1
Q
On what two fundamental aspects can the design of a research study be described?
A
Identification (grouping) of study subjects based on exposure or outcome, and the time course of the investigation (prospective, retrospective, or cross-sectional).
2
Q
How are the study subjects identified or grouped in research designs?
A
By differences in exposure (exposed vs. non-exposed) or differences in outcome (cases vs. controls).
3
Q
What are the possible time courses of investigation in research studies?
A
Prospective (forward from exposure to outcome), retrospective (backward from outcome to exposure), or cross-sectional (exposure and outcome measured simultaneously).
4
Q
What defines a prospective research study?
A
Subjects are assembled based on exposure, and outcomes are observed over time.
5
Q
What defines a retrospective research study?
A
Subjects are assembled after having already experienced the exposure of interest.
6
Q
What is the primary aim of descriptive studies?
A
To describe characteristics of a group without testing causal hypotheses or making comparisons.
7
Q
What do descriptive studies focus on reporting?
A
Characteristics present in the study group and their distribution.
8
Q
What are the simplest forms of descriptive studies?
A
Case reports and case series.
9
Q
What is a cross-sectional study?
A
A descriptive study where exposure and outcome are measured at a single point in time.
10
Q
What alternative name is used for cross-sectional studies and why?
A
Prevalence studies, because they assess attributes at one time point.
11
Q
What common research method is a type of cross-sectional study?
A
Surveys.
12
Q
What is a census in epidemiological research?
A
A cross-sectional study that includes the entire population rather than a subset.
13
Q
What is a prospective longitudinal study?
A
A study with repeated observations in the same population over an extended period.
14
Q
What are the advantages and disadvantages of longitudinal studies?
A
Useful for studying natural history, risk factors, and incidence, but costly and time-consuming.
15
Q
What defines analytical studies?
A
Comparative studies that measure associations between variables such as exposure and outcome.
16
Q
What are the common types of analytical studies?
A
Case control studies and cohort studies.
17
Q
What is the structure of a case-control study?
A
Subjects are selected based on outcome status (cases and controls), then past exposure is assessed.
18
Q
How is causality supported in a case-control study?
A
If exposure is more frequent among cases than controls.
19
Q
What is a cohort study?
A
An observational study following exposed and non-exposed groups over time to observe outcomes.
20
Q
What is an ecological study?
A
An analytical study where data are collected at the group or population level rather than the individual level.
21
Q
What assumption underlies ecological studies?
A
That all members of a population have similar exposure and risk.
22
Q
What is ecological fallacy?
A
Drawing individual-level conclusions from group-level data.
23
Q
What defines an experimental study?
A
A prospective study where exposure is deliberately assigned to observe effects.
24
Q
What distinguishes experimental studies from cohort studies?
A
Deliberate manipulation of exposure.
25
What are clinical trials also called?
Interventional studies.
26
What type of design is a case-control study?
Retrospective.
27
What has already occurred when a case-control study begins?
Both exposure and disease.
28
When is a case-control study particularly useful?
When exposure is common but disease is rare.
29
What are the main advantages of case-control studies?
Easy to conduct, inexpensive, time-efficient, suitable for rare diseases, no new exposure risk, multiple etiological risk factors studied, and no attrition.
30
What are the major disadvantages of case-control studies?
Susceptibility to bias, difficulty selecting controls, inability to measure incidence or relative risk, difficulty establishing temporality, and inability to prove causality.
31
Why is defining a case critical in case-control studies?
To ensure strict diagnostic and eligibility criteria.
32
What are acceptable sources of cases and controls?
Hospitals or general populations for cases; hospital patients, relatives, neighbors, or general population for controls.
33
What principle governs control selection?
Controls must be identical to cases except for disease status.
34
How does increasing the control-to-case ratio affect study power?
Power increases up to a ratio of 4:1 but not beyond.
35
Why is matching used in case-control studies?
To improve comparability between cases and controls.
36
What risk does excessive matching pose?
Reduced risk difference and inadvertent matching of causal factors.
37
What measure of association is used in case-control studies?
Odds ratio (OR).
38
Why is relative risk not calculated in case-control studies?
Incidence rates are unavailable.
39
When does odds ratio approximate relative risk?
When disease prevalence is less than 5%.
40
Why are log transformations used for odds ratios?
Because odds ratios range from zero to infinity.
41
What is a cohort?
A group of individuals sharing a common characteristic.
42
What is an exposure cohort?
Individuals exposed to a specific risk factor.
