OBJECTIVE: To develop a socioeconomic index for brief surveys that allows a classification in relation to a reference population with a parsimonious approach. MATERIALS AND METHODS: A socioeconomic index was created using assets and use of services indicators to predict income level within a national representative survey of income (ENIGH, National Survey of Households Income and Expenditures) and then tested with data from a survey among key populations (men who have sex with men/transvestite, transgender, transsexual/female sex workers) and compared to an already published indicator. RESULTS: The concordance for quintiles 1, 2, 3, 4, and 5 between the two indexes were 94, 94, 82, 83, and 89%, respectively. The Spearman's rank correlation coefficient was 0.85. CONCLUSIONS: The proposed parsimonious index captures the socioeconomic level heterogeneity in the MSM/TTT/FSW survey. The use of this index is suggested for short surveys because of the next advantages: a) it uses dichotomic variables; b) the variables used do not imply the fulfillment of any statistical assumption; c) it is easily calculated; d) it can be used for comparing groups.
OBJECTIVE: To develop a socioeconomic index for brief surveys that allows a classification in relation to a reference population with a parsimonious approach. MATERIALS AND METHODS: A socioeconomic index was created using assets and use of services indicators to predict income level within a national representative survey of income (ENIGH, National Survey of Households Income and Expenditures) and then tested with data from a survey among key populations (men who have sex with men/transvestite, transgender, transsexual/female sex workers) and compared to an already published indicator. RESULTS: The concordance for quintiles 1, 2, 3, 4, and 5 between the two indexes were 94, 94, 82, 83, and 89%, respectively. The Spearman's rank correlation coefficient was 0.85. CONCLUSIONS: The proposed parsimonious index captures the socioeconomic level heterogeneity in the MSM/TTT/FSW survey. The use of this index is suggested for short surveys because of the next advantages: a) it uses dichotomic variables; b) the variables used do not imply the fulfillment of any statistical assumption; c) it is easily calculated; d) it can be used for comparing groups.