Research Methods I Midterm

Dictionary definition for research

scholarly or scientific investigation or inquiry

Long definition of research

multiple, systematic strategies to generate knowledge about human behavior, human experience, and human environments in which the thought and action processes of the researcher are clearly specified so that they are logical, understandable, confirmable, and useful

NIH 3-part definition

-patient-oriented

-epidemiologic

-outcomes research

patient-oriented research

Studies conducted with human subjects to improve the understanding of diseases and disorders

epidemiologic

observational studies focused on describing patterns of disease and disability and identifying preventative factors

outcomes research

Studies to determine the impact of research on population health and utilization of EBI

clinical research is essential to

inform clinical judgments, as well as the organization and economics of practice

fundamental goal of research

To develop scientific knowledge of a profession, the end product is data connecting intervention with outcomes, demonstrating the effectiveness and efficiency of our practice

How is research linked to theory

Research yields findings that lead to a theory and that examine and refine existing theory

Where is research needed?

-develop tests and assessments

-examine clinical decision making

-connect therapy to outcomes

-improve the process of therapy

-explore novel service delivery models

Barriers to research

Material

-money

-materials

-people

Organizational

-environment

-time

-support

Attitudinal

-mindset

-beliefs

-opinion

-history

-flexibility

Overcoming barriers to research

-requires a new mindset

-not like learning a new technique

-is a natural function of practice

common denominator in the process is the problem-solving approach

Basic research

bench or preclinical research, directed toward the acquisition of new knowledge; may or may not have a direct clinical application (animal)

Applied research

clinical research to advance the development of new diagnoses, drugs, therapies, and prevention strategies (typically where OT focuses)

Explanatory research

uses experimental designs to compare two or more conditions or interventions; assess efficacy and effectiveness (RCTs, PCTs, quasi-experimental, single-subject designs)

Exploratory research

examines a phenomenon of interest in populations or communities; can be used to establish associations. Assesses validity and reliability (cohort studies, case-control studies, correlational studies)

Descriptive research

attempts to describe a group of individuals on a set of varibales ot document their characteristics. Allows researchers to classify and understand the scope of the clinical or social phenomena (developmental, case reports, qualitative research)

Translational research

application of basic scientific findings to clinically relevant issues and simultaneously the generation of scientific questions based on clinical dilemmas, 'bench to bedside'

Lag time

10-25 years for a discovery to reach publication, the average is 17 years

Translation continuum

T0: basic science research (animal)

T1: Does it work- translation to humans (new methods of diagnosis, treatment, and prevention)

T2: Can it work under ideal conditions- translation to patients (controlled studies leading to effective care)

T3: Will it work in real-world conditions- translation to practice (delivery of recommended and timely care to the right patient)

T4: Is it worth it- translation to community (true benefit to society)

Efficacy

performance of an intervention under ideal and controlled conditions; randomized controlled trials

Effectiveness

performance of an intervention under natural, real-world conditions; pragmatic clinical trials

Outcomes research

An umbrella term to describe studies that focus on the impact of the results of health care practices and interventions

-address both clinical and population-based questions with the intent of improving the benefit of health care to individuals and communities

Implementation studies

-focuses on understanding the influence of the environment and resources on whether research findings are actually translated into practice

-ways to change behavior through education, training, team-based efforts, community engagement, or systematic infrastructure redesign

-address: clinical performance audits, use of patient or provider alerts to remind them of guidelines, patient education, and interventions

Steps of the research process

1. Identify the research question

2. Design the study

3. Implement the study

4. Analyze the data

5. Disseminate the findings

6. Close the loop

Research questions

Problem: Utilize professional literature to find gaps and conflicts in the literature

Question: importance and feasibility, population

Rationale: to guide research design and interpretation

-identify variables of interest

-conduct a review of the literature

-develop a hypothesis

What makes a good research question?

