Neonatal mortality rates, characteristics, and risk factors for neonatal deaths in Ghana: analyses of data from two health and demographic surveillance systems

Figures & data

Box 1. Definition of variables

Outcome variable: Neonatal death: Neonatal death is defined as the death of a baby within 28 days of life (days 0–27). In this study it was dichotomised as Dead [coded as 1] if the baby died within 28 days or Alive [coded as Not dead = 0] if (s)he did not die within the study period or died after 28 days of birth. Deaths within seven days of birth (days 0–6) were classified as early neonatal deaths (END) and deaths after seven days (days 7–27) were considered late neonatal deaths (LND). Explanatory variables: Place of delivery: Refers to the setting where the baby was reported to have been delivered, categorised as hospital, clinic or home births. Hospitals were formal facilities with a midwife or nurse and a doctor. Clinics were formal facilities with either a midwife or nurse, but no doctor. It included health centres, community-based health planning services (CHPS) and maternity homes. Home births included babies delivered in either the traditional birth attendant’s (TBA) home or the mother’s home, with no formally trained midwife or doctor. Health facility birth: A collective term for hospital and clinic births Place of birth/ place of death: This was a derived variable based on where the baby was born and where (s)he died. For example, Hosp/Hosp referred to a baby who was born in a hospital and died in a hospital; Hosp/Home referred to a baby who was born in a hospital but died at home; Home/Hosp referred to a baby who was born at home but died in a hospital. Socioeconomic status (SES): SES was derived by statisticians at the research centre from principal component analyses (PCA) of household assets: type of housing; roofing and flooring material; structural condition of house; number of rooms; status of occupancy; electricity supply; type of cooking fuel; household facilities; source of water; health insurance enrolment status; land ownership and utilisation; type of assets in working conditions; income per week or month of household head and other income earners in the household; type of cooking salt; livestock and number of leave days. In this study SES was classified as five wealth quintiles for the study population: Poor, Next poor, Average, Next rich and Rich. Marital status: Marital status was dichotomised as married or not married. Data for marital status were only available in the Navrongo HDSS. Maternal age: Mother’s age at delivery was calculated as the difference between baby’s year of birth and mother’s year of birth. Mother’s age recorded as younger than 13 years old or older than 60 years was considered likely to be incorrect and treated as missing (n = 24). Education: Education referred to the highest level of formal education attained by the mother at the time of delivery. This was categorised as: No education: a mother who had received no form of schooling; Basic education: a mother who had attained up to junior high education; Secondary or more: a mother who had received secondary education or higher. This information was only available for Navrongo for the study period. Antenatal care (ANC) attendance: ANC attendance was classified as Yes or No, depending on whether a mother received at least one episode of ANC during pregnancy. This was recorded after delivery, based on maternal recall. In Navrongo, it was not recorded before 2012. Multiple births: Multiple births refer to a mother who delivered more than one baby at the end of a single pregnancy. It was dichotomised into singletons and multiple, for twins or more. If a multiple birth was recorded as more than four babies, it was judged to have been an error and treated as a missing value. First birth: First birth was derived from the birth order of the baby. This was dichotomised as Yes or No, if the baby was the first born or not. Parity: Parity refers to the total number of live births a woman had delivered, including the current birth, at the time of data collection. It was categorised into 1, 2, 3 and 4 or more, and the value was cross-checked with first birth for consistency (Parity = 1 if first birth = Yes; Parity>1 if first birth = No). If parity was greater than 1 and first birth = Yes, the value of parity was treated as missing. Sex: Male or female. Day of birth: Refers to the day of the week in which a baby was born. It was generated from the baby’s date of birth and categorised as weekend if a baby was born on Saturday or Sunday, and weekday if (s)he was born from Monday to Friday inclusive. Year of birth: Refers to the year the baby was born. We assessed for linearity of year of birth in the univariate models and observed that it was held for Navrongo, but not Kintampo. Therefore, year of birth was modelled as a continuous variable in Navrongo, and as a dummy in Kintampo. Age at death: Age at death was calculated as the difference, in days, between date of birth and date of death. It was categorised into 0 (deaths on the day of birth), 1–6 (early neonatal deaths, excluding deaths on the day of birth) and 7–27 days (late neonatal deaths). Cause of death*: Refers to the most probable cause of neonatal death as ascertained by at least two medical doctors from the VA narrative and checklist. *All the variables, except cause of death, were reported by the parent, usually the mother.

