Improving Growth of Very Low Birth Weight Infants in the First 28 Days
* Wesley Medical Center, Wichita, Kansas
Tacoma Regional, Tacoma, Washington
Harris Medical Center and Cook Children’s Center, Columbia, South Carolina
¶ Palmetto Baptist Medical Center, Columbia, South Carolina
|| Mercer Medical Center, Trenton, New Jersey
** Hospital Auxilio Mutuo, Hato Rey, Puerto Rico
Presbyterian Hospital, Oklahoma City, Oklahoma
Pediatrix Medical Group, Sunrise, Florida
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective. To increase weight gain in the first 28 days after birth for very low birth weight (VLBW) infants by isolating and sharing meaningful process differences between high- and low-weight-gain centers within a neonatal network.
Design/Methods. We identified weight gain as an important target for improvement in 1999 for our national group practice of neonatologists. Site-specific average weight gain during the first 28 days was the primary outcome measure. Our target population was defined as inborn infants who survived and remained in the hospital of birth, whose birth weights were between 401 and 1500 g (VLBW), and who were >22 weeks’ estimated gestational age. A team of 6 neonatologists and 1 nurse met, reviewed processes that might influence growth, and developed a structured observation guide for site visits. Weight gain data were obtained from an existing administrative database for the period January 1, 1997, through June 30, 1999. Centers were ranked and divided into upper, middle, and lower thirds. Seven team members visited 1 high- and 1 low-weight-gain center without being informed of the center’s performance. Following the site visits, the team isolated 16 meaningful differences between high- and low-weight-gain sites. Meaningful differences were defined as processes observed in all or virtually all (for this project, 6 or 7 of 7 centers) of the high and none or virtually none (for this project, 0 or 1 of 7) of the low centers. The meaningful differences were distributed to our medical directors in August 2000 along with their site-specific weight-gain performance. To document the impact of sharing this material, we compared weight gain in a baseline period of January 1 through December 31, 1999 and a posteducational intervention period of January 1 through September 30, 2001.
Results. Compared with neonates admitted to our national neonatal practice in 1999, neonates admitted in 2001 were similar in birth weight, gestational age at birth, exposure to antenatal steroids, and male gender. Average daily weight gain during the first 28 days increased from 10.4 ± 6 g for neonates cared for in 1999 to 12.5 ± 6 g for neonates cared for in 2001. Thirty-nine of 51 (76%) units noted improvements, 4 were unchanged and 8 noted a decrease in average weight gain. Despite similar average lengths of stay, the average discharge weight for neonates sent home increased from 2.15 ± 0.5 kg for 1999 to 2.29 ± 0.5 kg for 2001. There were no differences in frequencies of mortality or major morbidities, including severe intraventricular hemorrhage, retinopathy, or necrotizing enterocolitis, between the 2 time periods. An increase in the use of continuous positive airway pressure was noted in the post implementation period.
Conclusions. Variation in common processes can alter clinical outcomes. Although temporal trends in weight gain may be, in part, responsible for this trend, it appears that isolation and implementation of meaningful differences in processes can augment our desire to rapidly improve clinical outcomes.
Key Words: neonate • nutrition • quality improvement
Abbreviations: VLBW, very low birth weight • IUGC, intrauterine growth curve • NICU, neonatal intensive care unit
The American Academy of Pediatrics has suggested that to attain optimal nutritional support of the preterm neonate, postnatal growth approximating that of a normal fetus of the same postmenstrual age should be achieved.1,2 Nutrient intakes that meet the recommended dietary intakes take time to establish and are hard to maintain throughout hospital stays in preterm infants.3 This is especially true for the critically ill preterm neonate.4 As a result, a nutrient deficit typically accrues. Recommended dietary intakes are based on what is needed for maintenance and growth, with no provision to replace this accrued deficit.3
Very low birth weight (VLBW) infants typically fall below the intrauterine growth curve (IUGC) during the first 28 days after birth and then parallel the IUGC until discharge.4 This is important because early postnatal weight gain is linked to long-term weight and height status, as well as improvement in mental and developmental status.4–7 Because of the critical nature of this problem, in 1999 we selected improving neonatal growth as the top quality-improvement priority in our national network of neonatal health care providers.
