Deepthi Rajeev, MS, MSc

1
Evaluating the Impact of Electronic Disease
Surveillance Systems On Local Health Department
Work Processes

• Deepthi Rajeev, MS, MSc
• Department of Biomedical
  Informatics
• University of Utah

10/28/2009
2
Steps in the Reporting Process

• Laboratories, hospitals, doctors
• Identify condition
• Recognize that it is reportable
• Collect data and transmit clinical and laboratory
  information to public health
• Health Departments
• Receive clinical and laboratory reports
• Trigger an investigation if indicated
• Implement control measures to prevent further
  exposure and transmission

3
Problems for local health departments

• Manual reporting process (fax)
• Insufficient data in initial case report
• Manual triage of initial case report
• Time consuming, but not quantified
• Reports belong to other jurisdictions
• Duplicate reporting
• Lack of shared information

4
Former Reporting Process in Utah
Reporting entities
Local Health Dept
Phone
Manual entry
Fax
Fax
State Health Dept
Fax
Manual entry
Physician Infection Preventionist Others
No interface to receive electronic
information No shared public health records
5
New Reporting Process in Utah
Intermountain Healthcare
HL7 (RT-CEND)
Local Health Dept
Manual entry
Other reporting entities
Fax
Fax
Phone
State Health Dept
Fax
Physician
Manual entry
Infection Preventionist
Others
RTCEND electronic transmission of case
reports NEDSS shared public health records
6
Issues to consider

• Will the new electronic systems impact workflow?
• Who will be affected?
• Will the impact be positive or negative?

7
Research Objectives

• Identify metrics to monitor impact on workflow
  as new systems are developed and implemented
• Collect baseline data

8
Salt Lake Valley Health Dept
Study Location
9
Methods to select metrics

• Observation Study - observed tasks performed by
  various personnel at SLVHD
• Interviewed SLVHD personnel - Triage nurse, data
  entry, nurse, nurse manager
• Documented workflow associated with processing a
  case report and validated workflow
• Identified tasks that were frequent, important,
  and measurable
• Identified metrics to measure the selected tasks

10
Timestamps for timeliness evaluation
Onset of disease
Time to diagnose case
Case detected (date of lab results or diagnosis)
Reporting Time
Reported to public health
Time until case is triaged
Start triage process
Goal shorten this time interval
Time to review (establish jurisdiction and
reportable condition status) time for initial
data entry
Entry in surveillance database
Time until case is investigated
Investigation starts
Time until case investigation is completed
Investigation ends
11
SLVHD workflow
Forward to state health department
Stop
Start

• Does the report have all the information required
  to identify
• if the condition is reportable?
• if SLVHD is the responsible health department?

Triage Report
Archive Case Information
Review and assign case classification
Initial Data Entry
Identify if the report belongs to a new case or
is an update to an existing case
Investigation and implementation of control
measures
Assignment
12
Metrics for Triage Process

• Relevance of the reports received
• () of reports with new information including
• new unique (non-duplicate) cases
• updated information
• () of duplicate reports
• of out-of-county cases
• Follow-up
• of phone calls to gather additional information
  
• Type of additional information required
• of times data required was obtained
• of times forwarding of reports to data entry
  was delayed

• Time to review a report and determine that
  condition is reportable and relevant for Salt
  Lake County

13
Metrics for Data Entry Process

• Time required to identify whether information on
  a newly arrived report has previously been
  reported (i.e., new or existing case)
• Time required to enter data into the computer
• Number of reports entered each day and week

14
Baseline data collection
15
Methods

• Direct observations at Salt Lake Valley Health
  Department
• July 6 - 13, 2009
• Data collection form
• Extracted timestamps from NEDSS that were
  collected as part of routine work processes

16
Date Collection Form
17
Distribution of Reports Received
380 reports received for 33 different diseases
Out-of-County reports (n86)
New unique reports for Salt Lake County (n172)
Duplicate reports (n72)
76 reports from Utah Department of Health
Updated information (n50)
18
Number of reports triaged by day
19
Incomplete Reports

• Of 380 reports,
• 105 reports (32) required additional information
  
• 99 phone calls made
• 63 reports (60) were held for additional
  information and not forwarded to data entry
  immediately

20
Details on Missing Data
21
Time to Triage Reports

• Average 3 mins 31 sec / report
• 3 mins 30 sec for SLVHD cases
• 3 min 38 secs for Out-of-county cases
• Total time to triage cases (before forwarding to
  data entry) 122040 (hhmmss)
• 26 FTE

22
Interval between Report and Triage Date
23
Time for Initial Data Entry

• Observed 29th - 30th June 2009
• 62 reports entered
• Time to identify if report already exists
• in NETSS 12 seconds/ report
• in NEDSS 35 seconds/report
• Time to enter data
• in NETSS 49 seconds/ report
• in NEDSS 3 min 9 seconds/ report

During study, only part of the data was entered
in NETSS (NEDSS was the main system in use)
24
SLVHD Timeliness
Onset of disease
Time to diagnose case
Case detected (date of lab results or diagnosis
7 days
Reporting Time
6 days
Reported to public health
Time until case is triaged
1 day
Start triage process
Goal shorten this time interval
Time to triage
0 days
Entry in surveillance database
Time until case is investigated
7 days
Investigation starts
Time until case investigation is completed
Investigation ends
25
Next Steps

• Develop an ongoing monitoring system to evaluate
  impact of surveillance systems on workflow
• Issues
• Is this feasible with the existing
  infrastructure?

26
Acknowledgements

• CDC- Utah Public Health Informatics Center of
  Excellence (Grant 8P01HK000030)
• Rui Zeller
• Andrea Price
• Jon Reid
• Catherine Staes
• Ilene Risk
• Richard Kurzban
• Mary Hill
• Kris