Water, Sanitation, and Hygiene and Nutritional Risk Factors for Acute Respiratory Illness in the Democratic Republic of the Congo: REDUCE Prospective Cohort Study

ABSTRACT.

The objective of this cohort study was to examine the prevalence of acute respiratory illness among children under 5 years of age and to identify water, sanitation, and hygiene (WASH) and nutritional risk factors. This prospective cohort study was conducted in Walungu Territory, South Kivu, Democratic Republic of the Congo (DRC) and enrolled 512 participants. Spot checks of the household environment were conducted at baseline. Baseline minimum dietary diversity (MDD) was defined by consumption of five or more of the following food groups: 1) breast milk; 2) grains, roots, and tubers; 3) legumes and nuts; 4) dairy products; 5) flesh foods; 6) eggs; 7) vitamin A rich fruits and vegetables; and 8) other fruits and vegetables. Acute respiratory illness was defined as caregiver-reported rapid breathing, difficulty breathing, lower chest wall in-drawing, or coughing in the previous 2 weeks obtained at a 6-month follow-up. A total of 58% of children had acute respiratory illness, 19% had soap present in the household cooking area, and 4% in the defecation area, and 21% of children met MDD. A decreased odds of acute respiratory illness was associated with soap being present in the household cooking area (odds ratio [OR]: 0.49, 95% confidence interval [CI]: 0.38–0.88) and children with MDD (OR: 0.62, 95% CI: 0.38–1.00). These findings highlight the need for interventions targeting hygiene and improved dietary diversity among rural DRC households to reduce the rate of respiratory illnesses in children under 5 years.

INTRODUCTION

Acute respiratory infections remain a leading cause of morbidity and mortality among children under 5 years of age globally.¹ The burden of acute respiratory infections on child morbidity and mortality disproportionately impacts lower- and middle-income countries.² In the Democratic Republic of the Congo (DRC), which bears one of the highest burdens of under 5 mortality in both sub-Saharan Africa and the world, it is estimated that almost 300,000 children under 5 die each year.³ Risk factors for respiratory infections include low birth weight, lack of exclusive breastfeeding, incomplete immunization, household crowding, indoor air pollution, and undernutrition, factors that are often more common in impoverished communities.⁴

Acute respiratory infections can be reduced through water, sanitation, and hygiene (WASH) interventions. Handwashing with soap is strongly associated with reductions in respiratory illness.^5–7 Despite demonstrated effectiveness, difficulties with sustained behavior change have sometimes limited the effectiveness of handwashing interventions.^8,9 Recent large-scale WASH interventions have also shown mixed efficacy in reducing respiratory illness among young children.¹⁰ The WASH Benefits trial conducted in Bangladesh found that individual WASH interventions (water or sanitation or hygiene) significantly reduced respiratory illness in children.¹¹ However, the WASH Benefits trial in Kenya found no effect of comparable WASH interventions.¹² This finding suggests that the effectiveness of WASH interventions on respiratory illness may depend on several factors, such as study setting and intervention type, duration, and intensity.

Nutrition also plays a key role in child health.¹³ Malnutrition impairs immune system functioning and increases children’s susceptibility to illness.¹⁴ A meta-analysis found malnutrition to be associated with severe respiratory illness in young children.^15,16 Recent studies have investigated the combined impact of WASH and nutritional interventions on child health.^{11,12,17–19} The SHINE trial conducted in Zimbabwe found that infant and young child feeding in combination with WASH did not reduce stunting or anemia in children.¹⁷ Similarly, the WASH Benefits trials in Kenya and Bangladesh both found no additive benefit of combined WASH and nutrition programs for reductions in child stunting and diarrhea prevalence.^18,19 When examining the impact of combined WASH and nutritional interventions on acute respiratory illness, the WASH Benefits trial in Bangladesh found reductions in acute respiratory illness with the combined interventions compared with a nutrition intervention, but no additional benefit compared with the individual WASH arms.¹¹

Previous studies have found animal ownership to be associated with improved nutrition.²⁰ Contact with poultry, however, is also known to be a risk factor for respiratory infection among young children.²¹ Despite this, few studies have directly examined the association between animals in the home and respiratory illness in children in lower income settings. Investigating this association in Uganda, Ercumen et al.²² reported decreased respiratory infection prevalence with increased cow ownership as well as increased poultry ownership, compared with households that had less cow and less poultry ownership.

