Skip to main content

Table 2 Unadjusted and adjusted relationship between transactional sex and young women’s consumption of practical items (PRAI) and entertainment and birth control (EBC) items (compared to personal enhancement items) among sexually active women who reported some consumptiona

From: ‘It’s because I like things… it’s a status and he buys me airtime’: exploring the role of transactional sex in young women’s consumption patterns in rural South Africa (secondary findings from HPTN 068)

Outcome: PRAI (n = 398)b uOR 95% CI p-value aORd 95% CI p-value#,
 No Reference    Reference   
 Yes 1.32 0.72–2.44 0.37 1.09 0.57–2.12 0.79
Outcome: EBC (n = 486)c
 No Reference    Reference   
 Yes 1.67 1.03-2.72 0.04* 3.03 1.12-8.23 0.03*
  1. aAs separate analysis for each variable. Sample size for each analysis included in each category
  2. bPlease note: These numbers are the sum of the number of young women (that are sexually active) who are buying practical items + the number that are buying personal enhancement items. Records with missing data excluded. Total missing: n = 30 (7%) These are: work for money (n = 8 or 1.2%) missing), orphan status (n = 2 or 0.04% missing), type of primary caregiver (n = 1 or 0.02% missing), number of sex partners past 12 months (n = 19 or 4% missing)
  3. cPlease note: These numbers are the sum of the number of young women (that are sexually active) who are buying entertainment related and birth control items + the number that are buying personal enhancement items. Records with missing data excluded. Total missing: n = 40 (7.5%) These are: work for money (n = 7 or 1.3%) missing), orphan (n = 9 or 1.7% missing), number of sex partners past 12 months (n = 24 or 4.6% missing)
  4. uOR, unadjusted odds ratio aOR Adjusted odds ratio
  5. dAdjusted for all confounders – age, number of household members, being an orphan, type of primary caregiver, currently has a boyfriend, number of sexual partners last 12 months, working for money, per capita consumption
  6. #P-value estimation through likelihood ratio test (LRT); * P-values in bold are statistically significant at the < 0.05 level