| Show Linking records from two data sources Prior to stratifying the state mortality records by LDS church membership records from the state and from the LDS Church needed to be linked however Utah state death records and LDS church records do not sharecommon identifier that is unique to an individual In order to link records from these two sources we used the probabilistic linking program LinkPro Wajda et al 1991 This program calculates probabilities to determine whetherpair of records refer to the same person To calculate these probabilities we used eleven identifying characteristics that were common to the two data sets These eleven variables are last name first name middle name birth day birth month birth year marital status county of death and SOUNDEX versions of the three names SOUNDEX version ofname isnumerical value for that name its purpose is to account for different spellings of the same name In order forstate death record to be linked to an LDS membership record we required records to match on at least seven of the eleven variables Ambiguous matches which were relatively few were hand-coded There were 30 819 male deaths and 28 080 female deaths reported in Utah during 19941998 Using the probabilistic linking program with the criteria stated above 17 427 of the 18 005 96percent LDS male deceased membership records were matched tostate death record and 17 234 of the 17 859 96percent LDS female deceased membership records were matched tostate death record This percentage of linked records is very high and comparable to linkage results between two data sets when both data sets havecommon unique identifer likesocial security number Curb etal 1985 Ifstate death record was linked to an LDS church record then the individual was classified asmember of the LDS Church State death records that did not link to an LDS church record In an effort to were classified as non-LDS determine if any non-LDS were misclassified we sent name and birth date information from all unlinked records to the Membership and Records Department of the LDS Church where church personnel compared this information to the church complete membership database State death records that matched on birth date and first and last names to church records were reviewed by hand to ascertain whether or not they were an actual match Of the 24 233 unlinked records that were initially classified as non-LDS 215 17percent were matched tochurch record in this manner therefore of the 30 819 male deaths in Utah 19 657 63percent were classified as LDS and of the 28 080 female deaths 19 219 68percent were classified as LDS Causes of death with corresponding ICD-9 codes considered in this study are presented in Table Given the LDS proscription of tobacco use we divided the neoplasms into tobacco and non-tobacco related in order to obtainclearer comparison of cancer mortality between the LDS and non-LDS groups The following cancers were classified as tobacco-related lip oral cavity throat and esophagus stomach pancreas larynx trachea lung and bronchus kidney and bladder Results The ten leading causes of death in Utah for the years 1994-1998 by age are presented for LDS and non-LDS groups in Table males and Table females In addition to the cause of death the frequency and the percentage of all deaths explained by each cause are also included in the tables Because of the difficulty of correctly classifying newborns as either LDS ornon-LDS the 0-1 year age group corresponds to all deaths in this age group The leading causes of death and their ranks for males in the two oldest age groups are essentially the same for both LDS and non-LDS groups For males aged 40-59 years the top six leading causes and their ranks are the same for the two groups however there is only partial overlap in the remaining four causes of death between the LDS and non-LDS groups In the two younger age groups 1-19 20-39 there are relatively few differences in the 10 leading causes of death for the LDS and non-LDS males although several differences in their ranks exist For example in the 1-19 year age group the second and third leading causes of death are reversed for the two groups as are the fourth through sixth For males in the 20-39 year age group the top eight causes of death 41 Digital image 2005 Marriott Library University of Utah Al rights reserved |
