Clinical Reasoning

______________ = ethical concept in the design of clinical trials - we should only conduct trials under conditions of uncertainty when we do not already have sufficient evidence of the benefit of a treatment - placebos must be used judiciously (only when no proven intervention exists or when there is compelling rationale for use because the patients will not be as risk)

equipoise

what are the 4 basic types of research designs

experimental, descriptive, exploratory, integrative

common types of sample selection:__________: a sample that is deliberately chosen by using a sampling plan that screens out subjects with certain characteristics and/or selects only subjects with other relevant characteristics (inclusion and exclusion criteria)

selective sample

An (attribute/active) variable is one that is manipulated by the experimenter so that subjects are assigned to levels of the independent variables An (attribute/active) variable is a factor that cannot be assigned because it is an inherent characteristic of the participants (age, gender, diagnosis) - can't be manipulated by the experimenter

active, attribute

____________ = loss of study participants during a study

attrition

You are conducting a clinical experiment in which your null hypothesis is that there is no difference between means of a score on the Disability of Arm Shoulder and Hand (DASH) outcome measure for a group of patients receiving supervised therapeutic exercise and for a group who received a home exercise program. After analyzing the data from your study, you fail to reject the null hypothesis. Subsequently, the findings of two very similar (almost identical) studies suggest that there is a difference between the means of their groups. What is your assessment of your decision-making related to your study? a. You have committed a Type I error b. You have committed a Type II error c. You made the correct decision

b (You have committed a Type II error)

_____ occurs when a study's results are affected by unknown or unacknowledged errors resulting from the study's design or protocols - potential sources: natural history of a disease, placebo effect, drop outs, selection bias, info bias

bias

______________ is considered to try to control for observation bias, the participants knowledge of their treatment status of the investigators expectations can, consciously or unconsciously, influence the outcomes - the strongest approach is when neither the subjects nor the investigators known the identity of treatment groups until after data are collected

blinding

these are observational research comparing subjects who have a specific condition (the 'cases') with patients who do not have the condition but are otherwise similar (the 'controls') - no intervention is provided on part of the researchers. People who already have the condition of interest are compared to a group of people without the condition

case control

This research evidence describes practice. Often focuses on a patient or group of patients, but may also focus on facilities, edu programs, or other definable units. Topics often include patient/client management, ethical dilemmas, use of equipment or devices, etc. - can't prove effectiveness, test hypotheses, or prove cause & effect, and the outcomes that they report can't be generalized to patients or entities - Tend to receive few citations (measure of contribution to the field/science) aka falling out of favor - Benefits: critical initial step in a series of steps to higher-level evidence (agents for CHANGE) such as RCTs and Meta-analyses (MA) (lay the foundation/open the doorway to more and better research on a topic); Low expense (community setting, usually no cost); Appeal to clinicians (writers & readers)

case study (case report)(in order to be published has to "add something to the overall literature": novelty and innovation: either discuss a new approach or use old approach for a new thing; inform/advance practice; suggest testable hypotheses)

these are observational studies in which a defined group of people is followed over time. The outcomes of people in subsets of the defined group of people are compared; typically a comparison between people who were exposed to those who were not exposed to a particular intervention or other factor of interest also known as a longitudinal study

cohort study

one intervention is perceived as "better" by participants so those in the "worse" group (often the control group) take action to achieve good outcome. This can lead to crossover contamination

compensatory rivalry

_______________ sampling: oversampling of a group due to unequal distribution of subgroups within the population - Presents a challenge because the sample no longer reflects the true proportion of population diversity - To correct, sample weight is assigned to each person based on the number of people in the population represented by that subject (corrects for the bias in the sample groups)

disproportional

factors that would preclude someone from being a subject, usually due to being potentially confounding to the results and likely to interfere with interpretation of the findings

exclusion criteria

this type of research identifies cause and effect relationships among variables - uses comparison stats to identify the relationship (t-tests, ANOVA) - common types: RCTs, non-randomized trials, single-subject designs

experimental (research)

This type of clinical trial has a controlled design in a controlled environment. They tend to have smaller samples and focus specifically on outcomes.

explanatory (trials)

this type of research identifies relationships between variables - usually uses correlation stats - common types of this research: cohort & case control studies, methodological studies (reliability and validity studies)

exploratory (research)

