On 2017 Sep 18, Robert Garry commented:

A Correction to the May 18, 2016 Correction appeared September 7, 2017. https://doi.org/10.1371/journal.ppat.1005503

The Correction of the Correction states, "The location of Mangala and Boma were switched.”

In fact the locations of Mangala and Boma were not switched. Boma is correctly located in all three versions of the map near the first bend of the Congo River

The village of Mangala is incorrectly located to the east of Boma in the Original Figure 1 [September 27, 2012] and again in the September 7, 2017 Correction of the Correction. Mangala is located to the west of Boma closer to to mouth of the Congo River in the May 18, 2016 Correction. Neither location for Mangala is correct. Rather, Mangala is located about 30 kilometers to the north of Boma.

The Boma Health zone is also incorrectly labelled as the Boma Bungu Health zone in all three versions of Figure 1 and in Figure 7.

These facts are easily verified using a Médecins Sans Frontières (MSF) and other information in my August 11, 2017 post below.

On 2017 Aug 11, Robert Garry commented:

Neither location for Mangala in Grard, Fair, Lee et al., 2012 or the May 18, 2016 correction is correct.

The original version of Figure 1 of Grard, Fair, Lee et al., 2012 and the version from the May 18, 2016 correction show Mangala, the village of the 3 cases reported, to be located in different locations. Neither location is correct. Also, the Boma health zone is incorrectly labeled as the Boma Bungu health zone in Figure 1. Grard, Fair, Lee et al., 2012.

Grard, Fair, Lee et al., 2012 Fig. 1: http://i.imgur.com/pyHvZcU.png

Correction to Grard, Fair, Lee et al., Fig. 1: http://i.imgur.com/ImLkfJi.png

Further, while the authors fixed the country names for Sierra Leone, Liberia, Ivory Coast and Ghana in the “corrected” version, several other errors were not fixed. These include, among other errors, the fact that Sierra Leone has had Lassa fever cases (the most in the world). Guinea, Liberia, Mali, Ivory Coast Togo and Benin have also experienced Lassa fever cases or have been the source of exported cases. Ghana has never had a case of Ebola, but Ivory Coast was the site of a single isolation of Tai Forest virus, a related filovirus. Sudan is incorrectly labeled as North Sudan and has never had an Ebola or Sudan virus case.

The Boma Health Zone is also incorrectly referred to as the Boma Bungu Health Zone in Fig. 7A of Grard, Fair, Lee et al., 2012.

Fig. 7A of Grard, Fair, Lee et al., 2012: http://i.imgur.com/W7kByHN.png

A Médecins Sans Frontières (MSF) map shows that the Boma Health Zone includes the city of Boma (pop 200K), and the rural Boma Bungu Health Zone (HZ) incudes Mangala.

MSF map: http://i.imgur.com/Ru2Zk7s.png

Figure 1 (both versions) and Figure 7 of Grard, Fair, Lee et al. 2012 incorrectly label the Boma HZ as the Boma Bungu HZ.

Two independent sources provided me with the true location of the village of Mangala. First, researchers at the Luki Biosphere sent me a map. Mangala (blue arrow) is located near one of the entrances of the Biosphere occupying a bend in the Luki River on a road 30 Km north of Boma.

Mangala near Luki Biosphere:http://imgur.com/V4XHd16

The MSF Chief of Mission who investigated the Mangala outbreak reported that Mangala was 1) along the Luki river ; 2) along a ‘main’ road and 3) northwards from Boma and the Congo river.

Mangala’s actual location is fully consistent with the early report of this disease cluster.

http://www.promedmail.org/post/20090626.2326

“5 people out of almost a dozen ill individuals have died within a few days from an as yet unidentified disease in the village of Mangala, located some 30 kilometers [19 mi] from the city of Boma, the capital of the province of Bas-Congo [Kongo Central] in the west of the Democratic Republic of the Congo (DRC).”

On 2017 Jul 25, Robert Garry commented:

None

On 2017 Jul 24, Robert Garry commented:

None

On 2016 May 08, Robert Garry commented:

The authors should explain why the location of Mangala is different in the original Figure 1 and in the "corrected" Figure 1.

On 2016 Apr 10, Robert Garry commented:

Two children

I have asked for redacted/deidentified/anonymised copies of original case records used to compile Table 1 and to draw other conclusions surrounding the 2009 deaths of two children from the Democratic Republic of the Congo (DRC) in Grard, Fair, Lee et al., 2012.

