Increasing Cervical Cancer Screening Uptake Among Emergency Department Patients

Invasive cervical cancer is preventable with adequate screening but screening rates are considerably below national goals. Emergency departments care for a disproportionate number of women who are not up to date with recommended cervical cancer screening. This study will evaluate the effectiveness of a mobile technology based behavioral intervention (using text messaging prompts) to increase cervical cancer screening uptake among emergency department patients.

Cervical cancer (CC) is preventable. Screening for CC aims to identify treatable lesions prior to the development of invasive CC which, without screening, is often first detected at a stage when comfort care is all that can be offered. Still, only 80% of U.S. women age 21 - 65 report adherence to U.S. Preventive Services Task Force (USPSTF) CC screening recommendations. This rate is considerably below the Healthy People 2020 target of 93%8. Lower rates of screening have been particularly pronounced among racial and ethnic minorities and patients with lower education levels. The group most likely to be non-adherent with screening guidelines is women who use the emergency department (ED) for their usual source of care. The ED setting, therefore, is optimal for the deployment of an intervention to promote CC screening. Although designed to address acute illnesses and injuries, the ED is an effective environment to advance preventive health. Numerous preventive health interventions have been successfully deployed in the ED including smoking cessation, HIV testing, and depression screening. Interventions delivered in the ED often suffer from lack of follow up and hence low adherence. Novel approaches are needed to reinforce health messages delivered in the acute context of the ED. Short Message Service (SMS) on mobile phones, also known as text messaging, is a low-cost, scalable, and effective means of delivering health behavior interventions. According to the Pew Research Center, the vast majority of Americans - 95% - own a cell phone, and an estimated 98% of all cell phones have texting capabilities. Evidence-based SMS interventions for health behavior include smoking cessation,primary care attendance, breast cancer screening, and sun safety, among others. However, minimal research has been conducted applying SMS interventions to CC screening and no prior research, apart from our pilot work, has targeted a CC prevention intervention by leveraging the universal healthcare access setting of the ED. In 2018-19 the investigators conducted a randomized clinical trial pilot (ClinicalTrials.gov #NCT03483610; NIH Grant ID: UL1TR002001) among ED patients non-adherent with USPSTF recommendations comparing an SMS intervention group to a referral-only control group to improve CC screening uptake at follow-up. The intervention consisted of a series of text messages, grounded in behavioral change theory, aimed at generating intention to get screened. The results of this pilot: (A) demonstrated the feasibility of our approach, (B) showed preliminary evidence of efficacy, and (C) provided an effect size estimate. Our overarching goal is to develop a low-cost, scalable SMS intervention that increases CC screening uptake among ED patients and can be deployed in heterogenous ED settings. The proposed randomized controlled trial will test the efficacy of this intervention. Step 1 of our approach is to identify whether the participant is adherent with screening guidelines. Step 2 is to randomize non-adherent participants to one of the two treatment conditions: (1) referral only (control) or, (2) referral and an SMS-intervention consisting of a series of text messages, grounded in behavioral change theory, aimed at generating intention and autonomous motivation to get screened. To limit costs and increase the scalability of the intervention, determination of adherence with screening guidelines via a self-administered questionnaire on a tablet computer will be evaluated. A total of 1,460 non-adherent women, age 21-65, will be recruited from a high-volume urban ED and a low-volume rural ED, randomized among study conditions, and followed-up at 150 days to assess interval CC screening uptake. Aim 1: Compare CC screening uptake between SMS intervention and control groups. Hypothesis 1: Uptake of CC screening in the SMS intervention group will be greater than in the control group at 150-day follow-up.This aim will determine the efficacy of the intervention. Aim 2: Compare the impact of in-person (using research staff) versus self-administered (using a tablet) determination of CC screening adherence on the efficacy of the intervention. Hypothesis 2: The method of determining adherence will not impact the efficacy of the intervention. This aim will evaluate an approach to limit cost and increase the scalability of intervention implementation. Aim 3: Identify mediators and explore moderators of intervention effects on CC screening uptake at follow-up. Hypothesis 3: Uptake in the intervention group will be mediated by theory-posited proximal predictors. This aim will elucidate the mechanism of behavioral change resulting from the intervention and identify differential effects among study sub-groups. This project leverages the universal access setting of the ED to target women most at risk for non-adherence with CC screening guidelines. A low-cost, scalable intervention that increases CC screening uptake in this population would decrease CC incidence and save lives. Our next step would be a multi-site effectiveness trial using the NCI Community Oncology Research Program (NCORP) Network.

Randomized controlled clinical trial with a prospectively collected convenience sample of 1,460 women nonadherent with CC screening guidelines. Women aged 21-65 will be recruited from a high-volume urban ED and a low-volume rural ED, assigned to study conditions, and followed-up at 150 days to assess interval uptake of CC screening (primary outcome). Electronic Health Record (EHR) review will be conducted to assess screening results and subsequent clinical endpoints. A two-step randomization process will be used to achieve a 2 x 2 factorial design.

Treatment group (receives text messages), initial screening measures conducted in-person in the Emergency Department.

Treatment group (receives text messages), initial screening measures conducted on their own via an iPad.

SMS-intervention consisting of a series of text messages, grounded in behavioral change theory, aimed at generating intention and autonomous motivation to get screened.

The investigators will perform logistic regression comparing the treatment and control conditions on CC screening uptake (defined as either receiving CC screening or scheduling CC screening) at followup. The use of logistic regression, rather than a χ² test of independence, allows for (a) the inclusion of potential relevant covariates in the unlikely case of baseline group differences and (b) the examination of moderation/interaction in later primary analyses. Given the sample size of 1,460 non-adherent participants, the investigators will have power of 0.80 to observe an improvement in CC screening uptake of 19% as observed in the pilot study (i.e., 36% vs 43%). This is a conservative estimate of the anticipated effect size given the increased intervention intensity, incorporation of additional theoretically informed messages to participants, and the above-mentioned failures in randomization that occurred with our small pilot sample.

The investigators will perform a 2 X 2 χ² of individuals in the treatment condition (n = 730) examining differences in CC uptake across method of determination. These data will be analyzed as a non-inferiority trial, such that we will have power greater than 0.80 to show that the Self-Administered method for determining adherence does not negatively impact intervention efficacy (up to a 9.1% difference). The investigators will also perform logistic regression on the complete non-adherent sample using intervention condition, method for determining adherence, and the interaction between condition and method to predict CC screening uptake, with a null finding for the interaction (meaning method does not impact the effect of the intervention) supporting the case for non-inferiority of the Self-Administered method.

Inclusion Criteria: female age 21 - 65 years demonstrating decisional capacity to consent to participate Exclusion Criteria: past hysterectomy with cervical removal the absence of a cervix (i.e. in patient that is a transwoman known infection with HIV (as screening recommendations for women with HIV differ from the general population) inability to consent (e.g., lacking decisional capacity, intoxicated, or in distress) non-English/Spanish/American Sign Language (ASL) speaking (Spanish and ASL interpreter services are available 24/7 in our system and will be paid for through grant funds)