#100 Epidemiological Study Designs

(Core vs Variant )

#I. Observational Studies

#A. Descriptive Studies

1. Case report — 🔵 CORE

2. Case series — 🔵 CORE

3. Descriptive cross-sectional (prevalence) study — 🔵 CORE

4. Time–place–person descriptive study — 🟡 VARIANT

5. Descriptive ecological study — 🟡 VARIANT

#B. Analytical Observational Studies

1) Cross-Sectional / Correlational

6. Analytical cross-sectional study — 🔵 CORE

7. Correlational study — 🔵 CORE

8. Ecological study — 🔵 CORE

9. Multilevel cross-sectional study — 🟡 VARIANT

10. School-/institution-based cross-sectional study — 🟡 VARIANT

2) Case-Control and Related Designs

11. Classical case-control study — 🔵 CORE

12. Matched case-control study — 🟡 VARIANT

13. Unmatched case-control study — 🟡 VARIANT

14. Test-positive / test-negative case-control — 🟡 VARIANT

15. Nested case-control study — 🟡 VARIANT

16. Case-cohort study — 🟡 VARIANT

17. Two-stage case-control study — 🟡 VARIANT

18. Case–case study — 🟡 VARIANT

19. Case–population study — 🟡 VARIANT

3) Cohort / Longitudinal Studies

20. Prospective cohort study — 🔵 CORE

21. Retrospective cohort study — 🔵 CORE

22. Ambispective cohort study — 🟡 VARIANT

23. Fixed (closed) cohort — 🟡 VARIANT

24. Dynamic (open) cohort — 🟡 VARIANT

25. Historical cohort study — 🟡 VARIANT

26. Birth cohort study — 🟡 VARIANT

27. Occupational cohort study — 🟡 VARIANT

4) Self-Controlled / Time-Based Designs

28. Case-crossover study — 🔵 CORE

29. Self-controlled case series (SCCS) — 🔵 CORE

30. Case-time-control study — 🟡 VARIANT

31. Sequence symmetry analysis — 🟡 VARIANT

32. Self-controlled risk interval — 🟡 VARIANT

33. Time-trend study — 🟡 VARIANT

34. Longitudinal ecological study — 🟡 VARIANT

35. Panel / repeated-measures study — 🟡 VARIANT

#II. Experimental / Interventional Studies

#A. Randomized Designs

36. Randomized controlled trial (RCT) — 🔵 CORE

37. Clinical trial — 🔵 CORE

38. Field trial — 🔵 CORE

39. Community intervention trial — 🔵 CORE

40. Cluster randomized trial — 🟡 VARIANT

41. Stepped-wedge cluster RCT — 🟡 VARIANT

42. N-of-1 trial — 🟡 VARIANT

43. Factorial randomized trial — 🟡 VARIANT

#B. Quasi-Experimental / Natural Experiments

44. Non-randomized controlled trial — 🔵 CORE

45. One-group pre-test–post-test study — 🔵 CORE

46. Before–after study (no control) — 🔵 CORE

47. Controlled before–after study — 🔵 CORE

48. Interrupted time-series (ITS) study — 🔵 CORE

49. ITS with control — 🟡 VARIANT

50. ITS with multiple interruptions — 🟡 VARIANT

51. Regression discontinuity design — 🟡 VARIANT

52. Regression kink design — 🟡 VARIANT

53. Difference-in-differences study — 🟡 VARIANT

54. Synthetic control study — 🟡 VARIANT

55. Natural experiment study — 🟡 VARIANT

#III. Diagnostic, Screening & Measurement Designs

56. Diagnostic accuracy study — 🔵 CORE

57. Screening trial — 🔵 CORE

58. Validation study — 🟡 VARIANT

59. Reliability / reproducibility study — 🟡 VARIANT

60. Method comparison study — 🟡 VARIANT

61. ROC curve analysis study — 🟡 VARIANT

#IV. Genetic, Family & Special Designs

62. Twin study — 🔵 CORE

63. Family aggregation study — 🔵 CORE

64. Adoption study — 🔵 CORE

65. Genetic association study (GWAS) — 🟡 VARIANT

66. Phenome-wide association study (PheWAS) — 🟡 VARIANT

67. Mendelian randomization study — 🟡 VARIANT