-Important: What is the potential impact on treatments

-ethical: conforms to ethical standards in terms of protection of human subjects and researcher integrity

-feasible: should be practical to execute for a successful project

What is measurement

-Investigators collect data from a sample to answer a research question

-the process of assigning numerals to observed people, objects, or events according to a set of rules for the purpose of understanding them objectively without ambiguity

-an approach to detecting and documenting relative conditions or events

-basis for making decisions or drawing conclusions

Why do we measure?

-make decisions based on criteria or standard performance

-draw comparisons to choose between courses of action

-evaluate responses to assess a patient's condition by documenting change or progress

-discriminate among individuals who present different profiles related to their condition and characteristics

-predict outcomes to draw conclusions about relationships and consider expected responses

Benefits of measurement

-increase understanding about a given variable via the expression of qualitative/ quantitative information

-summarize existing evidence, identify gaps, and map relevant theories

-ensure a systematic approach to source selection and synthesis

quantification

The first part of the measurement is to assign a number to the variables

variable

characteristic being measured that varies among the persons, objects, or events being studied

-can denote quantity or an attribute

methodology selection

-allow us to distinguish between groups or conditions as well as predict future outcomes

-Measurement is the quantification of an observation using a standard, whereas assessment includes the interpretation of the obtained measure

direct measurement

directly observed (height, ROM, etc.)

indirect measurement

constructs assorted traits chosen to represent an otherwise abstract phenomenon (intelligence, health, QoL, pain, fatigue, etc.)

-must be operationally defined in research

dichotomous variable

when a variable continues only two discrete levels (true/false, gender)

discrete variable

limited to whole units (numbers, nominal or ordinal)

continuous variable

can be any value along a continuum within a specified range (fractional of whole numbers, interval, or ratio), height (cm), time (sec)

nominal scales

-categories of objects or subjects are assigned different labels, the lowest scale or measurement

-cannot be converted to any other scales, but other scales can be converted to nominal

-ex. gender, race, type of occupation, political affiliation, hand dominance, favorite type of music

ordinal scales

ranked categories based on the relative magnitude of the characteristic

-differences are not consistent or quantifiable

-military rank, low-medium-high, manual muscle test, am-pac, berg, barthel, etc.

-use nonparametric statistics

interval scale

-values with known and equal distances, an arbitrarily assigned zero point (zero doesn't truly mean 0)

-illustrates numerical differences between scores, but not absolute magnitude

-can add or subtract means but cannot compute or compare ratios

-second-highest measurement scale

-ex. temperature, years on calendar, sea level, goni, elevation

ratio scale

-interval scale with absolute zero point (indicates complete absence)

-negative numbers are possible

-highest measurement scale

-can perform all mathematical and statistical processes

-ex. height, weight, percentiles, t scores, quotients, grades, blood pressure

scales of measurement rules

can go down in level but not up (ordinal to interval), researchers should use the highest level of measure available

non-parametric measures

nominal: frequency only mode

ordinal: frequency, median, and range

-averages don't have true mathematical significance

parametric measure

interval: addition and subtraction mean and variance

Ratio: all mathematical and statistical procedures are allowed

scale

An ordered system based on a series of questions or items, resulting in a score that represents the degree to which a respondent possesses a particular attribute, value, or characteristic

-ordered system based on a series of questions that provides an overal rating (attitudes, function, health, QoL, pain, exertion

literature

all scholarly products including original research articles, editorials, position papers, reviews and meta-analysis, books and dissertations, conference proceedings and abstracts, and website materials

-helps in formulating a research question and planning the study

literature search

systematic and organized from the already published data to identify a breadth of good-quality references on a specific topic

retrieving available information relevant to a topic is

Vital to both conducting research and being a consumer of research, and is a key step in performing good, authentic research

scientific journals

publications that contains original reports of scientific research and evaluated for scientific quality and correctness, usually specialized by discipline: audience - other researchers

