Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

The Burden of Common Infectious Disease Syndromes at the Clinic and Household Level from Population-Based Surveillance in Rural and Urban Kenya

  • Daniel R. Feikin ,

    Affiliations International Emerging Infections Program-Kenya, Centers for Disease Control and Prevention-Nairobi and Kisumu, Nairobi and Kisumu, Kenya, Kenya Medical Research Institute/CDC Public Health and Research Collaboration, Kisumu and Nairobi, Kenya

  • Beatrice Olack,

    Affiliations International Emerging Infections Program-Kenya, Centers for Disease Control and Prevention-Nairobi and Kisumu, Nairobi and Kisumu, Kenya, Kenya Medical Research Institute/CDC Public Health and Research Collaboration, Kisumu and Nairobi, Kenya

  • Godfrey M. Bigogo,

    Affiliations International Emerging Infections Program-Kenya, Centers for Disease Control and Prevention-Nairobi and Kisumu, Nairobi and Kisumu, Kenya, Kenya Medical Research Institute/CDC Public Health and Research Collaboration, Kisumu and Nairobi, Kenya

  • Allan Audi,

    Affiliations International Emerging Infections Program-Kenya, Centers for Disease Control and Prevention-Nairobi and Kisumu, Nairobi and Kisumu, Kenya, Kenya Medical Research Institute/CDC Public Health and Research Collaboration, Kisumu and Nairobi, Kenya

  • Leonard Cosmas,

    Affiliations International Emerging Infections Program-Kenya, Centers for Disease Control and Prevention-Nairobi and Kisumu, Nairobi and Kisumu, Kenya, Kenya Medical Research Institute/CDC Public Health and Research Collaboration, Kisumu and Nairobi, Kenya

  • Barrack Aura,

    Affiliations International Emerging Infections Program-Kenya, Centers for Disease Control and Prevention-Nairobi and Kisumu, Nairobi and Kisumu, Kenya, Kenya Medical Research Institute/CDC Public Health and Research Collaboration, Kisumu and Nairobi, Kenya

  • Heather Burke,

    Affiliations International Emerging Infections Program-Kenya, Centers for Disease Control and Prevention-Nairobi and Kisumu, Nairobi and Kisumu, Kenya, Kenya Medical Research Institute/CDC Public Health and Research Collaboration, Kisumu and Nairobi, Kenya

  • M. Kariuki Njenga,

    Affiliations International Emerging Infections Program-Kenya, Centers for Disease Control and Prevention-Nairobi and Kisumu, Nairobi and Kisumu, Kenya, Kenya Medical Research Institute/CDC Public Health and Research Collaboration, Kisumu and Nairobi, Kenya

  • John Williamson,

    Affiliation Kenya Medical Research Institute/CDC Public Health and Research Collaboration, Kisumu and Nairobi, Kenya

  • Robert F. Breiman

    Affiliations International Emerging Infections Program-Kenya, Centers for Disease Control and Prevention-Nairobi and Kisumu, Nairobi and Kisumu, Kenya, Kenya Medical Research Institute/CDC Public Health and Research Collaboration, Kisumu and Nairobi, Kenya



Characterizing infectious disease burden in Africa is important for prioritizing and targeting limited resources for curative and preventive services and monitoring the impact of interventions.


From June 1, 2006 to May 31, 2008, we estimated rates of acute lower respiratory tract illness (ALRI), diarrhea and acute febrile illness (AFI) among >50,000 persons participating in population-based surveillance in impoverished, rural western Kenya (Asembo) and an informal settlement in Nairobi, Kenya (Kibera). Field workers visited households every two weeks, collecting recent illness information and performing limited exams. Participants could access free high-quality care in a designated referral clinic in each site. Incidence and longitudinal prevalence were calculated and compared using Poisson regression.


Incidence rates resulting in clinic visitation were the following: ALRI — 0.36 and 0.51 episodes per year for children <5 years and 0.067 and 0.026 for persons ≥5 years in Asembo and Kibera, respectively; diarrhea — 0.40 and 0.71 episodes per year for children <5 years and 0.09 and 0.062 for persons ≥5 years in Asembo and Kibera, respectively; AFI — 0.17 and 0.09 episodes per year for children <5 years and 0.03 and 0.015 for persons ≥5 years in Asembo and Kibera, respectively. Annually, based on household visits, children <5 years in Asembo and Kibera had 60 and 27 cough days, 10 and 8 diarrhea days, and 37 and 11 fever days, respectively. Household-based rates were higher than clinic rates for diarrhea and AFI, this difference being several-fold greater in the rural than urban site.


Individuals in poor Kenyan communities still suffer from a high burden of infectious diseases, which likely hampers their development. Urban slum and rural disease incidence and clinic utilization are sufficiently disparate in Africa to warrant data from both settings for estimating burden and focusing interventions.


Most childhood deaths today occur in Sub-Saharan Africa, and over three-quarters of these deaths are due to infectious diseases [1]. Adults in Africa, who bear a disproportionate share of the world's HIV/AIDS epidemic, also have the highest mortality rates in the world [2]. Characterizing burden of infectious disease syndromes in Africa is important for targeting and prioritizing use of limited resources for optimal curative and preventive services and for research and development of novel strategies and interventions. Moreover, longitudinal surveillance of disease burden is important in monitoring impact of these public health expenditures. In Africa, the majority of cases of acute infectious illnesses do not present at health facilities [3], [4], and most deaths occur at home [5]. Community-based, household surveillance can supplement facility-based surveillance to define disease burden and measure impact of public health interventions. Nonetheless, few community-based studies of disease burden in Africa exist, and most that have been done took place before the widespread introduction of interventions that substantially reduced disease burden, such as insecticide-treated bednets, Haemophilus influenzae type b vaccine, improvements in access to safe drinking water, and wider use of cotrimoxazole and anti-retroviral medications in HIV-infected persons [6][9].

