LESSONS LEARNED DURING ACTIVE ... - EBSCOhost

87 downloads 0 Views 570KB Size Report
Overall, 26 hemorrhagic epidemic outbreaks have been registered in 12 countries; 18 caused by the Ebola virus and eight by the Marburg virus. About 2551 ...
East African Journal of Public Health Volume 7 Numbe 1 March 2010

32

LESSONS LEARNED DURING ACTIVE EPIDEMIOLOGICAL SURVEILLANCE OF EBOLA AND MARBURG VIRAL HEMORRHAGIC FEVER EPIDEMICS IN AFRICA Yokouide Allaranga1, Mamadou Lamine Kone1, Pierre Formenty2, Francois Libama3, Paul Boumandouki3, Celia JI Woodfill1, Idrissa Sow1, Sambe Duale4, Wondimagegnehu Alemu1, Adamou Yada1

ABSTRACT Objective: To review epidemiological surveillance approaches used during Ebola and Marburg hemorrhagic fever epidemics in Africa in the past fifteen years. Overall, 26 hemorrhagic epidemic outbreaks have been registered in 12 countries; 18 caused by the Ebola virus and eight by the Marburg virus. About 2551 cases have been reported, among which 268 were health workers (9,3%). Methods: Based on articles and epidemic management reports, this review analyses surveillance approaches, route of introduction of the virus into the population (urban and rural), the collaboration between the human health sector and the wildlife sector and factors that have affected epidemic management. Findings: Several factors affecting the epidemiological surveillance during Ebola and Marburg viruses hemorrhagic epidemics have been observed. During epidemics in rural settings, outbreak investigations have shown multiple introductions of the virus into the human population through wildlife. In contrast, during epidemics in urban settings a single introduction of the virus in the community was responsible for the epidemic. Active surveillance is key to containing outbreaks of Ebola and Marburg viruses Conclusions: Collaboration with those in charge of the conservation of wildlife is essential for the early detection of viral hemorrhagic fever epidemics. Hemorrhagic fever epidemics caused by Ebola and Marburg viruses are occurring more and more frequently in Sub-Saharan Africa and only an adapted epidemiological surveillance system will allow for early detection and effective response.

Key words: Ebola, Marburg, epidemic, surveillance, collaboration, early detection, contacts follow-up, Africa. Introduction Outbreaks of viral hemorrhagic fever (VHF) epidemics, especially the ones caused by Ebola and Marburg viruses, are becoming a more and more frequent public health problem in sub-Saharian countries. Epidemics of Ebola and Marburg VHF are associated with a high mortality, ranging from 24% to 90% 1. The Marburg and Ebola viruses belong to the same family of Filoviruses. During their clinical progression, the VHF caused by Filoviruses are frequently accompanied by hemorrhagic signs and the virus is widely disseminated throughout the body. The blood, urine, vomit, feces, pus, sperm and saliva from patients affected by VHF are all infectious. Thus, the risk of transmission of the virus to health workers and to the community is high during provision of health care or during preparation of the deceased for the funeral. Overall, between 1976 and 2007, 26 epidemics of viral hemorrhagic fever have been registered in 12 countries, among which 18 were caused by Ebola and eight were caused by Marburg (Figure 1). Around 2855 cases of Ebola and Marburg infections have been reported during this time period, among which 268 are health workers (9.3%). The Ebola and Marburg viruses‘ hemorrhagic fever epidemics are becoming more and more frequent in certain parts of Africa. On average, there has been an yearly occurrence of Ebola or Marburg epidemic since 1994. In order to rapidly control these epidemics and reduce their impact, key interventions such as epidemiological surveillance, have been established. Epidemiological surveillance is one of the five essential components of the management of VHF Correspondence to Yokouidé Allarangar, e-mail: [email protected] 1