During our initial baseline assessment of this outcome, we noted significant unexplained variation in weight-gain performance among the sites within our network.8 This data suggested that meaningful differences in processes between sites existed. We hypothesized that identification of these meaningful differences and then implementation of an education program to raise the awareness of our practitioners to processes associated with the better weight-gain outcomes would lead to increased weight gain across our network. Previous work in emergency medicine, surgery centers, and neonatal intensive care unit (NICU)-acquired infection supported the potential for success using the Best Demonstrated Process methodology.9 This paper reports on the completion of our quality-improvement project on improving weight gain within our network.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Measurement and Identification of Site Variation
Primary Outcome Measure
Based on a survey of neonatologists within our network and a review of our administrative data set, weight gain for VLBW infants during the first 28 days was selected for this project. The metric was established as the average daily weight gain for all 401- to 1500-g inborn infants who were not transferred and survived to 28 days. Individual daily weight gain was calculated by subtracting birth weight from the reported weight on day 28 and dividing the difference by 28. Weights and other characteristics were obtained from our administrative data set which is derived de-identified data from our computerized note system.
Site Visit Selection
The initial review of sites (n = 74) included weight-gain data from infants admitted and cared for between January 1, 1997, and June 30, 1999. For each site, we calculated an average reported daily weight gain for the first 28 days. Based on a desire to observe sites where care for VLBW infants was common, we only ranked sites (n = 53) with >15 VLBW admissions reported in our administrative database. To further ensure that the centers we visited were at the top and bottom thirds of weight-gain performance, we created a linear regression model using the whole data set to predict the "expected weight gain" for each neonate based on gestational age and need for mechanical ventilator support during the first day of life (surrogate for severity of illness). The observed weight gain was divided by the expected weight gain and centers’ average ratio was calculated. A second set of site rankings were assigned by this observed/expected ratio. Eleven sites were classified in the top third and 11 sites were in the bottom third for both metrics. Outcomes as reported to the administrative data set were contrasted for these 11 high- and 11 low-weight-gain centers to increase confidence that high weight gain was not associated in this data set with increased rates of adverse outcomes. From the 22 sites, we selected 14 for site visits guided by average daily census to achieve a balance of large and small units.
Observation Guide
In March 2000, the observation team, which consisted of 7 volunteers, performed a critical literature review starting with a structured search (weight or growth in the title or abstract, limited to English language, and infants 0–1 year of age) from PubMed (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi). We identified information related to variations in nutritional support processes for the preterm infant and the importance of weight gain for this population. From this review, the team enhanced a flow chart of the system of neonatal intensive care with processes and subordinate processes. Using this flow chart, the team incorporated nearly 200 elements of care into a draft observation guide for site visits. Team members used the guide to observe in his/her own units and, in a follow-up meeting, we enhanced and refined the observation guide. The final guide contained over 170 specific processes and subordinate processes. Each team member visited and completed the observation guide at 1 low- and 1 high-weight-gain site. The visits were primarily observational in nature, complimented by discussions with care providers. The observers were not advised of the center’s performance until the isolation meeting.
Isolation of Meaningful Differences in Process
In April 2000, the site visit observations were individually reviewed in a 2-day meeting. Meaningful differences are defined as processes observed in all or virtually all of the low-weight-gain sites, and none or virtually none of the high-weight-gain sites, or vice versa. "Virtually" was defined as 6 of 7 or 1 of 7 sites. A consultant from The Lombardy Group, LLC, which holds the copyright to the Best Demonstrated Process methodology, facilitated the isolation process.
Sharing Meaningful Differences and Implementation
We distributed a summary of the observed meaningful differences to each of our center medical directors along with his/her center’s rank. We also sent an educational packet describing the meaningful differences to all neonatologists and neonatal nurse practitioners in the Pediatrix Medical Group, Inc, network in August 2000. The unit-by-unit response to the information was not monitored in detail as the culture for improvement work varies across our network. Active participation in this phase of the project was elective. An improvement target was clearly stated. Centers in the low-weight-gain sites (<9 g/d) were expected to move to the mean of the network (10.1 g/d), middle third centers were expected to move to the average of the high-weight-gain sites (12.7 g/d), and high-weight-gain sites were to maintain their performance or make modest improvements.
Monitoring Improvement
We measured changes in weight gain for all neonates reported to our administrative data set who had both a birth weight and a reported daily weight at 28 days during a baseline period (January 1 through December 31, 1999) and following the educational intervention (January 1 to September 30, 2001). We included only inborn neonates of birth weight 401 to 1500 g who survived and were not transferred before 28 days from centers with >15 qualifying infants per year. In addition to evaluating the overall improvement of the entire network, we looked at site-specific improvement for the pre- and posteducational intervention periods.