Despite the high burden of respiratory illness in children under 5 years, there is limited evidence examining the impact of WASH and nutritional risk factors on acute respiratory illness in the DRC. Based on previous literature, we hypothesized that soap presence would be associated with decreased respiratory illness in children under 5 years, whereas animal exposure may increase respiratory illness. The Reducing Enteropathy, Undernutrition, and Contamination in the Environment (REDUCE) program focuses on identifying exposure pathways to pathogens that are significant contributors to morbidity for young children in the DRC, as well as on developing and evaluating scalable interventions to reduce exposure from these pathways. The objective of this prospective cohort study was to determine the prevalence of acute respiratory illness symptoms in children under 5 in the DRC and to identify WASH and nutritional risk factors associated with decreased acute respiratory illness burden to target in future interventions in this population.

METHODS

Ethical approval.

Informed consent was obtained from a parent or guardian of each study participant. Study procedures were approved by the research ethical review committees of the University of Kinshasa (Protocol 043-2017) and the Johns Hopkins Bloomberg School of Public Health (Protocol 8057).

Study design.

This prospective cohort study was conducted in the Walungu Territory in South Kivu, DRC. The study included 512 children under 5 years of age in 223 households. The sample size was based on the number of children that could be enrolled between June 2018 and January 2019. To improve accuracy and standardization of results, the research team received training on collection of caregiver-reported data and spot check observations. Questions were asked using a standardized script.

To obtain data on household demographics and child dietary diversity and respiratory illness, caregivers were administered a questionnaire at baseline and a 6-month follow-up visit. At baseline, the questionnaire assessed child respiratory health and diet, with the 6-month questionnaire assessing only child respiratory health. Caregivers were asked to report presence of rapid breathing, difficulty breathing, lower chest wall in-drawing, and coughing in the past 2 weeks for each child under 5 years in their household.²³ Minimum dietary diversity (MDD) is a validated proxy for diet and nutritional adequacy in children.²⁴ MDD was assessed at baseline by research assistants using a WHO-recommended standardized definition.²⁵ MDD is a binary indicator measuring dietary diversity as consumption of food from at least five of the following groups in the previous 24 hours: 1) breast milk; 2) grains, roots, and tubers; 3) legumes and nuts; 4) dairy products; 5) flesh foods; 6) eggs; 7) vitamin A rich fruits and vegetables; and 8) other fruits and vegetables.

Unannounced spot checks were conducted after the baseline visit for all households to observe animal and soap presence and location on the household compound. The entire household compound, indoor cooking area, and child’s sleeping area were all assessed for the presence of animals including chickens, rabbits, and guinea pigs. Animal presence collected through unannounced spot checks was assessed to determine whether households had contact with animals—even those that they did not own. Soap and/or soapy water being present in the cooking area (within 10 steps of stove), and in the defecation area (within 10 steps of latrine/area used for defecation) was assessed. Spot checks occurred any time between 8:00 am and 1:00 pm so that households could not prepare for the arrival of research staff.

Statistical analysis.

The primary aim of this study was 2-fold: 1) to ascertain the 2-week prevalence of respiratory illness in children under 5 years of age in the DRC and 2) to identify WASH and nutritional risk factors for respiratory outcomes in this population. The WASH risk factors included in this study were as follows: 1) soap presence in the cooking and defecation area; 2) animal presence on the household compound, in the indoor cooking area, and in the child’s sleeping area; and 3) MDD.