-ex. AJOT, OTJR

-abstract, specific topic, peer-reviewed, academic journal

magazines

publications that include opinion piences, news, information on policy, and reports on a popular interest; readership- general public or professionals from a particular field

-ex. OT practice

-laymens terms, volume number, no abstract

government and professional websites

includes governmental agencies, nongovernmental organizations,

-ex. NIH, CDC, Red Cross, Doctors Without Borders

-CDC, not new research, sharing info, summarizing

Grey literature

anything not produced by a commercial publisher, including government documents, reports, fact sheets, practice guidelines, conference proceedings, and theses or dissertations

-poster session, no detailed methods

primary source

reported directly by the investigator

-original article, abstract

secondary source

a description or review of one or more studies presented by someone other than the original authors

-collecting information, health students, textbooks

Systematic reviews, meta-analyses, and clinical guidelines include a

critical analysis of published works that generate new knowledge

-secondary source, new knowledge, multiple sources reviewed

databases

an index of citations that is searchable by keywords, author, title, or journal

-most common databases in healthcare: Medline, CINAHL, and Cochrane for Systematic Reviews

PubMed

most used health science application, free, comprehensive, and readily available resource for biomedicine and health fields

MEDLINE

indexes more than 5,200 journal titles and over 24 million references in 40 languages

CINAHL

indexes references from nursing, biomedicine, health sciences, alternative/complementary medicine, consumer health, and 17 other disciplines

Cochrane Database of Systematic Reviews

contains citations of systematic reviews involving 37,000 contributors from 130 countries

Search engines

platforms that access one or more databases

-PubMed, OVID, EBSCO, Google Scholar

Successful searchers require

several iterations, revising keywords startefies, and using different databases

The search process can be guided by 4 overall questions

1. What am I looking for? (develop a PICO question)

2. Where can I find it? (making good predictions of where relevant information will be and using retrieved information to further define and articulate the search concept)

3. How do I access and retrieve it? (identify key words, search terms, generate search query, apply search syntax)

4. How well does it satisfy my information requirements, and what is the reliability of the information retrieved? (critical evaluation of information retrieved)

Too few articles

may need to remove some specificity from your topic

Too many citations

may need to be more specific in describing the population, the intervention, or the outcomes

limits and filters

can be used to refine a search

-ex. age groups, dates of publication

Sensitivity

The proportion of relevant articles identified by the search out of all relevant articles that exist on that topic

Specificity

proportion of relevant citations that the search is able to retrieve, or a measure of the ability of the search to exclude irrelevant articles

completing the search

-searched all relevant databases, using a variety of keywords and subject headings

-mined article bibliographies for their cited references

-looked in Google Scholar or Web of Science to see who has cited those articles

critical appraisal

The purpose is to determine the scientific merit of a research report and its applicability to clinical decision-making

Levels of Evidence: Quantitative Studies

Level 1: systematic reviews/meta-analyses

Level 2 of Evidence: Individual study with strong design

Level 3 of Evidence: Study with less rigorous design

Level 4 of Evidence: Case series

Level 5 of Evidence: Mechanistic reasoning

Grade up: strong effect size, strong design

Grade down: poor design, lack of consistency with other studies, small effect

Levels of Evidence - Qualitative Studies

1: generalizable studies

2: conceptual studies

3: descriptive studies

4: case studies

Appraisal Process

-first read the abstract to determine if the article addresses the clinical question

-should focus on the designs, the analysis, and the clarity of writing

Core Questions for Literature Appraisal

-Is the study valid? Is the research question clearly stated? What type of research is being done

-How was the study designed and implemented

-Are the results meaningful? What were the outcomes? How are the results interpreted

-Are the results relevant to my patient? Is clinical relevance discussed? Is the approach feasible in my setting? Is this approach worth the effort to incorporate into my treatment plan

-are applied for quantitative and qualitative studies

Critically Appraised Topic (CAT)

A brief summary of a search and critical appraisal of literature related ot a focused clinical question; evidence-based and patient-based tool

