The following report was submitted to the US Government in September 2010 for internal publication. This was official documentation of the existence of the world's first infectious disease forecasting center. The HEAS became much more during the course of operational engagement in Haiti.
T H E H A I T I E P I D E M I C A D V I S O R Y S Y S T E M
James M. Wilson V, M.D., Jane W. Blake, M.S., Heather E. Baker, M.S., and Heather Keatley, B.S.
Haiti Earthquake Intercept Team, Praecipio International, Seattle, Washington, USA
The 2010 earthquake in Haiti severely reduced medical and public health capacity and, like all catastrophic events, challenged continuity of government. After the event, representatives of the Haitian government and foreign government responders faced very difficult choices.
Answering Haiti’s calls for help, international aid agencies and non-government organizations sought to fill the void, donating medicines, equipment, and medical treatment. While foreign medical practitioners worked side by side with local doctors to perform triage, the authors of this paper volunteered their time to Haiti’s people as members of a new professional discipline, operational biosurveillance.
This is a report of the authors’ Haiti response – generating significant benefits solely from time volunteered by three experienced professionals drawing upon a very limited US$10,000 in donated expense funding.
During the 2009 Mexico influenza outbreak, the authors had shown themselves capable of delivering significant (more than two weeks) early warning of a major infectious disease event without the support of government infrastructure. In Haiti, the authors delivered accurate advanced forecasts, warnings, and alerts for a multitude of key infectious disease events that were often days, weeks, and months ahead of any alternative. This increased the preparedness and response capability of medical personnel and reduced threats of social disruption caused by disease.
The authors describe their process and lessons learnt in this report.
The 2010 Haiti earthquake was a catastrophic event. The earthquake epicentre was located near the town of Léogâne, approximately 25 km (16 miles) west of Port-au-Prince, Haiti's capital. The earthquake occurred at 16:53 local time (21:53 UTC) on Tuesday, 12 January 2010 and registered a magnitude of 7.0 Mw.
The earthquake caused major damage to Port-au-Prince, Jacmel and other settlements in the region. Many notable landmark buildings were significantly damaged or destroyed, including the Presidential Palace, National Assembly building, Port-au-Prince Cathedral, and main jail. The headquarters of the United Nations Stabilization Mission in Haiti (MINUSTAH), located in the capital, collapsed, killing many, including the Mission's Chief, Hédi Annabi. At least 56 aftershocks measuring 4.5 Mw or greater had been recorded.
An estimated three million people were affected by the quake; Prime Minister Jean-Max Bellerive stated on 3 February 2010 that 200,000 people had been identified as dead and estimated that 300,000 injured had been treated. Among those killed were Archbishop of Port-au-Prince Monsignor Joseph Serge Miot, and opposition leader Micha Gaillard. Resident medical and public health professionals in Haiti were also killed, which affected the ability of the public health and medical community to respond. Bellerive, in addition, estimated that 250,000 residences and 30,000 commercial buildings had collapsed or were severely damaged. Over 1,300 educational establishments and 50 hospitals and health centers collapsed or were rendered unusable by the earthquake. Health sector damages accounted for 6% of the total monetary costs attributed to the earthquake.
PHASE 1: ACTIVATION OF THE HAITI EPIDEMIC ADVISORY SYSTEM (HEAS)
Understanding that Haiti would need the immediate support from all sectors, our team – comprised of Dr. James Wilson V, Ms. Jane Blake, and Ms. Heather Baker – activated operations through our non-profit organization, Praecipio International, with a mission to use our professional skills and network to aid in disaster response and recovery in the area of operational biosurveillance.
Working initially with donations of time, our “Intercept Team” activated a Haiti Epidemic Advisory System (HEAS) as an adjunct infectious disease crisis anticipatory and early warning capability. The Intercept Team was aware of communication challenges between public health organizations outside of Haiti and their counterparts in-country due to the scale of destruction and loss of life.
Within minutes of notification of the earthquake by the Global Disaster Alert and Coordination System, the Intercept Team initiated online source scanning in English and other relevant languages, such as French and Spanish, to gain an initial understanding of infectious disease events reported in the affected areas of Haiti. We then conducted a search of peer-reviewed literature for relevant baseline knowledge of disease activity in Haiti. Shortly thereafter, the Intercept Team made publicly available our initial assessment in the earthquake’s immediate aftermath:
- Large-scale epidemics are not typically triggered by earthquakes;
- Seasonal endemic diseases are still observed such as malaria and dengue, which may complicate response; and,
- Corpses do not typically trigger epidemics.
The Intercept Team then sought to frame a biosurveillance assistance strategy for Haiti, first focusing on those infectious diseases capable of generating crisis or disaster level impact in their own right, separate from the physical destruction triggered by the earthquake.
To determine which of the infectious diseases endemic to Haiti were candidate disruptors, the Intercept Team reviewed all available peer-reviewed literature on Haiti indexed by the United States National Library of Medicine’s PubMED; all archival reports of infectious disease activity in Haiti and the Dominican Republic available in the Program for Monitoring Infectious Diseases (ProMED); French, Spanish, and English language archival media reporting available in Google News; and all available information in the Global Infectious Disease & Epidemiology Network (GIDEON). It was the Intercept Team’s observation that there was a relative low volume of peer-reviewed and open source publications regarding infectious disease endemic baselines in Haiti.
