Study selection
The online search yielded 1824 records through database searching. From this search, 1803 retrieved studies were removed through a step- by-step process due to the following reasons: 1665 excluded as irrelevant by title and abstract, 48 removed as duplicates, 5 full text were not accessible, 35 reflected different study population, 37 reflected different outcomes, 3 had unknown study area, and 10 of them were out of the study setting. One hundred and six full text articles were assessed for eligibility and finally 21 articles were included in the quantitative synthesis (meta-analysis) (Figure 1). Finally, the quality of included articles was assessed by using NOS criteria (Table 2) and discussed in detail in annex part (Table 4 and 5).
Study characteristics
In the review there were nineteen (19) cross-sectional studies and two (2) cohort studies, of which 21 had fair to good quality, containing (n=21,015) HIV positive women. The smallest sample size was a study done in Zimbabwe with 70 participants, (30) while the largest study was done in Uganda (31) and included 5198 individuals. Mean sample size for the included studies was 1001. The lowest coverage of cervical cancer screening uptake was estimated by Bulto G.et al.(32) at 2.1%, while the highest coverage was estimated by Cholli P. et al.(33) at 79.7%. There were 9 studies from Ethiopia (32, 34-41) and the remainders were from Kenya(42, 43), Zimbabwe(30, 44), Côte d’Ivoire(45, 46), Uganda(31, 47), Cameron(33), Nigeria(48), Malawi(49) and Republic of South Africa(50). Twelve of the included studies were conducted in both rural-urban settings simultaneously. Related to nature and type of screening; 1 programmatic study, 19 population based and 1 neither of them and in 3 of the study mass screening was conducted and clinic based screening was performed in 18 of the studies. With respect to type of screening, 16 of the studies used Visual Inspection with Acetic Acid (VIA) and Visual Inspection with Lugol's Iodine (VILI) and 5 of used Pap smears (Table 2). Regarding the barriers to CC screening, 15 of the studies were identified by quantitative methods, 5 included both quantitative and qualitative methods while one study were identified by a qualitative approach only (Table 1).
Synthesized outcomes
As showed by the I2 statistics, evidence of high heterogeneity was witnessed between the studies. Evidence of publication bias was also observed in this systematic review and meta-analysis through visual inspection of asymmetrical funnel plot; however, the quantitative Egger’s linear regression test P value (0.69) revealed insignificant publication bias.
In SSA, CC screening uptake was estimated to be 30% (95% CI: 19, 49, I2=100%) according to pooling estimates of included primary studies conducted in different parts of SSA (Figure 2). Due to significant heterogeneity between the included studies, we conducted meta regression to find out the possible source of heterogeneity. Accordingly, country where the study occurred was found as a possible source and explained the observed heterogeneity by 21.16%. We did and reported pooled estimates of uptake of CC screening through sub-group analysis of countries (i.e., western Africa, southern Africa, other eastern Africa and Ethiopia). In the sub-analysis the higher coverage of estimated pooled uptake of CC screening was in western Africa 53% (95% CI: 19, 87), other eastern African countries other than Ethiopia 40% (95% CI: 23, 57) compared to pooled data of Sub Saharan Africa (this review) 30% (95% CI: 19, 49, I2=100%), and southern Africa 28% (95% CI: 10, 55). The lowest coverage of uptake of CC screening among HIV positive women was in Ethiopia 16% (95% CI: 10, 22) (Figure 3). We did a sensitivity test and found two studies (31, 49) affected the result to the other studies. Nevertheless, they were not as influential and did not significantly sensitively affect the result (figure 4).
Barriers to uptake of cervical cancer screening (quantitative synthesis)
The most common reasons for not partaking in CC screening among HIV positive women were: lack of knowledge about cervical cancer and screening 49% [95% CI: 29, 68], perceived risk of cervical cancer (perception of I am healthy) 23% [95% CI: 19, 26], fear of test result as positive 18% [95% CI: 7, 30], and fear of screening and procedure is painful 11% [95% CI: 5, 17] (Table 3).
Barriers to uptake of cervical cancer screening (qualitative)
Knowledge about cervical cancer and screening
Poor knowledge of CC and poor awareness of CC screening were identified as the main reasons for not undertaking screening (barrier) in three studies contributing qualitative results (Table 1). In one study, a participant reported “Handizivi( literally meaning I don’t know)”(44). Most of HIV-positive respondents recognized that cervical cancer can be prevented, but all of them did not believe the disease can be effectively treated, once diagnosed(26, 52).
Risk perception of cervical cancer
Though most of HIV positive women perceived high risk of getting the disease, some of them would not participate in CC screening because of perception of “I am not sick or I do not have sign and symptoms of the disease”(26, 40, 44, 51). The statement by on HIV positive women summarized the potential lack of insight into risk with her comment that “I have not really had signs to show me that I might have it”(51).
Fear of test result and fear of screening procedure
Four studies revealed that fear of a positive test result was a potential hindrance for not using the service. Belete N. et al.2015(34) reported that, one participant stated “the word you have a cancer diagnosis is really irritating beside my HIV, I think I will get hopeless, if I am diagnosed as having cervical cancer.”HIV positive women were scared to screen for any disease including cervical cancer, because they frightened to add another stressful issue that would scare them and this disease comes with misconceptions of the need to remove organs (ovary, womb)(36, 51, 52). Fear of pain related to screening was also identified as a barrier for utilization of the service in four studies (36, 40, 44, 51).
Partner attitude and acceptance of the service
In most SSA countries, there is male dominated family model, with the husband deciding every issue of the family as the primary income earner. Hence, the male partner’s reluctance to have their wife undergo CC screening is often related to the belief that this procedure violates the pride and privacy of their partner (34, 36, 40).
Access to screening services
Lack of well-equipped health facilities and qualified personnel to carry out the procedure were identified as barrier for accessibility and uptake of CC screening. Trouble in navigating health care facilities, lack of awareness about location of screening facilities were also reported as barriers (44, 49, 52). Waiting times for screening and length of the procedure were also given as potential reasons for non-uptake of the service (36, 44, 51, 52).
Cost of screening and finance
Four studies reported that lack of financial support and associated costs for screening were barriers for uptake of CC screening. In areas where poverty is prevalent, payment for non-emergency health services like CC screening is a challenge. In addition, cost of transportation contributes to non-accessibility and non-use of the service (34, 49, 52).