#V. Implementation, Policy & Evaluation Designs

68. Implementation research study — 🔵 CORE

69. Process evaluation study — 🟡 VARIANT

70. Outcome evaluation study — 🟡 VARIANT

71. Impact evaluation study — 🟡 VARIANT

72. Policy evaluation study — 🟡 VARIANT

73. Realist evaluation — 🟡 VARIANT

74. Participatory action research — 🟡 VARIANT

#VI. Surveillance & Data-Based Designs

75. Surveillance study — 🔵 CORE

76. Passive surveillance — 🟡 VARIANT

77. Active surveillance — 🟡 VARIANT

78. Sentinel surveillance — 🟡 VARIANT

79. Syndromic surveillance — 🟡 VARIANT

80. Event-based surveillance — 🟡 VARIANT

81. Registry-based study — 🔵 CORE

82. Administrative/database study — 🟡 VARIANT

#VII. Evidence Synthesis Designs

83. Systematic review — 🔵 CORE

84. Meta-analysis — 🔵 CORE

85. Scoping review — 🟡 VARIANT

86. Umbrella review — 🟡 VARIANT

87. Network meta-analysis — 🟡 VARIANT

88. Individual participant data (IPD) meta-analysis — 🟡 VARIANT

89. Cumulative meta-analysis — 🟡 VARIANT

90. Living systematic review — 🟡 VARIANT

#VIII. Advanced Analytical / Hybrid Designs

91. Propensity score–matched study — 🟡 VARIANT

92. Propensity score–weighted study — 🟡 VARIANT

93. Instrumental variable study — 🟡 VARIANT

94. Target trial emulation — 🟡 VARIANT

95. Comparative effectiveness study — 🟡 VARIANT

96. Dose–response study — 🟡 VARIANT

97. Multiphase optimization strategy (MOST) — 🟡 VARIANT

#IX. Mixed & Emerging Designs

98. Mixed-methods epidemiologic study — 🟡 VARIANT

99. Hybrid effectiveness–implementation design — 🟡 VARIANT

100. Systems epidemiology / modeling study — 🟡 VARIANT

Lets go through each study design in brief.........................

#I. OBSERVATIONAL STUDIES

#A. Descriptive Studies

#1️⃣ Case Report — 🔵 CORE

Definition:
Detailed description of a single patient or unique medical event. Often first report of rare disease, complication, or novel presentation. No control group.

Outcome Analysis Plan:

• Qualitative description only

• Present timeline, symptoms, lab/imaging findings, and outcome

• No statistical testing

Sample Size Calculation:

• Not applicable (single case)

References: Riley DS, Barber MS, Kienle GS, et al. CARE guidelines for case reports: explanation and elaboration document. J Clin Epidemiol. 2017;89:218–235. Available from: https://www.jclinepi.com/article/S0895-4356(17)30037-9/fulltext

#2️⃣ Case Series — 🔵 CORE

Definition:
Description of a group of patients with a similar diagnosis or exposure to identify patterns. No control group.

Outcome Analysis Plan:

• Frequencies and proportions of features or complications

• Mean ± SD or median (IQR) for continuous variables

• Trends over time summarized in tables or figures

Sample Size Calculation:

• Not fixed; usually ≥5–10 patients for meaningful pattern recognition

References: Kooistra B, Dijkman B, Einhorn TA, Bhandari M. How to design a good case series. J Bone Joint Surg Am. 2009;91(Suppl 3):21–26. doi:10.2106/JBJS.H.01573. Available from: https://journals.lww.com/jbjsjournal/fulltext/2009/05003/how_to_design_a_good_case_series.5.aspx

#3️⃣ Descriptive Cross-Sectional (Prevalence) Study — 🔵 CORE

Definition:
Observes a population at a single point in time to measure the prevalence of disease, risk factors, or behaviors.