Children with baseline spot check and dietary diversity data and 6-month follow-up respiratory symptom data were included in the analysis. To be eligible, children had to have at least 5 months of surveillance data during the study period. We assessed three respiratory outcomes for their association with risk factors: acute respiratory illness, acute lower respiratory infection (ALRI), and coughing. Based on previous studies conducted in Bangladesh and Kenya,^11,12 acute respiratory illness was defined as at least one of the following symptoms: rapid breathing, difficulty breathing, lower chest wall in-drawing, and coughing. Acute lower respiratory infection was defined using the WHO integrated management of childhood illness criteria for classification of childhood pneumonia of the presence of at least one of the following: 1) coughing and rapid breathing; 2) coughing and lower chest wall in-drawing; 3) difficulty breathing and rapid breathing; and 4) difficulty breathing and lower chest wall in-drawing.²⁶ We also examined the association between risk factors and coughing alone, based on the previous studies.¹⁰

To assess the association between WASH and nutritional risk factors and respiratory outcomes, we performed logistic regression using generalized estimating equations with an exchangeable correlation structure to account for potential household level clustering and approximate 95% confidence intervals (CIs). To account for socioeconomic status and other household differences, all analyses were adjusted for child age, primary caregiver formal education, total number of individuals living in the home (crowding), animal ownership, and housing wall material (binary variable for mud walls and other wall materials). Analysis was completed using Stata 16 (Stata Corp LP, College Station, TX).

RESULTS

Baseline demographics.

A total of 512 participants were enrolled in the cohort study at baseline. Over the study period, 18% of participants (94/512) were lost to follow-up and did not receive a 6-month clinical surveillance visit. Those lost to follow-up and those retained in the study had similar baseline acute respiratory illness prevalence (P = 0.98). At baseline, 67% of children (343/512) had acute respiratory illness and 14% (72/512) had ALRI in the past 2 weeks (Table 1). Cough was the most common respiratory symptom at baseline, with 64% (328/512) of children having caregiver-reported coughing. Lower chest wall indrawing was the second most common, at 11% (54/512), followed by fast breathing and difficulty breathing, at 10% (51/512) and 8% (41/512) respectively. The median baseline age for participants was 19 months (standard deviation (SD): ±16, range: 0–60). Fifty percent of participants (256/512) were female. The median household size was six persons (SD: 2.3, range: 2–17). The primary water source type among households at baseline was a protected spring (71%, 366/512), followed by public tap (20%, 101/512), and unprotected spring (4%, 23/512). The median one-way distance to a water source was 15 minutes (SD: 15.1, range: 0–120).

Table 1.

Baseline demographic characteristics, WASH and nutritional risk factors, and respiratory illness in children < 5 years in the Democratic Republic of the Congo

	%	n	N
All participants			512
Households			250
Age in months (median ± SD) (min–max)	19 ± 16 (0–60)
Household size	6 ± 2.3 (2--17)
Female	50	256	512
Mud walls	66	336	512
Caregiver formal education	55	281	512
Unimproved latrine	99	506	512
Water source type
Protected spring	71	366	512
Public tap	20	101	512
Unprotected spring	4	23	512
Other	4	22	512
Distance to water source (minutes, one way)
Median ± SD (min–max)	15 ± 15.1 (0–120)
Piped water on household compound	5	27	512
Soap or soapy water in household compound	44	225	512
Soap in cooking area	19	97	512
Soap in defecation area	4	20	512
Minimum dietary diversity food groups
Group 1: Breast milk	51	263	512
Group 2: Grains, roots, and tubers	88	450	510
Group 3: Legumes and nuts	34	172	508
Group 4: Dairy products	12	60	509
Group 5: Flesh foods	67	341	510
Group 6: Eggs	5	25	511
Group 7: Vitamin A rich fruits and vegetables	78	395	508
Group 8: Other fruits and vegetables	20	102	511
Minimum dietary diversity (≥ 5 food groups)	22	113	512
Animal ownership	86	388	450
Animal presence in household	88	450	512
Chickens	71	366	512
Guinea pigs	43	218	512
Rabbits	21	107	512
Animal presence in cooking area	50	220	440
Chickens	24	106	440
Guinea pigs	37	163	440
Rabbits	17	75	440
Animal presence in sleeping area	22	91	412
Chickens	8	32	412
Guinea pigs	18	74	412
Rabbits	11	47	412
Acute respiratory illness	67	343	512
Cough	64	328	512
Fast breathing	10	51	512
Lower chest wall in-drawing	11	54	512
Difficulty breathing	8	41	512
Acute lower respiratory illness	14	72	512

