Practice Facilitation Improves Adolescent Reproductive Health Preventive Services in Primary Care

PFK has medical and financial responsibility for >330 000 pediatric Medicaid recipients throughout central and southeast Ohio. PFK has direct contracts with >2100 providers in independent and Nationwide Children’s Hospital–employed practices that care for pediatric Medicaid recipients. The majority of PFK physicians are salaried employees of the hospital or practice partners, whereas the remaining community physicians receive Medicaid fee-for-service rates plus incentive payments.³¹ 

PFK has an active QI coaching program composed of 5 quality improvement specialists (QISs) who use the practice facilitation model to support QI capacity building in affiliated primary care practices. In late 2016, the QI coaching program, supported by an adolescent medicine specialist, established the Reproductive Health Initiative (RHI) to have primary care practices focus on improving access to reproductive and sexual health care. The RHI was publicized to all practices providing care to ≥100 pediatric Medicaid recipients in central and southeastern Ohio (N = 157) through mailers and PFK’s Web site. Practices that previously expressed interest in adolescent sexual and reproductive health were invited via personal communication (n = 7). Interested practices could participate in the RHI at any time during the study period. Between November 2016 and December 2019, 6 practices elected to participate and completed a memorandum of understanding and business associate agreement, enabling data exchange between the PFK QISs and the practice. Each practice identified its own internal QI team, usually consisting of a physician leader along with nurses, medical assistants, and administrative staff.

All practice QI team members underwent a 3-hour training session based on the Institute for Healthcare Improvement Model for Improvement, including baseline measurement collection, project aim identification, and key driver diagram (KDD) development. Although the practices’ QI teams were responsible for implementing interventions, the PFK QISs held regular practice facilitation meetings to collect and interpret data, monitor progress, and address challenges. The PFK QI team also met with the adolescent medicine specialist routinely to review progress within individual practices and among the group of RHI-engaged practices.

RHI practices selected to focus QI efforts on (1) improving sexual and reproductive health assessment (RHAs) during adolescent WCVs and/or (2) increasing access to etonogestrel contraceptive implants within their practice. Practices identified at least 1 project aim that was specific, measurable, achievable, relevant, and time bound.³²  The PFK QISs facilitated internal team discussions and the use of relevant QI tools such as process maps and cause-and-effect diagrams to help each practice develop its own KDD. The PFK QI team and adolescent medicine specialist promoted shared learning opportunities, including webinars and site visits, among the RHI practices, similar to the approach in the Institute for Healthcare Improvement’s Breakthrough Series. Change concepts and/or interventions were tested and modified through plan-do-study-act cycles. Slightly different from the Breakthrough Series, the RHI emphasized customization of projects at the practice level, allowed for rolling enrollment over the study period, and let practices determine their own time line and pace of interventions. Supplemental Figures 3 and 4 depict KDDs created by 1 practice focused on RHAs and another on etonogestrel implants.

Patient-related factors, including promoting patient and family awareness about adolescent health and enhancing the patient’s clinic experience, were an important driver identified across all RHI projects. Interventions aimed at this driver included updating clinic policies and procedures to ensure that adolescents (1) attend WCVs yearly, (2) can meet confidentially with clinicians, and (3) receive verbal and written information on contraceptive options and STI prevention.

Clinical staff knowledge about and comfort with sexual and reproductive health-related topics was another driver that crossed both project measures. The adolescent medicine specialist and PFK-affiliated pharmacists offered educational sessions and materials on sexual health, contraceptive counseling, and contraceptive prescribing to interested providers and personnel. Interventions adopted by each practice are shown in Table 1.

To further support consistent RHA implementation during adolescent WCVs, interventions targeting clinic process standardization and enhancing clinical information systems were developed. Several practices provided patients with a comprehensive questionnaire on sexual history and reproductive health planning (Supplemental Information) that was then used to guide provider-patient discussions. Other practices had adolescents complete a brief screening survey and then solicited more information during the clinical encounter. Most practices found that having patients complete the questionnaire in the waiting room worked best for workflow, although 1 practice had its sole clinician administer the screen verbally during the WCV. Another workflow change pursued by some practices was specimen collection for STI testing and point-of-care pregnancy testing. One practice worked with its PFK QIS to review pharmacy claims data to track individual patient adherence to contraceptive prescription refills, with poorly adherent patients targeted for outreach.

The PFK QISs helped clinicians receive training in etonogestrel implant insertion as necessary. Strategies to develop new office operations and reimbursement procedures were critical to ensure practices could provide etonogestrel implants on-site. Interventions included technical assistance related to the provision of equipment for etonogestrel implant procedures, development of electronic medical record (EMR) procedure note templates, and addition of appropriate diagnosis and billing codes. One practice affiliated with a large health system partnered with nearby gynecology colleagues for assistance with program implementation, including patient referrals when appropriate.