This request was made under PLoS’ data access policy (http://journals.plos.org/plosone/s/data-availability). The first line of the policy reads, “PLOS journals require authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception.” The policy goes on to state, "The data policy was implemented on March 3, 2014. Any paper submitted before that date will not have a data availability statement. However for all manuscripts submitted or published before this date, data must be available upon reasonable request." Thus, the PLoS data policy applies to Grard, Fair, Lee et al., 2012.

The PLoS policy gives explicit “Guidance on sharing data sets that derive from clinical studies or other work involving human participants.” According to the PLoS policy, “For studies involving human participants, data must be handled so as to not compromise study participants' privacy.” PloS Pathogens rejected my data request stating, “… because the data requested includes records with confidential patient information, we will not pursue this matter any further as potential compromise of patient privacy is a stated exception to our data policy.”

It is important to note that the PLoS policy links to several websites that provide guidance on anonymisation of patient data. One of the links provided is to the UK Data Archive (http://www.data-archive.ac.uk/create-manage/consent-ethics/anonymisation). The UK Data Archive guidance states, “Direct identifiers are often collected as part of the research administration process but are usually not essential research information and can therefore easily be removed from the data.”

The data in Grard, Fair, Lee et al., 2012 certainly does not fall under one of the rare exceptions to PLoS’ data access policy. On the contrary, the PLoS policy is clear that patient data MUST be provided after proper anonymisation that protects patient privacy. Numerous direct identifiers of the human subjects discussed in this manuscript were provided by the authors, including the name of the village, age and in the case of the two children their date of death. The authors have also made several statements that contradict the original reports of these cases in the DRC. It is a trivial matter to remove other identifiers from the original patient records, such as their names, as I have requested. It is a reasonable request to see the original data about abdominal pain, epistaxis, conjunctival injection, oral hemorrhage, hemorrhagic vomiting, hemorrhagic bleeding etc. found in Table 1 of Grard, Fair and Lee 2012. This is in no way identifying information and should be made available under PloS’ data access policy.

Many scientists, organizations and journal editors are now calling for open data policies regarding serious public health threats, such as Zika virus (http://www.npr.org/sections/goatsandsoda/2016/04/02/472686809/scientists-say-its-time-to-end-parachute-research). Based on their analysis of the data summarized in Table 1 Grard, Fair, Lee et al., 2012 concluded that the two children who died in the DRC in 2009 and a nurse that attended then contracted a “typical” hemorrhagic fever. They also concluded that the Rhabdoviridae should now be included with the four other virus families with members that can cause hemorrhagic fever (Chiu, Fair and Leroy, Future Microbiology 8, 139-141, 2013). Bas-Congo virus, whose genome was sequenced from the blood of the nurse, is a rhabodovirus. PLoS has supported efforts to make data on Zika virus open access (http://blogs.plos.org/speakingofmedicine/2016/02/10/zika-emergency-puts-open-data-policies-to-the-test/). PLoS should follow suit and open access to data on Bas-Congo virus.

Robert Garry

On 2016 Apr 03, Robert Garry commented:

To the editors, administrators and board of PLoS Pathogens,

An erratum for this paper has now been published. Revised Figure 1 still contains important errors.

ERRORS (among others): Sierra Leone and other West African countries are now properly labeled, but omitted is the fact that Sierra Leone has experienced Lassa fever cases. In fact the incidence of Lassa fever in Sierra Leone is considered to be the world’s highest. Likewise, Guinea also has many documented Lassa fever cases, but this is omitted. Other the hand, Ghana has never had a documented Ebola virus disease case. Ivory Coast has had one documented case of human infection with Tai Forest virus, which is a filovirus in the same family as Ebola virus.

The authors took this opportunity to change the name of “Bas-Congo hemorrhagic fever” to “Mangala hemorrhagic fever.” I object to both terms as the authors have not established by virology community standards that any human subject discussed in their original manuscript actually had a viral hemorrhagic fever or that their illness was caused by Bas-Congo virus. The new term “Mangala hemorrhagic fever” was not used in the manuscript and its use in this still multiply flawed revised figure is not justified.

