Catastrophic health payments
Table 2 indicates the trend of the catastrophic headcount and the mean positive gap for two thresholds (10% and 25%). For both thresholds, there has been a decreasing pattern in terms of the catastrophic health payments headcount between 2005/06 and 2012/13. However, there was an increase between 2012/13 and 2016/17, irrespective of the thresholds. Concerning the mean positive gap, there has been a decreasing pattern for both thresholds, decreasing from 2005/06 to 2016/17.
Table 2: Household catastrophic health payments
|
10%
|
25%
|
Year
|
Headcount (%)
|
Mean positive gap (%)
|
Headcount (%)
|
Mean positive gap (%)
|
2005/06
|
22.4
|
11.5
|
5.9
|
13.1
|
2009/10
|
21.4
|
11.0
|
5.4
|
12.2
|
2012/13
|
13.8
|
8.9
|
2.6
|
10.9
|
2016/17
|
14.2
|
8.8
|
2.7
|
8.2
|
Source: Authors’ computation based on the UNHS 2005-2017 data
Table 3 shows catastrophic payments disaggregated by social-economic status, urban/rural location, and region of residence. The incidence of catastrophic health expenditure was higher among the richer quintiles when compared to the poorest quintile in the first three years. The reverse was true in 2016/17, where the poorer quintiles experienced a higher incidence of catastrophic payments. The incidence of catastrophic costs was much higher in the rural areas than in the urban areas in 2005/06 and 2009/10 with the pattern changing in 2012/13 and 2016/17. With regards to the regions, catastrophic health payments are highest in the Western and Central regions between 2005 and 2017.
There are some household characteristics associated with an increased likelihood of catastrophic health payments. As shown in Table 4, the factors which are significantly associated (5% level of significance) with an increased likelihood of catastrophic health expenditures are having a child, and an elderly member in the household.
Table 3: Disaggregation of catastrophic health expenditure (10% of total household expenditure)
Disaggregation variable
|
2005/06
|
2009/10
|
2012/13
|
2016/17
|
Total
|
22.4
|
21.4
|
13.8
|
17.1
|
Socio-economic status quintiles
|
|
|
|
|
Poorest
|
18.9
|
17.2
|
9.6
|
17.6
|
Second poorest
|
20.4
|
18.9
|
10.0
|
18.8
|
Middle
|
24.2
|
21.5
|
12.6
|
16.9
|
Second richest
|
26.5
|
24.2
|
18.1
|
17.6
|
Richest
|
22.0
|
25.0
|
18.7
|
14.7
|
Poverty Status
|
|
|
|
|
Non-poor
|
23.7
|
22.6
|
14.8
|
16.7
|
Poor
|
19.5
|
17.2
|
9.8
|
18.3
|
Residence
|
|
|
|
|
Rural
|
23.5
|
21.7
|
13.5
|
17.3
|
Urban
|
16.2
|
19.5
|
14.9
|
16.5
|
Region
|
|
|
|
|
Central
|
20.3
|
21.9
|
19.8
|
19.1
|
Eastern
|
21.1
|
21.6
|
9.1
|
15.9
|
Northern
|
20.2
|
18.3
|
13.1
|
14.9
|
Western
|
27.8
|
23.1
|
13.8
|
18.2
|
Source: Authors’ computation based on the UNHS 2005-2017 data
Table 4: Determinants of catastrophic health expenditure, 2016/17
Catastrophic health expenditure (10% of household expenditure)
|
Independent Variables
|
Odds-ratio (OR)
|
SE.
|
z
|
P>z
|
[95% CI]
|
Poverty (poor=1, non-poor=0)
|
0.4
|
0.0
|
-8.2
|
0.0
|
0.3
|
0.5
|
Residence (urban=1, rural =0)
|
0.8
|
0.1
|
-1.9
|
0.1
|
0.7
|
1.0
|
Region (R=central)
|
|
|
|
|
|
|
Eastern
|
0.8
|
0.1
|
-1.5
|
0.1
|
0.7
|
1.1
|
Northern
|
0.9
|
0.1
|
-0.6
|
0.6
|
0.8
|
1.2
|
Western
|
0.8
|
0.1
|
-1.6
|
0.1
|
0.7
|
1.0
|
Household size
|
1.1
|
0.0
|
2.7
|
0.0
|
1.0
|
1.1
|
Sex of household head
(male=1, female=0)
|
0.9
|
0.1
|
-1.0
|
0.3
|
0.7
|
1.2
|
Employment (R=formal)
|
|
|
|
|
|
|
Casual/Subsistence
|
1.1
|
0.1
|
0.5
|
0.6
|
0.9
|
1.3
|
Unemployed
|
1.3
|
0.2
|
1.9
|
0.1
|
1.0
|
1.6
|
Children below 5 (yes =1, no=0)
|
1.3
|
0.1
|
3.1
|
0.0
|
1.1
|
1.5
|
Adults above 60 (yes= 1, no=0)
|
1.4
|
0.2
|
3.3
|
0.0
|
1.2
|
1.7
|
Education (R= no formal education)
|
|
|
|
|
|
|
Primary level
|
1.1
|
0.1
|
0.7
|
0.5
|
0.9
|
1.4
|
Secondary level
|
1.0
|
0.1
|
-0.2
|
0.9
|
0.7
|
1.3
|
Tertiary
|
0.7
|
0.2
|
-1.7
|
0.1
|
0.5
|
1.1
|
Marital status
(married=1, not married=0)
|
1.3
|
0.2
|
2.0
|
0.0
|
1.0
|
1.7
|
_cons
|
0.1
|
0.0
|
-13.5
|
0.0
|
0.1
|
0.2
|
Log pseudo likelihood = -14632786
Number of obs = 15,349
Wald chi2(15) = 135.0
Prob > chi2 = 0.000
Pseudo R2 = 0.013
|
R= Reference category
Source: Authors’ computation based on the UNHS 2016/17 data
Impoverishment
The results in Table 5 show that OOP payments are impoverishing in Uganda as they increase the incidence and depth/intensity of poverty among the poor across all the time period. The pattern is similar for all the poverty lines considered. A decrease in the impoverishment headcount was observed from 2005/06 through to 2016/17, although the decline in the impoverishment headcount between 2012/13 and 2016/17 is minimal.
