SEARCH

SEARCH BY CITATION

Keywords:

  • neonatal;
  • mortality;
  • demographic;
  • surveillance;
  • Kassena–Nankana;
  • Ghana
  • néonatal;
  • mortalité;
  • démographique;
  • surveillance;
  • Kassena-Nankana;
  • Ghana
  • neonatal;
  • mortalidad;
  • seguimiento demográfico;
  • Kassena-Nankana;
  • Ghana

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Kassena–Nankana district
  5. Methodology
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Objectives  To describe the trend and causes of neonatal deaths in a rural district in northern Ghana.

Methods  Descriptive analysis of data collected from the Navrongo Demographic Surveillance System and verbal autopsies conducted on all neonatal deaths from 1995–2002.

Results  Of 1118 recorded neonatal deaths 1068 (95.5%) could be analysed. Only 13.2% of deaths occurred at the health facility; 62.7% occurred in the early neonatal period, with prematurity (38%) and birth injuries (19%) as leading causes. Infectious causes (66%) were the major contributors to late neonatal deaths. Infanticide accounted for 4.9% of all neonatal deaths. The cause-specific mortality rate for neonatal tetanus remained under 2.5% throughout the 8-year period. Overall, the neonatal mortality rate declined at an average of 2.5 per 1000 live births per year: Down by nearly 50% from 40.9 (95%C.I. 34.1–46.8) in 1995 to 20.5 (95%C.I.17.3–22.7) in 2002.

Conclusion  The various health interventions undertaken in this district have had the collateral effect of causing decline in neonatal mortality. Neonatal mortality could be further reduced by preventing and treating neonatal infections, having skilled attendance at delivery and the elimination of infanticide. Data from demographic surveillance sites may be useful in monitoring trends in child mortality.

Objectifs  Décrire les tendances et causes de décès néonataux dans un district rural du nord du Ghana.

Méthodes  Analyse descriptive de données collectées dans le Système de Surveillance Démographique de Navrongo et d'autopsies verbales menées sur tout décès néonatal de 1995 à 2002

Résultat  Sur 1118 décès néonataux enregistrés, 1068 (95,5%) ont pu être analysés. Seuls 13,2% des décès ont eu lieu dans un service de santé; 62,7% ont eu lieu dans la période juste après la naissance dont 38 % de prématurés et 19% par traumatisme de naissance comme cause majeures de décès. Les causes de décès par infection (66%) constituaient la majorité des décès néonataux tardives. L'infanticide comptait pour 4,9% de tous les décès néonataux. Le taux de mortalité cause-spécifique pour le tétanos néonatal restait en dessous de 2,5% durant toute la période de l’étude de 8 ans. Au total, la mortalité néonatale a décliné en moyenne de 2,5 pour 1000 vie naissances par an; ce qui correspond à presque 50% de dimunition passant de 40,09 (IC95%: 34,1–46,8) en 1995 à 20,5 (IC95%: 17,3–22,7) en 2002.

Conclusion  Les différentes interventions de santés entreprises dans ce district ont eu des effets collatéraux qui ont entraîné le déclin de la mortalité néonatale. La mortalité néonatale pourrait encore plus être réduite par la prévention et le traitement des infections néonatales, l'assistance qualifiée à l'accouchement et par l’élimination de l'infanticide. Les données des sites de la surveillance démographique peuvent être utilisées pour monitorer les tendances de la mortalité infantile.

Objetivos  Describir las tendencias y causas de las muertes neonatales en un distrito rural del norte de Ghana.

Métodos  Análisis descriptivo de los datos recolectados del sistema de seguimiento demográfico de Navrongo y de autopsias verbales realizadas para todas las muertes neonatales entre 1995–2002.

Resultados  De 1118 muertes neonatales registradas, 1068 (95.5%) pudieron ser analizadas. Solo 13.2% de las muertes ocurrieron en un centro sanitario; 62.7% ocurrieron durante un periodo neonatal temprano, siendo la prematuridad (38%) y las lesiones en el parto (19%) las causas principales. Las causas infecciosas (66%) fueron los principales contribuyentes a las muertes neonatales tardías. El infanticidio explica un 4,9% de las muertes neonatales. La tasa de mortalidad causa-específica para el tétanos neonatal se mantuvo debajo del 2.5% durante los 8 años. En general, la mortalidad neonatal disminuyó en promedio un 2.5 por 1000 nacidos vivos por año: casi un 50% de 40.9 (95% I.C 34.1-46.8) en 1995 a 20.5 (95% I.C 17.3-22.7) en 2002.