#Outcome Analysis Plan

• Categorical outcomes:

  • Report as proportion or prevalence (%)

  • Example: % of adults with hypertension

• Continuous outcomes:

  • Report as mean ± SD (if normally distributed)

  • Or median (IQR) (if skewed)

• Optional subgroup comparisons:

  • Chi-square test for categorical variables

  • t-test / ANOVA for continuous variables

#Sample Size Calculation

1. For single proportion (categorical outcome / prevalence):

n=Z² pq/d²

Where:

• Z = Z-score for desired confidence interval (e.g., 1.96 for 95%)

• p = expected prevalence

• q=1−p

• d = desired absolute precision

2. For continuous variable (mean ± SD):

n=Z² σ^2/d²

Where:

• Z = Z-score for confidence interval

• σ (sigma) = estimated standard deviation

• d = desired absolute precision

#1️⃣ Sample Size Calculation — Single Proportion (Prevalence)

Proposed title: Prevalence of Health Literacy among Adults in Biratnagar, Nepal: A Descriptive Cross-Sectional Study

Formula:

n=Z²pq/d²

Context from Published Study:
In a large cross‑sectional survey from Wuhan, China, the knowledge rate of health literacy was 19.3% (mean score 42.9 ± 11.7), indicating many adults have limited health literacy.SpringerLink

Assumptions for Our Proposal:

• Expected prevalence (p): 20% (0.20), based on available literature like Wuhan, China study.SpringerLink

• q = 1 − p: 0.80

• Desired precision (d): ± 5% (0.05)

• Confidence level: 95%; Z = 1.96

#Stepwise Calculation

n=(1.96)²×0.20×0.80/(0.05)²

Where,

1. Z²=1.96²=3.8416

2. p×q=0.20×0.80=0.16

3. Numerator: 3.8416×0.16=0.6147

4. Denominator: d²=0.0025

n=0.6147/0.0025=245.88≈246

👉 Required sample size = 246 adults

#2️⃣ Sample Size Calculation — Continuous Outcome (Mean Health Literacy Score)

Proposed title: Mean Health Literacy Score among Adults in Biratnagar, Nepal: A Descriptive Cross-Sectional Study

Formula:

n=Z²σ²/d²

Context from Published Study:
In the Wuhan study, the average health literacy score in adults was 42.9 with SD ± 11.7.SpringerLink

Assumptions for Our Proposal:

• Estimated SD (σ): 11.7

• Desired precision (d): ± 2 points

• Z = 1.96

#Stepwise Calculation

n=(1.96)²×(11.7)²/(2)²n

Where,

1. Z²=3.8416

2. σ²=11.72=136.89

3. Numerator: 3.8416×136.89=526.20

4. d²=4

n=526.20/4=131.55≈132

👉 Required sample size = 132 adults

d=? uff..What is this?

In sample size calculations—for both prevalence (categorical outcome) and mean (continuous outcome)—the symbol d represents the desired absolute precision of your estimate.

#1️⃣ In Prevalence / Single Proportion Studies

Formula:

n=Z²pq/d²

What is d?

• d = margin of error or acceptable difference between your estimated prevalence and the true population prevalence.

• Expressed as a proportion (not percentage) in formula.

• Determines how precise your estimate will be.

Example:

• Expected prevalence of limited health literacy in Biratnagar adults(p) = 20% → 0.20

• You want the prevalence estimate to be accurate ±5% → d=0.05

Interpretation:

• If calculated prevalence = 20%, then the true population prevalence is expected between 15% and 25% (20% ± 5%).

• Smaller d → more precise estimate → larger sample size

  Someone said Absolute vs Relative d – Oh !! What to add? BUT WHY?

