Implementing Parent Coaching in Hospital-Based Pediatric Occupational Therapy: A Multisite Quality Improvement Project

Early intervention is critical for children with cerebral palsy (CP) to maximize outcomes. Optimal neuroplasticity occurs in the first 2 to 3 yr of life (de Graaf-Peters & Hadders-Algra, 2006). Because children as young as age 5 months can now be reliably diagnosed with CP, early referral and rehabilitation interventions, including occupational therapy services, are possible (Byrne et al., 2017; Maitre et al., 2020; Novak et al., 2017). These early interventions can take place either in state-funded early intervention programs in the child’s home, in hospital-based outpatient services, and in many other settings. In the hospital-based outpatient setting, early intervention visits tend to be scheduled for approximately 30 to 60 min, once a week (Gmmash & Effgen, 2019). Therapy dosage includes the frequency, intensity, and timing with which the child spends practicing new skills as well as the type of therapy that the child receives (Gannotti et al., 2014). Unfortunately, 1 hr/wk of therapy is insufficient to effect change (Tinderholt Myrhaug et al., 2014). Although participation in more intensive therapy services may not be feasible for many families, it is possible to increase the dosage of therapy by increasing the time the child spends practicing new skills outside of therapy sessions. To increase the child’s therapy dosage outside the clinic, parents must be able to continue practice and environmental enrichment at home (Damiano & Longo, 2021). Collaborating with parents on these skills by actively involving them in therapy sessions through reflection, feedback, and joint planning is called parent coaching (Rush & Shelden, 2011).

Evidence shows that parent coaching improves motor skills in young children with motor delays (Tanner et al., 2020). Parent coaching is recommended by the most recent American Occupational Therapy Association practice guidelines for children with CP who are younger than age 5 yr (Clark & Kingsley, 2020), and it is considered best practice for children who are younger than age 2 yr (Morgan et al., 2021). Although the methodologies for individual studies informing these guidelines vary, many are grounded in the seminal handbook on early childhood coaching by Rush and Shelden (2011).

Although parent coaching is well integrated into state-funded Individuals with Disabilities Education Act (IDEA) Part C early intervention services, this model is not consistently used in outpatient practice for several reasons. First, the child is seen in a clinic rather than in their home, so the practitioner is not able to observe the child’s daily routines. Second, practitioners treating within the medical model are viewed as the experts during the appointment and thus may spend more time with their hands on the child rather than coaching parents through therapeutic activities. Finally, there may be situations in which a treatment modality with less parental involvement is indicated, particularly early in a therapy episode. Despite these potential challenges for outpatient practice, parent coaching provides a solid framework for engaging caregivers in therapy sessions to ensure maximal skill transfer. We developed a parent coaching toolkit to address the unique challenges of implementing this therapy mode in an outpatient hospital–based setting and initiated a quality improvement project with the specific goal of increasing the documented use of parent coaching in outpatient occupational therapy visits for children with CP who were younger than age 2 yr, from 0% to 80% by December 31, 2021, and sustaining the use of parent coaching for 6 mo.

This quality improvement project took place at three pediatric tertiary-care hospitals in the United States. At the time of project startup, all sites were part of or received referrals from the Cerebral Palsy Early Detection and Intervention Network, which is supported by the Cerebral Palsy Foundation. Individual analyses of strengths, weaknesses, opportunities, and threats (SWOT) conducted during the project startup revealed that all sites had some level of support for continuing education and evidence-based practice. However, none of the occupational therapy practitioners across sites had previously taken part in a formal quality improvement project (see Table 1 for site descriptions).

All developmental occupational therapy visits (including telehealth and in person) for patients who were younger than age 2 and who had a diagnosis of CP were included in the quality improvement project. Evaluation visits and splint fabrication sessions were excluded. For Sites B and C, the inclusion criteria were expanded for the second plan–do–study–act (PDSA) cycle to include the progress notes of patients with early clinical signs of CP (Boychuck et al., 2020). This decision was made because those sites experienced lower referral numbers of patients with the formal diagnosis of CP. Data were stratified for PDSA Cycle 2 for those two sites to investigate the effect of a diagnosis of CP compared with early clinical signs of CP.

