Comparison of Degree of Concordance between 2 Systems of Dengue Reporting in Thailand, - PowerPoint Presentation

1. Comparison of Degree of Concordance between 2 Systems of Dengue Reporting in Thailand, 2015 Pantila Taweewigyakarn(1)Darin Areechokchai(2)Yongjua Laosirithavorn(2)Evelyn O. Talbott(1)(1) Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh(2)Bureau of Epidemiology, Department of Disease Control, Thailand

2. Introduction: Dengue Dengue : An infection caused by Dengue VirusMode of transmission : Vector-borne (Aedes Aegypti mosquito)Symptoms : Wide range Asymptomatic Infection Mild Symptoms (Dengue Fever) : Fever, headache, nausea, vomiting Moderate – Severe symptoms (Dengue Hemorrhagic Fever, Dengue Shock Syndrome) : Abnormal bleeding, Plasma Leakage Symptoms, Shock, Hepatitis, Renal Failure Death Burden of the disease : Worldwide: 3.2 million cases reported in 2015(1) f (approx 43.83 / 100,000 pop) Thailand : 144,952 cases reported ( 222.58 / 100,000 population) (1) WHO. Dengue and Severe Dengue. http://www.who.int/mediacentre/factsheets/fs117/en/.

3. Study Question Bhatt, 2013. estimated burden of Dengue in 2010 Worldwide: 96 million people should have clinical symptoms while WHO showed only 2.2 million cases reported Thailand : 1.37 million people with clinical symptoms , while the surveillance system captured 116,947 reported casesTo capture all of symptomatic dengue cases is impossible. At least the surveillance should detect the cases visiting hospitals And reflect the correct distribution in terms of Time, Place, Person in order to indicate an appropriate responseThus, Is the surveillance system data concordant with the real situation?Bhatt S. et al. The global distribution and burden of dengue. Nature. 2013 Apr 25;496(7446):504-7

4. Communicable Disease Surveillance system in Thailand (R506) – Electronic Based District Health office ProvincialHealth office Regional office of Disease Prevention and Control Bureau of Epidemiology(National level)Government Hospitals Private Hospitals1-2 wk1-2 wk1 mo1 moA data collection process is done manually by hospital epidemiologists.The process might vary in each hospital.

5. The Electronic National Health Record (ENHR) Data Center atProvincial Health office Health Data Center Bureau of Policies and Strategies, Ministry of Public Health Government hospitalsPrivate hospitalsPrimary Care Unit (Subdistrict Level)Auto SyncAuto SyncAuto Sync1 moThe system collects every single visit of patients and automatically sends the data to the database at provincial health office The data collected in 43 different folders, depend on types of variables e.g. demographic data, diagnosis, lab study, medications, etc.

6. Difference between R506 and ENHRAspectsR506ENHRObjective/ Utilization Surveillance dataRoutinely used for surveillance purposeHealth information of all Thai population; from Birth to Death Not routinely used Used for research, annual summary for some certain diseases e.g. chronic disease, Data collection Only notifiable communicable diseasesOnly variables that are relevant for disease prevention and control Hospital epidemiologists are supposed to collect the cases that meet the definition to be reported All diseases or purposes of hospital visit Most of information of patients e.g. demographic data, diagnosis code, laboratory results, prescription history admission information, cost of servicesThe data is electronically sent from the hospitals to the provincial data center => closed to “the real world data”Number of folder A single folder contains relevant variables 43 folders: PERSON, ADDRESS, DIAGNOSIS_OPD, CHRONIC, …. Etc. Size of data- Small Big data