The Intercept Team monitored Hispaniola Island, which included the Dominican Republic and Haiti, as well as other regional countries exchanging air traffic with the disaster zone such as Puerto Rico, Cuba, and the United States. The team monitored not only reporting by residents of Haiti using SMS captured by InSTEDD and Ushahidi, but the multi-national responders who had provided commentary to the media, engaged in online discussion in Google groups and listservs, posted on blogs, and engaged in Twitter-based communication.
Praecipio International published an initial review of candidate targets for the Intercept Team to monitor in a Situation Report on 17 January (Day 5 post-earthquake), entitled Prioritized Diseases for Operational Biosurveillance During Disaster Response in Haiti.
After publication of the initial SitRep, the Intercept Team released Critical Information Requirements (CIRs) to the responder community, in the form of non-technical social indicators of possible infectious disease crises. It had been the team’s experience that efforts spent to promote recognition and reporting by non-medical personnel as well as clinicians on the ground would be in demand. These requirements were posted on 19 January and later updated on 3 April as the disaster situation in Haiti evolved.
PHASE 2: FOSTERING AN INFORMATION NETWORK
Prior to the 2010 Haiti Earthquake, the Intercept Team was fortunate to have worked together in government and private enterprise in the area of operational biosurveillance for nearly a decade. Through time, the Intercept Team had developed an appreciation for the power of open forums for the exchange of critical information.
On 4 April 2009, the Intercept Team, then working as lead analysts of a private operational biosurveillance group serving primarily corporate clients, had issued early alerts and warnings of the influenza outbreak in Mexico to the international community. This recent experience of exchange and working in emerging crisis situations informed our team’s approach in Haiti.
As a non-government organization directed by operational biosurveillance professionals with a recent and significant track record of success, Praecipio International was quickly able to foster an independent and peer-accepted open forum for an exchange of critical information through the HEAS.
Two approaches were engaged in fostering the HEAS network:
- The team approached individuals and organizations on the ground, in listservs, and in online forums that spontaneously formed during the response effort. Information and access to the team was provided collaboratively. This openness encouraged multi-lateral communication about infectious disease events.
- The team invited interested individuals and groups to a HEAS-specific forum. The Intercept Team provided support in connecting people to each other as well as to us. This resulted in cross-migration of interested parties into the HEAS forum where focused dialog about live infectious disease events could occur.
The Intercept Team utilized its international network of operational biosurveillance participants organized and used by the team during the process of warning and situational awareness in the 2009 H1N1 influenza pandemic. Crisis Mappers and the All Partners Access Network (APAN) were also utilized. Later, during Phase 3 (see below), information flow was augmented through direct face-to-face engagement with key contacts in Haiti who were selected based on their indigenous experience, subject matter expertise, and position within the social network on the ground involved in public health, veterinary medicine, agriculture, human clinical medicine, and humanitarian response.
Several Haitian public health and medical personnel were invited to join the HEAS partnership; however, their participation was limited due to lack of Kreyol language capacity within the Intercept Team, poor communications with resident Haitians due to downed infrastructure, and reluctance by the Intercept Team to interfere with Haitian officials’ authority to issue public proclamations of public health information. It is the assessment of the Intercept Team that additional budget, if it had been available, could have reduced these challenges.
Information flow evolved to a fusion of open source and trust network processing by astute analysts contextualized into the following formats:
- Critical Information Requirements. This was a focused list of non-technical indicators released to the participants through email, web portal, and Google group. The team initially contemplated releasing this guidance to the general Haitian public but Kreyol translation capacity was not present, and the team was reluctant to interfere with local authorities’ role in risk communication with their public. At the time, the main official-public communications platform was Kreyol radio broadcast.
- Tactical Reports. These were rapid releases of single reports at an initial stage of verification and contextualization. This information triggered dialog with networked responders, some of whom were known to the team, and some serendipitously approached us through either networked HEAS partners or via our online Twitter and website presence.
- Advisories. These reports represented contextualized rapid assessments designed to warn the responder community of emerging risk related to a particular infectious disease.
- Situation Reports (“SitReps”). These reports represented periodic, summary assessments of the overall monitoring environment, which included contextual issues such as current status of infectious disease reporting, medical infrastructure status, civil unrest as it pertained to infectious disease activity, and other issues relevant to operational biosurveillance. SitReps also contained forecast information regarding expected infectious disease activity over the subsequent 30-60-90 day period.
Again, the process of social networking was directly inspired by the Intercept Team’s experience and involvement with other similar social constructs that have been used in public health and health security surveillance and warning efforts. Lessons learned in these experiences have repeatedly indicated the strength of utilizing an informal network to compliment formal, hierarchical networks managed by ‘officials’. Specifically, we drew on lessons learned by the public health infectious disease crisis warning community through development of the World Health Organization (WHO) International Health Regulations and the Global Outbreak and Alert Response Network (GOARN), the National Biosurveillance Integration Center (NBIC) and the Biosurveillance Indication and Warning Analysis Community (BIWAC).
The social construct of informal network the team opted for with the HEAS had the following features:
- “Everyone is in charge, and no one is in charge”… but with facilitation. The team provided facilitation and moderation of comments within the HEAS forum in order to promote information flow. This became necessary, as the role of the team evolved quickly into neutral third-party brokers of information.