We developed a key driver diagram (see Figure 1) using pilot survey data from occupational therapy practitioners across all three sites. Before the first PDSA cycle, each site completed a SWOT analysis and a current state process flow diagram, both of which were used throughout the project to tailor interventions to each site, including educational materials and electronic medical record (EMR) phrases. Each site team included a program manager or lead, as well as peer coaches (occupational therapy practitioners with specific clinical expertise and/or interest in young children with CP) (see Table 2 for a summary of key interventions in each PDSA cycle by site).

For all practitioners employed at each site, the project team members (Kelly Tanner and project leads or peer coaches from each site) conducted the initial education sessions, which consisted of two sessions lasting 45 to 60 min. Session 1 described in detail the five components of parent coaching and rationale for this approach (Rush & Shelden, 2011). Session 2 focused on the application of these five components through a video-based case study, as well as practical implications such as documentation and billing. Team members also developed additional educational supports (referred to as quick references) to serve as ongoing reminders of how to use each of the components of parent coaching. Quick references included items such as a graphic organizer to enhance parent communication and a podcast-style interview with an experienced clinician using this approach.

Two EMR phrases were developed and adapted for use across sites: a therapist-facing (i.e., part of the clinical documentation for each session) parent coaching checklist and a parent-facing (i.e., able to be pasted into the parent education section of the EMR or printed out and handed to a caregiver) joint plan document. EMR chart reviews facilitated quality improvement interventions by notifying practitioners who were (or were not) documenting their use of parent coaching.

After reviewing data from PDSA Cycle 1, each site implemented changes as needed and initiated Cycle 2. Specific changes were made on the basis of data from the chart reviews. For example, Site A observed a high level of compliance with documentation, but this was with a high-burden quality improvement intervention. Therefore, Site A focused on decreasing support by no longer sending individual reminders to practitioners to document their use of parent coaching. Site B observed consistent documentation of the components of observation, action, feedback, and joint planning but not for reflection. In addition, a small population of patients received a formal diagnosis of CP, but many more showed clinical signs of CP. Therefore, Site B focused on additional education sessions with regard to one specific component of parent coaching— reflection—as well as broadening their criteria to include children with early clinical signs of CP. Site C observed that practitioners were inconsistently documenting parent coaching, but there was no particular component that seemed more challenging than another. Site C also had a relatively small population of patients who received a formal diagnosis of CP but many more with clinical signs of CP. Therefore, Site C focused on additional education sessions for all five components of parent coaching, as well as broadening their criteria to include children with early clinical signs of CP.

To evaluate the impact of this quality improvement project on the use of parent coaching in outpatient occupational therapy sessions, we developed a time- series design using weekly data points for the primary outcome measure.

The primary outcome measure was the documentation of parent coaching in outpatient sessions, (i.e., practitioner adoption of the practice), as evidenced by the percentage of documented components of parent coaching present in the EMR. This was selected as the outcome measure for the study as an implementation-level measure (Proctor et al., 2011). Our objective for this study was to change practitioner behavior (adoption), and we captured this through the audit of practitioner documentation of treatment sessions. Capturing clinical outcomes of individual patients was not within the scope of this project; our focus was on practitioner behavior change. Site leads collected this information through chart audits throughout the project’s duration and entered data weekly. Each individual who entered data into the RedCAP database had the opportunity to practice, ask questions, and receive feedback before collecting real data. Data were reviewed on a monthly basis, and any potential discrepancies were resolved through discussion with the site leads and repeat evaluation of audited charts.

Using an investigator-developed test that was delivered using REDCap (Harris et al., 2009), we assessed the practitioners’ knowledge of parent coaching, before and after the initial education sessions, as a key change in the system that may have affected the outcome measure. Use of the Parent Coaching checklist was also included as a process measure for Site A.

We summarized the data on parent satisfaction and experience pre- and postproject for each site using their institution’s standard patient satisfaction or experience surveys (e.g., Press Ganey). Because all patients were younger than age 2, all surveys were completed by parents or legal guardians; therefore, these survey results are collectively referred to as “parent satisfaction/experience.” Although parent satisfaction and experience can be considered different measurement constructs, both are intended to help managers and other staff members in institutions understand how care was received. In addition, administrators were surveyed after the project to determine whether there was a perceived loss of productivity among occupational therapy practitioners.