Statistical Analysis
Continuous variables (estimated gestational age and birth weight) were evaluated using 2-tailed t tests. Categorical variables (eg, race, gender, and intraventricular hemorrhage) were evaluated using a 2-tailed 
2 test. Nonparametric continuous data were assessed using a Kruskal-Wallis analysis of variance.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Measurement and Site Selection
Of the initial 74 sites, 53 reported >15 qualifying infants to our administrative database in the preobservation period and were initially ranked. The performance histogram is provided in Fig 1. A second set of rankings was assigned using the observed/expected ratio. Eleven sites ranked in the lower and upper third in the unadjusted and adjusted rankings. These 22 sites were used for the high-low comparisons. We selected 14 sites for visits from these 22 sites to achieve a balance of large and small units based on average daily census.
|
We compared selected demographic characteristics and other clinical outcomes at the 11 high- (n = 1161 neonates) and the 11 low-weight-gain sites (n = 834 neonates). The average weight gain at the 11 sites demonstrating higher weight gain was 12.7 ± 6 g (n = 1161 neonates) compared with 8.7 ± 5 g (n = 834 neonates) in the 11 low sites and the differences in performance were consistent across all birth weight groups (Fig 2). There were no differences in estimated gestational age, frequency of male gender, use of prenatal steroids, mortality, severe intraventricular hemorrhage, or severe retinopathy. Infants cared for at high-weight-gain centers had been more often exposed to postnatal steroids (41% vs 32%) and more often required some respiratory support at day 28 (59% vs 50%, P = .01). Although length of stay was similar for neonates who survived and were discharged from the hospital, discharge weight was heavier (2.05 ± 0.4 kg vs 2.33 ± 0.6 kg, P < .01) and head circumference was larger (31.6 ± 2 cm vs 32.4 ± 2 cm, P < .01) at high-weight-gain centers than at low-weight-gain centers.
|
Observation Guide
The final observation guide contained nearly 170 processes and subordinate processes. Several domains were identified in the flowchart including transporting the infant to a NICU bed; managing fluids and electrolytes, enteral nutrition, respiratory support, cardiovascular and vascular access, and the environment; laboratory testing and reporting; NICU leadership and coordination; faculty collaboration; and finally, managing discharge. An example of one element of the guide is included in the appendix (Appendix A).
Isolation and Sharing of Meaningful Differences
Fourteen site (7 high-weight-gain and 7 low-weight-gain) visits were completed and 16 meaningful differences were isolated (Appendix B). A complete report of these meaningful differences and the center-specific rank was circulated by e-mail to all medical directors in August 2000.
Monitoring Impact
Fifty-one sites reported enough infants (15 per year) in 1999 and 2001. There was no difference in birth weight, gestational age at birth, exposure to antenatal steroids, or male gender for neonates cared for in 1999 as compared with those cared for in 2001 (Table 1). Average daily weight gain during the first 28 days increased from 10.4 g ± 6 for neonates cared for in 1999 to 12.5 g ± 6 for neonates cared for in 2001 (P < .01; Table 2). These improvements occurred in each birth weight group (Fig 3). Eight sites demonstrated a reduction in average weight gain, 4 sites were unchanged, and 39 (76%) showed some improvement in average daily weight gain during the first 28 days (Fig 4).
|
|
|
|
Despite no increase in length of hospital stay or gestational age at discharge, discharge weights and head circumferences were also significantly increased (Table 2 and Fig 5). In our initial analysis, use of steroids was higher in high-weight-gain centers than in the low-weight-gain centers. Postimplementation, we noted a significant reduction in the use of postnatal steroids in the network as a whole (29% vs 19%, P < .01), and an increase in the percentage of infants who required respiratory support on day 28 (55% in 1999 vs 58% in 2001, P < .03). The increased need for respiratory support was small and primarily related to an increase in the use of continuous positive airway pressure support (Table 2). There were no differences between the 2 time periods in the frequency of severe intraventricular hemorrhage, retinopathy, necrotizing enterocolitis, or mortality (Table 3).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Variation in outcome and process has been noted in other networks that collect outcome and clinical practice data.9–19 It is widely recognized that the choices we make in our processes of clinical care can affect outcomes.13–15 Providing outcome reports alone has not been effective in generating improvements without a parallel knowledge of the systems and processes associated with each center’s outcomes.13 Only with this association can a practitioner have the information needed to consider changes in practice or process that might generate improvements.