We describe comprehensive findings from two years of longitudinal, population-based surveillance at field clinics and specific households in geographically defined locations in rural western Kenya and an urban informal settlement in Nairobi for respiratory illness, diarrhea and acute febrile illness.


Ethical review

Written informed consent was obtained for data collection at the clinics and households. For minors, written informed consent was obtained from their parent or guardian. The protocol and consent forms were reviewed and approved by the Ethical Review Boards of the Kenya Medical Research Institute (KEMRI) (# 932) and the Institutional Review Board of the U.S. Centers for Disease Control and Prevention (CDC) (# 4566).

Surveillance sites

The Kenyan International Emerging Infections Program of KEMRI/CDC has conducted population-based infectious disease surveillance (PBIDS) since late 2005 in two sites in Kenya: Asembo, a rural location in Nyanza Province in western Kenya along Lake Victoria and within the Kibera informal settlement in Nairobi (Table 1)[5], [10][13].

Table 1. Characteristics of study sites. Asembo, western Kenya and Kibera, Nairobi.

Household surveillance

PBIDS household surveillance methods have been described previously [10]. Briefly, all households in both surveillance sites are offered enrollment. Eligible households are those located within 5 km and 1 km from the designated referral clinics for the project in Asembo and Kibera, respectively. Eligible persons must have resided permanently in these areas for 4 calendar months or be a child born to a woman enrolled in PBIDS. Enrollment was continuous in both sites since the project's beginning. Community interviewers visit enrolled households every two weeks (“fortnightly ” visits.) Participants are asked standardized questions, in local language, about recent illnesses. For certain key symptoms—cough, fever and diarrhea—the exact days of occurrence are recorded. For older children (approximately over 12 years old) and adults, interviews of that person are done. If not at home or unable to answer questions, a proxy who is knowledgeable about the participant's health is interviewed. For children unable to answer for themselves, the mother or other primary caretaker of the child is interviewed. Abbreviated physical exams are carried out on ill persons present during the visit, including axillary temperature, 1 minute respiratory rate, evaluation for lower chest wall indrawing and stridor in ill children, and observation for signs of dehydration. Community interviewers are secondary school graduates, who undergo extensive training on data collection and physical examination led by KEMRI/CDC clinicians, including WHO Integrated Management of Childhood Illness (IMCI) training videos [14].

Clinic surveillance

Centrally located clinics were identified and enhanced in each surveillance area —in Asembo, St. Elizabeth Lwak Mission Hospital, operated by the Franciscan Sisters of St. Anna (henceforth referred to as “Lwak”), and in Kibera, Tabitha Clinic, operated by Carolina for Kibera (henceforth referred to as “Tabitha”)]. Enrolled participants receive free medical care at the study clinics for all acute, potentially infectious disease conditions, including syndromes not targeted for surveillance, by KEMRI/CDC-trained staff. Vital signs and symptoms are recorded by nurses or trained health facility recorders (secondary school graduates), and are verified by clinical staff. All temperatures are axillary and respiratory rates are measured for 1 minute using an audible countdown timer. Oxygen saturations are measured using fingertip pulse oximeters (NONIN Medical, Minnesota). Most examinations at both sites are carried out by clinical officers (similar to physician's assistants); two physicians at Tabitha Clinic also examined patients. Malaria blood smears are performed on all patients with history of fever or documented temperature ≥38.0°C. No x-rays were taken during the period described. Blood and stool for culture and combined nasopharyngeal and oropharyngeal swabs for PCR are collected from patients meeting clinical indications (data not presented). Lwak hospital has 40 inpatient beds and manages most hospitalizations, only referring complicated or surgical patients to the district hospital. Tabitha Clinic does not have inpatient capacity, and refers patients for hospitalization to a nearby district hospital.

Structured questionnaires detailing the current illness are completed for all sick visits at Lwak and Tabitha. At Lwak, scannable paper forms are used (TeleForm® software, Cardiff™, Vista, CA). Inconsistent or illogical data are returned to the clinic for correction. At Tabitha, data are entered directly into a custom-designed computerized system with internal quality control checks.

Data analysis

Case definitions for infectious disease syndromes in the clinics are provided in Table 2. For clinic visits, incidence rates were calculated as the number of syndrome-specific visits per person-year of observation. Revisits for the same illness episode were not counted as separate episodes. For denominator calculations, permanent residence status on enrolled participants in the surveillance area was used to determine start and stop dates for person-time contribution. Adjusted rates of clinic visitation were calculated for each syndrome accounting for the percentage of all clinic visits made for each syndrome that went to the study clinics (i.e. Lwak, Tabitha) as opposed to other area clinics, as determined from the household visits. Calculating case-fatality proportions in the 30 days following a clinic visit was possible by longitudinal follow-up of all participants at the households.

Table 2. Case definitions for major infectious disease syndromes from household and clinic surveillance in Asembo, western Kenya (Lwak Hospital) and Kibera, Nairobi (Tabitha Clinic), June 1, 2006 – May 31, 2008.

For household visits, slightly modified case-definitions were used (Table 2). Because examination findings were part of the household definitions for ALRI, adjustments to the household ALRI rate were made for persons with cough or difficulty breathing who did not have an examination, including temperature, carried out at the household visit, assuming the same percentage would have ALRI as those who had cough or difficulty breathing and an exam done. Adjustments were made separately for each age group. Reported, rather than documented, fever was used to define AFI for household visits, as reported fever is commonly used as a diagnostic criterion for malaria in endemic areas and due to widespread use of antipyretics [14].

