The Haiti Epidemic Advisory System (HEAS)

Ground Truth: 445 cases and 59 fatalities in Pestel

This is the dispatch we received last night from GlobalDIRT: 

 

Ok, we have all the cursory but CONFIRMED epidemiological info of confirmed cases from the cholera camps in our area (confirmed by our ground crew and the [Haitian physician on-site]) :

First cases started trickling in in October 2011 in the Pestel/Desrivaux/Tozia area:

• October 2011: 115 cases
• November 2011: 118 cases
• December 2011: 116 cases
• January 1st-13th: 86 cases

 

• There have been 59 deaths between December 2011-January 2012

• Total: 445 cases since October, 59 deaths (13% case fatality rate)

• Approximately 1/4 of cases are/were children (~111 cases)

• Villages affected: 
  • Hiya : 20 cases
  • Beaumont 31 cases
  • Nanfont: 14 cases
  • Cassanette: 15 cases
  • Sebarier: 14 cases
  • Amiel: 15 cases
  • Feranse: 14 cases
  • Larier: 14 cases
  • Chapette: 2 cases
  • Labondance: 3 cases
  • Ganier: 3 cases
  • Aucentre: 13 cases
  • Mouline: 17 cases
  • Lacadony: 6 cases
  • Nan pomme: 8 cases
  • Belecolle: 27 cases

We are still working on compiling lists of the water sources. The consensus is that right this moment the case volume is low, however, it is going to get A LOT worse over the next couple months.... Especially with the rainy season approaching. We need to be ready to respond to Zone 4 (about 2 dozen villages scattered within  a couple hours of each other the next few weeks/months) Only "hospital" around is on Pestil and poorly equipped... More info to come

Best,

Global DIRT

 

The current operational posture of MSPP / CDC / PAHO is they do not leave Port au Prince to investigate and respond unless there is laboratory confirmation.  We are fully aware it took them 72 hours to correct their own assertion there was no crisis in this region.

It is fortunate there are still NGOs with an ability to execute efficiently on the ground to help these villages.  Here we observe the gross operational disconnect between public health and tactical infectious disease forecasting, early warning, and tight coupling to response.  When one examines the amount of funding gone to public health's activities, that have been ongoing in Haiti without operational peer review, versus the reality of low cost, highly efficient "smart swarms" such as we are observing now converging on Pestel, it makes one wonder about where the United States' priorities are when expending money during a Recession.  One thing is clear: USAID, the American Red Cross, CDC, and the State Department failed to provide funding support to US organizations actually doing the mission.  Accountability, yet again, is quite a bitter pill to swallow.

At this point, HEAS partners are in the process of setting up four CTCs to handle the case load.

 

 

Eruption of New Cholera Activity in Pestel

On January 5th, the HEAS lit up abruptly with reports of 200 cases of cholera and 17 fatalities (CFR-apparent 8.5%) in Pestel, located in an extremely remote area of Haiti. Almost exactly one year after HEAS partners did an emergency helo drop of medical supplies.

The HEAS hit a new operational record of about two hours between initial warning and confirmation of an emergency response package made available by several members of the HEAS, including Direct Relief International.  Back and forth between officials at MSPP, PAHO, and CDC indicated they were unaware of the situation and vehemently denied the reports for 72 hours... but later came back to us and said it was not only 200 cases, but 300.

The ground team just arrived, confirming 300 cases in 2 of the 4 surrounding villages (survey of the other two villages pending).  They discovered 55 fatalities (CFR-apparent of 18%), not the previously reported 17.  The situation represents an IDIS Category 5. 

Additional HEAS partners are rallying assistance now...

 

HEAS SitRep 12.8.10

• On December 3rd, MSPP reported 91,770 hospital visits for cholera, 43,243 of which were admitted, and 2,071 fatalities as of that day. These retrospective statistics continue to represent gross underestimates of the true caseload and fatalities seen in Haiti thus far.
  • The greatest discrepancies in reporting are observed in the difficult to reach mountainous, rural areas. There are currently no credible statistics to account for these communities, which are thought to represent nearly 2/3 of Haiti by land area.
  • The election period may have disrupted official reporting.
  • Precipitation is forecasted for this weekend, particularly for the southern peninsula, which may result in enhancement of transmission. We hypothesize the dry period experienced by much of the country post-Tomas has inhibited transmission.

• The HEAS is primarily concerned with under-represented areas experiencing “first contact” sudden patient surge and high daily clinical mortality. We do not use the public health case fatality rate, which is defined as the number of individuals infected that die divided by the number of individuals infected. Instead, the daily clinical mortality is a reflection of psychosocial burden on the affected community, where it is calculated as the number of fatalities for a given disease observed by a given clinic divided by the number of admissions for that disease to the clinic. If five patients are admitted and those five patients all die, then the daily clinical mortality is 100%. This is representative of high and acute psychosocial impact to both the clinic staff and surrounding community as rumor of the event spreads. It is also representative of clinic capacity in terms of ability to exploit pre-event warning, pre-event education and training, materiel preparedness, and peri-event personnel and materiel capacity. It is an important marker for emergency response deployment. Below is an example at the departmental level, that of “first contact” daily clinical mortality:

 

• When considered by department, the most recent Haitian Ministry of Health (MSPP) statistics published as of December 4th highlight key areas experiencing “first contact” mortality.

