Electronic ISSN 2287-0237

VOLUME

IMPLEMENTATION OF REAL-TIME LABORATORY-BASED INFLUENZA SURVEILLANCE SYSTEM, A NETWORK OF 20 HOSPITALS IN BANGKOK DUSIT MEDICAL SERVICES PUBLIC COMPANY LIMITED, THAILAND

SEPTEMBER 2017 - VOL.13 | ORIGINAL ARTICLE

Influenza is one of the significant causes of morbidity and mortalityglobally. The World Health Organization (WHO) reports thatglobally influenza accounts for 3 to 5 million severe cases and250,000 to 500,000 deaths annually, with the majority of deathsoccurring among high risk groups especially elderly populations.1 Themovement of international trade and travel worldwide has had an increasinginfluence on global transmission of emerging infectious diseases.2Thailand, a commercial and travel hub for Southeast Asia, has a risk ofimported infectious diseases such as influenza. The long-standingnational influenza surveillance system in Thailand known as the Thainational influenza center (NIC) was established by WHO as a nationalinfluenza laboratory in 1972. The primary objective of the Thai NIC isto surveillance emerging influenza viruses in public hospitals and toregularly provide newly insolated strains with WHO. Previous studieshave demonstrated the benefit of laboratory surveillance and itscapability to accurately detect influenza outbreaks earlier thansyndromic surveillance.3-9 Current laboratory surveillance has anapproximately 4-week lag due to laboratory test turn-around time, datacollection and data analysis.

As part of strengthening influenza virus surveillance inresponse to the 2009 influenza A (H1N1) pandemic, thereal-time laboratory-based influenza surveillance system, theBangkok Dusit Medical Services Surveillance System(BDMS-SS), was developed in 2010 by the Bangkok HealthResearch Center (BHRC). The primary objective of the BDMS-SSis to alert relevant stakeholders on the incidence trends of theinfluenza virus. Type-specific results along with patientdemographic and geographic information were available tophysicians and uploaded for public health awareness within24 hours after patient nasopharyngeal swab was collected. Thissystem advances early warning and supports better decisionmaking during infectious disease events.2 The BDMS-SSoperates all year round collecting results of all routinelytested respiratory clinical samples from participating hospitalsfrom the largest group of private hospitals in Thailand. In thisarticle, we introduce the surveillance systems for infectiousdiseases in BDMS. We use influenza surveillance to describea comprehensive surveillance structure that integrates informationfrom relevant health information systems from our hospitalnetwork.

Figure 1: Bangkok Dusit Medical Services Surveillance System (BDMS-SS); System Architecture Design

Study protocol

The real-time laboratory-based influenza surveillancesystem, the BDMS-SS constitutes a collaborativepartnership between multidisciplinary teams within theBMDS. The partnerships involved in this surveillancesystem are the Bangkok Health Research Center (BHRC),Chief Information Office department of Bangkok HospitalHeadquarters, Infectious disease control department ofBangkok Hospital Headquarters, Business Intelligencedepartment of Greenline Synergy Co., ltd. and InformationTechnology department of National Healthcare Systems co.,ltd. (N-health). A protocol for influenza surveillance waswritten with the technical support of clinical epidemiologistsbased on CDC criteria.

Study setting

A network of 20 private hospitals representing all 5 regionsof Thailand is currently participating in the BDMS-SS. Thesurveillance sites were selected based on their accessibility topatients, adequate specimen storage capacity, and establishedtransportation system to the N-health.

Sample collection

Laboratory tests are requested by clinicians in charge ofpatient care from each hospital. Clinicians decide the test eachpatient needs, and which types of samples need to be taken,and when. The most common sample types reported to thissystem are nasopharyngeal aspiration, tracheal secretion andnasal and throat swab. All samples for influenza virus testingwere sent to N-health tested by rapid influenza diagnostic tests,multiplex real time polymerase chain reaction (PCR) andenzyme immunoassay (EIA) (approved by FDA and/or CE/IVD). Review meetings with all stakeholders were organizedevery four months, as part of a strategy to improve the surveillancesystem by identifying strengths and areas of concern duringthese meetings.

Data collection and management

The system systematically collected information from allinfluenza test results from N-health which provides medicallaboratory results and shared services for a network of 20hospitals. The test results were submitted automatically eachday into the Bangkok hospital data warehouse by the BusinessIntelligence department of Greenline Synergy Co., ltd.. Thenext step is for the Lab Influenza surveillance businessintelligence system, developed by the Chief Information Officedepartment of Bangkok Hospital Headquarters, to retrievethose test results and integrate these with health informationfrom the data warehouse, verify and transform the data andfinally display the information on the business intelligencesystem (Intranet). Data submission is carried out through asecure online data submission tool. The process of datacollection, management and application for BDMS-SS has beenapproved by the CIO of the Bangkok Hospital Headquarters.

Laboratory testing

All samples from participating hospitals were tested byN-Health using rapid test and reverse transcription polymerasechain reaction (r-PCR). All laboratories validated their assaysappropriately. Both positive and negative results of influenzaare submitted to N-health every day throughout the year. Inaddition, the influenza subtyping results were reported by all20 affiliated hospitals.