From November 1, 2016, to December 31, 2019, data were collected by retrospective review of each practice’s medical records and administrative medical and pharmacy claims data available to PFK. Because of practices engaging in RHI at different times, baseline periods varied across sites. Each practice’s performance on its outcome measure(s) for the most recent 12 months before project initiation was used to determine individual baselines. After project initiation, data were collected monthly and shared with the project team to chart progress. To track RHA administration, a PFK QIS reviewed medical records for every adolescent WCV performed during the previous month to determine if a complete RHA was documented during the visit. For etonogestrel access, the number of etonogestrel implant insertion procedures billed by the practice each month was obtained by using either the practice’s EMR system or PFK claims (querying for Current Procedural Terminology codes 11981 [insertion, nonbiodegradable drug delivery implant] or 11983 [removal with reinsertion, nonbiodegradable drug delivery implant]).

For practices focused on RHAs, the outcome measure was the proportion of adolescent WCVs that had documentation of a complete RHA. Documentation requirements included these elements: (1) pregnancy risk assessment, including sexual history, pregnancy intention, and female menstrual history; and (2) contraception and STI risk reduction counseling. For sexually active female patients, RHAs also must have included (1) Chlamydia trachomatis and Neisseria gonorrhoeae screening within the past year, (2) a contraception plan if the patient was interested, and (3) pregnancy testing, if clinically indicated. For practices focused on etonogestrel implants, the primary outcome was the number of etonogestrel implants inserted per month.

Specific characteristics of the engaged practices were described, including practice setting, size, ownership model, and patient volume. In unstructured interviews, the PFK QI team collected qualitative information to contextualize the successes and challenges in recruitment and project implementation. To track progress in outcome measures and evaluate the impact of interventions over time, Shewhart control charts were updated monthly for each practice separately as well as collectively among RHI practices for the 2 primary outcome measures.³³  For the 2 primary outcome measures, the Shewhart charts are presented as time-series analyses, with 12 months of baseline data before and 18 months after project initiation for participating practices. Control chart rules, including the 8-point rule and aggregate point rule, were applied to identify centerline shifts representing special cause variation (defined as performance change outside of what would be normally expected on the basis of historical activity).^33,34  The 2-sample test for proportions and 2-sample t test were used to assess differences in the primary outcome measures, compared with baseline for individual practices and the RHI overall, with a P value <.001 indicative of special cause variation.

Institutional review board approval was not required for this project because it is not research on human subjects and met our institution’s definition of QI work.

As of December 31, 2019, 6 independent primary care practices serving >7000 adolescents 12 to 19 years of age implemented reproductive health-related QI projects as part of the RHI. Another practice joined in October 2018 but closed its office within 3 months of project initiation and was not included in the analysis. The participating practices varied in patient volume and insurance payer mix, office size, practice model, and location (Table 2). Practices chose to pursue a QI project focused on improving rates of RHA completion during adolescent WCVs (n = 1), increasing access to etonogestrel implant insertions (n = 2), or both outcomes (n = 3).

Among practices focused on RHA administration, the overall proportion of adolescent WCVs with RHAs increased from a baseline of 0% to 50.1% after 8 months, with subsequent improvement to 65.8% over the next 10 months (Fig 1; P < .001 for each centerline shift). Within each practice, the proportion of WCVs with RHAs improved significantly over time compared with baseline (Table 3; P < .001 for each). All 4 participating practices implemented the same 3 interventions within the first month of project initiation. One practice only had 12 months of data available for inclusion during the study period.

Across all practices focused on increasing access to etonogestrel implants, there was an increase from 0 etonogestrel implant insertions per month at baseline to 8.5 insertions per month (Fig 2; P < .001). Special cause variation resulting in a centerline shift occurred in the first month of project initiation. Four practices started providing etonogestrel insertions for interested patients within 6 months of project initiation, whereas it took 1 practice >12 months to place its first implant. Two practices (practices A and C) were particularly successful in demonstrating improvement over time, with special cause variation resulting in shifts in the mean number of etonogestrel implant insertions performed monthly (Table 3). The other 3 practices showed modest improvement over time, but centerline shifts on Shewhart charts were not observed (data not shown). All 5 practices participating in this project implemented the same 2 interventions within the first month.