• The below figure represents the departments (Grand Anse, Nippes, Sud, Sud-Est, and Nord-Est) that are of greatest concern to the HEAS. These areas were forecasted to be the next “battlefront” by the HEAS; this forecast is now validated. These are sites experiencing either “first contact” patient surge high clinical mortality. WASH and Health Cluster resources are largely concentrated and in the process of mobilization for Jacmel and Les Cayes, however the situation remains highly dynamic. The rural mountainous areas of the southern peninsula are reporting cholera, however evaluation of these areas has been limited. The HEAS is most concerned about the status of Sud-Est due to its geographic isolation from the remainder of the country. As mentioned above, rain is forecasted for the coming weekend, with higher precipitation expected for the southern peninsula. This will likely exacerbate epidemic control efforts due to enhancement of transmission and inhibition of response mobilization.

• The next category of departments represent sites where there has been active and persistent transmission, hospitalizations, and moderate, persistent clinical mortality. Nord-Est is the only department in this category at-present:

• The final category represents sites where there has been active and persistent transmission, hospitalizations, and low but persistent clinical mortality. This category effectively represents the remainder of the country:

• The below figures highlight the results of 37 directly participating HEAS partners, two large social networks including the HEAS, 802 point to point email communications (with a daily peak of 284 emails), and a time delta between HEAS notification of Request For Emergency Assistance and materiel & personnel arrival on-site of 3 and 5 days, respectively.  In the social network plex figures, each colored diamond represents a separate HEAS partner, and the larger inverted triangles represent two large social networks, one of which is the HEAS itself.

The Borgne Emergency Ad Hoc Cholera Treatment Facility, Nov 22nd. 

• The reality of HEAS alert and response operations resides in a now well-established pattern of reporting. Communities along paved roads, which includes coastal areas, valleys, and mountain passes are the first to report cases followed later by observation that mountain rural communities above are the source of more critically ill cases. These cases are more ill upon arrival due to the multiple-hour trek by foot required to reach the nearest CTC / CTU. As reported before by the HEAS, many of these communities were caught completely by surprise having not even heard there was cholera in the country. These mountain rural communities represent approximately 2/3 of Haiti by land area and are extremely difficult to access, often only allowing access by foot or helo drop. Many do not have helo landing zones. The current national response effort does not contemplate accessing these difficult to reach areas. The HEAS continues to receive many requests for emergency assistance from these sites, however it is likely we are observing gross under-reporting. The below figure highlights the regions of concern from this perspective and reflects a substantially more challenging view of where response assets are versus the broader evolving need. Red areas indicate rural sites of greatest concern to the HEAS, followed by orange and yellow areas, respectively.

• Conservative estimates suggest more than 360,000 cases of cholera in Haiti to-date, the majority of which were subclinical. This is based on the 1 clinically apparent : 3 subclinical case ratios quoted by PAHO.

• In some areas of Haiti, we have had confirmation that in-patient statistics are under-reported by as much as 400%. In many areas of Haiti, we are documenting outbreaks that are not being accounted for in the official statistics. We therefore estimate the upper bound of estimated total (subclinical and clinically apparent) case counts to be one million. From a practical operations point of view, these estimates are academic, and we do not believe it likely the case counts are one million but rather closer to 500,000 total. The bottom line is the epidemic continues to spread without restraint.

• Initial conservative PAHO projections used an attack rate of 2% of a total estimated Haiti population of 10 million to estimate 200,000 total clinically apparent infections, the true community load will be closer to 800,000 if subclinical infections are counted. We will exceed these estimates well before the peak of the epidemic.

• We err on the side of over-estimating because this is a "virgin soil" epidemic and expected to aggressively spread throughout Hispaniola.

• We are now pursuing answers to the question of uptake by indigenous zooplankton and spread along oceanic currents that pass west of the Gonave Gulf, which is where the Artibonite River discharges, north and west along the northern Cuban coastline and north to the waters east of Florida.

• We assess the elections on the 28th facilitated further spread. To what degree remains uncertain.

• Multiple healthcare worker infections have been reported on the ground in Haiti with one report of a worker returning to the United States infected.

• There will be more cases in the United States; we believe it likely more cases are inside the US unreported.  Implications for the United States are non-significant.

• It is clear both the Health and WASH Cluster activities are unable to keep up with the speed and breadth of cholera transmission.

• Civil unrest triggered by announcement of the election results last night forced the closure of PAHO operations today.

• The entire HEAS surveillance / response grid is now overwhelmed and lacks the agility enjoyed in the beginning of the disaster. To meet this crisis, we are attempting to ascertain additional levels of funding and expertise.

