TB Notes Newsletter

No. 2, 2011

SURVEILLANCE, EPIDEMIOLOGY, AND OUTBREAK INVESTIGATIONS BRANCH UPDATE

The TB Genotyping Information Management System — an Assessment of Timeliness 6 Months after Launch

Background: The TB Genotyping Information Management System (TB GIMS) was launched in March 2010 to improve access to and use of genotyping data by state and local TB programs.  In order for genotyping results to be used in routine TB control activities, it must be available to TB GIMS users in a timely fashion. Additionally, genotyping results must be linked to patient surveillance records so that patient demographics and risk factors are also available for creating reports and maps. To identify ways to improve access and usefulness of genotyping data, we assessed the timeliness of genotyping and surveillance data flow in TB GIMS and attempted to identify barriers to timeliness within the system.

Methods: The National TB Controllers Association, with technical assistance from CDC, conducted a web-based survey of 337 TB GIMS users in September 2010. In addition to the survey, we also analyzed TB GIMS system-generated data from isolates received at the genotyping labs during an 8-week period (May–July 2010).  Over 1,000 records from 45 states were included in the analysis.

When analyzing data flow in TB GIMS, we used specific dates to define movement of information through the system.

• Genotype Report Date is the date that genotyping results were entered by the genotyping lab into TB GIMS.
• Surveillance Upload Date is the date that surveillance data (initially reported to CDC by individual jurisdictions) were uploaded from CDC to TB GIMS.
• Linking Date refers to the date the state case number (for a specific genotyped isolate) was entered into TB GIMS by the state TB program to link genotype results to surveillance records.
• Completion Date defines the date in which all three previous steps were complete and the record was available to TB GIMS users for use in TB control activities.

Results: Among active users, 189 (56%) responded from 48 states. Even though 71% of respondents reported they were satisfied or very satisfied with TB GIMS, only 49% were satisfied with the timeliness of the system. Analysis of system data showed that the median time from specimen collection until record completion date was 142 days; median time to linking was the longest of the three steps (Figure 1).

Figure 1. Time from specimen collection — TB GIMS system-generated data, May–July 2010 (N=1,059 records)

Text Version

The most common rate-limiting step was linking (42% of records), followed by upload of surveillance data (37%). Genotyping was the rate-limiting step for only 21% of records. To illustrate the effect of timely linking, we analyzed the time to record completion date, stratified by whether or not the record was linked at the time of isolate submission. Records that were linked at the time of submission had a median time of 96 days from specimen collection to completion date, versus 142 days for records linked after isolate submission, a difference of 1.5 months.

Finally, users were asked to name barriers to linking at the time of isolate submission. Seventy-nine percent of respondents stated that the state public health laboratory, which sends isolates to the genotyping labs, does not have access to the state case number. An additional 33% of respondents cited delays in verifying and reporting TB cases—some states wait until a case is officially verified before assigning a state case number.

Summary: Overall, users expressed a high level of satisfaction with TB GIMS but were less satisfied with the timeliness of the system. Complete records were available a median of 142 days after specimen collection, including the 4–6 week period necessary for TB culture. Genotyping was the least common rate-limiting step, and linking was the most common rate-limiting step. Records that were linked at the time of isolate submission had a dramatically reduced time to a complete record. Two areas that could improve timeliness of linking are assigning state case numbers earlier and improving communication between TB programs and state public health labs. DTBE will be working closely with state and local partners to determine the best ways of supporting improvements in these areas.

TB GIMS Update: One year after launch, the TB GIMS team is working to improve and expand the service.  As of March 2011, there were over 450 active users from 52 reporting areas and CDC. Four upgrades to TB GIMS have been released and several enhanced features are being developed for the future. We are dedicated to continuing to improve the system to best fit the needs of TB programs nationwide.

Acknowledgments: We would like to thank the National TB Controllers Association, the Washington Department of Health, the TB GIMS Development Team, survey respondents, and all TB GIMS users for their important contributions.

—Reported by Brian Baker, MD
EIS Officer, Molecular Epidemiology Activity, SEOIB
Div of TB Elimination

A Year in the Life of the Surveillance Report

Have you ever wondered how Reported Tuberculosis in the United States (informally known as the Annual Surveillance Report) is created each year? How do all the reports of TB cases evolve from raw numbers to morbidity tables? To provide you with an insider’s view of the process, I will describe a year in the life of the surveillance report.

Winter: The process starts with TB counted cases that are reported to CDC by state and local health departments. Health departments work diligently to finalize their numbers for the end of the year during this time period. After data from reporting jurisdictions are received at CDC, Surveillance team members work closely with members of the DTBE Data Management and Statistics Branch to review the data to determine whether there are any irregularities, particularly with variables that have been inconsistent in the past or have recently changed. In January, the Surveillance team asks sites for verbal case counts for the previous year, with a goal of developing a provisional data set ultimately used to create the World TB Day Morbidity and Mortality Weekly Report (MMWR). Bob Pratt reports, “The goal is to have our preliminary numbers be no more than 5-10 cases off from the final count, due in early March. It is rare for any surveillance group at CDC to report data from the year just concluded in March, and is a testament to the hard work of our partners that we are able to report our numbers so quickly.”

Spring: After the World TB Day MMWR has been published, the work of data formatting and report design begins. The Surveillance team holds a meeting to divide the assignments for producing the report. In the meantime, Carla Jeffries begins brainstorming to come up with ideas for the cover for the surveillance report. Using her background as a graphic artist, she then designs the cover that you see on the report. The 2010 Report will display Carla’s fourth cover design; although she is proud of all the covers she has designed, her favorite is the word cloud on the 2008 cover.

Summer: During the summer, the team double- and triple-checks the data to be included in the surveillance report. Bob Pratt creates Excel files containing the results for each table that will ultimately go into the report. Upon completing the drafts for the tables, he sends electronic copies to other surveillance team members. The team member assigned to checking the table will independently write SAS code to reproduce numbers needed for the table. This is done to avoid calculation or transcription errors.

Carla Jeffries noted that the production of the 2009 surveillance report presented unusual challenges for the team. Data checking was a particular challenge last year because it was the first year that the new RVCT form had been used. Therefore, the SAS code used to derive the data presented in each table had to be rewritten to reflect changes in the way the data were collected. In addition, many of the footnotes had changed, so the team held a meeting to discuss the definition of each table footnote. “Changes in footnotes and tables represented the biggest challenge in putting the report together,” Carla stated, “because changes in one table or footnote often have a waterfall effect on other tables, leading to multiple edits.” Meanwhile, with contributions from the team, Carla Winston wrote the 2009 Executive Summary, and Lilia Manangan wrote the Technical Notes.

By August most of the work of designing the report is complete. Carla Jeffries works with the Creative Services team at CDC to finalize the cover. Students or Surveillance team members create slide sets that describe the data in the report. These slides are among the most popular section of the report; since October 2010, the 2009 data slides have been viewed over 15,000 times. Once all these steps are completed and integrated into the full report, the report is entered into DTBE clearance.

Autumn: All the hard work pays off in autumn—the surveillance report is published on the web by the DTBE Communications Team. The greatest utility of the surveillance report is as a web-based product. The report has been downloaded over 8,000 times since its posting on the web, and viewed almost 50,000 times. About a month after the report is available on the web, 2,500 hard copies are printed for use by DTBE and other partners. Shortly thereafter, the process of finalizing the data for the next year begins. The Surveillance Team members thank all local and state health department staff, and people working on electronic data reporting, updates, and analysis, who make it possible to summarize TB trends each year.

—Reported by Suzanne Beavers, MD
Div of TB Elimination

Top of Page