- Participation was voluntary. This was a key element of trust. However, the caveat was accountability - the group inevitably evolved into a peer review construct where maladaptive behaviors such as deliberate lack of information sharing or aggressive behavior were discussed openly and promptly. This was a productive outcome, as it encouraged the evolution of transparency and trust among the participants.
These examples highlight several observations that supported the peer-reviewed literature:
- Emergence of informal networks, especially during crises, is inevitable.
- Cooperation between formal and informal networks can be extremely powerful and promote a productive atmosphere of trust.
- Formal-informal networks promote adaptive fitness- positive, growing and learning resilience over time.
- Accountability and peer-review are resultant characteristics.
The HEAS partnership ultimately included 237 partners associated with international organizations (IGOs), non-governmental organizations (NGOs), U.S. federal organizations, non-profit organizations, Haitian officials, and unaffiliated volunteer Haitian and non-Haitian responders by Day 150.
PHASE 3: FIELD DEPLOYMENT & BOLSTERING THE HEAS THROUGH ACTION
By March 2010, the Intercept Team had raised approximately US$10,000 in private donations to cover the modest travel expenses of a limited deployment to Haiti. Our Intercept Team had the following objectives for ground deployment in Haiti:
● Address the challenge of limited data through ground validation of information conducted via face-to-face informal interviews in Haiti with resident experts; and,
● Support the calibration of baseline infectious disease information for Haiti.
The Intercept Team’s Dr. Jim Wilson deployed beginning on 5 March (Day 52). Upon arrival, Dr. Wilson focused on the greatest areas of uncertainty. Assessments of threats posed by several diseases such as anthrax were adjusted after conversations with multiple resident subject matter experts and personnel who were directly involved with prevention efforts, including vaccinating livestock for anthrax.
The list of HEAS-focus diseases evolved over time as additional data was collected.
The list of prioritized diseases for Haiti was eventually divided into two general categories: disruptors and non-disruptors. Disruptors included those infectious diseases capable of generating a crisis or disaster (see below for definitions) in their own right. The emphasis was on those diseases capable of inducing significant strain on medical personnel and medical supplies, or generating significant and acute community disruption. Attempting to define crises within the context of a disaster response was challenging, as explained below. Non-disruptors included infectious diseases clinicians may see occasional cases of but were unlikely to overwhelm a clinic or hospital.
MEASURING THE RESULTS – (A) – HEAS PRODUCTIVITY
The driver for release of HEAS Situation Reports was receipt of credible information to indicate a change in risk status and advisory levels, or substantive updates regarding contextual information.
HEAS produced reports available to responders beginning on 17 January (Day 5) until beyond the period of analysis in this report (11 June, Day 150). During this time period, 14 Situation Reports and 115 Tactical Reports were produced. This was a greater number of releases than from ‘official’ sources.
During this time ‘official’ reports were released by the Haitian Ministry of Health (HMoH), the Pan American Health Organization (PAHO), and the United States Centers for Disease Control and Prevention (CDC).
● PAHO produced situation reports related to infectious disease surveillance from 15 January (Day 3) to 1 March (Day 48), for a period of coverage spanning 45 days. PAHO produced 9 Situation Reports during this period that contained information specifically related to infectious disease activity.
● On 10 March (Day 57), the first meeting was held of the Epidemiology SubCluster that ultimately led to the announcement of the newly created IDPSS (Internally Displaced Persons Surveillance System). At the suggestion of the Intercept Team, CDC created the IDPSS Google group in support of collaborative communications on 16 March (Day 63).
● PAHO’s reporting on infectious disease activity ceased completely when the IDPSS reports entered production on 17 March (Day 68). On 22 March (Day 69), the first IDPSS report was posted to the IDPSS Google group. IDPSS produced 11 total reports, including the 22 March report, from activation to Day 150.
Geographic coverage for the HEAS was biased towards the immediate earthquake-impacted areas of Haiti, as was PAHO and IDPSS reporting. Occasional reporting was received from other parts of Haiti such as the northern and southern peninsulas; however, such reporting was sporadic and was reflective of the greater response presence focused on the greater Port-au-Prince area.
Table 1 displays the information sources used to compile the summary timeline.
Key points on the timeline of infectious disease event reporting are listed below in Table 2:
MEASURING THE RESULTS – (B) – HEAS REPORTING DELTAS
Side-by-side comparison (see Table 3, below) of HEAS, PAHO and IDPSS SitReps shows substantial differences in reporting of infectious disease events in the majority of cases, with the HEAS providing more expansive disease coverage and advanced notice. In two cases, involving a chickenpox infectious disease event and another involving suspected measles, ‘official’ reporting was faster than the HEAS.
Several key infectious disease events were documented by HEAS partners not reported by public health officials such as infections, outbreaks, and fatalities due to infectious disease among aid workers; translocation of multiple infectious diseases to the United States; and other infectious disease either suspected or confirmed. These events represented important considerations for aid worker health, treatment of Haitian displaced persons and evacuated victims, and local and state public health officials receiving returning aid workers and evacuees from Haiti. In one example, translocation of dengue to Florida, HEAS partners reported cases of dengue seen in Florida hospitals in the first week of February; CDC subsequently issued a travel advisory for dengue infection in relief workers returning to the U.S. on 27 April 27, more than two months later.
