Early Warning Disease Surveillance After a Flood Emergency — Pakistan, 2010

During July–August 2010, Pakistan experienced extreme flooding that affected approximately 18 million persons. In response to the emergency, Pakistan's Ministry of Health and the World Health Organization (WHO) enhanced an existing disease early warning system (DEWS) for outbreak detection and response. This report summarizes surveillance results early after implementation, describes system usefulness, and identifies areas for strengthening. Daily disease counts were reported from health facilities in four provinces containing 98% of the flood-affected population. During July 29, 2010–September 15, 2010, approximately 5.6 million new patient visits were reported. The most frequent conditions reported were skin diseases (18.3%), acute respiratory infection (15.1%), and acute diarrhea (13.3%). A total of 130 outbreak alerts were documented, of which 115 (88.5%) were for acute watery diarrhea (AWD) (suspected cholera). Of these, 55 alerts (47.8%) had at least one microbiological sample with confirmed cholera. Overall, DEWS was useful in detecting outbreaks, but it was limited by problems with data quality. Improvements in DEWS have increased system usefulness in subsequent emergencies. This report highlights the need to follow updated WHO guidelines on early warning disease surveillance systems to improve their usefulness (1).

Background

In emergencies before the 2010 floods, the Pakistan National Institute of Health conducted outbreak surveillance in some provinces using an existing DEWS. Severe flooding in July and August of 2010 resulted in >1,700 deaths, damaged or destroyed 1.9 million homes, and left at least 10 million people without shelter. This led to the largest international appeal ever (USD 2 billion globally) for humanitarian assistance and a need for an expanded DEWS (2).

Postflood Implementation of DEWS

After flooding began, DEWS was expanded to a national system covering all flood-affected districts in the country. The primary objective of the system was early outbreak detection and control (Table 1). Disease reporting through this system began in July 2010. In August 2010, WHO requested CDC assistance to strengthen DEWS. Operational guides with standardized case definitions and reporting forms were distributed (3), and national and provincial surveillance staff members were trained. Fixed health facilities and mobile clinics in flood-affected areas were expected to report case counts of 13 conditions considered to be epidemic-prone or of public health importance.* Information was compiled daily at the district, provincial, and national levels, and a national epidemiological bulletin showing aggregated data was issued the following day (4).

DEWS also included an immediate disease alert and response component to meet its primary objective. Most diseases in DEWS had a defined alert threshold (Table 2) that triggered notification of surveillance staff members and outbreak investigation teams. Laboratory confirmation included onsite rapid diagnostic tests and microbiological testing at the national public health laboratory.

The rapid expansion of DEWS was supported by using surveillance personnel and mechanisms for disease reporting from existing provincial systems. Additional resources from vertical, field-based programs, communicable disease programs, and other health programs also were widely used. Lastly, provincial health departments actively supported DEWS implementation by facilitating the training of surveillance officers and mandating disease reporting from district health officers.

Surveillance Results

Daily reporting began on July 29, 2010. The average weekly number of reporting sites fluctuated between 958 and 1,948 sites for the first 6 weeks. By mid-September 2010, DEWS covered 81 (67.5%) flood-affected districts of the country's 120 districts.

During July 29–September 15, 2010, a total of 5,618,902 patient visits were reported to DEWS. Of those, 2,174,368 (38.7%) were for a reportable condition, primarily including 850,292 (15.1%) visits for acute respiratory infection, 745,532 (13.3%) for acute diarrhea, and 327,453 (5.8%) for unexplained fever. In some areas, data on additional conditions were collected using nonstandardized forms and included skin diseases, dog bites, snake bites, eye and ear infections, injuries, and heat stroke. Of these, skin diseases were the most commonly reported, with 1,029,942 (18.3% of total) visits.

In the same period, 130 outbreak alerts were generated, of which 115 (88.5%) were for AWD. Another seven (5.4%) disease alerts were for suspected measles, two (1.5%) were for acute flaccid paralysis, and two (1.5%) were for suspected meningitis. Of the AWD alerts, 82 (71.3%) had at least one microbiological sample submitted, with 55 (67.1%) of these samples testing positive for Vibrio cholerae. None of the cases of suspected measles, acute flaccid paralysis, or suspected meningitis were laboratory confirmed as measles, polio, or bacterial meningitis.

