Print this page Email this page Users Online: 974 | Click here to view old website
Home About us Editorial Board Search Current Issue Archives Submit Article Author Instructions Contact Us Login 

 Table of Contents  
Year : 2015  |  Volume : 28  |  Issue : 3  |  Page : 181-186

The promise of home visitation by community health workers in rural Kenya: A protective effect that reduces neonatal illness

1 Department of Pediatrics, School of Medicine, University of New Mexico, NM, USA
2 Department of Public Health and Community Medicine, University of New South Wales, New South Wales, Australia
3 Department of Orthopaedic Surgery, School of Medicine, University of Utah, Utah, USA
4 Department of Family and Community Medicine, School of Medicine, Albuquerque, NM, USA
5 Department of Community Health, School of Public Health, University of Nairobi, Nairobi, Kenya

Date of Web Publication11-Mar-2016

Correspondence Address:
Danielle N Mascarenas
UNM School of Medicine, MSC09 5040, 1 University of New Mexico, Albuquerque, NM 87131
Login to access the Email id

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/1357-6283.178600

Rights and Permissions

Background: Nearly three million neonates die each year, largely from preventable illnesses in developing countries. Sub-Saharan Africa has disproportionately high neonatal mortality. Community health worker (CHW) programs are a promising intervention to reduce mortality rates. This research evaluates a CHW newborn home visitation program in rural Kenya by assessing the frequency of health service utilization and overnight hospitalization. Methods: This quasi-experimental study encompassed two consecutive years and compared the rate of medical service utilization for neonatal illness in an intervention group with that of a control group using a household survey. Severity of neonatal illness was assessed by need for overnight hospitalization. Household, maternal, and child characteristics were collected. Results: A total of 489 surveys were completed and showed that 35% of not-visited families reported taking their infant to a healthcare facility, compared with 21% of CHW-visited families (P < 0.01). Rates of overnight hospitalization were 6% for not-visited infants and 1% for visited infants (P < 0.01). Few significant differences were found in household, maternal, and child characteristics. Discussion: This study found that a CHW newborn home visitation program in rural Kenya significantly decreased rates of health service utilization and overnight hospitalization, but did not have the power to detect an effect on neonatal mortality.

Keywords: Community health work, home visitation infant, hospitalization, Kenya, newborn

How to cite this article:
Mascarenas DN, Wurzburger R, Garcia BN, Tomedi A, Mwanthi MA. The promise of home visitation by community health workers in rural Kenya: A protective effect that reduces neonatal illness. Educ Health 2015;28:181-6

How to cite this URL:
Mascarenas DN, Wurzburger R, Garcia BN, Tomedi A, Mwanthi MA. The promise of home visitation by community health workers in rural Kenya: A protective effect that reduces neonatal illness. Educ Health [serial online] 2015 [cited 2022 Aug 17];28:181-6. Available from:

  Background Top

In 2012, there were 6.6 million deaths globally in children under five years of age. Of these, 44% occurred during the neonatal period, the first 28 days of life. Half of these neonatal deaths were within the first 24 hours.[1] The majority of infant deaths occur in developing countries and are secondary to preventable causes such as preterm birth complications, asphyxia, and severe infections like sepsis, pneumonia, and diarrhea.[2]

Between 1990 and 2012, under-five mortality decreased by 47%;[1] however, this trend is not also seen in neonatal deaths. In fact, the proportion of under-five deaths occurring in the neonatal period has increased—from 38.2% in 2000 to 40.3% in 2010.[3] Sub-Saharan Africa has the highest neonatal mortality rate (34 deaths per 1000 live births) worldwide, yet has achieved the slowest reduction in neonatal mortality.[4]

Child intervention programs have focused on treating preventable conditions such as pneumonia, malaria and diarrhea, which are important causes of mortality from two months to five years of age. Neonates, however, require different interventions for survival, including immunizations, proper breastfeeding techniques, good hygiene practices for mother and baby, and prevention of asphyxia. Thus, significant global reductions in neonatal mortality will depend on the prioritization of interventions that specifically target infants in the first four weeks of life.[2]

