PubMed Health⌕ Search
Home/Archive/PubMed

PubMed · 16632901

From Rasch scores to regression.

Karl Bang Christensen
View original source ↗Export metadata ↓

Abstract

Rasch models provide a framework for measurement and modelling latent variables. Having measured a latent variable in a population a comparison of groups will often be of interest. For this purpose the use of observed raw scores will often be inadequate because these lack interval scale properties. This paper compares two approaches to group comparison: linear regression models using estimated person locations as outcome variables and latent regression models based on the distribution of the score.

Explore related subjects

DenmarkHumansLinear ModelsModels, StatisticalPsychological TestsComparative StudyJournal Article

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Karl Bang Christensen. 2006. From Rasch scores to regression.. https://pubmed.ncbi.nlm.nih.gov/16632901/

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2026

Emerging trends in invasive Streptococcus dysgalactiae subsp. equisimilis infections in Denmark, 2014 to 2024: a nationwide genomic and registry-based study.

BACKGROUNDIncreasing incidence rates of invasive Streptococcus dysgalactiae subspecies equisimilis (iSDSE) have been detected worldwide.AIMWe aimed to investigate iSDSE infection incidence rates in Denmark during 2014-2024, and characterise the genomic population structure of a subset of iSDSE isolates and their antimicrobial resistance (AMR).METHODSUsing national register data, we estimated overall and sex-/age-stratified iSDSE incidences during 2014-2024, by retrospectively identifying cases of invasive infections with group C and G streptococci or S. dysgalactiae (including specified as subspecies equisimilis). From the voluntary national beta-haemolytic streptococci laboratory surveillance system, whole-genome-sequenced isolates from August 2020-September 2022 were used to investigate the iSDSE genomic population structure. Susceptibility to penicillin, erythromycin and clindamycin was determined and AMR genes identified.RESULTSDuring 2014-2024, iSDSE incidence rates increased significantly (linear trend analysis p&#x2009;<&#x2009;0.001) with mean annual incidence ranging between 10.3 and 16.4 per 100,000, peaking in 2023. Incidence was higher in males, increasing with older age. Nearly 75% of the&#x2009;1,223 iSDSE isolates belonged to four of 14 genetic clusters. Sequence types (STs) ST20 and ST17 were most prevalent, while emm-type stG62647, a variant associated internationally with higher virulence, dominated. All isolates were phenotypically susceptible to penicillin but approximately 10% were respectively erythromycin and clindamycin resistant. High erythromycin resistance prevalence (57%;&#x2009;39/68), coinciding with gene ermA, occurred in one genetic cluster.CONCLUSIONThe findings illustrate the need for national registry-based surveillance to detect epidemiological changes and potential outbreaks. Further, continuous genomic surveillance can monitor the occurrence and expansion of genetic clades and AMR genes.

Denmark↗
PubMed2007

[Primary ovarian cancer. A comparison of registrations in the Danish Cancer Registry and the Pathology Data Bank].

INTRODUCTION: Ovarian cancer (OC) registrations in the Danish Cancer Registry (CR) and the Danish Pathology Data Bank (PDB) 1978-99 were compared in order to assess the possibility of identifying OC cases from another data source than the CR. MATERIAL AND METHODS: A total of 13,320 OC were identified in the CR/PDB and comparisons of the registrations were performed. RESULTS: The frequency of cases registered in both registries increased during 1978-99. Agreement in registration of histological subtypes in the two registries was found in 82% of the cases. Diagnosis among the rest was most often consistent; however, a more specific diagnosis was often found in one or the other registry. The frequency of women only registered with OC in the CR decreased from 81% to 35%, as the number of women with no diagnoses at all in the PDB declined from 1978 to 1997, whereafter it stabilised, primarily because all pathology departments now reported data to the PDB. In 1997-99, 29% of OC registered in the CR with a histologically verified OC diagnosis was still not registered with cancer in the ovaries in the PDB. Most of these, however, could be identified in the PDB with diagnoses consistent with metastases from OC. In total, 749 OC were not registered in the CR, though explanations for their missing registrations could be found in most cases. CONCLUSION: Data from the PDB may not be useful in complete identification of ovarian cancer cases for research purposes. However, it is a good supplement to data from the CR.

Denmark↗