Reporting Challenges

During its rapid implementation, DEWS encountered several challenges common to disease early warning systems established during disasters (5,6). First, application of nonstandard case definitions varied. For example, in Punjab, 311,882 patients with suspected malaria were recorded as having unexplained fever and only 772 confirmed cases were reported as suspected malaria. In contrast, in Sindh, suspected and confirmed malaria cases both were reported as suspected malaria (n = 168,302). This affected national estimates. Disease misclassification made the aggregated national data inadequate for identifying and monitoring disease trends.

Second, acceptance of standardized reporting forms varied because some diseases considered important by provincial authorities were not specifically included on DEWS forms (e.g., skin diseases). Hence, provinces used nonstandard forms, which led to the reporting of multiple disease categories inconsistent with standardized DEWS case definitions. A prominent example was diarrhea. In practice, diarrhea was captured as acute diarrhea, bloody diarrhea, AWD, suspected cholera, gastroenteritis, or other diarrhea, depending on the reporting location.

Third, data rarely were analyzed at the district level or lower because staff members were fully occupied fulfilling daily reporting requirements. Data were analyzed and reported nationally, but most outbreak alerts were based on reports of small numbers of cases reported immediately by telephone or e-mail at the local level.

Fourth, disease reporting was difficult to monitor. Lack of reliable information on functioning health facilities and their catchment populations made it difficult to determine timeliness and coverage. Sites reporting fluctuated daily and late or missing reports were difficult to track because hundreds of sites reported daily. Although DEWS covered most government health facilities, not all partners delivering health services (e.g., nongovernmental organizations) participated in the system. Data were analyzed using proportionate morbidity and case counts, but with uneven reporting, trends reported in epidemiological bulletins were of limited usefulness.

Recent Situation

In 2011 and 2012, Pakistan again experienced heavy monsoon flooding, which affected >4 million persons each year (7,8). National DEWS continued to operate with weekly reporting and captured 45,510,570 patient visits and 5,752 disease outbreak alerts in 2011 (4). Weekly bulletins were expanded to include subnational trends and outbreak investigation results. Despite early challenges, DEWS remained important for outbreak detection in the absence of other outbreak detection systems.

Reported by

Guido Sabatinelli, MD, Suzette Rene Kakar, MD, Musa Rahim Khan, MD, Mamunur Malik, MBBS, World Health Organization, Pakistan. Birjees Mazher Kazi, MBBS, Pakistan National Institute of Health. Jahanzeb Khan Aurakzai, MBBS, Pakistan National Emergency Preparedness and Response Network. Michelle Gayer, MBBS, Disease Control in Humanitarian Emergencies, World Health Organization. Farah Husain, DMD, Muireann Brennan, MD, Oleg Bilukha, MD, PhD, Irshad Shaikh, MD, PhD, Susan Cookson, MD, Div of Global Disease Detection and Emergency Response, Center for Global Health; Cyrus Shahpar, MD, EIS Officer, CDC. Corresponding contributor: Cyrus Shahpar, [email protected], 770-488-3920.

Editorial Note

This report describes the implementation of a postdisaster DEWS in Pakistan. The challenges of DEWS implementation mirror those of other early warning alert and response network (EWARN) surveillance systems, which have been documented in many emergencies, including Sudan, Darfur, Haiti, and Pakistan, and discussed at two WHO technical workshops (9,10).

The primary objective of an EWARN system is early detection of and response to epidemic-prone diseases. In a major emergency, EWARN systems should be implemented expeditiously and should focus on that objective. Alerts of typically rarer diseases should trigger timely investigation and control measures. In practice, however, EWARN systems frequently include monitoring of other infectious diseases of public health importance, that occur more frequently, as they did in Pakistan. This is problematic because reporting of these more common diseases can overwhelm resources, negatively affect data quality, and potentially detract from outbreak detection.

