 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ARTICLE |
a Division of Neonatology, Children's Hospital, Boston, Massachusetts
b Division of Neonatology, Connecticut Children's Medical Center, Hartford, Connecticut
c Institute for Clinical Research and Health Policy Studies
d Division of Newborn Medicine, Tufts–New England Medical Center, Boston, Massachusetts
e Department of Neonatology, Beth Israel-Deaconess Medical Center, Boston, Massachusetts
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONAPPENDIX: NICU STAFF NURSE...REFERENCES |
|---|
METHODS. We surveyed nurses in US NICUs with neonatal-perinatal fellowships in 2004. Data collected included pulse oximeter saturation limits targeted by nurses and by NICU policy when present, nurses' opinions about appropriate pulse oximeter saturation limits, and NICU and nurse characteristics. Factors associated with pulse oximeter saturation limits targeted by nurses were identified with hierarchical linear modeling.
RESULTS. Among those eligible, 2805 (45%) nurses in 59 (60%) NICUs responded. Forty (68%) of 59 NICUs had a policy that specified a pulse oximeter saturation target range for extremely preterm infants. Among 1957 nurses at NICUs with policies, 540 (28%) accurately identified the upper and lower limits of their NICU's policy and also targeted these values in practice. NICU-specific SDs for individual nurse target limits were less at NICUs with versus without a policy for both upper and lower limits. Hierarchical linear modeling identified presence of pulse oximeter saturation policy, NICU-specific nurse group opinion, and individual nurse opinion as factors significantly associated with individual pulse oximeter saturation target limits. For each percentage point increase in individual opinion upper limit, the individual target upper limit increased by 0.41 percentage point at NICUs with a policy compared with 0.6 percentage point at NICUs with no policy.
CONCLUSIONS. Presence of policy-specified pulse oximeter saturation limits, nurse group opinion, and individual nurse opinion were independently associated with individual nurse pulse oximeter saturation target limits during routine care of extremely preterm infants. The presence of a policy reduced the influence of individual nurse opinion on targeted pulse oximeter saturation limits and reduced variation among nurse target limits within NICUs.
Key Words: oxygen saturation • prematurity • policy • oxygenation • pulse oximetry • NICU
Abbreviations: EPI—extremely preterm infant • SpO2—pulse oximeter oxygen saturation
Extremely preterm infants (EPIs), born before 28 weeks' gestation, are at greatest risk for oxygen-related morbidities such as retinopathy of prematurity and bronchopulmonary dysplasia. Pulse oximeter oxygen saturation (SpO2) is the standard, noninvasive, continuous method used to estimate arterial oxygen saturation in neonates and to guide their oxygen therapy.1 Cohort studies suggested that lower rather than higher SpO2 ranges may reduce severe retinopathy of prematurity and bronchopulmonary dysplasia without increasing adverse outcomes2–6; however, there is no conclusive evidence regarding which targeted SpO2 range will maximize beneficial effects for EPIs while minimizing toxicity.7–9 As a result, NICU policies regarding oxygen management in EPIs include a spectrum of SpO2 target range limits.10,11
In the present climate of uncertainty, personal opinions of NICU nursing staff regarding appropriate oxygen management and other NICU and nurse factors may influence the SpO2 limits targeted and achieved in EPIs during daily care. These factors, in turn, could influence outcomes of EPIs or results of clinical trials in which oxygen therapy is an intervention or outcome. Factors that influence the SpO2 ranges targeted by individual nurses when managing oxygen saturation in EPIs have not been determined. The objectives of this study were (1) to compare SpO2 limits targeted by individual nurses during routine care of EPIs in the first 4 weeks of life with nurse opinions regarding appropriate SpO2 limits and with SpO2 limits specified by their NICU's SpO2 policy and (2) to examine the influence of individual nurse opinion and other factors on SpO2 limits targeted by individual nurses.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONAPPENDIX: NICU STAFF NURSE...REFERENCES |
|---|
We used e-mail and telephone to contact nurse managers at eligible NICUs to invite their NICU's participation. Nurse managers of participating NICUs advertised the survey among their staff nurses. To maximize NICU participation, we made multiple attempts to contact nurse managers who failed to respond to the initial invitation.
Data Collection
At each NICU, the survey was conducted during a 2-week period. Each nurse answered questions regarding his or her individual characteristics, the presence of a written SpO2 policy in the NICU, and NICU policy-specified SpO2 target limits for EPIs. Each nurse also responded to questions regarding the upper and lower SpO2 limits that, in their opinion, were most appropriate for EPIs and to questions regarding the upper and lower SpO2 limits that they attempted to target and maintain in EPIs in their personal practice (see Appendix for full survey).
