From the *Rwanda Ministry of Health, Rwanda; †Université Libre de Bruxelles, Brussels, Belgium; ‡Rwanda Biomedical Center, Rugenge, Kigali, Rwanda; §Hôpital Universitaire des Enfants Reine Fabiola (HUDERF), Bruxelles, Belgium; ¶Centers for Disease Control and Prevention, Atlanta, USA; ‖World Health Organization, Regional Office for Africa, Brazzaville, Congo; **World Health Organization, Kigali, Rwanda; and ††Harvard University School of Public Health, Boston, MA.
Accepted for publication, June 5, 2013.
The authors have no conflicts of interest to disclose.
Address for correspondence: Fidele Ngabo, Maternal and Child Health Unit, Ministry of Health, Kicukiro, Kigali, KG 631, Rwanda. E-mail: email@example.com.
As rotavirus vaccine is introduced into routine childhood immunization programs in Africa, understanding its impact on diarrheal disease burden is important. The objective of this analysis was to determine whether routinely collected health information on national diarrhea hospitalizations, in-hospital deaths and outpatient visits would be useful to monitor rotavirus vaccine impact.
We analyzed data for all-cause, nonbloody diarrheal disease among children <5 years of age from the routine health management information system (HMIS) in Rwanda from January 2008 through December 2011. We described trends in absolute numbers of inpatient admissions, in-hospital deaths and outpatient visits by year, age and setting.
All-cause, nonbloody diarrheal hospitalizations and outpatient visits among children <5 years of age in Rwanda from 2008 to 2011 peaked during the June to August dry season, coinciding with the rotavirus season. The bulk of the diarrheal disease burden occurred in children <1 year of age. Health centers provided many care to children with diarrhea including 60–72% of hospitalizations and 97–99% of outpatient visits. Many in-hospital diarrheal deaths (84%) occurred in district hospitals.
Given the stable and consistent trends and the prominent seasonality consistent with that of rotavirus, HMIS data should provide a useful baseline to monitor rotavirus vaccine impact on the overall diarrheal disease burden in Rwanda. Active, sentinel surveillance for rotavirus diarrhea will help interpret changes in diarrheal disease trends following vaccine introduction. Other countries planning rotavirus vaccine introduction should explore the availability and quality of their HMIS data.
In concordance with the World Health Organization’s recommendation to introduce rotavirus vaccine in all countries globally, and particularly in those countries with high diarrhea mortality,1 Rwanda became the first African country to introduce the pentavalent rotavirus vaccine (RV5; RotaTeq, Merck and Co, Inc, West Point, PA) into its national childhood immunization schedule in May 2012. Compared with the high efficacy of rotavirus vaccines (85–98%) that was observed in clinical trials in high- and middle-income countries,2,3 clinical trials in Africa found modest efficacy (50–70%) against severe rotavirus diarrhea.4–6 However, despite this lower efficacy, the public health impact of a rotavirus vaccination program in terms of the burden of severe rotavirus disease prevented by vaccinating a given number of infants is substantially greater in developing compared with developed countries due to the substantially higher disease burden in developing countries.4
As a vaccine may perform differently under conditions of routine use than under the controlled conditions of a clinical trial, monitoring vaccine performance and impact, particularly in early adopter countries like Rwanda, is important to understand the full public health benefit of the vaccine. This information will help the Rwanda Ministry of Health to make evidence-based decisions to sustain use of rotavirus vaccines, to advocate for further resources, to maintain the rotavirus vaccination program and to identify the predominant barriers known to affect the performance of rotavirus vaccines in real-world settings. While the vaccine targets diarrhea due to rotavirus, the most common cause of severe diarrhea in children <5 years of age, the impact of the vaccine should be evident by a reduction in all-cause diarrhea hospitalizations due to the high proportion of diarrheal disease due to rotavirus.
The objective of this analysis was to determine whether routinely collected health information on national diarrhea hospitalizations, in-hospital deaths and outpatient visits can be used to monitor the impact of rotavirus vaccine. We used data from the health management information system (HMIS) in Rwanda to describe trends in all-cause, nonbloody diarrhea hospitalizations and outpatient visits among children <5 years of age from 2008 to 2011 before vaccine introduction. These data will provide a baseline to monitor the impact of rotavirus vaccine on the overall diarrheal disease burden among children in Rwanda.
We analyzed data for inpatient admissions, in-hospital deaths and outpatient visits for all-cause, nonbloody diarrhea among children <5 years of age from the routine HMIS in Rwanda from January 2008 through December 2011. Data on disease morbidity and mortality was compiled monthly from registers and patient forms by monitoring and evaluation personnel at each of health facility. Data from health centers were sent to the local district hospital. District hospitals compiled the data from health centers in their catchment area and their own hospital’s data and sent it to the HMIS Unit at the Ministry of Health. The HMIS database does not include data from private health facilities or the national referral hospitals. A data quality assessment was done every quarter to evaluate quality and timeliness of the reporting system. All health centers and districts hospitals in Rwanda provide clinical services on treatment and reporting of diarrhea cases.