n refers to the number of participants with the given characteristic; acute respiratory illness defined as rapid breathing, difficulty breathing, lower chest wall indrawing, and coughing in the past 2 weeks; acute lower respiratory illness defined as presence of at least one of the following: coughing and rapid breathing, coughing and chest indrawing, difficulty breathing and rapid breathing, difficulty breathing and lower chest wall indrawing.

Table 2.

Association between respiratory illness and soap presence, minimum dietary diversity, and animal presence

				Cough	Acute respiratory illness*	ALRI†
Risk factor	%	n	N	6-month follow-up odds ratio (95% CI)	6-month follow-up odds ratio (95% CI)	6-month follow-up odds ratio (95% CI)
Soap present in household	43	181	418	0.62 (0.41–0.95)	0.58 (0.38–0.88)	0.88 (0.45–1.72)
Soap in cooking area	19	80	414	0.52 (0.31–0.89)	0.49 (0.29–0.84)	0.43 (0.16–1.18)
Soap in defecation area	2	10	418	1.14 (0.30–4.38)	1.05 (0.27–4.03)	–
Minimum dietary diversity	21	86	418	0.64 (0.40–1.03)	0.62 (0.38–1.00)	0.57 (0.22–1.48)
Animal presence in household	84	352	418	1.34 (0.72–2.48)	1.41 (0.76–2.62)	1.95 (0.56–6.75)
Guinea pig	43	181	418	0.93 (0.60–1.45)	0.95 (0.61–1.48)	1.39 (0.64–3.02)
Chicken	73	304	418	1.09 (0.65–1.80)	1.18 (0.71–1.98)	0.53 (0.23–1.22)
Rabbit	22	90	418	0.89 (0.53–1.50)	0.83 (0.49–1.40)	0.71 (0.31–1.64)
Animal presence in sleeping area	24	85	352	1.14 (0.65–1.99)	1.18 (0.66–2.08)	0.81 (0.34–1.93)
Guinea pig	18	65	352	0.98 (0.53–1.81)	0.97 (0.52–1.79)	0.95 (0.37–2.43)
Chicken	9	31	352	0.47 (0.21–1.06)	0.43 (0.19–0.99)	0.26 (0.04–1.82)
Rabbit	10	34	352	1.00 (0.44–2.31)	0.92 (0.40–2.12)	0.51 (0.13–2.05)
Animal presence in cooking area	53	202	381	1.15 (0.73–1.83)	1.12 (0.70–1.80)	1.13 (0.50–2.57)
Chicken	26	98	381	0.83 (0.51–1.34)	0.77 (0.47–1.25)	0.61 (0.24–1.54)
Guinea pig	39	149	381	0.94 (0.59–1.50)	0.90 (0.56–1.45)	1.10 (0.46–2.61)
Rabbit	19	72	381	0.83 (0.47–1.45)	0.78 (0.45–1.38)	0.58 (0.20–1.64)

Bolded text indicates significance at P < 0.05. All analyses adjusted for participant age, caregiver formal education, number of individual living in household, animal ownership, and wall material. An odds ratio for soap presence in the defecation area could not be calculated due to only 10 households with soap present in the defecation area.