Nurse managers or their designees separately submitted confidential responses to questions regarding nursing staff and NICU characteristics, including presence of a NICU SpO2 policy for EPIs and, when present, the target SpO2 limits of their policy. We considered nurse manager responses as the true standard for questions regarding nursing staff and NICU characteristics. We contacted nurse managers several times during the 2-week survey period to maximize compliance and staff participation. For each submitted survey, all answers were registered in real time in a central database that was developed and maintained by informatics personnel of the General Clinical Research Center and Institute for Clinical Research and Health Policy Studies at Tufts-New England Medical Center.
Data Analysis
We summarized NICU and nurse characteristics submitted by nurse managers and survey responses submitted by each nurse. For categorical questions, all nonblank responses were considered valid. Numbers ranging from 70 to 100 were considered valid responses regarding values for SpO2 limits. For each question, all valid responses were included in analyses. Categorical variables were described with frequencies, and continuous variables were described with mean and SD. The Student's t test or Mann-Whitney U test was used to compare characteristics of NICUs with and without a policy-specified SpO2 target range.
Hierarchical Linear Modeling
Because nurse opinions and SpO2 limits targeted by nurses were likely to correlate with those of other nurses at the same NICU, we performed comparative and regression analyses by using 2-level hierarchical linear models.12 Modeling was accomplished with HLM 6.0 (Scientific Software International, Inc, Lincolnwood, IL). Two-level models were created to compare paired nurse responses, such as individual opinion limits compared with individual target limits, to test the hypothesis that the difference between paired responses was 0 while adjusting for clustering.
For identification of factors that were associated with individual nurse SpO2 target limits in multivariate analysis, individual target limits were modeled in several steps as a function of individual nurse and NICU characteristics. Separate models were created to examine factors that were associated with upper and lower limits of individual nurse target ranges. Initially, unelaborated random-effects models were created in which individual target range upper or lower limit was described as a single mean value with random components attributable to variation between NICUs and between nurses within NICUs. Independent variables were introduced subsequently. For each NICU, data regarding individual opinions were centered on the NICU mean so that model intercepts would yield clinically meaningful values.
At the nurse level (level 1), individual target limits were modeled as a function of nurse characteristics. Nurse-level characteristics tested included primary shift worked, years of NICU experience, hours worked per week, proportion of work time caring for EPIs, and individual opinion regarding appropriate upper and lower oxygen saturation limits for EPIs.
In the second level of modeling, the level 1 model intercept (representing the mean value for individual nurse target limit) was treated as an outcome and modeled as a function of NICU characteristics. NICU-level characteristics tested included presence or absence of a policy regarding SpO2 levels for EPIs, policy-defined upper and lower SpO2 limits when a policy was in place, number of NICU beds, and EPI admissions per year. The mean opinion of nursing staff regarding most appropriate upper and lower SpO2 limits was calculated for each NICU and was tested as a NICU-level characteristic representing nursing staff group opinion. The presence versus absence of a NICU policy was tested as a NICU-level determinant of each NICU's intercept for individual nurse target limits. Presence versus absence of a policy was also tested as a NICU-level determinant of the regression coefficient for individual opinion to test the effect of the presence of a policy on the relationship between individual opinion and individual target limits. Random effects were tested for each significant characteristic identified to assess whether its effect on individual target range limits was constant or varied among NICUs. The resulting 2-level models included nurse- and NICU-level estimates of residual variance. The percentage of variation explained by the final models was obtained by comparing the residual variance for each full 2-level model with the variance of the corresponding unelaborated random-effects model described already.12
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONAPPENDIX: NICU STAFF NURSE...REFERENCES |
|---|
Of 6251 eligible nurses at participating NICUs, 2805 (45%) submitted surveys. Center-specific nurse response rates ranged from 5% to 100% (median: 42%; mean: 48%). Eighty-eight percent of responding nurses submitted complete surveys with valid responses to all questions. Response rates to individual questions ranged from 91% to 100%. More responding nurses worked day than evening/night shift (49% vs 39%), and 72% worked 
36 hours per week. NICU experience among participants averaged 11 ± 9 years (range: 0–40 years).
