International Journal of Infectious Diseases 64 (2017) 80–84

Contents lists available at ScienceDirect

International Journal of Infectious Diseases

journal homepage: www.elsevier.com/locate/ijid

Knowledge and practices related to plague in an endemic area of

Uganda
Kiersten J. Kugelera,* , Titus Apangub , Joseph D. Forrestera , Kevin S. Griffitha ,
Gordian Candinib , Janet Abarub , Jimmy F. Okothb , Harriet Apiob , Geoffrey Ezamab ,
Robert Okellob , Meghan Bretta , Paul Meada
a
Bacterial Diseases Branch, Division of Vector-Borne Diseases, National Center for Emerging and Zoonotic Infectious Diseases, Centers for Disease Control and
Prevention, Fort Collins, CO, USA
b
Uganda Virus Research Institute, Plague Program, Arua and Entebbe, Uganda

A R T I C L E I N F O A B S T R A C T

Article history: Background: Plague is a virulent zoonosis reported most commonly from Sub-Saharan Africa. Early
Received 28 June 2017 treatment with antibiotics is important to prevent mortality. Understanding knowledge gaps and
Received in revised form 1 September 2017 common behaviors informs the development of educational efforts to reduce plague mortality.
Accepted 5 September 2017
Methods: A multi-stage cluster-sampled survey of 420 households was conducted in the plague-endemic
Corresponding Editor: Eskild Petersen, Aar-
hus, Denmark
West Nile region of Uganda to assess knowledge of symptoms and causes of plague and health care-
seeking practices.
Results: Most (84%) respondents were able to correctly describe plague symptoms; approximately 75%
Keywords:
Plague
linked plague with fleas and dead rats. Most respondents indicated that they would seek health care at a
Yersinia pestis clinic for possible plague; however plague-like symptoms were reportedly common, and in practice,
Knowledge persons sought care for those symptoms at a health clinic infrequently.
Practices Conclusions: Persons in the plague-endemic region of Uganda have a high level of understanding of
Behaviors plague, yet topics for targeted educational messages are apparent.
Africa Published by Elsevier Ltd on behalf of International Society for Infectious Diseases. This is an open access
article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction 1997; Prentice and Rahalison, 2007; Mead, 2015). A lack of early
and effective treatment increases the risk of death and likelihood
Plague is a highly virulent zoonosis caused by the bacillus of epidemic spread (Mead, 2015; Kugeler et al., 2015).
Yersinia pestis (Perry and Fetherston, 1997; Pollitzer, 1954; Prentice Sub-Saharan Africa, including the Democratic Republic of the
and Rahalison, 2007). When an epizootic occurs, high mortality Congo (DRC) and Uganda, accounts for most human cases of plague
among rodents leaves infectious fleas searching for alternative reported in recent decades (WHO, 2010; Kilonzo, 1999). In Uganda,
hosts. Most human infections result from an infectious flea bite plague is endemic in the northwestern West Nile region, bordered
and present as bubonic plague, characterized by fever and a by the Nile River to the East, South Sudan to the North, and DRC to
swollen, tender lymph node. The dissemination of Y. pestis through the West and South (WHO, 2010; Kilonzo, 1999; Orochi-Orach,
the bloodstream and into the lungs results in pneumonic plague, a 2002). Persons in Uganda often seek health care from both
fulminant pneumonia characterized by bloody sputum. Pneumon- allopathic medical practitioners and traditional healers or
ic plague is the only clinical form of the illness that can be directly community drug shops (Nuwaha and Muganzi, 2008). The
transmitted from person to person; patients near death can infect widespread use of traditional medicine and drug shops is
close contacts through respiratory droplets (Pollitzer, 1954; Kool, attributed to accessibility, affordability, and cultural beliefs (
2005). Although mortality in untreated plague infections ranges WHO, 2002; Barnes-Dean, 1986), but these practices may delay
between 50% and 100% depending on the clinical presentation, it seeking care and antibiotics from a health clinic, and can
can be treated successfully if appropriate antimicrobials are contribute to plague mortality.
administered within 72 h of symptom onset (Perry and Fetherston, General knowledge of plague symptoms and health care-
seeking practices in the West Nile region of Uganda has not been
quantified. These knowledge gaps hinder effective educational and
interventional efforts to reduce plague mortality in the region. A
* Corresponding author.
survey of household caregivers was performed in the two districts
E-mail address: [email protected] (K.J. Kugeler).