7. Methods : Data Preparation Electronic National Health Records – using MySQL 6 folders selected : DIAGNOSIS_OPD(outpatient), DIAGNOSIS_IPD(inpatient), DEATH, PERSON, ADDRESS, HOMESelected Dengue cases from DIAGNOSIS_OPD : variable diagnosis code contains A90 or A91 (ICD-10)DIAGNOSIS_IPD : variable diagnosis code contains A90 or A91 (ICD-10)DEATH: 6 variables of disease/condition contribute to death contains A90 or A91(ICD-10) Join the cases from 3 folder together (Union command)Removal duplication of records using ID Keep the record of the day of last diagnosis, regardless of patients type (OPD/IPD)If the day of diagnosis is the same but the diagnosis is different between OPD/IPD, keep IPD diagnosis Join the non-duplicated dataset to the folder PERSON, ADDRESS, HOME (Left join Command)Selected variables : ID, Sex, Birthdate, Diagnosis code, Date of Diagnosis, Death status (Yes/No), province

8. Methods : Data Preparation Data from Communicable disease surveillance system (R506)Single folder De-identified data Selected variables : ID, Sex, Age in year, Disease code, Date of Visit, Death status (Yes/No), province

9. Methods: Data Analysis Time Period : Jan 1 – Dec 31, 2015 Inclusion criteriaAll records were included Exclusion Criteria When analyze each variables, the missing record would be excluded We study the distribution of the cases in term of sex, age group, race, months of diagnosis, address (province level).

10. Results Figure1. Number of Patients Diagnosed with Dengue recorded in the ENHR and the R506 surveillance databases by Area

11. Results Figure2 Number of Patients diagnosed Dengue by Disease severity  Figure3. Number of Patients diagnosed Dengue by Sex

12. ResultsFigure4 Number of Patients diagnosed Dengue Syndrome by Age Group

13. Results Figure7. Number of Patients diagnosed Dengue by Month of diagnosis

14. ResultsFigure8. Number of dengue cases by month and region (Bangkok excluded)

15. Discussion Similar distribution between data from the surveillance system and ENHRsex, age group, race, disease severity, and type of treatment Different distribution Geographical data : missing data in ADDRESS and HOME folder in ENHRTime of diagnosis : missing data in ENHR (an effect of fiscal year activities) The Surveillance system (R506) captured approximately 30% less than the cases recorded in ENHR Human-dependent process in the surveillance system The patients in NEHR might have the diagnosis changed later (The information in NEHR might be overestimated) Surveillance system detected more dead cases than ENHRPhysicians incorrectly reported a cause of death Huge missing data problem found in ENHR in ADDRESS, HOME folders Less problem seen in other folder such as DIAGNOSIS, PERSON and DEATH

16. Conclusion The data from R506 represent the similar pattern of distribution of dengue to the data from ENHR in terms of sex, age, race, age groups at national level The distribution of dengue cases in terms of time of diagnosis and places were not be able to be compared due to poor data quality of the ENHR. ENHR requires improvement of data quality so that the system is able to an effective data source for health study and research. Acknowledgement We would like to express our appreciation to the center of epidemiological informatics (CEI), Bureau of Epidemiology, Ministry of Public Health, Thailand for a great cooperation.

17. References World Health Orgainization(2016). Dengue and Sever Dengue. Retrieved from http://www.who.int/mediacentre/factsheets/fs117/en/ Rojanapithayakorn W.(1998). Dengue Hemorrhagic Fever in Thailand. Dengue Bulletin. 22.60-72Hammon, WM. (1973). Dengue Hemorrhagic Fever – Do we know its cause?. The American Journal of Tropical Medicine and Hygiene. 22(1).83-91Bhatt S. (2013). The global distribution and burden of dengue.Nature.496(7446):504-7Limpakanjanarat K.,Thiraratkul A.,Ungchusak K.(n.d.) Three decades of DHF Surveillance in Thailand, 1958-1987. Retrieved from http://apps.who.int/iris/bitstream/10665/146020/1/dn1991v16p11.pdfBureau of Epidemiology.(n.d.). Concept of Surveillance system adjustment. Retreived from http://www.boe.moph.go.th/files/meeting/506.pdf