Table 1 compares NICUs with and without SpO2 policies with respect to nurse opinions and individual target limits. At NICUs without a policy, SpO2 values for individual opinions and individual target range limits were higher than at NICUs with a policy. NICU-specific SDs for individual nurse target limits were less at NICUs with a policy compared with NICUs with no policy for both upper (P = .01, Mann-Whitney U) and lower (P = .004) limits, indicating less variation among individual nurse target range limits at NICUs with policies. At NICUs with policies, individual target range upper limits exceeded the NICU policy upper SpO2 limit by an average of 0.9 percentage point (P = .002, adjusted for clustering).
|
2 percentage points above or below the policy value. The policy-specified lower limit was given as the individual lower limit by 945 (48%), and 917 (47%) gave an individual lower limit 2 percentage points above or below the policy value.
|
2 percentage points for the majority of nurses who were unable to identify policy-specified target limits or were unaware that a policy was present. Nurses at NICUs with policies reported reasons that they thought that EPIs on supplemental oxygen were not maintained within the range of their NICU's SpO2 policy. Among the 1258 nurses who knew that their NICU had a SpO2 policy, 888 (71%) cited infant lability with frequent desaturations, 13% thought that their NICU's target range was too narrow, 6% believed that the target range was too high or too low, and 7% cited other reasons.
Hierarchical Linear Modeling
NICU SpO2 policy presence, NICU-specific nurse group opinion, and individual nurse opinion were factors identified by hierarchical linear modeling to have significant and independent effects on both upper (Table 3) and lower (Table 4) limits targeted by individual nurses. As shown in Table 3, presence of a NICU policy was associated with a reduced mean individual target upper limit of 0.81 percentage points, after adjustment for other factors. Presence of a SpO2 policy was also associated with a reduced individual opinion-target slope, reflecting reduced influence of individual opinion on individual target upper limit (Fig 1). At NICUs with a policy, each percentage point increase in individual opinion upper limit was associated with an increase in individual target upper limit of 0.41 percentage point after adjustment for other factors, in contrast to 0.6 percentage point at NICUs with no policy (P < .001 for difference between policy and no policy). NICUs varied significantly in the strength of the relationship between individual opinion and target upper limit, even after accounting for the presence of a policy and group opinion.
|
|
|
Shift worked, hours worked per week, percentage of time caring for EPIs, number of nurses on staff, EPIs admitted annually, and number of NICU beds were not significantly associated with nurse target range upper or lower limits after adjustment for the factors in these models. Years of NICU experience was significantly associated with individual lower limit (Table 4), but this effect was clinically inconsequential.
Explained Variance
The initial unelaborated random-effects models demonstrated that approximately two thirds of variability in individual target limits was attributable to variability in target limits among nurses within NICUs (Table 5). This was especially true at NICUs with no policy, where variability among nurses within NICUs accounted for 75% of variability in individual target upper limits and 99% of variability in individual lower limits. The full 2-level models explained 52% of overall observed variability in individual target upper limits for all NICUs and 66% of observed variability in individual lower limits. Despite the many nurse-level variables tested, most variation in SpO2 target range limits among nurses within NICUs remained unexplained.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONAPPENDIX: NICU STAFF NURSE...REFERENCES |
|---|
At NICUs with a SpO2 policy, fewer than one third of nurses identified their unit's policy-specified SpO2 target limits and also gave the policy's limits as their individual target limits. Nearly half reported individual target limits that were 2 percentage points different from their NICU's policy. Differences between individual target limits and policy were attributable to inaccurate knowledge of policy presence or policy-specified values, choice of target limits different from known policy-specified values, or a combination of incorrect knowledge and choice.
This study further demonstrates that the opinions of individual nurses and the group opinion of nurses at each NICU are strongly associated with SpO2 limits targeted by individual nurses. At NICUs with a policy, the policy-specified limits had a significant, independent association with individual SpO2 target limits; however, the effects of individual opinion and group opinion remained significant after accounting for the effect of the policy-specified target limits. The influence of individual opinion on individual target limits varied significantly among NICUs regardless of whether a policy was present; however, at NICUs with a policy, the variation among individual target limits was significantly less, the influence of individual opinion on individual target limits was significantly less, and individual target upper limits were significantly lower than at NICUs with no policy.
The variety of policy-specified target limits seen in this study is consistent with 2 earlier studies.6,10 A 1992 survey of nurse managers from US NICUs with fellowship programs found that 15% of NICU SpO2 policies accepted an upper SpO2 higher than published recommendations of SpO2 of 98%, and 13% accepted SpO2 of 100% for infants born at <1500 g.10 A 2001 survey of 142 US level III NICUs reported that 87% had SpO2 guidelines.6 The overall range of policy-specified SpO2 target limits in the 2001 survey was 82% to 100%, with average target lower limit SpO2 of 89% and average upper limit of 95%. In our survey, the mean ± SD of SpO2 lower limits was 86 ± 3% and 94 ± 2% for upper limits. Although a direct comparison of our results with the 2001 survey does not account for differences between the studies, it is interesting to speculate whether SpO2 limits may be trending downward or that NICUs with training programs may maintain lower SpO2 values than level III NICUs as a whole.