43
What is an inception cohort?
Patients assembled at a narrow time window based on a common factor.
44
What is the defining feature of a cohort study at inception?
Exposure has occurred, but disease has not.
45
When are cohort studies particularly useful?
When exposure is rare but disease risk is high.
46
What requirements exist for cohort feasibility?
A stable, cooperative cohort and a comparable non-exposed control group.
47
What types of controls can be used in cohort studies?
Internal or external controls.
48
What is a retrospective cohort study?
A cohort identified from records and followed forward in time using existing data.
49
What measures of association can be calculated in cohort studies?
Relative risk (RR) and odds ratio (OR).
50
What is a nested case-control study?
A case-control study conducted within a cohort.
51
What are the defining features of an RCT?
Experimental intervention, randomisation, and control comparison.
52
What is the purpose of randomisation?
To eliminate bias, allow probabilistic inference, facilitate blinding, and balance baseline characteristics.
53
What types of randomisation exist?
Simple, block, stratified, cluster, minimisation, and quasi-randomisation.
54
Why is quasi-randomisation discouraged?
It is predictable and fails to ensure equal distribution.
55
Does randomisation guarantee precision?
No; precision depends on sample size.
56
What is a crossover trial?
Participants receive both intervention and control sequentially with a washout period.
57
When are crossover trials inappropriate?
When the disease can be cured by treatment.
58
What is an N-of-1 trial?
A double-blind trial conducted in a single patient.
59
What is a factorial trial?
A design testing multiple interventions and their combinations simultaneously.
60
What is a patient preference trial?
A trial where participants with preferences receive a chosen treatment, and others are randomised.
61
What is Zelen's modified RCT?
Randomisation occurs before consent is obtained.
62
What is a non-inferiority trial?
A trial designed to show that a new intervention is not worse than standard treatment beyond a predefined margin.
63
What is an uncontrolled trial?
A single-group study using historical controls.
64
What is a before-and-after trial?
A study measuring outcomes before and after an intervention in the same group.
65
What is a multi-arm trial?
An RCT with more than one intervention group.
66
What is meant by efficacy in clinical research?
How well a drug works under optimal, controlled conditions.
67
What is meant by effectiveness in clinical research?
How well a drug works under usual, real-world practice conditions.
68
How does efficiency measure in healthcare evaluation?
Whether healthcare resources are used to maximise value for money.
69
What distinguishes efficacy trials from effectiveness trials?
Their objectives and motivation; efficacy trials support regulatory approval, while effectiveness trials aim to demonstrate real-world usefulness.
70
What are pragmatic or practical clinical trials (PCTs)?
Effectiveness trials designed to answer decision-makers' questions regarding real-world use.
71
What is an example of a pragmatic PCT?
CATIE.
72
How are interventions handled in pragmatic trials?
Detailed intervention protocols are not rigidly predefined and are adjusted according to patient needs.
73
What are the defining features of pragmatic RCTs?
Comparison of relevant alternatives, diverse participants, heterogenous settings, broad outcome measures, and focus on real-world effectiveness.
74
What do pragmatic RCTs primarily inform?
Effectiveness, not efficacy, of interventions.
75
What are common limitations of pragmatic trial designs?
High dropout rates, failed blinding, less rigorous methodology, and higher nonspecific treatment effects.
76
What are the core components of randomisation in an RCT?
Sequence generation, assignment to study arms, and maintenance of assigned groups until outcome measurement.
77
Why can knowledge of treatment allocation compromise an RCT?
It can bias clinician behavior, patient reporting, and outcome assessment.
78
What is blinding in an RCT?
Keeping treatment allocation concealed after randomisation to prevent bias.
79
What is an open-label study?
A study in which all parties know the treatment allocation.
80
What defines a single-blind study?
Either the patient or clinician (usually the patient) is unaware of allocation.
81
What defines a double-blind study?
Both patient and investigator are unaware of allocation.
82
What defines a triple-blind study?
Patient, investigator, and outcome assessor are unaware of allocation.
83
What is unblinding?
Planned or unintended disclosure of treatment allocation.
84
What is allocation concealment?
Keeping the randomisation sequence secret until assignment occurs.
85
Why is allocation concealment critical?
To prevent selection bias before group assignment.
86
How can allocation concealment be achieved?
Separate sequence generation from recruitment, centralised randomisation, or SNOSE.
87
What is the intention-to-treat (ITT) principle?
Participants are analysed in the groups to which they were originally randomised, regardless of compliance.
88
What is the rationale for ITT analysis?