Table 1. Summary of maternal and neonatal characteristics, neonatal deaths, and overall NMRs in Navrongo (2004–2012) and Kintampo (2005–10)

	Navrongo				Kintampo
	Total livebirths	Dead	NMR	p-value	Total livebirths	Dead	NMR	p-value
	n (%)	n (%)			n (%)	n (%)
Total	17,016	320	18.8	-	11,207	140	12.5	-
Place of birth
Home	7,002 (41.1)	139 (43.4)	19.9	5,783 (51.6)	58 (41.4)	10.0
Hospital	7,306 (42.9)	150 (46.9)	20.5	0.01	3,634 (32.4)	64 (45.7)	17.6	0.003
Clinic	2,708 (15.9)	31 (9.7)	11.4	1,790 (16.0)	18 (12.9)	10.1
Maternal age
<20 years	1,875 (11.0)	52 (16.3)	27.7	0.07	963 (8.6)	8 (5.8)	8.3	<0.001
20–29	8,357 (44.1)	141 (44.1)	16.9	5,344 (47.9)	55 (39.6)	10.3
30–39	5,229 (30.7)	106 (33.1)	20.3	3,994 (35.8)	54 (38.9)	13.5
40+	1,555 (9.1)	21 (6.6)	13.5	863 (7.7)	22 (15.8)	25.5
Missing	-	-	43	1
Socioeconomic status
Poor	3,855 (24.3)	61 (26.0)	15.8	0.34	1,811 (16.2)	24 (17.3)	13.3	0.40
Next poor	2,916 (18.3)	47 (20.0)	16.1	2,138 (19.1)	22 (15.8)	10.3
Average	2,777 (17.5)	37 (15.7)	13.3	2,311 (20.7)	24 (17.3)	10.4
Next rich	3,205 (20.2)	49 (20.9)	15.3	2,509 (22.5)	35 (25.2)	13.9
Rich	3,146 (19.8)	41 (17.5)	13.0	2,406 (21.5)	34 (24.5)	14.1
Missing	1,117	85	32	1
Marital status
Not married	2,736 (16.1)	65 (20.3)	23.8	0.04	-	-	-
Married	14,280 (83.9)	255 (79.7)	17.9	-	-	-
Education
None	4,926 (29.6)	99 (31.6)	20.1	0.11	-	-	-
Basic	9,026 (54.3)	176 (56.2)	19.5	-	-	-
Secondary+	2,682 (16.1)	38 (12.1)	14.2	-	-	-
Missing	382	7	-	-
Sex
Female	8,455 (49.7)	138 (43.1)	16.3	0.02	5,475 (48.9)	61 (43.6)	11.1	0.21
Male	8,561 (50.3)	182 (56.9)	21.3	5,732 (51.2)	79 (56.4)	13.8
First birth
No	12,132 (71.3)	191 (59.7)	15.7	<0.001	8,651 (77.3)	117 (83.6)	13.5	0.07
Yes	4,884 (28.7)	129 (40.3)	26.4	2,541 (22.7)	23 (16.4)	9.1
Missing	-	-	15	0
Parity
1	4,884 (29.0)	129 (40.6)	26.4	<0.01	2,396 (21.9)	19 (14.4)	7.9	0.02
2	3,639 (21.6)	47 (14.8)	12.9	1,872 (17.1)	19 (14.4)	10.1
3	3,114 (18.5)	54 (17.0)	17.3	1,721 (15.7)	25 (18.9)	14.5
4+	5,185 (30.8)	88 (27.7)	17.0	4,975 (45.4)	69 (52.3)	13.9
Missing	194	2	243	8
Multiple birth
Single	16,396 (96.4)	276 (86.3)	16.8	<0.001	10,658 (96.2)	126 (90.7)	11.8	0.001
Multiple	614 (3.6)	44 (13.8)	71.7	423 (3.8)	13 (9.4)	30.7
Missing	6	0	126	1
Day of birth
Weekday	12,149 (71.4)	215 (67.2)	17.7	0.09	8,015 (71.5)	96 (68.6)	12.0	0.44
Weekend	4,867 (28.6)	105 (32.8)	21.6	3,192 (28.5)	44 (31.4)	13.8
ANC
Yes	-	-	-	8,142 (86.8)	103 (82.4)	12.7	0.15
No	-	-	-	1,241 (13.2)	22 (17.6)	17.7
Missing	-	-	-	1,824	15

- Data not available.

Figure 1. Distribution of neonatal deaths in Navrongo and Kintampo by age at death

Table 2. Multivariable logistic regression analyses of factors associated with neonatal mortality in Navrongo (2006–2012)

