Study Designs: Considerations and Application

Study Designs: Considerations and Application Sabrina W. Cole, Pharm.D., BCPS Intermountain Healthcare Salt Lake City, Utah

Study Designs: Considerations and Application

Study Designs: Considerations and Application Sabrina W. Cole, Pharm.D., BCPS Intermountain Healthcare Salt Lake City, Utah

ACCP Updates in Therapeutics® 2015: Pediatric Pharmacy Preparatory Review Course 1-49


Learning Objectives 1. Distinguish between observational and experimental study designs. 2. Describe the differences between superiority, equivalence, and noninferiority study designs. 3. Describe the differences between parallel and crossover study designs. 4. Determine the utility, benefits, and limitations of various study designs (e.g., case study, case-control, cohort, cross-sectional, randomized controlled clinical trials, systematic review, meta-analysis). 5. Reconstruct the hierarchy of evidence for various study designs. 6. Define the concepts of internal and external validity, randomization, bias, and confounding, and describe the advantages and limitations of each. 7. Differentiate between the types of data, including continuous, nominal, and ordinal. 8. Define estimates of effect including relative risk (RR), relative risk reduction (RRR), absolute risk reduction (ARR), number needed to treat (NNT), and number needed to harm (NNH), and given a data set, calculate and interpret the measure.

Self-Assessment Questions Answers and explanations to these questions can be found at the end of this chapter. 1. A study was designed to evaluate the efficacy and safety of risperidone long-acting injectable (LAI) for preventing the recurrence of mood episodes in patients with type I bipolar disorder. After a 12-week open-label period with risperidone LAI, patients who did not experience a recurrence entered an 18-month randomized, double-blind period with risperidone LAI, placebo, or oral olanzapine. The primary end point was time to recurrence of any mood episode for risperidone LAI versus placebo in the study period. Which type of data is being evaluated? A. Interval. B. Ordinal. C. Nominal. D. Ratio.

2. Identify the trial design from the following description: It is hypothesized that proton pump inhibitors (PPIs) contribute to development of hospital-onset Clostridium difficile. A study was conducted to test this hypothesis. Medical charts from a group of patients with C. difficile were compared with medical charts from a group of patients without C. difficile. The groups were matched by age, sex, date of admission, and other confounding factors, such as antibiotic use. The use of PPIs in each group was assessed and compared. A. Cohort study. B. Case-control study. C. Randomized controlled trial. D. Meta-analysis. 3. Identify the trial design from the following description: After the 1986 accident at the Chernobyl nuclear power plant in the Soviet Union, epidemiologists examined differences in the frequency of thyroid tumors between children living downwind of the reactor and children living upwind of the reactor. During the 5 years after the accident, the frequency of thyroid tumors increased only among children exposed to the radiation. A. Cohort study. B. Case-control study. C. Randomized controlled trial. D. Meta-analysis. 4. Identify the study design from the following description: A paper documents emerging posttraumatic obsessive-compulsive disorder in 13 Israeli military veterans with a diagnosis of both obsessive-compulsive disorder (OCD) and posttraumatic stress disorder, for whom the onset of OCD was clearly associated with the trauma. The data presented include four detailed patient histories that delineate the relationship between the symptomatologies of the two disorders. The clinical and theoretical implications of the data are discussed. A. Case control. B. Case report. C. Case series. D. Cross-sectional.



5. A study is conducted of a new oral chemotherapy agent (group A) compared with an existing oral chemotherapy agent (group B). Patients are assigned to group A or B and are unaware of treatment assignment. The primary outcome is tumor regression, as assessed by a radiologist. All study investigators are unaware of the treatment assignment. This study is: A. Single-blinded. B. Double-blinded. C. Triple-blinded. D. Double-dummy. 6. Identify the trial design from the following description: A study was conducted to estimate the incidence of rhabdomyolysis in patients treated with different statins and fibrates, alone and in combination, in the ambulatory setting. Claims data from managed care health plans were used to identify statin and fibrate users. Which type of study design is this? A. Cohort study. B. Case-control. C. Randomized controlled trial. D. Meta-analysis. 7. Identify the trial design from the following description: A literature review was conducted and a study was performed on data retrieved from studies that followed patients with inflammatory bowel disease (IBD) who received immunosuppressive therapy for more than 1 year. The incidence of newly developed malignancy was documented. Then, a literature search was performed using MEDLINE and the Cochrane Library. Nine cohort studies met the inclusion criteria for this study. Analysis of these studies showed no discernible difference in the incidence of any type of malignancy in patients with IBD who received immunosuppressive therapy compared with those who did not receive immunosuppressants. A. Cohort study. B. Case-control. C. Randomized controlled trial. D. Meta-analysis.