The authors provided "raw, uncropped blots” used previously to prepare Supplemental Figure 2. They acknowledge that the original blot used to generate this figure 1) contained additional lanes that had been removed, 2) the incorrect lane 4 was inserted into the published figure and 3) a black-and-white inversion and global gamma correction was applied to the entire image. This is an incomplete accounting of the image manipulations that are now apparent upon inspection of the original image. It would also be appropriate for the authors to acknowledge that further unannounced image manipulations occurred in the published figure resulting in duplications of background artifacts (for example, artifacts resembling question marks near the control band) and other incongruences, such a lighter background around the control band.

The authors of this paper should make available for review the data used to derive Table 1 in Grard, Fair, Lee et al., 2012. Table 1 includes very explicit information about abdominal pain, epistaxis, conjunctival injection, oral hemorrhage, hemorrhagic vomiting, hemorrhagic bleeding, etc. This is an incredible amount of clinical detail to come from a village health unit, particularly in a retrospective analysis of cases. As an expert in viral hemorrhagic fevers, I made a reasonable open data request under the PLoS guidelines to review the patient charts used to compile Table 1.

There are several reasons why I believe that an independent review of this data by an independent expert is necessary and important:

It is extremely difficult to make a diagnosis of any viral hemorrhagic fever on clinical grounds alone, yet the authors wrote definitively that these represented a “ cluster of three human cases of typical acute hemorrhagic fever.” This can only be verified by examining the patient records.

For background I suggest the article found at:

http://www.sciencedirect.com/science/article/pii/S1473309915001607

Regarding the accuracy of clinical diagnosis of Ebola virus disease, these authors write, “Furthermore, the specificity of the WHO case definition was strikingly low at 31·5% (95% CI 26·0–37·6). This implies that 68% of patients who would be selected for admission to a holding unit would not actually have Ebola virus disease (false-positive results)—indeed, a toss of a coin might have yielded a better chance of an accurate result.”

This begs the question that if Ebola clinical diagnosis at the height of the still ongoing West African outbreak was worse than a guess, how can “Mangala hemorrhagic fever” [sic] be diagnosed years later from patient records found in a small African village public health unit?

As for whether the paper meets community standards for causation, it does not. I can do no better to justify this statement than one of the co-authors of Grard, Fair, Lee et al., 2012.

From Future Microbiol. 2010;5(2):177-189: "The discovery of a novel virus in a clinical sample from an individual with an acute or chronic disease does not imply causation or even bona fide association of the virus with the disease."

Bas-Congo virus has been isolated from only a single individual. You cannot have proof of causation with n=1.

Another reason to review the data records is to resolve a discrepancy between the findings reported in Grard, Fair, Lee et al., 2012 who only reported 3 cases and the original reports of the cases, which stated that there were more cases, including others that died. This cluster of cases was first reported in a June 22, 2009 ProMED report stating, “5 people out of almost a dozen ill individuals have died within a few days from an as yet unidentified disease in the village of Mangala.” The original newswire reports, “Patients are being treated with a combination of antibiotics, antacids and other tonics, as well as transfusion and rehydration suggesting that possible bacterial etiologies are being considered.” In this regard, it is worth noting that case three in Grard, Fair, Lee et al. 2012 from whom Bas-Congo virus was isolated recovered quickly after antibiotic treatment. I am interested in examining all the clinical data to determine why other cases were excluded from Grard, Fair, Lee et al. 2012.

My prior request to PLoS Pathogens to see original patient data used to derived Table 1 data was rejected because, “In addition, because the data requested includes records with confidential patient information, we will not pursue this matter any further as potential compromise of patient privacy is a stated exception to our data policy.”

While the PLoS journals should be commended for strengthening their data access policies, in particular to extend their open data requirement to all PLoS manuscripts including those published in 2012, confidentiality is absolutely no grounds to reject my request to see the key elements of the data (the signs and symptoms listed in Table 1) from the patient records in Mangala village in DRC in 2009. The parents of the two children who died (cases one and two) and the nurse that attended them (case three) gave consent to give use private information as case reports in the paper. Age, date of death of the children, and the name of the village (Mangala) are all provided in the paper. It is a trivial matter to redact all of the identifying information from these patient records.

I am publicly requesting access to the patient records used to compile Table 1 of Grard, Fair, Lee et al., 2012, with all identifying information redacted so as to completely avoid confidentiality issues.