Although we asked about illness in the two weeks preceding home visit, we previously showed a significant decay of the recall of symptoms over that two week period, slightly impacted by age [10]. Therefore, to calculate the most accurate rates from the household visits avoiding decay of recall with time, we only used symptoms reported on the day of visit and the 3 previous days (days 0–3) for children <5 years and the day of visit and the 4 previous days (days 0–4) for persons ≥5 years, and adjusted the person-days calculations accordingly. Longitudinal prevalence rates were calculated as total number of days of a syndrome over total number of days of eligible observation, converted to person-years of observation. For calculation of longitudinal prevalence, the number of days of cough or difficulty breathing (ARI), diarrhea, and fever (AFI) in persons meeting the case definitions was used. Only days in which there was information as to whether the symptom was present or absent were included in denominators [10]. Prevalence rates for ALRI were not calculated because only cough, one component of the ALRI definition, was available for all days under surveillance. Incidence rates from household visits were calculated as the number of new episodes per person-year. We used a symptom-free interval for diarrhea of 3 days and for cough and fever of 7 days to define new episodes of the same syndrome [15], [16]. Denominators were calculated as for longitudinal prevalence, but, in addition, only days in which a new episode could be counted were included; therefore, subsequent days of an episode after the first day and the days during the symptom-free interval were excluded from the denominator [17].

For data collected from June 1, 2006 to May 31, 2008, we modeled incidence and longitudinal prevalence of illnesses at the household visits, and compared rural and urban rates, using Poisson regression (PROC GENMOD, SAS version 9.1, SAS Institute, Cary, North Carolina, USA). We controlled for clustering of symptoms at the household level using generalizing estimating equations (GEE). For the clinic visits, incidence rate ratios between rural and urban sites were calculated using Breslow's test for 95% confidence intervals (PEPI, Sagebrush publishers). For comparisons of adjusted rural and urban rates, the delta method was used [18].

We used incidence rates to project the number of cases of ALRI, diarrhea, and AFI in Nyanza Province and among the urban slum population in Nairobi, populations with similar epidemiology to our surveillance populations. We used 2007 population projections based on the 1999 national census, applying the same percentages of the national population that was residing in Nyanza and Nairobi Provinces, and assumed 60% of Nairobi residents were residents of informal settlements [19][21]. Because our surveillance sites were likely not representative of other provinces of Kenya in terms of certain risk factors (e.g., HIV and malaria prevalence), we did not make national projections during this analysis.


General morbidity

There were 20,830 visits to the referral clinic in Asembo and 38,857 visits to Kibera during the study period, yielding overall rates of clinic visitation of 0.43 per person-year and 0.69 visits per person-year, respectively [Rate Ratio (RR) - 0.63, 95% CI 0.62–0.64]. Clinic visit rates were highest in children <5 years old (0.77 visits/person-year in Asembo, 2.4 visits/person-year in Kibera.) Clinic visit rates were higher for all age groups <35 years in Kibera than Asembo, similar in persons 35–49 years and higher in Asembo for persons ≥50 years.

During the study period, 1,181,788 interviews of household members (or proxies) were completed in Asembo and 710,997 in Kibera. For children <5 years, illness episodes were reported during the last two weeks in 40.3% of household visits in Asembo compared with 13.7% in Kibera (RR = 2.9, 95% CI 2.9–3.0). For persons ≥5 years, illness episodes were reported during the last two weeks in 28% of household visits in Asembo compared with 5.1% in Kibera (RR = 5.4, 95% CI 5.4–5.5).

Acute Respiratory Infections

ARI was the most common syndrome in both sites (Table 3). ARI rates were highest in young children, with cough or difficulty breathing being reported for almost two full months per year in Asembo and one month per year in Kibera for children <5 years old. ARI incidence was 5.6 times higher at the household visit than at the clinic for children in Asembo and 5.1 times higher for adults. In contrast, in Kibera the incidence detected for ARI in the household was 0.89 times and 0.71 times the incidence in the clinic for children and adults, respectively.

Table 3. Rates of Acute Respiratory Illness by age group from household and clinic surveillance in Asembo, western Kenya (Lwak Hospital) and Kibera, Nairobi (Tabitha Clinic), June 1, 2006 – May 31, 2008.

Acute lower respiratory tract infections.

The rate of ALRI resulting in clinic visitation was lower in Asembo than Kibera among children <5 years old (Table 4). Among children <5 years old, 0.41% and 0.89% of ALRI episodes in the clinic occurred in neonates in Asembo and Kibera, respectively. Persons ≥5 years had much lower rates of ALRI than children in both Asembo and Kibera. In contrast to all other syndromes, ALRI rates were higher in the clinic than in the household. Among children with ALRI seen in the clinic, 8.5% in Asembo and 2.0% in Kibera had auscultatory wheezing reported on exam at the clinic. Among children with ALRI in the clinic, 56% in Asembo and 42% in Kibera met IMCI criteria for severe or very severe pneumonia (Table 4). In Asembo, 3% of children with ALRI had oxygen saturation <90% and only 1% had oxygen saturation <90% without other criteria for IMCI pneumonia of any severity; 7.5% of adults with ALRI had oxygen saturation <90%. In Kibera, 11% of children with ALRI had oxygen saturation <90% but none had oxygen saturation <90% without other criteria for IMCI pneumonia of any severity; 23% of adults with ALRI had oxygen saturation <90%. At Lwak Hospital, 64% of children and 34% of persons ≥5 years with ALRI were hospitalized.

Table 4. Rates of Acute Lower Respiratory Illness by age group from household and clinic surveillance in Asembo, western Kenya (Lwak Hospital) and Kibera, Nairobi (Tabitha Clinic), June 1, 2006 – May 31, 2008.