HEAS Operations Update

We currently have several “onion layers” of operations in-place now to facilitate the movement of alerts to response.  These layers follow the steps we have to go through before getting emergency aid to a site in need:

Initial report.  This is the first report of a site either reporting cases of cholera for the first time, sudden surge, or “first contact” mortality.  The degree of alarm to us is related to this sequence.  That said, we prefer to engage as quickly as possible to avoid “first contact” mortality, however this is often difficult to a number of factors such as degree of site isolation (i.e. a community’s access to communications), pre-contact awareness there is a cholera disaster in-country (we have had communities who had no awareness of the epidemic!), and social thresholds of tolerance for adverse clinical outcomes.  The general rule of thumb is people typically alert for mortality, not non-mortal cases.  We encourage reporting of at least “first surge” or the window of opportunity to avert unnecessary fatalities may be lost.

Verification of event features.  We generally see the majority of reports validate, however the actual case counts may vary tremendously between the HEAS reporter, the officials, and any on-site NGO(s).  We want to make it clear that accuracy in case counts is not important to us.  What IS important is verification of impact such as the ability of local medical responders (if there are any) to manage the demand of patients.  This is important to because we are looking for specific, prioritized items of need involving materiel or personnel (see below).  Collapse of a clinic or high daily clinic mortality is associated with a distinct psychosocial impact on the involved community that does play a role in civil stability dynamics.  In short, we seek verification of impact first, accuracy of case counts last.

There are several key caveats to keep in mind.  There MUST be a false positive rate when reporting or we are not doing our job.  While we as warning analysts are very tolerant of false positive reporting, those that manage deployment of response are NOT.  This highlights the crucial importance of verification as an operational bridge.

Now, we do monitor sources of reporting and will “dial back” reports from a source to the HEAS if we observe a high rate of unverifiable reporting.  We see this rarely, but it does happen.  There is also a triage of reporting, where sources unknown to us typically require more attention to verification than trusted sources.  That said, we have observed many instances where highly trusted, credible sources have either misinterpreted their information or were simply wrong.  It happens to the best of us, as we are all human.

When verifying information, we are seeking to pin down several aspects of event information.  The name of the site is often problematic for us, as there are often different names for the same site in Kreyol versus French or English.  GPS coordinates or Google map links are also very problematic, especially for the rural areas, because these sites are often not even well known to NGOs who work in the region unless they have worked there directly.  We have found this can be problematic to lock down even for indigenous Haitians.  The effect of multi-source reporting often distorts information as elements are passed from person to person.  In general, the more steps you are away from direct ground observation, the more probable distortion has occurred above and beyond the basic limits of human observation.  And of course, two people standing at the same site observing the same thing will often come away with two different interpretations based on life experiences, for instance.  If multiple NGOs are managing a given community’s needs, we often see conflicting reporting depending on the relationship between those NGOs. 

Bottom line, verification can be a difficult and time-consuming process.

The firehose of information now streaming through the HEAS community is at once an incredible accomplishment for the community but also a challenge from an information management perspective.  To address this, we have broken off a much smaller group of HEAS partners that manage event reports, verification, triage of those reports to the HEAS Community Watchboard as potential sites for response consideration, and ground logistics and response.  Events on the HEAS Community Watchboard are triaged according to a simple scale of need:

Level 1= Request For Emergency Assistance triggered; requires IMMEDIATE attention via deployment of emergency supplies or personnel (time delta= hours-48 hrs)

Level 2= Request For Emergency Assistance triggered; requires URGENT attention via deployment of emergency supplies or personnel (time delta= multiple days)

Level 3= No Request For Emergency Assistance triggered yet, but Request For Verification triggered

Level 4= Awareness of active cholera treatment noted but no action required at this time

Validation of emergency need is often entwined with the verification process.  Here we ask for a strict prioritization of need, which sometimes indicates how much trouble a site is in.  For instance, a site requesting supplies but no personnel appears less urgent to us than a site requesting body bags over other supplies and “as many people as we can spare”.  There is an implied anxiety level in the requesters that we also monitor.

Some use a case count cut-off of 50 cases or above to prompt mobile response.  We have struggled with this guideline because we are looking at the psychosocial dynamic from the viewpoint of expected local capacity: i.e. the local medical clinic.  One community, as mentioned above, may have a high psychosocial tolerance for morbidity but become tremendously anxious upon observation of 5 fatalities.  Another community may be very resistant to high fatalities and wait to report until the last possible moment.  This is of course due to a whole range of factors.  Bottom line, it is a difficult and terrible triage to decide which site, all things considered, should be prioritized for response.

We have observed now several instances where the HEAS core team did not need to directly push for response, but the HEAS community self-triggered response on its own, offline from the main group communications.  This is a tremendous positive development and indicates the "living entity" that is the HEAS.  Response coordination, however, has been a serious challenge for us in that we have seen a general trend of no response report from organizations planning to mobilize to a site.  The problem arises when multiple organizations arrive with duplicative assets due to lack of communication. This becomes a tremendous problem because when organizations perceive possible duplication of effort, hesitation results.  Hesitation may cost lives. This is the purpose of the HEAS Google group and Community Watchboard: to assist in coordination.  So, the HEAS is experiencing "growing pains" that we are addressing as they arise.