8. A 5-year parallel, randomized, double-blind study compared metformin with glipizide in patients with type 2 diabetes mellitus. In this study, an intentionto-treat analysis would: A. Be appropriate to clearly establish the effect of completing therapy with the medications. B. Be appropriate to estimate the effects of the two medications in a clinical setting. C. Not be appropriate because it would overestimate the efficacy of the medication. D. Not be appropriate because this type of analysis excludes subjects with protocol violations. 9. Investigators conducted a clinical trial to evaluate the impact of nebulized hypertonic saline compared with nebulized 0.9% normal saline on admission rates and length of stay in infants with bronchiolitis. Which strategy performed at study enrollment ensures that subjects have an equal chance of receiving the active intervention? A. Blinding. B. Exclusion criteria. C. Randomization. D. Inclusion criteria. 10. Investigators conducted a trial to evaluate the optimal blood glucose concentrations in critically ill patients. Within 24 hours of admission to an intensive care unit (ICU), adults who were expected to require treatment in the ICU for 3 or more consecutive days were randomly assigned to undergo either intensive glucose control, with a target blood glucose range of 81–108 mg/dL, or conventional glucose control, with a target of 180 mg/dL or less. The primary end point was defined as death from any cause within 90 days after randomization. All patient outcomes were included in the final analysis, even if there were deviations from the protocol. Which type of control group was used in this study? A. Placebo. B. Active. C. Historical. D. Open-label.



11. Investigators conducted a trial to evaluate the optimal blood glucose concentrations in critically ill patients. Within 24 hours of admission to an ICU, adults who were expected to require treatment in the ICU for 3 or more consecutive days were randomly assigned to undergo either intensive glucose control, with a target blood glucose range of 81–108 mg/dL, or conventional glucose control, with a target of 180 mg/dL or less. The primary end point was defined as death from any cause within 90 days after randomization. All patient outcomes were included in the final analysis, even if there were deviations from the protocol. Which type of variable is the primary outcome in this study? A. Continuous. B. Ordinal. C. Nominal. D. Interval. 12. A study was designed to assess the effect of hypertonic saline on mucus flow in infants with acute bronchitis and evaluate a change in bronchiolitis severity score. Patients were randomized to receive either 7% saline or 0.9% saline, both with epinephrine. Which type of study is this? A. Crossover. B. Parallel. C. Cohort. D. Case-control. 13. Which type of analysis risks underestimating the treatment effect observed in a clinical trial? A. Intention-to-treat. B. As-treated. C. Per-protocol. D. Historical.



I. INTRODUCTION A. Pediatric Pharmacy Specialty Content Outline 1. Domain 1: Patient Management (58% of examination) a. Pathophysiology, epidemiology, risk factors, diagnosis, prevention, and evidence-based treatment of common diseases and conditions in pediatric patients 2. Domain 3: Information Management and Education (18% of examination) a. Knowledge of: i. Research design, methodology, and statistical analysis ii. Clinical application and limitations of published data and reports iii. Medical literature publication and review process iv. Opportunities for disseminating pediatric knowledge and scholarly activity (e.g., presentations, manuscripts, newsletters, abstracts, posters) v. Appropriate pediatric-specific references B. Why Read and Evaluate? 1. Clinical experience can be misleading. 2. The FDA (U.S. Food and Drug Administration) is variable in its requirements for approval. 3. Learn from the experience of others. 4. Stay current in your practice area. C. Common Problems Encountered in the Literature 1. Flawed study design 2. Invalid statistical analysis 3. Fraud, deception, misrepresentation 4. Unintentional errors 5. Poorly conducted research 6. Poorly written manuscripts 7. Data dredging D. Primary Goals of Interpretation 1. Establish the significance or importance of the trial (“This is where we were…”). 2. Relate the results to the original objectives of the trials (“This is what we know now…”). 3. Compare data from the trial with data obtained from other trials (“This is what it means…”)