Through unstructured interviews with practice team members to evaluate project performance, we observed that 4 practices with an identified project champion (practices, A, C, E, and F) successfully implemented more interventions over the initial 12 months, compared with 2 practices (practices B and D) without champions. Practices with strong project leadership described greater buy-in and support for the interventions among all office staff. Members of the 2 practices without clear project champions reported minimal progress on interventions between meetings with the QIS, and their projects were less successful. For practices that lacked an EMR, which was an identified barrier for effective data collection, the QIS had to perform manual chart reviews.

The use of practice facilitation to support QI initiatives substantially improved the delivery of adolescent preventive health services in pediatric community practices partnering with a large accountable care organization. Despite well-recognized challenges to the provision of reproductive health care to adolescents, all 4 practices focusing on RHA administration demonstrated significant improvement, with 2 achieving their aim. Of the 5 practices focusing on etonogestrel insertions, 2 practices successfully met their aim. Our real-world findings across diverse clinic settings suggest that pediatric practices interested in improving reproductive health care delivery can achieve their goals through a QI-driven approach.

Clear identification of project champions, regardless of their role in the office, was critical for helping practices implement interventions successfully and observe improvement in process and/or outcome measures over time. Whereas providers in most practices self-identified as project leads, the office manager in 1 practice was particularly effective in supporting this work. The importance of team leaders was further demonstrated by the inability of 2 practices that lacked clear on-site leaders to improve on access to etonogestrel implant insertion. These practices had difficulty maintaining project activity between meetings with the PFK QIS and thus struggled with demonstrating improvement over time.

Challenges with data collection were a barrier for some practices. One practice lacked an EMR, requiring its team to either perform manual reviews or wait for time-delayed administrative claims. The PFK QISs helped support this practice by performing on-site chart reviews so that the project team could focus on implementing interventions. Assessment and documentation of sexual activity in pediatrics has been found to be variable,³⁰  with no common criterion across practices and health systems. Improvements in the systematic measurement of adolescent sexual activity coordinated across EMR platforms would facilitate QI measurement in projects like ours that are attempting to improve adolescent reproductive health care delivery.

This study reveals that practice facilitation using QI coaching is an effective way to increase adherence to national guidelines for adolescent reproductive health preventive services and improve contraception provision, especially highly effective contraceptives, within primary care pediatrics.

In our analysis, we did not differentiate among individual intervention effects to determine their relative impact on progress in the primary outcome measures, so the effectiveness of all interventions is not known. However, practices typically implemented the same few interventions within the first month of project initiation, resulting in most practices seeing rapid improvement. Other interventions that were pursued later in the project had variable impact, as they were implemented differently in each practice. Similarly, we did not have access to patient-level variables, such as race or ethnicity and socioeconomic status, which may have influenced completion of RHAs and contraception provision. Patient-level outcomes other than etonogestrel implant insertion were not assessed; however, one might reasonably expect there were additional unmeasured desirable outcomes, including increased STI screening, prescribing of other contraceptives, and preconception care.

Our study has certain limitations. We used a nonexperimental design, and our approach was explicitly iterative, so participating practices were not blinded to outcomes. Generalizability might be limited by selection bias of early adopters to voluntarily participate in this project. Moreover, results may not be generalizable outside our geographic area or with practices that do not have a similar partnership with an organization such as PFK. Balancing measures, such as examination of etonogestrel implant removal rates and assessment of patient satisfaction or person-centered contraception counseling, were not assessed in participating practices but represent important variables to capture moving forward. Finally, our QI coaching approach rooted in practice facilitation requires a substantial investment of resources and the availability of technical QI expertise. Some health systems and practices will not have the capacity to use this approach to improve adolescent reproductive health care.

We credit the project’s success to the practice-specific customization of interventions and ongoing, hands-on support from the QISs. Further study using practice facilitation to improve adolescent reproductive health care across different settings will support the generalizability of the approach. Additional examination of individual interventions and patient-level factors may allow practices to better prioritize their efforts.

Building on the success of initial collaborations between the PFK QI coaching team and RHI-engaged practices, we plan to continue supporting practices through practice facilitation to enable them to expand their QI efforts. Further opportunities for project growth may include evaluation of the adolescents’ experiences with different forms of contraceptives and likelihood for adherence, the impact of sexual health counseling on STI rates, and development of targeted interventions to address potential health disparities by race and ethnicity in access to and uptake of reproductive health services among adolescents.

We thank Dr Stephen Cardamone for initial support of this project and Dr Anup Patel, Dr Veronica Mruk, and Jahnavi Valleru for their careful article editing.

EMR

electronic medical record

KDD

key driver diagram

PFK

Partners For Kids

QI

quality improvement

QIS

quality improvement specialist

RHA

reproductive health assessment

RHI

Reproductive Health Initiative

STI

sexually transmitted infection

WCV

well-care visit