Inexpensive and feasible interventions can potentially reduce neonatal mortality, but such strategies are rarely found in the places that they would benefit most. Several low-income countries have successfully implemented low-cost interventions for reducing newborn death rates, including widespread access to antenatal and postnatal care, and skilled birth attendance.[5] One such program is a home visitation program, in which health workers visit newborns at home to assess their health status. Studies in Bangladesh, India and Pakistan, all developing countries with high rates of infant mortality, showed that a home visitation program can reduce neonatal death rates by as much as 61%.[6],[7],[8],[9]

Based on the success of these programs, WHO and UNICEF currently recommend at least two home visits for all births, one within 24 hours of the delivery and another on day three.[5] A third visit before the end of the first week of life is ideal. Such home visits promote early and exclusive breastfeeding, encourage skin to skin contact, provide protective measures to keep the baby warm, delay bathing of the infant, increase hygiene by hand washing and umbilical cord care, provide information on signs of newborn illness and counseling on when to seek medical attention. Maternal health is also a component of some home visitation programs, in which mother's well-being, postpartum bleeding and maternal nutrition are assessed. These home visits should be done by skilled health workers, but in areas where there is a lack of healthcare providers, trained community members, called community health workers (CHWs), are used instead.[5] These community health workers are trained to perform basic preventative and curative care, and to assist families in seeking necessary care at a health care facility.[5]

In 2011 a two-day CHW training program was developed and implemented in the Machakos and Kitui counties of southeastern Kenya. The neonatal mortality rate in this region is estimated to be 31 deaths per 1000 live births, compared to the global rate of 22 deaths per 1000 live births.[10],[11] These districts are characterized by limited access to health care due to their geographic isolation and extreme poverty. The training program was based on a training module of the Ministry of Health and Family Welfare, Government of India, UNICEF [12] and a Lancet study [12] of the most sensitive and specific clinical signs that predict severe illness in children under two months of age.

Since the training in 2011, the CHWs have performed newborn home visits in approximately 75 villages. This study is part of an ongoing investigation of the effectiveness of this CHW home visitation program in preventing neonatal mortality. Specifically, it assesses the impact of the program on neonatal healthcare utilization as an indicator of neonatal health status.

  Methods Top

Population and Study Site this study was conducted in southeastern Kenya. CHWs were trained to provide newborn home visits to 75 villages within four adjacent sub-locations of Machakos County and one sub-location of Kitui County. By definition, a sub-location is a cluster of villages within a defined governmental geographic area. Based on estimates from the local officials collecting yearly census data, approximately 900 to 1,000 infants are born per year in the five sub-locations.

CHW training program

In 2007–2008 a non-governmental organization, Global Health Partnerships (GHP), funded a CHW program for the target region of this study. One hundred and fifty CHWs, two for each of the 75 villages of the target area, were chosen by their community based on their education level and willingness to volunteer. The CHWs elected five CHW leaders—one for each sub-location. The training program for these 150 CHWs consisted of a five-day course that was taught by the district Ministry of Health officers and was based on the 'Kenya Essential Package for Health: A Manual for Training Community Health Workers'. The topics covered included methods of malaria control, childhood respiratory tract infections and diarrheal disease, domestic and personal hygiene, child nutrition, reproductive health, HIV/AIDS, and communication skills. The training methodologies consisted of introductory lectures, group discussions, audiovisuals, demonstrations, and role-playing. In 2011, GHP developed and funded another training program for a CHW newborn home visitation project in order to address the high neonatal mortality rate in the area. The CHW leaders selected the four most active CHWs in each of their sub-locations to participate in this new project.

A total of 20 CHWs (five from each sub-location, including the CHW leaders), were trained to recognize the signs and symptoms of neonatal illness on a routine home visit during the first week of life, and to determine whether a referral to a health facility was necessary. Using the training models established by the WHO and UNICEF,[12],[13] the CHW training program consisted of several modules taught over two days that focused on: (1) Instructions for first home visit; (2) assessment of the neonate; (3) addressing feeding problems; (4) providing home care advice for the mother and family; and (5) tips for effective communication. Group discussions throughout the training covered the causes of neonatal illness and mortality within the area. These discussions were important for addressing any misconceptions or questions the CHWs had regarding neonatal illness.[14]