Two diseases were assessed to be potential disruptors and thus, of priority concern to the team: diarrheal and respiratory disease in children. This was due to the intersection of limited pediatric-specialized care, disrupted access to safe drinking water and sanitation, and overcrowded camps. One disease, malaria, was not assessed to be a likely disruptor but provided a brief challenge to this assessment as seasonal activity increased and briefly generated limited patient demand and concern in the ad hoc medical community. The team balanced uncertainty in risk assessments due to lack of peer-reviewed data, limited reporting infrastructure, and changing human and environmental conditions, such as sanitation challenges in the context of heavy rainfall, when deciding when to upgrade or downgrade alerts intended to sensitize medical responders.
Figure 1 displays the timeline of diarrheal disease event reporting and alerts issued by HEAS and public health officials. Diarrheal disease was a well-known endemic problem in Haiti responsible for significant pediatric morbidity and mortality. However, specific data regarding geotemporal patterns of pathogen activity was extremely limited. Assessments were based on available epidemiological information, historical environmental and meteorological data, and open source information regarding prior flood crises where diarrheal disease was reported. These assessments were initially complicated by historical reports of cholera many years prior without clear indication in the peer-reviewed literature or other credible sources to indicate Haiti was truly cholera-free. The endemic presence of cholera would have represented a more serious consideration than other agents of diarrheal disease. Reporting by credible non-profit and international organizations that had been present in Haiti for many years indicated a concern for “cholera” and watery disease, which further complicated attempts to gain initial clarity regarding true risk.
Figure 1. Timeline of risk and tactical reporting of diarrheal disease and typhoid activity in Haiti, as reported by HEAS, PAHO, and IDPSS.
HEAS reporting on 21 January (Day 9) indicated high and increasing prevalence of diarrheal disease 7 days ahead of public health reporting, as shown in Figure 1. Typhoid activity reported by HEAS on 22 January (Day 10) was 2 months ahead of public health reporting on 22 March. HEAS partners reported suspected typhoid-related intestinal perforation to be an important cause of pediatric mortality. Mortality due to infectious disease in general was not reported by public health officials in the first 150 days post-earthquake.
The decision to activate graded alerts for diarrheal disease was complicated by a paucity of peer-reviewed literature regarding non-cholera diarrheal disease or typhoid morbidity and mortality statistics following seasonal flooding in Haiti. This was further complicated by the uncertainty presented by the unprecedented numbers of displaced people in overcrowded camps. The team opted for a conservative approach, erring on the side of caution given the perceived high risk of the victims. We acknowledged that other factors might mitigate the activity of diarrheal disease and typhoid activity, such as effective provision of safe drinking water by aid organizations and herd immunity. Transient increases in diarrheal disease activity were noted in the initial weeks following the earthquake and throughout the period, suspicious for possible sporadic outbreak activity of both watery diarrhea and typhoid.
Respiratory disease in general was not considered a disruptor, however the prospect of pandemic H1N1 transmission in the post-earthquake setting was a concern. An uncertainty encountered was the robust impression of resident healthcare providers and public health officials that a wave of pandemic H1N1 had not been observed in Haiti, which was a point of confusion given many other countries in the region had, including Haiti’s neighbor, the Dominican Republic. Poor clinical and social outcomes had been observed by the team during the emergence of pandemic H1N1 in 2009 among populations who were geographically isolated, had poor access to healthcare, or were impoverished. Pregnancy and young age represented special populations of concern. In 2009, the team had reported outbreaks of pandemic H1N1 in Taiwan refugee shelters after typhoon-induced flooding. One report of an American physician who landed in Haiti and then was bed ridden with an influenza-like illness raised the question of if pandemic H1N1 had not transmitted yet in Haiti, could an outbreak be triggered through introduction by aid workers? This question was raised again as Cuba reported increased pandemic H1N1 transmission on 24 April (Day 102). As with diarrheal disease and typhoid, the concern was the massive displacement of impoverished populations juxtapositioned with poor sanitation and limited access to specialized healthcare. While considered an unlikely scenario, prior experience by the team with the 2009 pandemic and seasonal influenza in tropical areas prompted a conservative and cautious posture in alerting the responder community, as shown in Figure 2.
Reporting of increased respiratory disease activity on 5 February (Day 24) by HEAS partners preceded public health reporting by 17 days. Limited laboratory testing in May and June conducted by MSPP indicated no positive samples for pandemic H1N1. Although transient increases of respiratory disease were reported, with occasional reports of fatalities due to pneumonia, there was no definitive information to support or refute the presence of pandemic H1N1 in the first 150 days post-earthquake. Reporting throughout the first 150 days was suspicious for sporadic outbreak activity; however, forensic epidemiological assessments were not performed to the knowledge of the team due to limited resources.
Figure 2. Timeline of risk and tactical reporting of respiratory disease and typhoid activity in Haiti, as reported by HEAS, PAHO, and IDPSS.
