Recruitment of Minority Adolescents and Young Adults into Randomised Clinical Trials: Testing the Design of the Technology Enhanced Community Health Nursing (TECH-N) Pelvic Inflammatory Disease Trial

Abstract

Purpose

Pelvic inflammatory disease (PID) disproportionately affects adolescent and young adult (AYA) women and can negatively influence reproductive health trajectories. Few randomized controlled trials (RCTs) have focused on strategies to improve outpatient adherence or to reduce reproductive morbidity in this population. This paper describes the research methods and preliminary effectiveness of recruitment, retention, and intervention strategies employed in a novel RCT designed to test a technology-enhanced community-health nursing (TECH-N) intervention among urban AYA with PID.

Methods

AYA women aged 13–25 years were recruited during acute PID visits in outpatient clinics and emergency departments (ED) to participate in this IRB-approved trial. Participants completed an audio-computerized self-interview (ACASI), provided vaginal specimens, and were randomized to standard treatment or the intervention. Intervention participants received text-messaging support for 30 days and a community health nurse (CHN) interventionist performed a home visit with clinical assessment within 5 days after enrollment. All patients received a full course of medications and completed research visits at 14-days (adherence), 30 days and 90 days with by an outreach worker. STI testing performed at the 30-and 90-day visits. Exploratory analyses using descriptive statistics were conducted to examine recruitment, retention, and follow-up data to test the overall design of the intervention.

Results

In the first 48 months, 64% of 463 patients were eligible for the study and 81.2% of 293 eligible patients were recruited for the study (63.3%); 238 (81.2%) of eligible patients were enrolled. Most participants were African American (95.6%) with a mean age of 18.6 (2.3). Ninety-four percent of individuals assigned to the TECH-N intervention completed the nursing visits. All completed visits have been within the 5-day window and over 90% of patients in both arms have been retained over the 3-month follow-up period. Biological data suggests a shift in the biological milieu with the predominance of Chlamydia trachomatis, Mycoplasma genitalium, and Trichomonas vaginalis infections.

Conclusions

Preliminary data from the TECH-N study demonstrated that urban, low-income, minority AYA with PID can effectively be recruited and retained to participate in sexual and reproductive health RCTs with sufficient investment in the design and infrastructure of the study. Community-based sexual health interventions appear to be both feasible and acceptable in this population.

Introduction

Pelvic Inflammatory Disease (PID) is a serious reproductive health disorder and disease rates remain unacceptably high among minority adolescent girls and young adult (AYA) women.^{1, 2} Each episode of this upper reproductive tract infection, which is usually caused by a sexually transmitted infection (STI), increases the risk for sequelae such as tubal infertility, ectopic pregnancy, and chronic pelvic pain (CPP). ^3–8 Care recommendations for PID has shifted from inpatient to the outpatient settings in response to efficacy data from the Pelvic Inflammatory Disease Evaluation and Clinical Health (PEACH) study.⁹ Unfortunately, regardless of the treatment strategy, recent age-stratified analysis from the PEACH data demonstrated that AYA women have high rates of repeat STIs and CPP. Twenty percent of women less than 25 years experienced a repeat or persistent STI at 30 days and/or repeat PID over the 7-year follow-up period, with a 5-fold risk of CPP.¹⁰ Further, our local data has shown poor adherence to the 3-day follow-up visit recommended by the Centers for Disease Control and Prevention (CDC), and increased risk for an STI at 90-days. While brief interventions in the acute care setting have had a positive effect, adherence to self-management is problematic.¹¹

Inpatient treatment for PID is expensive without incremental increases in effectiveness compared with outpatient treatment, therefore^12–15 cost-effective outpatient PID care supports are needed to improve reproductive health outcomes for this vulnerable population.^{10,16, 16, 17} Previous work from our group demonstrated that young women perceive significant health related quality of life reductions and desire more closely monitored care for PID by clinicians was informative for strategically guiding the intervention we advanced herein.^{14, 18}

Research has demonstrated that community health nurse (CHN) interventions can increase access to resources, enhance health care utilization, and promote risk-reducing behavior.^19–21 We postulate that integrating a technology component with CHN care would increase appeal to AYA women given data showing that use of SMS messaging enhances improves attention to medical visits, medication adherence, and health communication.^{22, 23} We hypothesized that repackaging the recommended CDC-follow-up visit using a technology-enhanced community health nursing intervention (TECH-N) with integration of an evidence-based STI prevention curriculum and SMS communication would reduce repeat infection by improving adherence to PID treatment and reducing unprotected intercourse and would be more cost-effective compared with standard of care.