  ---

  #1️⃣ Absolute d

  • Let’s say d=5%

  • Interpretation: True prevalence will be within 20% ± 5% → 15% to 25%

  • Fixed margin, independent of the prevalence

  Formula in sample size calculation: n=Z² pq/d²​(d in same units as p)

  ---

  #2️⃣ Relative d

  • Relative precision = 10% of expected prevalence

  • Absolute d=10%×20%=2%

  • Interpretation: True prevalence within 20% ± 2% → 18% to 22%

  • Scales with prevalence

  Formula: dabsolute=drelative×p

  ---

  #Visual Table

  Type	Value	Interpretation
  Absolute d	5%	True prevalence 15–25%
  Relative d	10% → 2%	True prevalence 18–22%

  💡 Tip:

  • Use absolute d for standard prevalence calculations. Most textbooks and standard formulae (e.g., n=Z²pq/d² use absolute d.

  • Use relative d when prevalence is very low (<10%) or very high (>90%), to avoid extremely large or tiny sample sizes and to make the precision meaningful.

#2️⃣ In Mean / Continuous Outcome Studies

Formula:

n=Z² σ²/d²

What is ddd?

• d = desired precision around the mean

• It is the maximum acceptable difference between the sample mean and the true population mean

• Same effect: smaller ddd → more precise estimate → larger sample size

Example:

• Estimated SD (σ\sigmaσ) = 12 points

• Desired precision ±2 points → d=2

• Interpretation: The sample mean will be within 2 points of the true population mean with the specified confidence (e.g., 95%).

#✅ Key Points About d

1. Always express d in the same units as your outcome:

   • Proportion for prevalence

   • Score/measurement unit for continuous outcome

2. Smaller d → higher precision → larger sample size

3. Larger d → less precision needed → smaller sample size

#4️⃣ Time–Place–Person Descriptive Study — 🟡 VARIANT

Definition:
Describes disease distribution by person, place, and time, often used in outbreak investigations.

Outcome Analysis Plan:

• Person: age, sex, occupation → frequency/proportion, mean ± SD

• Place: geographic distribution → maps, clustering

• Time: epidemic curve, trend analysis

• Typically descriptive; inferential statistics optional

Sample Size Calculation:

• If prevalence is estimated → use cross-sectional formula

• For outbreak investigations → include all available cases

#5️⃣ Descriptive Ecological Study — 🟡 VARIANT

Definition:
Uses population-level (group) data to describe exposure–outcome relationships (unit of analysis = group, not individual).

Outcome Analysis Plan:

• Group-level summaries: mean, median, range

• Scatter plots or maps to visualize exposure–outcome relationships

• Correlation/regression analysis at group level

Sample Size Calculation:

• Depends on number of groups (≥10–20 groups recommended)

• Individual-level sample size not applicable

#B. Analytical Cross-Sectional / Correlational Studies

#1️⃣ Analytical Cross-Sectional Study — 🔵 CORE

Definition:
Measures exposure and outcome simultaneously in a population and analyzes association between them.

Outcome Analysis Plan:

• Binary outcome: odds ratio (OR), chi-square, logistic regression

• Continuous outcome: mean ± SD, t-test/ANOVA, linear regression

Sample Size Calculation:

• Two proportions (binary):

n=(Zα/22pˉ(1−pˉ)+Zβp1(1−p1)+p2(1−p2))2(p1−p2)2n = \frac{(Z_{\alpha/2}\sqrt{2\bar{p}(1-\bar{p})} + Z_\beta\sqrt{p_1(1-p_1) + p_2(1-p_2)})^2}{(p_1 - p_2)^2}n=(p1​−p2​)2(Zα/2​2pˉ​(1−pˉ​)​+Zβ​p1​(1−p1​)+p2​(1−p2​)​)2​

• Two means (continuous):

n=2(Zα/2+Zβ)2σ2(μ1−μ2)2n = \frac{2 (Z_{\alpha/2} + Z_\beta)^2 \sigma^2}{(μ_1 - μ_2)^2}n=(μ1​−μ2​)22(Zα/2​+Zβ​)2σ2​

#2️⃣ Correlational Study — 🔵 CORE

Definition:
Measures strength and direction of association between two continuous variables at one point in time.