-useful to inform an aspect of patient care related to intervention, diagnosis, prognosis, or harm

-limitation: have a short shelf-life, often limited in scope, and include fewer references

5 steps in CAT process

-ask a question

-acquire relevant literature

-appraise the quality of the study

-apply findings to patient care

-assess the success of the clinical application

descriptive research

-observed approach that documents traits, behaviors, attitudes, and conditions of individuals, groups, and populations

-takes advantage of naturally occurring events or available information to generate new information through inductive processes

-exploratory elements

basic descriptive information

used to indicate norms, trends, needs, and circumstances that inform practice

-used to characterize subjects and other relevant circumstances that surround the research

univariate research

Data is collected on a single variable or a series of single variables and then characterized with descriptive statistics

-Characterize the sample or circumstances that make up any study

-Characterize a problem or phenomenon that is of interest to the fields

-document incidence and prevalence of health-related conditions

-establish norms to provide a basis for prescribing corrective intervention or predicting future performance

-document developmental phenomena: growth or change over time within selected segments of the population

-document case studies: in-depth descriptions of the experiences or behaviors of a particular individual or a series of individuals

correlational research

an examination of relationships between 2 or more variables

-can provide evidence that is consistent or inconsistent with causal assertions to inform the development of theoretical propositions

-can be useful first steps to understanding causal relationships and precursors ot ecperimental studies

-limitation: does not have an experimental control, possibly a variable, so there is no way to establish causality

developmental studies

describes developmental change and the sequencing of behaviors in people over time

Longitudinal Studies

-Collect data at prescribed intervals on the cohort to determine developmental changes over time

-studies patterns of change

cross-sectional research

-Collect data at one point in time on cohorts at different developmental stages

-describes differences in levels or various stages of life

-often seen with survey research and exploratory intervention studies

-sample often includes a wider range of subjects that allow for a variety of analyses

normative studies

-describes typical or standard values for characteristics of a particular population

-mean and standard deviation

-samples must be large, random, and representative of the population's heterogeneity

-important for determining need for intervention, developing treatment goals and protocols

case reports/studies

-in-depth description of a person's condition or response to treatment

-thorough analysis of a single situation

-want to end with a hypothesis that can be tested experimentally

-typically seen with unusual medical conditions

-inherent lack of control and limited generalizability

historical research

-critical review of events, documents, literature, and other sources of data to reconstruct the past in an effort to understand how past events occurred

-synthesizing data from the past

-can include primary and secondary sources

parameters

the measure of population characteristics

statistics

a systematic conversion of observation into numbers and then using those numbers to draw inferences about the world

descriptive statistics

used to characterize the shape, central tendency, and variability within a set of data, often with the intent to describe a sample or population

general goal of descriptive statistics

-organize

-summarize

-analyse

-group outcomes

frequency distribution

a table of rank ordered scores that shows the number of time each value occured, or its frequency

-refers to the total set of scores or measures

-often reported as n

population

-all possible measurements (N)

-measurements are called parameters

sampling distribution

-a subset of the population under study (n)

-measurements are called statistics

The most common ways to graph frequency distributions

histograms, line plots, and stem-and-leaf plots

-described by three characteristics: shape, central tendency, and variability

Negative skew

-most scores are positive

-more people die at old age than when they're young

-The mean is smaller than the median

positive skew

-most are negative

-annual income In the US, the number of people at the high end is small, and the majority of people cluster around the low end

-The mean is larger than the median

mean

The sum of a set of scores divided by the number of scores

-used for numerical data and symmetric distributions

median

of a series of observations is that value above which there are as many scores as below it, dividing a rank order distribution into two equal halves

-unaffected by extreme scores

-useful for skewed distribution or ordinal scores

mode

the score that occurs most frequently in a distribution

-more than one score with the highest frequency = bimodal or multimodal

-most useful in the assessment of categorical data on nominal or ordinal scales

-limited application for continuous data