Barriers and facilitators to uptake of cervical cancer screening among women in Uganda: a systematic review

This systematic review was modelled on the PRISMA guidelines [13]. A systematic literature search was performed using Ovid MEDLINE, EMBASE, PsycINFO and SCOPUS in October 2017. The subject search and text word search were performed separately in all databases and then combined with ‘OR’ and ‘AND’ operators. The MeSH (Medical Subject Headings) terms included ‘cervical cancer’, ‘cervical neoplasms’, ‘cervical cancer screening’, ‘HPV testing’, ‘pap smear’, ‘visual inspection with acetic acid’, ‘barriers’, ‘facilitators’, ‘utilisation’, ‘Uganda’, ‘East Africa’, ‘Sub-Saharan Africa’. The search was limited to the year 1990 onwards and to English language, full-text articles. The database search was supplemented with searches on Google scholar, Proquest Theses and Dissertations database, manually examining reference lists of included articles and querying content experts. The last search was completed on 30th May 2019. The search outputs were saved where possible on databases and the authors received notification of any new searches meeting the search criteria.

The initial database search returned 207 articles after removing unrelated titles, and 3 additional articles were identified through Google Scholar and from reference lists (see Fig. 1). The 69 duplicates were removed. The abstracts of the 141 articles were read and 115 studies excluded for not meeting the inclusion criteria. Following full-text review, 12 additional articles were excluded, as they did not specifically address barriers or facilitators to cervical cancer screening in Uganda. For example, some studies included data collected in Uganda as part of a larger study of African countries, but did not specify which findings were from Uganda.

Data were extracted from the remaining 14 papers independently by two of the authors (EB and FH). Discrepancies were resolved by discussion and consensus. For quantitative studies, data extracted included barriers and facilitators that were significantly associated with CCS intention or uptake, as well as proportions of participants reporting a barrier or facilitator. For qualitative studies, data extracted included all reported barriers/facilitators. Due to the heterogeneity in study designs, participants, and outcomes, a meta-analysis was not feasible. Instead, data from the studies was used to form a narrative analysis of barriers and facilitators to cervical cancer screening based on emergent themes.

Quantitative and qualitative studies examining barriers and/or facilitators to uptake of CCS among women in Uganda (any age) were included. Quantitative studies were included to identify associations between various factors and screening uptake, while qualitative studies were included to explore barriers and facilitators to screening that were reported by women or health care workers (HCWs). Studies that described the views of, or measured data from HCWs were included as it was anticipated they would have relevant insights into factors related to health systems and resources. Exclusion criteria were as follows: studies published prior to 1990, not in the English language or not available in full text, and those that did not specifically address barriers or facilitators to uptake of cervical cancer screening among Ugandan women. Studies that focused on barriers faced by women with HIV were not included given that this group of women face their own, unique challenges to accessing screening services [14].

Included studies were subjected to a quality assessment using an appraisal method designed and evaluated by Sirriyeh and colleagues for use in studies with diverse designs [15]. The tool uses a 16-item scale with a 4-point scoring system and allows for an assessment of the overall quality of mixed qualitative and quantitative data. Given the small number of included studies, no studies were excluded based on their quality score.

Overall, 14 studies were included in the final analysis. Eight were cross sectional, five were qualitative studies using focus group discussions (FGDs) and key informant interviews (KIIs), and one was a mixed methods study. Table 1 provides information on the author, publication year, region/study site, sample size, research methods, and the type of statistical analysis used for quantitative studies. It also specifies which type of CCS (if relevant) was addressed in the study, and what proportion of women in the study had been screened (if measured). The 14 studies were published between 2006 and 2017 and comprise of a mix of urban and rural study populations, with at least two districts per region represented (see Fig. 2 below). Five studies focused on visual inspection methods (VIA/VILI), two on HPV self-collection, one on cytology, and six looked generally at screening without specifying a particular screening method. The studies covered a total of 4386 women and 350 HCWs. The proportion of women ever screened was measured in six studies and ranged from 4.8 to 35.1%. The highest screening rates were found among studies recruiting from women already attending health clinics [16, 17], and consequently these findings are not representative of the Ugandan population. These higher rates possibly reflect the recruitment and sampling design of these studies, whereby participants may have been encouraged and/or referred by HCWs to attend the health clinics. In studies conducted at the household level [8, 18, 19] and where multi-stage sampling was used [8, 18], the proportion of women screened was lower.

