Skip to main content

Table 5 Effects of cash transfers on older people (30–54 years) compared to the original and synthetic comparison groups, Manicaland Cash Transfer Trial, Manicaland, Zimbabwe, 2010–2011

From: Spillover HIV prevention effects of a cash transfer trial in East Zimbabwe: evidence from a cluster-randomised trial and general-population survey

Outcome:

Males (30–54)

Females (30–54)

Ctrl

CT

CT vs. ctrl (ref.)

CT vs. s. ctrl (ref.)

Ctrl

CT

CT vs. ctrl (ref.)

CT vs. s. ctrl (ref.)

n/N (%)

n/N (%)

ATE

(95% CI)

ATE

(95% CI)

n/N (%)

n/N (%)

ATE

(95% CI)

ATE

(95% CI)

Primary outcomes

Had sex in past 30 days

84/92 (91.3)

207/239 (86.6)

−5.05

(−12.4, 2.25)

−0.25

(− 8.56, 8.05)

152/300 (50.7)

334/696 (48.0)

−1.61

(−8.29, 5.07)

−3.04

(− 10.3, 4.24)

Condom use (last sex)

22/92 (23.9)

48/239 (20.1)

−2.30

(−12.7, 8.08)

7.64

(−0.57, 15.8)

67/300 (22.3)

160/695 (23.0)

1.16

(−5.09, 7.42)

5.95

(1.46, 10.4)

Multiple partnersa

13/92 (14.1)

24/240 (10.0)

−2.84

(−10.9, 5.26)

−2.54

(−8.82, 3.73)

5/300 (1.67)

12/695 (1.73)

−0.40

(−1.71, 0.91)

0.02

(−0.44, 0.47)

Secondary outcomes

Psychological distressb

16/94 (17.0)

41/243 (16.9)

− 4.42

(− 15.7, 6.82)

1.27

(−5.33, 7.88)

118/301 (39.2)

254/701 (36.2)

−3.97

(− 11.7, 3.75)

3.00

(−1.62, 7.63)

Alcohol usec

38/94 (40.4)

104/243 (42.8)

3.57

(−11.2, 18.3)

−10.7

(−22.7, 1.22)

2/301 (0.66)

9/700 (1.29)

NAd

 

NAd

 

Smokes cigarettes

32/94 (34.0)

80/243 (32.9)

−2.06

(−14.8, 10.7)

−4.05

(−16.4, 8.27)

2/300 (0.67)

5/700 (0.71)

NAd

 

NAd

 

Takes recreational drugs

35/94 (37.2)

63/242 (26.0)

−10.5

(−23.5, 2.59)

−4.80

(−15.1, 5.53)

1/301 (0.33)

8/697 (1.15)

NAd

 

NAd

 
  1. Ctrl Control, CT Cash transfer, s. ctrl Synthetic control, ref. Reference group, ATE Average treatment effect, CI Confidence interval
  2. Sample: Older people (30–54 years) who had sex before for primary outcomes and all young people for secondary outcomes. The synthetic comparison group was determined through propensity score matching of individuals from the Manicaland Cohort to treatment-group individuals from the Manicaland Trial
  3. Numbers are sample sizes of individuals reporting the outcome among everyone with data on the outcome (n/N) together with percentages (%) in the original control and treatment (CT) groups of the Manicaland Trial and estimated ATEs with 95% CIs when comparing the treatment (CT) groups against the original and synthetic comparison groups. ATEs for each outcome were estimated from separate mixed-effects logistic regression models, controlling for age and wealth index quarters (not shown), with study site and treatment cluster random effects. ATEs for comparisons against synthetic comparison groups were estimated from mixed-effects logistic regression models specified in the same way as for the original comparison group after propensity score matching. Propensity score matching for the control group was implemented with replacement. Full results, including sample sizes for synthetic comparison groups, can be found in Additional file 3 (section 1).
  4. a Having had more than one sexual partner in the past 12 months
  5. b Reporting at least 7 symptoms of psychological distress of a 25-item scale
  6. c Having been to a beer hall (bar) in the past month or drinking more than 3 drinks when drinking alcohol
  7. d No regression model was estimated due to sample size limitations
Back to article page

BMC Public Health

ISSN: 1471-2458

Contact us