Conclusión  Las diferentes intervenciones en salud realizadas en este distrito han tenido el efecto colateral de disminuir la mortalidad neonatal. Dicha mortalidad podría reducirse aún más previniendo y tratando las infecciones neonatales, teniendo una atención cualificada durante el parto y eliminando el infanticidio. Los datos de lugares con seguimiento demográfico pueden ser útiles para monitorizar las tendencias en mortalidad infantil.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Kassena–Nankana district
  5. Methodology
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Achieving the millennium development goal of reducing under-five mortality by two-thirds between 1990 and 2015 would only be possible through significant reduction in neonatal deaths in sub-Saharan Africa. Four million children die within the first 28 days of life, and almost two-thirds die in the first week. Neonatal deaths in sub-Saharan Africa account for approximately one quarter of all deaths in children aged 1–3 months (Lawn et al. 2005; Zupan & Aahman 2005). Generally, the neonatal mortality rate (NMR) in sub-Saharan Africa is estimated to range from 20 to 40 per 1000 live births (Greenwood 2003). There is little or no information on the causes of death of neonates and early infants in many developing countries as the majority of these deaths occur at home, and in infants who have not been seen in the formal health sector during their final illness (Greenwood 2003). The breakdown in vital registry further ensures that neither births nor deaths occurring outside of the formal health sector are recorded (Cooper et al. 1998; Child Health Research Project 1999; Yu 2003; Lawn et al. 2005). The result of this tragic state of affairs is a near absence of interventions specifically aimed at reducing neonatal mortality. So acute is the problem that many agencies do not mention any causes of neonatal death on the list of major ‘child killers’. By making available the epidemiological information regarding the time, place and causes of neonatal deaths, it is hoped that greater visibility will be given to neonatal deaths in policies and programme planning (Lawn et al. 2004).

Vital registrations systems in the developing world are unlikely to see dramatic improvements in the near future. The verbal autopsy (VA) technique, which involves questioning the family of a dead infant about the features of the child's final illness and supplementing this account with any clinic or hospital records that are available provides the only means of obtaining information on the cause of death (COD) (Snow et al. 1992; Taylor 1992; Greenwood 2003). VA continues to be used extensively in large-scale demographic and health surveys, community-based surveillance and community intervention trials. Most sites that maintain demographic surveillance system (DSSs) routinely conduct VA. By definition they also routinely collect data on vital events such as pregnancies, deliveries and deaths. Potentially therefore, these sites could provide fairly accurate information on the magnitude and causes of neonatal deaths (Anker et al. 2001; Kynast-Wolf et al. 2002; Black et al. 2003).

Kassena–Nankana district

  1. Top of page
  2. Summary
  3. Introduction
  4. Kassena–Nankana district
  5. Methodology
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Kassena–Nankana district (KND) in northern Ghana covers a land area of 1685 km2 and has an estimated population of 141 000. Most parts of KND are rural. It is only the small central area, Navrongo (the district capital) that has sub-urban character. Majority of the inhabitants are subsistence farmers who live in small, scattered settlements. The educational level of people in the district is very low, with over 70% being illiterate. Most of the people in KND hold firmly to traditional beliefs and customs. Women in the district enjoy little autonomy. Infanticide, locally referred to as ‘chichuru’ is a customary practice among some communities within the district. The practice originates from the belief in the ‘spirit child’ phenomenon – that some children are born with supernatural powers that could be harmful to the family into which they are born and this could only be averted by killing such children (Sena 1993; Allotey & Reidpath 2001).

The health system in KND is modelled along the district health care concept. There are four health centres strategically located in the four geographical cardinals and a district referral hospital in Navrongo. Medical assistants and community health nurses are in charge of the health centres. These centres offer antenatal, delivery and child welfare services. These are conducted through static and outreach clinics. Malaria, diarrhoea, acute respiratory tract infections and meningitis are the leading causes of child morbidity and mortality in KND. The HIV prevalence is estimated at 4.4% (National HIV/AIDS Control Program 2004). Total fertility rate is estimated at 3.9. Only about 20% of births are attended to by skilled attendants (Navrongo Health Research Center 2004).

