12/20/2013

I.  ABSTRACT

The Ohio Perinatal Quality Collaborative (OPQC) is a learning network that implements ongoing quality improvement collaborative projects with perinatal health care practitioners and providers in Ohio including physicians, nurses, other clinicians, parents, payers and policy makers. The OPQC mission is “through collaborative use of improvement science methods, reduce preterm births and improve perinatal and newborn outcomes in Ohio as quickly as possible.” OPQC membership is at the institution level; each participating institution identifies a lead physician and nurse manager as well as additional clinical faculty and staff who work as members of local site-specific and topic-specific improvement team(s).

The James M. Anderson Center for Health Systems Excellence at Cincinnati Children’s Hospital Medical Center (CCHMC) serves as the administrative hub and data coordinating center for the OPQC, meaning that the contract funding is issued to the Anderson Center at CCHMC, central project management staff are employed by the institution, and data sets are submitted to the Anderson Center for analysis. The Division of Biomedical Informatics at has received separate funding to develop and administer the perinatal quality improvement database.

Participant sites share limited data sets containing patient-level data with the Anderson Center. The purpose of this data sharing is to facilitate quality improvement activities. In addition, the network has a research interest in understanding whether and how the use of improvement science and collaborative projects is able to improve outcomes at a population level. Encrypted data is transferred electronically via a secure, password-protected, web-based extranet from participating institutions to the Anderson Center. The HIPAA-specified, “limited data set” procedures are used as part of all OPQC improvement projects. The Anderson Center’s data team uses statistical process control (SPC) methods and develops run charts and control charts that contain only aggregate and de-identified data. Site-specific graphs are provided to each of the participating institutions. Each improvement team has visibility only to their site-specific graphs and to graphs that include aggregate data for the all sites participating in a given project. While the Anderson Center will receive limited data sets, the graphs produced and returned to teams contain only de-identified aggregate data.

OPQC will also continue to work with the Ohio Department of Health Office of Vital Statistics to secure access to de-identified birth registry data for specific projects as needed.

OPQC will continue to engage in multiple simultaneous improvement projects with new projects starting and old projects ending depending on success achieved and resources available for taking on new projects. All projects will use only limited data sets. Improvement efforts will occur concurrently with care. For example, if projects are designed to assure that physician orders are reliably followed, improvement efforts will occur simultaneously during the care of the patients for whom these orders are made. Lessons learned as part of OPQC improvement efforts will be published in peer-reviewed journals.

II: BACKGROUND

Evidence-based guidelines exist to address factors that can improve pregnancy outcomes. These guidelines are not implemented consistently.

Ohio ranks 40th in Black prematurity rates (2010) and 47nd in infant mortality (2011) and 49th in Black infant mortality (2011). In Ohio, 60-70% of deaths in the first year of life are associated with extreme prematurity (<29 weeks gestational age). In addition to the markedly elevated mortality risk, it is estimated that 50% of lifelong handicapping conditions including blindness, deafness, developmental delay and cerebral palsy, with onset in infancy, is associated with extreme prematurity. Only 1% of births are extremely premature (<28 weeks), but these births account for approximately 25% of spending for perinatal care in Ohio.

Unexplained variation in risk-adjusted prematurity rates in Ohio strongly suggests that opportunities for prevention exist. There is much to learn about the causes of preterm birth, yet there is substantial evidence to support a variety of interventions to reduce prematurity, including use of progesterone in women with a prior preterm birth, appropriate management of uterine malformations, management of inter-pregnancy interval for women with prior preterm birth, use of effective smoking cessation programs, and appropriate use of ovulation induction and assisted reproductive technologies.

Quality improvement activities have been shown to be effective in increasing the use of evidence-based clinical practices (Horbar, 2000; Horbar, Carpenter, & Buzas, 2004; Horbar, Rogowski, & Plsek, 2001; Iams & Donovan, 2011; Kaplan, Lannon, Walsh, & Donovan, 2011) and improving clinical outcomes. These methods include the use of benchmarking, collaborative learning, using common metrics, and application of rigorous improvement methods such as the Associates for Process Improvement “Model for Improvement” (Langley, Nolan, Nolan, Norman, & Provost, 1996).

III: PURPOSE OF STUDY

The purpose of OPQC is to improve perinatal newborn and infant outcomes in Ohio. OPQC will enhance existing quality improvement activities at participating practitioner sites by helping develop capacity to promote quality improvement, by helping develop and apply comparable quality measures, and by helping establish collaborative relationships and a broad-based coalition that will support these goals.

