571667
research-article2015
PENXXX10.1177/0148607115571667Journal of Parenteral and Enteral NutritionJadcherla et al.
Original Communication
Impact of Process Optimization and Quality Improvement Measures on Neonatal Feeding Outcomes at an All-Referral Neonatal Intensive Care Unit
Journal of Parenteral and Enteral Nutrition Volume XX Number X Month 201X 1–11 © 2015 American Society for Parenteral and Enteral Nutrition DOI: 10.1177/0148607115571667 jpen.sagepub.com hosted at online.sagepub.com
Sudarshan R. Jadcherla, MD, FRCPI, DCH, AGAF1,2,3,4; James Dail, MBA, LSSBB5; Manish B. Malkar, MD, MPH1,2,3,4; Richard McClead, MD, MHA4,5; Kelly Kelleher, MD, MPH4,6; and Leif Nelin, MD2,4
Abstract Aim: We hypothesized that the implementation of a feeding quality improvement (QI) program among premature neonates accelerates feeding milestones, safely lowering hospital length of stay (LOS) compared with the baseline period. Methods: Baseline data were collected for 15 months (N = 92) prior to initiating the program, which involved development and implementation of a standardized feeding strategy in eligible premature neonates. Process optimization, implementation of feeding strategy, monitoring compliance, multidisciplinary feeding rounds, and continuous education strategies were employed. The main outcomes included the ability and duration to reach enteral feeds–120 (mL/kg/d), oral feeds–120 (mL/kg/d), and ad lib oral feeding. Balancing measures included growth velocities, comorbidities, and LOS. Results: Comparing baseline versus feeding program (N = 92) groups, respectively, the feeding program improved the number of infants receiving trophic feeds (34% vs 80%, P < .002), trophic feeding duration (14.8 ± 10.3 days vs 7.6 ± 8.1 days, P < .0001), time to enteral feeds–120 (16.3 ± 15.4 days vs 11.4 ± 10.4 days, P < .04), time from oral feeding onset to oral feeds–120 (13.2 ± 16.7 days vs 19.5 ± 15.3 days, P < .0001), time from oral feeds–120 to ad lib feeds at discharge (22.4 ± 27.2 days vs 18.6 ± 21.3 days, P < .01), weight velocity (24 ± 6 g/d vs 27 ± 11 g/d, P < .03), and LOS (104.2 ± 51.8 vs 89.3 ± 46.0, P = .02). Mortality, readmissions within 30 days, and comorbidities were similar. Conclusions: Process optimization and the implementation of a standardized feeding strategy minimize practice variability, accelerating the attainment of enteral and oral feeding milestones and decreasing LOS without increasing adverse morbidities. (JPEN J Parenter Enteral Nutr. XXXX;xx:xx-xx)
Keywords premature neonate; feeding management; quality improvement; feeding outcomes; multidisciplinary approach
Clinical Relevancy Statement This article offers insight into the benefits of implementing a standardized feeding guidelines quality improvement initiative within the neonatal intensive care unit setting.
Background Overview of the Problem Advances in neonatal care and transport of sick premature infants are resulting in increased survival rates, along with increases in aerodigestive morbidities, resource utilization, and hospital length of stay (LOS).1 Successful enteral and oral feeding progression among hospitalized sick premature neonates in tertiary care hospitals remains a major challenge and a primary contributor to LOS. Feeding- and nutritionrelated morbidities, LOS, and the economic burden continue to be on the rise globally. Variability with approaches to feeding among multidisciplinary providers in the neonatal intensive care unit (NICU) may be at least contributory to the
staggering healthcare costs for preterm birth estimated at more than $26 billion each year in the United States.1,2
Outline of the Problem Feeding difficulties in the premature neonate are complex, and solutions require special attention to the process details. Feeding difficulties become apparent with age and develop through the enteral feeding phase, transition to oral feeding phase, and/or oral feeding phase. Measures of feeding difficulties may be attributed to delayed acquisition of independent enteral or oral feeding milestones or to a lack of progressive tolerance of increasing feeding volumes. These difficulties may be related to immaturity, regurgitation or emesis, suspicion of necrotizing enterocolitis (NEC) or gastroesophageal reflux disease (GERD), gastroparesis or gastrointestinal dysmotility, life-threatening events, and aspiration syndromes.3,4 Regardless, these difficulties further prolong the hospital course, resulting in an increased need for central venous access, parenteral nutrition (PN), intragastric feeding (ie, gastrostomy
2 tube placement), and/or oxygen supplementation. Thus, evaluation and management of feeding problems in a preterm infant are ongoing processes that change with time and from patient to patient. Therefore, development of standardized feeding approaches must include consideration of the unique physiology of the preterm infant.