The Navrongo Health Research Centre has maintained a DSS and used VA in the KND for the past 11 years. The system involves a 90-day cycle of recording vital events such as births, deaths, migrations, marriages and pregnancies. Trained community key informants, who are resident in all communities, pick up and record vital events (births, deaths, pregnancy and migrations) that occur in-between the census rounds.

Verbal autopsies are conducted on all deaths in the district. Questionnaires specifically designed by the Navrongo DSS are used for neonatal, post-neonatal and adult deaths (available on request). All compounds within the district have been uniquely and conspicuously marked to make for easy identification (Binka et al. 1996). Until the introduction of the standard INDEPTH tool in 2003, information on stillbirths was not routinely collected as part of the DSS.

This paper presents a review of the trend and causes of neonatal deaths in KND from 1995 to 2002. It uses data collected through the Navrongo DSS and the VA conducted on deaths.

Methodology

  1. Top of page
  2. Summary
  3. Introduction
  4. Kassena–Nankana district
  5. Methodology
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

We report an analysis of data collected with the neonatal VA questionnaires for all deaths in the district from 1995 to 2002. The questionnaire had both open and closed-ended questions and included portions that provided for respondent's verbatim account of the circumstances leading to the death of the child. The VA interviews were conducted on an average of 6 months after death had occurred. Trained fieldworkers and their supervisors, who had a minimum of high-school qualification, conducted the interviews. Immediate caregivers (most often mothers) were the primary target respondents. Ten per cent of interviews were re-conducted for quality assurance and where marked discrepancies were detected, interviews conducted by particular fieldworkers were re-conducted. Fieldworkers supplemented information obtained from the respondents with information in hospital or patients records, if available. Dates of births and deaths were ascertained with the aid of information available in the DSS database, facility records and a vital events calendar. Three physicians independently reviewed the VA forms and a diagnosis was established if at least two of them agreed on the principal COD. Where, however, all three disagreed, the form was submitted to two physicians who, sitting together, discussed the available information, and arrived at a single principal COD. Where there was enough VA information, but no principal COD could be agreed upon, the case was declared undetermined. When little or no information was available to enable an assignment of COD, the diagnosis was taken to be unknown. All physicians used the same locally developed COD list and coding was done directly to this list. Conditions indicated on the COD list were informed by local knowledge of common diseases in the district. Infanticide is specifically indicated in the list. The practices and procedures in the Navrongo DSS were essentially the same throughout the 8-year review period from 1995 to 2002.

Neonatal death was defined as death of a live-born child occurring within 28 days of life. Early neonatal death referred to death occurring within 7 days of life while death occurring on or after 7 days but before 28 days was referred to as late neonatal deaths. Neonatal mortality rates were calculated as risks, i.e. the number of neonatal deaths, divided by the number of live births in the given year, expressed per 1000 live births. Cause-specific mortality rates were similarly calculated by dividing the number of deaths due to a particular cause by the total number of live births in the given year, expressed per 1000 live births. Epi-Info version 6.01 was used to run frequencies for events such as place of delivery, attendant at delivery, material used in dressing the umbilical cord etc. During analysis, all deaths determined to be of infectious origin were summed together as infectious causes. These included respiratory tract infections, diarrhoea, neonatal sepsis, meningitis and malaria. The five leading causes of mortality were analysed to show trends over the 8-year period and their percentage contribution to early and late neonatal deaths. The proportion of deaths due to neonatal tetanus, a vaccine-preventable disease for which VA has high specificity, was also computed.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Kassena–Nankana district
  5. Methodology
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

Over the 8-year period, 1118 neonatal deaths were recorded. Fifty cases were dropped during data cleaning due to date-related discrepancies. Analysis was therefore, based on 1068 (95.5%) ascertained neonatal deaths. The majority (86.1%) of deaths occurred at home, with only 141(13.2%) occurring at the health facility (and could possibly have been captured in the vital registry of the district). There were significantly more male (56.4%; C.I. 53.4–59.4%) than female (43.6%; C.I. 40.6–46.6%) deaths. Most (62.7%) deaths occurred in the early neonatal period. Prematurity (38%) and birth injuries (19%) were the major causes of early neonatal deaths (Figure 1) while infectious causes (66%) were the major contributors to late neonatal deaths (Figure 2). Overall, the four leading causes of neonatal deaths were infections (39.2%), prematurity (26.0%), birth injuries (14.4%) and neonatal infanticides (4.9%) (Table 1).

image

Figure 1. Causes of early neonatal deaths in Kassena-Nankana district, 1995–2002 (n = 670).