^*Acute respiratory illness defined as rapid breathing, difficulty breathing, lower chest wall in-drawing or coughing in the past 2 weeks.

^†Acute lower respiratory infection (ALRI) defined as cough and rapid breathing, cough and lower chest wall in-drawing, difficulty breathing and rapid breathing, or difficulty breathing and lower chest wall indrawing in the past 2 weeks. Number of participants (N) varies because of inclusion of variable collection at spot checks and ability to view household area.

Soap presence, MDD, and animal presence.

Overall, 19% (97/512) of children’s homes had soap present in the cooking area, and only 4% (20/512) had soap present in the defecation area. Twenty-two percent (113/512) of children met the MDD criteria. Grains, roots, and tubers at 88% (450/512), vitamin A rich fruits and vegetables at 78% (395/508), and flesh foods at 67% (341/510) were the most common food categories consumed by children in the last 24 hours. Baseline animal presence on the household compound was high at 88% (450/512); 14% (62/450) of these households did not report animal ownership. Presence of chickens on the household compound was 71% (366/512), guinea pigs at 43% (218/512), and rabbits at 21% (107/512). Fifty percent of households (220/440) had animals in the indoor cooking areas, and 22% (91/412) in child sleeping areas. In the indoor cooking area, chickens were present in 24% (106/440) of homes, guinea pigs for 37% (163/512), and rabbits for 17% (75/440). In the sleeping area, chickens were present in 8% (32/412) of children’s households, guinea pigs for 18% (74/412), and rabbits for 11% (47/412).

Regression analyses.

A total of 418 children under 5 years of age who had baseline spot check and questionnaire data and 6-month follow-up clinical surveillance data were included in this analysis. At the 6-month follow-up, 58% of children under 5 years had caregiver-reported acute respiratory illness (244/418). In addition, 57% (237/418) of children had coughing and 9% (39/418) had ALRI. All regression models were adjusted for baseline child age, primary caregiver formal education, total number of individuals in the household, animal ownership, and housing wall material. Children living in households with soap present in the cooking area had significantly lower acute respiratory illness (odds ratio [OR]: 0.49, 95% CI: 0.29–0.84, P = 0.009) and coughing (OR: 0.52, 95% CI: 0.31–0.89, P = 0.017) (Table 2). Acute respiratory illness (OR: 0.58, 95% CI: 0.38–0.88, P = 0.011) and coughing (OR: 0.62 95% CI: 0.41–0.95, P = 0.027) were also lower for children in households with soap or soapy water present. MDD was protective for acute respiratory illness, reducing the odds of acute respiratory illness by 38% for children who consumed food from at least five MDD food groups compared with those who consumed food from four or less food groups (95% CI: 0.38–1.00, P = 0.048). A sensitivity analysis was performed with breastfeeding in the model. This did not change the association between acute respiratory illness and dietary diversity (OR: 0.60, 95% CI: 0.37--0.98, P = 0.039). The presence of chickens in the child’s sleeping area was associated with a significant reduction in acute respiratory illness (OR: 0.43, 95% CI: 0.19, 0.99, P = 0.047). There were no other associations seen between animal presence and acute respiratory illness or coughing. There were no significant associations between any risk factors and ALRI.

DISCUSSION

In the REDUCE prospective cohort study conducted in rural eastern DRC, we investigated the association between respiratory illness and the presence of soap on the household compound, animal presence, and dietary diversity. The presence of soap in the cooking area was associated with a decreased odds of acute respiratory illness and coughing among children under 5. We observed that meeting MDD was associated with decreased odds of acute respiratory illness. These findings emphasize the need for interventions targeting hygiene and increasing dietary diversity to reduce respiratory illness among children under 5 years in rural DRC.