Rates for diarrhea resulting in clinic visit were lower in Asembo than Kibera among children <5 years (Table 5). Among children <5 years old, 0.42% and 0.37% of diarrhea episodes in the clinic occurred in neonates in Asembo and Kibera, respectively. On average, children <5 years had 9.5 and 8.1 diarrhea days per year in Asembo and Kibera, respectively. Persons ≥5 years had much lower rates of diarrhea than children in both Asembo and Kibera. Numbers of diarrhea episodes as reported in the household were 6-fold and 2-fold higher than those seen in the referral clinics in Asembo and Kibera, respectively. Among children with diarrhea visiting the referral clinics, 40% in Asembo and 11% in Kibera had a sign of dehydration or an IMCI danger sign. At Lwak, 26% of children and 20% of persons ≥5 years with diarrhea were hospitalized.

Table 5. Rates of diarrhea by age group from household and clinic surveillance in Asembo, western Kenya (Lwak Hospital) and Kibera, Nairobi (Tabitha Clinic), June 1, 2006 – May 31, 2008.

Acute Febrile Illness.

In contrast to the other 3 syndromes, clinic visitation rates for AFI were higher in Asembo than Kibera among children <5 years, likely due to greater malaria prevalence in rural western Kenya (Table 6). Among children <5 years old, 0.34% and 0.42% of AFI episodes in the clinic occurred in neonates in Asembo and Kibera, respectively. On average, children <5 years had 37 and 11 fever days per year in Asembo and Kibera, respectively. Persons ≥5 years had much lower rates of AFI than children in both Asembo and Kibera. Among children with AFI in the clinic who had a blood smear done, 68% in Asembo and 23% in Kibera had malaria parasites seen. At Lwak Hospital, 40% of children and 28% of persons ≥5 years with AFI were hospitalized.

Table 6. Rates of acute febrile illness (AFI) by age group from household and clinic surveillance in Asembo, western Kenya (Lwak Hospital) and Kibera, Nairobi (Tabitha Clinic), June 1, 2006 – May 31, 2008.


For illness episodes presenting at the clinic, the case-fatality ratio was higher for all illnesses for children and adults in Asembo than Kibera (Table 7). Overall, mortality rates were also higher for all groups in Asembo than Kibera, particularly at the extremes of age (Figure 1)

Figure 1. Overall mortality rates by age for Asembo and Kibera in areas of population-based surveillance.

Table 7. Case-fatality ratio (# deaths/# cases) by syndrome, within 30 days of clinic visit date.

Burden projections.

The projected number of episodes of illness in 2007 was substantial in both Nyanza province (population of 4.3 million in 1999) and the informal settlements of Nairobi, ranging up to almost 27 million episodes of AFI in Nyanza Province (Table 8).

Table 8. Number of episodes of ALRI, diarrhea and AFI estimated to have occurred in Nyanza Province and informal Nairobi settlements in 2007 based on household (HH)- and clinic-based surveillance, Kenya.


Our longitudinal, population-based surveillance provides one of the few recent, comprehensive estimates of infectious disease burden in Africa, being unique in measuring burden for multiple infectious disease syndromes using common methodology in both rural and urban settings among children and adults, and in surveying at both clinics and in the household. Household visits served two purposes. First, they were used to encourage ill participants to access the referral clinic. Second, household visits were useful in interpreting and adjusting incidence rates. Shops and drug-sellers are more common sources of healthcare than licensed clinics in both areas [22], [23]. Multiple factors influence healthcare utilization at clinics, including cost, distance, quality of care, and severity of syndrome [24][27]. Even when offering free high-quality care as part of surveillance, many ill persons did not access our designated referral clinics, particularly in the rural area, where only 30–40% of ill participants sought care at any clinic and only a half of these at the project's designated referral clinic, percentages that did not vary substantially during the surveillance period [28]. Rates of ARI, diarrhea and AFI, but not ALRI, were higher when calculated from household visits than from clinic data. The higher rates of ALRI in the clinic than the household is likely due to differences in the case definitions used in the clinic and household, as well as a higher percentage of persons with ALRI visiting clinics due to the greater severity of illness than the other syndromes. The smaller difference between clinic and household rates of disease in Kibera than Asembo is because healthcare utilization overall, and particularly to the designated referral clinic, was higher in the urban site [22], [23], [28]. Distance of residence to clinic clearly influences health utilization [26]; those distances are much less in Kibera. In addition it is possible that people living in Kibera are more educated and comfortable with use of licensed clinic facilities than residents in Asembo, a remote rural area. Nonetheless, our findings suggest that a substantial burden of infectious disease illness does not present to the clinic, especially in rural Africa, highlighting the need for preventive measures (e.g. vaccines, insecticide-treated bednets) and non-clinic-based case management strategies in such settings [29][34].

Community-based studies of clinically-defined ALRI in African children have yielded varied incidences, with our estimates (0.29 per person-year rural, 0.19 per person-year urban) falling in the middle of the range when compared with peri-urban Nigeria (1.3 episodes/person-year), rural Gambia (0.45 episodes/person-year), urban Nigeria (0.08–0.22 episodes/person-year), rural central Kenya (0.21 episodes/person-year), and rural Ghana (0.06 episodes/person-year) [8], [9], [35][38]. The reason for higher rates of wheezing in children in the rural area when compared with urban in our study might be related to greater exposure to indoor air pollution from cooking with firewood in the rural area (Table 1), but could also be related to the frequency of certain viral infections (e.g. RSV) [39]. For ARI in children, the incidence we found (7.9 episodes/person-year rural, 3.5 episodes/person-year urban) is similar to rates from other community-based African studies – rural Kenya, 6.6 episodes/person-year, urban Nigeria 8.1 episodes/person-year [9], [37], [38].