It appears that systems-of-care vary in many ways that are not driven by patient need but rather by habit or local history.10,12,15,17,18 These process differences traverse patient populations and can affect patient care. Practicing evidence-based medicine is important, but not all of our processes can be tested. In the absence of clinical trial data, the uncertainty in what constitutes "best" practice is typically managed by what used to work or what has always worked in that unit (ie, the homegrown way). Therefore, much of what we do continues to be based on local or regional experience and not on evidence from clinical trials.15
The quality improvement method we report provides a way to contrast what processes are in place in units with the best and worst results for a targeted clinical outcome. The observed meaningful differences in process represent multiple-site observations and not opinions, administrative intent, policies, or perceptions of a single benchmarking center.
Improvement in each center was dependent on local implementation and some chose not to participate in this project. High-weight-gain sites that received a visit did not appear to improve during the measurement period, whereas other high-weight-gain sites noted unexpected gains. This is most likely explained by the adoption of the meaningful differences in addition to an already high performance system and the existence of other important performance factors. Low-weight-gain sites that were visited had varied responses primarily based on their willingness to implement the meaningful differences.
Providing this information alone is not enough; the implementation phase is critical to success. Setting priorities for units in our network remains the domain of the medical director in partnership with the staff of the facility. Unlike the NIC/Q Collaborative of the Vermont Oxford Network, our post-project survey noted that all centers in our lower third that actively sought implementation of these meaningful differences noted improvement.13 This project reinforces the notion that reports documenting unit performance are not sufficient; there must be local investment of time and energy to implement specific changes in process and a set of effective recommendations for change.
Weight gain is a competing priority. This project allowed us to offer our units a "safe way" to restructure their priorities and redesign processes-of-care that resulted in more rapid growth of our VLBW infants in the first 28 days. Although it is tempting to assign relative value to the meaningful differences based on one’s experience or frustration with current processes, we made no attempt to do so. After reviewing the list of observations about meaningful differences, care providers evaluate their current processes and select from the list those changes that they are willing and able to implement. Ongoing monitoring allows them to recognize improvement, stagnation, or even worsening results. If performance is stagnant, they may return to the list and select additional other items for implementation.
Regression to the mean, reduced use of postnatal steroids and a temporal trend in weight gain are competing explanations for our success. We note from the 1999 and 2000 Vermont Oxford annual reports that with no change in average length of stay for infants discharged from the hospital (1999; 61 ± 31 vs 2000; 61 ± 31), discharge weight gain had increased an average of 0.9 g per day (1999; 2194 ± 527 vs 2000; 2250 ± 525). Without a change in length of stay in our network, we noted a higher weight gain of 2.6 g/d difference between these 2 periods (data not shown). Compounding the positive influence of several changes over this period of time with no increase in adverse outcomes is reassuring. We suggest, but cannot verify, that this method in isolation was responsible for the improvement noted. The long-term impact of this project will be measured by longitudinal monitoring, the use of which will enable us to conclude that the improvements we observed are unique to units responsive to our findings.
|
|
|
|
|
|
FOOTNOTES |
|---|
Received for publication Jun 6, 2002; Accepted Nov 21, 2002.
Reprint requests to (B.T.B.) Wesley Medical Center, 550 North Hillside, Wichita, KS 67214-4976. E-mail: barrybloom{at}aol.com
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
- American Academy of Pediatrics, Committee on Nutrition. Nutritional needs of low-birth-weight infants.
Pediatrics.1985; 75
:976
–986
[Abstract/Free Full Text] - Kashyap S, Ohira-Kist K, Abildskov K, et al. Effects of quality of energy intake on growth and metabolic response of enterally fed low-birth-weight infants. Pediatr Res.2001; 50 :390 –397[Web of Science][Medline]
- Embleton NE, Pang N, Cooke RJ. Postnatal malnutrition and growth retardation: an inevitable consequence of current recommendations in preterm infants?
Pediatrics.2001; 107
:270
–273
[Abstract/Free Full Text] - Ehrenkranz RA, Younes N, Lemons JA, et al. Longitudinal growth of hospitalized very low birth weight infants.
