Limits...
Refining hypertension surveillance to account for potentially misclassified cases.

Peng M, Chen G, Lix LM, McAlister FA, Tu K, Campbell NR, Hemmelgarn BR, Svenson LW, Quan H, Hypertension Outcomes Surveillance Te - PLoS ONE (2015)

Bottom Line: Accumulation of false positive cases over time using the 1H2P method could result in the overestimation of hypertension prevalence.A Bayesian method was then used to adjust the prevalence estimated from the reclassification method.The reclassification method with Bayesian adjustment produced similar prevalence estimates as the 1H2P method.

View Article: PubMed Central - PubMed

Affiliation: Department of Community Health Sciences, Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada.

ABSTRACT
Administrative health data have been used in hypertension surveillance using the 1H2P method: the International Classification of Disease (ICD) hypertension diagnosis codes were recorded in at least 1 hospitalization or 2 physician claims within 2 year-period. Accumulation of false positive cases over time using the 1H2P method could result in the overestimation of hypertension prevalence. In this study, we developed and validated a new reclassification method to define hypertension cases using regularized logistic regression with the age, sex, hypertension and comorbidities in physician claims, and diagnosis of hypertension in hospital discharge data as independent variables. A Bayesian method was then used to adjust the prevalence estimated from the reclassification method. We evaluated the hypertension prevalence in data from Alberta, Canada using the currently accepted 1H2P method and these newly developed methods. The reclassification method with Bayesian adjustment produced similar prevalence estimates as the 1H2P method. This supports the continued use of the 1H2P method as a simple and practical way to conduct hypertension surveillance using administrative health data.

No MeSH data available.


Related in: MedlinePlus

The diagram of methods developed and used in the study.*1H2P method: at least 1 hospitalization or 2 physician claims within two year-period for hypertension coded in International Classification of Diseases.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4372561&req=5

pone.0119186.g001: The diagram of methods developed and used in the study.*1H2P method: at least 1 hospitalization or 2 physician claims within two year-period for hypertension coded in International Classification of Diseases.

Mentions: Using the administrative health data, we defined patients with hypertension using 1H2P and reclassification methods, respectively. Reclassification method was applied at each fiscal year to identify the hypertension cases in the study population without carrying forward the case definitions into the following year. Bayesian method was then used to adjust the results reported from the reclassification (see Fig. 1).


Refining hypertension surveillance to account for potentially misclassified cases.

Peng M, Chen G, Lix LM, McAlister FA, Tu K, Campbell NR, Hemmelgarn BR, Svenson LW, Quan H, Hypertension Outcomes Surveillance Te - PLoS ONE (2015)

The diagram of methods developed and used in the study.*1H2P method: at least 1 hospitalization or 2 physician claims within two year-period for hypertension coded in International Classification of Diseases.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4372561&req=5

pone.0119186.g001: The diagram of methods developed and used in the study.*1H2P method: at least 1 hospitalization or 2 physician claims within two year-period for hypertension coded in International Classification of Diseases.
Mentions: Using the administrative health data, we defined patients with hypertension using 1H2P and reclassification methods, respectively. Reclassification method was applied at each fiscal year to identify the hypertension cases in the study population without carrying forward the case definitions into the following year. Bayesian method was then used to adjust the results reported from the reclassification (see Fig. 1).

Bottom Line: Accumulation of false positive cases over time using the 1H2P method could result in the overestimation of hypertension prevalence.A Bayesian method was then used to adjust the prevalence estimated from the reclassification method.The reclassification method with Bayesian adjustment produced similar prevalence estimates as the 1H2P method.

View Article: PubMed Central - PubMed

Affiliation: Department of Community Health Sciences, Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada.

ABSTRACT
Administrative health data have been used in hypertension surveillance using the 1H2P method: the International Classification of Disease (ICD) hypertension diagnosis codes were recorded in at least 1 hospitalization or 2 physician claims within 2 year-period. Accumulation of false positive cases over time using the 1H2P method could result in the overestimation of hypertension prevalence. In this study, we developed and validated a new reclassification method to define hypertension cases using regularized logistic regression with the age, sex, hypertension and comorbidities in physician claims, and diagnosis of hypertension in hospital discharge data as independent variables. A Bayesian method was then used to adjust the prevalence estimated from the reclassification method. We evaluated the hypertension prevalence in data from Alberta, Canada using the currently accepted 1H2P method and these newly developed methods. The reclassification method with Bayesian adjustment produced similar prevalence estimates as the 1H2P method. This supports the continued use of the 1H2P method as a simple and practical way to conduct hypertension surveillance using administrative health data.

No MeSH data available.


Related in: MedlinePlus