From late January to early March, the team processed multiple reports of “increasing malaria activity” with no indication of pharmaceutical stock depletions specific to the treatment of malaria, clinic strain due to patient demand, or anxiety among medical responders, resident Haitians, or officials in reaction to these reports, as shown in Figure 3. This information was judged by the team to not represent activity indicative of a potential disruptive process (i.e. emerging crisis) and therefore did not warrant an advisory. In early April, reliable HEAS partners in several locations in Port-au-Prince and other sites both north and south of Port-au-Prince began to report a noticeable increase in suspect malaria caseloads being seen in their clinics. One clinic reported being strained by the case-load. Several of these clinics had been operating in Haiti for many years and routinely performed either blood smears or Rapid Diagnostic Tests (RDT) to guide treatment. Therefore, the information was deemed credible. At this time, group dynamic within the HEAS consortium changed quickly, where not only did more partners contribute similar observations, but the frequency of information exchange increased. This was a sociological observation familiar to the team, where if a signal is validated across multiple sites, multi-lateral information exchange increases. Based on this observation, the team decided to change the status of malaria to ‘potential disruptor’ (see below) and activate an advisory on 8 April (Day 86). A definitive statement from public health officials regarding increased malaria was not presented to the responder community until nearly two weeks later on 21 April (Day 99), after increased malaria activity was generally acknowledged.
Figure 3. Timeline of risk and tactical reporting of malaria activity in Haiti, as reported by HEAS, PAHO, and IDPSS.
Based on available information, malaria activity seasonally appeared in late March as predicted by the HEAS forecast issued on 8 February (Day 27), more than 6 weeks prior. This forecast had estimated seasonally increased activity beginning in March. It was apparent to the team that risk assessments and forecast capabilities would improve with more field experience.
Other non-infectious disease events were reported by HEAS partners such as suicide clusters involving Haitian healthcare workers, suspected toxic and environmental exposures, rape, kidnappings, and outbreaks of violence were captured in HEAS reporting.
CONTEXT – CATASTROPHE CHALLENGES TRADITIONAL INFRASTRUCTURE
The challenges to traditional public health surveillance and response activities given the scale of destruction were clear. Below is a short list of challenges to the public health infrastructure, as observed by the Intercept Team from the outside.
Along with the deaths of personnel, a substantial number of resident Haitian medical and public health critical infrastructure was destroyed by the earthquake, which severely impaired the limited pre-existing capacity to definitively diagnose, investigate, and accurately report infectious disease activity. As disaster response convergence occurred within the public health community, there were other critical issues warranting priority attention such as taking inventory of functional medical infrastructure, managing a stockpile of medical supplies called PROMESS, and promoting preventive measures such as vaccination. The capacity for rapid public health responsiveness to reported infectious disease events was limited as a result.
In post-earthquake Haiti, a substantial portion of the responders had never worked in Haiti before. The response community included a large number of individuals inexperienced in treating patients in Haiti and diagnosing many of the infectious diseases of concern, thereby presenting a surveillance challenge due reporting inaccuracy.
● During the first week of April, for example, the Intercept Team was aware that the majority of malaria reported was clinically suspected but not definitively confirmed due to lack of expertise, available microscopes, or access to test kits. Two of the busiest diagnostic laboratories in Port-au-Prince, at Haiti’s University and Educational Hospital (HUEH) and St. Francois de Sales, were severely rate-limited due to lack of air-conditioning. This meant all hematology and blood chemistries were performed manually because the air temperatures exceeded the operational limits of the automated analyzers. Thus, tactical infectious disease event reporting consisted of responders reporting clinical suspicion but without a means to provide laboratory confirmation. In the case of malaria, patients would be treated based on clinical impression, positive blood smear, or positive Rapid Diagnostic Test (RDT).
Because of the heterogeneity in responder experience and access to diagnostics, reliability of the information was also a constant challenge for the Intercept Team. For example, multiple statements to the international media from otherwise credible non-profit organizations claimed either risk or presence of cholera were at odds with historical baseline epidemiological assessments for diseases such as cholera, which had not been reported in Haiti for many years.
● In one report received by the HEAS, an American nurse claimed there was an outbreak of cholera in Montroius, which in turn the team communicated the event given the potential severe consequence of actually identifying true cholera in the post-earthquake environment - even though the report was considered of lowest credibility. There were also reports from highly credible physicians from such institutions such as Rush and Stanford that claimed identification of measles and mumps among the refugees. The team was unable to verify these reports, and the Haiti National Laboratory’s diagnostic testing did not identify positive cases. However, the team recognized a Kreyol and French language ambiguity where the same word was used to describe both chickenpox and measles, which may explain why all of the investigated claims of measles were found to be chickenpox.
In regards to participation in syndromic surveillance, the team observed a high degree of variance in regards to participant experience and understanding of completing epidemiological surveillance forms. This, in addition to challenges involving laboratory diagnostic capacity, statistically ambiguous baselines, and uncertainty in regards to post-disaster conditions influencing disease incidence later triggered questions of credibility regarding the data produced by public health officials. The team was challenged to accurately report “outbreaks” or “epidemics” without statistically valid information, which was difficult to acquire in the resource constrained post-earthquake situation. This necessitated a different descriptive mechanism for reported infectious disease events, the Infectious Disease Impact Scale (IDIS).
DEVELOPMENT OF THE INFECTIOUS DISEASE IMPACT SCALE (IDIS)
The operational sensitivities and operations tempo of ‘official’ vs. HEAS disease surveillance were different.
At the time of the HEAS activation, there was no functioning public health infrastructure in Haiti, which presented a challenge to defining the social construct for organized decision making.
A multitude of individuals and organizations were deployed in Haiti to treat victims both in the immediate disaster zone and in sites throughout the country that received refugees. The majority of responders operated without access to seasonal endemic disease baseline information, tactical situational awareness, or anticipatory information.