Sincerely,

Robert F. Garry, PhD

On 2016 Apr 03, Robert Garry commented:

To the editors, administrators and board of PLoS Pathogens,

An erratum for this paper has now been published. Revised Figure 1 still contains important errors.

ERRORS (among others): Sierra Leone and other West African countries are now properly labeled, but omitted is the fact that Sierra Leone has experienced Lassa fever cases. In fact the incidence of Lassa fever in Sierra Leone is considered to be the world’s highest. Likewise, Guinea also has many documented Lassa fever cases, but this is omitted. Other the hand, Ghana has never had a documented Ebola virus disease case. Ivory Coast has had one documented case of human infection with Tai Forest virus, which is a filovirus in the same family as Ebola virus.

The authors took this opportunity to change the name of “Bas-Congo hemorrhagic fever” to “Mangala hemorrhagic fever.” I object to both terms as the authors have not established by virology community standards that any human subject discussed in their original manuscript actually had a viral hemorrhagic fever or that their illness was caused by Bas-Congo virus. The new term “Mangala hemorrhagic fever” was not used in the manuscript and its use in this still multiply flawed revised figure is not justified.

The authors provided "raw, uncropped blots” used previously to prepare Supplemental Figure 2. They acknowledge that the original blot used to generate this figure 1) contained additional lanes that had been removed, 2) the incorrect lane 4 was inserted into the published figure and 3) a black-and-white inversion and global gamma correction was applied to the entire image. This is an incomplete accounting of the image manipulations that are now apparent upon inspection of the original image. It would also be appropriate for the authors to acknowledge that further unannounced image manipulations occurred in the published figure resulting in duplications of background artifacts (for example, artifacts resembling question marks near the control band) and other incongruences, such a lighter background around the control band.

The authors of this paper should make available for review the data used to derive Table 1 in Grard, Fair, Lee et al., 2012. Table 1 includes very explicit information about abdominal pain, epistaxis, conjunctival injection, oral hemorrhage, hemorrhagic vomiting, hemorrhagic bleeding, etc. This is an incredible amount of clinical detail to come from a village health unit, particularly in a retrospective analysis of cases. As an expert in viral hemorrhagic fevers, I made a reasonable open data request under the PLoS guidelines to review the patient charts used to compile Table 1.

There are several reasons why I believe that an independent review of this data by an independent expert is necessary and important:

It is extremely difficult to make a diagnosis of any viral hemorrhagic fever on clinical grounds alone, yet the authors wrote definitively that these represented a “ cluster of three human cases of typical acute hemorrhagic fever.” This can only be verified by examining the patient records.

For background I suggest the article found at:

http://www.sciencedirect.com/science/article/pii/S1473309915001607

Regarding the accuracy of clinical diagnosis of Ebola virus disease, these authors write, “Furthermore, the specificity of the WHO case definition was strikingly low at 31·5% (95% CI 26·0–37·6). This implies that 68% of patients who would be selected for admission to a holding unit would not actually have Ebola virus disease (false-positive results)—indeed, a toss of a coin might have yielded a better chance of an accurate result.”

This begs the question that if Ebola clinical diagnosis at the height of the still ongoing West African outbreak was worse than a guess, how can “Mangala hemorrhagic fever” [sic] be diagnosed years later from patient records found in a small African village public health unit?

As for whether the paper meets community standards for causation, it does not. I can do no better to justify this statement than one of the co-authors of Grard, Fair, Lee et al., 2012.

From Future Microbiol. 2010;5(2):177-189: "The discovery of a novel virus in a clinical sample from an individual with an acute or chronic disease does not imply causation or even bona fide association of the virus with the disease."

Bas-Congo virus has been isolated from only a single individual. You cannot have proof of causation with n=1.

Another reason to review the data records is to resolve a discrepancy between the findings reported in Grard, Fair, Lee et al., 2012 who only reported 3 cases and the original reports of the cases, which stated that there were more cases, including others that died. This cluster of cases was first reported in a June 22, 2009 ProMED report stating, “5 people out of almost a dozen ill individuals have died within a few days from an as yet unidentified disease in the village of Mangala.” The original newswire reports, “Patients are being treated with a combination of antibiotics, antacids and other tonics, as well as transfusion and rehydration suggesting that possible bacterial etiologies are being considered.” In this regard, it is worth noting that case three in Grard, Fair, Lee et al. 2012 from whom Bas-Congo virus was isolated recovered quickly after antibiotic treatment. I am interested in examining all the clinical data to determine why other cases were excluded from Grard, Fair, Lee et al. 2012.