Table 2 summarizes the quality assessment findings. Scores for quantitative studies ranged from 11 to 40, and qualitative scores ranged from 12 to 39. Four studies based their investigation on an applied theoretical framework. All studies gave a clear description of the research setting and 12 of the studies completely identified their objectives. The sample was broadly representative of the target population in seven studies. Data collection procedures were described well by six studies although among the quantitative studies only a few reported assessment of reliability and validity of the survey tool. Most studies provided a fair explanation of their choice of analysis method. A few of the qualitative studies used a range of methods to assess reliability, but two studies did not report on this item.

19 barriers and 21 facilitators emerged from thematic analysis. The number and type of studies in which these were reported is summarized in Table 3. For quantitative studies, a distinction is made between those studies that reported proportions/other results, and those that identified statistically significant associations. Barriers reported by the greatest number of studies were embarrassment, fear of the screening procedure, fear of outcome, residing in a remote or rural area, limited resources/health infrastructure, and limited access to screening care. Being recommended to attend screening was the facilitating factor most consistently reported across studies. One barrier was statistically significant and this was having limited access to CCS facilities. Knowledge of CCS, perceiving oneself as at risk of CC, and being recommended to attend screening were statistically significant facilitators in two studies each. Because of the wide range in methodologies, sample sizes and study scopes of the included studies, it is not possible to draw conclusions about the most significant barriers or facilitators. Nevertheless, these results illustrate that certain factors have been identified as important barriers or facilitators by numerous studies. These factors merit further evaluation by future studies.

Barriers and facilitators were categorized into three main categories for the purpose of this review: individual, sociocultural, and structural factors.

Poor knowledge of CC was a barrier in four qualitative studies, and poor awareness of CCS was a barrier in three qualitative studies. Some women did not know about the cause of CC and many women did not know of any screening method. HCWs felt that low screening uptake could be attributed to poor knowledge of CC. Conversely, adequate knowledge of at least one screening method was significantly associated with having been screened [8] or having intention to screen [18] in two quantitative studies. However, it is not possible to ascertain the direction of causality and it is possible that women are knowledgeable as a result of having been screened rather than it being the reason for screening.

In three quantitative studies women with low risk perception were less likely to report intention to screen [11, 18, 19]. Some women had not been screened because they believed it was unnecessary in the absence of symptoms. Conversely, women who felt at risk were twice as likely to report intention to screen [19], and feeling at risk was significantly associated with willingness to collect an HPV sample [20].

Experiencing signs and symptoms of CC was a trigger to seeking CCS amongst women in two qualitative studies [12, 21] and one quantitative study [8]. Ndejjo and colleagues reported that signs and symptoms were the strongest trigger to accessing CCS among the women in their study [12].

Five qualitative studies reported that embarrassment related to the intimate nature of VIA/pap smears was a deterrent to screening. Self-collected HPV testing was regarded as embarrassing by women in one qualitative study [22]. Location of screening and whether privacy was afforded also affected willingness to screen in two quantitative and two qualitative studies. Importantly, Teng and colleagues found that women universally agreed that embarrassment would not be a major deterrent to screening if they were well informed about the need to screen, and if a private place for self-collection of HPV swabs was available [22].

Five qualitative and two quantitative studies reported on fear related to the screening procedure. In many cases this related to fear of pain. Fear of becoming infected through non-disposable speculums or poor sanitary practices was reported in three qualitative studies [14, 21, 23]. Fear that the procedure might cause cancer [22], lead to ‘enlargement of the sexual parts’, [23] or ‘pull out the uterus’ [21] were also reported.

Fear of being diagnosed with CC, often coupled with a sense of fatalism regarding prognosis and implications, was a reported barrier in five qualitative studies. Notably, in Paul et al’s qualitative study, fear of receiving a CC diagnosis motivated some women to attend screening [21]. Women who reported being unafraid of receiving a diagnosis were significantly more likely to have intention to screen in one study [19].

In one quantitative study, HCWs reported that lack of spousal emotional and financial support was a barrier to CCS [24]. Conversely, women in Teng et al’s qualitative study universally stated that they would attend CCS regardless of whether or not their spouse approves [22], and spousal approval did not influence women’s willingness to self-collect HPV samples in Mitchell et al’s cross sectional study [20].

Encouragement from family members to attend screening, particularly spousal encouragement, was an important motivator for women in Paul et al’s qualitative study [21]. Women who reported discussions with their husbands about screening were more likely to report intention to screen in one quantitative study [19].