II. OVERVIEW OF STUDY DESIGNS A. Perspectives 1. Prospective: Followed forward over time 2. Retrospective: Reviewed back over time B. Descriptive Studies 1. Quantitative reports that describe naturally occurring events 2. Document and communicate experience 3. Begin to search for explanations 4. Do not provide detailed explanations of the cause of disease or offer the type of evidence needed to evaluate the efficacy of new treatments



5. Examples a. Case reports b. Case series 6. For example, descriptive studies can alert us to possible adverse effects or unusual complications of disease, and the observations noted often lead to further confirmatory work. C. Explanatory Studies 1. Examine etiology, efficacy, cause using a comparison strategy 2. Parallel versus crossover design 3. Parallel a. Subjects receive either the study medication or the control medication throughout the study. b. Two or more groups receive treatment at the same time. Placebo Treatment arm –I Treatment arm –II

4. Crossover a. Each subject serves as his or her own control. b. Subject receives both the study and the control treatments. c. Washout period Wash out period Placebo Placebo Period –I

Period –II

Active Active Treatment Treatment

5. Experimental a. Evaluate the efficacy of an intervention b. Investigators control the allocation of treatment c. Examples: i. Clinical trial ii. Educational intervention



Disease or Outcome (+) Experiment Intervention Study Population Control

Disease or Outcome (–) Disease or Outcome (+) Disease or Outcome (–)

Present

Future

Figure 1. Experimental study design. d. Superiority i. Primary objective of showing that the response to the investigational product is superior to a comparator

Figure 2. Superiority study design.

Committee for Proprietary Medicinal Products [July 2000, cited November 2006]. Points to Consider on Switching Between Superiority and Non-inferiority. Available at www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC500003658.pdf. Accessed February 20, 2015.

e. Equivalence i. Primary objective of showing that the response to two or more treatments differs by an amount that is clinically unimportant ii. Usually demonstrated by showing that the true treatment difference is likely to lie between a lower and an upper equivalence margin of clinically acceptable differences



Figure 3. Equivalence study design.


f. Noninferiority i. Primary objective of showing that the response to the investigational product is not clinically inferior to a comparator product ii. Noninferiority margin - Prespecified maximum allowable difference between the new and the standard treatment

Figure 4. Noninferiority study design.


iii. Considerations for switching the objective (a) Noninferiority to superiority (1) Properly designed and carried out with strict requirements for noninferiority (2) Actual p values for superiority are presented to allow independent assessment (3) Analysis according to intention-to-treat principle (b) Superiority to noninferiority (1) Protocol must contain a prospectively defined noninferiority margin. (2) Intention-to-treat and per-protocol analyses should show similar findings. (3) Trial was properly designed and carried out with requirements of a noninferiority trial. ACCP Updates in Therapeutics® 2015: Pediatric Pharmacy Preparatory Review Course 1-56


6. Observational Investigator observes the natural course of health events. a. Case-control i. Think “disease.” ii. Retrospective iii. Patients with the disease (cases) and without the disease (controls) are compared to determine their exposure to the risk factor in question.

Exposure (+) Cases

Exposure (–) Exposure (+)

Control

Exposure (–) Past

Present

Figure 5. Case-control study design. iv. Strengths (a) Quicker and less expensive than cohort or experimental studies (b) Optimal for evaluating rare diseases (c) Appropriate for evaluating diseases with long latency periods v. Limitations (a) Inefficient for evaluating rare exposures (b) Prone to bias (c) Do not allow for direct calculation of incidence rates (d) Temporal sequence of exposure and outcome is not always known b. Cohort i. Think “exposure.” ii. Retrospective or prospective iii. Design may involve the evaluation of risk factors for disease development in a specific patient population.



Disease or Outcome (+) Risk or Factor (+)

Study Population (free of outcome)

Risk or Factor (+)

Disease or Outcome (–) Disease or Outcome (+) Disease or Outcome (–)

Present

Future

Figure 6. Cohort study design.

Risk Factor (+)

Study Population

Risk Factor (–) Past

Present

Disease or Outcome (+) Disease or Outcome (–) Future

Figure 7. Prospective cohort study design.