WHO, Regional office for Africa, Brazzaville, Congo, 2WHO Headquarters, Geneva, Switzerland, 3Ministry of Health, Brazzaville, Congo, 4Africa‘s Health in 2010 Project, Tulane University School of Public Health and Tropical Medicine, New Orleans, USA

epidemics (Figure 2). The other four components, contributing to the efficiency of epidemiological surveillance, are (1) coordination, (2) social mobilization, (3) psychosocial support, and (4) logistic and security support3. During an epidemic, the coordinating body organizes and orients the work of all the technical teams present in the field, including the surveillance team. Social mobilization is an essential element directing the community to adhere to the activities set up in the field , which is critical for the control of the epidemic The same remarks are true for psychosocial support. Finally, without logistic and security support, the epidemiological surveillance team could not carry out its activities. Each of these components are represented by a sub-committee. The epidemiological surveillance sub-committee also directs laboratory activities2. The laboratory is an essential supporting component during the surveillance of VHF epidemics. Laboratory tests allow classification of alert and suspected cases as either confirmed cases or negative cases. Recent capacity building of laboratories in the African region has greatly facilitated confirmation of VHF outbreaks; the epidemics in Congo and in the Democratic Republic of Congo were confirmed by the International Center of Medical Research of Franceville in Gabon in less than three days after samples were taken. Until a few years ago, all samples had been sent to the USA, Canada or Europe and results were not available before 2 to 3 weeks or more. The development of sub-regional reference laboratories for VHFs, and above all mobile laboratories, has revolutionized epidemiological surveillance as the laboratory confirmatory results can now be obtained within a few hours. The aim of this article is to review the experiences and lessons learned during active epidemiological surveillance of Ebola and Marburg viruses hemorrhagic fever epidemics in the African Region during the past 15 years.

East African Journal of Public Health Volume 7 Numbe 1 March 2010

Methodology The principal objective of surveillance during epidemics is early detection of all suspected cases of VHF in humans. As with any epidemiological surveillance, surveillance during hemorrhagic fever epidemics, mainly due to Ebola and Marburg viruses, is essentially based on definition of cases (alert (reported?) case, suspect case, probable case and confirmed case). This case definition is adapted from the technical guidelines for integrated disease surveillance of the World Health Organization (WHO) Regional Office for Africa, in order to take into account socio-demographic characteristics4,5. We reviewed reports and publications concerning Ebola and Marburg VHF epidemics which took place between 1994 and 2007. The reports were mainly written by the response operational teams and the articles were written by several teams after the resolution of these epidemics. We have analyzed the different epidemiological surveillance approaches, the factors that may have affected these surveillances, the introduction of the virus in rural and urban settings and the collaboration between the health sector and the wildlife sector. The following table sums up the chronology of viral hemorrhagic fever outbreaks1.

the end of the surveillance and the follow-up ending date. This form allows the daily follow-up of all contact cases of one index case until the end of their follow-up. All suspected cases that have been identified will be notified on a form which allows collection of all useful information about this case (socio-demographic data, geographic data, origin of the contamination, clinical symptoms, and laboratory analysis). In addition to this notification case form, there is a list which routinely sums up all registered cases. Table 1 : Distribution of viral hemorrhagic fever outbreaks caused by Ebola and Marburg viruses, classified by year, cases, deaths and lethality, 1976-2007. Epidemic

Year 1976

1977

Ebola

1979 1994 1995

Results

1996

Active epidemiological surveillance approaches and epidemic management During Ebola and Marburg VHF epidemics, epidemiological surveillance in the field is done by mobile teams and at surveillance focal points in the hospital facilities. Specifically, the surveillance included: - Active case-finding of suspected cases for immediate management and confinement; - Investigate each reported case; - Identify all new contact cases; - Register the cases, the deaths and the contact cases; - Follow-up daily with mobile teams all contact and registered cases and organize medical management if needed; - Establish epidemiological links between cases and describe epidemiological patterns of the epidemic; - Create a detailed data base of cases and contact cases ; - Publish an epidemiological bulletin in the newspaper. Follow-up of contact cases is made with a standardized form. The follow-up is carried on for 21 days after the last exposure date. For a more efficient surveillance, teams use the door-to-door strategy in villages and city districts. In order to allow contact cases to attend their daily duties, such as working in the fields (in rural settings), the mobile teams follow-up the cases early in the morning. This strategy allows coverage of almost 100% of the cases on a daily basis6. A surveillance form has been designed for the contact cases‘ follow-up (Table 2). This form allows collection of the following data: locality, index case, number of days of follow-up, number of days left before

33

20002001 20012002 20022003 2004 2005 2007 Total 1967

Marburg

1975 1980 1987 19982000 20042005 2007 Total

Country

Cases

Deaths

Sudan DRC (former Zaïre) DRC (former Zaïre) Sudan Gabon Ivory Coast Liberia DRC (former Zaïre) Gabon South Africa Uganda