We are currently enrolling AYA diagnosed with PID and randomizing them to receive CHN clinical support using a single post-PID face-to-face clinical evaluation and SMS communication support during the 30-day period following diagnosis (intervention group) or standard of care (control group). In this paper, we discuss the design and methods of this novel randomized clinical trial (RCT). We also describe the preliminary 48-month effectiveness of recruitment and retention strategies being employed in this research designed to reduce recurrent STIs that result in adverse reproductive health outcomes after PID. ^{10, 17, 24}

Conceptual Framework

This work builds on our previous studies with AYA with PID demonstrating inadequate interface with the healthcare system and poor adherence to treatment.¹¹ For this work, we employ the integrative model of behavioral prediction²⁵ to frame our approach. Our previous work has allowed us to demonstrate that the distal (demographics) and intermediate factors (self-efficacy) do not sufficiently allow us to predict behaviors. Despite issues with adherence, AYA express high treatment value and self-efficacy.²⁶ Capitalizing on the positive attitudes towards treatment, we are attempting to change behavior by 1) affirming positive outcome expectations, 2) enhancing sexual health skills e.g. condom use), and 3) removing environmental constraints (e.g. transportation).⁹ We employ comprehensive structured CHN care that includes: 1) standardized prevention case-management components, 2) an effective one-on-one intervention for STI behavior change^{27, 28} and 3) PID-related health SMS messages as core components of the TECH-N intervention.

Methods

Study Overview

The TECH-N team is actively enrolling 350 AYA with mild-moderate PID in a 2-arm, single-blinded RCT. Participants are recruited from a large, academic urban hospital system located in an STI-prevalent community in the United States. Patients enrolled in Arm 1 (TECH-N intervention group) receive standard of care according to the CDC guidelines, receive a full course of medication, a welcome SMS message followed by daily medication reminders and tri-weekly messages following the 14 day treatment period up to 30 days, community-based visits by a TECH-N at 3–5, 14, and 30 days following their PID diagnosis. In Arm 2 (control group), adolescents will also receive standard of care treatment including a full course of medications comparable to the intervention group, but are expected to arrange their own 3–5 day follow- up according to standard practice. All participants participate in follow-up interviews at 14, 30, and 90 days. STI testing occurs at the 30-day and 90-day study visits in accordance with timelines for expected disease clearance and re-infection.²⁹

Setting and Participants

The study is being conducted in the Baltimore, Maryland metropolitan area, which currently ranks 10^th in the nation for incident HIV/AIDs infection²⁹ and 17^th for Chlamydia trachomatis (CT) among its citizens.^30,31 The Maryland Department of Health and Mental Hygiene has determined that every county is a Chlamydia hotspot and county maps demonstrate that Baltimore is the most densely affected.³¹ Patients are recruited at the time of diagnosis from the Pediatric and Adult Emergency Departments (ED), General Pediatric, and AYA outpatient clinics.

Inclusion and Exclusion Criteria

At enrollment, trained research staff performs a detailed review of the study as a part of the informed consent process and patients are screened for eligibility. Inclusion criteria include age 13–25 years, diagnosis of mild-moderate PID with a disposition to outpatient treatment, residence in the Baltimore Metropolitan area, and willingness to sign informed consent and to be randomized. Patients who are pregnant, have a concurrent diagnosis of sexual assault, or are unable to communicate with staff due to cognitive, mental, or language difficulties are ineligible.

Enrolled participants complete an audio computerized assisted self-interview (ACASI) to collect baseline information and are randomized to the TECH-N or control group using a permutated block design.³² Patients also provide a collection of a vaginal swab for STI testing. After completion of initial study procedures and instruction, patients receive remuneration of $10 and a 14-day course of medication dispensed by the clinician.

TECH-N Intervention

The TECH-N intervention includes follow-up by a CHN trained in clinical assessment, indications for physician referral, disease intervention protocols for partner notification and treatment, the CDC guidelines, and the TECH-N protocol. The CHN uses the Sister-to- Sister Teen Intervention® that has a 20-minute one-on-one module to guide the patient through skill-based risk reduction counseling^27,28 during home visits 3–5 days after enrollment consistent with other published PID trials.^33–35 Given that only 20% of AYA adhere to follow-up visits, ¹¹ this window proved to be justifiable. Intervention participants are also enrolled in Reify Health Responsive SMS system® that delivers a welcome message, a prompt to schedule the appointment for the 3–5 day CHN, and daily medication reminders for 14 days. After 14 days, the patient also receives positive health reminders for the rest of the month. While 95% of AYA women in our prior research had a cell phone for personal use, many low-income patients have difficulty with monthly maintenance. ^{22, 23} Participants who are assigned to the intervention and do not have a cell phone are provided one for temporary use.