http://dx.doi.org/10.1016/j.ijid.2017.09.007
1201-9712/Published by Elsevier Ltd on behalf of International Society for Infectious Diseases. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).
 K.J. Kugeler et al. / International Journal of Infectious Diseases 64 (2017) 80–84 81

of the plague-endemic West Nile region to (1) determine the level Methods
of knowledge regarding the causes and symptoms of plague; (2)
assess the frequency of possible behavioral risk factors for human In June 2013, a cross-sectional, stratified, multi-stage cluster-
plague; (3) estimate the background frequency of symptoms sampled household survey was performed in West Nile, Uganda as
consistent with bubonic and pneumonic plague; and (4) part of an ongoing collaboration between the Uganda Virus
describe health care-seeking behavior of persons with plague-like Research Institute (UVRI) Plague Program and the US Centers for
symptoms. Disease Control and Prevention (CDC). This region is composed of
two districts (Arua and Zombo) with two predominant tribes, with

Figure 1. Location of households surveyed to assess knowledge and health care-seeking practices related to plague in Arua and Zombo districts, West Nile region, Uganda.
82 K.J. Kugeler et al. / International Journal of Infectious Diseases 64 (2017) 80–84

corresponding languages (Lugbara and Alur, respectively) and (91%) compared to Zombo (81%) (p = 0.018). Overall, 377 (87%)
customs. reported seeking care from drug shops; frequency of drug shop
First, all villages within Arua and Zombo districts as of 2009 usage was similar between the two districts. Fewer respondents
were stratified according to district and categorized according to (17%) indicated that household members utilized traditional
whether suspected or confirmed human plague cases had or had healers, but reported usage was higher in Arua District (28%)
not occurred among residents during 1999 to 2012. All villages than in Zombo District (12%) (p = 0.0003). A travel time of less than
were then classified according to parish (the next largest 2 h to a health clinic was reported by 283 (82%) households, to a
administrative unit). Sampling was restricted to parishes in which chemist/drug store by 328 (87%) households, and to a traditional
50% of villages had a history of plague. A random number healer by 65 (78%) households. Discrete factors why respondents
generator was used to select 30 eligible villages from each district, chose not to seek care from a health clinic were rarely reported.
for a total of 60 villages. One re-selection was necessary because Length of the wait, cost, the type of illness, and personal preference
the village initially was determined to be too inaccessible for were reported reasons for fewer than 3% of respondents.
survey administrators. The median number of census-enumerated
households per selected village was 72 (range 31–108). In a design Knowledge about plague
similar to that recommended by the World Health Organization for
immunization coverage surveys (Lemeshow and Robinson, 1985), Nearly all caregivers (n = 400; 95%) had heard of plague. Most
seven households in each of the 60 selected villages were surveyed, caregivers (n = 328; 84%) were able to describe plague as a febrile
for a total of 420 households. The corresponding sampling rate of illness with a painful swelling in the groin or armpit; 5% believed
villages in the restricted sampling area used for weighted the primary symptoms were headache and confusion, 8% did not
calculations was 37.5% for Arua District and 12% for Zombo District. identify any symptoms of plague, and fewer thought it to be a
Local teachers, community development officers, and health diarrheal illness (2%) or a febrile illness with rash (1%).
care providers trained by study personnel in interview techniques Most respondents also understood that plague was an illness
and fluent in one of the two primary languages of the region served transmitted by fleas (n = 330; 76%) and was associated with rats
as survey administrators. Upon arrival in selected villages, survey dying in large numbers (n = 314; 72%). Nevertheless, 19% indicated
administrators approached village leaders to explain the study and that people got plague from drinking bad water, and 12% believed
seek permission to conduct surveys of seven primary caregivers in that plague could be transmitted by mosquitoes. The frequency of
each village. The primary caregiver of the household was defined as these responses did not differ by district. Almost half of surveyed
