Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
75 Cards in this Set
- Front
- Back
what is evidence based medicine?
|
use of current best evidence in making decisions about the care of individual patients
|
|
Steps to providing EBM
|
ask - clinical question
acquire (best scientific evidence) appraise - evaluate evidence apply ***** - evaluate performance |
|
practice of EBM
|
using... to improve patient outcomes
1. patient's values and expectations 2. individual clinical expertise 3. best available clinical evidence |
|
Research Methods - approach
|
quantitative** vs qualitative
or mixed methods |
|
quantitative research strategies
|
experimental > intervention under control of the researcher
observation > not under control |
|
study designs in clinical research
|
|
|
prospective vs retrospective vs cross-sectional
|
PROSPECTIVE
1. experimental > randomized or controlled clinical trial 2. analytical > cohort study, comparison groups 3. descriptive > cohort study, single group RETROSPECTIVE 1. analytical > cohort study, comparison group OR case control 2. descriptive > cohort study, single group * retro cannot be randomized CROSS-SECTIONAL > analytical or descriptive |
|
deciding on type of study
|
factors?
- state of current knowledge - availability of comparison group - ethics of randomization - frequency of events - research question - feasibility |
|
What is a clinical trial?
|
form of clinical research that evaluates the effects of >1 health-related interventions on health outcomes
- randomized controlled trial is the primary method for evaluating therapeutic interventions |
|
TCPS2
|
tri-council policy statement-2
|
|
Randomized Controlled Trials
|
- experimental method
- treatment and control groups are assigned by chance rather than chosen by subject/doctor/researcher > attempts to make both groups the same except for intervention key features: - ethics - randomization - concealment of allcoation - blinding - complete follow-up |
|
Clinical Drugs Trials - Phases
|
1. "first into man" - small number of patients > figuring out safety and doses
2. "proof of principle" > studies involving patients with specific disease, first venture into safety and efficacy 3. large-scale comparative studies > 3a) demonstrative of efficacy and safety; 3b) new indications 4. surveillance and post-marketing studies |
|
Hierarchy of evidence
|
best to least
1. meta-analyses and systematic reviews 2. randomized controlled trials 3. cohort studies (prospective better than retro) 4. case-control studies 5. case reports (write article about somebody with particular side effect); case series (several case reports) 6. editorials, expert opinion (only when there is no other evidence available) |
|
Makes a good trial
|
1. high internal validity
- low systematic error > bias and confounding (systemic deviation from the truth) 2. precision > low random chance error 3. high external validity - generalizabiltiy > extent in which results can be applied to other individuals or settings > AKA EBM: can this be applied to my patient? |
|
Strengths of RCT
|
- difference at the end = due to intervention >> only study that can prove causality, and not just correlation
- groups similar for known and unknown confounders |
|
Randomization in RCTs
|
randomization > sorting participants > intervention OR control group > outcome
|
|
Randomization
|
- allocation of treatment by chance
- reduces SELECTION BIAS - achieved by random sequence generation, coin toss, computer...always reported in published study |
|
Allocation Concealment
|
- researcher doesn't know the patient that is being recruitment
- not usually described in published papers Sequentially Numbered Opaque Sealed Envelopes (SNOSE) |
|
Blinding
|
- disguise treatment
- blind to avoid unequal co-intervention in treatment groups and unequal ascertainment of outcome - look for descriptions in published papers about who was blinded > will affect interpretation |
|
allocation concealment VS blinding
|
AC = blinding before randomization > can always be implemented > aims to prevent selection bias
B = blinding after randomization > goal is to prevent unequal treatment of groups |
|
cohort
|
group of individuals having a statistical factor in common
|
|
Cohort Study
prospective vs restrospective |
proceed from exposure to factor of interest (now or in the past); then FOLLOW FORWARD in time to determine outcome
PROSPECTIVE - start assembling cohort and collects data in the future RETROSPECTIVE - use existing database to assemble cohort at a time in the past and follow patients forward |
|
Case Control Study
|
- looking backwards
- looks at group with a certain outcome and selects another group without the outcome - information is compared on whether subjects have been exposed to the treatment |
|
Cross-sectional study
|
looking at now/snap-shot
- examines relationship between outcomes and exposures in a defined population at one particular time |
|
Following up on RCTs
|
- ensure that all participants are accounted for > and if they dropped out, why?