It estimates the real-world impact of allocating an intervention rather than ideal adherence.
89
How does ITT affect treatment effect estimates?
It provides a conservative estimate and enhances generalisability.
90
When may ITT be less suitable?
When investigating pure biological effects of a drug.
91
How does ITT influence equivalence or non-inferiority trials?
It may favor equivalence due to dilution of differences.
92
What challenge does attrition pose in ITT analysis?
Missing outcome data for participants lost to follow-up.
93
What is LOCF (Last Observation Carried Forward)?
A method where the last recorded value is used for subsequent missing observations.
94
What are key limitations of LOCF?
Ignores natural remission and may bias results.
95
What is worst-case scenario analysis?
Assuming all lost participants experienced treatment failure.
96
What are common data imputation methods?
Mean substitution, hot deck imputation, regression-based prediction, growth curve analysis, random effects modelling, and multiple imputation.
97
What is per-protocol (PP) analysis?
Analysis limited to participants who adhered sufficiently to the protocol.
98
What is treatment-received (TR) analysis?
Analysis based on actual treatment received rather than assigned group.
99
What is the default analytical approach in RCTs?
Intention-to-treat (ITT) analysis.
100
When should an N-of-1 trial be considered?
When treatment benefit, optimal dose, or adverse effects are uncertain for an individual patient.
101
When is an N-of-1 trial feasible?
When patient is cooperative, treatment effects are rapid and reversible, outcomes are measurable, and blinding is possible.
102
How is an N-of-1 trial conducted?
Double-blinded cycles of active and control treatment with random allocation and repeated outcome measurement.
103
What is a key limitation of N-of-1 trials?
Results cannot be generalised to other patients.
104
What is the primary value of an N-of-1 trial?
Optimising therapy for an individual patient.
105
What are Koch's postulates intended to establish?
Causality between a microorganism and an infectious disease.
106
Why are Koch's postulates limited?
They are not generalisable to non-infectious diseases.
107
What is the purpose of Bradford Hill's criteria?
To assess causal relationships in epidemiology.
108
What is Susser's key contribution to causality assessment?
Emphasis on directionality, time order, and association as essential criteria.
109
What is Rothman's sufficient-component cause model?
Disease results from a complete set of component causes forming a sufficient cause.
110
What is a necessary cause?
A component present in every sufficient cause for an outcome.
111
What is the 'web of causation'?
A model illustrating complex interactions among multiple risk factors.
112
What is a confounder?
A third variable that distorts the true association between exposure and outcome.
113
What criteria must a confounder meet?
Association with exposure, association with outcome, not on causal pathway, and unequal distribution between groups.
114
How can confounding be controlled?
Restriction, matching, randomisation, stratification, and multivariate analysis.
115
What is an effect modifier?
A variable that changes the strength of an exposure-outcome relationship.
116
How much does confounding affect study validity?
It threatens internal validity by introducing bias.
117
What defines internal validity?
Accuracy of results in reflecting the truth.
118
What defines external validity?
Applicability of results to other settings or populations.
119
What are the main threats to study validity?
Random error, systematic error (bias), and confounding.
120
What is random error?
Chance-related imprecision due to sampling variability.
121
What is systemic error (bias)?
Predictable error from flawed design, measurement, or analysis.
122
Can bias be corrected at the analysis stage?
No.
123
What are the main stages at which bias can occur?
Participant selection, measurement, and data analysis.
124
What is selection bias?
Systematic differences between study groups unrelated to exposure.
125
What is measurement bias?
Non-uniform data collection across groups.
126
What is analysis bias?
Bias arising from contamination or attrition during analysis.
127
When is qualitative research appropriate?
When phenomena cannot be meaningfully quantified.
128
What questions does qualitative research address?
Why and how people think, behave, and experience events.
129
What are the main types of qualitative research designs?
Phenomenology, ethnography, grounded theory, case study, and participant action research.
130
What are the primary data collection methods in qualitative research?
Interviews, observation, and document analysis.
131
What is coding in qualitative analysis?
Transforming raw data into categories or themes.
132
What methods assess validity in qualitative research?
Triangulation, respondent validation, reflexivity, and deviant case analysis.
133
What is secondary research?
Research conducted using data from previously published studies rather than recruiting individual patients.
134
What is recruited in secondary research instead of patients?
Eligible studies.
135
What are 2 major types of secondary research?
Narrative reviews and systematic reviews.
136
Who typically conducts narrative reviews?
Experts in the relevant field.
