Relationships between statistical measures of agreement: sensitivity, specificity and kappa

Feuerman, Martin; Miller, Allen R.

doi:10.1111/j.1365-2753.2008.00984.x

Cited by 61 publications

(51 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The kappa term ranges from Ϫ1 to 1 and can be negative if the agreement is less than what would be expected by chance. The following labels were assigned to the corresponding ranges of kappa strength: poor agreement, Ͻ0; slight, 0.0 to 0.20; fair, 0.21 to 0.40; moderate, 0.41 to 0.60; substantial, 0.61 to 0.80; and almost perfect, 0.81 to 1.00 (12,13). All statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC).…”

Section: Methodsmentioning

confidence: 99%

Survey of Culture, GoldenGate Assay, Universal Biosensor Assay, and 16S rRNA Gene Sequencing as Alternative Methods of Bacterial Pathogen Detection

et al. 2013

View full text Add to dashboard Cite

Section: Methodsmentioning

confidence: 99%

Survey of Culture, GoldenGate Assay, Universal Biosensor Assay, and 16S rRNA Gene Sequencing as Alternative Methods of Bacterial Pathogen Detection

et al. 2013

View full text Add to dashboard Cite

“…Although there is a long discussion about the problems of using the kappa coefficient (e.g., [6,15,51]), it is usually reported in those studies. However, no further correction is performed, since the resulting estimation model is not well-identified, making the derivation of a precise-valued estimator impossible.…”

Section: Introductionmentioning

confidence: 97%

Partially identified prevalence estimation under misclassification using the kappa coefficient

Küchenhoff

Augustin

Kunz³

2012

International Journal of Approximate Reasoning

View full text Add to dashboard Cite

We discuss prevalence estimation under misclassification. That is we are concerned with the estimation of a proportion of units having a certain property (being diseased, showing deviant behavior, etc.) from a random sample when the true variable of interest cannot be observed, but a related proxy variable (e.g. the outcome of a diagnostic test) is available. If the misclassification probabilities were known then unbiased prevalence estimation would be possible. We focus on the frequent case where the misclassification probabilities are unknown but two independent replicate measurements have been taken. While in the traditional precise probabilistic framework a correction from this information is not possible due to non-identifiability, the imprecise probability methodology of partial identification and systematic sensitivity analysis allows to obtain valuable insights into possible bias due to misclassification. We derive tight identification intervals and corresponding confidence regions for the true prevalence, based on the often reported kappa coefficient, which condenses the information of the replicates by measuring agreement between the two measurements. Our method is illustrated in several theoretical scenarios and in an example from oral health on prevalence of caries in children.

show abstract

“…The differences in the prevalence of sufficient levels of physical activity between questionnaires were evaluated by the intersection of CI 95%. To evaluate the agreement between measures obtained by the simplified questionnaire and detailed questionnaire to correctly classify adolescents as the level of physical activity (sufficiently physically active vs insufficiently physically active), the Kappa index was used (values up to 0.19 were classified as poor; from 0.20 to 0.39, slight agreement; from 0.40 to 0.59, moderate agreement; from 0.60 to 0.79, substantial agreement and values above 0.80, almost perfect 20 ) and sensitivity and specificity measures, adopting as reference the measured obtained with the detailed questionnaire 21 . Statistical analyses were performed using Stata 13 software.…”

Section: Methodological Proceduresmentioning

confidence: 99%

“…(20,1%, IC95%: 18,(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)6) foi inferior a do questionário detalhado (50,2%, IC95%: 48,7). A concordância entre as medidas dos questionários foi leve (k variando de 0,21 a 0,34).…”

unclassified

Prevalência de níveis suficientes de atividade física em adolescentes: comparação entre a medida obtida de um questionário simplificado e a de um detalhado

Júnior

Barbosa

Mendes

et al. 2016

Rev. Bras. Cineantropom. Desempenho Hum.

View full text Add to dashboard Cite

-This study aimed was to compare the prevalence of sufficient levels of physical activity produced by a measure of the simplified questionnaire compared to a detailed questionnaire and analyze the concordances between these measures in adolescents. This is cross-sectional study with 2,859 adolescents high school students (57.8% female, 14-19 years old), from the city of João Pessoa (PB), Northeast, Brazil. Physical activity was measured by two questionnaires: a simplified instrument, containing a single question and a detailed questionnaire. Adolescents engaging in physical activity ≥300 minutes/ week were classified as sufficiently physically active. Agreement between the two measures was evaluated by the Kappa statistic (k) and sensitivity and specificity measures using the results of the detailed questionnaire as reference. The prevalence of adequate levels of physical activity was lower in the simplified questionnaire (20.1%; 95%CI: 18.6-21.6) compared to the detailed questionnaire (50.2%; 95%CI: 48.4-52.7). Inter-questionnaire agreement was slight (almost all with k<0.30). Sensitivity of the simplified questionnaire was low (from 26.4%, 95%CI: 22.9-30.0; to 40.9%, 95%CI: 37.0-45.0), and specificity was high (from 92.3%, 95%CI: 86.9-96.0; to 94.3%, 95%CI: 92.0-96.1). Conclusions: the agreement between the prevalence of sufficient levels of physical activity measured by a simplified and detailed questionnaire was low. Simplified questionnaires should be used with caution because they can underestimate the prevalence of levels of physical activity in adolescents.Key words: Adolescent; Motor activity; Sensitivity; Specificity. (20,1%, IC95%: 18,(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)6) foi inferior a do questionário detalhado (50,2%, IC95%: 48,7). A concordância entre as medidas dos questionários foi leve (k variando de 0,21 a 0,34). A sensibilidade da medida de atividade física do questionário simplificado foi baixa (variou de 26,4%, IC95%: 22,0; a 40,9%, IC95%: 37,0) e a especificidade alta (variou de 92,3%, IC95%: 86,0; a 94,3%, IC95%: 92,1

show abstract

Relationships between statistical measures of agreement: sensitivity, specificity and kappa

Abstract: The analytic formulas and graph could be potentially useful to clinicians and biostatisticians in better interpreting the outcomes of an alternative diagnostic test whenever the measures sensitivity, specificity and kappa are employed together.

Cited by 61 publications

References 10 publications

Survey of Culture, GoldenGate Assay, Universal Biosensor Assay, and 16S rRNA Gene Sequencing as Alternative Methods of Bacterial Pathogen Detection

Survey of Culture, GoldenGate Assay, Universal Biosensor Assay, and 16S rRNA Gene Sequencing as Alternative Methods of Bacterial Pathogen Detection

Partially identified prevalence estimation under misclassification using the kappa coefficient

Prevalência de níveis suficientes de atividade física em adolescentes: comparação entre a medida obtida de um questionário simplificado e a de um detalhado

Contact Info

Product

Resources

About