Risk Factor (+)

Disease or Outcome (+)

Risk Factor (–)

Disease or Outcome (–) Past

Study Population

Present

Figure 8. Retrospective cohort study design. iv. Strengths (a) Appropriate for evaluating rare exposures (b) Temporal sequence of exposure and outcome is clear. (c) Selection and recall bias is minimized. (d) Allow for direct calculation of incidence rates



v. Limitations (a) Time and cost tend to be more than with case-control. (b) Subjects may be lost to follow-up during the study. (c) Not efficient for evaluating rare diseases c. Cross-sectional i. Also known as prevalence or slice-in-time study ii. Gathers information about the prevalence of health-related states and conditions iii. Does not identify cause-and-effect relationships iv. Identifies existence of health problems simultaneous Risk or Factor (+) Free of disease or outcome Study Population Have disease or outcome

Risk or Factor (–) Risk or Factor (+) Risk or Factor (–)

Present Figure 9. Cross-sectional study design. D. Reasons for Predefining a Trial 1. Ensures that comparator treatments, doses, patient populations, and end points are appropriate 2. Allows sample size estimates based on correct power calculations 3. Ensures that equivalence and noninferiority criteria are chosen, if needed 4. Permits appropriate plans for analysis 5. Ensures sufficient sensitivity to achieve objectives E. Meta-analysis 1. Process of systematically evaluating and combining the results of clinical trials that have been conducted 2. Types of meta-analyses a. Type 1 – Combining raw data to increase study power b. Type 2 – Combining conclusions (e.g., odds ratio [OR]) of single trials to create and overall average odds ratio c. Type 3 – Summating individual results (e.g., stating that 7 of 10 trials support that drug A is more effective compared with placebo) 3. Purpose a. Increase statistical power b. Address research or clinical questions when controversy exists because of conflicting data ACCP Updates in Therapeutics® 2015: Pediatric Pharmacy Preparatory Review Course 1-59


c. Answer research or clinical questions not previously asked in a systematic way d. Better quantity the magnitude of certain responses e. Decrease biases in addressing specific clinical issues by being more objective, rigorous, and systematic f. Ethical considerations preclude the conduct of additional placebo-controlled trials g. Minimize costs of conducting large clinical trials 4. Limitations a. Heterogeneity of clinical trials b. Number of trials does not necessarily move you closer to the truth c. Garbage in equals garbage out 5. Design issues a. Results are highly dependent on criteria for inclusion and exclusion of the previous studies and statistical methods to ensure validity. b. Literature search conducted for inclusion of trials c. How treatment was assigned d. Tests for heterogeneity e. Criteria for pooling data and how they are measured f. Funnel plot: Assesses publication bias g. Forest plot: Summarized study results F. Other 1. N-of-1 2. Survey research 3. Quality-of-life 4. Postmarketing surveillance 5. Pharmacoeconomic G. Strength of Evidence Meta-analysis Randomized Clinical Trial Cohort Study Case-control Study Case Series Case Report Letter to the Editor III. CONSIDERATIONS IN STUDY DESIGN A. Population vs. Sample 1. Population – Every individual in the universe with the specific characteristic(s) or disease state under study 2. Sample – Group of individuals chosen as representatives from the population under study a. Inclusion criteria: Characteristics that the subject must possess to enroll in the study b. Exclusion criteria: Characteristics that prevent study participation c. Patient demographics: Baseline characteristics of patients must be as similar as possible or controlled for in order to reduce confounding effect (e.g., age, concomitant disease state, medications) ACCP Updates in Therapeutics® 2015: Pediatric Pharmacy Preparatory Review Course 1-60


B. Controls 1. Treatment used for comparison in a study 2. Types a. Placebo b. Historical c. Crossover d. Standard treatment e. Within patient comparison f. No control C. Randomization 1. Each subject has an equal and independent chance of being in any of the treatment arms. 2. Decreases, but does not eliminate, the possibility that study and control groups will differ according to factors that affect prognosis 3. All inclusion and exclusion criteria must be met before randomization occurs. 4. Types of randomization a. Simple b. Block (cluster) c. No randomization 5. Stratification D. Blinding 1. Purpose a. Prevents clinicians from assessing/treating one group differently from the other b. To overcome the “placebo effect” c. To ensure equal patient adherence 2. Limitations a. May be difficult to blind certain interventions b. Expensive and time-consuming 3. Types of blinding schemes a. Open-label b. Single-blind c. Double-blind d. Triple-blind e. Double-dummy E. Bias 1. Systematic error that enters a clinical trial and distorts the data obtained 2. Types (examples) a. Selection b. Observation i. Recall ii. Interviewer iii. Misclassification c. Channeling 3. Controlling for bias a. Minimize confounders b. Proper selection of patients c. Objective method(s) of selecting data ACCP Updates in Therapeutics® 2015: Pediatric Pharmacy Preparatory Review Course 1-61