137
How are narrative reviews shaped?
They are aligned with the author's opinions, supported by selectively chosen evidence.
138
How focussed is the clinical question in narrative reviews?
Broad and nonspecific.
139
Why are narrative reviews prone to bias?
Literature searching is unsystematic and subjective.
140
Are narrative reviews reproducible?
No, except potentially by the same author.
141
What are the strengths of narrative reviews?
They introduce topics, stimulate interest and controversy, and generate new hypotheses.
142
What defines a systematic review?
A review using rigorous, predefined methods to identify, appraise, and synthesise evidence.
143
What type of research question does a systematic review address?
A narrow, focussed question.
144
What are the core characteristics of systematic reviews?
Focussed question, comprehensive search, uniform selection criteria, and quantitative analysis where possible.
145
Is quantitative synthesis mandatory in systematic reviews?
No, it is optional if data are not combinable or of insufficient quality.
146
What is a meta-analysis?
Statistical combination of data from individual studies within a systematic review.
147
What does the GIGO principle mean?
'Garbage in, garbage out', or poor quality studies lead to meaningless results.
148
What factors determine the quality of individual trials?
Patient samples, outcomes, follow-up duration, treatment comparability, and methodology.
149
Which methodological factors are critical for trial quality?
Adequate randomisation, allocation concealment, intention-to-treat (ITT) analysis, and blinded outcome assessment.
150
What databases should be used in a proper literature search?
Multiple databases, not just MEDLINE or CINAHL.
151
Why is reference cross-checking important?
To identify studies missed by database searches.
152
What is meant by hand searching?
Manual identification of studies not indexed electronically.
153
Why consult experts during a literature search?
To identify missing or unpublished studies.
154
What is gray literature?
Unpublished material such as conference abstracts and posters.
155
What study characteristics must inclusion criteria specify?
Study design, subjects, publications, language, interventions, and time frame.
156
What is the primary use of meta-analysis?
Assessing clinical effectiveness of healthcare interventions.
157
Which study design is most commonly included in meta-analyses?
Randomised controlled trials (RCTs).
158
What types of data can meta-analyses use?
Summary data or patient data.
159
Why is individual patient data meta-analysis uncommon?
High cost and difficulty obtaining original datasets.
160
What are the 4 basic steps of a meta-analysis?
Literature search, selection criteria, data extraction, and statistical analysis.
161
How are results combined in meta-analysis?
Using weighted averages.
162
What determines study weighting?
Sample size, precision, external validity, and methodological quality.
163
Is weighting influenced by study outcomes?
No.
164
What does a fixed effects model assume?
All studies share a single common treatment effect.
165
What does a random effects model assume?
True treatment effects vary across studies.
166
Which model gives more weight to smaller studies?
Random effects model.
167
What is a limitation of random effects models?
Greater susceptibility to publication bias and wider confidence intervals (CIs).
168
When do fixed and random effects models give similar results?
When heterogeneity is absent.
169
What is heterogeneity in meta-analysis?
Variation in results across included studies.
170
What is clinical heterogeneity?
Differences in interventions, patients, severity, or treatment duration.
171
is clinical heterogeneity measurable?
No, it is describable but not quantifiable.
172
What is methodological heterogeneity?
Differences due to study design or conduct.
173
What is statistical heterogeneity?
Inconsistency across effect sizes or direction across studies.
174
How can heterogeneity be assessed visually?
Forest plots, L’Abbé plots, and Galbraith plots.
175
What does a L’Abbé plot display?
Control event rate vs experimental event rate for each study.
176
What statistical tests assess heterogeneity?
Cochran's Q and I² statistics.
177
What does the I² statistic measure?
Percentage of variability due to heterogeneity rather than chance.
178
What is publication bias?
Preferential publication of statistically significant results.
179
Why are small studies more prone to publication bias?
Non-significant small studies are less likely to be published.
180
What is a funnel plot used for?
Detecting asymmetry suggestive of publication bias.
181
What is fail-safe N?
Number of unpublished null studies needed to negate meta-analysis findings.
182
What is sensitivity analysis?
Testing how results change when assumptions or inputs are altered.
183
Why is sensitivity analysis important?
To assess robustness of findings under uncertainty.
184
What does each square or blob represent in a forest plot or blobbogram?
Effect estimate from an individual study.
185
What does the horizontal line represent in a forest plot?
Confidence interval and precision.
186
What does the diamond represent in a forest plot?
Pooled effect estimate.
187
What effect size measures are used for binary outcomes?