Control Condition

In the control condition, patients receive standard of care per the CDC guidelines.³⁶ The institution has integrated standard PID management protocols into usual clinical care as a part of continuous quality improvement protocols.³⁶

Disposition

After completion of informed consent, enrollment procedures, baseline ACASI survey, and specimen collection, the clinician dispenses study-supplied doxycycline 100 mg twice daily for 14 days with or without metronidazole 500 mg twice daily for 14 days and provides discharge instructions for self-care. Clinicians also have access to Azithromycin 1 g given once orally q week to be administered with metronidazole 500 mg taken orally twice daily for 14 days through the study, for the rare patient with doxycycline allergy.^{9, 37}

Measures

Interview Data

All participants are interviewed by an outreach worker at 14 days post-enrollment to assess residual symptomatology and adherence to the recommendations for self-care. ^{9, 37} A pill count is performed if the original medication bottle is available at the time of the interview. Patients with ongoing symptoms are discussed with a TECH-N team clinician and referred for care as needed.

In addition to the baseline visit, ACASI is used to collect data at the 30-day and 90 day visits. The measures on the baseline survey contain: demographics, reproductive and sexual history, PID adherence self-efficacy and perceived barrier scales ^{26, 7} social provisions scale,³⁸ and the short-form survey instrument (SF-12) as a measure of health-related quality of life. ^39,40 The 30-and 90-day ACASI surveys captures interim sexual behavior, condom use, and contraceptive use, interim diagnosis of STIs and pregnancy, and health status.

TECH-N CHN Visit Data

The CHN nurse records contact tracing (e.g. number of contacts to arrange and complete visit) and clinical assessment data including a pain rating scale, abdominal exam, medication usage, supportive care, side effects, activity level, and patient teaching outcomes.

Costs

The costs of administering the intervention is based on hours worked, miles traveled, other supplies, and overhead. The primary cost components for patients in both arms include the initial treatment for PID, follow-up therapy for treatment failure, treatment for recurrent STIs and PID, CPP, ectopic pregnancy, time lost from work/school/household management, travel to PID-related care, and therapy or evaluation for infertility. These data will be used to estimate the direct (medical costs) and indirect (employment, impairment) costs associated with PID for the TECH-N and control arms. Health service utilization will be derived from the longitudinal interviews, care utilization data, prescription and over-the-counter medication use, laboratory services, and radiological services. Productivity loss will be derived using the Workplace Productivity and Activity Impairment (WPAI).^41,42

Biological Measurements

Participants are asked to provide a self-collected vaginal swab (SVS) for Mycoplasma genitalium (MG) and Trichomonas vaginalis (TV) at baseline to accompany routine testing at the clinical site. Neisseria gonorrhoeae (GC), Chlamydia trachomatis (CT), TV, and MG are also obtained at 30 days and 90 days and processed at the Johns Hopkins University International STD and Respiratory Research Laboratory. Vaginal specimens are processed and tested according to the APTIMA Combo 2 manufacturer’s package inserts for CT, GC, and TV.⁴³ The APTIMA urine-based assay was tested using the GEN-PROBE transcription mediated amplification (TMA) research assay in the same manner as for the other APTIMA assays. This assay targets MG rRNA for detection in genital specimens.^44–48

All participants with positive STI tests are contacted by a research nurse practitioner, notified of their positive test result, and referred for treatment. Positive cases of GC/CT are reported to the local health department. The TECH-N encourages partner notification and treatment and provides support and resources to assist all participants.

Incentives

Participants receive a $10 gift card at enrollment and $10 for each completed face- to-face research visit (14, 30, and 90 days) plus an addition $10 for each STI sample (N=3). AYA in the TECH-N intervention group do not receive remuneration for the 3–5-day clinical care visit because adherence to the CHN clinical visit is a non-incentivized behavior under study. Participants receive $5 remuneration for notifying us of changes in cell phone number and/or other contact information.

Statistical Analyses

Sample Size

Sample size calculations were driven by the results of our prior research using longitudinal data to examine repeat STIs/PID after an initial diagnosis of PID. At 3 months, the STI positivity rate was 25%²⁴ and thus we assumed this infection rate for the control group. Preliminary data also suggested that we will be able to recruit approximately 175 participants in each arm for a total of 350 participants. We anticipated about 30% attrition over the study period thus each study arm will have an effective sample size of 122.5 participants, for a total of 245 participants. Using these assumptions, our power calculations suggest that we will have 80% power to detect a relative risk of 1.66.

Data Analysis

The TECH-N study will compare the effectiveness of the intervention to standard of care using an intention-to-treat analysis. We will conduct a logistic regression analysis 1) to determine whether the intervention group showed greater short-term adherence rate compared to the control group and 2) to determine whether the intervention group showed a lower rate of recurrent STIs at 90 days compared to the control group, including potential factors that may affect the difference in rates for both models.