284 318

151 280

Lethality (%) 53 88

1

1

100

34 52 1 1 315

22 31 0 0 250

65 60 0 0 81

97 1 425

66 1 224

68 100 53

Gabon Congo Congo

65 59 178

53 44 157

82 75 88

Sudan Congo DRC Uganda

37 12 264 149 2293 31

12 10 187 37 1536 7

32 83 71 24 67 22

3 2 1 154

1 1 1 128

33 50 100 83

Angola

164

150

91

Uganda

3 358

1 282

33 78

Germany and Yugoslavia South Africa Kenya Kenya DRC

Table 2: Sample form of contact cases‘ follow-up for surveillance by localities and index cases., March 2002 au Gabon. Department

I. Zadie

II. Ivindo

Localities

Index case

Ilahounéné

Y. F N. A. Y. J. N. Y. B. W.

Mékouma Mékambo TOTAL Makokou TOTAL

L. A.

No. of contact cases 35 4 1 1 1 42 20 20

No. of days follow-up J16 J11 J16 J2 J5

No. day left 5 10 5 19 16

follow-up ending date

J20

1

16 march 02

20 march 02 26 march 02 20 march 02 03 april 02 31 march 02

In the community, the surveillance carried on by the mobile teams relies on a community network including

East African Journal of Public Health Volume 7 Numbe 1 March 2010

the village leaders, the community-based health workers and the local International Red Cross volunteers. Further surveillance is conducted in households if there is an indication or rumor of a suspected case. The suspected cases are managed in confined facilities or at home in case of refusal by the family. Members of the family who assist the suspected case at home should be informed of protection methods and protection equipment (gloves, bleach)4. Factors that have affected the active epidemiologic surveillance and the management of epidemics. Implementing active epidemiologic surveillance during the course of hemorrhagic fever epidemics of the Ebola and Marburg viruses was handicapped by a number of factors that included the following: The inaccessibility of localities that need surveillance: impractical roads in the villages that do not permit normal tracking of cases and contacts. There is too often insufficient logistical resources; this was the case in Uige in Angola, Mekambo in Gabon and Kelle in Congo where the surveillance teams often had to adjust their surveillance timetables. The use of logistical air support (such as helicopters) was not always possible (as in Angola). The follow up of contact subjects in the urban environment was often quite difficult to organize, particularly when there are a large number of subjects to follow every day. This was difficult because of the multiple locations of where people assemble and the difficulty in using public transport systems. The adherence of populations to surveillance activities, the existence of several trails through rural areas in the forest makes it easy for the populations to move from one locality to another have made it difficult to control the epidemics. The lack of motivation from the nursing personnel who believed fervently that they could infect themselves over the course of the epidemiologic surveillance activities via the cases and contact, and who found themselves displaced from their workplace for a long period of time without rational means of subsistence. The security of the epidemiologic surveillance teams was often threatened and we observed a phonomenon of stigmatisation of the medical corps and agressive behaviors (verbal and physical threats) towards them by native populations. Transmission of the Hemorrhagic Fever Viruses in the Population The results of the epidemiologic surveillance over the course of hemorrhagic fever epidemics that have occurred during this time period presented two profiles: either it was an epidemic in a rural environment or an epidemic of an urban environment.

Transmission of the Virus in a Rural Setting The epidemiologic and virologic research on epidemics that took place in rural settings has provided evidence of multiple introductions of the virus into the human population from wildlife2 as shown in Table 3. Often, hunters (professional or amateur) or their