Odds ratio (OR) and relative risk (RR).
188
When can odds ratios overestimate effects?
Especially when event rates exceed 20%.
189
What effect size is preferred in Cochrane reviews?
Relative risk (RR).
190
What are effect size measures for continuous data?
Mean difference and standardised mean difference.
191
What is the purpose of economic analysis?
To inform decision-makers about cost-effectiveness and benefits.
192
What are the 5 main types of economic evaluations?
Cost minimisation, cost effectiveness, cost utility, cost benefit, and cost consequences.
193
What are direct costs?
Costs of treatment delivery such as staff, drugs, and facilities.
194
What are indirect costs?
Productivity losses and caregiver-related expenses.
195
What are intangible costs?
Pain, suffering, and stigma.
196
What is opportunity cost?
Value of what is given up to choose something else/'best alternative foregone'.
197
What is ICER?
Incremental cost-effectiveness ratio.
198
How is ICER calculated?
Difference in costs divided by difference in effects.
199
What determines if an intervention is cost-effective?
Whether ICER falls below the decision-maker's willingness-to-pay threshold.
200
What does psychiatric epidemiology study?
Frequency, risk factors, and burden of psychiatric disorders.
201
What is descriptive epidemiology?
Description of disease patterns.
202
What is analytical epidemiology?
Hypothesis-driven examination of associations.
203
What is incidence?
Number of new cases in a population over a specified time.
204
What is prevalence?
Number of existing cases at a point or period in time.
205
How are incidence and prevalence related?
Prevalence = incidence x duration
206
What is crude mortality rate?
All-cause deaths divided by total population.
207
What is case-fatality rate?
Proportion of affected individuals who die from the disease.
208
What is YPLL?
YPLL = Year of Potential Life Lost
Measure of premature mortality weighted towards younger deaths.
209
What is a DALY?
DALY = Disability-Adjusted Life Year
One year of healthy life lost due to death or disability.
210
What is quality improvement (QI)?
A scientific, iterative approach to improving healthcare quality and safety.
211
How does QI differ from Quality Assurance (QA)?
QI is proactive and improvement-focussed; QA is retrospective and fault-focussed.
212
What is the core QI cycle?
Plan-Do-Study-Act (PDSA).
213
What is the Model for Improvement (MFI)?
A widely used NHS framework that extends PDSA by requiring reflective questions before initiating improvement cycles.
214
How does MFI relate to PDSA?
MFI builds on PDSA by emphasising preparatory reflection prior to testing change.
215
What is Root Cause Analysis (RCA)?
A retrospective quality method conducted after a sentinel event to identify causal factors underlying performance variation.
216
What type of events typically trigger an RCA?
Sentinel events, usually adverse events such as inpatient suicide.
217
What is the primary aim of RCA?
To identify putative fundamental causes explaining why a specific adverse event occurred.
218
What are the main limitations of RCA?
Focus on individual events, hindsight bias, and limited applicability to system-wide improvement.
219
What is Failure Modes and Effects Analysis (FMEA)?
A prospective method for identifying potential system failures that could lead to patient harm.
220
How does FMEA differ from RCA and SEA?
FMEA is conducted prospectively, whereas RCA and SEA (Significant Event Auditing) are conducted after adverse events.
221
What key question does FMEA address?
'How could the system fail?'
222
What key question does RCA address?
'Why did the system fail?'
223
What is the focus of FMEA?
Processes within a system rather than individual adverse events.
224
What are the 5 steps of FMEA?
Team selection, process identification, process flow diagram, failure mode identification and risk prioritisation, and action planning.
225
What is a failure mode?
A weak point in a system that can break down and adversely affect outcomes.
226
What features must be appraised when evaluating a QI project?
Defined change with predicted direction, multiple iterations, small-scale testing, understanding temporal variation, and documentation.
227
What is the SQUIRE statement?
A 19-item reporting checklist for formal healthcare quality improvement studies.
MRCPsych Paper B
flashcards
Decks in class (15)
# Cards
-
Research Methods227
-
Advanced Statistics229
-
Evidence-Based Medicine96
-
Psychiatric Services46
-
Adult Psychiatry209
-
Emergency Psychiatry24
-
Liaison Psychiatry13
-
Perinatal Psychiatry23
-
Old Age Psychiatry45
-
Psychotherapy86
-
Child Psychiatry79
-
Addiction Psychiatry148
-
Forensic Psychiatry81
-
Intellectual Disability79
-
Epidemiological Data4