Outline of Context Nationwide Children’s Hospital NICU is the only level IV NICU in Franklin County, Ohio, but it does not provide a delivery service. Thus, this is an all-referral unit providing tertiary care for transported sick outborn preterm neonates. The unit houses 42 level IV beds, 28 level III chronic NICU beds, and 16 level II step-down beds. Overall, the providers include 55 neonatologists, 40 neonatal nurse practitioners, more than 300 registered nurses, 9 registered dietitians, 5 lactation consultants, 5 occupational therapists, and 2 pharmacists, all influencing feeding management, thus contributing to potential variability in practice implementation. Consequently, this diverse talent of providers can influence implementation of feeding-related decisions because of the variability in individual practice attitudes, knowledge, and interdisciplinary communications. This problem is not only local to our institution but also occurs at multiple other large referral institutions, as evidenced by the variability in hospital outcomes among premature infants.5-8
Assessment of Feeding Problems Approach to Planning the Intervention We observed variability in feeding methods because individual clinical providers determined the timing of gut priming feeds (trophic feeds), incremental advances in feedings, transitioning to oral feedings, and establishing full oral feedings. We examined the institutional practices and processes, physiological and pathophysiological evidence, and theory and research literature so as to provide a basis for the development
Journal of Parenteral and Enteral Nutrition XX(X) of a simplified feeding program. Few published data exist to benchmark feeding outcomes for this unique population of sick transported premature infants to an all-referral tertiary care unit. Lack of benchmarks is understandable since the levels of care at the referring hospital and comfort levels of the staff providing early aerodigestive support to the premature infant vary widely. Thus, this unique situation demands a personalized approach at the tertiary care all-referral NICU (such as ours). Given the diverse talent at our institution and the variation in feeding approaches within the institution, we embarked on this quality improvement (QI) initiative to minimize variability with feeding management practices among the providers in our NICU. The Nationwide Children’s Hospital Institutional Review Board approved this project as a QI Project, and informed consent was not deemed necessary. The study complied with the Health Insurance Portability and Accountability Act.
Objective of Intended Improvement Our approach to improvement was driven by the baseline data (see below sections) and with the consideration that our overall objectives were to develop an all-inclusive providerand infant- driven feeding program and to evaluate the impact of this simplified, individualized, milestone-targeted, pragmatic, longitudinal, and educational (SIMPLE) feeding strategy on the overall feeding and outcome metrics while monitoring balancing measures (Figure 1). Specifically, we targeted progression to enteral feeding, transition to oral feeding, and oral feeding to discharge process and measured the enteral and oral feeding milestones, growth trends, length of hospitalization, and deaths and comorbidities during the NICU hospitalization.
Feeding QI and Safety Hypothesis We tested the hypothesis that the standardized SIMPLE feeding approach would be superior in the quality metrics measured compared with the nonstandardized feeding approach
From the 1The Neonatal and Infant Feeding Disorders Program, 2Center for Perinatal Research, and 3Innovative Infant Feeding Disorders Research Program, The Research Institute at Nationwide Children’s Hospital, Columbus, Ohio, USA; 4Department of Pediatrics, The Ohio State University College of Medicine, Columbus, Ohio, USA; 5Neonatal Quality Improvement Service, Nationwide Children’s Hospital, Columbus, Ohio, USA; and 6 Center for Innovative Pediatric Health, The Research Institute at Nationwide Children’s Hospital, Columbus, Ohio, USA. Financial disclosure: Nationwide Children’s Hospital Neonatal Service Line. Received for publication November 17, 2014; accepted for publication December 18, 2014. Simplified, Individualized, Milestone-targeted, Pragmatic, Longitudinal, and Educational (SIMPLE) feeding strategy © Sudarshan R. Jadcherla, MD Corresponding Author: Sudarshan R. Jadcherla, MD, FRCPI, DCH, AGAF, Professor of Pediatrics & Associate Division Chief, Principal Investigator, Innovative Feeding Disorders Research Program, Center for Perinatal Research, WB 5211, The Research Institute at Nationwide Children’s Hospital, Nationwide Children’s Hospital, 575 Children’s Crossroads, Columbus, OH 43215, USA. Email: [email protected]
Jadcherla et al
3 Aim & Key Drivers Design Changes & Intervenons Key Drivers Provider educaon and parcipaon Specific Aim 1) Prevent Feeding Failure by 38 wks PMA 2) Decrease LOS from 96 days to 86 days over a 12 month program
Parent educaon and involvement Mul-disciplinary collaboraon
Feeding rounds Educaon Discuss Science
Simplified feeding program
Inadequate Database Compliance to the program
Enhance muldisciplinary Rounds Provide educaonal resources Ensure concepts reinforcement Parent learning resources Clarify expectaons vs. reality Weekly feeding updates to parents
Modify feeding protocols Develop focused guideline Personalized approach Develop acceptable database Verificaon of data Weekly feeding rounds Follow-up & updates Core group meengs
Figure 1. Smart aim and key drivers encompassing simplified, individualized, milestone-targeted, pragmatic, longitudinal, and educational (SIMPLE) feeding strategy. SIMPLE feeding strategy © Sudarshan R. Jadcherla.
used in the baseline period. We compared the prospectively collected feeding milestones and safety outcomes with data from the baseline cohort admitted prior to the institution of the SIMPLE feeding approach.
several groups of providers, with our academic practice group caring for about 70% to 80% of admissions. During the baseline period, there were 188 patients admitted to our practice, and during the program period, there were 264 admissions to our practice.