the person who usually makes health care choices for members of caregivers (49%) reported that plague could be transmitted from
the household. Household members were defined as all persons touching a person who is sick with, or has died from, plague. This
who ate from the same cooking pot daily. To select households, the response was more common in Zombo district (57% vs. 28%;
survey administrator dropped a writing implement on the ground p = 0.001). Only four respondents indicated that they believed
in the center of the village and proceeded to the nearest house in plague to be the result of a curse specifically, yet 20 (6%) believed
the direction indicated by the implement. The nearest neighboring plague occurred when there was discord among village or family
households were subsequently selected until the total number of members. Roughly half of respondents (55%) regardless of district
households surveyed reached seven. Two attempts at reaching the were aware that the risk of plague was linked to the rainy season.
primary caregiver of a household were made; if unsuccessful, the Overall, friends and family were the most commonly reported
nearest neighboring household was approached. Following docu- source of information on plague (53%). The radio (37%), health
mented informed consent, a 41-question survey was administered clinics (32%), and village health workers (32%) were slightly less
to each caregiver in Lugbara or Alur. Responses that indicated a common sources of information, followed by drug shops (10%),
history of plague in the village were cross-referenced to historical posters (3%), and traditional healers (1%). Persons in Arua District
passive and active clinic surveillance records in the region from reported more varied sources of plague information than those in
1999 through June 2013. Zombo District. Radio was much more commonly reported as a
Hard copy survey forms were entered into Epi Info v.7 (CDC, source of plague information in Arua District (73% vs. 22%;
Atlanta, GA, USA). Survey data were analyzed using sample survey p < 0.0001), as were drug shops (30% vs. 2%; p < 0.0001) and
analytic procedures within SAS (SAS Institute, Cary, NC, USA). The traditional healers (4% vs. 1%; p = 0.008).
inverse of the sampling fraction for each household observation
within each village cluster was assigned as the sampling weight. Frequency of potential plague risk factors
Weighted percentages are presented unless otherwise indicated.
Rao–Scott Chi-square tests were used for comparison between In both districts, most respondents (n = 328; 80%) reported
respondents in Arua and Zombo districts, where appropriate. This seeing live rats in their village more than five times in the past
study was approved by the institutional review boards at UVRI, the month. People commonly reported taking measures to eliminate
Uganda National Council for Science and Technology, and CDC. rats: 79% used either chemicals or poison, 52% used traps, and 25%
reported ‘catching’ rats in other ways. Overall, 27% of respondents
Results indicated that they tried to store food in a way that rats cannot
access. Most respondents noticed dead rats infrequently; 10%
Study population and general health care-seeking practices indicated that they had seen dead rats one or two times during the
prior month.
Among the 420 respondents, 210 were from Vurra County in Most households reported someone sleeping on the floor;
northern Arua District and 210 were from Zombo County in Zombo however, this was more common in Zombo District (93%) than in
District (formerly Okoro County of Nebbi District) (Figure 1). The Arua District (83%) (p = 0.003). In 20% of households, everyone
median age of respondents was 40 years (range 12–82 years); 224 slept on the floor. Among those households where some but not all
(55%, unweighted) were male. family members slept on the floor, children most commonly did so
When asked about overall utilization of different health care (65%), and slightly more frequently in Zombo (71%) than in Arua
options during the prior year, most respondents (n = 359; 84%) District (52%) (p = 0.008). Women slept on the floor in 35% of
indicated that someone in their household had sought care from a households, also more frequently in Zombo (42%) than in Arua
health clinic; the percentage was slightly higher in Arua District (16%) (p < 0.0001). Men reportedly slept on the floor in only nine
 K.J. Kugeler et al. / International Journal of Infectious Diseases 64 (2017) 80–84 83

