicting nonrandom change in one characteristic (or characteristics) based on change in another characteristic (or characteristics) – Example: Patients’ rates of social media usage and their ratings of providers
- 19. © 2017 American Health Information Management Association Correlational Research (cont.) • Strength of association – Strength of 0.00 means absolutely no association – Strength between 0.00 and +1.00 or between 0.00 and –1.00 means that the variables sometimes, but not always, move together – Strength of 1.00 or –1.00 means a perfect association, with the variables moving exactly in tandem • Cannot establish causal relationship – Unknown variable could be creating apparent association – Confounding variable also called extraneous or secondary variable • Shortcomings of self- report – Subjects’ biases – Selective memory – Social desirability
- 20. © 2017 American Health Information Management Association Linear Relationships • Positive (direct) linear relationship – Exists when the scores for variables proportionately move in the same direction – As one increases, so does the other; or as one decreases, so does the other • Negative (inverse) linear relationship – Exists when the scores for variables proportionately move in opposite (inverse) directions – As one increases, the other decreases; or as one decreases, the other increases
- 21. © 2017 American Health Information Management Association Examples of Relationships Positive (direct) linear relationship Negative (inverse) linear relationship Curvilinear relationship (s-curve)
- 22. © 2017 American Health Information Management Association Observational Research • Observational research is exploratory research that identifies factors, contexts, and experiences through observations in natural settings – Focus is participants’ perspective of their own feelings, behaviors, and perceptions – Provides insights into what subjects do, how they do it, and why they do it through rich data, artifacts, and triangulation – Nonparticipant observation, participant observation, and ethnography – Example: Case study of how well portals convey information to patients
- 23. © 2017 American Health Information Management Association Observational Research (cont.)
- 24. © 2017 American Health Information Management Association Evaluation Research • Evaluation research is the systematic application of criteria to assess the value of objects – Can use any of the other research designs; it is the purpose—evaluation—that classifies the design as evaluation – Objects evaluated in terms of merit, worth, quality, or combination of these attributes – Evaluated objects include policies, programs, technologies (including procedures or implementations), products, processes, events, conditions, organizations, and others – Evaluation criteria include conceptualization, design, components, implementation, usability, effectiveness, efficiency, impact, scalability, and generalizability
- 25. © 2017 American Health Information Management Association Evaluation Research (cont.) • Terms for evaluation research – Outcomes research – Health services research – Health technology assessment – Comparative effectiveness research – Usability testing • Types of evaluation research – Needs assessment – Process evaluation – Outcome evaluation – Policy analysis Example: Study that evaluated the impact of the Regional Extension Center program
- 26. © 2017 American Health Information Management Association Experimental Research • Experimental research is conducted to establish cause-and-effect (causal) relationships – Strict procedures, random assignment of subjects to groups, manipulation of subjects’ experience, and measurement of resulting physical, behavioral, or other changes – Strictly controlled situations and environments – Four features • Randomization (random sampling, experimental [study] group [arm]) • Observation before and after treatment (dependent variable) • Presence of control group(s) (arm) • Treatment manipulating independent variable – Test hypotheses following protocols – Example: RCT evaluating effectiveness of HIT in medication safety
- 27. © 2017 American Health Information Management Association Experimental Research (cont.) • Control environments and subjects – Control: Processes used to maintain uniform conditions in order to eliminate sources of bias, variations, and any extraneous factors that might affect research’s outcome – Allows researchers to conclude that independent variable caused change in dependent variable without any other possible explanations related to bias, variation, or unknown factors
- 28. © 2017 American Health Information Management Association Experimental Research (cont.)
- 29. © 2017 American Health Information Management Association Quasi-experimental Research • Quasi-experimental research searches for plausible causal factors or indicates that a causal relationship could exist – Approximate environment of true experiments – Investigations of possible cause-and-effect relationships – Often, randomization, is absent – Quasi implies design is “similar to” or “almost” experimental – Also called causal-comparative research or ex post facto – Example: Study assessing effectiveness of decision support systems for various diseases
- 30. © 2017 American Health Information Management Association Quasi-experimental Research (cont.) • Quasi-experimental research design appropriate situations prevent random assignment, such as when variables – Cannot be manipulated (gender, age, race, birth place) – Should not be manipulated (accidental death or injury, child abuse) – Represent differing conditions that have already occurred (medication error, heart catheterization performed, smoking) • Lack of randomization creates potential for bias • Lack of control creates potential for introduction of confounding variable
- 31. © 2017 American Health Information Management Association Time Frame as an Element of Research Design • Retrospective – Looks back in time • Prospective – Follows subjects into the future • Cross-sectional – One point in time – Snapshot – May be unrepresentative time • Longitudinal – Three or more waves – Days, weeks, months, years, or lifetimes – Duration varies dependent upon topic
- 32. © 2017 American Health Information Management Association Review • Research is systematic process of inquiry • Research frames are overarching structures including theories, models, assumptions, methods, and analytical tools • Theories explain or predict phenomena and provide definitions, relationships, and boundaries • Models are idealized representations that abstract and simplify situations • Basic and applied research are on a continuum • Scientific inquiry involves inductive and deductive reasoning • Selection of a research design depends upon the researcher’s purpose • Common research designs are historical, descriptive, correlational, observational, evaluation, experimental, and quasi-experimental • Time frame is an element of research design