Lyme Disease Surveillance Using Sampling Estimation: Evaluation of an Alternative Methodology in New York State

Abstract

In the 14-year period from 1993 to 2006, New York State (NYS) accounted for over one-quarter (27.1%) of all confirmed Lyme disease (LD) cases in the United States. During that time period, a nine-county area in south-east NYS accounted for 90.6% of the reported LD cases in the state. Based on concerns related to diminishing resources at both the state and local level and the increasing burden of traditional LD surveillance, the NYS Department of Health (DOH) sought to develop an alternative to traditional surveillance that would reduce the investigative workload while maintaining the ability to track LD trends by developing a system to estimate county-level LD cases based on a 20% random sample of positive laboratory reports. Estimates from this system were compared to observed cases from traditional surveillance for select counties in 2007-2009 and 2011. There were no significant differences between the two methodologies in six of nine evaluations conducted. In addition, in 93 of 98 (94.9%) demographic, symptom and other variable proportion comparisons made between the two methodologies in 2009 and 2011, there were no significant differences found. Overall, using sampling estimates was accurate and efficient in estimating LD cases at the county level. Use of case estimates for LD should be considered as a useful surveillance alternative by health policy makers for states with endemic LD.

Surveillance for Lyme Disease — United States, 2008–2015

“…Other states have begun to employ statistical methodology to estimate the number of cases each year. For example, several counties in New York have implemented a system in which 20% of positive laboratory reports are sampled and investigated to determine what proportion can be confirmed; these results are extrapolated to the remaining unsampled laboratory reports to arrive at an estimate of Lyme disease case counts in those counties ( 22 ). Several states are considering adopting similar methodologies to better manage public health surveillance for Lyme disease ( 23 ).…”

Tick bite and Lyme disease‐related emergency department encounters in New Hampshire, 2010–2014

“…Lyme disease is a high‐volume reportable disease in NH and in certain other parts of the country. The number of LD investigations in highly endemic states like NH combined with underreporting and resource limitations has made it difficult to accurately track LD incidence and collect useful data in a timely fashion; this has led some states to consider alternative LD surveillance strategies (Lukacik, White, Noonan‐Toly, DiDonato, & Backenson, ). Additionally, many more people may receive care for tick bites that put them at risk for LD and other tick‐borne illnesses than are reflected in traditional case‐based surveillance data (Nelson et al., ; Schiffman et al., ; White et al., ).…”

Validation of Random Sampling as an Estimation Procedure for Lyme Disease Surveillance in Massachusetts and Minnesota

“…How to address the burden in a consistent way across jurisdictions is a topic still under discussion with various methodologies being investigated and validated. Implementation of this estimation procedure for LD surveillance has been shown to be a valid approach to assessing disease burden in MA, MN and NY (Lukacik et al., ) while still meeting the primary goals of LD surveillance. Estimation provides an alternative approach that, when necessary, permits redistribution of funding and personnel to LD prevention efforts and investigation of other endemic and emerging vector‐borne diseases.…”

“…These deviations from the national case definition, as well as the degree of underreporting that occurs with LD, result in national case counts that are inaccurate and challenging to compare between states. In 2007, the New York State Department of Health (NYSDOH) sought to develop an alternative surveillance method that would be as accurate as the traditional surveillance system, yet more efficient (Lukacik et al., ). In their method, county‐level LD cases were estimated using a 20% random sample of positive laboratory reports to reduce the surveillance workload of clinical follow‐up.…”

“…In their method, county‐level LD cases were estimated using a 20% random sample of positive laboratory reports to reduce the surveillance workload of clinical follow‐up. Evaluation of this alternative surveillance method showed that the estimation procedure was efficient and accurate in estimating LD cases at the county level; 19 counties in New York State outside of New York City currently utilize this surveillance approach (Lukacik et al., ). This study describes a validation of this alternative LD surveillance strategy.…”