Study Design and Timeline Details Measurement of the problem was first ascertained by collecting retrospective data from neonates admitted from January 1, 2009, to March 15, 2010 (before the institution of the SIMPLE feeding program), referenced as the baseline group. Electronic medical records were scrutinized by trained neonatal registered nurses and analyzed by a QI analyst. There were 92 neonates during this baseline period who qualified based on the inclusion and exclusion criteria given below. During the baseline period, the feeding strategy was prescribed by the individual attending physician and implemented by the clinical care providers. An action medical research model9,10 was applied to evaluate the impact of the SIMPLE feeding program prospectively. For comparison, the feeding program subjects also included 92 neonates admitted to the same NICU but between June 15, 2010, and April 30, 2012 (SIMPLE feeding program). We aimed to compare the same sample size (N = 92) in both groups. The difference in duration of the baseline and SIMPLE feeding program is due to a decreased number of eligible infants in our NICU during the program period. In addition, only infants from our academic practice group were included in this study. Our NICU is an open unit composed of
Subject Inclusion and Exclusion Criteria Inclusion criteria were similar for baseline and feeding program groups and included those preterm neonates born ≤32 weeks of gestational age at birth, ≤34 weeks of postmenstrual age (PMA) on admission, and who survived to discharge. Infants were excluded if they were admitted with a diagnosis of medical or surgical NEC, neonatal abstinence syndrome, genetic/chromosomal defects, and congenital birth defects or if they had undergone any form of gastrointestinal or neurologic surgery.
Strategies for QI and Change Process Optimization and Development of Guidelines for Neonatal Feeding Management The overwhelming response to a local survey given to stakeholders in the NICU was to minimize the variability in feeding practices among neonatal feeding providers. Therefore, the following steps were taken: (1) We first developed goals and
4 defined process changes and measures (Figure 1). (2) To implement the process optimization of the program, a dedicated core group of feeding champions were enlisted as members of the task force. (3) Since there is no benchmark available for defining and targeting feeding milestones in the context of sick premature infants with aerodigestive and feeding problems transported to an all-referral tertiary care NICU, we developed pragmatic feeding milestone targets based on previous studies,1 Cochrane reviews,11-19 neonatal GI physiology,9,10,20 neonatal gastrointestinal and pharyngo-esophageal motility,9,10,20 and safe feeding strategies.4,21 (4) Upon fulfilling the inclusion and exclusion criteria, each subject in this QI initiative was followed during focused multidisciplinary feeding rounds (see below), and targeted milestones were personalized for each neonate and displayed at the crib side. Feeding providers and the neonatal care team were therefore consistently reminded of these targets. This approach was acceptable to all of the stakeholders and feeding providers offering care in the NICU. The feeding milestones targets were defined as follows: (a) start of trophic feeding as 10–20 mL/kg/d within 3 days of admission and continued for a duration of 3–14 days overall. The provision of trophic or gut priming feeds and the duration was dependent on the age at referral, sickness or reasons for referral, and level of acuity. This approach was agreed upon because of varied gestational ages and diverse critical clinical states on admission. (b) Progression to enteral feeds–120, defined as enteral gavage feeding of 120 mL/kg/d, by day of life 14–28. (c) First oral feeding by 33–34 weeks PMA. (d) Attainment of oral feeds–120, defined as an oral intake of 120 mL/kg/d, by 36–38 weeks PMA. (e) Ad lib oral feeding at discharge. Because our institution is an outborn NICU, we have high referral rates of extremely premature infants with a higher incidence of bronchopulmonary dysplasia (BPD). Since most BPD infants are on feeding volume restriction of 120–130 mL/ kg/d, we chose attainment of oral feeds–120 as the milestone for the purpose of standardization.
Quality Compliance Monitoring Rounds Because of variability in the level of acuity and in referral patterns, and to monitor compliance with the QI initiative, multidisciplinary feeding rounds were conducted twice a week on each infant to (1) provide team education regarding factors that are helping or impeding feeding progress, (2) address the key drivers, (3) monitor compliance, and (4) provide personalized guidance in situations of feeding delays. Physician-led multidisciplinary feeding rounds included neonatology, neonatal nurse practitioners, nurses, nutritionists, lactation consultants, occupational therapists, and parents of individual infants. The approach taken at the multidisciplinary feeding rounds is akin to action medical research.22,23 Infant-focused and providerfocused feeding-related discussions were provided to include (1) appropriate personalization of feeding strategies, (2) educational opportunities for the providers regarding mechanisms of
Journal of Parenteral and Enteral Nutrition XX(X) dysfunctional feeding patterns, and (3) engagement of parents by providing anticipatory guidance.
Methods of Data Evaluation and Continuous Feedback Demographic and morbidity factors as well as feeding milestone data gathered for the baseline group (collected by retrospective chart review) were summarized. Data from the feeding program group were collected prospectively, weekly. Data included demographic and disease characteristics, risk factors, feeding milestones, and hospital LOS. Education of the feeding champions, reiteration of process optimization, and discussion of smart aims and key drivers were reviewed each quarter along with the data from enrolled patients.