34

porters are the ones who are infected first at the time of slaughter, handling or transport of primates or other infected animals collected in the forest. These hunters or porters, in turn, contaminate their colleagues and friends over the course of the first few weeks of the epidemic. Then family members infect themselves while treating the sick or from the mortuary toilet or during the funeral of their loved ones. If measures are not put into place immediately to control the epidemic, then it spreads throughout the community. It is transmitted especially among the women of a given population. In Gulu 63% of the cases were women7 and in Kikwit 166 women versus 149 men8. In Congo and in Gabon the number of women cases appeared equally high at the end of the epidemic as shown in Figure 4 below. In rural settings, traditional healers have played a direct role several times in transmission of the infection and amplification of the epidemic (Kikwit in 1995, Booué et Mayibout in 1996, Mbomo n 2002, Uige in 2005). Indeed, the sick often resort to traditional healers at the beginning of their sickness. Some of these ―traditional healers‖ treat by laying their hands on the sick, which presents an opportunity for transmission. Furthermore, the fact that the traditional healers often keep infected individuals at their home for several days, in contact with members of their families, could be a factor responsible for the spread of the disease in the community. Nevertheless, the main mode of contamination remains intrafamilial transmission at the time of caring for the sick and during funerals. Epidemics in rural settings are usually of small magnitude with about thirty cases among which there may be one or two health care workers2. This type of epidemic is, therefore, very difficult to detect except when it spreads to the level of a health center. Table 3: Number of introductions of Ebola Virus in the human population during the epidemics in Gabon and Congo Year 1994 2001-2002 2003

Town/Country Minkebe /Gabon Mayibout/Gabon Mekambo/Gabon Mbomo/Congo Kelle/Congo

Number of Introductions 3 2 5 3 3

Transmission in an Urban Setting Epidemiologic and virologic analyses suggest that a single introduction into the community was responsible for each urban epidemic; this was the case in Kikwit in 1995 and in Gulu in 20002. An epidemic in an urban setting is always spread within hospital settings that do not respect, due to lack of resources, the universal hygiene rules. Epidemics are rapidly controlled once the healthcare workers put in place protective measures while treating their patients. Health personnel pay a heavy toll during the course of these urban epidemics. Nosocomial infections affect all cadres of personnel: doctors, midwives, matrons, nurses, medical personnel, laboratory personnel, pharmacists, and paramedical personnel. In Kikwit, 70%

East African Journal of Public Health Volume 7 Numbe 1 March 2010

of the first wave of patients were health workers, while at the end of the epidemic they represented 26 % (Figure 5)8,9 and in Gulu, 31 cases were health workers among which 17 died8. During the Marburg epidemic in Angola, a total of 18 health workers were infected and died10,11. This demonstrates the vulnerability of the health workers, especially at the beginning of an epidemic, due to infection control programs in the health facilities9. Here surveillance in the hospital setting has a great importance. Improvement of viral hemorrhagic epidemics‘ management through collaboration between the health sector and the wildlife sector during active surveillance Considerable progress has been made in the active epidemiological surveillance of VHF outbreaks. Overall, the detection of Ebola epidemics is being made earlier than before. The epidemics among animals precede the human epidemics, thus animal deaths are used as early alert in the surveillance system of Ebola fevers. The alert is given earlier and can even precede the human epidemic, as it has been the case in Kéllé in 20032. The management of the fourth epidemic of Ebola virus hemorrhagic fever in Congo in 2005 is a good example (Figure 3). This epidemic was detected very early and controlled in 5 weeks12 with 12 cases and 10 deaths compared to 178 cases and 157 cases which occurred in the 2002-2003 epidemic. The latter lasted 4 months.

35

The second epidemic of Marburg in Uganda in 2007 is another example. Only 3 cases and 1 death were reported compared to 425 cases and 224 deaths during the first Ebola epidemic in 2000, which lasted 3 months and a half7, 13. This success can be attributed to the implementation of the integrated disease surveillance and response. In the two countries, priority was given to areas affected previously in order to reinforce capacities in active surveillance and response. Finally, success can also be attributed to a better collaboration between the human health services and the wildlife services with fairly regular meetings and exchange of information. The same was true for the network of laboratories who collaborated with the WHO for VHF. The excellent collaboration between the teams from the Projet d‘Ecosystème Forestier de l‘Afrique Centrale – ECOFAC (the Forest Eco-system of Central Africa Project) and the health teams from the Province de la Cuvette Ouest (Western Basin Province) over the course of epidemics in Congo was proof of this improved exchange of information. This collaboration permitted the launch of an alert for the majority of epidemics that appeared in Congo.

Figures

Figure 1 : Distribution of Ebola and Marburg epidemics in Africa between 1976 and 2007.

East African Journal of Public Health Volume 7 Numbe 1 March 2010

36

Figure 2 : Control strategy of viral hemorrhagic fever epidemics.