While facility-based data on respiratory illness in adults exist, there have been no community-based data [40][45]. The facility-based rates of ALRI we estimate of 67 and 26 per 1,000 person-years in rural and urban Kenya, respectively, are on the higher end of estimates found in other studies in Africa, being more comparable with rates in HIV-infected populations (37–69 per 1000) [40], [43], [45]. Among HIV-negative or HIV-status unknown adult populations in Africa, ALRI rates have ranged from 2–10 per 1,000 [40], [41], [44], [46], [47]. There are several reasons likely responsible for the somewhat higher rate of adult ALRI we found. First, both our populations have high rates of adult HIV infection (15–18%, KEMRI/CDC data.) Second, we used a clinical ALRI definition that might have captured upper respiratory and influenza-like illness, in addition to ALRI [48]. Third, we collected longitudinal data in relatively small populations, and provided free, high-quality healthcare, which could have led to higher levels of health facility utilization than in other African studies.

The incidence rate of diarrhea we found in children of 1.9–2.2 episodes per child-year is lower than other estimates in Africa from the 1980s [7]. In a study from urban Nairobi, among children <3 years, 3.5 diarrheal episodes per year were recorded, whereas in Guinea-Bissau and Zaire 10.4 and 6.3 episodes per year, respectively, were recorded in children <5 years [7], [49][51]. It is possible that in the two decades since most earlier studies were undertaken that diarrhea disease rates have decreased in Africa due to interventions such as safe-water promotion, better hygiene, increased measles vaccination coverage, and reduced malnutrition [6]. Alternatively, differences in methodology, such as frequency of visits, or cultural concepts about what constitutes “diarrhea” could have contributed. Few community-based studies estimate diarrhea rates in African adults. In rural Uganda during weekly household visits, HIV-positive persons ≥5 years not on cotrimoxazole prophylaxis had 194 diarrheal episodes per 100 person-years, while HIV-negative persons ≥10 years living in the same households experienced 15.5 episodes per 100 person-years [52], [53]. We found diarrhea rates in adults (rural, 55 per 100 person-years; urban, 13 per 100 person-years) closer to the HIV-negative Ugandan population.

Rates of AFI are not usually calculated in African settings, despite fever usually being the most common presenting symptom [54][57]. More often, rates of specific etiologies of febrile illness are reported. Invasive bacterial disease rates among children range from 505 to 1,100 episodes per 100,000 person-years in various African settings [35], [58], [59]. These were estimates from hospital-based studies, and likely underestimate the true rate. A majority (up to 80%) of malaria illness never gets diagnosed in the community because of low levels of healthcare seeking and lack of laboratory testing [60]. Estimates of malaria rates in children are obtained from control groups during recent studies of candidate malaria vaccines, where symptomatic malaria incidence ranged from 17–26 per 100 person-years [29], [33]. We found clinic rates of AFI among children that were lower than the symptomatic malaria incidence rates from these trials, but higher than the invasive bacterial disease rates. At the household visits, we found AFI incidence was several-fold higher, with children experiencing 10 febrile episodes a year in the rural site and 3 in the urban site.

Among adult African populations, even less incidence data on AFI are available. From control arms of clinical trials in malaria-endemic areas of western Kenya, incidence rates of symptomatic and asymptomatic malaria parasitemia were 1–2 episodes per person-year [61], [62]. One review estimated the typhoid rate to be 39/100,000 in East Africa [54]. The incidence of AFI defined in the household for persons ≥5 years in our surveillance was several-fold higher than either these malaria or typhoid rates, suggesting that these diseases, particularly malaria, may be under-diagnosed and, perhaps more importantly, broader diagnostic testing is needed to elucidate the multiple causes of fever in older children and adults in Africa.

The mortality pattern in our surveillance areas, with highest rates among young children then rising again in young adults, is typical of parts of Africa with high levels of HIV [1], [2]. Mortality and case-fatality ratios following clinic visits were higher in Asembo than Kibera. There are several potential reasons for this difference. First, Tabitha clinic sees more patients than Lwak Hospital, likely reflecting greater healthcare use for milder illness [26]. Because of this, illness episodes might be treated at an earlier stage preventing progression to severe disease, which in turn would result in lower overall mortality. Second, more challenging access to quality health care (because of distance and poor roads, especially during rainy seasons) might have resulted in later presentation of illness in the rural setting. Third, some diseases, particularly malaria, are more prevalent in the rural area. Lastly, there might have been surveillance bias. Many residents of Kibera consider their ancestral home to be elsewhere in Kenya, and it is culturally important to be home for death and burial; thus, it is possible that pre-terminal patients would go “up-country” to die, and not be reported or counted as deaths in the Kibera surveillance system.

Besides incidence, we also report on longitudinal prevalence from the household visits, which is the cumulative number of days over a given time period in which a person is ill [63], [64]. Longitudinal prevalence might be a better measure of illness burden, because unlike incidence, it accounts for both the number and duration of illness episodes. Longitudinal prevalence has been more closely correlated with growth faltering and mortality in children with diarrhea than has incidence [63], [64]. Longitudinal prevalence of fever and lower respiratory tract infections in children has also been associated with growth faltering, but has not been evaluated as a disease burden measure in adults [65].

We encountered several challenges when comparing rates of disease in the clinic and the community. The specificity of the case definitions for the same syndrome likely differed for household- and clinic-based surveillance. While trained field workers evaluated participants in the household visits, clinicians did so in the clinics, where they could verify reported symptoms based on their clinical assessment of the patient. Also, patients were likely to be sicker when seen in the clinic than at the household visit and so had a greater prior probability of having the syndromes under surveillance. Our decisions whether to use reported versus documented findings was based on previously used diagnostic criteria [8], [14]. While reported symptoms in the household might have had reduced specificity (e.g. AFI), using an exam-based case definition in the household introduced other complications because a person must be at home and acutely ill at the time of the biweekly household visit to capture abnormal signs. In using an exam-based definition at the household for ALRI, we adjusted rates to compensate for individuals who might have had ALRI who did not have an exam done, an adjustment that could have been prone to bias if those without exams were more or less likely to have ALRI than those with exam.