d. Blinding e. Use of a control group f. Reliable data sources F. Types of Analyses 1. Intention-to-treat a. Compares outcomes according to the intended initial subjects’ assignments b. Determines the effect of treatment under usual conditions (e.g., gives a better idea of how the treatment will work in the real world) c. Gives a conservative estimate of the treatment difference d. No data should be eliminated. 2. As-treated a. Analyzes subjects according to which intervention the subjects actually received b. No data should be eliminated. c. Cleaner data regarding method effectiveness compared with intention-to-treat analysis 3. Per-protocol a. Analyzes the subjects who precisely followed the protocol b. Gives superior information regarding method effectiveness compared with other analyses c. Provides a more generous estimate of the differences between treatments d. Problematic if patient adherence is related to prognosis G. Validity 1. Internal validity a. Within the confines of the study i. Methods and analysis used were appropriate. ii. Investigators’ interpretation is supported. iii. Results appear accurate. b. Affected by controls, randomization, blinding, confounders c. Strong trial design = internal validity. 2. External validity a. Generalizability b. Ability to extrapolate study results to the population of interest c. NO internal validity = NO external validity. H. Reliability - Reproducibility of measurements I. Other Methodological Considerations 1. Attrition 2. Interim analyses 3. Early stop rules 4. Adherence assessment J. Variables 1. Dependent variable: Outcome of interest within the study 2. Independent variable: Intervention or what is being manipulated 3. Confounding variable: Factor affecting the patient’s conditions and associated statistically with the intervention being evaluated



K. Study End Points 1. DOE (disease-oriented evidence) a. Measurement of intermediate outcomes in a study, such as a surrogate marker or end point b. Aimed at increasing understanding of a disease and its incidence, prevalence, diagnosis, treatment, and prognosis i. Intermediate outcomes ii. Surrogate markers iii. Disease markers 2. POE (patient-oriented evidence) Outcomes that patients care most about: a. Mortality b. Morbidity c. Quality of life

IV. TYPES OF DATA A. Continuous Variables 1. Data with constant and defined units of measure 2. Equal distance between increments a. Interval: Arbitrary zero point assigned (e.g., degrees Fahrenheit) b. Ratio: Absolute zero point (e.g., blood pressure, degrees Kelvin, heart rate) c. Note: Interval and ratio variables are analyzed using the same statistical procedures. B. Ordinal 1. Limited number of categories that DO have an implied rank or order 2. Order of numbers is meaningful, but the distance between categories is not equal. 3. For example, Likert-scale questionnaires, Glasgow Coma Scale C. Nominal 1. Named categories that have no implied rank or order 2. No mathematical relationship between classifications 3. Subjects are categorized on the basis of some mutual characteristics, which the entire group possesses. 4. For example, sex; hair color; the presence or absence of disease

V. ESTIMATES OF EFFECT A. Also called Measures of Association B. Helpful When Determining the Clinical Applicability of Trial Results C. Methods for Assessing Treatment Effect and the Importance of the Results to the Study Sample and Target Population Outcome (yes)

Outcome (no)

Total

Exposure (experimental group)

A

B

A+B

No exposure (control group)

C

D

C+D



D. Risk Ratio (RR) 1. Ratio of risk of an event occurring in the experimental group compared with the control group 2. Probability that a condition will occur 3. Interpretation a. Value less than 1 = therapy decreased the risk. b. Value = 1 = no treatment difference between treatments. c. Value greater than 1 = therapy increased the risk. 4. RR = [A/(A + B)]/[C/(C + D)]. E. Odds Ratio 1. Ratio of the odds of an event occurring in the experimental group compared with in the control group 2. Association with retrospective studies or case-control studies 3. OR = A*D/C*B. F. Relative Risk Reduction (RRR) 1. Estimates the percentage of baseline risk that is removed as a result of therapy 2. Percent reduction in the experimental group event rate compared with the control group event rate 3. RRR = 0 indicates that there is no effect of the treatment compared with control. 4. RRR = 1-RR*100 or RRR = % placebo − % treatment/% placebo. G. Absolute Risk Reduction (ARR) 1. The difference in the event rate between a control group and an experimental group 2. Provides the percentage of patients spared the events as a result of the treatment 3. ARR = 0 indicates no difference between comparison groups 4. ARR = % control group − % treatment group. H. Number Needed to Treat (NNT) 1. Number of patients who require treatment to prevent one event 2. Assumes the baseline risk is the same for all patients 3. Does not extrapolate beyond the study end points 4. Reciprocal of the ARR 5. NNT = 1/ARR. I. Number Needed to Harm (NNH) 1. Number of people needed to be at risk for a defined period for an adverse effect to occur with one person 2. Reciprocal of the absolute risk increase with a drug adverse effect