The Intercept Team sought to help fill a void for guidance on dynamic seasonal risk of disease activity in the displaced Haitian population, with associated implications for managing an ad hoc medical response infrastructure in an environment of high uncertainty in regards to expected patient demand. Some members of the ad hoc medical infrastructure complained of limited access to situational awareness, which left medical responders exposed to sudden, unexpected shifts in patient demand. The HEAS sought to address this need.
The Intercept Team began to characterize selected infectious diseases as “disruptors”: those capable of generating a crisis or disaster in their own right that may stress or even induce collapse of the nascent ad hoc medical system.
Brief Background: Defining Crises Within Disasters
The Intercept Team chose the following definition of ‘the manifestation of a crisis’ to inform this work:
… where decisions are made prior to non-routine public health organizational involvement. Although it may be a crisis for a physician, his patient, and their family, a true infectious disease crisis, as defined from a sociological perspective, begins with time-sensitive, non-routine public health organizational decision making where such decisions may affect a local community’s activities of daily living. It is more common such decision-making falls within the organizational roles and responsibility of a public health institution than a public or private hospital or individual healthcare provider. This becomes a community level decision-making activity in countries where there is no public health capacity.
Taken to the point beyond a crisis and manifestation of a disaster:
… features of an infectious disease event that would be most likely to imply a designation of “disaster” would include unpredictable, unexpected, abrupt appearance, rapid spread, pervasive involvement of the community, lethal, and uncontrollable. An infectious disease crisis making a transition to being considered a “disaster” may be observed when crisis mode decision making by a public health institution fails to result in situational control either from an informational or response consideration. This failure of control may then result in substantial social disruption that is associated with features of community disintegration.
It is was the Intercept Team’s experience that a manifestation of community disintegration would include the so-called “panic” evacuation, which is a behavior often observed, for example, in outbreaks of Ebola hemorrhagic fever in Africa, dengue hemorrhagic fever in South America, and measles in undeveloped areas of the world. It is the Intercept Team’s experience that this is a transient phenomenon and may even represent a locally routine practice in some communities who believe evacuation to be the sole practical countermeasure available to them.
The Intercept Team assessed if one of the infectious disease disruptors were to transmit in an uncontrolled manner in Haiti, it was more likely to see crisis rather than disaster conditions result. This represented an analytic challenge of recognizing an infectious disease crisis within a natural disaster scenario, which required the necessary activities of detection, assessment, and recognition of emerging risk, as well as issuance of warning to the responder community by the HEAS. The Intercept Team’s working assumptions were the following:
- A novel community baseline associated with new activities of daily living would emerge in post-earthquake Haiti. This included development of a transplant community of responders who themselves would adopt new routine activities of daily living.
- Infectious disease crises would be reported in rumor networks.
- Should an infectious disease crisis emerge, there would be potential for disruption of the new baseline of community activities of daily living which would also be discoverable through rumor networks.
Regarding the potential to actually see the emergence of an infectious disease crisis, the Intercept Team adopted a conservative operational position that embraced a high degree of sensitivity. The reason for this philosophy was while it was acknowledged the peer-reviewed literature and humanitarian community experience suggested large-scale epidemics were not typically triggered by earthquakes, the formation of large Internally Displaced Person (IDP) camps in the context of the coming minor and major rainy seasons did represent a significant public health risk due to limited access to safe drinking water and adequate sanitation. The deployed humanitarian community robustly echoed this concern. The IDP camps that formed in Haiti also facing the onset of the minor and major rainy seasons, which were expected to further complicate their access to safe drinking water and adequate sanitation. The risks posed by seasonal rains were also conveyed by many of the deployed humanitarian responders.
Understanding and Anticipating Risk of an Infectious Disease Crisis
Information about baseline endemic seasonality of infectious disease in Haiti was limited, as was information about extraneous factors that may contribute to unexpected infectious disease activity. There was much concern among responders about the potential for infectious disease crises emerging in the refugee camps and in the context of the rainy season. There were few peer-reviewed publications demonstrating substantial evidence that the conditions seen in Haiti would or would not result in infectious disease crises. Therefore, two key issues emerged for the HEAS to address: defining risk in the context of a disaster the size of that seen in Haiti, and placing risk in context with the indigenous Haitian environment.
The Intercept Team reviewed the available literature, which had predominantly focused on humanitarian experiences in Africa and Asia, and then conducted a review of available archival media information dating back to the 1950s to ascertain whether infectious disease activity was observed to increase following previous disasters in Haiti. Flooding was the predominant disaster mentioned, and the evidence was suggestive of increases in diarrheal and vector-borne disease observed in the post-flood period.
An attempt to construct a matrix of risk was then undertaken. Several factors were considered, including but not limited to the following:
1. Infectious disease hazards, for example:
- What infectious agents were endemic or could be introduced to Haiti via inbound responders
- Which of the agents were zoonotic, vaccine-preventable, or treatable with antibiotics
- Which of the agents were clinically familiar to responders, especially American responders
- Which of the agents were associated with high morbidity, mortality, or highly efficient transmission
- What was the differential risk for indigenous Haitians (especially IDPs) versus non-indigenous responders
2. Responder community characteristics, for example:
- Number of available clinically-competent responders in the disaster response zone
- Intrinsic and extrinsic access to auxiliary medical personnel and supplies such as antibiotics, intravenous fluids, ventilators, and other key materiel
- Access to and intrinsic dependence on specialized care such a neonatal, pediatric, and adult critical care services, ventilators, and emergency medical transport
- Degree of experience specific to disease management in Haiti of both the responders and organization they were working within in
- Access to timely situational awareness in regards to infectious disease activity within the disaster response zone
- Ability to effectively bridge warning information to proactive preparedness or response
3. Indigenous patient population characteristics, for example:
- Social resilience to disease and expectation for access to care
- Access to healthcare
- Herd immunity levels and vaccination status
- Level of education
- Maternal-child feeding patterns as a proxy for risk of exposure to diarrheal agents
In summary, an examination was conducted of the intersection between the infectious hazard and the new indigenous baseline vulnerabilities represented by the post-disaster situation.