	Complete case model (N = 12,138)		Multiple imputations model (N = 13,327)
	Multivariable OR (95% CI)*	p-value	Multivariable OR (95% CI)*	p-value
Place of birth
Home	1.00	1.00
Hospital	1.08 (1.02–1.14)	0.01	1.14 (1.03–1.25)	0.01
Clinic	0.67 (0.63–0.72)	<0.001	0.67 (0.65–0.68)	<0.001
Maternal age
<20 years	1.18 (0.75–1.83)	0.48	1.14 (0.78–1.65)	0.50
20–29	1.00	1.00
30–39	1.46 (1.04–2.05)	0.03	1.42 (0.98–2.06)	0.06
40+	0.72 (0.58–0.90)	<0.01	0.76 (0.64–0.91)	<0.01
Socioeconomic status
Poor	1.47 (1.22–1.78)	<0.001	1.36 (0.97–1.90)	0.07
Next poor	1.33 (0.72–2.44)	0.36	1.25 (0.85–1.84)	0.27
Average	1.13 (0.77–1.68)	0.53	1.09 (0.65–1.81)	0.75
Next rich	1.20 (1.03–1.39)	0.02	1.14 (0.82–1.59)	0.42
Rich	1.00	1.00
Education
None	1.13 (1.06–1.21)	<0.001	1.12 (1.03–1.21)	0.01
Basic	1.00	1.00
Secondary+	0.76 (0.69–0.84)	<0.001	0.76 (0.67–0.87)	<0.001
Parity
1	1.00	1.00
2	0.41 (0.21–0.78)	0.01	0.41 (0.23–0.74)	<0.01
3	0.54 (0.32–0.92)	0.02	0.58 (0.39–0.85)	0.01
4+	0.49 (0.31–0.77)	<0.01	0.52 (0.34–0.81)	<0.01
Boys	1.38 (1.18–1.61)	<0.001	1.39 (1.20–1.60)	<0.001
Multiple birth	5.06 (4.40–5.82)	<0.001	5.35 (5.16–5.55)	<0.001
Weekend birth	1.05 (0.79–1.39)	0.75	1.05 (0.75–1.45)	0.79
Married	1.03 (0.69–1.53)	0.89	1.17 (0.85–1.59)	0.33
Year of birth	0.98 (0.97–0.99)	<0.01	0.97 (0.97–0.98)	<0.001

*Adjusted for place of delivery, year of birth, maternal age, SES, marital status, education, sex, multiple birth, parity and day of birth.

*Data for multivariable analyses cover the years 2006–2012.

*m = 50.

Table 3. Multivariable logistic regression analyses of factors associated with neonatal mortality in Kintampo (2005–2010)

	Complete case model (N = 9,162)		Multiple imputations model (N = 11,207)
	Multivariable OR (95% CI)#	p-value	Multivariable OR (95% CI)#	p-value
Place of birth
Home	1.00	1.00
Hospital	1.72 (1.57–1.88)	<0.001	1.76 (1.55–2.00)	<0.001
Clinic	1.27 (1.24–1.31)	<0.001	1.12 (1.06–1.18)	<0.001
Maternal age
<20 years	1.07 (0.90–1.25)	0.42	0.95 (0.75–1.20)	0.65
20–29	1.00	1.00
30–39	1.35 (1.00–1.82)	0.05	1.28 (1.02–1.62)	0.04
40+	2.64 (1.15–6.08)	0.02	2.62 (1.44–4.78)	<0.01
Socioeconomic status
Poor	1.20 (0.91–1.59)	0.21	0.96 (0.82–1.14)	0.67
Next poor	0.94 (0.78–1.12)	0.48	0.76 (0.65–0.88)	<0.001
Average	0.99 (0.60–1.65)	0.98	0.75 (0.58–0.96)	0.03
Next rich	1.28 (0.86–1.91)	0.23	1.01 (0.79–1.29)	0.93
Rich	1.00	1.00
Parity
1	1.00	1.00
2	1.62 (1.06–2.48)	0.03	1.17 (0.51–2.67)	0.72
3	1.99 (1.74–2.29)	<0.001	1.55 (0.92–2.63)	0.10
4+	1.31 (0.82–2.10)	0.26	1.13 (0.54–2.39)	0.74
Boys	1.33 (0.86–2.04)	0.20	1.24 (0.66–2.33)	0.51
Multiple birth	2.12 (0.66–6.77)	0.20	2.48 (1.08–5.66)	0.03
Weekend birth	0.93 (0.71–1.23)	0.63	1.14 (0.96–1.36)	0.14
ANC attendance	0.65 (0.50–0.84)	0.001	0.66 (0.49–0.90)	0.01
Year of birth	0.69 (0.48–0.81)	0.01	0.71(0.67–0.92)	0.04

#Adjusted for place of delivery, year of birth, maternal age, SES, sex, multiple birth, parity, antenatal care attendance and day of birth.

#Data for multivariable analyses cover the years 2005–2010.

#m = 40.

Figure 2. Forest plot of adjusted* odds ratio for the association between place of birth and neonatal mortality in Navrongo, according to age at death. (Left- Hospital births, Right- Clinic births, Reference line- Home births) Adjusted for place of delivery, year of birth, maternal age, SES, marital status, education, sex, multiple birth, parity, day of birth and year of birth

Figure 3. Forest plot of adjusted* odds ratio for the association between place of birth and neonatal mortality in Kintampo, according to age at death. (Left- Hospital births, Right- Clinic births, Reference line- Home births)

Figure 4. Number of neonatal deaths by place of birth/place of death according to age at death

Figure 5. Cause of neonatal death by place of birth/place of death