IV. SIGNIFICANCE OF STUDY IN RELATION TO HUMAN HEALTH

Quality Improvement projects led by OPQC can lead to a simultaneous decrease in rates of prematurity and infant mortality and neurodevelopmental, sensory and physical morbidity in Ohio over time, a decrease in the length of stay in the neonatal intensive care unit (NICU), and enhance the parents’ experience of their infant’s care and therefore decrease subsequent acute care health care use for surviving infants. In addition, lessons learned by OPQC and research focused on understanding the mechanisms of improvement and organizational behavior change in QI projects can help improve collaborative projects throughout the healthcare system

V. PREVIOUS WORK DONE IN THIS AREA

Over the last five years, OPQC has successfully implemented the following quality improvement projects, each of which followed a modified BTS methodology (as described in the Methods section):

1. Decreasing Late-Onset Infections in Infants born 22 – 29 weeks gestation

This project aimed to decrease infections in neonates born at 22 to 29 weeks gestation, through three specific changes to practice – implementing an evidence-based catheter insertion bundle, implementing an evidence-based catheter maintenance bundle, and increasing the use of human milk as medicine in this population. A total of 24 teams/sites participated in this initiative.

Our work and that of others suggests that reductions in infection are not seen consistently until reliability of processes is achieved and maintained. As maintenance bundle reliability increased from <30% to 80% in the first phase of OPQC (2008-2010), we documented a 20% decrease in infections in this population from 18% to 14% (Kaplan, et al., 2011).

We estimate that since OPQC work began in in 2008, 390 infections have been averted in these tiny babies, resulting in an approximate cost savings exceeding $6,500,000. Having attained consistent level 1 reliability of the catheter maintenance bundle, OPQC focused on attaining level 2 reliability (>95%) with monthly action calls, and active support to teams through December 2013. OPQC will then transition this project to a “sustain” mode. Research has shown that without ongoing monitoring and periodic refreshers teams lose their impact within 90- 120 days of the work. Thus it is important to continue monitoring and periodic check-ins. With a transition to sustain mode, infection data will continue to be monitored on a quarterly basis by OPQC leadership to ensure that improvements are sustained.

2. Decreasing Non-Medically Indicated Scheduled Births Prior to 39 Weeks Gestation

Since 2008, OPQC has worked to reduce unnecessary scheduled births at 360/7 to 386/7 weeks gestational age in the obstetrical units of 20 of the largest hospitals in Ohio with the 20 charter-member OB teams. Following execution of an initial successful project in the 20 largest maternity hospitals in Ohio with sustained decreases observed, OPQC received additional funding from the Ohio Department of Health (ODH), Ohio Medicaid and the Center for Disease Control and Prevention (CDC) to 1) design a systematic implementation plan to engage all Ohio maternity hospitals in application of the successful scheduled delivery changes; 2) pilot this plan with a small number of maternity hospitals and 3) disseminate this work to remaining Ohio maternity hospitals. This project also involved the development of a plan to utilize the state regional BEACON (Best Evidence for Advancing Childhealth in Ohio Now) Quality Improvement Coordinators (QICs) in recruitment, outreach, and performance feedback efforts.

In coordination with the Ohio Department of Health, and with the help of two BEACON QICs, OPQC recruited 15 hospitals in Ohio for a pilot dissemination project (January 2012 – December 2012). We developed and implemented a collaborative learning effort for these maternity hospitals, which included an initial Learning Session, data feedback using birth certificate data, and regular webinars. Selection criteria for the pilot sites ensured that this cohort included many of the hospitals with the highest number of Medicaid births, as indicated from birth certificate data, as well as geographic diversity. The pilot project concluded in December 2012.

Lessons learned from the pilot project were used to shape the recruitment strategy, dissemination plan, and project curriculum for a dissemination project. Recruitment strategy involved targeting the remaining non-OPQC maternity hospitals in the state of Ohio, which were been divided into three approximately equivalent cohorts; a letter was sent on behalf of OPQC leadership, ODH-Vital Statistics leadership, and the Ohio Hospital Association leadership to CEOs included in the first and second wave of dissemination hospitals, followed up with recruitment phone calls. The final wave of sites is on target to complete the improvement project by April 2014.