How well do these studies compare to my situation/question are the findings applicable beyond that particular study

external validity

what are the two major threats to internal validity in experimental research studies

history (events between baseline and post-test that may impact the outcome/dependent variable), Maturation (process that is a natural change in the outcome due to the passing of time)

describes the primary traits of the target and accessible populations that will make someone eligible to participate - Must consider the variety of characteristics present in the population in terms of clinical findings, demographics, and geographic factors, as well as whether they are important to the question being studied

inclusion criteria

characteristic or intervention that will predict or cause an outcome, or change in the dependent variable (researchers manipulate these as they see fit; the "cause"); need to define how these will be manipulated in research question.

independent variable

this type of research rigorously integrates finding from more than one study on the same topic - statistical methods vary - types: evidence-based clinical guidelines, meta-analysis, systematic reviews

integrative (research)

One way to handle non-compliance and drop outs when doing data analysis is ______________ which consists of data are analyzed according to original random assignments, regardless of the treatment subjects actually received, if they dropped out of were non-compliant (so data is analyzed according to the way the researchers intended to treat the subjects). This analysis ideally includes all subjects. - 2 main purposes of this approach = 1) it guards against the potential for bias if dropouts are related to outcomes or group assignment, and preserves the original balance of random assignment, & 2) it reflects routine clinical situations, in which some patients will be non-compliant. This is most relevant to pragmatic trials, where the design is intended to incorporate real-world practice. - This analysis runs a higher risk for bias towards type II errors (failing to reject the null hypothesis when there really is a difference between groups; false negative; can happen if you conclude that the difference between groups is so small that it doesn't matter much or is very hard to detect, OR concluding that the difference is big enough to care about, but your sample size was just too small to tell you much; Clinical implication: failing to use an effective treatment)

intention to treat (ITT)

______________ variable changes with exposure and contributes to the outcome; often a mechanism - an intervening variable in causal pathway between the intervention and the outcome

mediator (i.e. summer/warm weather is an independent variable that is directly related to the outcome of more drownings occurring. A mediator in this relationship would be pool use, summer causes more pool use, which then causes more drowning)

assuming the results are valid as presented, the authors draw inappropriate conclusions about the meaning of those findings/explanation of results (classic example: correlation)

misinterpretation

_______________ = effect of the treatment depends on the level (or value) of another variable (the modifier) - changes the magnitude and/or direction of the intervention --> outcome relationship; can be direct or indirect effect on outcome; may determine WHO will respond best to an intervention

moderator (i.e. for the summer and drowning example, a moderator could be age, this modifies the magnitude of the effect -drowning-, but isn't being driven by the independent variable -summer/warm weather-) (Another example of a moderator would be in the relationship between self-rated memory function and depressive symptoms - if an older adult can self-identify that their memory is declining that may lead to increased depressive symptoms, but depression can also lead to decreased memory function. A moderator in this relationship is the person's level of self-efficacy. So, a person may identify that their memory is declining, but they come up with a plan to work out and improve their health/brain function and know they have a social support system behind them [high self-efficacy and thus have fewer depressive symptoms compared to someone who may actually have better memory function but no plan or social support [low self-efficacy] leading them to have more depressive symptoms.)

One way to handle non-compliance and drop outs when doing data analysis is ______________ which consists of eliminating data from any subjects who did not get or complete their assigned treatment and include only those subjects who sufficiently complied with the trial's protocols. This usually results in a bias that favors a treatment effect, as those who succeed at treatment are able to tolerate it well and are most likely to stick with it. - increased risk for Bias towards Type I error (false positive; concluding there is a difference between groups, when in fact, there is no difference; Clinical implication: applying an ineffective treatment)

per protocol analysis

larger group to which the research results will be applied; the aggregate of persons or objects that meet a specified set of criteria and to whom we wish to generalize results of a study

population

- Research questions should define & include PICO(T), what does this acronym stand for?

population (patient, problem), intervention, comparison, outcome, time

the ability to find significant effects when they exist; influenced by variability within a sample (standard deviation), anticipated magnitude of the effect, and the sample size - commonly accepted value for this: 0.80 or higher - the lower this value is, the higher the risk of committing a Type II error

power (smaller sample size tends to decrease a study's power)