We adjusted clinic rates to avoid underestimation of the true burden that would result from basing calculations on only one clinic in sites where multiple clinics were available to the population. We assumed that those illnesses episodes in which care was sought at another local clinic were likely to be of equivalent severity. Similar adjustments to incidence rates based on persons not sampled have been made before for estimates of acute febrile illness in Egypt [55]. It is possible though that those illness episodes that resulted in a visit to the referral clinic were not the same as those that went to other clinics. In Asembo, the bias might have been for more severe illnesses to visit Lwak, where people knew admission would be free. Whereas in Kibera, the opposite might have been the case, whereby ill persons who felt they would need hospitalization would bypass Tabitha and go directly to an inpatient hospital.

In conclusion, disease burden estimates such as this derived from local surveillance data can provide credible data to decision makers about the need for targeted interventions and what health and economic impact such interventions might yield. We demonstrated that disease rates in the community are higher in rural than urban Kenya, although clinic-based rates tend to be higher in the urban setting, where health utilization is better. We also showed that in rural Kenya many illness episodes are missed at the clinic and the burden of infectious diseases is likely several-fold higher than can be measured at the clinic. Such data suggest interventions to improve health care access or to implement non-facility based treatment strategies, such as effective home-management of malaria and diarrhea, might have a large public health impact in rural Africa [30]. Both communities had an intolerable burden of illness, with children experiencing illness in up to one-quarter of the year. Proven interventions, such as insecticide-treated bednets and the pneumococcal conjugate and rotavirus vaccines, are urgently needed to alleviate this burden [31], [32], [34]. Without more effective interventions to control these common infectious disease syndromes, continued high morbidity and mortality in poor African settings will continue to hamper development of these communities.

Author Contributions

Conceived and designed the experiments: DF HB RB. Performed the experiments: DF BO GB HB MKN RB. Analyzed the data: DF AA LC BA JW RB. Wrote the paper: DF BO GB AA LC BA HB MKN JW RB.