My prior request to PLoS Pathogens to see original patient data used to derived Table 1 data was rejected because, “In addition, because the data requested includes records with confidential patient information, we will not pursue this matter any further as potential compromise of patient privacy is a stated exception to our data policy.”

While the PLoS journals should be commended for strengthening their data access policies, in particular to extend their open data requirement to all PLoS manuscripts including those published in 2012, confidentiality is absolutely no grounds to reject my request to see the key elements of the data (the signs and symptoms listed in Table 1) from the patient records in Mangala village in DRC in 2009. The parents of the two children who died (cases one and two) and the nurse that attended them (case three) gave consent to give use private information as case reports in the paper. Age, date of death of the children, and the name of the village (Mangala) are all provided in the paper. It is a trivial matter to redact all of the identifying information from these patient records.

I am publicly requesting access to the patient records used to compile Table 1 of Grard, Fair, Lee et al., 2012, with all identifying information redacted so as to completely avoid confidentiality issues.

Sincerely,

Robert F. Garry, PhD

On 2016 Apr 10, Robert Garry commented:

Two children

I have asked for redacted/deidentified/anonymised copies of original case records used to compile Table 1 and to draw other conclusions surrounding the 2009 deaths of two children from the Democratic Republic of the Congo (DRC) in Grard, Fair, Lee et al., 2012.

This request was made under PLoS’ data access policy (http://journals.plos.org/plosone/s/data-availability). The first line of the policy reads, “PLOS journals require authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception.” The policy goes on to state, "The data policy was implemented on March 3, 2014. Any paper submitted before that date will not have a data availability statement. However for all manuscripts submitted or published before this date, data must be available upon reasonable request." Thus, the PLoS data policy applies to Grard, Fair, Lee et al., 2012.

The PLoS policy gives explicit “Guidance on sharing data sets that derive from clinical studies or other work involving human participants.” According to the PLoS policy, “For studies involving human participants, data must be handled so as to not compromise study participants' privacy.” PloS Pathogens rejected my data request stating, “… because the data requested includes records with confidential patient information, we will not pursue this matter any further as potential compromise of patient privacy is a stated exception to our data policy.”

It is important to note that the PLoS policy links to several websites that provide guidance on anonymisation of patient data. One of the links provided is to the UK Data Archive (http://www.data-archive.ac.uk/create-manage/consent-ethics/anonymisation). The UK Data Archive guidance states, “Direct identifiers are often collected as part of the research administration process but are usually not essential research information and can therefore easily be removed from the data.”

The data in Grard, Fair, Lee et al., 2012 certainly does not fall under one of the rare exceptions to PLoS’ data access policy. On the contrary, the PLoS policy is clear that patient data MUST be provided after proper anonymisation that protects patient privacy. Numerous direct identifiers of the human subjects discussed in this manuscript were provided by the authors, including the name of the village, age and in the case of the two children their date of death. The authors have also made several statements that contradict the original reports of these cases in the DRC. It is a trivial matter to remove other identifiers from the original patient records, such as their names, as I have requested. It is a reasonable request to see the original data about abdominal pain, epistaxis, conjunctival injection, oral hemorrhage, hemorrhagic vomiting, hemorrhagic bleeding etc. found in Table 1 of Grard, Fair and Lee 2012. This is in no way identifying information and should be made available under PloS’ data access policy.

Many scientists, organizations and journal editors are now calling for open data policies regarding serious public health threats, such as Zika virus (http://www.npr.org/sections/goatsandsoda/2016/04/02/472686809/scientists-say-its-time-to-end-parachute-research). Based on their analysis of the data summarized in Table 1 Grard, Fair, Lee et al., 2012 concluded that the two children who died in the DRC in 2009 and a nurse that attended then contracted a “typical” hemorrhagic fever. They also concluded that the Rhabdoviridae should now be included with the four other virus families with members that can cause hemorrhagic fever (Chiu, Fair and Leroy, Future Microbiology 8, 139-141, 2013). Bas-Congo virus, whose genome was sequenced from the blood of the nurse, is a rhabodovirus. PLoS has supported efforts to make data on Zika virus open access (http://blogs.plos.org/speakingofmedicine/2016/02/10/zika-emergency-puts-open-data-policies-to-the-test/). PLoS should follow suit and open access to data on Bas-Congo virus.