Outcome Analysis Plan:

• Pearson correlation (r) for normally distributed variables

• Spearman rank correlation for non-normal data

• Simple/multiple linear regression for effect estimation

Sample Size Calculation:

n=(Zα/2+Zβ)2(0.5⋅ln⁡1+r1−r)2+3n = \frac{(Z_{\alpha/2} + Z_\beta)^2}{(0.5 \cdot \ln \frac{1+r}{1-r})^2} + 3n=(0.5⋅ln1−r1+r​)2(Zα/2​+Zβ​)2​+3

Where rrr = expected correlation coefficient

#3️⃣ Ecological Study — 🔵 CORE

Definition:
Analyzes population- or group-level data to examine exposure–outcome associations.

Outcome Analysis Plan:

• Group means, medians, proportions

• Scatter plots, maps, linear regression at group level

• Correlation coefficient optional

Sample Size Calculation:

• Based on number of groups ≥10–20

• More groups → more statistical power

• No individual-level sample size

#4️⃣ Multilevel Cross-Sectional Study — 🟡 VARIANT

Definition:
Cross-sectional study accounting for hierarchical data (e.g., students nested in schools).
Can separate individual-level and group-level effects.

Outcome Analysis Plan:

• Mixed-effects regression (random intercept, random slope)

• ICC (intra-class correlation) to measure clustering

• Adjust for confounders at different levels

Sample Size Calculation:

• Adjust for design effect due to clustering:

neffective=n1+(m−1)⋅ICCn_{\text{effective}} = \frac{n}{1 + (m-1)\cdot ICC}neffective​=1+(m−1)⋅ICCn​

Where:

• mmm = average cluster size

• ICCICCICC = intra-cluster correlation

• Then use cross-sectional formula for neffectiven_{\text{effective}}neffective​

#5️⃣ School-/Institution-Based Cross-Sectional Study — 🟡 VARIANT

Definition:
Cross-sectional study conducted within defined schools or institutions.
May introduce cluster effect if multiple institutions included.

Outcome Analysis Plan:

• Same as analytical cross-sectional study

• Adjust for clustering if multiple schools → multilevel or mixed-effects model

Sample Size Calculation:

• Adjust for design effect:

nadjusted=nsimple×DEn_{\text{adjusted}} = n_{\text{simple}} \times DEnadjusted​=nsimple​×DEDE=1+(m−1)⋅ICCDE = 1 + (m-1) \cdot ICCDE=1+(m−1)⋅ICC

• mmm = average students per school, ICCICCICC = intra-cluster correlation

#B. Case-Control and Related Designs

#1️⃣ Classical Case-Control Study — 🔵 CORE

Definition:
Selects cases with the outcome and controls without the outcome, then looks backward for exposure.

Outcome Analysis Plan:

• Odds Ratio (OR) for association

• Logistic regression for adjustment of confounders

• Chi-square / Fisher exact test for categorical variables

Sample Size Calculation:

• Binary exposure:

n=(Zα/22pˉ(1−pˉ)+Zβp1(1−p1)+p2(1−p2))2(p1−p2)2n = \frac{(Z_{\alpha/2}\sqrt{2\bar{p}(1-\bar{p})} + Z_\beta\sqrt{p_1(1-p_1) + p_2(1-p_2)})^2}{(p_1 - p_2)^2}n=(p1​−p2​)2(Zα/2​2pˉ​(1−pˉ​)​+Zβ​p1​(1−p1​)+p2​(1−p2​)​)2​

• Continuous exposure: two means formula

#2️⃣ Matched Case-Control Study — 🟡 VARIANT

Definition:
Cases and controls matched on confounders (e.g., age, sex) to reduce bias.

Outcome Analysis Plan:

• Conditional logistic regression

• Paired analysis for matched variables

Sample Size Calculation:

• Similar to classical, adjusted for correlation due to matching (smaller n usually)

#3️⃣ Unmatched Case-Control Study — 🟡 VARIANT

Definition:
No matching of controls; simpler design, risk of confounding.