this reduces ATTRITION BIAS (bias arising from exclusion of participants from study groups) - can look in paper to see who dropped out and reasons |
|
Patient flow diagram
|
assigned for eligibility > patients declined > rest randomized > allocated to 1) intervention or 2) control > numbers tallied for a) died before discharge, b) wished to be excluded, c) lost to follow-up and d) excluded from analysis > reported number of patients analyzed
|
|
historical documents
|
1. hippocratic oath > earliest known code of medical ethics (5th century BC) > "first do no harm"
2. nuremberg code - "voluntary consent" > US report post-Nazi 3. declaration of Helsinki > world medical association (1964) > cornerstone for research ethics 4. Belmont report > National Commission for the protection of Human Subjects of Biomedical and Behavioural Research > summarizes ethics for research with human subjects > "respect, beneficence, justice" |
|
General Requirements for RCTs - lecture 8
|
medically, scientifically and ethically justifiable
- appropriate treatment and comparison groups - benefit vs risk - compatibility with health care needs - sufficient sample size to find a difference - reasonable doubt about efficacy |
|
responsible conducts of research
|
- ethics/human subjects protection
- accountability of researchers |
|
rules and regulations of clinical research
|
REGULATIONS : Health Canada Food and Drug Regulations for Clinical Trials
Guidelines - Tri-council Policy Statement-2 (TCPS2) - ethical conduct for research involving humans (Canada) - International Conference on Harmonization of pharmaceuticals for human use Good Clinical Practice (ICH-GCP) guidelines |
|
Drug Approval process in Canada
|
approvals based on drug's safety, efficacy and quality
1. Therapeutic Products Directorate (TPD) > reviews trials for pharmaceuticals and medical devices 2. Biologics and Genetic Therapies Directorate (BGTD) > reviews trials for biologics and radiopharmaceutics 3. Natural Health Product Directorate (NHPD) > reviews trials for natural health products |
|
Clinical Trials
|
- investigation of drug for use in humans that involve human subjects > intended to verify clinical, pharmacological or pharmacodynamic effects
- Phase 1, 2, 3 studies involving human subjects must be reviewed and approved by Health Canada - in Canada, all studies must follow ICH-GCP guidelines as well as be reviewed by Research Ethics Board (REB) > who follows TCPS2 |
|
ICH-GCP
|
2 important elements
1. patient safety 2. study credibility standard for clinical trials that provides assurance that data and reported results are credible , while keeping the integrity and confidentiality of study subject are protected |
|
Research Ethics Board (REB)
|
- independent peer review body
- >5 members, both medical and lay - reviews protocol to ensure safety and well-being of subjects - research begins AFTER APPROVAL is granted - usually more than one institution involved |
|
TCPS2
|
core principles
1. respect for persons - autonomy, informed consent > subject voluntarily confirms willingness to participate ; may be delegated to an authorized third party (ex. parent for child) 2. concern for welfare - beneficence, maximum benefit/minimum risk 3. justice - fairness |
|
threats to voluntary participation
|
1. undue influence > from a person of authority
2. coercion > involving harm or punishment 3. incentives > not the same as compensation/reimbursement |
|
Capacity (research)
|
participants must understand information about the research and appreciate the consequence of their decision to participate
"assent" obtained for children |
|
Stopping or withdrawing from study
|
1. researchers may stop a study if it doesn't meet safety/efficacy rules
2. researchers may pull participant > worsening health, better therapy available, non-adherence 3. participant removes themselves |
|
research without consent
|
allowed !
- involves no more than minimal risk - impossible to carry out research with consent - won't affect person's welfare - not a therapeutic intervention - consent is obtained later, if appropriate |
|
concern for welfare
|
research risks
1. physiological > ex side effects 2. psychological > ex stress 3. social > ex embarrassment 4. economic > ex monetary |
|
privacy and confidentiality
|
privacy risks related to being able to identify participants and associated potential harms
|
|
clinical equipoise
|
genuine uncertainty exists on the part of the relevant expert community about what therapy or therapies are most effect for a given condition
|
|
placebo controlled trials
|
ethically acceptable if...
- used to establish the efficacy or safety of the intervention - doesn't compromise safety/health of patient - compelling scientific justification placebos used when... - no established effective therapies - doubt about benefit of available therapies - add-on trial - patients have provided refusal of effective therapy and withholding therapy will not cause serious/irreversible harm - patients resistant to available therapies due to treatment/medical history |
|
potential conflict of interest by being a clinician-researcher?
|
yes!