We analyzed the primary outcome using a statistical process control chart that was displayed on a p-chart. Although adherence to parent coaching was initially scored dichotomously as present (with all five components) or absent (with four or fewer components), this definition was amended during PDSA Cycle 2 to capture, instead, the overall percentage of documented components. We calculated the percentage of documented components of parent coaching for each site per week (numerator) and divided it by the total number of potential components (denominator). This change was made because of the low numbers of patients younger than age 2 with CP who had been seen weekly at Sites B and C. We used Nelson’s rules for statistical process change (Langley et al., 2009; Provost & Murray, 2003) to determine the presence of special cause variation.

Change in therapist knowledge was analyzed in Excel using a one-tailed t test. Change in parent satisfaction was reported using percentages, and change in practitioner productivity was reported by each institution’s outpatient occupational therapy program manager.

This study was determined to not be human subjects research and, therefore, exempt from review by Site A’s institutional review board. Lead team members from Sites B and C submitted the letter of exemption through their appropriate institutional channels to gain permission to engage in this project.

The primary intervention for PDSA Cycle 1, practitioner education, was modified for each site to reflect its institutional EMR and to include the peer coaches as presenters. Any questions from practitioners that arose within Session 1 were addressed formally in Session 2. Parent coaching knowledge scores from pre- to postsession for practitioners who completed surveys at both time points indicated an increase in parent coaching knowledge from 83.1% to 87.9%, t(56) = 3.289, p = .001. One unexpected benefit that each site experienced was the use of parent coaching in disciplines and departments other than those targeted in the project. For example, at Site A, outpatient physical therapists and home-care occupational, physical, and speech therapists attended the education sessions per staff request.

Time-series analysis of the outcome measure for this cycle revealed a modest improvement in the use of parent coaching across sites after the education sessions (see Figure 2). Site A attained the goal of 80% use of parent coaching when an opt-out checklist was entered into all progress notes; however, this required an additional step for practitioners to start a new note for their recurring patients. Sites B and C were not able to implement a similar opt-out approach in their respective EMRs. Sites A and B also provided direct feedback to practitioners throughout the chart review process, but site C did not.

For Sites B and C, there were several weeks when patients who met the criteria were not seen at any of the clinical sites because of low patient volumes.

During PDSA Cycle 2, Site A aimed to sustain their rates of parent coaching use while withdrawing individual reminders. Sites B and C aimed to increase their patient volumes by increasing referrals and broadening their criteria to include patients with early clinical signs of CP who had not yet received a formal diagnosis (Boychuck et al., 2020). Another benefit of this project resulted from increased professional networking within sites as they worked to increase their referrals. For example, Site B reported working with additional clinical units within the hospital to further improve coordination of care for young patients with CP. Site C reported improved communication with its main referring clinic, resulting in a more streamlined referral process for both parties.

Parent satisfaction/experience and productivity data are presented in Table 3. Managers across all sites qualitatively reported no changes in perceived parent satisfaction/experience or therapist productivity that may be attributed to implementation of parent coaching.

This multicenter quality improvement initiative demonstrates that occupational therapy practitioners can effectively implement and sustain the use of parent coaching for their youngest patients with CP. Parent satisfaction/experience and practitioner productivity were not affected by this initiative.

Occupational therapy practitioners were able to implement parent coaching successfully through a combination of education and additional supports. There was a statistically significant improvement in practitioner knowledge of parent coaching after education coupled with a center-line shift in the control chart for the primary study outcome measure. However, the goal of 80% documentation of parent coaching was not attained until after the initiation of individual reminders, which confirms the importance of targeted knowledge translation strategies that extend beyond education. This is consistent with the findings of previous studies, which demonstrated that targeted knowledge translation efforts result in change for specific areas of occupational therapy practice such as hand therapy (MacDermid & Graham, 2009), mental health (Bazyk et al., 2015), and stroke (Doyle & Bennett, 2014; Hansen et al., 2016). Because education alone does not produce behavior change for occupational therapy practitioners, additional supports (e.g., administrative, EMR) and data collection are required to understand the effects of our knowledge translation efforts.