The model of the U.S. National Weather Service was utilized to express risk, where Warnings, Watches, and Advisories were defined as:
- Warning: Indication of increasing incidence of an infectious disruptor agent, with alignment of optimized conditions to support transmission
- Watch: Indication of increasing incidence of an infectious disruptor agent, without alignment of optimized conditions to support transmission
- Advisory: Presence of an infectious disruptor agent reported
The term “infectious disruptor agent” was then selected to refer to a pathogen or infectious disease capable of generating crisis or disaster level impact. These definitions served as guidelines, as declaration of a given alert level was at the judgment of the Intercept Team.
The Intercept Team then chose issue graded advisories to the response community as a means of influencing sensitivity to risk. This change in sensitivity provoked dialog about direct ground observations as it related to validity of the advisory in question, as well as key discussions about preparedness. Discussions of preparedness led in several instances to reviews of medical supply chains and ready access to pharmaceuticals, intravenous kits and solution, and oral rehydration salts. For example, as risk levels for pediatric diarrheal disease increased, advisories were upgraded, and some organizations responded by shoring up stocks of intravenous kits and oral rehydration salts.
The Infectious Disease Impact Scale
On 14 April (Day 92), the Infectious Disease Impact Scale was introduced by the Intercept Team. The IDIS was intended to serve as a qualitative assessment and guide to the responder community for expected and reported impact of infectious disease events. The threshold for a true disruptive infectious agent was the interface between an IDIS Category 3 and 4 event, respectively. No infectious disease event was assessed to have reached this level of impact in the first 150 days post-earthquake.
Table 4 provides a summary of infectious diseases expected, reported, and assessed by IDIS level. Substantial concern, attention, and resources was directed to the provision of safe drinking water and adequate sanitation, which likely prevented gastrointestinal, diarrheal disease, and typhoid from becoming a truly disruptive event (IDIS Category 4). Malaria was not assessed to be a likely disruptor; however, this assessment was briefly challenged as explained earlier in this report. Despite increased malaria activity reported in multiple locations throughout Haiti in April, conditions did not reach an IDIS Category 4. Dengue was not assessed to be a likely disruptor, and the team did not observe indicators of time-sensitive decision making in direct response to reports of dengue activity.
Table 4. IDIS levels of reported infectious disease events in Haiti.
The operational model employed by the HEAS focused on the medical responder community. The Intercept Team sought to support their need for anticipatory information and tactical situational awareness. The team recognized that routine operational activities would be established once each ad hoc medical clinic or hospital was functional for a period of time. The HEAS was concerned with preparedness and mitigation against sudden challenges to the functional but fragile medical system.
As a result, the operational default position was to report unless there was overwhelming evidence to suggest the information was truly not credible given the consequences of failing to recognize an emerging infectious disease risk. This information was then presented to HEAS partners for open debate, which occasionally supported or refuted the credibility of the reported event.
In addition, in regards to explicit reporting of named diseases and syndromes, the team also considered indirect, non-technical indicators to be of equal importance when monitoring the impact zone. The focus was on preserving response capacity and promoting the adaptive fitness of the ad hoc medical infrastructure as opposed to building a database to inform public health policy and longer term reconstruction objectives. The number of observers in Haiti, and accessed by HEAS, that were capable of reporting issues such as IDP migrations triggered by reports of an epidemic or a medical clinic experiencing strain due to a sudden increase of disease activity were far greater than full validation and laboratory confirmation of a given named infectious disease. Confirmation of such non-technical information was also less challenging than investigation and laboratory confirmation of a given infectious disease. Immediate tactical advisories would be disseminated to the response community based on these non-technical indicators, which may or may not trigger a public health investigation.
The Intercept Team also sought to take advantage of informal associations. Literature and experience suggest that individual behavior is conditioned by pre-existing social roles and personal attributes such as emotional composition and preferences. Individuals that are part of pre-existing and defined formal groups have been observed to assume defined, predictable behaviors as part of a larger predictable group behavior. This is dependent heavily on the group’s experience and pre-defined responsibilities in crisis scenarios. Informal, uncoordinated response is a prominent feature in the minutes, days, and even weeks before official engagement has begun and hierarchical command and control systems are in-place. This was the very phenomenon exploited by the Intercept Team to support the HEAS.
We note that the Intercept Team sought to develop a network of trust amongst the unique group of stakeholders that arrived on the ground in Haiti. An axiom “trust doesn’t surge” comes sharply into play in those communities experiencing a crisis where such integrative and inclusive social networks are not present.