Statistical Analysis Data were compared between the baseline and the feeding program groups using analysis of variance for assessing variability within and between groups, unpaired t test to test continuous variables, and χ2 tests to compare categorical variables and proportions. Statistical control charts were used for comparing cumulative LOS between baseline and comparison groups. Data are presented as percentages, mean ± SD, median (interquartile range), or as stated. A P value of 2 BPD, n (%) PDA medical treatment, n (%) PDA surgical treatment, n (%) Duration of ventilation, d n Median (IQR) Duration of CPAP, d n Median (IQR) NEC medical treatment, n (%) NEC surgical treatment, n (%) Deaths, n (%)a
Baseline Program (n = 92)
Feeding Program (n = 92)
92 26.0 (23.0–32.0)
92 27.5 (23.0–32.0)
92 862.5 (691.0–1156.0) 9 (10) 82 (89)
87 965.0 (750.0–1326.0) 8 (9) 85 (92)
26 (28) 4 (4) 53 (57) 22 (24) 30 (33)
23 (25) 0 (0) 53 (57) 31 (34) 16 (17)
55 23.0 (2.0–123.0)
45 15.0 (2.0–106.0)
77 30.0 (1–99) 2 (2) 6 (7) 10 of 102 (10)
61 23.0 (1–95) 3 (3) 4 (4) 3 of 95 (3)
P Value .07 .07 .89 .45 0.62 1.00 .14 .17 .14 .02 .65 .52 .06
BPD, bronchopulmonary dysplasia (defined as oxygen requirement at 36 weeks postmenstrual age); CPAP, continuous positive airway pressure; IVH, intraventricular hemorrhage; NEC, necrotizing enterocolitis; PDA, patent ductus arteriosus; RDS, respiratory distress syndrome; SGA, small for gestational age. a Denominator includes deaths and survivors.
87% at start of enteral feeds, 79% at full enteral, 46% at oral feeding start, and 33% at full oral feeds.
Comparisons of Hospital LOS and Balancing Measures The SIMPLE feeding program was effective in decreasing LOS, as shown in the control chart (Figure 3). Special cause variation effect is observed as there are 11 consecutive points (>8 points24) below the baseline from April 2011. A shift in baseline (April 2011 onward) due to special cause variation is depicted (Figure 3). The DOL at admission to our NICU did not differ between the 2 groups: 9.3 ± 9.9 for the baseline group and 13.1 ± 14.8 for the program group (P = .17). The rate of readmission was 22% (20 of 92) for the baseline group and 11% (10 of 92) for the feeding program group (P = .07). Other balancing measures evaluated were weight gain velocity (24.0 ± 6.1 g/d baseline group vs 27.3 ± 11.4 g/d program group, P < .03), duration of PN (20.2 ± 14 days vs 16.7 ± 12 days, P = .08), duration of peripherally inserted central catheter line (26.2 ± 20.3 days vs 21.8 ± 15 days, P = .32), NEC rates (no difference, as shown in Table 1), and proven sepsis rates (14% vs 7%, P = .28), respectively.
Discussion Potential Benefits and Mechanisms for Change Variability in neonatal feeding strategies is due to both infantdependent factors and provider-dependent factors. Rigid feeding protocols, individual variability, and subjectivity in the interpretation of the symptoms pertinent to neonatal feeding difficulties complicate feeding progression. In this study, we describe our QI efforts to develop a standardized yet practical approach to neonatal feeding among such sick premature infants referred from delivery hospitals. The salient features of our feeding program are (1) simplification and optimization of our neonatal feeding paradigm, which allowed for successful and timely acquisition of feeding milestones; (2) by adopting individualized care, provider-dependent variability was minimized; and (3) compliance monitoring at feeding rounds was necessary to overcome the challenges of attaining feeding milestones. Educational activities were key to achieving our goals and targeted to (1) feeding methods by clarifying when, how, and why to advance feeds to attain targeted milestones; (2) educating providers by discussing and reviewing the science behind neonatal feeding; (3) engaging parents by early
6
Journal of Parenteral and Enteral Nutrition XX(X)
Enteral Feeds -120
Oral Feeds -120
p = 0.01
p = 0.02
100%
100%
90%
90%
80%
80%
70%
70%
60% 50%
81% On- me Improvement
40% 30%
Other
20%
Late Milestone
10%
On- me Milestone
0%
Baseline
Program Group
% Population
% Population
A
60%
72% On-Time Improvement
50% 40% 30%
Other
20%
Gastrostomy
10%
Late Milestone On- me Milestone
0%
Baseline
Program Group
B
Figure 2. Proportion of infants achieving the milestone of enteral feeds–120 and oral feeds–120 is shown (A). Comparison of postmenstrual ages (PMAs) depicting maturation at the attainment of enteral feeds–120, first oral feed, oral feeds–120, and discharge (B).
and consistently providing support, education, and anticipatory guidance on the need for rational feeding strategies; and (4)
developing innovative feeding strategies in a timely manner for those with feeding difficulties.
Jadcherla et al
7
Table 2. Chronological Age (DOL), Maturational (PMA) Age, and Duration to Acquisition of Targeted Feeding Milestones.a Characteristic
Baseline Program (n = 92)
Feeding Program (n = 92)
P Value
3.5 (3.0–5.0) (n = 34) 29.0 (26.0–30.0) (n = 34) 16.0 (5.0–22.0) (n = 66) 17.5 (9.0–23.0) (n = 66) 29.0 (27.0–31.0) (n = 66) 9.0 (6.0–21.0) (n = 79) 28.0 (18.0–39.0) (n = 79) 31.0 (29.0–32.0) (n = 79)
3.0 (2.0–5.0) (n = 72) 27.4 (26.1–30.1) (n = 72) 5.0 (1.5–11.5) (n = 72) 9.5 (5.0–17.0) (n = 72) 29.0 (27.0–31.0) (n = 72) 8.0 (5.0–14.0) (n = 81) 19.0 (11.0–29.0) (n = 81) 30.0 (29.0–32.0) (n = 81)
.27 .26
research-article2015
PENXXX10.1177/0148607115571667Journal of Parenteral and Enteral NutritionJadcherla et al.