_____________ = a disadvantage of repeated measures - learning that can take place when one individual repeats a task over and over

practice effects

the particular group of participants you are testing: selected from the population - must be representative (have the same characteristics) of the population - serves as a reference group for estimating the characteristics of and drawing conclusions about the population

sample

this type of bias is caused when individuals have different probabilities of being selected in a study according to relevant study characteristics a bias that is caused by some kind of problem in the process of selecting subjects initially - or in a longitudinal study - in the process that determines which participants drop out of the study

selection bias (Sample Bias = when individuals selected for a sample overrepresent or underrepresent certain population attributes that are related to the phenomenon under study - conscious or unconscious)

alternative approach to RCTs, which allow for continuous analysis of data as they become available, instead of waiting until the end of the experiment to compare groups. Results are accumulated as each subject is tested so that the experiment can be stopped at any point as soon as the evidence is strong enough to determine a significant different between treatments. - Major advantage of this approach is the potential ability to make a decision about treatment effectiveness earlier than in a fixed sample study, leading to a substantial reduction in the total number of subjects needed to obtain valid statistical outcomes and avoiding unnecessary administration of inferior treatments.

sequential clinical trials

_____________ trials examine the effect of a treatment or intervention on a particular disease or condition

therapeutic

_____________ designs: repeated measures that are great for establishing patterns/trends; ideally you would take repeated pre-intervention measurements to establish a static baseline, provide intervention, and then do repeated post-intervention measurements - also helps establish dose-response relationship - when doing repeated measures with multiple interventions you have to be careful about latency effects from previous interventions potentially being the source of a patient's change in performance level (order effects)

time series

Rejecting null hypothesis when it is true (false positive) = what type of error (concluding there is a difference between groups, when in fact, there is no difference)

type I error

failing to reject the null hypothesis when there really is a difference between groups (false negative) = what type of error - can happen if you conclude that the difference between groups is so small that it doesn't matter much or is very hard to detect, OR concluding that the difference is big enough to care about but your sample size was just too small to tell you much (uninformative null finding)

type II error

response or effect that is presumed to vary with the independent variable (i.e. disability, pain, QoL, strength levels) (these variables are not manipulated in any way but instead simply measured; outcome of interest; the "effect"); need to define how you plan to measure this.

dependent variable

this type of research describes characteristics of groups of people or other phenomena - often uses questionnaires, interviews, and/or direct observation - common types of this research: developmental, normative (establishes norms of specific variables), qualitative, and evaluation research

descriptive (research)

phenomenon in which subjects in behavioral studies change their performance in response to being observed

Hawthorne effect

this is considered by many to be the gold standard of evidence for a research study - involves randomly assigned members of the study sample to either receive the intervention or a control treatment - randomizing subjects ensures the two groups are equal on factors that may influence the study outcomes (not biased, even unconsciously, by the assigner) - having a control group ensures that the changes between the beginning and end of the study are due to the treatment

RCT (examines pre- vs post-treatment perfomance [cause and effect] reveals a causal relationship between independent and dependent variables employs repeated and reliable measurement within & between subject comparisons to control for major threats to internal validity)

outliers tend to move towards the average/mean over time; part of random error; more likely with increasing measurement error (inversely related to reliability of an outcome measure; when follow-up is only examined in a sub-sample), & not guaranteed to happen in every case

RTM (regression toward the mean) (Can confound decisions in medicine [if baseline measurements are made when a patient's values are fluctuated towards the tails/further away from their 'mean' i.e. their HDL levels are low at baseline and the doctor tells them to change their diet, start exercises, etc. and the next time they see the patient their HDL levels are up which may really just be due to them regressing towards their mean and not from the doctor's advised intervention])

the average amount that each of the individual scores varies from the mean of the set of scores; the most commonly used measure of variability

SD (standard deviation) (1 SD = 68% of your data, 2 SD = 95%, 3 SD = 99.7%) (only applies to a NORMAL model [normal distribution]; in reality data doesn't fall in a normal distribution)

One way to handle missing data when doing data analysis is ______________ which involves excluding subjects and using only the data from those who complete the study; represents the efficacy of an intervention for those who persist with it but may be open to serious bias. This method can be justified if the number of incomplete/excluded cases is small and if the data is missing completely at random (MCAR - data missing because of completely unpredictable circumstances that have no connection to the variables of interest or the study design) - throwing out cases will make the sample size smaller and likely lower the power of the study. - most direct approach to handling missing data

completer analysis

_________ are related to the outcome, and perhaps the intervention, but not always (!!) a causal pathway - present threats to internal validity because they offer competing explanations for the observed relationship between the independent and dependent variables; that is, they interfere with cause-and-effect inferences. - Confusion of effects. - Ways to control for these: sample accordingly (inclusion/exclusion criteria), stratify during sampling or randomization, and measure as covariate for consideration in analysis

confounders (Covariates: explain part of the variability in outcome, but are not influences by intervention)