Grid level response has been tricky at times due to grid overload.  That said, we HAVE seen an increased awareness on the ground that 2/3 of Haiti is likely under-supported due to extreme terrain and lack of access or communications.  Thus, there has been a concerted effort on our part to emphasize the consideration of helo and rapid mobile medical team support.  This has been tremendously challenging, as many of the mountain sites do not have a suitable landing zone for helo support, resulting in a many-hour hike across terrain that makes hauling mass quantities of IV fluids impossible.  As we have seen so many times throughout the world, the more isolated a community is, the more likely one sees adverse outcomes in the context of an infectious disease crisis or disaster.

We conclude with an important observation: that the HEAS has now evolved into a self-regulating, distributive, complex, adaptive, net-centric warning and response community.  This is a major, positive development as we continue to struggle to support the people of Haiti.

HEAS SitRep 12.1.10

On December 1st, the Haiti Health Cluster reported 77,208 hospital visits for cholera, 34,248 of which were admitted, and 1,751 fatalities as of November 27th. These continue to represent gross underestimates of the true caseload and fatalities seen in Haiti thus far. The below represents areas of specific concern and active monitoring by the HEAS:

Red- where we continuously receive emergency requests for assistance due to "first contact" high mortality and have facilitated several mobile emergency responses.  Currently, we have report that MSF’s facilities are overrun with patients with no auxiliary staff.  We are expecting overflow surge into surrounding clinics that are already struggling.   Both the Cap Haitien and NW regions are seriously strained, with the NE region becoming increasingly problematic to manage.  Little situational awareness is provided for the remote mountainous regions.  We suspect very high mortality for those villages at >3 hrs walking distance from the nearest medical facility able to support cholera treatment.

Orange- where we occasionally receive notification of "first contact" high mortality but confirmation is difficult; we suspect there is far greater mortality in these mountains but it goes unreported because of very low NGO presence and difficulty of access.  As mentioned above, we suspect very high mortality for those villages at >3 hrs walking distance from the nearest medical facility able to support cholera treatment.

Yellow- where we have confirmed cholera activity at multiple sites throughout the southern peninsula, however the majority of international response is focused on PaP, Artibonite, and recently the north… the south is left relatively untouched in terms of response.  As with the orange areas, we suspect far more activity that is what reported because of very low NGO presence.  We also note very large time lags in official reporting from this area.  The southern peninsula is the next ‘battlefront’, where we expect the 'war' to go very badly given the low availability of response assets.

Case fatality rates such as the most recent Health Cluster report of 3.6% (in-hospital) is representative of gross national level aggregation of information available to officials, which represents a substantial bias towards CTC/CTUs staffed by experienced teams such as MSF.  What is not reflected is the continually documented "first contact" pattern of daily clinical mortality seen by rural communities and urban environments such as Gonaives several weeks ago reported by officials do not reflect the true impact of cholera at the community level.  The daily mortality we have documented on multiple occasions may range from 10 to 100%.  We often see sudden overwhelming of local capacity to the point of back-loading corpses for burial, having run out of body bags.

Conservative estimates suggest more than 300,000 cases of cholera in Haiti to-date, the majority of which were subclinical.  This is based on the 1 clinically apparent: 3 subclinical case ratios quoted by PAHO.

In some areas of Haiti, we have confirmation that in-patient statistics are under-reported by as much as 400%. In many areas of Haiti, we are documenting outbreaks that are not being accounted for in the official statistics.  We therefore estimate the upper bound of estimated total (subclinical and clinically apparent) case counts to be one million.  From a practical operations point of view, these estimates are academic.  The bottom line is the epidemic continues to spread without restraint.

Initial conservative PAHO projections used an attack rate of 2% of a total estimated Haiti population of 10 million to estimate 200,000 total clinically apparent infections, the true community load will be closer to 800,000 if subclinical infections are counted.   We will exceed these estimates well before the peak of the epidemic.

We err on the side of over-estimating because this is a "virgin soil" epidemic and expected to aggressively spread throughout Hispaniola.

We are now pursuing answers to the question of uptake by indigenous zooplankton and spread along oceanic currents that pass west of the Gonave Gulf, which is where the Artibonite River discharges, north and west along the northern Cuban coastline and north to the waters east of Florida.  

We assess the elections on the 28th facilitated further spread.  To what degree remains uncertain.

Multiple healthcare worker infections have been reported on the ground in Haiti with one report of a worker returning to the United States infected.

There will be more cases in the United States; we believe it likely more cases are inside the US unreported.  Implications for the United States are non-significant.

The below represents the USAID view of the situation, which highlights multiple areas of mismatch between emergency need and CTC/CTU placement.  The HEAS is working with PAHO to close these gaps.

The below represents the WASH view of the situation, which also does not completely agree with our situational awareness (see below notes):

It is clear both the Health and WASH Cluster activities are unable to keep up with the speed and breadth of cholera transmission.

Donor fatigue continues to impair response efforts, as does periodic civil unrest.  There is a significant lack of volunteerism from the international medical community, where levels of volunteerism observed post-quake are not observed currently.  The response grid is in jeopardy of collapse without additional assistance.

