Prevalence and risk factors for khat use among youth students in Ethiopia: systematic review and meta-analysis, 2018

This systematic review and meta-analysis were registered at the International Prospective Register of Systematic Reviews. The following represents the registration number: CRD-42017081886.

Three investigators (WG, TA and WW) independently screened the selected articles using their titles and abstracts before retrieving the full-text papers. We use pre-specified inclusion criteria to screen the full-text articles. Disagreements between the investigators were discussed during a consensus meeting with a fourth reviewer (SS) in order to select the studies to be included in the systematic review and meta-analysis.

All the articles accessed using the databases and search engines were exported to EndNote (version 6), and we excluded duplicate articles. The remaining articles were evaluated based on the topic, language and study area. Next, studies conducted outside of Ethiopia, those not published in English and those on irrelevant topics were excluded. There were no time restrictions among the included studies. Finally, the abstracts and full text of the remaining articles were reviewed.

Current khat use is defined as the proportion of students who are chewing for different purposes within 3 months of prior to data collection.

After extracting and documenting the data in a Microsoft Excel spreadsheet, we exported it to Stata (version 14) for further analysis. All the analyses were conducted using Comprehensive Meta analysis software (version 3) [42]. The overall pooled prevalence of khat chewing was estimated using a random-effects meta-analysis [43]. First, using a fixed-effects model, heterogeneity among the studies was determined. A Q test and an I² heterogeneity test [43] were used to declare heterogeneity at p < 0.05. The prevalence of statistical heterogeneity between the studies was assessed using I² statistics, with 25%, 50% and 75% representing low, medium and high heterogeneity, respectively [44]. The quality of the included studies was evaluated using the Newcastle–Ottawa Scale [45] and tested based on sample size and representativeness, comparability between participants, ascertainment of khat chewing and statistical quality. To test the agreement between the three reviewers, the actual agreement and agreement beyond chance (unweighted kappa) were used. The values of reviewers’ result 0, 0.01–0.20, 0.21–0.40, 0.41–0.60, 0.61–0.80, and 0.81–1.00 were used to represent poor, slight, fair, moderate, substantial, and almost perfect agreements, respectively [46]. A random-effects model was used in the analysis. Meta-regression was conducted to explore the probable source of heterogeneity. A leave-one-out sensitivity analysis was also conducted to assess which studies majorly impacted between-study heterogeneity. A funnel plot and Egger’s regression test were used to measure publication bias.

The systematic literature search generated a total of 284 articles. In total, 55 were duplicated and 167 were irrelevant, and, as such, these were excluded. In addition, 44 were excluded (not measure outcome, measure not current use). The remaining 18 articles were used to determine the pooled prevalence of khat chewing in Ethiopia. All 18 articles were cross-sectional studies (Fig. 1).

All the included studies were conducted in Ethiopia (Table 1). Finally, the levels of agreements between the reviewers about the levels of bias for studies included in this meta-analysis we got from moderate to almost perfect (Kappa statistic range 0.60–1) (Table 2).

The pooled prevalence of khat chewing in Ethiopia was 16.7% (13.7–19.7: I² = 63.8%, p ≤ 0.001) (Fig. 2).

In the random-effects model, the subgroup analysis by setting on high school and university is different which is 14.61% (11.09, 18.13:I² = 11.9%, p = 0.338) and 17.56% (13.52, 21.61:I² = 71.6%, p = 0.001), respectively (Fig. 3).

By region showed that the highest prevalence of khat chewing was observed in the Oromia Region (OR: 21.1 (15.5, 26.7)). A comparable prevalence was observed in the Amhara Region (OR: 14.8 (10.6, 18.9)) and the southern Nation Nationality people of Ethiopia (OR: 14.3 (10.3, 18.3)) (Fig. 4).

There was no evidence of bias, as observed in the funnel plot. An Egger’s regression test confirmed this (p = 0.53) (Fig. 5).

Khat use is common among high school, college and university students. There are many risk factors, but just we include those reported in more than one study. Being male is more likely to chew khat than females (OR: 2.43 (1.73, 3.13). In addition, khat chewing is predominantly practised by Muslims (OR: 2.22 (1.6, 2.8). Students with a family member who had a history of khat use or who currently uses khat are more likely to chew khat than students without such a family member (OR: 1.8 (1.4, 2.2).