The economic evaluation will examine the net difference in costs between TECH-N and control groups in relation to prevention of 1) recurrent STIs and 2) PID-related complications. The difference in costs will account for the net difference between the outpatient strategies in use of medical services and indirect costs for PID treatment. We will also consider savings in medical-care costs and indirect costs from prevention of the outcome variables. A cost-benefit analysis can be performed comparing the excess costs of TECH-N with the medical care savings and productivity gains from cases averted. We can extend the analysis to a cost-effectiveness evaluation if TECH-N remains more expensive even after accounting for medical care costs and productivity loss averted. In this case, we will measure the dollars spent per case of STI averted over the relevant period.

Study Update

In the first 48 months, 293 of 463 patients were eligible for the study (63.3%) and 238 (81.2%) of eligible patients were enrolled. (Figure 1) The mean age of participants was 18.6 (SD 2.3). Most participants were low-income (Medicaid/self-pay (77%), African American (95.6%), and resided in female-headed households with maternal education level as high school or less. Patients who declined study participation were older (19.8.versus 18.6 years (p<0.05), but otherwise were demographically similar. To date, only 14 of intervention patients (11.6%) required a study cell phone for participation. To date, 19 patients were lost to follow-up, but were demographically indistinct (age, race/ethnicity, insurance status) from those who completed the intervention. Ninety-four percent of individuals assigned to the TECH-N completed the intervention. The average number of contacts to execute the CHN visits was 2.46. Almost all eligible participants completed their 14-day, 1-month, and 3-month visits [95%, 96% and 93%, respectively]. Mean number of days to complete the 14-day research visit was 1.1 [SD 0.55] days. Baseline laboratory results reveal infection rates that are highest for CT (26%) and MG (21%), followed by TV (16%) and GC (8%). While MG is associated with PID,³³ recent studies have suggested low rates of infection in adolescents⁴⁹ and there is currently no commercial laboratory test for MG to diagnose patients in the United States outside of the research environment. Our data will allow for a longitudinal analysis that may drive differential approaches to testing and therapy.

Figure 1.

Outcomes of recruitment, randomization, and retention efforts within the TECH-N Study

Summary

PID remains a common reproductive health disorder disproportionately affecting AYA women in urban minority communities. We have demonstrated that using the TECH-N design, AYA can effectively be recruited during acute PID treatment visits to participate in a longitudinal RCT and can be retained over time. Further, sexual health counseling and clinical follow-up interventions for STI management can be reliably administered in the home by CHNs. While PID is a polymicrobial disorder, the baseline rates of STI positivity suggest a shifting biological milieu in this cohort of AYA women. Moreover, we are reaching the target population of AYA at high risk for recurrent STIs for intervention.

The strengths of this study include utilization of evidence based intervention materials (Sister-to-Sister Teen®) and innovative communication tools (Reify® SMS)) for intervention delivery and embedding the study in an existing clinical infrastructure. This includes having access to electronic health records to track patient visits during recruitment and the follow-up of results and care outcomes, full time research staff to work with clinical staff in recruitment sites, and service provision (outreach, STI testing, CHN visits) as a part of the research protocol. As such, TECH-N research staff is viewed as an extension of care.

Staffing of the study also has been a critical component for the success of this program. In our prior research patients with PID, we demonstrated high recruitment rates, but retained only 70% of the sample.¹¹ Investment in both recruitment and outreach staff has ensured that patients can effectively be tracked in the field without undermining the onsite team. Staff training and team building has resulted in a knowledgeable and high functioning interdisciplinary team. The team is also diverse, but most field staff members are women, further elevating the ‘Sister-to-Sister’ concept. Our research staff has been described as personable, culturally sensitive, patient, nonjudgmental, supportive, committed, and persistent in terms of their execution of the protocol and follow-through with patients in complex social circumstances.

We also recognize the limitations due to the limited generalizability of our sample due to the demographics and epidemiology of STIs of Baltimore.^29–31 Despite this; our work has the potential to identify an alternative, cost-conscious strategy for addressing the observed disparities for AYA with PID. As with the PEACH trial,³⁴ the confounding effects of race/ethnicity and socioeconomic status are difficult to parse out because the burden PID is predominately borne by low-income minority women. The PEACH trial group found that younger and lower income women were hard to recruit and often excluded.³⁴ We actively seek out these women and attempt to overcome the latter by optimizing our design.

Ultimately, this research demonstrates that inclusion of urban, minority AYA women in large community-based sexual health RCTs is highly feasible. It also suggests that adequate investments in developing the research infrastructure are critical for execution of RCTs designed to reduce PID-related health disparities among vulnerable populations in communities with high STI prevalence.