The entire HEAS surveillance / response grid is now overwhelmed and lacks the agility enjoyed in the beginning of the disaster.  To meet this crisis, we are attempting to ascertain additional levels of funding and expertise.

HEAS SitRep 11.9.10

Summary

Current official stats are more than 9,500 cases and 583 fatalities.  In some areas of Haiti, we have confirmation that in-patient statistics are under-reported by as much as 400%.

There is no question of under-reporting.  If clinically apparent case counts are assumed to be 1/4 the true community load, then nearly 36k cases are shedding pathogen into the environment.  We believe the true statistic to be closer to more than 50k based on the degree of under-reporting. This is an uncontrolled, uncontained epidemic of cholera that has exceeded public health capacity to investigate and assess every site reported and every sample received.  Substantial political interference and resultant lack of proper coordination at the Health Cluster and WASH levels continues to severely compromise timely reporting of information critical to save lives.  This has been a problem since the earthquake in January 2010.  Haitian officials have declared the cholera epidemic to be an issue of national security, which may further contribute to political interference in information sharing.

Evidence now suggests the epidemic has crossed the border into the Dominican Republic, which was expected.

Key Updates

Cholera has now seeded the Haitian environment in more sites than can be properly assessed.  Evidence of community transmission is present in multiple sites such as the northwest, greater Port au Prince area, and strongly suspected in the southern peninsula.  Transmission modes include waterborne, food contamination, and human-to-human contact.

Transmission in the original epicenter of the epidemic along the Artibonite River has decreased substantially and shifted to surrounding rural areas, extending to the northern coast.  As feared, cholera transmission in communities along the coastal highway between St Marc and Cite Soleil likely resulted in dozens of indigenous cases now identified in Cite Soleil, with dire implications for Port au Prince.  Multiple confirmed cases without travel history to Artibonite have been documented in both Cite Soleil and Port au Prince, and hundreds more suspect cases have been declared inside Port au Prince.  Significant political interference prevents full clarity of the situation inside the city.  As of this report, HUEH and other major medical centers inside Port au Prince are stable and able to handle the patient flow.

Sporadic, unconfirmed rumor suggests cholera has reached Carrefour, however this information is treated with a high degree of uncertainty.  Eventual confirmation in Carrefour is expected.  Other areas have reported cholera such as Grand and Petit Goave; confirmed or not, this would not be unexpected.  All suspect cholera cases tested in Leogane have reportedly been negative, however it would not be unexpected to see confirmation in the very near future.  Suspect cholera cases have been reported in Les Cayes, Jacmel, and several other rural communities in the southern peninsula.  We assess it is highly likely the epidemic has indeed extended to the southern peninsula.

The issue of “suspected” versus “confirmed” reporting and decision points for action have been hotly debated inside the HEAS community.  As was observed in the United States during the 2009 H1N1 influenza pandemic, the number of samples collected is disproportionate to the ability of LNSP to process them in a timely fashion and publicly declare confirmation status.  The entire nation is now sensitized to report cholera, however many have never seen cholera before.  This implies means the false positive rate is likely increased.  Diarrheal disease not due to cholera is caused by a wide variety of other endemic pathogens and is seen this time of year during the major rainy season.  All of this acknowledged, hesitation to verify or assume “cholera until proven otherwise” may mean the difference between 1 and 10 fatalities for a given community.  This is a tremendous challenge to the HEAS community.

We find that the pattern of reporting, particularly with rural villages in the north with minimal baseline medical capacity, initial IDIS ratings reach Category 5 and then are quickly downgraded to Cat 4 once medical response is mobilized and responders become habituated to clinical care for cholera patients.  Much of Artibonite Valley is now rated at IDIS 3 or lower. 

Port au Prince is currently rated at IDIS Cat 1 to 3, depending on the specific site, with strong potential to reach Cat 5 in the coming days.  Cite Soleil is likely to reach Cat 5 conditions before Port au Prince.

The closest site reporting cholera to the border with the Dominican Republic (DR) is approximately 10 miles and is not on a major roadway.  That said, 3 suspect cases have been reported with positive travel history from Haiti to DR, and no acknowledgement has been provided to-date by DR officials.  We assess it is highly likely cholera is inside DR’s borders.  The HEAS has coordinated with DR officials to provide situational awareness to enable preparedness.

The major rainy season normally peaks in October and persists through November.  We assess the pass-through of Hurricane Tomas provided environmental enhancement of cholera transmission due to flooding.  We expect to see continued effects for the next several days.

Status of the Biosurveillance Grid

The HEAS is a community of more than 600 online and offline ground medical responders, international experts in cholera ecology, operational biosurveillance analysts, meteorologists, veterinarians, sanitation specialists, and public health officials.  The membership includes international NGOs, IGOs, UN, US agencies, and private citizens with specific, relevant expertise. 

The HEAS Mid Action report, based on the initial 150 days of operations post-earthquake, displays the power of operational biosurveillance-facilitated distributive networking for early warning and rapid response.  The HEAS community has now evolved into a self-regulating body that processes tactical event warning and forecast information to prompt verification and rapid “swarm” medical response activities.  It is the largest group of its kind operating in Haiti.

