From the Division of Neonatology, Department of Pediatrics, Baylor University Medical Center, Dallas, Texas.

Corresponding author: Jonathan Whitfield, MBChB, Division of Neonatology, Department of Pediatrics, Baylor University Medical Center, 3500 Gaston Avenue, Dallas, Texas 75246.

VERMONT OXFORD NETWORK

BUMC participates in the Vermont Oxford Network, a voluntary network of neonatal nurseries formed in 1989 (3). This network grew to 325 nurseries worldwide by 1999. The purpose of the network is to track the outcome of very low birthweight infants (401 to 1500 g). Quarterly reports enumerating key outcome variables are developed and then sent to member nurseries. Although the results for each center are confidential, comparisons at each center can be made quarter to quarter and year to year. Further, submitted data from all nurseries are pooled. This permits each nursery to compare itself with the network as a whole. For example, nosocomial infection with coagulase-negative Staphylococcus in very low birthweight infants is very common. In 1998, the percentage of infants with coagulase-negative Staphylococcus at BUMC was 12%. The network as a whole reported a 14% incidence for the same year (4).

Comparing a member nursery with the pooled network data represents a form of “benchmarking.” Yet, hidden in pooled network outcome data are comparable nurseries with very diverse outcomes--e.g., mortality rates vary enormously from nursery to nursery (4) (Figure 1). Some NICUs in the Vermont Oxford Network consistently perform in the best quartile; that is, only 25% or less of all reporting Vermont Oxford Network members have better outcomes. Clearly, learning the processes of care in these better-performing nurseries might be helpful in improving the effectiveness of care in one's own NICU. A good example of collaborative learning was seen in the Northern New England Cardiovascular Project, where complications of coronary bypass surgery were substantially reduced as a result of the multidisciplinary and multi-institutional collaboration (5).

NEONATAL INTENSIVE CARE QUALITY IMPROVEMENT COLLABORATIVE

To focus on superior processes of care, the Vermont Oxford Network formed a pilot Neonatal Intensive Care Quality Improvement Collaborative (NICQ) in 1994 consisting of 10 volunteer NICUs. Multidisciplinary teams from the nurseries collaborated in site visits and study groups to develop potentially better practices. As an example of the success of the early pilot collaborative, one participating NICU reduced its nosocomial infection rate in infants <1500 g from 39% in 1994 to 13% in 1997 (personal communication, Dr. William Edwards, 1999).

Spurred on by the early success of NICQ, a new active collaboration began in 1998 and is currently under way with 34 NICUs in the USA. The goal of NICQ 2000 is to improve the effectiveness and efficiency of neonatal intensive care in 3 areas: clinical, operational, and organizational. Four key habits are being taught (Figure 2):

• The habit for change. The key to more effective care is the willingness of staff to accept new ideas. New ideas may come from many sources--e.g., the literature, an analysis of internal and external processes, benchmarking with superior performers, site visits to other NICUs, and researchers' own experience and thinking. Resistance to new ideas and changes in process of care is common. Acceptance of change depends to some extent on the culture of the NICU: a unit may be strong in group culture, stressing affiliation between staff members, teamwork, and participation; developmental culture, stressing risk taking and willingness to change; hierarchical culture, stressing establishment and maintenance of bureaucratic work patterns; or rational culture, stressing efficiency and achievement. Although no NICU will reflect only 1 type of culture, the unit culture in the NICU at Baylor in 1998 was predominantly group and developmental. Baylor had mean scores of 33% group culture, 27% developmental culture, 22% rational culture, and 18% hierarchical culture. Units with strong group and developmental cultures, such as BUMC, are considered receptive to quality improvement efforts.
• The habit for understanding the processes of care. The multiple steps involved in providing care are often not appreciated. The formulation of a flow process chart can be very helpful in identifying problems. For example, to understand why hypocapneic ventilation was occurring in very low birthweight infants admitted to the NICU at BUMC, we mapped the process of care in a flow chart (Figure 3). After educating the health care team about the importance of avoiding hypocapnia and the optimal way of selecting ventilator settings, we observed a significant decrease in variation of the first arterial Pco₂ obtained after delivery (Figure 4).
• The habit for collaborative learning. By working with other NICUs with similar interests and carefully evaluating the care practices in better-performing NICUs with multidisciplinary site visits, potentially better practices can be formulated. Two focus groups formed at BUMC collaborated with other similar-minded groups among NICQ 2000 members. One focus group has concentrated on reducing nosocomial infection and has developed a potentially better practices work list. The other focus group has collaborated in developing strategies to reduce neonatal intraventricular hemorrhage and periventricular leukomalacia and also has produced a potentially better practices work list. Many of these practices will be implemented in each of the member NICUs after discussion with the local medical and nursing staff. Not all practices from a potentially better practices work list are appropriately applied in every nursery. As an example, a reduction in skin punctures and vascular line entries (known risk factors for infection) has been implemented at BUMC as a potentially better practice. The results of our early experience are shown in Figure 5.
• The habit for using evidence-based practice. Many of the routine practices in NICUs have not been well studied. Those interventions that have been subjected to randomized controlled trials need to be reviewed and, if appropriate, utilized in the NICU. Careful and critical evaluation and judicious application in one's own care practices may be very helpful in improving effectiveness of care. The National Perinatal Epidemiology Unit at Oxford has pioneered this approach to practice. The Cochrane Collaboration is an outgrowth of this effort (6). Neonatology has a separate section within the Cochrane database that can be accessed free of charge on the Internet at http://www.nichd.nih.gov/cochraneneonatal/. This site and other Internet resources are available to the staff on NICU-based computer terminals at BUMC.

