Cover; Half title; Chapman & Hall/CRC Biostatistics Series; Title; Copyrights; Dedication; Contents; Preface; 1 Introduction; 1.1 Control and self-control in epidemiology; 1.2 Self-controlled methods; 1.3 Guide to contents; 1.4 Computer package and data; 2 Epidemiological overview; 2.1 Genesis of the SCCS method; 2.2 Rationale for the SCCS method; 2.2.1 Case series; 2.2.2 Self-control; 2.2.3 Data requirements; 2.3 Some illustrations; 2.3.1 Using only cases; 2.3.2 Controlling confounding; 2.4 Assumptions and alternatives; 2.4.1 Assumptions of the SCCS method

2.4.2 What if the assumptions are not satis ed?2.5 Bibliographical notes and further material; 3 The SCCS likelihood; 3.1 Why start with the likelihood?; 3.2 Likelihood for the standard SCCS model; 3.3 Properties of the SCCS likelihood; 3.4 Example: MMR vaccine and aseptic meningitis; 3.5 The general SCCS likelihood; 3.6 MMR vaccine and aseptic meningitis: derivation of the SCCS likelihood; 3.7 Assumptions of the SCCS method; 3.7.1 Assumption 1: Poisson or rare events; 3.7.2 A counter-example: negative binomial events*; 3.7.3 Assumptions 2 and 3: validity of conditioning

3.7.4 A more formal demonstration*3.7.5 Assumption 4: independent ascertainment; 3.8 Derivation of the SCCS likelihood*; 3.9 Bibliographical notes and further material; 4 The standard SCCS model; 4.1 Proportional incidence models; 4.2 Fitting the standard SCCS model; 4.3 The R package SCCS: standard SCCS model; 4.3.1 A single point exposure: MMR vaccine and ITP; 4.3.2 Reshaping the MMR vaccine and ITP data; 4.3.3 Extended exposures: antidepressants and hip fracture; 4.4 Data formats for repeated exposures; 4.4.1 Intermittent treatments: NSAIDs and GI bleeds

4.4.2 Multiple vaccine doses: convulsions and DTP vaccine4.5 Multiple exposure types; 4.5.1 Exposures of several types: convulsions, Hib and MMR vaccines; 4.5.2 Multiple exposures of several types: NSAIDs, antidepressants and GI bleeds; 4.5.3 Multiple doses of di erent vaccines: convulsions, DTP and Hib vaccines; 4.5.4 Overlapping risk periods: convulsions and DTP; 4.6 Comparing models: likelihood ratio tests; 4.6.1 Comparing models: ITP and MMR vaccine; 4.6.2 Combining multinomial categories*; 4.7 Interactions: e ect modi cation and strati cation; 4.7.1 Interactions: sex, ITP and MMR vaccine

4.7.2 Interactions between exposures: GI bleeds, NSAIDs and antidepressants4.8 Inde nite and extremal risk periods; 4.8.1 Curtailed observation: antidiabetics and fractures; 4.8.2 Inde nite risk periods: MMR vaccine and autism; 4.8.3 Initial risk periods: NRT and MI; 4.9 SCCS analyses with temporal e ects; 4.9.1 Calendar time: GBS and in uenza vaccine; 4.9.2 Seasonal SCCS model: OPV and intussusception; 4.10 Parameterisation of the standard SCCS model*; 4.11 Bibliographical notes and further material; 5 Checking model assumptions; 5.1 Rare disease assumption for non-recurrent events