Conclusion

The cholera epidemic in Haiti proceeds in an uncontrolled, uncontained fashion and will likely encompass all of Haiti within a matter of weeks.  Ecological establishment will be pervasive regardless of ongoing response efforts.

Efforts to improve access to safe drinking water and sanitation are as challenging as they were post-quake, and now additional vast areas of Haiti will require similar attention.  It is unlikely effort will be mobilized quickly enough to forestall the spread of cholera.  Public health intervention in the areas of education, hygiene, and training of medical responders will continue to be important.  However, the default operational position now is not to prioritize effort to prevent spread but to prevent or mitigate unnecessary fatalities.  The HEAS is narrowly focused on early, actionable warning coupled to “swarm” medical response.  We have observed time deltas between warning and imminent healthcare facility collapse can be as short as 24 hrs.  The HEAS has encouraged high false positive rate reporting to ensure communications with involved medical facilities and staff can be established and resources mobilized quickly enough should they be required.  This process has already assisted several overwhelmed facilities and likely prevented unnecessary loss of life.

It is notable the function of the HEAS, as facilitated by operational biosurveillance analysts, is distinct from public health operations with a different focus, scope, and operations tempo.  Although complimentary information sharing between the two communities would likely be more productive than maintaining a separation of operations, we have found public health involvement aside from consultative input to be largely irrelevant.  This is a similar observation to what was observed during the 2009 H1N1 influenza pandemic and likely reflects a well-understood lack of agility in public health function during rapidly evolving infectious disease crises and disasters.

The HEAS coordinates with other foreign nations about the situation in Haiti.  As discussed above, DR will likely be the second country in the region to experience epidemic conditions.  It is difficult to anticipate the next involved country thereafter.  Advisories were sent by the HEAS to key elements of the United States medical community on 10/21, 10/22, and 11/4 in an effort to sensitize clinical awareness for travelers or patients transferred from Haiti.  The discovery of cholera in travelers or patients from Haiti inside the United States will not be unexpected.

Continued independent "smart swarming", utilizing distributive networks for situational awareness, is strongly recommended with an expectation for a protracted period of response operations. 

Eventual regionalization of cholera in the Caribbean is a strong possibility but not a certainty if the pandemic of the early 1990s is a guide.

Current Map of Cholera-Affected Areas

We will adjust this map as more information becomes available. 

Updates: Official stats, estimates of true case loads, updated map.

While the UN OCHA maps and official MSPP reporting tends to focus on Artibonite and points north, there are other areas routinely not included on the OCHA maps that have reported cholera.  We opt on the side of "cholera until proven otherwise" or when political sensitivity is such that full disclosure and transparency allows.

Current official stats are more than 72,017 cases that have sought treatment and 1,648 fatalities.  We note large discrepancies between HEAS partner information and officially reported statistics, reflective of official under-reporting.

• Case fatality rates such as the most recent Health Cluster report of 2.3% is representative of gross national level aggregation of information available to officials, which represents a substantial bias towards CTC/CTUs staffed by experienced teams such as MSF.  What is not reflected is the continually documented "first contact" pattern of daily clinical mortality seen by rural communities and urban environments such as Gonaives several weeks ago reported by officials do not reflect the true impact of cholera at the community level.  The daily mortality we have documented on multiple occasions may range from 10 to 100%.  We often see sudden overwhelming of local capacity to the point of backloading corpses for burial, having run out of body bags.
• Conservative estimates therefore suggest more than a quarter-million cases of cholera in Haiti to-date, the majority of which were subclinical. 
• In some areas of Haiti, we have confirmation that in-patient statistics are under-reported by as much as 400%. In many areas of Haiti, we are documenting outbreaks that are not being accounted for in the official statistics.  We therefore estimate the upper bound of estimated total (subclinical and clinically apparent) case counts to be nearly one million.  From a practical operations point of view, these estimates are academic.  The bottom line is the epidemic continues to spread completely out of control.
  
• Although PAHO projections use an attack rate of 2% of a total estimated Haiti population of 10 million to estimate 200,000 total clinically apparent infections, the true community load will be closer to 800,000 if subclinical infections are counted. 
  
• We err on the side of over-estimating because this is a "virgin soil" epidemic and expected to aggressively spread throughout the country and across the border into the Dominican Republic to cause massive disruption.
• We are now pursuing answers to the question of uptake by indigenous zooplankton and spread along oceanic currents that pass west of the Gonave Gulf, which is where the Artibonite River discharges, north and west along the northern Cuban coastline and north to the waters east of Florida.   
  
• We assess the elections on the 28th facilitated further spread.  To what degree remains uncertain.
  
• Multiple healthcare worker infections have been reported on the ground in Haiti with one report of a worker returning to the United States infected.
• There will be more cases in the United States; we believe it likely more cases are inside the US unreported.  Implications for the United States are neglible.

The below represents areas of specific concern and active monitoring by the HEAS:

 

• Red- where we continuously receive emergency requests for assistance due to "first contact" high mortality and have facilitated several mobile emergency responses.
• Orange- where we occasionally receive notification of "first contact" high mortality but confirmation is difficult; we suspect there is far greater mortality in these mountains but it goes unreported because of very low NGO presence and difficulty of access.
  