In this study, we found a high baseline prevalence of acute respiratory illness (58%) for children under 5 years of age, based on a definition of difficulty breathing, fast breathing, lower chest wall in-drawing or coughing in the past 2 weeks. This finding is similar to the baseline prevalence of acute respiratory illness reported in the control arm of the WASH Benefits trial in Kenya (48%).¹² For ALRI, we found a 14% prevalence among children under 5 years. This is slightly higher than the 10% ALRI prevalence reported for DRC by the Global Health Observatory.²⁷

Through this study, we identified lack of soap in the cooking area to be associated with increased respiratory illness. It is well-established that handwashing with soap is an effective measure to reduce respiratory illness.^5,6 Soap presence in cooking and defecation areas of the home is often used as a proxy measure for handwashing with soap and has been associated with increased handwashing.²⁸ Having soap at a close distance to cooking and defecation areas can help to facilitate handwashing with soap behaviors.²⁹ In their 2018 systematic review of the impact of WASH interventions, Gera et al.¹⁰ identified six studies that assessed the effect of hygiene interventions on acute respiratory infections, estimating a 24% reduction in the risk of developing an acute respiratory infection after hygiene interventions. We were only able to identify one study investigating the association between respiratory illness and presence of soap that was not an intervention study.³⁰ This study, in rural Kenya, did not find an association between home soap presence and reduced acute respiratory illness prevalence.³⁰ The authors noted that the very high proportion of households with soap (97%) may have limited their ability to detect modest changes in respiratory illness. However, the presence of soap and water at household handwashing stations has been associated with lower respiratory illness in a previous study in Bangladesh,³¹ supporting our finding that soap presence in the cooking area is protective against acute respiratory illness. The baseline characteristics of our study population, where the average roundtrip time to collect water was 15 minutes, also highlight the challenges of providing both water and soap for handwashing. Reductions in acute respiratory illness symptoms with soap presence in the cooking area is an important finding. Washing hands before eating, serving, or preparing food are considered key times for handwashing for soap to prevent disease transmission.³² The presence of soap in the cooking area likely facilitated handwashing at these times, thereby reducing the transmission of microbes that cause acute respiratory illness symptoms. In contrast, the lack of the presence of soap in the defecation area was not a significant risk factor for respiratory illness. This was an unexpected finding that should be investigated in future studies. The low proportion of households with soap present in the defecation area (4%) suggests the need for interventions to increase soap availability at this location in the home.

This study is a part of the REDUCE program, which focused on identifying risk factors for morbidity among young children in the DRC, with the goal of developing and evaluating scalable interventions to reduce exposure from these factors. For the REDUCE program, we created theory-driven and evidence-based WASH behavior change communication modules informed by our prospective cohort study findings as well as 2 years of formative research, including 91 semi-structured interviews, six focus group discussions, and a pilot study of 102 households.³³ The developed REDUCE Baby WASH modules targeted reducing pediatric exposure to pathogens and included a module on handwashing with soap. The cost of soap emerged as a major barrier for handwashing with soap during our qualitative formative research, consistent with our quantitative findings that showed very few households had soap present in cooking areas and places of defecation on the household compound. Therefore, as part of the REDUCE Baby WASH modules, a pictorial flipbook on how to prepare soapy water as a low-cost alternative to bar soap was delivered.³³ Handwashing with soapy water is as effective as handwashing with bar soap at reducing microbes on hands.³⁴ Soapy water provision may be an important intervention to facilitate increased and sustained handwashing behavior, as it is less expensive than other soaps. The REDUCE Baby WASH module also included instructions on how to construct a tippy tap to provide a low-cost enabling technology to facilitate handwashing with soap. Currently, the REDUCE Baby WASH module on handwashing with soap has been delivered to over 1 million people in the South Kivu and Tanganyika provinces of DRC. Given the finding that the presence of soap is protective against acute respiratory illness, the module may have the potential to contribute to reductions in acute respiratory illness prevalence among children under 5 in these areas. A future study is needed to evaluate the impact of the REDUCE Baby WASH modules on acute respiratory illness.