Original Communication
Impact of Process Optimization and Quality Improvement Measures on Neonatal Feeding Outcomes at an All-Referral Neonatal Intensive Care Unit
Journal of Parenteral and Enteral Nutrition Volume XX Number X Month 201X 1–11 © 2015 American Society for Parenteral and Enteral Nutrition DOI: 10.1177/0148607115571667 jpen.sagepub.com hosted at online.sagepub.com
Sudarshan R. Jadcherla, MD, FRCPI, DCH, AGAF1,2,3,4; James Dail, MBA, LSSBB5; Manish B. Malkar, MD, MPH1,2,3,4; Richard McClead, MD, MHA4,5; Kelly Kelleher, MD, MPH4,6; and Leif Nelin, MD2,4
Abstract Aim: We hypothesized that the implementation of a feeding quality improvement (QI) program among premature neonates accelerates feeding milestones, safely lowering hospital length of stay (LOS) compared with the baseline period. Methods: Baseline data were collected for 15 months (N = 92) prior to initiating the program, which involved development and implementation of a standardized feeding strategy in eligible premature neonates. Process optimization, implementation of feeding strategy, monitoring compliance, multidisciplinary feeding rounds, and continuous education strategies were employed. The main outcomes included the ability and duration to reach enteral feeds–120 (mL/kg/d), oral feeds–120 (mL/kg/d), and ad lib oral feeding. Balancing measures included growth velocities, comorbidities, and LOS. Results: Comparing baseline versus feeding program (N = 92) groups, respectively, the feeding program improved the number of infants receiving trophic feeds (34% vs 80%, P < .002), trophic feeding duration (14.8 ± 10.3 days vs 7.6 ± 8.1 days, P < .0001), time to enteral feeds–120 (16.3 ± 15.4 days vs 11.4 ± 10.4 days, P < .04), time from oral feeding onset to oral feeds–120 (13.2 ± 16.7 days vs 19.5 ± 15.3 days, P < .0001), time from oral feeds–120 to ad lib feeds at discharge (22.4 ± 27.2 days vs 18.6 ± 21.3 days, P < .01), weight velocity (24 ± 6 g/d vs 27 ± 11 g/d, P < .03), and LOS (104.2 ± 51.8 vs 89.3 ± 46.0, P = .02). Mortality, readmissions within 30 days, and comorbidities were similar. Conclusions: Process optimization and the implementation of a standardized feeding strategy minimize practice variability, accelerating the attainment of enteral and oral feeding milestones and decreasing LOS without increasing adverse morbidities. (JPEN J Parenter Enteral Nutr. XXXX;xx:xx-xx)
Keywords premature neonate; feeding management; quality improvement; feeding outcomes; multidisciplinary approach
Clinical Relevancy Statement This article offers insight into the benefits of implementing a standardized feeding guidelines quality improvement initiative within the neonatal intensive care unit setting.
Background Overview of the Problem Advances in neonatal care and transport of sick premature infants are resulting in increased survival rates, along with increases in aerodigestive morbidities, resource utilization, and hospital length of stay (LOS).1 Successful enteral and oral feeding progression among hospitalized sick premature neonates in tertiary care hospitals remains a major challenge and a primary contributor to LOS. Feeding- and nutritionrelated morbidities, LOS, and the economic burden continue to be on the rise globally. Variability with approaches to feeding among multidisciplinary providers in the neonatal intensive care unit (NICU) may be at least contributory to the
staggering healthcare costs for preterm birth estimated at more than $26 billion each year in the United States.1,2
Outline of the Problem Feeding difficulties in the premature neonate are complex, and solutions require special attention to the process details. Feeding difficulties become apparent with age and develop through the enteral feeding phase, transition to oral feeding phase, and/or oral feeding phase. Measures of feeding difficulties may be attributed to delayed acquisition of independent enteral or oral feeding milestones or to a lack of progressive tolerance of increasing feeding volumes. These difficulties may be related to immaturity, regurgitation or emesis, suspicion of necrotizing enterocolitis (NEC) or gastroesophageal reflux disease (GERD), gastroparesis or gastrointestinal dysmotility, life-threatening events, and aspiration syndromes.3,4 Regardless, these difficulties further prolong the hospital course, resulting in an increased need for central venous access, parenteral nutrition (PN), intragastric feeding (ie, gastrostomy
2 tube placement), and/or oxygen supplementation. Thus, evaluation and management of feeding problems in a preterm infant are ongoing processes that change with time and from patient to patient. Therefore, development of standardized feeding approaches must include consideration of the unique physiology of the preterm infant.