Figure 3 : Distribution of the cases and deaths from Ebola according to the starting date of the disease and of deaths in Congo in 2005.

East African Journal of Public Health Volume 7 Numbe 1 March 2010

37

Figure 4: Average number of women and men with cases of Ebola hemorrhagic fever in Congo and in Gabon, from week 44, 2001 to week 11, 2002.

Figure 5 : Ebola epidemic in Kikwit, Zaïre, distribution of cases by date of beginning of the symptoms and by professional status, 1995. Discussion The surveillance of VHF caused by Ebola and Marburg viruses during inter-epidemic periods is based on principles and classical methods of epidemiological surveillance. However during the epidemic period, for the sake of efficiency, the methods used during epidemics are adapted to local conditions and socioeconomic characteristics in order to take into account

active case discovery and the follow-up of contact cases14. The high literacy rate in urban settings facilitates the acceptance of measures set up by the health teams to fight against the epidemic, thus facilitating surveillance during these epidemics2. This is not the case in rural settings where the reticence of illiterate populations is frequent. After the occurrence of several VHF epidemics these past few years, epidemiological surveillance has

East African Journal of Public Health Volume 7 Numbe 1 March 2010

been reinforced. As a matter of fact, these epidemics have driven the national responsible persons, such as in Gabon, Congo Republic and Uganda, to strengthen the surveillance teams by organizing training sessions, setting up surveillance tools in the field and preparing emergency stocks of personal protective equipment and other consumables. The identification of suspected cases of viral hemorrhagic fever being the main concern, it is important that the health workers, and mainly staff members from health districts, are trained and regularly supervised in order to early detect cases and confine them to avoid the further spread of the disease and confirm the cases9. The door-to-door follow-up approach of contact cases has aided both early detection of contact cases who are developing the first signs of the disease and their isolation in a timely manner. This approach has minimized intra-familial contaminations that occurs when the patients stay too long at home with the family. During the epidemic period, contact should be constantly maintained with the staff responsible for wildlife surveillance. The fact of finding dead primates in the forests during the inter-epidemic, epidemic and postepidemic periods is an indication that the virus is circulating among the animals14. Good surveillance requires an excellent collaboration between the epidemiological surveillance teams and the wildlife surveillance teams. The involvement of national and local authorities in the management of epidemics in collaboration with the health teams, in particular on cases and contact cases‘ follow-up, was also essential and allowed the teams to overcome the difficulties surrounding the surveillance of these epidemics. Due to the introduction of multiple viruses during VHF epidemics in rural settings, collaboration with the staff responsible of the protection of wildlife is essential in order to detect these epidemics as early as possible, as well as monitoring the emergence of other diseases such as bird flu. It is crucial that the countries at high risk have a program to respond to VHFs, including on-going and permanent epidemiological surveillance as part of the integrated disease surveillance and response system. The medical teams should be on permanent alert and should pursue sensitization campaigns among the populations. As soon as the end of an epidemic is declared, the postepidemic and inter-epidemic surveillances should resume. VHF epidemics due to Ebola and Marburg viruses are more and more frequent in sub-Saharan Africa, and only an adapted epidemiological surveillance will allow timely detection and efficient response.

Angola, Dr. Diallo Fatoumata Binta, in Congo, Dr. Lamine cissé Sarr, in Gabon, Dr. Alain C. Brun, in DRC, Dr. Jean Baptiste Roungou, and in Uganda, Dr. George Melville for their support and encouragements. We would also like to thank our colleagues from the Ministries of Health in Angola, Congo, Gabon, DRC and Uganda with whom we have worked during the different viral hemorrhagic fever epidemics caused by Ebola and Marburg viruses. We would also like to thank Dr. Helen Perry for giving us the opportunity to finalize the manuscript of this article. Additionally we thank Dr. M.E. Penaranda, J. Kyle and M. Baker, volunteers of the Sustainable Sciences Institute, for proofreading this manuscript We also thank our colleagues from the Global Epidemic Alert and Response Network (GOARN), from the International Center of Medical Research in Franceville (Gabon), from the Control Disease Center (CDC) in Atlanta (USA), in Winnipeg (Canada), from the National Institute of communicable disease (NICD) in Johannesburg (South Africa), from the Institute Bernhard-Nocht in Hamburg (Germany), from the Tropical Medicine Institute in Antwerp (Belgium) and from the Forest Ecosystem of Central Africa (ECOFAC) Project References 1.

