TWIN:Family
From Schemalet Wiki
Contents |
Generic data model
Story
Need generic data structure which can be used to handle all kind of family relationships including mono- and dizygotic twins.
Class diagram
Description
The data structure is from National Public Health Institute (KTL) where it is used to handle group of individuals in epidemiological studies. Subjects can have different roles (father., mother, child..) in different groups (like family or mzpair groups).
Use case examples
Proof
Model was not no practical for the eutwin-stud due to legacy reasons. Most of twin registries/ studies were using indexing system where twin pairs share same code (see below). Family data is expressed in de facto standard linkage format where subject object has ids of mother and father (ids are null if father/mother do not exist in cohort/study)
Practical solution - TWIN ID
Schema
Country Code The country codes, giving data origin will be according to the ISO 3166 standard.
Randomized number In the database there will be twins and non twins. Each twin pair will share the same randomized number, a none twin will receive an own randomized number. The none twin randomized number will only occurrence once, but the randomized number for twins, will occur two times if twins, three times if triplets and four times if quadruplets. The country code is part of the EUidnumber, and this allows each country to administrate their own randomized numbers. How this will be done is up to each country as long as it generates a unique number for each individual and contains of 7 numbers.
Identification number The EUidnumber needs indicator weather or not this is a twin. To obtain this information and still have unique numbers the EUidnumber should end (except the checksum) with the following numbers:
1 - Twin 1 2 - Twin 2 3 - Triplet 4 - Quadruplet 0 - Non twins
Checksum The checksum is calculated by GUMM algorithm (H. Peter Gumm: A new class of check digit methods for arbitrary number systems, IEEE Transactions on information theory, 31 (1985), 102-105)
Note
In genral, coding information into primary IDs (on which the informatics system is based on) is not good thing. Here we cannot update twin information once the ID is assigned.
Description
Type on twin relationship (mono- or dizygotic) is flagged in subject table. The model is based on requirements given by partners based on scope of the study. The model do not allow to have mixed triplets (e.g mzpair and one dizygotic).