Outcome Analysis Plan:

• Logistic regression

• Chi-square or t-test for comparisons

Sample Size Calculation:

• Same as classical case-control study

#4️⃣ Test-Positive / Test-Negative Case-Control — 🟡 VARIANT

Definition:
Cases = test-positive; controls = test-negative; often used in vaccine effectiveness studies.

Outcome Analysis Plan:

• Odds ratio for exposure

• Stratified analysis for confounders

Sample Size Calculation:

• Two-proportion formula using expected vaccine effectiveness (VE)

#5️⃣ Nested Case-Control Study — 🟡 VARIANT

Definition:
Case-control within a cohort; exposure measured before outcome occurs.

Outcome Analysis Plan:

• Conditional logistic regression

Sample Size Calculation:

• Include all cases + random sample of controls

• Based on cohort event rate and expected OR

#6️⃣ Case-Cohort Study — 🟡 VARIANT

Definition:
Random subcohort selected as controls; all incident cases included.

Outcome Analysis Plan:

• Cox proportional hazards

• Logistic regression for binary outcomes

Sample Size Calculation:

• Subcohort size based on expected event rate and desired power

#7️⃣ Two-Stage Case-Control Study — 🟡 VARIANT

Definition:
Screening + confirmatory stage; improves efficiency and reduces cost.

Outcome Analysis Plan:

• Logistic regression

• Adjust for selection probabilities

Sample Size Calculation:

• Stage 1: determine n for initial screening

• Stage 2: adjust for sampling fraction

#8️⃣ Case–Case Study — 🟡 VARIANT

Definition:
Compares two sets of cases with different outcomes/exposures.

Outcome Analysis Plan:

• Logistic regression to compare exposures

• Chi-square for categorical variables

Sample Size Calculation:

• Two proportions or two means depending on type of exposure

#9️⃣ Case–Population Study — 🟡 VARIANT

Definition:
Uses population-level controls instead of individual controls.

Outcome Analysis Plan:

• OR estimation comparing cases to population exposure prevalence

Sample Size Calculation:

• Based on expected exposure prevalence in cases and population

#C. Cohort / Longitudinal Studies

#1️⃣ Prospective Cohort Study — 🔵 CORE

Definition:
Follows exposed and unexposed groups forward in time to measure incidence of outcome.

Outcome Analysis Plan:

• Binary outcome: Risk Ratio (RR), OR, risk difference

• Time-to-event: Kaplan-Meier, Cox proportional hazards

• Continuous outcome: mean difference, linear regression

Sample Size Calculation:

• Binary:

n=(Zα/2+Zβ)2[p1(1−p1)+p2(1−p2)](p1−p2)2n = \frac{(Z_{\alpha/2} + Z_\beta)^2 [p_1(1-p_1)+p_2(1-p_2)]}{(p_1 - p_2)^2}n=(p1​−p2​)2(Zα/2​+Zβ​)2[p1​(1−p1​)+p2​(1−p2​)]​

• Continuous: two means formula

#2️⃣ Retrospective Cohort Study — 🔵 CORE

Definition:
Uses existing records to identify exposed/unexposed groups; looks forward from past exposure to outcome.

Outcome Analysis Plan:

• Same as prospective cohort

• Check for completeness/quality of records

Sample Size Calculation:

• Same as prospective cohort

#3️⃣ Ambispective Cohort Study — 🟡 VARIANT

Definition:
Combines retrospective and prospective data in one cohort.

Outcome Analysis Plan:

• Same as cohort

• Can analyze retrospective and prospective periods separately

Sample Size Calculation:

• Same as prospective cohort

#4️⃣ Fixed (Closed) Cohort — 🟡 VARIANT

Definition:
Cohort membership fixed at baseline; no new entrants during follow-up.

Outcome Analysis Plan:

• Incidence, RR, Cox regression

Sample Size Calculation:

• Standard cohort formula

#5️⃣ Dynamic (Open) Cohort — 🟡 VARIANT

Definition:
Members can enter or leave the cohort; follow-up time varies.