The CHWs were educated on signs of illness in neonates, including lethargy, low body temperature, and fever. CHWs were taught to assess for possible bacterial infections by asking about convulsions, obtaining a respiratory rate, looking for severe chest retractions and nasal flaring, listening for grunting, assessing the umbilicus for redness or draining pus, looking for skin pustules, recognizing lethargy and unconsciousness, taking a temperature, and noting abnormal movements for an infant. Photographs were used to demonstrate a normal versus abnormal umbilicus and skin pustules. The CHWs were also trained in the steps for referring a child to a healthcare facility, which included explaining the need for treatment to the mother and family, calming fears, solving problems such as transportation, and writing a referral card for the mother to take with her to the hospital.[14]

In regard to feeding issues, CHWs were trained to ask mothers if there were any difficulties with feeding, how often they breastfed in 24 hours, and if it was via breast, bottle, or cup. They also learned to ask about any other foods or drinks the infant received, pain with breastfeeding and signs of proper latching. CHWs were trained to look for sore nipples, engorged breasts and breast abscesses. Videos and photographs were provided to teach the CHWs the correct form of latching and signs of infection in the breast.[14] CHWs were also instructed on giving advice to the family about home care. This module included umbilical care, proper hygiene techniques, keeping the baby warm, and written precautions. Finally, CHWs were also given tips on how to effectively communicate with the families, such as using understandable language, asking open-ended questions about well-being of mother and baby and checking for understanding with questions.[14] There was no treatment component to the program. A copy of the training manual can be provided upon request.

Sample size

The target sample size was 268 in the intervention group and 114 in the control group, based on the allotted 2-week window for data collection and previous survey completion rates in the region. This sample size could detect a 6.67% difference in hospital visits between infant groups, with 5% significance level and 80% power.

Study design

This was a quasi-experimental study that compared the rate of healthcare utilization for neonatal illness in an intervention group with the rate of healthcare utilization in a control group. The use of medical services was measured using a household survey (available upon request). Severity of neonatal illness was assessed by the need for overnight hospitalization.

There were no predetermined factors or instructions to the CHWs to determine which infants would or would not be included in the home visitation program. CHWs were notified of a pregnancy or delivery by the families or birth attendants. CHWs attempted to keep track of all pregnant women and births in their area, so that they could visit all infants in the target region in the first days of life.

The CHWs visited 1,399 newborn infants over a two-year period from March 1, 2011 to February 28, 2013. This group formed our experimental group, and a list of these infants was maintained by the program coordinator, who resided in the study intervention region. Of those 1,399 CHW-visited infants, a random sample of 400 infants was selected for surveying by using Microsoft Excel 2013 to generate a random list of numbers, with each number corresponding to an infant. The number 400 was chosen based on the predicted difficulties in completing surveys in the catchment area (relocation, mothers not at home, etc.), to meet the minimum desired sample size of 268 in the intervention group.

Some infants were not visited by CHWs because their mothers received prenatal care elsewhere, moved into the village after giving birth, the CHW was not aware of the pregnancy, or the CHW was notified of the birth beyond the first week of life. These not-visited infants comprised the control population. The control group population was found and recruited in a door-to-door household census conducted for two consecutive years in all villages of the catchment area of the project. From that census a list of all children under 12 months of age was compiled (starting in March 2011 when the CHW newborn home visits began). The CHW-visited children were removed from this census list, leaving a total of approximately 300 not-visited infants that qualified for the control group. While the intention was to get an equally matched 400 children for the control group, the census did not identify this many not-visited infants, therefore, all of these children were included as controls. The exact number of control infants was not obtained.

Data collection

Over a two-week period in March of 2012, and another two-week period in March of 2013, an investigator with a native KiKamba interpreter, or a native KiKamba-speaking investigator, administered the survey to the mother or caretaker of the infantsin the intervention and control groups. The households were located with the assistance of a village guide, usually the local CHW. Data was collected using a tablet computer. The survey included questions on the incidence of illness and utilization of healthcare facilities, the frequency of overnight hospital stays, household characteristics and maternal characteristics. The survey questions can be provided upon request.

Data analysis

Data analysis was completed using SAS, version 9.3, software (Statistical Analysis System (SAS) Version 9.3. SAS Institute Inc., Cary, North Carolina, USA). Fisher's exact test, Student's t-test, and Wilcoxon test were used to compare maternal, child and household demographic variables between the control and intervention groups. Fisher's exact tests were also used to compare proportions of infants taken to a health facility and proportions of infants hospitalized overnight.