Minimum dietary diversity was associated with a reduced odds of acute respiratory illness in this study. Dietary diversity is a known indicator of nutritional adequacy in children.²⁴ Undernutrition predisposes children to infections through weakening immune function.³⁵ Adequate nutrition helps to provide important dietary components (micro and macro nutrients) that are required for immune system functioning. In our study, MDD was likely reflective of nutritional adequacy, leading to higher immune system functioning and less respiratory illness through infections. Our finding adds to the limited literature demonstrating that improved nutrition is associated with reductions in respiratory illness.¹⁵ There is only one other study we are aware of that found a significant association between dietary diversity and respiratory illness.³⁶ This study among HIV-exposed infants in Dar es Salaam, Tanzania, found that low-dietary diversity was associated with increased risk of acute respiratory illness.³⁶ The study, used an Infant and Child Feeding Index composite score that included a dietary diversity component measuring the number of food groups consumed in the previous day (cereals, dairy, flesh foods, and fruits and vegetables). Additional research is needed on dietary diversity and respiratory illness.

In this study, we found that the presence of chickens in a child’s sleeping space was protective against acute respiratory illness. This finding is in contrast to the literature indicating that contact with chickens is associated with increased risk for respiratory illness, however, it aligns with Ercumen et al.,²² who found that children in households owning more poultry had decreased respiratory infections compared with households with less poultry. The relationship between animals and disease outcomes can be confounded by the associations between animal ownership and increased socioeconomic status in low-income households.^20,22 One study in rural Ethiopia reported that although poultry ownership was associated with increased child growth, close exposure was a risk factor for undernutrition.³⁷ Therefore, our positive association between presence of chickens and acute respiratory illness may be because children had chickens came from households with higher socioeconomic status. An alternative explanation for our finding is that the presence of chickens in the sleeping area was less harmful in our rural study setting compared with, for example, an urban setting. In rural settings unlike urban settings, chickens are often free roaming to forage for food. Therefore, they do not spend all day inside the home, decreasing the time spent close to children in their sleeping space.³⁸ Despite our findings, we continue to advocate for interventions that reduce contact between animals and children because of associations we found within this same cohort that contact with rabbits and guinea pigs was associated with linear growth faltering.³⁹

Our study has some limitations. First, we did not use physician-diagnosed respiratory illness. Caregiver-reported symptoms may be subject to recall bias given the 2-week recall period. When clinical confirmation is not feasible, caregiver-reported symptoms are commonly used in epidemiological studies. Second, difficulty breathing, fast breathing, lower chest wall in-drawing and coughing are nonspecific and may be indicative of other conditions, such as allergic airway disease. Coughing, in particular, is associated with air pollution and allergens.⁴⁰ Third, although our analysis was prospective with baseline risk factors and respiratory illness measured at a 6-month follow-up, we included respiratory illness assessed at a single 6-month timepoint rather than investigating the association between risk factors and respiratory illness at several timepoints over the study period. Future studies should include respiratory symptom surveillance at frequent intervals over the study period.

There are several strengths of this study. First, the prospective cohort design used allowed us to examine the impacts of risk factors on respiratory illness prospectively. Second, this study includes measures of animal presence on the household compound, and in indoor cooking and sleeping areas as opposed to relying solely on animal ownership as a proxy for animal contact. Third, the inclusion of the dietary diversity measure allowed us to investigate the association between diversity of the diet and respiratory illness.

In the REDUCE prospective cohort study, the presence of soap in the cooking area and high dietary diversity were associated with decreased acute respiratory illness among young children in rural DRC. The findings presented here contributed to the development of the evidence-based REDUCE Baby WASH modules targeting these exposure pathways, which are currently delivered to over 1 million beneficiaries in the South Kivu and Tanganyika provinces of DRC. Our findings highlight the need to establish effective and inexpensive approaches to improve household access and use of soap to reduce respiratory illness among young children, and the importance of nutrition intervention focused on dietary diversity in the DRC.