Table 5: Impoverishment indicators using the international poverty line
|
Pre-payment poverty (%)
(A)
|
Post-payment poverty (%)
(B)
|
Absolute difference (%)
(B-A)
|
2005/06 (PPP=513.9492)
|
|
|
|
Poverty headcount
|
51.8
|
57.0
|
5.2
|
Normalised mean positive poverty gap
|
35.2
|
37.0
|
|
2009/10 (PPP=741.3262)
|
|
|
|
Poverty headcount
|
46.3
|
50.8
|
4.5
|
Normalised mean positive poverty gap
|
33.4
|
34.9
|
|
2012/13 (PPP=1043.083)
|
|
|
|
Poverty headcount
|
64.0
|
67.2
|
3.2
|
Normalised mean positive poverty gap
|
39.4
|
40.2
|
|
2016/17 (PPP=1161.989)
|
51.8
|
57.0
|
5.2
|
Poverty headcount
|
35.2
|
37.0
|
|
Normalised mean positive poverty gap
|
|
|
|
Source: Authors’ computation based on the UNHS 2016/17 data
Table 6: Impoverishment indicators using Uganda’s national poverty line
|
Pre-payment poverty (%)
(A)
|
Post-payment poverty (%)
(B)
|
Absolute difference (%)
(B-A)
|
2005/06
|
|
|
|
Poverty headcount
|
31.1
|
35.6
|
4.6
|
Normalised mean positive poverty gap
|
35.2
|
37.0
|
|
2009/10
|
|
|
|
Poverty headcount
|
23.2
|
27.2
|
4.0
|
Normalised mean positive poverty gap
|
27.6
|
28.3
|
|
2012/13
|
|
|
|
Poverty headcount
|
19.7
|
21.7
|
2.0
|
Normalised mean positive poverty gap
|
26.4
|
26.7
|
|
2016/17
|
|
|
|
Poverty headcount
|
21.5
|
24.1
|
2.5
|
Normalised mean positive poverty gap
|
5.3
|
6.0
|
|
Source: Authors’ computation based on the UNHS 2016/17 data
Table 7 shows the disaggregation of impoverishment effect by socio-economic status, residence and region. The results show that the impoverishment effect is mainly concentrated in the middle and second richest quintiles of socio-economic status. It is also largely concentrated in the Central and Western regions. The distribution of impoverishment by residence is less clear-cut, showing a mixed pattern over the different years considered.
Table 7: Disaggregation of impoverishment headcount
Disaggregation variable
|
2005/06
|
2009/10
|
2012/13
|
2016/17
|
Total
|
5.2
|
4.5
|
3.2
|
2.7
|
Socio-economic status quintiles
|
|
|
|
|
Poorest
|
0.0
|
0.0
|
0.0
|
0.0
|
Second poorest
|
0.0
|
0.0
|
0.0
|
0.0
|
Middle
|
16.7
|
18.6
|
0.0
|
11.0
|
Second richest
|
8.3
|
2.5
|
14.9
|
2.5
|
Richest
|
1.0
|
1.5
|
1.2
|
0.0
|
Residence
|
|
|
|
|
Rural
|
5.6
|
4.9
|
2.9
|
3.1
|
Urban
|
2.9
|
2.2
|
4.1
|
1.6
|
Region
|
|
|
|
|
Central
|
5.4
|
3.4
|
4.8
|
2.5
|
Eastern
|
5.2
|
5.3
|
2.0
|
2.5
|
Northern
|
2.7
|
4.4
|
2.7
|
3.0
|
Western
|
6.9
|
4.9
|
3.4%
|
3.0
|
Source: Authors’ computation based on the UNHS 2005-2017 data