Robert Garry

On 2016 May 08, Robert Garry commented:

The authors should explain why the location of Mangala is different in the original Figure 1 and in the "corrected" Figure 1.

On 2017 Jul 24, Robert Garry commented:

None

On 2017 Jul 25, Robert Garry commented:

None

On 2017 Aug 11, Robert Garry commented:

Neither location for Mangala in Grard, Fair, Lee et al., 2012 or the May 18, 2016 correction is correct.

The original version of Figure 1 of Grard, Fair, Lee et al., 2012 and the version from the May 18, 2016 correction show Mangala, the village of the 3 cases reported, to be located in different locations. Neither location is correct. Also, the Boma health zone is incorrectly labeled as the Boma Bungu health zone in Figure 1. Grard, Fair, Lee et al., 2012.

Grard, Fair, Lee et al., 2012 Fig. 1: http://i.imgur.com/pyHvZcU.png

Correction to Grard, Fair, Lee et al., Fig. 1: http://i.imgur.com/ImLkfJi.png

Further, while the authors fixed the country names for Sierra Leone, Liberia, Ivory Coast and Ghana in the “corrected” version, several other errors were not fixed. These include, among other errors, the fact that Sierra Leone has had Lassa fever cases (the most in the world). Guinea, Liberia, Mali, Ivory Coast Togo and Benin have also experienced Lassa fever cases or have been the source of exported cases. Ghana has never had a case of Ebola, but Ivory Coast was the site of a single isolation of Tai Forest virus, a related filovirus. Sudan is incorrectly labeled as North Sudan and has never had an Ebola or Sudan virus case.

The Boma Health Zone is also incorrectly referred to as the Boma Bungu Health Zone in Fig. 7A of Grard, Fair, Lee et al., 2012.

Fig. 7A of Grard, Fair, Lee et al., 2012: http://i.imgur.com/W7kByHN.png

A Médecins Sans Frontières (MSF) map shows that the Boma Health Zone includes the city of Boma (pop 200K), and the rural Boma Bungu Health Zone (HZ) incudes Mangala.

MSF map: http://i.imgur.com/Ru2Zk7s.png

Figure 1 (both versions) and Figure 7 of Grard, Fair, Lee et al. 2012 incorrectly label the Boma HZ as the Boma Bungu HZ.

Two independent sources provided me with the true location of the village of Mangala. First, researchers at the Luki Biosphere sent me a map. Mangala (blue arrow) is located near one of the entrances of the Biosphere occupying a bend in the Luki River on a road 30 Km north of Boma.

Mangala near Luki Biosphere:http://imgur.com/V4XHd16

The MSF Chief of Mission who investigated the Mangala outbreak reported that Mangala was 1) along the Luki river ; 2) along a ‘main’ road and 3) northwards from Boma and the Congo river.

Mangala’s actual location is fully consistent with the early report of this disease cluster.

http://www.promedmail.org/post/20090626.2326

“5 people out of almost a dozen ill individuals have died within a few days from an as yet unidentified disease in the village of Mangala, located some 30 kilometers [19 mi] from the city of Boma, the capital of the province of Bas-Congo [Kongo Central] in the west of the Democratic Republic of the Congo (DRC).”

On 2017 Sep 18, Robert Garry commented:

A Correction to the May 18, 2016 Correction appeared September 7, 2017. https://doi.org/10.1371/journal.ppat.1005503

The Correction of the Correction states, "The location of Mangala and Boma were switched.”

In fact the locations of Mangala and Boma were not switched. Boma is correctly located in all three versions of the map near the first bend of the Congo River

The village of Mangala is incorrectly located to the east of Boma in the Original Figure 1 [September 27, 2012] and again in the September 7, 2017 Correction of the Correction. Mangala is located to the west of Boma closer to to mouth of the Congo River in the May 18, 2016 Correction. Neither location for Mangala is correct. Rather, Mangala is located about 30 kilometers to the north of Boma.

The Boma Health zone is also incorrectly labelled as the Boma Bungu Health zone in all three versions of Figure 1 and in Figure 7.

These facts are easily verified using a Médecins Sans Frontières (MSF) and other information in my August 11, 2017 post below.