NICQ 2000 has promoted a very helpful technique for reviewing study results in a more rigorous fashion; the network calls this technique a critically appraised topic, or CAT (7). Part of the process of critical appraisal is the calculation of the number of patients that have to be treated with the experimental or new treatment to avoid the event in question. This number needed to treat (NNT) is calculated as follows: Determine the percentages of subjects in the experimental and control groups that experience the event in question (experimental event rate, EER; control event rate, CER). Then calculate the relative risk reduction (RRR) and absolute risk reduction (ARR) by using the following formulas: RRR = (CER - EER)/CER and ARR = CER - EER. By following this stepwise process of evaluation, the NNT can be calculated by applying the formula NNT = 1/ARR. The NNT value is especially useful in drawing conclusions about the clinical effectiveness of the proposed intervention. This topic has been well reviewed by Soll and Andruscavage (8).

Improving the quality of care in BUMC's NICU is a journey. We have begun that journey. Our destination is effective, efficient neonatal care.

Acknowledgments

We thank Alice Morrow, RN, MSN, Pam McKinley, RN, Deana Black, RN, Rachel Cody, RN, and Jobeth Pilcher, RN, for their support and the entire staff of the NICU, who are dedicated to improving the care of neonates at BUMC.

1. Lorenz JM. Survival of the extremely preterm infant in North America in the 1990s. Clin Perinatol 2000;27:255-262.
2. Baylor University Medical Center 1999 Annual Neonatology Report. Dallas, Tex: Baylor University Medical Center, 1999.
3. Horbar JD. The Vermont Oxford Network: evidence-based quality improvement for neonatology. Pediatrics 1999;103(1 Suppl E):350-359.
4. Vermont Oxford Neonatal Network 1998 Annual Report. Burlington, Vt: Vermont Oxford Network, 1998.
5. O'Connor GT, Plume SK, Olmstead EM, Morton JR, Maloney CT, Nugent WC, Hernandez F Jr, Clough R, Leavitt BJ, Coffin LH, Marrin CA, Wennberg D, Birkmeyer JD, Charlesworth DC, Malenka DJ, Quinton HB, Kasper JF. A regional intervention to improve the hospital mortality associated with coronary artery bypass graft surgery. The Northern New England Cardiovascular Disease Study Group. JAMA 1996;275:841-846.
6. The Cochrane Library. Update Software Ltd. Oxford, England, 1998.
7. Sauve S, Lee HN, Meade MO, et al. The critically appraised topic: a practice approach to learning critical appraisal. Annals of the Royal College of Physicians and Surgeons of Canada 1995;28:396-398.
8. Soll RF, Andruscavage L. The principles and practice of evidence-based neonatology. Pediatrics 1999;103(1 Suppl E):215-224.