Existing nonemergency surveillance systems can be used to capture information on the more common diseases, or reporting from select sentinel sites might suffice to assess trends. In this case, reporting sites should be chosen based on reliability of reporting, representativeness, and other factors that maximize data quality and the ability to respond in a timely manner.

Implementation and coordination of EWARN systems must be improved. Data collection forms with standardized case definitions should be developed in consultation with local partners to maximize acceptance and should be widely distributed in paper form and electronically. Systems should be designed to include available technologies (e.g., short message service data collection), but also have contingency plans should infrastructure fail. Multiple training sessions are required because of high staff turnover in emergencies. Local staff members should be trained to enter and analyze data and receive frequent feedback, so they can use the information they report for public health action and appreciate the benefits of reporting.

In 2010, DEWS exemplified the value and challenges of early warning disease surveillance, and it was a functional system despite the massive scope of the emergency. In 2012, WHO released updated operational guidelines on EWARN implementation, based on evidence gained from prior implementations in Pakistan and other countries (1). These guidelines target many of the documented challenges of EWARN implementation. Further evaluations are needed to determine whether adherence to new guidelines results in improvements in the quality and usefulness of surveillance data.

References

1. World Health Organization. Outbreak surveillance and response in humanitarian emergencies: WHO guidelines for EWARN implementation; 2012. Geneva, Switzerland: World Health Organization; 2012. Available at http://www.who.int/diseasecontrol_emergencies/publications/who_hse_epr_dce_2012.1/en/index.html. Accessed March 30, 2012.
2. United Nations Office for the Coordination of Humanitarian Affairs. Pakistan floods relief and early response recovery plan revision, November 2010. Islamabad, Pakistan: United Nations Office for the Coordination of Humanitarian Affairs; 2010. Available at http://reliefweb.int/report/pakistan/pakistan-floods-relief-and-early-recovery-response-plan-revision-november-2010. Accessed December 10, 2011.
3. World Health Organization. Outbreak surveillance and response, disease early warning system, flooding response in Pakistan, operational guidance, August 2010. Geneva, Switzerland: World Health Organization; 2010. Available at http://www.who.int/hac/crises/pak/pakistan_operational_guidance_flooding_august2010.pdf. Accessed February 8, 2011.
4. National Institute of Health, World Health Organization. Disease early warning system and response in Pakistan. Wkly Epidemiol Bull 2012:3(1).
5. CDC. Launching a national surveillance system after an earthquake—Haiti, 2010. August 6, 2010. MMWR 2010;9:933–8.
6. CDC. Rapid establishment of an internally displaced persons disease surveillance system after an earthquake—Haiti, 2010. MMWR 2010;9:939–45.
7. US Agency for International Development. Pakistan—floods. Fact sheet no. 1, fiscal year (FY) 2012. October 3, 2011. Washington, DC: US Agency for International Development; 2012. Available at http://reliefweb.int/sites/reliefweb.int/files/resources/10.03.11%20-%20usaid-dcha%20pakistan%20floods%20fact%20sheet%20%231%20-%20fy%202012.pdf. Accessed October 11, 2011.
8. United Nations Office for the Coordination of Humanitarian Affairs. Pakistan: monsoon 2012 situation report no. 2. October 3, 2012. Islamabad, Pakistan: United Nations Office for the Coordination of Humanitarian Affairs; 2012. Available at http://reliefweb.int/sites/reliefweb.int/files/resources/OCHA_Pakistan_Monsoon_2012_Sitrep2-3Oct2012.pdf. Accessed December 7, 2012.
9. World Health Organization. Early warning surveillance and response in emergencies. Report of the WHO technical workshop. December 7–8, 2009. Geneva, Switzerland: World Health Organization; 2009. Available at http://whqlibdoc.who.int/hq/2010/who_hse_gar_dce_2010.4_eng.pdf. Accessed February 1, 2011.
10. World Health Organization. Early warning surveillance and response in emergencies. Report of the second WHO technical workshop. May 10–11, 2010. Geneva, Switzerland: World Health Organization; 2010. Available at http://whqlibdoc.who.int/hq/2010/who_hse_gar_dce_2010.4_eng.pdf. Accessed March 1, 2012.