We described trends in absolute numbers of inpatient admissions, in-hospital deaths and outpatient visits for all-cause, nonbloody diarrhea by year, age (<1 year vs. 1–4 years) and setting (district hospital vs. health center). We also described the seasonality of nonbloody diarrhea by age and setting.
The number of all-cause, nonbloody diarrhea hospitalizations among children <5 years of age ranged from 13,153 in 2011 to 17,254 in 2009 (Fig. 1A), which accounted for 93–94% of all diarrheal (bloody and nonbloody) hospitalizations in this age group. Almost half of all these hospitalizations (44–52%) occurred among children <1 year of age (Table 1). Many diarrhea hospitalizations occurred in health centers with the proportion of children <5 years of age hospitalized for all-cause, nonbloody diarrhea at health centers compared with district hospitals ranging from 72% in 2008 to 60% in 2011. All-cause, nonbloody diarrhea hospitalizations exhibited a distinct seasonality with the number of cases peaking in June, July and August, which corresponds to the dry season in Rwanda(Fig. 2A).
The number of in-hospital deaths due to all-cause, nonbloody diarrhea among children <5 years of age decreased from 133 in 2008 to 59 in 2011.(Table 1) This large decrease in in-hospital deaths was largely driven by a decrease in deaths among children <1 year of age which decreased from 91 in 2008 to 27 in 2011. Many in-hospital deaths occurred in district hospitals and the proportion of deaths that occurred in district hospitals increased from 78% in 2008 to 92% in 2011. In-hospital diarrheal deaths did not show the same pronounced seasonality as was observed among diarrheal hospitalizations, but a slight increase in deaths was observed during the June to August dry season.
The number of all-cause, nonbloody diarrhea outpatient visits among children <5 years of age ranged from 138,365 in 2011 to 166,397 in 2010 (Fig. 1B), which accounted for 92–93% of all diarrheal outpatient visits in this age group. Over one-third of these outpatient visits (36–40%) occurred among children <1 year of age (Table 1). Many outpatient visits (97–99%) for all-cause, nonbloody diarrhea among children <5 years of age occurred at health centers. Similar to hospitalizations, all-cause, nonbloody diarrhea outpatient visits also peaked during the dry season with the greatest number of cases between May and September (Fig. 2B).
All-cause, nonbloody diarrheal disease among children <5 years of age in Rwanda from 2008 to 2011 peaked during the dry season with an increase in such hospitalizations and outpatient visits during these months. The bulk of the disease burden occurred in children <1 year of age with almost half of diarrheal hospitalizations and over one-third of outpatient visits occurring in this age group. Health centers provided many care to children including 60–72% of hospitalizations and 97–99% of outpatient visits for all-cause, nonbloody diarrhea. However, many in-hospital deaths occurred in district hospitals suggesting that district hospitals treat the more severe cases of all-cause, nonbloody diarrhea.
The strong consistent seasonality and stable trends in diarrheal disease in Rwandan children suggest that routinely collected HMIS data will provide a strong baseline with which to monitor the impact of rotavirus vaccine on diarrhea disease in Rwanda. Although no published data regarding the prevalence of rotavirus among all-cause diarrhea disease in Rwanda are available, data from other African countries suggest that rotavirus accounts for 30–40% of all-cause diarrheal hospitalizations and seasonal peaks in rotavirus disease corresponding with the dry season have also been observed in some countries.7 A reduction in rotavirus disease burden following the introduction of the vaccine will likely be evident in all-cause diarrhea hospitalizations and outpatient visits due to the large contribution of rotavirus to all-cause diarrhea among children.
Key epidemiologic features of rotavirus diarrhea will help interpret whether changes in diarrhea disease trends following rotavirus vaccine introduction are likely due to the vaccine. First, diarrhea disease is highly seasonal in Rwanda. Thus, assuming that the peaks in diarrheal disease are due to peaks in rotavirus disease, the largest decreases in diarrheal illness should occur during the June to August dry season when rotavirus is most prevalent. Second, rotavirus causes severe gastroenteritis and is therefore likely to account for a higher proportion of visits for all-cause, nonbloody diarrhea among inpatients than outpatients. Similarly, the vaccine is more effective against severe disease than milder disease. Thus, following vaccine introduction, a greater decrease in all-cause, nonbloody diarrhea cases should be seen among inpatients compared with outpatients. Third, as the vaccine is rolled out, the greatest impact during the first year following vaccine introduction should be seen among children <1 year of age as only this birth cohort is directly protected by vaccination with high coverage. While older children may be indirectly protected by reduced rotavirus transmission among children <1 year of age, the reduction in disease among older children should be significantly smaller than that observed among children <1 year of age. Finally, given the decreasing trend in all-cause, nonbloody, in-hospital diarrhea deaths, continued reductions in in-hospital mortality should be interpreted with caution and may not be able to be attributed to the vaccine using these data.