Hierarchical linear modeling of these data was necessary because nurse opinions and individual target limits were clustered within NICUs. In addition to adjusting for the clustering of nurse responses, hierarchical linear modeling allowed evaluation of community influences on individual responses. Thus, the presence of policy-specified target limits in the NICU "community" was demonstrated to reduce but not eliminate the effect of individual opinion on individual target limits.
Nurse group opinions, individual nurse opinions, and individual target limits that exceed their NICU's policy-specified target upper limit may contribute to higher SpO2 values during routine care of EPIs. Collectively, these findings may partly explain variable compliance with policy-specified SpO2 target limits reported in recent studies. The Achieved Versus Intended Pulse Oximeter Saturation in Infants Born Less Than 28 Weeks' Gestation (AVIOx) study reported that successful maintenance of intended SpO2 range varied substantially among centers, among infants within centers, and for individual infants over time.11 Others documented varying compliance with policy-specified SpO2 target limits.13 It is plausible that differences in individual nurse target limits and the factors that influence individual target limits contribute to such variation.
It is noteworthy that one third of level III NICUs that participated in this 2004 survey had no SpO2 policy for EPIs. At these NICUs, higher SpO2 values were reported for individual nurse opinions and target limits than those reported by nurses at NICUs with SpO2 policies. At NICUs with policies, individual nurse target limits exceeded the upper SpO2 limits of NICU policies, suggesting that studies that accept policy-specified target limits as a proxy may underestimate oxygen exposure.
In this analysis, we could not account for much of the variation among nurse target limits within NICUs. It is possible that factors such as time since SpO2 policy implementation, variable nurse input into policy development, variation in educational initiatives associated with policy implementation, or variable emphasis on maintenance of SpO2 targets between NICUs may account for some of this variability. Such unexplored factors are among many aspects of oxygen delivery that may contribute to risk for departures from policy-specified SpO2 targets and exposure to hyperoxemic oxygen levels in EPIs. Recent studies suggested that variable compliance with oximeter alarm settings or discretionary supplemental oxygen administration during procedures or transport also may contribute to departures from oxygenation goals.14,15 Additional research is needed regarding the causes of variation among nurse SpO2 target limits and other factors involved in achieving oxygen saturation goals for EPIs.
Regardless of the SpO2 ranges determined by ongoing trials to maximize beneficial effects and minimize toxicity in EPIs, the findings of this survey will be relevant in efforts to optimize supplemental oxygen therapy during routine care of EPIs. Until such time that automatic oxygen control systems are available to assist neonatal oxygen delivery, our options to improve SpO2 target range compliance are to exercise control over factors that are amenable to change.16,17 In efforts to improve quality of care, this may include establishing a NICU SpO2 policy for EPIs at units where one does not currently exist, addressing oxygen delivery during all aspects of neonatal care. In addition, nursing educational efforts should be ongoing and include evidence-based education and training in an effort to stress current unit target SpO2 values. We hope that the results of this study will increase awareness of all neonatal care providers regarding factors that affect targeted SpO2 limits and guide nurse education, NICU quality improvement initiatives, and implementation of changes in NICU clinical practice.
|
APPENDIX: NICU STAFF NURSE SURVEY |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONAPPENDIX: NICU STAFF NURSE...REFERENCES |
|---|
There are no scientifically proven correct answers to the following questions. In order to gain the most insight on people's beliefs and practices, we need your honest answers. Please complete this survey individually and without discussion or collaboration with others.
By completing and submitting the survey, you are consenting to participate in the survey.
Please answer the following questions regarding infants born at <28 weeks gestation who are receiving oxygen therapy.
The following questions will pertain to your personal nursing practice.
The following questions will pertain to your NICU's policy and your personal opinion regarding oxygen saturation ranges.
|
ACKNOWLEDGMENTS |
|---|
We gratefully acknowledge the NICU nurses and nurse managers who participated in this study.
|
FOOTNOTES |
|---|
Address correspondence to Tuyet-Hang Nghiem, MD, Division of Newborn Medicine, Children's Hospital, Boston, 300 Longwood Ave, Boston, MA 02115. E-mail: hang.nghiem-rao@tch.harvard.edu
The authors have indicated they have no financial relationships relevant to this article to disclose.
| What's Known on This Subject Extremely preterm infants are at greatest risk for retinopathy of prematurity and bronchopulmonary dysplasia. Studies suggest that lower rather than higher oxygen saturation ranges may reduce these oxygen-related morbidities. NICU policies accept varying oxygen saturation target range limits.
|
| What This Study Adds This study documents clinically important variability among nurses with respect to the oxygen saturation ranges that they target during routine care and demonstrates that nurse opinion, nurse group opinion, and NICU policy are independently associated with individual nurse target saturation ranges.
|
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONAPPENDIX: NICU STAFF NURSE...REFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2008 American Academy of Pediatrics. All rights reserved. |
||
|
||