Download figure to PowerPoint

image

Figure 2. Causes of late neonatal deaths in Kassena-Nankana district, 1995–2002 (n = 398).

Download figure to PowerPoint

Table 1.  Number, rates and leading causes of neonatal deaths in the KND of northern Ghana as determined by VA 1995–2002.
 YearTotal (%*)
19951996199719981999200020012002
  1. * Percentage of deaths.

  2. ** NMR – Neonatal mortality rate.

  3. *** CSNMR – Cause-specific neonatal mortality rate.

Number of live Births4350427343413868413644114057400333439
Number of deaths (NMR**)178 (40.9)186 (43.5)132 (30.4)128 (33.1)98 (23.7)128 (29.0)136 (33.5)82 (20.5)1068
Infective causes (CSNMR***)61 (14.3)77 (18.0)52 (12.0)51 (13.2)38 (9.2)52 (11.8)46 (11.3)42 (10.5)419 (39.2%)
Prematurity (CSNMR)43 (10.3)43 (10.1)33 (7.6)32 (8.3)25 (6.0)38 (8.6)49 (12.1)15 (3.7)278 (26.0%)
Birth injury (CSNMR)27 (6.2)30 (6.6)24 (5.5)19 (4.9)16 (3.9)17 (3.9)13 (3.2)8 (2.0)154 (14.4%)
Infanticide (CSNMR)12 (2.8)4 (1.4)6 (1.4)7 (1.8)8 (1.9)4 (0.91)10 (2.5)2 (0.5)53 (4.9%)
Other (CSNMR)13 (3.0)9 (2.1)14 (3.2)9 (2.3)11 (2.7)16 (3.6)17 (4.2)8 (2.0)97 (9.1%)
Unknown/Undetermined2223310011767 (6.3%)

All four leading causes of mortality showed a modest decline in cause-specific mortality rates from 1995 to 2002. The greatest decline was recorded in the proportion of deaths attributable to infections (average drop of 0.8/1000 live births per year), birth injuries (0.6) and prematurity (0.4). A sharp increase in the proportion of deaths due to prematurity was recorded in 2001 (Figure 3). The cause-specific mortality rate for neonatal tetanus (a vaccine-preventable disease) remained under 2.5% throughout the 8-year period. The highest of 2.3% was recorded in 1995. High levels of home delivery (78%), unskilled attendants at delivery, (62%) and low-birth weight (30%) characterized the history of the cases of neonatal deaths over the review period. Overall, the NMR declined at an average of 2.5 per 1000 live births per year, down by nearly 50% from 40.9 (95%C.I. 34.1–46.8) in 1995 to 20.5 (95%C.I.17.3–22.7) in 2002 (Figure 3).

image

Figure 3. Trends in overall and cause-specific neonatal mortality rates in the Kassena-Nankana district 1995–2002 (n = 1068).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Kassena–Nankana district
  5. Methodology
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

The lack of data on neonatal mortality in sub-Saharan Africa has made country-specific trends in neonatal mortality difficult to ascertain. Wide variations in reported rates exist. Figures ranging from 13 per 1000 live births in Kenya to 108 in Senegal have been documented (Child Health Research Project 1999; Hyder et al. 2003). In the Upper River Division of The Gambia, NMR was estimated at 39 per 1000 live births in 1994 (Leach et al. 1999).

The design of the Navrongo DSS, i.e. 3-monthly census cycle, use of community-based key informants to pick up vital events in between census rounds and in-built quality checks, go far to ensure that very few vital events are missed (Binka et al. 1996). However, in view of the fact that the DSS was, until 2003, not routinely recording stillbirths, and some mothers might have reported early neonatal deaths as stillbirths (Kramer et al. 2002), we are inclined to consider our estimates of NMRs as probably underestimated. However, we expect this bias to be consistent across the years and not substantially affect the trend. We therefore, consider our findings as a fair representation of the trend in neonatal mortality in this district over the 8-year old period. A review of the available literature on population-based estimates of NMR suggests that the extent of decline observed in this district is exceptional. It is akin to the dramatic reduction in neonatal (44 to 15 per 1000 live births), infant (162 to 36) and under-five mortality rates (397 to 66) reported of three villages in The Gambia that were monitored over a 50-year period (Rayco-Solon et al. 2004).