REFERENCES 1. Berensen NM. Overview of statistics. In: Chumney ECG, Simpson KN. Methods and Designs for Outcomes Research. Bethesda, MD: American Society of Health-System Pharmacists, 2006. 2. DeYoung G. Biostatistical applications. In: DeYoung G, Klepser T, Smith C, et al., eds. Updates in Therapeutics: The Pharmacotherapy Preparatory Course, 2006 ed. Kansas City, MO: American College of Clinical Pharmacy, 2006:(III)1-(III)13. 3. DeYoung G. Biostatistics: a refresher. In: DeYoung G, Klepser T, Smith C, et al., eds. Updates in Therapeutics: The Pharmacotherapy Preparatory Course, 2006 ed. Kansas City, MO: American College of Clinical Pharmacy, 2006:(IV)1-(IV)19. 4. DeYoung G. Clinical trial design. In: DeYoung G, Klepser T, Smith C, et al., eds. Updates in Therapeutics: The Pharmacotherapy Preparatory Course, 2006 ed. Kansas City, MO: American College of Clinical Pharmacy, 2006:(V)1-(V)19. 5. Gehlbach SH. Interpreting the Medical Literature. New York: McGraw-Hill, 2006. 6. Lang T, Secic M. How to Report Statistics in Medicine: Annotated Guidelines for Authors, Editors, and Reviewers, 2nd ed. Philadelphia: American College of Physicians, 2006. 7. Mansfield L. The reading, writing, and arithmetic of the medical literature, part 2: critical evaluation of statistical reporting. Ann Allergy Asthma Immunol 2005;95:315-22. 8. Riegelman RK. Studying a Study & Testing a Test. Baltimore: Lippincott Williams & Wilkins, 2013.



ANSWERS AND EXPLANATIONS TO SELF-ASSESSMENT QUESTIONS 1. Answer: D Continuous data have a constant and defined unit of measure, with an equal distance between increments, and include both interval and ratio data. Ratio data have an absolute zero point (e.g., time, as in this question) (Answer D is correct). Interval data have an arbitrary zero point assigned (e.g., degrees Fahrenheit) (Answer A is incorrect). Ordinal data have an implied rank or order, but they are limited in the number of categories (Answer B is incorrect). Nominal data do not have an implied rank or order and are categorical (Answer C is incorrect). 2. Answer: B Case-control studies are observational studies that serve to compare patients with a disease or outcome to those without the disease or outcome to evaluate the effect of an exposure to a risk factor (Answer B is correct). Although a cohort study is similar to a case-control study in its observational approach to examination, a cohort study is used when subjects are selected on the basis of their exposure to a risk factor (Answer A is incorrect). A randomized controlled trial evaluates the impact of an intervention (Answer C is incorrect). A meta-analysis includes the results from several trials (Answer D is incorrect). 3. Answer: A A cohort study is appropriate when subjects are selected on the basis of their exposure to a risk factor (e.g., radiation, as in this case) (Answer A is correct). Although casecontrol studies are observational, they serve to compare patients with a disease or outcome to those without the disease or outcome to evaluate the effect of exposure to a risk factor (Answer B is incorrect). A randomized controlled trial evaluates the impact of an intervention (Answer C is incorrect). A meta-analysis includes the results from several trials (Answer D is incorrect). 4. Answer: C A case series is a descriptive, not an experimental, design that allows documentation and communication of an experience of several patients (Answer C is correct). A case report describes a single patient (Answer B is incorrect). A case-control design is an explanatory design appropriate for examining the etiology, efficacy, and cause using a comparison strategy (Answer A is incorrect). A cross-sectional study is appropriate for gathering