Statistical analysis and feedback

The proportion of samples positive (positivity) for virusesunder surveillance is calculated (based on weekly samplestested) by virus type, sex, age group (<2, 2-4, 5-14, 15-49,50-64 and >65 years) and residential area using weekly numberof positive detections divided by the weekly total number ofsamples tested. The data are analyzed to determine the trendsand predominant virus types. This information is essential forrisk assessment and estimation of morbidity, mortality, and thepotential impact of the influenza on the community and onhealthcare services.

To detect outbreaks as early as possible, an alarm wastriggered when the weekly count of proportion of positivecases was higher than 2 SDs of the mean of 5 year historicaldata. A statistical alarm is triggered if the observed value issignificantly different from the expected value. After checkingbiological criteria, alarms were assessed as confirmed orunconfirmed by a senior epidemiologist during the weeklysurveillance meeting. Alarms that were escalated into an alert(after further investigations that included descriptive analysisof cases in terms of time, place, and population) led to furtherepidemiologic investigations, which then had to be declaredto the infectious unit of those sites if a real outbreak wasconfirmed.

For feedback, a weekly epidemiologic report is producedby the BHRC and addressed to all participating hospitals andThai-MOPH. This report is based on the BDMS-SS data ofinfluenza, to track the weekly number of positive results andweekly proportion of positives (positivity) by sampling week,age group, sex, location and virus type. This report is alsoprovided on the website www.bangkokhealth.com.

 

From October 2010 to April 2017, a total of 482,789subjects were tested and 86,110 (17.8%) cases of influenzawere identified. Of those who tested positive for influenza theywere aged 65 years old (3.7%). Approximately 50% ofsubjects were male and female. Of these, 40,552 (47.0%) wereinfluenza type B, 31,412 (36.4%) were influenza A unspecifiedsubtype, 6,181 (7.2%) were influenza A H1N1, 4,001 (4.6%)were influenza A H3N2, 3,835 (4.4%) were influenza Aseasonal and 196 (0.4%) were respiratory syncytial virus(RSV) (Figure 2).

The number of influenza-positive specimens reported bythe real-time influenza surveillance system were from week40, 2015 to week 39, 2016. A total of 117,867 subjects weretested and 17,572 (14.91%) cases tested positive for the influenzavirus (Figure 3). Based on the long-term monitoring of collectedinformation, this system can delineate the epidemiologic patternof circulating viruses in near real-time manner, which clearlyshows annual peaks in winter dominated by influenza subtypeB in 2015-1016 season. This surveillance system helps toprovide near real-time reporting, enabling rapid implementationof control measures for influenza outbreaks.

This article presents the findings of a new laboratory-basedinfluenza surveillance system, which was developed andimplemented across 20 hospitals of BDMS. This novel,real-time laboratory-based surveillance system was automaticallyuploaded and an aggregated influenza test correlated withdemographic information. The system provides useful informationin a timely fashion which has contributed to describe theepidemiology of the influenza virus. The system was alsosuccessfully used for real-time, daily surveillance amonghospitals of the BDMS group. This daily surveillance operationcan pick up early signs of increased activity of influenza andthis increases the ability to undertake a prompt response andto take appropriate action. Surveillance for influenza iscommon practice in many countries but near real-timesurveillance is not common.10 This system thus provides a newmechanism to monitor the epidemiology of influenza in atimely fashion in Thailand. Expansion of this near real-timecapability to a public health agency could improve both localand national surveillance and response lead to reduce the needfor syndromic influenza surveillance. With a large sample sizefrom the BDMS networking hospitals, this system provides arobust supplementary mechanism, through the collection ofroutinely available laboratory data at minimum extra cost, tomonitor influenza in private hospitals, Thailand.

This surveillance system was the first real-time, dailyreporting surveillance system to report on the largest data baseof private hospitals in Thailand and provides timely reportsand feedback to all stakeholders. It provides an importantsupplement to the routine influenza surveillance system inThailand. This illustrates a high level of awareness and willingnessamong the BDMS hospital network to report emerging infectiousdiseases, and highlights the robust and sensitive nature ofBDMS’s surveillance system. This system demonstrates theflexibility of the surveillance systems in BDMS to evaluate toemerging infectious disease and major communicablediseases. Through participation in the Thailand influenzasurveillance network, BDMS can more actively collaboratewith national counterparts and use its expertise to strengthenglobal and regional surveillance capacity in Southeast Asia,in order to secure advances for a world safe and secure frominfectious disease. Furthermore, this system can be quicklyadapted and used to monitor future influenzas pandemics andother major outbreaks of respiratory infectious disease, includingnovel pathogens.

Figure 2: Laboratory confirmed influenza cases, Bangkok Hospital Group, Bangkok Dusit Medical Services Public Company Limited, October 2010 to April 2017 (RSV: Respiratory syncytial virus)

Figure 3: Weekly distribution of number of collected samples with influenza viruses, Bangkok Hospital Group, Bangkok Dusit Medical Services Public Company Limited, September 2015 to October 2016