Estimating the effectiveness of self-help groups on the adoption of secondary preventive measures by people living with HIV in Central America, 2012

Sanchez-Dominguez, Mario Salvador; Lamadrid-Figueroa, Hector; Leyva-Flores, Rene; Infante-Xibille, Cesar

doi:10.1186/s12913-020-05235-0

Table 1 Bias reduction tests before and after matching based on the Kernel algorithm

From: Estimating the effectiveness of self-help groups on the adoption of secondary preventive measures by people living with HIV in Central America, 2012

Variable		Intervention	Control	%bias	%bias reduction	t test	P value
Sex (Male)	Without matching	0.548	0.510	7.7	–	1.97	0.048
Sex (Male)	Kernel	0.548	0.561	− 2.5	67.0	− 0.89	0.372
Age (years)
Youth (18–24)	Without matching	0.122	0.092	10.0	–	2.48	0.013
Youth (18–24)	Kernel	0.122	0.126	−1.1	89.4	−0.35	0.729
Young adult (25–44)	Without matching	0.623	0.627	−0.9	–	−0.23	0.819
Young adult (25–44)	Kernel	0.623	0.626	−0.6	33.9	−0.21	0.836
Mature adult (45–59)	Without matching	0.209	0.243	−8.3	–	−2.15	0.031
Mature adult (45–59)	Kernel	0.209	0.215	−1.5	82.3	−0.53	0.600
Older adult (≥60)	Without matching	0.046	0.038	4.1	–	1.03	0.302
Older adult (≥60)	Kernel	0.046	0.034	6.1	−48.4	2.19	0.029
Schooling
None	Without matching	0.115	0.066	17.2	–	4.17	0.000
None	Kernel	0.115	0.097	6.4	62.5	2.09	0.036
Literate	Without matching	0.029	0.032	−2.2	–	−0.56	0.573
Literate	Kernel	0.029	0.038	−5.2	−140.8	−1.76	0.079
Completed elementary school	Without matching	0.367	0.449	−16.6	–	−4.28	0.000
Completed elementary school	Kernel	0.367	0.357	2.0	87.9	0.71	0.475
Completed middle school	Without matching	0.383	0.376	1.4	–	0.35	0.724
Completed middle school	Kernel	0.383	0.384	−0.2	87.8	−0.06	0.953
University	Without matching	0.106	0.077	10.2	–	2.52	0.012
University	Kernel	0.106	0.125	−6.4	37.3	−2.01	0.044
Average monthly income	Without matching	205.760	182.180	10.4	–	2.61	0.009
Average monthly income	Kernel	205.760	216.680	−4.8	53.7	−1.73	0.083
Yes, economic dependents	Without matching	0.634	0.593	8.4	–	2.16	0.031
Yes, economic dependents	Kernel	0.634	0.597	7.6	10.0	2.65	0.008
Sexual orientation
Heterosexual	Without matching	0.789	0.749	9.5	–	2.47	0.014
Heterosexual	Kernel	0.789	0.706	19.6	−106.2	6.68	0.000
Bisexual	Without matching	0.071	0.066	1.9	–	0.49	0.621
Bisexual	Kernel	0.071	0.088	−7.0	− 258.3	−2.27	0.023
Homosexual	Without matching	14,029.000	0.185	−12.2	–	−3.21	0.001
Homosexual	Kernel	0.140	0.205	−17.6	−44.6	−6.04	0.000
Yes, received counseling when tested for HIV.	Without matching	0.851	0.826	6.7	–	1.75	0.080
Yes, received counseling when tested for HIV.	Kernel	0.851	0.793	15.6	−130.7	5.25	0.000
Yes, has suffered from tuberculosis.	Without matching	0.142	0.094	14.9	–	3.66	0.000
Yes, has suffered from tuberculosis.	Kernel	0.142	0.137	1.4	90.6	0.45	0.650
Yes, receives family support.	Without matching	0.744	0.769	−5.8	–	−1.48	0.139
Yes, receives family support.	Kernel	0.744	0.751	−1.6	72.2	−0.56	0.576
Yes, has used drugs.	Without matching	0.109	0.179	−19.9	–	−5.34	0.000
Yes, has used drugs.	Kernel	0.109	0.174	−18.5	6.8	−6.51	0.000
Yes, is currently in a stable relationship.	Without matching	0.452	0.432	3.9	–	1.01	0.312
Yes, is currently in a stable relationship.	Kernel	0.452	0.438	2.8	29.1	0.98	0.328
Future time perspective	Without matching	7.310	7.285	1.8	–	0.46	0.647
Future time perspective	Kernel	7.310	7.180	9.1	− 410.0	3.12	0.002
Country
El Salvador	Without matching	0.119	0.334	−52.9	–	−14.84	0.000
El Salvador	Kernel	0.119	0.110	2.4	95.4	1.08	0.281
Belize	Without matching	0.102	0.021	34.0	–	7.67	0.000
Belize	Kernel	0.102	0.079	9.8	71.1	2.85	0.004
Costa Rica	Without matching	0.087	0.287	−53.1	–	−15.25	0.000
Costa Rica	Kernel	0.087	0.095	−2.2	95.9	−1.00	0.320
Guatemala	Without matching	0.253	0.060	55.0	–	12.60	0.000
Guatemala	Kernel	0.253	0.237	4.5	91.7	1.30	0.195
Honduras	Without matching	0.223	0.241	−4.3	–	−1.11	0.267
Honduras	Kernel	0.223	0.230	−1.6	61.7	−0.58	0.561
Panama	Without matching	0.216	0.057	47.6	–	10.95	0.000
Panama	Kernel	0.216	0.250	−10.2	78.7	−2.81	0.005

Source: prepared by the author based on data from the following project: Elaboration of the risk profile of people living with HIV in the beneficiary countries of the REDCA+ Regional Program
Information was obtained running the Stata 13.1 pstest command. Data were matched using the Kernel method after estimating the propensity score through a probit model. Matching took into account the region of common support and the balance of the variables presented in this table. Mean bias dropped from 15.6 to 6.2

Back to article page

ISSN: 1472-6963

Contact us

General enquiries: journalsubmissions@springernature.com

BMC Health Services Research

Contact us