information relevant to the prevalence of health-related states and conditions (Answer D is incorrect). 5. Answer: C In a triple-blind design, the subject, investigator, and third party (e.g., the radiologist) are all blinded to treatment assignment (Answer C is correct). Single-blind and double-blind designs are appropriate for the blinding of one or two groups (e.g., subjects, investigators) in a study, respectively (Answers A and B are incorrect). A doubledummy design is appropriate for comparing treatments available in different dosage forms (e.g., oral dosage form vs. subcutaneous injection) to blind the treatment allocation (Answer D is incorrect). 6. Answer: A A cohort study is appropriate when subjects are selected on the basis of their exposure to a risk factor (e.g., medication exposure, as in this case) (Answer A is correct). Although case-control studies are observational, they serve to compare patients with a disease or outcome to those without the disease or outcome to evaluate the effect of an exposure to a risk factor (Answer B is incorrect). A randomized controlled trial evaluates the impact of an intervention (Answer C is incorrect). A meta-analysis includes the results from several trials (Answer D is incorrect). 7. Answer: D A meta-analysis describes the systematic evaluation of the results of a group of clinical trials and the combining of the results (Answer D is correct). A cohort study is appropriate when subjects are selected on the basis of their exposure to a risk factor (Answer A is incorrect). Case-control studies are observational studies that compare patients with a disease or outcome to patients without the disease or outcome to evaluate the effect of an exposure to a risk factor (Answer B is incorrect). A randomized controlled trial evaluates the impact of an intervention (Answer C is incorrect). 8. Answer: B An intention-to-treat analysis compares an outcome according to the intended initial subjects’ assignments and determines the effect of a treatment under usual conditions (Answer B is correct). A per-protocol analysis is appropriate to clearly establish the effect of completing



therapy with the medications as prescribed because only data from subjects who precisely followed the protocol are included in the final analysis (Answer A is incorrect). An intention-to-treat analysis provides a conservative estimate of the true treatment difference (Answer C is incorrect). Intention-to-treat analyses include all data points, whereas per-protocol analyses exclude subjects from the final analysis if there are protocol violations (Answer D is incorrect). 9. Answer: C Randomization is used to ensure that each subject has an equal chance of being in any of the treatment arms (Answer C is correct). Blinding is used to prevent the placebo effect and to reduce investigator bias from assessing/treating one patient group differently from the other (Answer A is incorrect). Inclusion and exclusion criteria are defined to select the most appropriate patient population for the study (Answers B and D are incorrect). 10. Answer: B An active control is used in this study, given that all patients received active treatment (Answer B is correct). A placebo control is an inactive comparator (Answer A is incorrect). A historical control is selected from a comparison group that was observed in the past or can be evaluated from patient records (Answer C is incorrect). Open-label does not describe which type of control used the study; instead, it describes the lack of blinding of treatment assignment (Answer D is incorrect).

or her own control and receives both the study and control treatments throughout (Answer A is incorrect). Cohort and case-control are both observational study types and do not incorporate the use of an intervention (Answers C and D are incorrect). 13. Answer: A An intention-to-treat analysis compares outcomes according to the subjects’ intended initial assignments and gives a conservative estimate of the true difference, but provides a better idea of how the treatment will perform in clinical practice (Answer A is correct). An as-treated analysis compares outcomes according to the treatment subjects received in the study and provides a more accurate estimate of the true treatment difference compared with intention-to-treat (Answer B is incorrect). A per-protocol analysis compares outcomes according to subjects who precisely followed the study protocol and provides the most accurate estimate of the true treatment difference (Answer C is incorrect). Historical is a type of control, not a type of study analysis (Answer D is incorrect).

11. Answer: C Nominal data are named categories with no implied rank of order—in this example, death from any cause or survival (Answer C is correct). Continuous data have a constant and defined unit of measure, with an equal distance between increments (Answer A is incorrect). Ordinal data have an implied rank or order but are limited in the number of categories (Answer B is incorrect). Interval data are continuous with a predetermined order and an arbitrary zero point assigned (Answer D is incorrect). 12. Answer: B Parallel and crossover describe study perspectives. A parallel design, as in this case, describes a study in which the subject receives either the study treatment or the control treatment throughout the study (Answer B is correct). By contrast, in a crossover design, each subject serves as his ACCP Updates in Therapeutics® 2015: Pediatric Pharmacy Preparatory Review Course 1-67