Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

A comparison of methods to estimate case number for Lyme disease in Allegheny County, Pennsylvania, 2014

Chen, Xirun (2016) A comparison of methods to estimate case number for Lyme disease in Allegheny County, Pennsylvania, 2014. Master Essay, University of Pittsburgh.

[img] Microsoft Word
Submitted Version
Available under License : See the attached license file.

Download (413kB)
[img] Plain Text (licence)
Available under License : See the attached license file.

Download (1kB)


Background: Lyme disease is the most commonly reported tick-borne illness in the United States. The Allegheny County Health Department is challenged by progressively increasing reports of Lyme cases recently. Most reports require follow-up investigation by public health officials with laboratories and providers to obtain missing clinical information. Due to this burgeoning burden, it is imperative to seek alternative surveillance approaches to monitor Lyme disease. The purpose of this study was to examine the accuracy of sampling and modeling methods for predicting, the number of cases for Lyme disease in Allegheny County, using 2014 data. Methods: A 20% sample of the laboratory reports was selected by R program. The estimate of sampling method was calculated by multiplying the number of cases resulting from sample by five. Univariate logistic regression models were conducted to determine the significant predictors related to final case status. The estimate of modeling method was generated from predicted probabilities which were computed by multivariate logistic regression model based on a 20% sample. Both methods were simulated for 1,000 times to reduce variance. Results: Univariate logistic regression analysis demonstrated that greater odds of being classified as “confirmed” were revealed for investigations with more positive test results (OR = 2.29, CI 1.91-2.73), and those involving patients aged 60 or older (OR = 1.71, CI 1.31-2.26). Sampling and modeling methods yielded identical (758 with a standard deviation of 36.2) estimated numbers of confirmed and probable cases. For both methods, 96.7% simulated case counts were within the 10% margin of error of the 758 true cases. Modeling method did not improve the accuracy of the sampling results. 20% sampling rate was found better suited than rates of 10% and 33% in terms of reducing follow-up burden and tracking trends of Lyme cases. Conclusions: Sampling estimation was efficient and accurate in estimating case occurrences and alleviating investigative burden. The public health significance was highlighted as future application of sampling could conserve limited surveillance resources by reducing follow-up efforts while providing precise estimates. SAS macros were provided to implement this approach for 2015 Lyme surveillance.


Social Networking:
Share |


Item Type: Other Thesis, Dissertation, or Long Paper (Master Essay)
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Chen, Xirunxirun.chen@pitt.eduXIC62
ContributionContributors NameEmailPitt UsernameORCID
Committee ChairGlynn, Nancy Wglynnn@edc.pitt.eduUNSPECIFIEDUNSPECIFIED
Committee MemberKurland, Brenda FBFK10@pitt.eduUNSPECIFIEDUNSPECIFIED
Committee MemberFiddner, JenniferJennifer.Fiddner@alleghenycounty.usUNSPECIFIEDUNSPECIFIED
Date: 2016
Date Type: Publication
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Publisher: University of Pittsburgh
Institution: University of Pittsburgh
Schools and Programs: School of Public Health > Epidemiology
Degree: MPH - Master of Public Health
Thesis Type: Master Essay
Refereed: Yes
Uncontrolled Keywords: Lyme
Date Deposited: 07 Sep 2016 17:50
Last Modified: 26 Jul 2022 10:55


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item