The Navrongo DSS has been a major attraction to epidemiologists and other researchers wishing to conduct community-based intervention trials. Diverse social and biomedical studies have been undertaken in this district. The include the Vitamin A Supplementation Trial (VAST) from 1988 to 1992 (Ghana VAST Study Team 1993) the Insecticide-Treated Bednets Trial from 1993 to 1995 (Binka et al. 1996) and the Community Health and Family Planning Project from 1995 to 2004 (Binka et al. 1995). These and other community-based interventional studies have invariably been accompanied by numerous health education campaigns, the effects of which appears to be diverse. It includes improved access to basic health services, high levels of antenatal attendance, use of impregnated bed-nets and antitetanus immunization coverage for pregnant women. A decline of 15% (representing one birth) in fertility has been reported as an effect of the community health and family planning project (Debpuur et al. 2002). Given the fact that no intervention has specifically targeted reduction in neonatal mortality, we believe that the observed decline in NMR is a collateral effect of the multiple health research activities undertaken in the district over the years. It would be of importance to examine in detail specific factors that have contributed to decline in NMR in KND and the three villages in the Gambia (referred to above). Among others factors, the effect of increased birth intervals, improved access to health facilities etc would need to be examined. Confounding factors would include the diversity in study designs, time periods and coverage of the various interventions.

A major barrier to action on neonatal health has been the erroneous perception that only expensive, high-level technology and facility-based care can reduce mortality (Darmstadt et al. 2005; Knippenberg et al. 2005). The observed decline in neonatal mortality in KND and the three villages in The Gambia (Rayco-Solon et al. 2004) are suggestive of the fact that even in resource-constrained settings, high neonatal mortality rates can be reduced.

Only few studies have validated the use of VA to ascertain the causes of neonatal deaths (Kalter et al. 1999; Marsh et al. 2003; Freeman et al. 2005). The findings of these studies have been generally favourable. VA have been found to be able to identify several of the most common causes of neonatal illness and death (Kalter et al. 1999; Marsh et al. 2003) although generalisability of this position is limited by differences in the methods and procedures used in different settings (Freeman et al. 2005). For similar reasons, the conclusion that neonatal VA describes well the group cause-specific proportionate mortality and individual misclassifications ultimately cancel themselves out (Marsh et al. 2003) cannot be generalized. In the classification of causes of death ascertained using VA, it is usually possible to differentiate deaths resulting from obstetric problems, prematurity or an infection, although differentiating between the various forms of the latter is often difficult (Greenwood 2003). Although VA are highly dependent on the techniques and tools used, thus making validation studies very important, discussions about their advantages and limitations need to be balanced with the prevailing reality. VA is just about the only available tool for ascertaining the causes of neonatal deaths in virtually all of the communities that contribute 99% of worldwide neonatal deaths. To the extent that procedures and methods for the conduct of VA vary from site to site, findings from VA need to be interpreted within the context of the methods and procedures adopted, and comparisons across studies made with caution.

The finding that about 40% of all neonatal deaths are due to infectious causes compares with 57% in rural Gambia (Leach et al. 1999) and 45% in rural Tanzania (Hinderaker et al. 2003). The trends in the observed case-specific NMRs (Figure 3) suggest that, the reduction in infections causes must have contributed greatly to the over decline in NMR in the district. We are unable explain fully the sudden increase and decrease in cause-specific mortality rate for prematurity in 2001. Researchers, nurse and community health workers in the district attribute it to a strike action by nurses and community health workers in 2001. However, the general trend of decline is little affected and not put in doubt. It appears that further reduction in overall NMR will require interventions aimed at providing prompt and adequate management of neonatal infections and reduction in cases of prematurity. Home-based neonatal care for infections has been advocated, and there is evidence to suggest the home management of neonatal sepsis is feasible, acceptable and could substantially contribute to reducing in neonatal mortality. Reduction by as much as 50% was achieved in a community trial in rural India (Bang et al. 1999). The concern however is that home use of antibiotics by mothers in the management of neonatal infections would accelerate the development of antimicrobial resistance (Kallander et al. 2004).

