Time-series analysis of weather and mortality patterns in Nairobi's informal settlements

Figures & data

Table 1. Cause of death profiles for the study period 2003–2008

Underlying cause of death	n	%
HIV related
AIDS	281	11.18
HIV + TB	127	5.05
TB	225	8.95
Non-communicable
Cancers	55	2.19
Diabetes	30	1.19
Hypertension	19	0.76
Other NCDs	209	8.32
Acute infections
Malaria	90	3.58
Measles	61	2.43
Meningitis	77	3.06
Other acute infection	191	7.6
Pneumonia	223	8.87
Other causes
Maternal-related death	58	2.31
Malnutrition	50	1.99
Perinatal death	70	2.79
Prematurity/pre-term	29	1.15
Indeterminate	718	28.57

Table 2. Distribution of mortality by age, gender, and cause of death

	n	Percent	Daily average	Daily maximum
Age
All ages	2,512	100	1.19	13
0–4 years	847	33.7	0.40	4
5–19 years	131	5.2	0.06	2
20–49 years	1,186	47.2	0.56	10
50+ years	348	13.9	0.16	4
Sex
Female	1,180	47.0	0.56	7
Male	1,332	53.0	0.63	9
Cause of death
HIV	632	25.2	0.30	5
NCDs	313	12.5	0.15	4
Pneumonia	223	8.9	0.11	3
Acute infections	419	16.7	0.20	4
Other causes	925	36.8	0.44	9

Fig. 1.  Seasonal distribution of all-cause (weekly) mortality and temperature (°C).

Fig. 2.  Annual seasonal variation plots for all-age and under-five mortality. The vertical axes show the log (relative risk) and the horizontal axis show the month starting with January. Confidence intervals (95%) are shown as dotted lines.

Fig. 3.  Smooth functions of temperature for all and under-five mortality, and rainfall for all ages allowing lags of 0–13 days and 14–29 days. The vertical axes show the log (relative risk) and the horizontal axis show the scale of the explanatory variable. Confidence intervals (95%) are shown as dotted lines.

Table 3. Percentage change associated with 1°C decrease in temperature below 25th percentile, 1°C increase in temperature above 75th percentile and 25.4 mm (1 inch) increase in amount of rainfall

	Temperature		Rainfall
	25th percentile	75th percentile	Lag 0–13 days	Lag 14–29 days	Cumulative
All deaths	3 (−5, 13)	0 (−1, 1)	2 (−1, 5)	1 (−1, 4)	3 (0, 7)
Age groups
0–4 years	3 (−9, 16)	1 (0, 2)	2 (−2, 6)	0 (−3, 4)	2 (3, 8)
5–49 years	2 (−8, 14)	0 (−1, 1)	2 (−1, 6)	1 (−2, 4)	3 (−1, 8)
50+ years	9 (−6, 28)	1 (−1, 2)	3 (−2, 8)	3 (−2, 7)	5 (−1, 13)
Gender
Female	−3 (−8, 13)	0 (−1, 2)	5 (1, 8)	2 (−1, 6)	7 (2, 12)
Male	10 (−1, 22)	0 (−1, 1)	0 (−3, 3)	0 (−3, 4)	0 (−4, 5)
Cause of death
HIV	2 (−11, 17)	0 (−2, 1)	1 (−3, 5)	−1 (−5, 3)	0 (−5, 6)
NCDs	−9 (−22, 6)	1 (0, 3)	6 (1, 12)	6 (1, 11)	12 (5, 20)
Pneumonia	6 (−21, 11)	1 (−1, 2)	13 (7, 19)	10 (5, 15)	24 (16, 33)
Acute	13 (−2, 30)	1 (−1, 2)	−2 (−7, 2)	0 (−4, 4)	−2 (−8, 4)
Other	1 (−11, 14)	0 (−1, 2)	2 (−2, 6)	1 (−3, 4)	2 (−3, 8)

Fig. 4.  Sensitivity analyses, with increasing degrees of freedom, of the percentage increase in mortality for an increase in temperature of 1°C at lag 0–1 days for both cold and heat effects. The left vertical axis shows the heat effect and right vertical axis shows cold effect in percentages.