Ethics approval

This study was approved by the affiliated universities' human research protection offices and committees.

  Results Top

Between March 1, 2011 and February 28, 2013, the 20 CHWs visited 1,399 newborn infants. A total of 1,336 of these infants (95%) had three visits, and 1,384 infants (99%) had two visits or more. Twenty-eight infants were referred to a health facility for treatment. There were 12 neonatal deaths during the first week of life.

Of the approximately 300 control infants and the randomly selected 400 infants from the intervention group, the household survey interviews were successfully conducted for 326 neonates in the intervention group and 176 neonates in the control group, for a total of 502 surveys. Of these 502 surveys, 13 were dropped because the surveys were not filled out completely, leaving a total of 489 completed surveys for analysis.

Household, maternal, and child characteristics

Maternal and child characteristics are shown in [Table 1]. More CHW-visited families were in the poorest quintiles (43%) than the not-visited families (34%), however this difference was not statistically significant (P = 0.16). Other differences that did not reach statistical significance included number of children who were born alive but later died, weight of the youngest child and distance from the family's house to the house of the nearest CHW. The CHW-visited infants were more likely to have mothers with fewer years of education, with 59% of CHW-visited mothers having less than 9 years of schooling compared to 51% of not-visited mothers (P < 0.01). Other significant differences between CHW-visited and not-visited infants included mother's age in years (P = 0.01) and mother's total number of births (P < 0.01) [Table 1].
Table 1: Comparison of selected baseline characteristics of the mothers and infants in both the community health worker -visited and not-visited groups

Click here to view

Healthcare utilization

As shown in [Table 2], a significant difference in health care utilization was found between CHW-visited and not-visited families. Among those families not visited, 35% reported taking their youngest child to a health care facility, compared with 21% of families visited by a CHW (P < 0.01). The rates of overnight hospitalization reported were also significantly higher among not-visited infants compared to those visited (6% vs. 1%, respectively; P < 0.01). Of the infants who were brought to a health facility, 18% of not-visited infants required overnight hospitalization compared to 6% of visited infants (P = 0.05) [Table 2].
Table 2: Comparison of rates of healthcare utilization and overnight hospitalization between community health worker -visited and not-visited infants

Click here to view

  Discussion Top

This study found that a significantly lower number of infants enrolled in the CHW home visitation program were taken to a health facility and hospitalized for their illness as compared to infants not visited by a CHW. It is possible that because CHWs educated families to recognize illnesses that require medical attention, visited families may be better able to differentiate between severe and mild illness. Without the guidance of CHWs, not-visited families may seek care for less serious infant conditions more often than visited families. However, this would not account for increased rates of overnight hospitalization among not-visited infants. A more plausible explanation is that visitation by a CHW actually has a protective effect for the infant, reducing the rates of illness that would require healthcare utilization and overnight hospitalization. The interventions CHWs are trained to teach, such as hygiene and breastfeeding, may actually be producing the desired behavioral changes among visited families that lead to lower rates of infection, and thus lower rates of neonatal health facility utilization and hospitalization.

While the CHWs have been very successful in carrying out the goals of the CHW training program in assessing neonates, it appears they are also incorporating knowledge from the initial CHW training done in 2007-2008, as described in the Methods section. In a meeting with the CHWs in July 2013, it became apparent that the CHWs have actually been counseling families beyond what they learned in the 2-day training course. Not only were they performing the required tasks at home-visits, but they also discouraged specific common but harmful practices in the local communities, such as the application of ash or soot to the umbilical stump, or feeding water to the infant instead of exclusive breast feeding. Further, the CHWs reported that they encouraged women to attend prenatal care clinics, to go to a health facility for delivery and to avoid medicines and drugs unless prescribed by a doctor.

This study has several important limitations. As described, there was no randomization process for determining which households were originally chosen for visitation by the CHW. The CHWs attempted to visit all children, but we can only speculate as to why some were visited while others were not. Therefore, the possibility that the differences found between the intervention and control groups were an artifact of some undetermined household characteristic, selection bias or confounding variables, cannot be excluded. Another difficulty in our data interpretation was that the exact number of children in the control group was not known and was estimated to be 300. The actual number of not-visited infants in the region was not known and some were likely missed by the census. Therefore, the survey response rate for the control group could not be calculated for comparison with the intervention group, where a possible difference may exist.