• Yellow- where we have confirmed cholera activity at multiple sites throughout the southern peninsula, however the majority of international response is focused on PaP, Artibonite, and recently the north… the south is left relatively untouched in terms of response.  As with the orange areas, we suspect far more activity that is what reported because of very low NGO presence.  We also note very large time lags in official reporting from this area.  The southern peninsula is the next battle front, where we expect the 'war' to go very badly given the low availability of response assets.

 

 

View Overall Haiti Cholera Epidemic Map in a larger map

HEAS SitRep 11.1.10

Summary

As of Friday, October 29th, PAHO has reported more than 4,700 cases and more than 330 fatalities (apparent case fatality rate approximately 7%).  PAHO was unable to report updated statistics because MSPP had not released them on Monday, Nov 1st.  Statistics quoted by the media over the weekend indicate fatalities approaching 340.  If approximately 25% of cholera cases are clinically apparent, then a highly conservative estimate places the true case count at nearly 19,000.  We estimate the case counts to be closer to 40,000 based on the level of under-reporting and rural transmission occurring in areas not being serviced by responders.  Cholera epidemics often follow a waxing and waning pattern of transmission and therefore can be unpredictable.

Key Updates

Further evidence of cholera transmission in Cite Soleil has not been reported, which is a key sentinel site for the greater Port au Prince area.  Vigilance and sensitization for cholera in Cite Soleil has been elevated to the maximum extent possible.

We are aware of 12 confirmed cases of cholera reported within PAP but not publicly reported by officials.

Suspect case samples have been sent from Leogane to LNSP for processing, indicative of a high index of suspicion and a highly sensitized local surveillance grid.

Field teams operating in Artibonite have reported back today that the true case counts are far higher than is officially reported due to the following factors:

1.  Evidence of pervasive transmission throughout the Artibonite region along rivers and tributaries. Seeding of tributaries has been reported, which again strongly raises the question of ecological establishment, and a "rolling" pattern of transmission from village to village downriver has been reported.

2.  Rural community transmission is “off the grid” and often have little to no access to safe drinking water or sanitation- conditions that existed prior to the earthquake.

3.  Lack of appropriate infrastructure to support situational awareness and strongly suspected political interference to report accurate case counts.

4.  Little presence of public health officials or epidemiologists in rural areas.

The Health Cluster has, to some degree of success, mobilized medical response (Cholera Treatment Centers, CTCs) for key patient catchment sites such as St Marc and Arachaie with surge capacity created in PAP, Carrefour, and Leogane.  However, rural transmission continues in areas distant from these CTCs. Laboratory diagnostic capacity is severely limited and localized to main patient catchment areas, limiting an ability to rapidly confirm and plan intervention at sites throughout the countryside.  We recommend a high index of suspicion and to not wait for official confirmation to respond.

The major rainy season normally peaks in October and persists through November.  The prospect of Tropical Storm Tomas gaining strength to become a Category 1 hurricane before causing potential damage and heavy rain in Haiti further complicates attempts to gain control of the epidemic.

Status of the Biosurveillance Grid

The U.S. Centers for Disease Control and Prevention have officially declined to provide funding and operational collaboration to the 400+ overlapping social networks of responders represented by the HEAS.
This represents the second time CDC has declined to fund and participate.  The HEAS Mid Action report based on the initial 150 days of operations post-earthquake displays the power of operational biosurveillance-facilitated distributive networking for early warning and rapid response.

The Internally Displaced Persons Syndromic Surveillance system (IDPSS) has once again been resurrected by CDC in support of MSPP to provide situational awareness response support.  This system collapsed several times and was unable to produce actionable early warning during the post-quake period. The IDPSS did not provide early warning of the current cholera epidemic and has not provided continuous situation reports or data outputs to the response community.

CDC has decided to activate a cholera-specific surveillance system predicated on commune reporting through mobile teams.  It is unlikely this approach will reach operational maturity quickly enough to enable proactive and strategic response.  Without integrative collaboration, both the IDPSS and the cholera-specific surveillance systems will provide only a partial view into the crisis and will not promote maximally efficient deployment of surveillance and response resources.

Conclusion

Current evidence of under-reporting particularly from rural areas and case foci that appear and disappear without on-site public health evaluation, lack of timely situational awareness reports from public health, and refusals to interface with medical responders indicate an overwhelmed public health capacity complicated by political sensitivities.

Continued independent "smart swarming", utilizing distributive networks for situational awareness, is strongly recommended with an expectation for a protracted period of response operations.

Haiti Epidemic Advisory System SitRep: 21 OCTOBER

Updated 1246 GMT, 21 OCTOBER 2010

NOTE: The Advisory System has been reactivated now to support the current cholera epidemic.

 

Mission Objective.  Our objective is to identify those infectious diseases that may abruptly overwhelm the ad hoc medical infrastructure or be perceived by the Haitian public to be unusual or unexpected.  In other words, identification and prioritization of those infectious diseases that may become a crisis or disaster.