This study has several limitations. First, HMIS data are collected passively and depend on regular reporting from health centers and district hospitals. Diagnoses of nonbloody diarrhea are based on clinical assessment and no standard case definition for diarrheal illness is used. However, the consistency of reporting from all facilities and the regular seasonality suggests that these data are stable during the period of our study. Second, HMIS only collects data from public hospitals and health centers. Children who seek care for diarrheal disease at private facilities will not be captured by this system. However, many healthcare in Rwanda is provided through public facilities. For example, only 1% of women delivered in a private facility and only 4% use a private facility as their supply for modern contraceptive use.8 Thus, the proportion of children seeking care for diarrheal illness at a private facility is also likely small. Furthermore, trends in diarrhea disease among children at private facilities will likely be similar to trends in public facilities. Third, the management of diarrheal disease in Rwanda has changed during the study period with an increased focus of managing diarrheal cases in health centers and community using the World Health Organization’s Integrated Management of Childhood Illness guidelines. Under Integrated Management of Childhood Illness guidelines, only severe diarrhea cases are referred to district hospitals for care. The overall decreasing trend of in-hospital deaths coupled with an increased proportion of in-hospital deaths occurring at district hospitals likely reflects this change in management of diarrheal disease in Rwanda. However, the in-hospital deaths likely underestimate diarrheal mortality in Rwanda as children who die in the community or in private facilities or referral hospitals will not be captured by the HMIS. Finally, HMIS data only describe trends in all-cause diarrhea illness as testing for rotavirus was not routinely performed. However, the consistent peaks in all-cause, nonbloody diarrheal disease that occur during the dry season suggest a seasonal pattern of rotavirus disease in Rwanda. Testing for rotavirus at sentinel sites will help establish the seasonality of rotavirus disease.
Given the stable and consistent trends in all-cause, nonbloody diarrheal disease, HMIS data will provide a useful baseline to monitor the impact of rotavirus vaccine on the overall diarrheal disease burden in Rwanda. Other countries that plan to introduce rotavirus vaccines into their national immunization programs should explore the availability and quality of their HMIS data. Additionally, active, sentinel surveillance for rotavirus diarrhea will help to interpret changes in diarrheal disease trends following vaccine introduction and these active surveillance sites can serve as a platform for case-control studies to measure the field effectiveness of the vaccine. As Rwanda is 1 of the early adopters of rotavirus vaccine in Africa, documenting the impact on the overall burden of diarrheal disease in children is important for sustaining vaccine use in Rwanda and for other countries to see the benefits of vaccine introduction.
1. Rotavirus vaccines:an update. Wkly Epidemiol Rec. 2009; 84:533–540
2. Vesikari T, Matson DO, Dennehy P, et al. Rotavirus Efficacy and Safety Trial (REST) Study Team Safety and efficacy of a pentavalent human-bovine (WC3) reassortant rotavirus vaccine. N Engl J Med. 2006; 354:23–33
3. Ruiz-Palacios GM, Pérez-Schael I, Velázquez FR, et al. Human Rotavirus Vaccine Study Group Safety and efficacy of an attenuated vaccine against severe rotavirus gastroenteritis. N Engl J Med. 2006; 354:11–22
4. Madhi SA, Cunliffe NA, Steele D, et al. Effect of human rotavirus vaccine on severe diarrhea in African infants. N Engl J Med. 2010; 362:289–298
5. Armah GE, Sow SO, Breiman RF, et al. Efficacy of pentavalent rotavirus vaccine against severe rotavirus gastroenteritis in infants in developing countries in sub-Saharan Africa: a randomised, double-blind, placebo-controlled trial. Lancet. 2010; 376:606–614
6. Zaman K, Dang DA, Victor JC, et al. Efficacy of pentavalent rotavirus vaccine against severe rotavirus gastroenteritis in infants in developing countries in Asia: a randomised, double-blind, placebo-controlled trial. Lancet. 2010; 376:615–623
7. Mwenda JM, Ntoto KM, Abebe A, et al. Burden and epidemiology of rotavirus diarrhea in selected African countries: preliminary results from the African Rotavirus Surveillance Network. J Infect Dis. 2010; 202:(suppl)S5–S11
8. National Institute of Statistics of Rwanda (NISR) [Rwanda], Ministry of Health (MOH) [Rwanda], and ICF International Rwanda Demographic and Health Survey 2010. 2012; Calverton, Maryland, USA NISR, MOH, and ICF International