Another problem encountered was an inability to locate all of the infants selected for each sample. An attempt was made to visit every household from the randomly selected lists, but the investigators had only a two-week window to obtain surveys and many of the families were not at home to be surveyed or were reported by neighbors or other family to have moved away from the village. Lastly, there were 13 surveys dropped from the analysis because they were not complete. It is unknown if there was something different about this cohort of infants, or if their responses could have influenced the results of our data analysis.

In regard to confounding variables, birth weight is an important factor in predicting the health status and probability of survival of an infant. Low birth weight is associated with increased morbidity and mortality.[5] Accurate birth weight data were not available for this study population, but using the survey data available, no significant difference was found in the prevalence of low birth weight between visited and not-visited infants. Another potentially important difference in the two populations is that a greater proportion of the visited families were from poorer quintiles. Since the neonates of the poorest families may have a higher risk of morbidity and mortality,[2] this difference would likely favor the outcome of the not-visited group. Further, the mothers of the visited infants had significantly less education than the not-visited group. Maternal education has been found to be an important predictor of child survival,[15] so this difference between the groups would also likely favor the not-visited infants. Despite these differences in wealth index and maternal characteristics, which one would expect to favor the not-visited infants, CHW-visited infants had significantly less healthcare utilization and hospitalizations. This further supports the theory that CHW visitation has a protective effect for neonates in the area.

No other important differences were found in baseline characteristics, but there remains the possibility of the influence of unmeasured or unknown confounders. In fact, an important part of this study was to determine what others factors could have contributed to the differences in the visited and not-visited groups and the use of healthcare facilities because very little data about the not-visited group had been obtained previously.

This study found that a CHW newborn home visitation program was able to significantly decrease rates of healthcare facility utilization and need for hospitalization, but it was not designed, and did not have the power, to detect an effect on neonatal mortality rate (NMR). However, the small number of early (less than seven days) neonatal deaths suggests that a reduction may have been achieved. Of the 1,399 visited infants there were 12 deaths during the first week of life, a crude early neonatal mortality rate (ENMR) of 8.6 per 1,000 live births. In Kenya the neonatal mortality is 31 deaths per 1,000 live births,[11] and an estimated 85% of these neonatal deaths occur in the first seven days,[16] giving an estimated ENMR of 26 deaths per 1,000 live births. An estimated 20% of the 31 neonatal deaths are from asphyxia [2] and likely occur within the first few hours after birth, and therefore may not be reported to the CHW as a live birth. If these deaths were excluded, the “expected” ENMR in our intervention population would be 20 per 1,000 live births, far higher than the observed rate of 8.6 per 1,000.

  Conclusion Top

This conclusion is based on a historical comparison and a small number of neonatal deaths, and therefore is speculative. However, the interventions taught are certainly beneficial from a medical standpoint, and the CHWs have demonstrated significant investment in the program and in improving neonatal survival for their communities. Further research will help determine the true efficacy of this intervention towards the ultimate goal of reducing neonatal mortality, but this study is an indication of the promise the CHW home visitation program holds.


The authors would like to thank Betty Skipper and Cheryl Schmitt for their assistance with the statistical analysis, Nicholas Mutuku, Vincent Makau, Holly Maloy and Adriana Dhawan for their invaluable help in data collection, and the community health workers of the Kisesini Community Health Project, without whom this project would not be possible.

Financial support and sponsorship

The funding for this study was provided by the University of New Mexico.

Conflicts of interest

There are no conflicts of interest.