Pediatrics.1999; 104
:280
–289
[Abstract/Free Full Text] - Carver JD, Wu PY, Hall RT, et al. Growth of preterm infants fed nutrient-enriched or term formula after hospital discharge.
Pediatrics.2001; 107
:683
–689
[Abstract/Free Full Text] - Cooke RJ, Embleton ND, Griffin IJ, Wells JC, McCormick KP. Feeding preterm infants after hospital discharge: growth and development at 18 months of age. Pediatr Res.2001; 49 :719 –722[Web of Science][Medline]
- Lucas A, Fewtrell MS, Morley R, et al. Randomized trial of nutrient-enriched formula versus standard formula for postdischarge preterm infants.
Pediatrics.2001; 108
:703
–711
[Abstract/Free Full Text] - Clark RH, Bloom BT, Thomas P, Peabody J. Application of the best demonstrated process method to a neonatal growth quality-improvement project. J Perinatol.2000; 20 :488
- Hoffenberg S, Hill MB, Houry D. Does sharing process differences reduce patient length of stay in the emergency department? Ann Emerg Med.2001; 38 :533 –540[CrossRef][Web of Science][Medline]
- Perfect SS, Dodd E, Thomas P, Peabody J, Clark R. Home apnea monitor use in preterm infants discharged from newborn intensive care units. J Pediatr.2001; 139 :245 –248[CrossRef][Web of Science][Medline]
- Al-Aweel I, Pursley DM, Rubin LP, Shah B, Weisberger S, Richardson DK. Variations in prevalence of hypotension, hypertension, and vasopressor use in NICUs. J Perinatol.2001; 21 :272 –278[CrossRef][Medline]
- Eichenwald EC, Blackwell M, Lloyd JS, Tran T, Wilker RE, Richardson DK. Inter-neonatal intensive care unit variation in discharge timing: influence of apnea and feeding management.
Pediatrics.2001; 108
:928
–933
[Abstract/Free Full Text] - Horbar JD, Rogowski J, Plsek PE, et al. Collaborative quality improvement for neonatal intensive care. NIC/Q Project Investigators of the Vermont Oxford Network.
Pediatrics.2001; 107
:14
–22
[Abstract/Free Full Text] - Synnes AR, Chien LY, Peliowski A, Baboolal R, Lee SK. Variations in intraventricular hemorrhage incidence rates among Canadian neonatal intensive care units. J Pediatr.2001; 138 :525 –531[CrossRef][Web of Science][Medline]
- Lee SK, McMillan DD, Ohlsson A, et al. Variations in practice and outcomes in the Canadian NICU network: 1996–1997.
Pediatrics.2000; 106
:1070
–1079
[Abstract/Free Full Text] - Horbar JD. The Vermont Oxford Network: evidence-based quality improvement for neonatology. Pediatrics.1999; 103(suppl) :350 –359
- Bednarek FJ, Weisberger S, Richardson DK, Frantz ID, Shah B, Rubin LP. Variations in blood transfusions among newborn intensive care units. SNAP II Study Group. J Pediatr.1998; 133 :601 –607[CrossRef][Web of Science][Medline]
- Ringer SA, Richardson DK, Sacher RA, Keszler M, Churchill WH. Variations in transfusion practice in neonatal intensive care.
Pediatrics.1998; 101
:194
–200
[Abstract/Free Full Text] - Horbar JD, Badger GJ, Lewit EM, Rogowski J, Shiono PH. Hospital and patient characteristics associated with variation in 28-day mortality rates for very low birth weight infants. Vermont Oxford Network.
Pediatrics.1997; 99
:149
–156
[Abstract/Free Full Text]
PEDIATRICS (ISSN 1098-4275). ©2003 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  R. H. Clark, D. H. Chace, A. R. Spitzer, and for the Pediatrix Amino Acid Study Group Effects of Two Different Doses of Amino Acid Supplementation on Growth and Blood Amino Acid Levels in Premature Neonates Admitted to the Neonatal Intensive Care Unit: A Randomized, Controlled Trial Pediatrics, December 1, 2007; 120(6): 1286 - 1296. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J Cockerill, S Uthaya, C J Dore, and N Modi Accelerated postnatal head growth follows preterm birth Arch. Dis. Child. Fetal Neonatal Ed., May 1, 2006; 91(3): F184 - F187. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
