Association between patient-reported HIV status and provider recommendation for screening in an opportunistic cervical Cancer screening setting in Jos, Nigeria

Of the half million new cases of invasive cervical cancer (ICC) reported globally each year, over 80% occur in Low-and Middle Income Countries (LMICs) [1]. Nigeria is one of these countries with a high burden of ICC incidence and mortality [2]. The Global Burden of Cancer 2013 ranked cervical cancer the 2nd most common in incidence and mortality among all cancers in Nigeria [3].

CCS is an important health care service intervention for reducing ICC incidence and mortality and its benefits are evident from data in developed countries, where organized CCS programs have resulted to a substantial decreases in ICC incidence and mortality [4‐10]. However, in Nigeria and other LMICs where organized CCS programs are lacking, the opportunity to have a screening test likely depends on several factors ranging from availability of screening, offering screening recommendations by providers, to health system support to overcome barriers to accessing services. The literature on cancer screening suggest that it is a process of care, consisting of several steps and interfaces between patients, providers, and health care organizations [11]. In this context, screening rates are largely driven by strategies that limit the number of interfaces across organizational boundaries; recruiting patients, promoting referrals, and facilitate appointment scheduling; and promote continuous patient care [11]. The organizational capability of the health care system to address these boundaries likely explains higher CCS rates (83%) in the US [12], in comparison to Nigeria and other LMICs in sub-Saharan Africa with much lower CCS rates between 6 and 8% [13, 14].

Indeed, we have an established body of literature on the effectiveness of provider recommendation for screening on CCS participation [11, 15‐18]. Also, we have evidence in Human Immunodeficiency Virus (HIV) infected populations, suggesting that women’s awareness that HIV infection increases the risk of ICC and having a strong provider-patient relationship were significant facilitators for CCS utilization [19]. Therefore, provider-patient discussions about CCS and offering referrals for such screening are critical because there is a high burden of HIV in Nigeria and also a high burden of HIV-associated precancerous abnormalities of the cervix and ICC in HIV infected population [20‐22]. We, however, do not understand the relationship between patient-reported HIV infection and provider behavior in providing a CCS recommendation and referral during the care process particularly in settings where CCS is largely opportunistic. In brief, opportunistic screening is dependent on a woman or her healthcare provider taking the initiative to do a pap test [23]. Indeed, strategies to improve early detection of cervical cancer through screening have focused either on opportunistic screening requested by a provider or an individual, or organized CCS in which a defined population is contacted and invited to screen at regular scheduled intervals [24]. We therefore hypothesized that in opportunistic screening settings, women with reported HIV infection are more likely to receive a provider-referral for CCS than women who are HIV uninfected. This manuscript reports on patient-reported HIV status and other socio-demographic factors associated with provider-referral in women at first CCS in an opportunistic screening in Nigeria.

During the 10-year study period (2006 to 2016), there were 17,016 records of women who received CCS services using the Pap test at the OSCC. Since this cross-sectional analysis focused on data records of women at first CCS, multiple follow-up entries were excluded from this analysis (see Fig. 3 for details of sample derivation).

Therefore, a final study sample of 14,088 women was utilized for this analysis (Table 1 provides detailed descriptive characteristics of the sample population at first CCS). The median age at first CCS in the sample population was 37 years (IQR, 30–45) and a mean of 38.1 years ±10.1. A total of 703 out of 14,088 women reported their HIV status as infected leading to patient-reported HIV prevalence of 5.0% (95% CI: 4.6, 5.4). The proportion of women who received provider referral was 50.1% (95% CI: 49.2, 50.5), while women who received CCS by self-referral was 49.9% (95% CI: 49.1, 50.8). When we compared the proportion of women who received a provider referral by patient reported HIV status, we found that 68.7% (95% CI: 65.3, 72.1) of women with HIV received a provider referral compared to 49.2% (95% CI: 48.4, 50.1) of women who were HIV uninfected (p-value < 0.001). The baseline characteristics of the study sample is summarized in Table 2.

Also, the unadjusted odds ratio for receiving a provider referral if HIV infected was 2.27 (95% CI: 1.92, 2.68) compared to being HIV uninfected. The unadjusted odds ratio for receiving a provider referral for other patient-reported socio-demographic and risk factors have been summarized in Table 3. In the final model, adjusting for smoking, age at first coitus < 22 years, age at first CCS ≥ 35 years, parity ≥5, history of condom use for sex, and 7–12 years of completed education, women who reported HIV infected were 2.35 times more likely to receive a provider referral for their first CCS compared to women who were HIV uninfected (aOR = 2.35; 95% CI: 1.95, 2.82, p = 0.001). Other socio-demographic factors that were independently associated with provider-referral for first cervical cancer screening were age ≥ 35 years (aOR = 1.25, 95% CI: 1.15, 1.35), parity ≥5 (aOR = 1.18, 95% CI: 1.09, 1.28), age at first sex ≤22 years (aOR = 1.27; 95% CI: 1.16, 1.39), smoking history (aOR = 3.20; 95% CI: 1.67, 6.12) and use of condoms (aOR = 1.47; 95% CI: 1.28, 1.70). We also found that women who reported completing 7–12 years (grade 7 to high-school) of education were less likely to receive a provider-referral than women with less than 7 years (equivalent to grade 6 or less) of completed education (aOR = 0.77; 95% CI: 0.71, 0.84). These results have been summarized in Table 3.