Since the original “39-Weeks” project kick-off in 2008, over 36,200 births that would have been expected to occur prior to 39 weeks gestation have shifted to 39 weeks gestation. A total of 105 sites have participated in the quality improvement initiative.

3. Increasing the use of Antenatal Corticosteroids (ANCS) in women at risk for preterm birth.

Maternal administration of antenatal corticosteroids (ANCS) to reduce the frequency and severity of complications of prematurity for infants born between 24 and 34 weeks of gestation has been recommended for almost 40 years yet gaps exist. This project focused reducing perinatal morbidity and mortality by ensuring >90% of infants born between 240/7th and 340/7th weeks of age receive ANCS. BTS methods were used to engage 19 OPQC maternity hospitals.

VI: SELECTION OF PARTICIPANTS

Patient Subject Selection

Individual patients are not targeted or recruited to participate in OPQC efforts. Rather, populations of patients for whom evidence-informed care processes will improve care will benefit from OPQC improvement efforts.

Participating sites will decide on aspects of care to improve. Populations of patients (pregnant women and infants) whose outcomes might benefit from that improved care will be targeted for care improvements. Practice sites will target groups of patients that meet improvement project eligibility criteria. As new improvement objectives are developed, different improvement projects will have different eligibility requirements. Because practice sites are working collaboratively, the same eligibility criteria will be used across sites for each improvement project.

Practice Site Selection

All perinatal practice sites in Ohio may voluntarily choose to participate in OPQC improvement projects. Each site will be responsible for obtaining local IRB approval for participation or will agree to allow CCHMC to serve as the IRB of record; each site will sign data sharing agreements as appropriate.

VII. RECRUITMENT

Patient Subject Recruitment

Individual patients are not targeted or recruited to participate in OPQC improvement efforts. All patients who meet pre-specified, project-specific criteria will be included in improvement projects. Because all OPQC improvement projects are evidence-informed and apply only interventions that have been endorsed by professional and governmental organization as the standard of care, none of the interventions are experimental. Individual patient-level consent will not be obtained. OPQC adds proven-effective quality improvement methods, rapid data analysis and collective experiential learning to accelerate quality improvement at participating sites.

Practice Site Recruitment

All perinatal practice sites in Ohio may voluntarily choose to participate in OPQC improvement projects.

VIII. RANDOMIZATION

Patients will not be randomized.

IX. STUDY PROCEDURES

OPQC will be directly engaged in facilitating the implementation of perinatal quality improvement activities at participating provider sites in Ohio. OPQC will support ongoing development of a multidisciplinary stakeholder group consisting of leaders from hospitals with neonatal care units in the state, key obstetrics leaders, and representatives from the Ohio Department of Health, the Ohio Chapters of the American Academy of Pediatrics (AAP) and American College of Obstetricians and Gynecologists (ACOG) and existing family support groups in each of the participating NICUs. The OPQC Steering Committee with representatives from perinatal caregivers from across the state will review and endorse recommendations and measures for improving care at the participating sites. Participating providers will report on their performance monthly throughout the project period.

Improvement teams from participating provider groups, together with clinical experts, quality improvement experts and peers at other provider sites will exchange data about results. An organizational team at each site will plan specific changes they will make at the local level. Lessons learned among the various participating provider sites will be shared using the OPQC communication infrastructure (e.g. face-to-face meetings twice/year, monthly webinars, listserv). OPQC technical staff will provide regular coaching on data collection and reporting, improvement methods and spread strategies by teleconference on a monthly basis.

A full time OPQC Project Manager will be responsible for managing timelines for improvement at each site. With the leadership of the OPQC Steering Committee, the data management team funded by the state and based at CCHMC will collate monthly reports from each provider site. OPQC sites will participate in statewide improvement collaborative meetings each year. OPQC will produce and distribute monthly reports from participating provider sites that include aim statements, key driver analyses and annotated run charts.

Selected improvement projects will be implemented by provider sites that agree to participate. For example, sites may decide to improve the delivery of pulmonary surfactant to eligible infants. Participating sites will submit both individual patient process and outcome data that are relevant to the project to OPQC via a secure, password-protected web site. Consistent with the HIPAA-specified Limited Data Set, the only protected health information (PHI) that will be collected and retained at the Anderson Center will be dates of service. Patient-level data will be aggregated to provide site-specific progress reports back to the site where the improvement project is being conducted. Unless there is specific written agreement from the site, reports that are shared with other participating sites will not identify the site from which the data derive.