Figure 1.
The system architecture and data flow of infectious disease case reporting in Taiwan before (Panel A) and after (Panel B) the implementation of AEIS.
National Notifiable Disease Surveillance System a(NDSS) receives infectious disease (ID) cases reporting from btwo sources: passive case finding by physicians or active case identification by public health officers. (A) Prior to AEIS: Taiwan CDC managed ID-related health information by using three information systems: NDSS, central microbiological laboratory diagnostic system, and National Immunization Information System c(NIIS); for epidemic investigation, dpre-designed digital questionnaires were available for 18 notifiable diseases, and paper-based questionnaires had to be generated as needed each time for all other notifiable diseases. (B) After the application of AEIS: Template questionnaires (A2) for use in investigating all 56 notifiable diseases are available in digital format. For EIDs and specific epidemiological scenarios, eTW-CDC ISWG (Taiwan CDC investigation system working group) adopts the available template questions in the “question bank” (A1) to generate the template questionnaire (A2), herein, the time interval needed for this process is termed response time for generating template questionnaire (QRT), [fPSI (personnel-system interface)+SRT]. AEIS contains three system-operated galgorithms to filter and consolidate redundant questions (B1), generate a preliminary questionnaire (B2) and re-order questions into the final version of the questionnaire (C), and integrate data management (B3) and analysis; the time period is termed system response time (SRT), . The centralized database (E) which integrated clinical, epidemiological, and laboratory data from NDSS, NIIS, and AEIS, can perform real-time analysis upon request (F). Direction of information flow is indicated with arrows in solid lines; blue indicates system-operated automatic information flow, and black indicates information flow between personnel activity (reporting and data entry) and the system. The personnel response time, hPRT, is from having the final version of the questionnaire to the completion of the epidemic investigation and data entry (D), PRT. Results of the first-tier analysis will be fed back to three stake-holders (shown by dotted blue lines): (1) to Taiwan CDC ISWG for policy decision making, (2) to public health officers for fine-tuning on-site control measures, and (3) for border quarantine through the Autonomous Health Management System i(AHMS).
Figure 2.
Epicurve of the 51 acute notifiable infectious diseases and its mean personnel response time by week, 2006∼2008.
Left axis is the total weekly case number of the 51 notifiable acute IDs in Taiwan indicated in blue bars, January 1st 2006 through December 31st 2008; the right axis is the AEIS-related weekly mean personnel response times (PRT) shown in red line, February 13th 2006 through December 31st 2008; orange dotted line indicates the beginning of AEIS implementation. Several seemingly high mean PRT for the period coincided with the ongoing large Dengue outbreaks (marked with black arrows) occurring in 2006 through 2007; two more peaks of mean PRT (marked with grey arrows) were due to newly added one case each of HBV and HCV which were newly added to the list of IDs requiring epidemiological investigation.
Figure 3.
Personnel Response Time of (A) diseases category by transmission routes, and (B) areas by various population densities, February 13th, 2006 to December 31st, 2008.
The personnel response time (PRT) (line) was defined as the response time from case report/confirmation to epidemiologic investigation and data entry. The horizontal bar indicates ± two standard errors (SE). An asterisk demonstrates a p-value of <0.0001. (A) PRT by routes of transmission including respiratory, fecal-oral, vector-borne, blood-borne and contact transmission. (B) PRT by areas of various population densities. The population density was classified as high (3,000 or more people/km2), medium (637 to 2,999 people/km2) and low (636 or less people/km2).
Table 1.
Mean Response Times (days) during the epidemiologic investigation of shigellosis cases, before and after AEIS application.
Table 2.
Proportion of shigellosis and rubella identified by epidemiologic investigation before and after the application of AEIS.
Figure 4.
Rubella Cases in Taiwan (A) Geographical distribution of cluster cases, (B) Weekly Cases and (C) Duration of each cluster vs. mean PRT, 2007∼2008.
Duration of each cluster: the time interval between the dates of onset of the index case and the last case. For an asymptomatic rubella case, the date of taking the specimen was taken as the date of onset. (A) Six spatial-temporal clusters of confirmed rubella cases were identified during 2007 and 2008 (pie charts). All clusters were located in counties with incidence higher than 0.25/105 and foreigner density higher than 23/105. Within each pie, cases identified by epidemiologic investigation are demonstrate in red, and cases reported by a physician in light blue. (B) The 6 rubella clusters numbered according to the time of symptom onset of the respective index case of each cluster. The weekly case numbers are presented with red bars indicating cases identified by epidemiologic investigation and light blue bars as reported by a physician according to their onset dates in relation to the onset date of the index case. The mean PRT of each cluster with standard deviation and the duration of the cluster occurrence are presented at the bottom. The date of onset for each confirmed case to the date of index onset are shown in red (cases identified by epidemiologic investigation), and light blue (cases reported by a physician) bars.