(2%) households. More than half of surveyed households (n = 269; Additionally, drug shop owners in Arua District have received
68%) reported that people slept in huts where food was stored and substantial health training and are well-connected with the
149 (38%) reported that people slept in huts where food was district health office, whereas their counterparts in Zombo have
prepared. less training and are solely business owners (T. Apangu, personal
communication). Lastly, traditional healers in Arua District are
Frequency of plague like-symptoms and associated health care- most often herbalists, whereas in Zombo District, they are often
seeking practices considered to invoke the supernatural (T. Apangu, personal
communication). Because visits to healers believed to have
Nearly all caregivers (n = 383; 94%) reported that they would go supernatural powers can carry some level of community stigma,
to a health clinic first if they had plague, and similarly 366 (84%) respondents in Zombo may have been less likely to report that they
caregivers reported that they would go to a health clinic first if they had visited traditional healers to the survey administrators. The
were sick with fever and painful swelling in their groin area. There identification of cultural or geographic differences in care-seeking
were no significant differences by district. A similar proportion of behavior can inform options for the future dissemination of
respondents indicated that they would visit a health clinic first for educational messages.
other illnesses: headache and confusion, fever and rash, and Despite a high level of understanding of plague in these
diarrhea and vomiting. Approximately 13% reported that they communities, opportunities for targeted educational messages are
would seek care from a community drug shop first for these apparent. First, improving the recognition of possible plague
symptoms. among alternative health providers such as drug shop owners and
When asked about the frequency of plague-like symptoms, 29 traditional healers could serve as a means of secondary prevention.
(8%) reported that someone in their household had been sick with Even if the patients are initially seen outside of a health clinic, rapid
fever and a swelling in their groin, neck, or armpit in the prior year; recognition and referral to a health clinic could improve outcomes.
14 (5%) reported that someone in their household had been sick Second, although overall knowledge of plague was quite high,
with fever and had coughed up blood during the same time period. confusion about the causes of plague (bad drinking water,
Of the 29 respondents who indicated fever and lymph node community discord) remain obstacles to prompt recognition and
swelling in the prior year, 20 (69%, unweighted) reportedly sought treatment (Barnes-Dean, 1986; Centers for Disease Control and
care in a health clinic, but only 12 (41%, unweighted) sought care at Prevention, 2009). Village-specific education by village health
a health clinic first. Similarly, among the 14 respondents who volunteers may foster increased recognition of possible plague
indicated that someone had been sick with a bloody cough, most when dead rats are found or when persons fall ill with plague-like
(n = 11; unweighted 79%) sought care from a health clinic for symptoms. Sleeping on the floor and sleeping in huts where food is
treatment, but only one (7%, unweighted) of these patients went to stored or cooked may increase the risk of plague. Increased
the health clinic first. education regarding plague risk in villages with a history of plague
may additionally motivate behavioral measures to mitigate human
Discussion plague risk. Additionally, this survey revealed that members of
these communities are motivated to keep rats out of their
Based on survey responses, caregivers in the plague-endemic dwellings, and spend some of their minimal available resources
West Nile region of Uganda have a strong understanding of the to purchase poison or traps. This suggests that an available cost-
clinical features and epidemiology of plague. Most understood effective prevention method to reduce both rats and fleas would be
plague as a disease characterized by fever and lymph node used. Lastly, this survey demonstrated that in the event of a plague
swellings associated with fleas and rats requiring treatment at a epidemic in the region, a ready-made message for radio distribu-
health clinic. Nevertheless, there appears to be some discrepancy tion may be the most efficient way of public education. Emphasis of
between caregivers’ theoretical care-seeking practices and what those messages should not be on plague symptoms, as there is
they reported doing when household members were recently sick already a high level of associated understanding in the population,
with plague-like symptoms. Reasons for this discrepancy should be but on the importance of seeking care early from the health clinic
pursued further, as they will affect the success of plague in order to save lives.
educational messages in the region. Given the relatively high There were several limitations to this study. First, because of
frequency of fever and lymph node swelling in the population, as vague wording that could have been misinterpreted, it was not
revealed by this survey, it is possible that caregivers are able to possible to adequately assess the participant’s understanding of
distinguish the severity of illness due to plague from other less the transmissibility of pneumonic plague. Further surveys with
virulent etiologies. Messages that urge immediate assessment at more discrete wording easily translatable into both languages
the health clinic for fever and swelling may require modification to should be conducted, as understanding practices associated with
increase the specificity for plague. Regardless, village health pneumonic plague is an important component of reducing
volunteers should be reminded of the importance of plague epidemic spread and decreasing plague mortality. Second,
recognition and rapid referral to reduce the potential for future restricting sampling to parishes where at least 50% of the villages
plague outbreaks. had a history of plague limits the generalizability of findings to
Plague occurred in what is now Zombo District for several other parts of Arua and Zombo districts where plague is less
decades before spreading north into Arua District in the early common (MacMillan et al., 2011; MacMillan et al., 2012). Third,
1990s (Orochi-Orach, 2002). While overall knowledge of plague recall bias could have influenced the reporting of various practices,
was similar in both districts, several differences became apparent, and observer bias could have influenced respondents to answer
specifically sources of plague knowledge and the reported use of with what they perceived to be the ‘correct’ answer, rather than the
traditional healers. Persons in Arua District reported more varied most truthful one reflecting their actual practice. This latter bias
sources of plague information, including radio and drug shops. may potentially have contributed to the observed discrepancy
Several radio stations in Arua include daily health messages run between theoretical health care-seeking with plague versus more
during every major news hour. In contrast, Zombo has only one accurate reporting of what people did when they or a family
primary radio station with limited geographic coverage and fewer member had fever and a swelling. Finally, as mobile phones and
health message broadcasts, which focus primarily on malaria and Internet access become more common, these could be sources of
diseases other than plague (T. Apangu, personal communication). health-related information even in rural Sub-Saharan Africa;
84 K.J. Kugeler et al. / International Journal of Infectious Diseases 64 (2017) 80–84