  1. 1. Black RE, Cousens S, Johnson HL, Lawn JE, Rudan I, et al. (2010) Global, regional, and national causes of child mortality in 2008: a systematic analysis. Lancet 375: 1969–1987.
  2. 2. World Health Organization (2010) World Health Statistics. Geneva: World Health Organization. Available:[Indicator].[SDEC].Members. Accessed: May 1, 2009.
  3. 3. Amin AA, Marsh V, Noor AM, Ochola SA, Snow RW (2003) The use of formal and informal curative services in the management of paediatric fevers in four districts in Kenya. Trop Med Int Health 8: 1143–1152.
  4. 4. Central Bureau of Statistics (CBS) [Kenya] Ministry of Health Kenya, and ORC Macro (2004) Kenya Demographic and Health Survey 2003: Key Findings. Calverton, Maryland, USA: Central Bureau of Statistics (CBS) [Kenya], Ministry of Health (MOH) [Kenya], and ORC Macro.
  5. 5. Adazu K, Lindblade KA, Rosen DH, Odhiambo F, Ofware P, et al. (2005) Health and demographic surveillance in rural western Kenya: a platform for evaluating interventions to reduce morbidity and mortality from infectious diseases. Am J Trop Med Hyg 73: 1151–1158.
  6. 6. Jones G, Steketee RW, Black RE, Bhutta ZA, Morris SS (2003) How many child deaths can we prevent this year? Lancet 362: 65–71.
  7. 7. Kosek M, Bern C, Guerrant RL (2003) The global burden of diarrhoeal disease, as estimated from studies published between 1992 and 2000. Bull World Health Organ 81: 197–204.
  8. 8. Rudan I, Tomaskovic L, Boschi-Pinto C, Campbell H (2004) Global estimate of the incidence of clinical pneumonia among children under five years of age. Bull World Health Organ 82: 895–903.
  9. 9. Selwyn BJ (1990) The epidemiology of acute respiratory tract infection in young children: comparison of findings from several developing countries. Coordinated Data Group of BOSTID Researchers. Rev Infect Dis 12: Suppl 8S870–888.
  10. 10. Feikin DR, Audi A, Olack B, Bigogo GM, Polyak C, et al. (2010) Evaluation of the optimal recall period for disease symptoms in home-based morbidity surveillance in rural and urban Kenya. International Journal of Epidemiology 39: 450–8.
  11. 11. Ngongo CJ, Mathingau FA, Burke H, Brieger W, Frick K, et al. (2007) Community participation to refine measures of socio-economic status in urban slum settings in Kenya. Int Q Community Health Educ 28: 33–49.
  12. 12. Mcelroy PD, Terk Kuile FO, Hightower AW, Hawley WA, Phillips-Howard PA, et al. (2001) All-Cause Mortality Among Young Children In Western Kenya. Vi: The Asembo Bay Cohort Project. Am J Trop Med Hyg 64: 18–27.
  13. 13. Phillips-Howard PA, Nahlen BL, Wannemuehler KA, Kolczak MS, ter Kuile FO, et al. (2003) Impact of permethrin-treated bed nets on the incidence of sick child visits to peripheral health facilities. Am J Trop Med Hyg 68: 38–43.
  14. 14. World Health Organization (2008) Integrated Management of Childhood Illness. World Health Organization. Available: Accessed: January 28, 2009.
  15. 15. Baqui AH, Black RE, Yunus M, Hoque AR, Chowdhury HR, et al. (1991) Methodological issues in diarrhoeal diseases epidemiology: definition of diarrhoeal episodes. Int J Epidemiol 20: 1057–1063.
  16. 16. Wright JA, Gundry SW, Conroy R, Wood D, Du Preez M, et al. (2006) Defining episodes of diarrhoea: results from a three-country study in Sub-Saharan Africa. J Health Popul Nutr 24: 8–16.
  17. 17. Rothman K, Greenland S (1998) Modern Epidemiology. Philadelphia: Lippincott-Raven Publishers.
  18. 18. Bickel PJ, Doksum KA (2001) Mathematical Statistics: Basic Ideas and Selected Topics. Upper Saddle River, New Jersey: Prentice Hall, Inc.
  19. 19. World Health Organization (2010) WHO Statistical Information System. Available: Accessed May 2, 2010.
  20. 20. Syagga PM, Mitullah WV, Gitau SK (2001) Slum Upgrading: Lessons Learned in Nairobi. Nairobi: UNCHS (Habitat). Available: Accessed July 24, 2010.
  21. 21. Central Bureau of Statistics (2001) Population Distribution by Administrative Areas and Urban Centres. Nairobi: Kenya 1999 Population and Housing Census. pp. 2–48.
  22. 22. Breiman RF, Olack B, Schultz A, Roder S, Kimani K, et al. Healthcare Utilization for Major Infectious Disease Syndromes in an Informal Settlement, Nairobi, Kenya. Journal of Health Population and Nutrition. (in press).
  23. 23. Burton DC, Flannery B, Onyango B, Larson C, Alaii J, Zhang X, Hamel MJ, Breiman RF, Feikin DRHealth-seeking behaviour for common infectious disease illnesses in western Kenya. Journal of Health Population and Nutrition. (in press).
  24. 24. Armstrong Schellenberg JR, Adam T, Mshinda H, Masanja H, Kabadi G, et al. (2004) Effectiveness and cost of facility-based Integrated Management of Childhood Illness (IMCI) in Tanzania. Lancet 364: 1583–1594.
  25. 25. Bhattarai A, Ali AS, Kachur SP, Martensson A, Abbas AK, et al. (2007) Impact of artemisinin-based combination therapy and insecticide-treated nets on malaria burden in Zanzibar. PLoS Med 4: e309.
  26. 26. Feikin DR, Nguyen LM, Adazu K, Ombok M, Audi A, et al. (2009) The impact of distance of residence from a peripheral health facility on pediatric health utilisation in rural western Kenya. Trop Med Int Health 14: 54–61.
  27. 27. Hayes JA (1990) Hedging opportunities in health care. Health Care Strateg Manage 8: 12–14.
  28. 28. Bigogo G, Audi A, Aura B, Aol G, Breiman RF, et al. (2010) Health-seeking patterns among participants of population-based morbidity surveillance in rural western Kenya: implications for calculating disease rates. Int J Infect Dis 14: e967–973.
  29. 29. Bejon P, Lusingu J, Olotu A, Leach A, Lievens M, et al. (2008) Efficacy of RTS,S/AS01E vaccine against malaria in children 5 to 17 months of age. N Engl J Med 359: 2521–2532.
  30. 30. Hopkins H, Talisuna A, Whitty CJ, Staedke SG (2007) Impact of home-based management of malaria on health outcomes in Africa: a systematic review of the evidence. Malar J 6: 134.
  31. 31. World Health Organization (2007) Pneumococcal conjugate vaccine for childhood immunization – WHO position paper. Weekly epidemiological record 82: 93–104.
  32. 32. World Health Organization (2009) Meeting of the immunization Strategic Advisory Group of Experts, April 2009 - conclusions and recommendations. Weekly epidemiological record 84: 220–236.
  33. 33. Sacarlal J, Aide P, Aponte JJ, Renom M, Leach A, et al. (2009) Long-term safety and efficacy of the RTS,S/AS02A malaria vaccine in Mozambican children. J Infect Dis 200: 329–336.