Outcome Analysis Plan:

• Incidence rate (events/person-time)

• Poisson regression for time-to-event

Sample Size Calculation:

• Person-time formula:

n=(Zα/2+Zβ)2[p1(1−p1)/t1+p2(1−p2)/t2](p1−p2)2n = \frac{(Z_{\alpha/2}+Z_\beta)^2 [p_1(1-p_1)/t_1 + p_2(1-p_2)/t_2]}{(p_1-p_2)^2}n=(p1​−p2​)2(Zα/2​+Zβ​)2[p1​(1−p1​)/t1​+p2​(1−p2​)/t2​]​

#6️⃣ Historical Cohort Study — 🟡 VARIANT

Definition:
Exposure and outcome occurred in the past; uses existing records.

Outcome Analysis Plan:

• Same as retrospective cohort

Sample Size Calculation:

• Same as retrospective cohort

#7️⃣ Birth Cohort Study — 🟡 VARIANT

Definition:
Follows individuals from birth onward to examine outcomes over life course.

Outcome Analysis Plan:

• Incidence, prevalence, growth trends, or survival analysis

Sample Size Calculation:

• Standard cohort formula; larger n needed for rare outcomes

#8️⃣ Occupational Cohort Study — 🟡 VARIANT

Definition:
Cohort defined by workplace or occupation.

Outcome Analysis Plan:

• Incidence of disease, RR

• Adjust for workplace clustering if multiple sites

Sample Size Calculation:

• Same as cohort, with optional design effect adjustment for clustering

#D. Self-Controlled / Time-Based Designs

#1️⃣ Case-Crossover Study — 🔵 CORE

Definition:
Each case serves as its own control; compares exposure during “hazard” vs “control” period.

Outcome Analysis Plan:

• Conditional logistic regression

• OR estimation

Sample Size Calculation:

• Event-driven; usually ≥50–100 events for sufficient power

#2️⃣ Self-Controlled Case Series (SCCS) — 🔵 CORE

Definition:
Only cases included; follow-up divided into risk and control periods.

Outcome Analysis Plan:

• Conditional Poisson regression

• Risk ratios within subjects

Sample Size Calculation:

• Event-driven; based on expected number of events

#3️⃣ Case-Time-Control Study — 🟡 VARIANT

Definition:
Adjusts case-crossover for time trends in exposure.

Outcome Analysis Plan:

• Conditional logistic regression

• Control for secular trends

Sample Size Calculation:

• Event-driven; more events needed than case-crossover

#4️⃣ Sequence Symmetry Analysis — 🟡 VARIANT

Definition:
Uses sequence of drug dispensing or exposure to detect association with outcome.

Outcome Analysis Plan:

• Odds ratio via sequence ratio

• Paired analysis

Sample Size Calculation:

• Event-based; number of sequences determines power

#5️⃣ Self-Controlled Risk Interval — 🟡 VARIANT

Definition:
Compares risk period vs control period within same individual.

Outcome Analysis Plan:

• Conditional Poisson/logistic regression

Sample Size Calculation:

• Event-driven; sufficient number of events required

#6️⃣ Time-Trend Study — 🟡 VARIANT

Definition:
Analyzes population-level trends in outcome over time.

Outcome Analysis Plan:

• Regression analysis for trends

• Graphical depiction of time series

Sample Size Calculation:

• Based on number of time points and expected effect size

#7️⃣ Longitudinal Ecological Study — 🟡 VARIANT

Definition:
Ecological (group) data collected over multiple time points.

Outcome Analysis Plan:

• Linear regression, time-series analysis

• Group-level trends

Sample Size Calculation:

• Number of groups × time points

• More groups/time points → higher power

#8️⃣ Panel / Repeated-Measures Study — 🟡 VARIANT

Definition:
Same individuals measured repeatedly over time.

Outcome Analysis Plan:

• Repeated-measures ANOVA, mixed-effects model

• Adjust for within-subject correlation

Sample Size Calculation:

• n per time point, adjusted for correlation between repeated measures (design effect for repeated measures)