Although the goal of 80% documentation was exceeded and a new baseline of 87.82 was maintained for 4 mo, the centerline shifted down to 79.1% and was maintained for the final 7 mo of the project. This shift coincided with the start of PDSA Cycle 2, during which Site A discontinued the use of individual therapist reminders to document use of parent coaching. Because Site A accounted for the largest percentage of visits included in the data set, it is possible that this lack of continued direct support may have contributed to the downward shift. In addition, hiring patterns (i.e., new staff replacing staff who were initially trained, use of contingent therapists) and seasonality (i.e., the start of the winter holiday season, staffing trends, COVID trends that affect staffing) may also have contributed to this shift. Focusing on system-based interventions rather than on individual interventions could be explored as a strategy to further sustain change (Pittet et al., 2000).

We further hypothesize that some components of parent coaching might have been more challenging to implement than others, yielding a potential “ceiling” of 80% (four of five components). For example, during PDSA Cycle 2, Site B was observed to have difficulty specifically with documenting the “reflection” component of parent coaching. We hypothesize that this component might have been more challenging to influence than others, such as observation that requires the parent to be present and looking at the therapist but not necessarily engaging actively. Additionally, the reflection portion typically takes place toward the end of a session and might, therefore, be subject to time constraints. Finally, practitioners are trained to document observable behaviors. Although the concept of reflection was thus modified to “repeat back in their own words,” practitioners might still have found this to be too subjective to state in the EMR.

This study has resulted in long-term change for the people and systems involved. For example, Site A has since developed several more quality improvement initiatives that affect similar patient populations, expanding capacity for further practice improvement. Site B has been able to refine its clinical offerings for young children with CP and further expand their early diagnosis and referral program. Site C has opened up direct channels of communication with its primary referral source for young children with CP, allowing for more children to receive outpatient services. Additional multisite efforts around these initiatives will help to ensure that this high quality of care is sustained.

There are several limitations of this study. First, although the three institutions that participated were geographically diverse, all were large, tertiary-care facilities with strong cultures for evidence-based practice and professional development. This may limit generalization to other settings with varying levels of resources and supports. The customization of our implementation package for each site may also limit generalization to other settings. In addition, although the same EMR was not used across all institutions, the ability to change and augment the EMR was a key intervention that may not be possible everywhere. The primary outcome was measured by means of an EMR chart audit, which may also be considered a limitation. Because we were not able to videotape and code sessions for fidelity within the scope of this project, the EMR audit was selected as the most objective measure available.

Parent coaching has been identified as an evidence-based practice to improve motor skills for young children with, or at high risk for, CP. Occupational therapy practitioners can use parent coaching across practice settings. A parent coaching implementation toolkit may help practitioners to adopt this style of intervention in a hospital-based outpatient setting. This multicenter quality improvement undertaking demonstrates how quality improvement methodology can support implementation of occupational therapy best practices.

• ▪ The parent coaching implementation toolkit consists of educational and EMR supports. EMR supports facilitate improved adherence after education sessions.

• ▪ Institutions or practitioners who are considering implementing parent coaching in their practice can use quality improvement tools to plan and measure their progress.

This multi-site quality improvement project demonstrates that parent coaching was implemented in three different hospital-based pediatric occupational therapy settings to address motor development for young children with or at high risk for CP. Practitioners learned about parent coaching concepts through educational sessions, and then documented use of parent coaching principles through their clinical documentation. The objective of 80% documentation was exceeded when practitioners were given high levels of support, but then fell slightly lower when direct supports were removed from one site. Occupational therapy practitioners can use quality improvement strategies to implement best practices for their clients.

This project was funded by the American Occupational Therapy Foundation. The authors acknowledge the support of the Cerebral Palsy Foundation (CPF) and the CPF Early Detection Network. The authors also thank the practitioners at each site, especially those who served as peer coaches and leaders for their teams.