responsibility of health care provider is patient VS researcher's is research |
|
clinical trial registration
|
public registries to provide a record of ongoing clinical trials
intended to increase transparency and accountability |
|
Adverse events
|
- untoward medical occurrence in patient or subject
- does not necessarily need to have a casual relationship with treatment or usage - includes adverse reactions which are noxious and unintended response to a medicinal product or to a medical or surgical device |
|
serious adverse event + reporting
|
SERIOUS ADVERSE EVENT
- is life threatening or results in death - results in persistent or significant disability/incapacity - is a congenital anomaly or birth defect - results in hospitalization REPORTING - reporting requirements will vary by local research ethics board and regulatory authority |
|
adverse drug reaction
|
ADR is a type of adverse event with a reasonable possibility of being related to the treatment
|
|
research misconduct
|
1. plagiarism - misrepresenting someone else's work as one's own
2. fabrication - report measurements that have not been performed 3. falsification - ignore or change relevant data that contract reported findings |
|
Clinical trial protocol / research plan
|
1. research problem - background and significance
- area of concern 2. research question - objective/hypothesis - what? - primary questions = focus of the study - secondary questions = add valuable information 3. design - structure and procedures 4. statistical issues - sample size, analytical approach |
|
elements of the research question
|
PICOT
population intervention comparison outcome timeframe |
|
null hypothesis (for superiority trials)
|
superiority trials - when you test whether a drug is better than another drug
null hypothesis has to be a testable statement > there is no difference in whatever is being tested |
|
Alternative/Study hypothesis
|
opposite of a null hypothesis
|
|
hypothesis testing
superiority vs equivalence vs non-inferiority |
1. superiority - intended to determine that new treatment is better than control
2. equivalence - new and control treatment are therapeutically similar 3. non-inferiority - new is no worse than the comparison WHEN CHOOSING... - consider effective treatment and ethics of placebo-controlled trial - pharmacological profile and other factors (ex. convenience, cost,...) |
|
Trial designs - participants
|
when choosing...
1. selection criteria - what kinds of pts are best suited for research question 2. sampling procedure - process of picking subgroup to be part of study |
|
Sampling the target population
|
target population > all members of particular group that results are intended to be generalized to
study sample > subset of popn included in study |
|
selection/admission criteria
|
1. ethical rationale > don't want to deny treatment of impose contraindicated treatment
2. scientific rationale > must fulfill same admission criteria to optimize study results |
|
Inclusion criteria to maximize
|
- rates of outcomes
- likely benefit of intervention - generalizability - ease of recruitment |
|
exclusion criteria to minimize
|
- harm
- ineffectiveness - non-adherence - loss to follow-up - practical problems |
|
what is a "representative sample"
|
sample that is similar to popn on all charateristics
|
|
Sampling methods
|
1. probability/random > all members of popn have equal chance > rarely done
simple - each individual is randomly chosen stratified - popn divided into homogenous subgroups before simple sampling is applied cluster - relatively homogenous group by dividing popn up and selecting one of these groups 2. non-probability / non-random systematic - ordering popn and then selecting members at regular intervals convenience - patients selected because of convenience purposive - patients selected based on a variety of criteria |
|
Selection / sampling bias
|
1. individuals have unequal chance of being selected
2. intervention and comparison group differ from each other |
|
Trial Design - exposure to intervention (4)
|
parallel - each subject just receives one of the treatments - each subject receives all treatments
cross-over cluster - groups of subjects (rather than individuals) are randomized to interventions factorial |
|
Longitudinal study
|
follows subjects forwards over time
|
|
Sample size - based on number of participants
|
1. fixed size - prior sample size calculatio
2. mega trial - large sample 3. sequential - variable size 4. N-of-1 - single subject |
|
allocation ratio
|
ratio of participants in each intervention group
|
|
interventions
|
1. intervention/treatment/experimental group
2. comparator/control > intervention can only be measured against an alternative |
|
choice of intervention
|
1. generalizability - feasibility
2. complexity - incorporation of clinical decision-making 3. strength - reasonably affects outcome |
|
choice of comparison to intervention
|
1. placebo
2. active/positive control 3. same intervention, diff regimen 4. absence of treatment 4. standard care control 1, 4 - negative control |
|
head to head VS add on
|
1. head to head - comparing active treatments
2. add-on - comparing the effect of adding on an active treatment |
|
allocation methods + randomization methods
|
1. randomization - chance > key of RCT
2. quasi-pseudo-randomization RANDOMIZATION METHODS 1. simple - all subjects are randomized 2. blocked - randomize subjects within a block 3. stratified - randomize subgroups sharing similar characteristics 4. cluster - randomly allocate group rather than individuals (ex hospital) |
|
blinded studies
|
blinding methods
1. 1 or more groups don't have knowledge of treatment assignment 2. open trial - all participants and investigators are aware of treatment assignment blinding is important to minimize risk that groups are treated differently during trial or that outcomes are reported differently - intended to reduce bias after randomization |
|
measurement bias
|
1. performance bias - systematic differences between groups not due to intervention
2. contamination bias - control group gets the intervention 3. co-intervention bias - treatments other than intervention applied unequally 4. detection bias - outcomes assessed unequally due to preconceived notions |
|
types of outcomes
|
1. primary and secondary (ex. death, disease, pain...)
- easy to observe - free of measurement error - clinically relevant - chosen before starting the study - can be observed independent of treatment assignment 2. Composite outcome - a few outcomes adding together 3. Surrogate outcome - associated with relevant clinical outcome, but by myself it doesn't matter |
|
outcome evaluation
|
1. efficacy/explanatory/fastidious
2. effectiveness/management/pragmatic |