Biosurveillance Grid Status.  HEAS has now been reactivated in light of the current cholera epidemic.

Social Order-Health Security Nexus.  Findings from other countries suggest massive social disruption potential during epidemics of cholera.  As with any infectious disease crisis or disaster, provokation of social outcry and challenge to established authority may occur.

Meteorology.   Haiti is currently in the midst of the majory rainy season, which is expected to persist through November, having peaked in October.

Medical Response Grid Status.  Should an infectious disease crisis become manifest, the risk of inundation and overwhelming medical logistics during a rapid response campaign has remained high since the earthquake.  This concern is magnified in the context of the current major rainy season.  Current observations of the cholera epidemic in Artibonite validate this assessment, as reports indicate hospital capacity has been exceeded and patients are being diverted to other clinics (see below).
 

Current Advisories and Situation Status.

 

 

 

Op Notes- Disruptors

Cholera.  The current situation is UNSTABLE.  Current monitoring indicates a rapidly evolving situation, and we have indications of an infectious disease event due to cholera in Haiti (Artibonite Valley) rated now at a possible IDIS Category 4 infectious disease event transitioning to a Category 5, defined as:

IDIS Category 4. Infectious disease event associated with social disruption. Category 4 events highlight when organized response has occurred, yet significant social disruption has been documented.

IDIS Category 5. Infectious disease event associated with disaster indicators.

Key observations as of the date/time of this message:

• Non-routine occurrence of diarrheal disease, described by Dr. Claude Surena, President of the Haitian Medical Association, as "according to the results of the analysis carried out in the laboratory it is cholera" to AFP.
• We note that laboratory-confirmed cholera has not been reported since the early 1990s
• Photographs and direct observations from St Nicholas Hospital in St Marc and comments from Surena indicate the hospital is overwhelmed and now in the process of divesting patients to other clinics for treatment- indicative/suggestive of local medical capacity collapse; photographs show multiple patients on IV therapy ORS is being used and is being mobilized.
• PROMESS aware, however logistics status unknown.
• Local infrastructure to respond in Artibonite is severely limited, with evidence of poor information sharing and alerting capacity. Public health resources are much more limited than in Port-au-Prince
• Significant community anxiety noted; indigenous Haitians claiming the presence of "cholera" and surging advice via Twitter for proper food and water handling / sanitation precautions
• International NGOs are mobilizing, and the UN Clusters are mobilizing around the issue such as WASH and Health.
• Statements to-date/time from WHO/PAHO and MSPP emphasize no laboratory confirmation; we expect official statements soon
• Tremendous and abrupt international sensitization as evidenced by Twitter and HEAS web portal hit counts 

Transition of this event to a Category 5 is expected at this point; immediate international response is required.

 

Operational Definitions

   

Infectious disruptor agent.  An infectious disease capable of triggering crisis or disaster conditions.

Infectious disease crisis: time-sensitive, non-routine organization-level decisions that may affect a local community’s activities of daily living. It is more common such decision-making falls within the 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.  The majority concern here is pediatric diarrheal disease in Haiti.

Infectious disease disaster: when crisis mode decision making by public health officials or institution fails to control the situation, either from an informational or response perspective and substantial social disruption associated with features of community disintegration occurs as a result.  The term "community disintegration" refers to the dissolution of a community as a social unit. This is classically exemplified by the evacuation of a village due to an uncontrolled epidemic. Ebola is the archetype disruptor capable of generating this kind of social response in undeveloped areas of Africa.  We do not consider any of the infectious disease in Haiti to represent potential disasters.

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. 

Infectious Disease Impact Scale (IDIS).  The IDIS is a model that serves as a guide to understanding the impact of acutely disruptive infectious disease events through the lens of disaster sociology.  Up to this point, event descriptors of infectious disease events have focused on use of terms like "outbreak" and "epidemic".  This is problematic in the operational setting when valid epidemiological data is often sparse or unreliable.  Here we consider a different perspective that focuses on the interface between an infectious disease hazard and indigenous vulnerability. 

IDIS Category 0.  Unreported infectious disease event.  Daily, routine infectious diseases are handled at this level, and provision of warning about these diseases is not deemed 'relevant'.

IDIS Category 1.  Reported infectious disease event.  The typical Category 1 infectious disease event reported by a community reflects a sensitivity to public health or medical significance.  No other significant features indicative of immediate public health or medical infrastructure impact, public anxiety, or civil unrest triggered by the event are noted.

IDIS Category 2.  Infectious disease event associated with routine organized response.  Category 2 events often reflect locally well-known diseases that nevertheless generate a demand for organization-level time-sensitive action.  This action is local routine. 

IDIS Category 3.  Infectious disease event associated with non-routine organized response.  Category 3 events are essentially the beginnings of a community crisis.

IDIS Category 4.  Infectious disease event associated with social disruption.  Category 4 events highlight when organized response has occurred, yet significant social disruption has been documented.

IDIS Category 5.  Infectious disease event associated with disaster indicators. 

IDIS Category 6.  Infectious disease event associated with apocalyptic indicators.

 

 

 

 

 

Haiti Epidemic Advisory System SitRep: 21 OCTOBER