Outline of Context Nationwide Children’s Hospital NICU is the only level IV NICU in Franklin County, Ohio, but it does not provide a delivery service. Thus, this is an all-referral unit providing tertiary care for transported sick outborn preterm neonates. The unit houses 42 level IV beds, 28 level III chronic NICU beds, and 16 level II step-down beds. Overall, the providers include 55 neonatologists, 40 neonatal nurse practitioners, more than 300 registered nurses, 9 registered dietitians, 5 lactation consultants, 5 occupational therapists, and 2 pharmacists, all influencing feeding management, thus contributing to potential variability in practice implementation. Consequently, this diverse talent of providers can influence implementation of feeding-related decisions because of the variability in individual practice attitudes, knowledge, and interdisciplinary communications. This problem is not only local to our institution but also occurs at multiple other large referral institutions, as evidenced by the variability in hospital outcomes among premature infants.5-8
Assessment of Feeding Problems Approach to Planning the Intervention We observed variability in feeding methods because individual clinical providers determined the timing of gut priming feeds (trophic feeds), incremental advances in feedings, transitioning to oral feedings, and establishing full oral feedings. We examined the institutional practices and processes, physiological and pathophysiological evidence, and theory and research literature so as to provide a basis for the development
Journal of Parenteral and Enteral Nutrition XX(X) of a simplified feeding program. Few published data exist to benchmark feeding outcomes for this unique population of sick transported premature infants to an all-referral tertiary care unit. Lack of benchmarks is understandable since the levels of care at the referring hospital and comfort levels of the staff providing early aerodigestive support to the premature infant vary widely. Thus, this unique situation demands a personalized approach at the tertiary care all-referral NICU (such as ours). Given the diverse talent at our institution and the variation in feeding approaches within the institution, we embarked on this quality improvement (QI) initiative to minimize variability with feeding management practices among the providers in our NICU. The Nationwide Children’s Hospital Institutional Review Board approved this project as a QI Project, and informed consent was not deemed necessary. The study complied with the Health Insurance Portability and Accountability Act.
Objective of Intended Improvement Our approach to improvement was driven by the baseline data (see below sections) and with the consideration that our overall objectives were to develop an all-inclusive providerand infant- driven feeding program and to evaluate the impact of this simplified, individualized, milestone-targeted, pragmatic, longitudinal, and educational (SIMPLE) feeding strategy on the overall feeding and outcome metrics while monitoring balancing measures (Figure 1). Specifically, we targeted progression to enteral feeding, transition to oral feeding, and oral feeding to discharge process and measured the enteral and oral feeding milestones, growth trends, length of hospitalization, and deaths and comorbidities during the NICU hospitalization.
Feeding QI and Safety Hypothesis We tested the hypothesis that the standardized SIMPLE feeding approach would be superior in the quality metrics measured compared with the nonstandardized feeding approach
From the 1The Neonatal and Infant Feeding Disorders Program, 2Center for Perinatal Research, and 3Innovative Infant Feeding Disorders Research Program, The Research Institute at Nationwide Children’s Hospital, Columbus, Ohio, USA; 4Department of Pediatrics, The Ohio State University College of Medicine, Columbus, Ohio, USA; 5Neonatal Quality Improvement Service, Nationwide Children’s Hospital, Columbus, Ohio, USA; and 6 Center for Innovative Pediatric Health, The Research Institute at Nationwide Children’s Hospital, Columbus, Ohio, USA. Financial disclosure: Nationwide Children’s Hospital Neonatal Service Line. Received for publication November 17, 2014; accepted for publication December 18, 2014. Simplified, Individualized, Milestone-targeted, Pragmatic, Longitudinal, and Educational (SIMPLE) feeding strategy © Sudarshan R. Jadcherla, MD Corresponding Author: Sudarshan R. Jadcherla, MD, FRCPI, DCH, AGAF, Professor of Pediatrics & Associate Division Chief, Principal Investigator, Innovative Feeding Disorders Research Program, Center for Perinatal Research, WB 5211, The Research Institute at Nationwide Children’s Hospital, Nationwide Children’s Hospital, 575 Children’s Crossroads, Columbus, OH 43215, USA. Email: [email protected]
Jadcherla et al
3 Aim & Key Drivers Design Changes & Intervenons Key Drivers Provider educaon and parcipaon Specific Aim 1) Prevent Feeding Failure by 38 wks PMA 2) Decrease LOS from 96 days to 86 days over a 12 month program
Parent educaon and involvement Mul-disciplinary collaboraon
Feeding rounds Educaon Discuss Science
Simplified feeding program
Inadequate Database Compliance to the program
Enhance muldisciplinary Rounds Provide educaonal resources Ensure concepts reinforcement Parent learning resources Clarify expectaons vs. reality Weekly feeding updates to parents
Modify feeding protocols Develop focused guideline Personalized approach Develop acceptable database Verificaon of data Weekly feeding rounds Follow-up & updates Core group meengs
Figure 1. Smart aim and key drivers encompassing simplified, individualized, milestone-targeted, pragmatic, longitudinal, and educational (SIMPLE) feeding strategy. SIMPLE feeding strategy © Sudarshan R. Jadcherla.
used in the baseline period. We compared the prospectively collected feeding milestones and safety outcomes with data from the baseline cohort admitted prior to the institution of the SIMPLE feeding approach.
several groups of providers, with our academic practice group caring for about 70% to 80% of admissions. During the baseline period, there were 188 patients admitted to our practice, and during the program period, there were 264 admissions to our practice.