however, the Internet was not assessed as a source of information Nene Terence for their assistance in the preparation and execution
in this survey. of the survey.
As with many other epidemic-prone infectious diseases, plague
is relatively rare and highly episodic. Even in plague-endemic References
regions of Uganda, the illness is sometimes absent for several
years. This unpredictable pattern likely contributes to delays in Barnes-Dean V. Lugbara illness beliefs and social change. Africa 1986;56:334–51.
Centers for Disease Control and Prevention. Bubonic and pneumonic plague —
recognition and poses a challenge to public health authorities Uganda, 2006. MMWR 2009;58:778–81.
seeking to anticipate and control outbreaks. The framing and Kilonzo BS. Plague epidemiology and control in eastern and southern Africa during
dissemination of plague educational messages is further chal- the period 1978 to 1997. Cent Afr J Med 1999;45:70–6.
Kool JL. Risk of person-to-person transmission of pneumonic plague. Clin Infect Dis
lenged by varying cultural nuance and limited communications 2005;40:1166–72.
infrastructure. Knowledge of how people living in West Nile Kugeler KJ, Staples JE, Hinckley AF, Gage KL, Mead PS. Epidemiology of human
understand plague and where opportunities exist for improving plague in the United States, 1900–2012. Emerg Infect Dis 2015;21:16–22.
Lemeshow S, Robinson D. Surveys to measure programme coverage and impact: a
knowledge and modifying current practice should be used to
review of the methodology used by the expanded programme on immuniza-
inform appropriate public health interventions. tion. World Health Stat Q 1985;38:65–75.
MacMillan K, Enscore RE, Ogen-Odoi A, Borchert JN, Babi N, Amatre G, et al.
Landscape and residential variables associated with plague-endemic villages in
Disclaimer
the West Nile region of Uganda. Am J Trop Med Hyg 2011;84:435–42.
MacMillan K, Monaghan AJ, Apangu T, Griffith KS, Mead PS, Acayo S, et al. Climate
The findings and conclusions in this report are those of the predictors of the spatial distribution of human plague cases in the West Nile
authors and do not necessarily represent the official position of the region of Uganda. Am J Trop Med Hyg 2012;86:514–23.
Mead PS. Yersinia species (Including Plague). In: Bennett JE, Dolin R, Blaser MJ,
Centers for Disease Control and Prevention. editors. Principles and practice of infectious diseases. 8th ed. Philadelphia:
Elsevier; 2015. p. 2607–18.
Funding Nuwaha F, Muganzi E. Predictors of use of traditional medicine by patients with
sexually transmitted infections in southwest Uganda. J Altern Complement Med
2008;14:733–9.
This work was supported by a US CDC Cooperative Agreement Orochi-Orach S. Plague outbreaks the gender and age perspective in Okoro County,
with the Uganda Virus Research Institute. Nebbi District, Uganda. Uganda: Uganda Agency for Accelerated Regional
Development; 2002.
Perry RD, Fetherston JD. Yersinia pestis–etiologic agent of plague. Clin Microbiol Rev
Conflict of interest 1997;10:35–66.
Pollitzer R. Plague. Geneva, Switzerland: World Health Organization; 1954.
Prentice MB, Rahalison L. Plague. Lancet 2007;369:1196–207.
None.
WHO. Human plague: review of regional morbidity and mortality, 2004–2009.
Wkly Epidemiol Rec 2010;6:40–5.
Acknowledgements World Health Organization. Traditional medicine strategy 2002-2005. Geneva:
World Health Organization; 2002.

We thank Aimee Geissler, Maryekmungu Immaculate, Ayakaka

Charity, Alionzi Christopher, Odama Bosco, Pithua Ephraim, and