A point of debate within the team emerged around whether the Kreyol-speaking Haitian public should be included in the distribution of both the CIRs and operational output of the HEAS. On the one hand, distribution of advisories would have encouraged greater participation by the public in preventive measures. Conversely, such action may have superseded the Haitian Ministry of Health’s authority, as they struggled to recover from the disaster. The latter was an important consideration due to the desire to promote civil stability and show support for Haitian government authority. The final outcome was to strike a balance where information was shared publicly in English on the Internet instead of in Kreyol via SMS and radio announcement in Haiti.
The 2010 Haiti earthquake caused catastrophic losses.
In the wake of the devastation, the Intercept Team had a unique opportunity to contribute to disaster response and advance new approaches to operational biosurveillance. This contribution occurred in the context of catastrophic destruction of Haiti’s already limited medical and public health sector.
During the 2009 Mexico influenza outbreak, the Intercept Team had shown itself capable of delivering significant (over two weeks) early warning of a major infectious disease event without the support of government infrastructure. In response to the 2010 Haiti earthquake, the Intercept Team had an opportunity to do so again.
The Intercept Team chose a different approach to infectious disease outbreak surveillance in the wake of a disaster as large as the 2010 Haiti earthquake. Primary objectives were to address the unmet needs of medical responders in crisis, with focus on identifying infectious disease events that could themselves pose a disruptive threat to the work being undertaken by the Haiti Ministry of Health, PAHO, and the CDC to maintain a functioning official infrastructure.
The Intercept Team’s established track record and operations apart from official channels created opportunities to access and engage new sources of information. While the reliability and credibility of all sources incorporated into the HEAS was not confirmed during the rapid response efforts following the earthquake, the Intercept Team delivered accurate advanced forecasts, warnings, and alerts for a multitude of key infectious disease events that were often days, weeks, and months ahead of any alternative.
The Intercept Team encourages the consideration of how these approaches may be ‘formalized’ to support the Haitian people in the years ahead, and incorporated into the established orthodoxy of catastrophe response.
It is our assessment that the model for which the Haiti Epidemic Advisory System (HEAS) is based, along with the accompanying Infectious Disease Impact Scale (IDIS) would have broad applications for supporting disaster response and health threat mitigation around the world.
The Haiti Earthquake Intercept Team would like to thank the following individuals and organizations for their tremendous assistance and collaboration during the post-disaster response:
Mr. Sean Penn, Lady Alison Thompson, Ms. Aleda Frishman, Mr. Barry Frishman, Mr. Stephen Duvall and the Jenkins-Penn Haiti Relief Organization (J/P HRO); Dr. Michael McDonald and Global Health Initiatives (GHI); the Pan American Health Organization; the U.S. Centers for Disease Control and Prevention (CDC); the Haitian Ministry of Health (MSPP); the United States Southern Command (USSOUTHCOM); Dr. Eric Rasmussen, Mr. Eduardo Jezierski, Mr. Nico Di Tada, and Mr. Luke Beckman of InSTEDD; Patrick Meier and Rosalind Sewell of Ushahidi, the communities of Crisis Mappers and the All Partners Access Network (APAN), the anonymous private donors to Praecipio International and the anonymous Haiti Epidemic Advisory System partner community.
 The Global Disaster Alert and Coordination System (GDACS), http://www.gdacs.org/reports.asp?eventType=EQ&ID=72809&system=asgard&alertlevel=Red&glide_no=EQ-2010-000009-HTI&location=HTI&country=Haiti
 Laing, A. Haiti earthquake: presidential palace among buildings demolished. Telegraph.co.uk. http://www.telegraph.co.uk/news/worldnews/centralamericaandthecaribbean/haiti/6988391/Haiti-earthquake-presidential-palace-among-buildings-demolished.html
 United States Geological Survey (USGS). http://earthquake.usgs.gov/earthquakes/recenteqsww/Quakes/us2010rja6.php#summary
 United Nations Office for the Coordination of Humanitarian Affairs (OCHA). Haiti Earthquake Situation Report #19, 8 February 2010. http://www.reliefweb.int/rw/RWFiles2010.nsf/FilesByRWDocUnidFilename/MYAI-82H8LM-full_report.pdf/$File/full_report.pdf
 Haiti Earthquake, Post-Disaster Needs Assessment (PDNA), Assessment of damage, losses, general and sectoral needs, 2010, http://siteresources.worldbank.org/INTLAC/Resources/PDNA_Haiti-2010_Working_Document_EN.pdf
 Ushahidi. http://haiti.ushahidi.com/
 World Health Organization. International Health Regulations. http://www.who.int/ihr/en/
 World Health Organization. Global Outbreak and Alert Response Network (GOARN). http://www.who.int/csr/outbreaknetwork/en/
 Statement by James M. Wilson V, MD Research Faculty of Department of Pediatrics and Director of Division of Integrated Biodefense, Imaging Science and Information Systems (ISIS) Center at Georgetown University Before the Senate Homeland Security & Government Affairs Subcommittee on Oversight of Government Management, the Federal Workforce, and the District of Columbia October 4, 2007, found at http://hsgac.senate.gov/public/index.cfm?FuseAction=Files.View&FileStore_id=3b5da0a2-5b3a-463e-aed5-4b2e7ec64cbf
 Pan American Health Organization (PAHO). Promess Warehouse: Matching the world's medical donations to Haiti's needs. January 31, 2010. http://www.reliefweb.int/rw/rwb.nsf/db900sid/MYAI-82939N?OpenDocument&RSS20&RSS20=FS