2.

3.

4.

5. 6. 7.

8. 9.

10.

11.

Contribution of the authors Y. Allarangar has initiated, conceived and developed the manuscript; P. Formenty, I. Sow and S. Duale equally participated in the conception and review of the manuscript; K.M. Lamine, W. Alemu, A. Yada, F. Libama, P. Boumandouki and C. Woodfill reviewed the manuscript. Acknowledgements We are very grateful to WHO representatives in

38

12. 13. 14. 15.

WHO/AFRO - Procédures Opérationnelles Standard pour le contrôle des Epidémies de Fièvres Hémorragiques à virus Ebola et Marburg, Brazzaville (Congo), mars 2007. Y, P. Formenty, F Libama, A. Epelboin,Y. Allarangar, E. Leroy, H. Moudzeo, P. Tarangonia, A. Molamou, M. Lenzi, K. Ait-Ikhlef, B. Hewlett, C. Roth, T Grein et l‘Equipe de lutte contre l‘épidemie d‘Ebola, au Congo. L‘épidemie de fièvre hémorragique à virus Ebola en République du Congo, 2003: Une nouvelle strategie ?, Med Trop2003; 63 : 291-295 Formenty P, Libama F, Epelboin A, Allarangar, E. Leroy, H. Moudzeo, et al L‘épidémie de fièvre hémorragique à virus Ebola en République du Congo, 2003: une nouvelle stratégie? Médecine Tropicale. 2003;63:291-295. Nkoghe D, Formenty P, Leroy E, Nnegue S., Obame edou S, Ibaba j., Allarangar Y., Cabore J., Bachy C., Andraghetti R., Benoist A. De, Galanis E., Rose A., Bausch D., Reynolds M., Rollin P., Choueibou C., Shongo rR, Gergonne B., Kone L, Yada A., Roth C., Toungmve M Plusieurs épidémies de fièvre hémorragiques due au virus Ebola au Gabon, d‘octobre 2001 a avril 2002. Bulletin de la Société de Pathologie Exotique. 2005;98(3):224229. WHO/AFRO. Guide technique de surveillance intégrée de la maladie et la riposte, 2001. Allarangar Y. Rapport de gestion de l‘épidémie de fièvre hémorragique à virus EBOLA au Gabon, Mars 2002. WHO/AFRO Okware, S, Omaswa F. ; Zaramba S. Opio A., Lutwama J, Kamugisha J, Rwaguma E. Kagwa P, Lamunu M : An Outbreak of Ebola in Uganda, Tropical Medicine & International Health. 7(12):1068-1075, December 2002. WHO. Fièvre hémorragique a virus Ebola au Zaïre. Relevé Epidémiologique Hebdomadaire.1995;70:241-248. Lloyd ES, Zaki SR, Rollin PE, Tshioko K, Bwaka MA, Ksiazek TG, et al. Long-Term Disease Surveillance in Bandundu Region, Democratic Republic of Congo: A model for Early Detection and Prevention of Ebola Hemorrhagic Fever. J Infect Dis. 1999;179 (suppl 1);S274-280. Jeffs B, Roddy P, Weatherill D, de la Rosa O, Dorion C, Iscla M, et al. The Médecins Sans Frontières Intervention in the Marburg Hemorrhagic Fever Epidemic, Uige, Angola, 2005. Lessons learned in the Hospital. J Infect Dis. 2007;196(2):S154-161. Guigui-Zoundi MT. Rapport de gestion de l‘épidémie de fièvre hémorragique à virus Marburg en Angola, mai 2005. WHO/AFRO Allarangar Y and Koné ML. Rapport de gestion de l‘épidémie de fièvre hémorragique à virus EBOLA au Congo, Juin 2004. WHO/AFRO WHO , Outbreak of Marburg Haemorrhagic Fever: Uganda, June–August 2007 Weekly Epidemiological Record, No. 43, 26 October 2007, 82nd Year , 381– 384 WHO. Flambée de fièvre hémorragique a virus Ebola au Congo et Gabon, octobre 2001-juillet 2002. Relevé Epidémiologique Hebdomadaire. 2003;78, 223-228.

Copyright of East African Journal of Public Health is the property of East African Public Health Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.