Common types of sample selection:_____________: a sample in which the subjects are chosen on a strict "first come, first chosen" basis. All individuals who are eligible should be included as they are seen (inclusion criteria)

consecutive sample

common types of sample selection:_____________: a sample where the patients are selected, in part or in whole, at the convenience of the researcher. The researcher makes no attempt, or only a limited attempt, to ensure that this sample is an accurate representation of some larger group or population

convenience sample

One way to handle missing data when doing data analysis is ______________ which attempts to maintain the integrity of a sample and fulfill assumptions for intention to treat, missing data points are replaced with estimated values that are based on observed data using the process called imputation. Analyses can then be run on the full dataset with the missing values "filled in" and conclusions drawn about outcomes based on those estimates - always some uncertainty in how well the imputed scores accurately represent the unknown true scores.

data imputation

subjects receiving less desirable treatment come to resent the situation, may perform even lower which inflates the apparent "treatment" effect, if any, for the treatment group

demoralization

2 Major complications that affect data analysis =

non compliance, missing data

____________ trials: focuses on demonstrating that the effect of a new intervention is the same as standard care, in an effort to show that it is an ~acceptable substitute~. This approach is based on the intent to show no difference, or more precisely that the new treatment is "no worse" than standard care.

non-inferiority

point that lies outside of the pattern formed by the majority of the data; data point that is very different from the rest of the data in the set - Can be defined by means/SD (outside of 2 or 3 SD), or by median (using a Boxplot) - Why do they occur? Error (in measurement/recording), small sample sizes (too small to generate a full range of observations), subjects that are inappropriately included in a sample, etc.

outlier

__________ clinical trial is a clinical trial that focuses on correlation between treatments and outcomes in real-world health system practice rather than focusing on proving causative explanations for outcomes, which requires extensive deconfounding with inclusion and exclusion criteria so strict that they risk rendering the trial results irrelevant to much of real-world practice - flexible design - heterogeneous samples (more challenging to interpret data, but more applicable to clinical practice) - large samples - diverse practice settings - meaningful outcomes (pt satisfaction, QoL, mobility and function, etc.)

pragmatic

using knowledge and reasoning to estimate the outcome after therapy

predict (i.e. predict the response of a patient's LBP to your manual intervention)

assess the outcome regardless of interventions, based on signs/Sx/biomarkers

prognosticate (i.e. prognosticate the natural history of a disease)

intensive, prospective study of a single patient or small group (1-3 ppl, 3 is ideal to demonstrate experimental effect on at least 3 occasions) - intervention is systematically manipulated in experimentally-controlled manner across multiple discrete phases - repeated, frequent measurement of the behavior targeted for change (must repeat through each phase of design) - Methodology: Phases. At a minimum there should be a baseline (A) and an intervention (B). Phase A controls for personal characteristics and analysis of this phase is critical for interpretation of Phase B. Recording a minimum of 5 points are recommended within a phase - Important to look at the stability (consistency over time) of points and trend/slope (rate and direction of change) in the dependent variable - threats to internal validity in this type of study design: history, testing, maturation, spontaneous recovery

single case experimental design (SCED)

________ __________ sampling: identifying relevant population characteristics and partitioning members of a population into homogeneous, non-overlapping subsets, or strata, based on these characteristics; Random samples are then drawn from the stratum to control for bias - allows researchers to obtain a sample population that best represents the entire population being studied - involves dividing the entire population into homogeneous groups called 'strata' - differs from simple random sampling (which involves the random selection of data from an entire population, so each possible sample is equally likely to occur)

stratified random (stratify samples based on characteristics that have high potential to confound results. Must be mutually exclusive-same person can't belong to more than one group)

____________ trials: research design is looking to show that a new intervention is more effective than a placebo or standard treatment - aim is to demonstrate that one treatment is "superior: to another - The process typically involves the statement of a null hypothesis and an alternative hypothesis asserting that there will be a difference (directional or non-directional) with the intent to reject the null and accept the alternative in favor of one treatment.

superiority

what are the 5 criteria to look at when hypothesizing cause and effect relationships from observational studies

temporal clues (exposure to proposed causative agent must occur before the outcome), strong association (between proposed causative agent and outcome), reproducibility (+ consistency, of the relationship across samples and conditions), credible postulated mechanism (to explain how exposure to the agent could result in the outcome), dose response relationship (if A causes B, then more A should cause more B)