The principal findings in this analysis showed that women who reported being HIV infected were more than 2 times more likely to be referred by a provider at the time of first CCS compared to women who were HIV uninfected (aOR = 2.35; 95% CI: 1.95, 2.82). We also found that women who had completed at least 7–12 years of education were less likely to received provider-referral compared to women of lower educational attainment (aOR = 0.77; 95% CI: 0.71, 0.84). Other socio-demographic factors that were significantly associated with provider-referral for first CCS were age ≥ 35 years, parity ≥5, age at first sex ≤22 years, smoking history, and use of condoms.

The finding that patient-reported HIV was significantly associated with provider-referral at first CCS is particularly important since several studies have found that HIV infected women are at greater risk of developing precancer and ICC [21, 30‐34]. Because of this risk, studies have demonstrated the critical role of health care providers in linking HIV infected women to important reproductive health services including CCS in areas with a high HIV burden [35]. The finding that patient-reported HIV was associated with provider-referral in our study population is not surprising giving that the OSCC is located in one of the tertiary health institutions supported by Presidential Emergency Plan for AIDS Relief (PEPFAR) and one of the largest facilities that offers HIV care and treatment in West Africa. It is possible that providers offering HIV care to these women are aware of both the risk of cervical cancer in these women and the availability of screening services in the facility, thereby offering referral to such high-risk population. Prior reports on the role of medical care providers facilitating CCS for HIV infected women receiving care in the same facility with a gynecologic care provider has been well documented in a U.S. HIV population [36], and it has been recommended that HIV care be integrated with gynecologic care, and educating clinicians to recommend CCS to these women could significantly improve adherence and utilization of CCS [37, 38].

Although, the prevalence of smoking is reportedly very low (0.6%) in our study population, women with reported history of smoking were more likely to have received a provider-referral for their first CCS. This could be related to provider’s knowledge that smoking is a risk factor for cervical cancer [39‐43] thereby initiating counselling and offering screening referrals for such high-risk women. Other demographic factors such as age and parity have been documented in previous studies as epidemiologic risk factors for cervical precancer and progression to ICC [44]. These demographic factors were also found to be significantly associated with a higher likelihood for receiving a provider-referral at the time of first CCS in our study population.

The finding that more educated women were less likely to receive provider-referral for first CCS is interesting. This possibly implies that more educated women were more likely to be aware of CCS and more likely to self-refer themselves for screening compared to the less educated. This explanation has some plausibility given the findings of recent systematic review of qualitative studies on the barriers to utilization of Pap screening in sub-Saharan Africa [45]. The systematic review reported client factors such as lack of knowledge and awareness about Pap smear among the barriers for low utilization of Pap screening [45]. Also related to provider-referral, the study found that provider barriers such as failure to inform or encourage women to screen were important provider factors contributing to low Pap utilization in sub-Saharan Africa [45]. A related systematic review also recommended improvement in cervical health education, addressing cultural beliefs and practices, spousal support, provider attitude and addressing the problems of cost and physical access to CCS services as interventions to improve screening utilization in sub-Saharan Africa [46].

Our findings suggest that women with known potential risk factors for cervical cancer such as HIV infection, multiparity, and smoking were more likely to be referred for screening by providers. This is an important finding for opportunistic CCS, and implementation of screening guidelines in such settings should encourage providers to assess potential risk factors in women accessing routine clinical care and those with such reported risk factors should be given screening referrals and encouraged to receive CCS. Studies have also identified that interventions that increased discussions between providers and women, educating women on the benefits of screening, and allaying their fears on possible screening outcomes are significantly associated with participation in screening [47].

The strength of our study findings is related to the relatively large sample size of our study population spanning a decade of CCS services offered in an opportunistic setting in a cosmopolitan Nigerian city that also offers care to a large population of HIV infected adults in West Africa. To the best of our knowledge this is the first secondary analysis of CCS data in Nigeria focusing on understanding the contributions of providers in utilizing CCS services. We feel that our findings could be generalized to other settings in West Africa with ongoing HIV care and availability of opportunistic CCS services. We however, recognize and acknowledge the limitations of self-reported risk factors in this analysis. It is possible that some women may have concealed some information that could affect the internal validity of our estimates.

Our future research will focus on elucidating provider and patient perspectives on the facilitators and barriers to CCS in an opportunistic screening setting using qualitative research methodology. For instance, it will be appropriate to understand the perspectives of providers on implementation of CCS guidelines in Nigeria and the factors that could enhance adherence to such guidelines in various practice settings. It is also important to understand patient perspectives on the acceptability of male providers performing pelvic examinations and alternative screening methods such as self-sampling for HPV testing in various cultural settings in Nigeria.

Women who reported having known risk factors for cervical cancer such as HIV infection, multiparity, and smoking are significantly more likely to received recommendation and referral for CCS from a health care provider during routine clinical care visits. More educated women with at least 7–12 years of education are more likely to be aware of CCS service and self-refer for screening than women who have lower education. Despite the limitations identified in this secondary data analysis, healthcare providers in Nigeria should be encouraged and supported to make CCS counseling, recommendations and referral during routine clinical care to eligible women who have not had a CCS. This is particularly important in the current Nigerian setting where HPV vaccination is not supported and the opportunity for screening is largely dependent on provider-initiated counseling and screening either by the provider or through referral to facilities offering a screening service.