Study Design and Timeline Details Measurement of the problem was first ascertained by collecting retrospective data from neonates admitted from January 1, 2009, to March 15, 2010 (before the institution of the SIMPLE feeding program), referenced as the baseline group. Electronic medical records were scrutinized by trained neonatal registered nurses and analyzed by a QI analyst. There were 92 neonates during this baseline period who qualified based on the inclusion and exclusion criteria given below. During the baseline period, the feeding strategy was prescribed by the individual attending physician and implemented by the clinical care providers. An action medical research model9,10 was applied to evaluate the impact of the SIMPLE feeding program prospectively. For comparison, the feeding program subjects also included 92 neonates admitted to the same NICU but between June 15, 2010, and April 30, 2012 (SIMPLE feeding program). We aimed to compare the same sample size (N = 92) in both groups. The difference in duration of the baseline and SIMPLE feeding program is due to a decreased number of eligible infants in our NICU during the program period. In addition, only infants from our academic practice group were included in this study. Our NICU is an open unit composed of
Subject Inclusion and Exclusion Criteria Inclusion criteria were similar for baseline and feeding program groups and included those preterm neonates born ≤32 weeks of gestational age at birth, ≤34 weeks of postmenstrual age (PMA) on admission, and who survived to discharge. Infants were excluded if they were admitted with a diagnosis of medical or surgical NEC, neonatal abstinence syndrome, genetic/chromosomal defects, and congenital birth defects or if they had undergone any form of gastrointestinal or neurologic surgery.
Strategies for QI and Change Process Optimization and Development of Guidelines for Neonatal Feeding Management The overwhelming response to a local survey given to stakeholders in the NICU was to minimize the variability in feeding practices among neonatal feeding providers. Therefore, the following steps were taken: (1) We first developed goals and
4 defined process changes and measures (Figure 1). (2) To implement the process optimization of the program, a dedicated core group of feeding champions were enlisted as members of the task force. (3) Since there is no benchmark available for defining and targeting feeding milestones in the context of sick premature infants with aerodigestive and feeding problems transported to an all-referral tertiary care NICU, we developed pragmatic feeding milestone targets based on previous studies,1 Cochrane reviews,11-19 neonatal GI physiology,9,10,20 neonatal gastrointestinal and pharyngo-esophageal motility,9,10,20 and safe feeding strategies.4,21 (4) Upon fulfilling the inclusion and exclusion criteria, each subject in this QI initiative was followed during focused multidisciplinary feeding rounds (see below), and targeted milestones were personalized for each neonate and displayed at the crib side. Feeding providers and the neonatal care team were therefore consistently reminded of these targets. This approach was acceptable to all of the stakeholders and feeding providers offering care in the NICU. The feeding milestones targets were defined as follows: (a) start of trophic feeding as 10–20 mL/kg/d within 3 days of admission and continued for a duration of 3–14 days overall. The provision of trophic or gut priming feeds and the duration was dependent on the age at referral, sickness or reasons for referral, and level of acuity. This approach was agreed upon because of varied gestational ages and diverse critical clinical states on admission. (b) Progression to enteral feeds–120, defined as enteral gavage feeding of 120 mL/kg/d, by day of life 14–28. (c) First oral feeding by 33–34 weeks PMA. (d) Attainment of oral feeds–120, defined as an oral intake of 120 mL/kg/d, by 36–38 weeks PMA. (e) Ad lib oral feeding at discharge. Because our institution is an outborn NICU, we have high referral rates of extremely premature infants with a higher incidence of bronchopulmonary dysplasia (BPD). Since most BPD infants are on feeding volume restriction of 120–130 mL/ kg/d, we chose attainment of oral feeds–120 as the milestone for the purpose of standardization.
Quality Compliance Monitoring Rounds Because of variability in the level of acuity and in referral patterns, and to monitor compliance with the QI initiative, multidisciplinary feeding rounds were conducted twice a week on each infant to (1) provide team education regarding factors that are helping or impeding feeding progress, (2) address the key drivers, (3) monitor compliance, and (4) provide personalized guidance in situations of feeding delays. Physician-led multidisciplinary feeding rounds included neonatology, neonatal nurse practitioners, nurses, nutritionists, lactation consultants, occupational therapists, and parents of individual infants. The approach taken at the multidisciplinary feeding rounds is akin to action medical research.22,23 Infant-focused and providerfocused feeding-related discussions were provided to include (1) appropriate personalization of feeding strategies, (2) educational opportunities for the providers regarding mechanisms of
Journal of Parenteral and Enteral Nutrition XX(X) dysfunctional feeding patterns, and (3) engagement of parents by providing anticipatory guidance.
Methods of Data Evaluation and Continuous Feedback Demographic and morbidity factors as well as feeding milestone data gathered for the baseline group (collected by retrospective chart review) were summarized. Data from the feeding program group were collected prospectively, weekly. Data included demographic and disease characteristics, risk factors, feeding milestones, and hospital LOS. Education of the feeding champions, reiteration of process optimization, and discussion of smart aims and key drivers were reviewed each quarter along with the data from enrolled patients.