The high level of unsupervised deliveries in the district is likely to be a contributory factor to the level of neonatal deaths due to birth injuries. Birth injuries accounted for 19 and 14% of early and all neonatal deaths, respectively. As part of the Community Health and Family Planning project undertaken by the research centre, trained nurses were deployed to reside within communities. These nurses offered basic curative and preventive services including deliveries services. The success of this approach in expanding the coverage of maternal and child health services has caught national attention. The Government of Ghana is presently engaged in scaling-up the intervention nationwide. There would be great benefit to ensure that deployed nurses have basic midwifery skills that make them competent enough to supervise uncomplicated deliveries. It would also be important to put in place a good referral system. Additionally, continuous and focused health education campaigns need to be undertaken to positively influence the firmly-held desire of pregnant mothers to deliver with traditional birth attendants (TBAs). This can be achieved if community-based nurses and midwives work closely with TBAs within their catchment areas, and even more importantly, incorporate some of their positive attributes in their practices.

Maternal immunization is another strongly advocated strategy to reduce neonatal mortality. It follows on the successful reduction and elimination of neonatal tetanus as a major cause of mortality. It is a strategy that is likely to enjoy good appeal from public health policy makers in sub-Saharan Africa. High levels of antenatal attendance and breastfeeding are conditions that make maternal immunization a viable strategy to reduce the contribution of infections to neonatal and other infant deaths in sub-Saharan Africa (Mulholland et al. 1996; Greenwood 2003). Among the various vaccines in the production pipeline, pneumococcal vaccines are the strongest candidates for large-scale evaluation in developing countries. It is hoped that soon results will be available on trials that evaluate the effect of maternal immunization on the incidence of serious infections and/or deaths in early infancy.

The contribution of undesirable customary practices to infant mortality in sub-Saharan Africa is exemplified in the profile of causes of neonatal deaths in KND. A study conducted in 1995 found that infanticide accounts for 15% of all deaths in children less than 3 months in KND (Allotey & Reidpath 2001). The findings of this study influenced the inclusion of infanticide in the COD list used by the research centre. Although validation studies are required to ascertain the sensitivity and specificity of the use of VA to diagnose infanticide, we believe that this would be fairly high given the open cultural practices and rituals that accompany the procedure (Allotey & Reidpath 2001). It involves calling in a soothsayer who performs some rituals within the compound of the suspected spirit child. The latter is then forced to take in a potion of bitter herbs. Although infanticide has for a long time been recognized as an ‘important risk factor’ for infant mortality within this district (Sena 1993; Allotey & Reidpath 2001), no intervention studies have been specifically directed as stopping the practice. Detailed anthropological studies are needed to guide the development of interventions.

In the conduct of clinical trials, a fair idea of the prevalence of the end-point is needed to estimate the adequate sample size. In the absence of site-specific data, it is accepted practice to make extrapolations of information available in-country or within sub-region. With disproportionate decline in mortality however, there is the danger that such approach would lead to gross underestimation of required sample sizes. We believe that the observed NMR in KND is unrepresentative of typical rural northern Ghana, but more an outcome of the health interventions instituted in KND. It is therefore, important that the estimation of sample sizes in trials proposed for demographic surveillance sites such as Navrongo be based on data collected within the district, and as close to the time of study as possible. The good thing is, the data needed to do this, is available and continually updated as part of the surveillance system.

This review of the trend and causes of neonatal mortality using data collected at the Navrongo DSS has demonstrated how routinely collected data at DSS sites could be used to assess trends on important demographic and health indicators. More validation studies are needed. An approach towards standardization of tools and procedures that recognize the peculiarity of sites is recommended. This would make meaningful the pooling of data from DSS and non-DSS sites to inform the broader picture on progress towards reducing neonatal deaths and the achievement of goals set within the millennium development agenda.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Kassena–Nankana district
  5. Methodology
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References

We acknowledge the support received from the INDEPTH Network in putting together this manuscript. We are also grateful to staff of the NDSS and Computer units at NHRC who managed data collection and conducted the verbal autopsies. Gratitude is extended to the chiefs and people of the Kassena–Nankana district for their continued collaboration with the centre.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Kassena–Nankana district
  5. Methodology
  6. Results
  7. Discussion
  8. Acknowledgements
  9. References