Concern about how screening was perceived by community members and family was a barrier reported by four qualitative studies. A common preoccupation was that CCS might also reveal one’s HIV status, leading to societal rejection. In one qualitative study women were concerned their spouse might leave them if they were found to have CC due to resultant treatment expenses [12].

Having known somebody with CC, or somebody who had undergone CCS, was a motivating factor for women to access screening in one quantitative and two qualitative studies. Some women related that loss of a family member to CC had motivated them to be screened [12, 14]. In one cross sectional study, women who knew someone who had ever been screened or diagnosed were significantly more likely to have been screened [8].

Being recommended to attend screening by HCWs was a significant facilitator in Ndejjo et al’s study, where women who had been recommended by a HCW were 87 times more likely to have been screened for CC compared with their counterparts [8]. Osingada et al. found that women who had never received encouragement to screen from HCWs were 84% less likely to have been screened [16].

In one study, several HCWs reported that many women first seek healthcare from traditional practitioners because of the perception that CC is caused by witchcraft. This was described as being a barrier to CCS in that it delays screening among women who first look for traditional cures [12].

In one study, women with postsecondary education were significantly more likely to have been screened than their less educated counterparts [23]. Formal employment was seen to significantly facilitate screening [19], and women whose households earned more than 40 US dollars per month had a significantly higher level of intention to screen [18]. In one cross sectional study, respondents who lived in households with five or less members were twice as likely to have undergone CCS compared with their counterparts [8]. Living in a remote or rural area was a barrier to screening in four qualitative and one quantitative study.

Women found it difficult to present for screening when health facilities were not nearby, as reported in three quantitative and two qualitative studies. Waiswa and colleagues found that 32.9% of the women who had never been screened attributed this to not having a nearby facility [17]. Ndejjo et al. found that women who lived within a 5 km radius of a health facility where CCS was offered had a higher intention to screen [18].

Four qualitative and one quantitative study reported staffing shortages, lack of pathology services and limited health infrastructure as barriers to provision of CCS [11, 12, 14, 21, 24]. Lack of speculum equipment in some cases meant that women who presented for screening had to be turned away [21].

Time constraints and prohibitively long waits at health facilities were barriers in one quantitative and two qualitative studies. In Li et al’s cross sectional study, 27.8% of the women who refused screening did so because of time constraints [25].

Women in Ndejjo et al’s qualitative study reported that rude or insensitive HCWs were a disincentive to attend screening [12]. In one quantitative study, women who were not concerned about the gender of the HCWs performing the screening were 5 times more likely to have been screened compared with those who were [16]. HCWs reported that lack of training and skills for CCS among some of the clinical staff was a barrier to provision of CCS [24].

Financial costs associated with screening were a barrier for women in four included studies, and related either to the cost of the service or to associated transport/food costs [12, 19, 21, 24]. Twinomujuni et al. found that total costs for services were reported as prohibitive by 89.7% of the women in their survey [19].

In one quantitative study, women who had attended community outreach services for CCS were significantly more likely to have engaged with screening services [16]. There was no reference to outreach services in any of the other included studies.

To our knowledge, this is the first systematic review to focus on barriers and facilitators to uptake of CCS among women in Uganda. Data on factors that enable women to access screening is required to provide information about how CCS uptake may be improved and is of particular importance given that CCS uptake in Uganda is low in the setting of high CC and HPV incidence. This review focuses on the views of women as well as HCWs and thus contributes valuable information regarding the perspective of the target group for screening, as well as insights from professionals who provide this care. The collective evidence may guide the development of health promotion programs that incorporate the views of the target group.

While this review found general agreement among the HCWs and women in the included studies, and between women living in different regions, the small number of included studies limited a deeper understanding of district-specific barriers or facilitators. For example, post-conflict Northern Uganda has a large proportion of internally displaced women who likely have different competing priorities and may face different barriers to women in other districts. The small number of studies included in the review also meant that some barriers/facilitators were not identified. Furthermore, questionnaire types were often pre-established questions determined by the investigators, and some studies did not provide details regarding questions asked or themes explored. Depending on the way that questions were structured, relevant barriers and facilitators may not have emerged.

The quality of included studies was highly variable. In some studies the investigation of barriers or facilitators was not the primary outcome of the study. More studies specifically designed to address barriers and facilitators among women in Uganda are needed. Statistical assessment of reliability and validity of measurement tools was not at all evident, or only slightly evident, in eight of the quantitative studies. This limits the quality of findings from this review and signals a need for more rigorous study design in future studies.