  References Top

Wardlaw T, You D, Newby H, Anthony D, Chopra M. Child survival: A message of hope but a call for renewed commitment in UNICEF report. Reprod Health 2013;10:64.  Back to cited text no. 1
Lawn JE, Cousens S, Zupan J; Lancet Neonatal Survival Steering Team. 4 million neonatal deaths: When? Where? Why? Lancet 2005;365:891-900.  Back to cited text no. 2
Liu L, Johnson HL, Cousens S, Perin J, Scott S, Lawn JE, et al. Global, regional, and national causes of child mortality: An updated systematic analysis for 2010 with time trends since 2000. Lancet 2012;379:2151-61.  Back to cited text no. 3
Blencowe H, Cousens S. Addressing the challenge of neonatal mortality. Trop Med Int Health 2013;18:303-12.  Back to cited text no. 4
World Health Organization and United Nations International Children's Emergency Fund. Home Visits for the Newborn Child: A Strategy to Improve Survival. Geneva: WHO; 2009. Available from: . [Last accessed on 2013 Dec 27].  Back to cited text no. 5
Baqui AH, El-Arifeen S, Darmstadt GL, Ahmed S, Williams EK, Seraji HR, et al. Effect of community-based newborn-care intervention package implemented through two service-delivery strategies in Sylhet district, Bangladesh: A cluster-randomised controlled trial. Lancet 2008;371:1936-44.  Back to cited text no. 6
Kumar V, Mohanty S, Kumar A, Misra RP, Santosham M, Awasthi S, et al. Effect of community-based behaviour change management on neonatal mortality in Shivgarh, Uttar Pradesh, India: A cluster-randomised controlled trial. Lancet 2008;372:1151-62.  Back to cited text no. 7
Bhutta ZA, Memon ZA, Soofi S, Salat MS, Cousens S, Martines J. Implementing community-based perinatal care: Results from a pilot study in rural Pakistan. Bull World Health Organ 2008;86:452-9.  Back to cited text no. 8
Bang AT, Bang RA, Baitule SB, Reddy MH, Deshmukh MD. Effect of home-based neonatal care and management of sepsis on neonatal mortality: Field trial in rural India. Lancet 1999;354:1955-61.  Back to cited text no. 9
The Inter-agency Group for Child Mortality Estimation. Levels and Trends in Child Mortality Report. New York; 2012. Available from: . [Last accessed on 2013 Dec 27].  Back to cited text no. 10
Kenya National Bureau of Statistics. Kenya Demographic and Health Survey 2008-09. Calverton, Maryland; 2010. Available from: . [Last accessed on 2013 Dec 27].  Back to cited text no. 11
Young Infants Clinical Signs Study Group. Clinical signs that predict severe illness in children under age 2 months: A multicentre study. Lancet 2008;371:135-42.  Back to cited text no. 12
Government of India, Ministry of Health and Family Welfare. IMNCI training module for workers. India: Ministry of Health and Family Welfare, Government of India; 2003.  Back to cited text no. 13
Global Health Partnerships and Kenya Ministry of Public Health and Sanitation. Home Visitation for Newborns: Training Module for Community Health Workers. Yatta District Health Management Team; 2011.  Back to cited text no. 14
Bhutta ZA, Darmstadt GL, Hasan BS, Haws RA. Community-based interventions for improving perinatal and neonatal health outcomes in developing countries: A review of the evidence. Pediatrics 2005;115 2 Suppl: 519-617.  Back to cited text no. 15
Hill K, Choi Y. Neonatal mortality in the developing world. Demogr Res 2006;14:429-52.  Back to cited text no. 16


  [Table 1], [Table 2]

This article has been cited by
1 Training programs to improve identification of sick newborns and care-seeking from a health facility in low- and middle-income countries: a scoping review
Alastair Fung, Elisabeth Hamilton, Elsabé Du Plessis, Nicole Askin, Lisa Avery, Maryanne Crockett
BMC Pregnancy and Childbirth. 2021; 21(1)
[Pubmed] | [DOI]
2 Sisters in Birth: Improving Birth Outcomes in Mississippi
Getty Israel
The Journal of Perinatal Education. 2020; 29(4): 168
[Pubmed] | [DOI]
3 Knowledge of low birth weight care as a source of coping strategies for mothers: Cross sectional study in perinatology WARDæS
Eviana S. Tambunan,Hadi Pratomo,Ella Nurlaella Hadi,Yeni Rustina
Journal of Neonatal Nursing. 2020;
[Pubmed] | [DOI]


Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

  In this article
Article Tables

 Article Access Statistics
    PDF Downloaded368    
    Comments [Add]    
    Cited by others 3    

Recommend this journal