  34. 34. Masanja H, de Savigny D, Smithson P, Schellenberg J, John T, et al. (2008) Child survival gains in Tanzania: analysis of data from demographic and health surveys. Lancet 371: 1276–1283.
  35. 35. Campbell H, Byass P, Lamont AC, Forgie IM, O'Neill KP, et al. (1989) Assessment of clinical criteria for identification of severe acute lower respiratory tract infections in children. Lancet 1: 297–299.
  36. 36. Fagbule D, Parakoyi DB, Spiegel R (1994) Acute respiratory infections in Nigerian children: prospective cohort study of incidence and case management. J Trop Pediatr 40: 279–284.
  37. 37. Oyejide CO, Osinusi K (1990) Acute respiratory tract infection in children in Idikan Community, Ibadan, Nigeria: severity, risk factors, and frequency of occurrence. Rev Infect Dis 12: Suppl 8S1042–1046.
  38. 38. Wafula EM, Onyango FE, Mirza WM, Macharia WM, Wamola I, et al. (1990) Epidemiology of acute respiratory tract infections among young children in Kenya. Rev Infect Dis 12: Suppl 8S1035–1038.
  39. 39. Rudan I, Boschi-Pinto C, Biloglav Z, Mulholland K, Campbell H (2008) Epidemiology and etiology of childhood pneumonia. Bull World Health Organ 86: 408–416.
  40. 40. Corbett EL, Churchyard GJ, Charalambos S, Samb B, Moloi V, et al. (2002) Morbidity and mortality in South African gold miners: impact of untreated disease due to human immunodeficiency virus. Clin Infect Dis 34: 1251–1258.
  41. 41. Fedson DS, Scott JA (1999) The burden of pneumococcal disease among adults in developed and developing countries: what is and is not known. Vaccine 17: Suppl 1S11–18.
  42. 42. Feikin DR, Feldman C, Schuchat A, Janoff EN (2004) Global strategies to prevent bacterial pneumonia in adults with HIV disease. Lancet Infect Dis 4: 445–455.
  43. 43. French N, Nakiyingi J, Carpenter LM, Lugada E, Watera C, et al. (2000) 23-valent pneumococcal polysaccharide vaccine in HIV-1-infected Ugandan adults: double-blind, randomised and placebo controlled trial. Lancet 355: 2106–2111.
  44. 44. Gilks CF, Ojoo SA, Ojoo JC, Brindle RJ, Paul J, et al. (1996) Invasive pneumococcal disease in a cohort of predominantly HIV-1 infected female sex-workers in Nairobi, Kenya. Lancet 347: 718–723.
  45. 45. Wiktor SZ, Sassan-Morokro M, Grant AD, Abouya L, Karon JM, et al. (1999) Efficacy of trimethoprim-sulphamethoxazole prophylaxis to decrease morbidity and mortality in HIV-1-infected patients with tuberculosis in Abidjan, Cote d'Ivoire: a randomised controlled trial. Lancet 353: 1469–1475.
  46. 46. Murray C, Lopez A (1996) Global Health Statistics. Cambridge: Harvard School of Public Health on behalf of the World Health Organization and World Bank. pp. 286–287.
  47. 47. Tornheim JA, Manya AS, Oyando N, Kabaka S, Breiman RF, et al. (2007) The epidemiology of hospitalized pneumonia in rural Kenya: the potential of surveillance data in setting public health priorities. Int J Infect Dis 11: 536–543.
  48. 48. Ortiz JR, Sotomayor V, Uez OC, Oliva O, Bettels D, et al. (2009) Strategy to enhance influenza surveillance worldwide. Emerg Infect Dis 15: 1271–1278.
  49. 49. Manun'ebo MN, Haggerty PA, Kalengaie M, Ashworth A, Kirkwood BR (1994) Influence of demographic, socioeconomic and environmental variables on childhood diarrhoea in a rural area of Zaire. J Trop Med Hyg 97: 31–38.
  50. 50. Mirza NM, Caulfield LE, Black RE, Macharia WM (1997) Risk factors for diarrheal duration. Am J Epidemiol 146: 776–785.
  51. 51. Molbak K, Jensen H, Ingholt L, Aaby P (1997) Risk factors for diarrheal disease incidence in early childhood: a community cohort study from Guinea-Bissau. Am J Epidemiol 146: 273–282.
  52. 52. Mermin J, Lule J, Ekwaru JP, Downing R, Hughes P, et al. (2005) Cotrimoxazole prophylaxis by HIV-infected persons in Uganda reduces morbidity and mortality among HIV-uninfected family members. Aids 19: 1035–1042.
  53. 53. Mermin J, Lule J, Ekwaru JP, Malamba S, Downing R, et al. (2004) Effect of co-trimoxazole prophylaxis on morbidity, mortality, CD4-cell count, and viral load in HIV infection in rural Uganda. Lancet 364: 1428–1434.
  54. 54. Crump JA, Luby SP, Mintz ED (2004) The global burden of typhoid fever. Bull World Health Organ 82: 346–353.
  55. 55. Crump JA, Youssef FG, Luby SP, Wasfy MO, Rangel JM, et al. (2003) Estimating the incidence of typhoid fever and other febrile illnesses in developing countries. Emerg Infect Dis 9: 539–544.
  56. 56. Perkins BA, Zucker JR, Otieno J, Jafari HS, Paxton L, et al. (1997) Evaluation of an algorithm for integrated management of childhood illness in an area of Kenya with high malaria transmission. Bull World Health Organ 75: Suppl 133–42.
  57. 57. Van Hemelrijck MJ, Lindblade KA, Kubaje A, Hamel MJ, Odhiambo F, et al. (2009) Trends observed during a decade of paediatric sick visits to peripheral health facilities in rural western Kenya, 1997–2006. Trop Med Int Health 14: 62–69.
  58. 58. Berkley JA, Lowe BS, Mwangi I, Williams T, Bauni E, et al. (2005) Bacteremia among children admitted to a rural hospital in Kenya. N Engl J Med 352: 39–47.
  59. 59. Enwere G, Biney E, Cheung YB, Zaman SM, Okoko B, et al. (2006) Epidemiologic and clinical characteristics of community-acquired invasive bacterial infections in children aged 2-29 months in The Gambia. Pediatr Infect Dis J 25: 700–705.
  60. 60. Breman JG (2001) The ears of the hippopotamus: manifestations, determinants, and estimates of the malaria burden. Am J Trop Med Hyg 64: 1–11.
  61. 61. Hamel MJ, Greene C, Chiller T, Ouma P, Polyak C, et al. (2008) Does cotrimoxazole prophylaxis for the prevention of HIV-associated opportunistic infections select for resistant pathogens in Kenyan adults? Am J Trop Med Hyg 79: 320–330.
  62. 62. Polhemus ME, Remich SA, Ogutu BR, Waitumbi JN, Otieno L, et al. (2009) Evaluation of RTS,S/AS02A and RTS,S/AS01B in adults in a high malaria transmission area. PLoS One 4: e6465.
  63. 63. Morris SS, Cousens SN, Kirkwood BR, Arthur P, Ross DA (1996) Is prevalence of diarrhea a better predictor of subsequent mortality and weight gain than diarrhea incidence? Am J Epidemiol 144: 582–588.
  64. 64. Schmidt WP, Luby SP, Genser B, Barreto ML, Clasen T (2007) Estimating the longitudinal prevalence of diarrhea and other episodic diseases: continuous versus intermittent surveillance. Epidemiology 18: 537–543.
  65. 65. Morris SS, Santos CA, Barreto ML, Cousens SN, Strina A, et al. (1998) Measuring the burden of common morbidities: sampling disease experience versus continuous surveillance. Am J Epidemiol 147: 1087–1092.