Statistical Analysis Data were compared between the baseline and the feeding program groups using analysis of variance for assessing variability within and between groups, unpaired t test to test continuous variables, and χ2 tests to compare categorical variables and proportions. Statistical control charts were used for comparing cumulative LOS between baseline and comparison groups. Data are presented as percentages, mean ± SD, median (interquartile range), or as stated. A P value of 2 BPD, n (%) PDA medical treatment, n (%) PDA surgical treatment, n (%) Duration of ventilation, d n Median (IQR) Duration of CPAP, d n Median (IQR) NEC medical treatment, n (%) NEC surgical treatment, n (%) Deaths, n (%)a
Baseline Program (n = 92)
Feeding Program (n = 92)
92 26.0 (23.0–32.0)
92 27.5 (23.0–32.0)
92 862.5 (691.0–1156.0) 9 (10) 82 (89)
87 965.0 (750.0–1326.0) 8 (9) 85 (92)
26 (28) 4 (4) 53 (57) 22 (24) 30 (33)
23 (25) 0 (0) 53 (57) 31 (34) 16 (17)
55 23.0 (2.0–123.0)
45 15.0 (2.0–106.0)
77 30.0 (1–99) 2 (2) 6 (7) 10 of 102 (10)
61 23.0 (1–95) 3 (3) 4 (4) 3 of 95 (3)
P Value .07 .07 .89 .45 0.62 1.00 .14 .17 .14 .02 .65 .52 .06
BPD, bronchopulmonary dysplasia (defined as oxygen requirement at 36 weeks postmenstrual age); CPAP, continuous positive airway pressure; IVH, intraventricular hemorrhage; NEC, necrotizing enterocolitis; PDA, patent ductus arteriosus; RDS, respiratory distress syndrome; SGA, small for gestational age. a Denominator includes deaths and survivors.
87% at start of enteral feeds, 79% at full enteral, 46% at oral feeding start, and 33% at full oral feeds.
Comparisons of Hospital LOS and Balancing Measures The SIMPLE feeding program was effective in decreasing LOS, as shown in the control chart (Figure 3). Special cause variation effect is observed as there are 11 consecutive points (>8 points24) below the baseline from April 2011. A shift in baseline (April 2011 onward) due to special cause variation is depicted (Figure 3). The DOL at admission to our NICU did not differ between the 2 groups: 9.3 ± 9.9 for the baseline group and 13.1 ± 14.8 for the program group (P = .17). The rate of readmission was 22% (20 of 92) for the baseline group and 11% (10 of 92) for the feeding program group (P = .07). Other balancing measures evaluated were weight gain velocity (24.0 ± 6.1 g/d baseline group vs 27.3 ± 11.4 g/d program group, P < .03), duration of PN (20.2 ± 14 days vs 16.7 ± 12 days, P = .08), duration of peripherally inserted central catheter line (26.2 ± 20.3 days vs 21.8 ± 15 days, P = .32), NEC rates (no difference, as shown in Table 1), and proven sepsis rates (14% vs 7%, P = .28), respectively.
Discussion Potential Benefits and Mechanisms for Change Variability in neonatal feeding strategies is due to both infantdependent factors and provider-dependent factors. Rigid feeding protocols, individual variability, and subjectivity in the interpretation of the symptoms pertinent to neonatal feeding difficulties complicate feeding progression. In this study, we describe our QI efforts to develop a standardized yet practical approach to neonatal feeding among such sick premature infants referred from delivery hospitals. The salient features of our feeding program are (1) simplification and optimization of our neonatal feeding paradigm, which allowed for successful and timely acquisition of feeding milestones; (2) by adopting individualized care, provider-dependent variability was minimized; and (3) compliance monitoring at feeding rounds was necessary to overcome the challenges of attaining feeding milestones. Educational activities were key to achieving our goals and targeted to (1) feeding methods by clarifying when, how, and why to advance feeds to attain targeted milestones; (2) educating providers by discussing and reviewing the science behind neonatal feeding; (3) engaging parents by early
6
Journal of Parenteral and Enteral Nutrition XX(X)
Enteral Feeds -120
Oral Feeds -120
p = 0.01
p = 0.02
100%
100%
90%
90%
80%
80%
70%
70%
60% 50%
81% On- me Improvement
40% 30%
Other
20%
Late Milestone
10%
On- me Milestone
0%
Baseline
Program Group
% Population
% Population
A
60%
72% On-Time Improvement
50% 40% 30%
Other
20%
Gastrostomy
10%
Late Milestone On- me Milestone
0%
Baseline
Program Group
B
Figure 2. Proportion of infants achieving the milestone of enteral feeds–120 and oral feeds–120 is shown (A). Comparison of postmenstrual ages (PMAs) depicting maturation at the attainment of enteral feeds–120, first oral feed, oral feeds–120, and discharge (B).
and consistently providing support, education, and anticipatory guidance on the need for rational feeding strategies; and (4)
developing innovative feeding strategies in a timely manner for those with feeding difficulties.
Jadcherla et al
7
Table 2. Chronological Age (DOL), Maturational (PMA) Age, and Duration to Acquisition of Targeted Feeding Milestones.a Characteristic
Baseline Program (n = 92)
Feeding Program (n = 92)
P Value
3.5 (3.0–5.0) (n = 34) 29.0 (26.0–30.0) (n = 34) 16.0 (5.0–22.0) (n = 66) 17.5 (9.0–23.0) (n = 66) 29.0 (27.0–31.0) (n = 66) 9.0 (6.0–21.0) (n = 79) 28.0 (18.0–39.0) (n = 79) 31.0 (29.0–32.0) (n = 79)
3.0 (2.0–5.0) (n = 72) 27.4 (26.1–30.1) (n = 72) 5.0 (1.5–11.5) (n = 72) 9.5 (5.0–17.0) (n = 72) 29.0 (27.0–31.0) (n = 72) 8.0 (5.0–14.0) (n = 81) 19.0 (11.0–29.0) (n = 81) 30.0 (29.0–32.0) (n = 81)
.27 .26
