Abstract
OBJECTIVE: We examined the impact of a nurse-driven clinical pathway on length of stay (LOS) for children hospitalized with asthma.
METHODS: We conducted a randomized controlled trial involving children hospitalized with asthma. Nurses of children in the intervention group weaned salbutamol frequency using an asthma scoring tool, whereas physicians weaned salbutamol frequency for the control group patients as per standard care. The primary outcome was LOS in hours. Secondary outcomes included number of salbutamol treatments administered, ICU transfers, unplanned medical visits postdischarge, and stakeholders’ pathway satisfaction. Research staff, investigators, and statisticians were blinded to group assignment, except for research assistants enrolling participants. Qualitative interviews were done to assess acceptability of intervention by physicians, nurses, residents, and patients.
RESULTS: We recruited 113 participants (mean age 4.9 years, 62% boys) between May 2012 and September 2015. Median LOS was 49 hours (21–243 hours) and 47 hours (22–188 hours) (P = .11), for the control and intervention groups, respectively. A post hoc analysis designed to deal with highly skewed LOS data resulted in a relative 18% (95% confidence interval 0.68–0.99) LOS reduction for the intervention group. There was no difference in secondary outcomes. No significant adverse events resulted from the intervention. The 14 participants included in the qualitative component reported a positive experience with the pathway.
CONCLUSIONS: This nurse-driven pathway led to increased efficiency as evidenced by a modest LOS reduction. It allowed for care standardization, improved utilization of nursing resources, and high stakeholder satisfaction.
Asthma is the most common childhood chronic disease.1 In 2004, asthma, when listed as 1 of the first 5 diagnoses on hospitalization records, contributed to 10% and 8% of all Canadian hospital admissions for children 0 to 4 years and 5 to 14 years, respectively.2
The regular administration of salbutamol is a key element of treatment of children with asthma exacerbations. In most hospitals, physicians prescribe salbutamol and determine administration frequency. Children whose salbutamol treatment can be weaned are often considered stable and do not require immediate medical attention. This can lead to delays in reassessments and administration of unneeded salbutamol treatments. Additionally, physicians’ assessments of children with asthma and their decisions to wean salbutamol frequency are not standardized and vary among physicians.
Clinical pathways are an “operational version of a clinical guideline that designates the timing and sequence of desired care.”3 They may reduce resource utilization and care delivery variation and improve patient education and care quality. Several retrospective studies examining inpatient asthma clinical pathways have shown a decrease in length of stay (LOS).4–6 One randomized controlled trial published in 2000 demonstrated an average 13-hour decrease in LOS in the pathway group, with no increase in adverse events.7 Nurses of children in the pathway group identified patients ready to wean their asthma medication. However, nurses still needed to wait for physicians’ orders to advance to a different stage of the pathway, potentially resulting in delays. Therefore, to improve efficiency and standardize patient assessments, we developed a clinical pathway allowing nurses to wean salbutamol for children hospitalized with asthma. Our pathway was adapted from one previously used in the inpatient setting in another Canadian tertiary care institution (unpublished document) and is based on the Pediatric Respiratory Assessment Measure (PRAM) score, an extensively studied scoring system that discriminatively measures asthma severity and treatment response.8,9 Although not validated for inpatient settings, the PRAM score has been used to assess pediatric patients with asthma exacerbations in emergency departments (ED) for 48 hours,10 a time frame approaching that of inpatient hospitalization. We hypothesized that children randomized to the nurse-driven pathway would have a significantly shorter LOS than children receiving standard care.
Methods
Trial Design
The trial was a 2-arm, parallel, randomized controlled study. Qualitative interviews were done to assess acceptability of intervention by stakeholders. Because of the nature of the intervention, study participants and their care team members were not blinded. All research staff, investigators, and statisticians were blinded to group assignment, with the exception of research assistants (RA) enrolling participants. The Children’s Hospital of Eastern Ontario (CHEO) Research Ethics Board approved this study.
Participants
All children admitted to CHEO’s medical inpatient units were screened for eligibility. CHEO, located in Ottawa, Canada, is a stand-alone tertiary care pediatric hospital. During the study period, CHEO admitted 205 children yearly for asthma, representing 7% of total medical admissions. Respiratory therapists are not routinely involved in inpatient asthma care but are available if needed. Children aged 2 to 17 years admitted with acute asthma exacerbations were initially eligible. Eligibility criteria were changed and Research Ethics Board approval was obtained in April 2015 to include children 18 to 24 months with a history of previous wheezing. This change was based on Canadian guidelines stating that children as young as 1 year can be diagnosed with asthma once they have had 2 episodes of asthma-like symptoms.11 Children were ineligible to participate if they were diagnosed with (1) congenital heart disease; (2) chronic lung disease other than asthma, including cystic fibrosis and bronchopulmonary dysplasia; (3) severe neurologic impairment; or (4) other significant comorbid disorders. Patients whose parents could not understand English or French were excluded, as well as patients with social issues (involvement of child protective services due to parental discord, patients living in foster care or in shelters, with inconsistent caregivers, or with no phone access) because phone follow-up was felt to be too complex with these groups. We recruited from May 2012 to September 2015 to reach our target sample size.
Study Protocol
A computer-generated randomization schedule was composed of randomly permuted blocks of 4 or 6. Group assignment was concealed in sequentially numbered envelopes kept in a secure location. An RA regularly reviewed the admission lists to identify children admitted with asthma. The RA completed eligibility screening, obtained informed and written consent, and allocated patients to study groups based on the randomization schedule. Group allocation was performed as soon as possible after admission.
Intervention Group
Children randomized to the intervention group received standard medical care for asthma at CHEO. This includes close monitoring by nurses and physicians, supportive care (oxygen and fluids), and systemic steroids. Physicians determined the initial salbutamol administration frequency and ordered all additional therapies, including intravenous hydration, antibiotics, systemic and inhaled corticosteroids, or other types of nebulized or inhaled therapies. Physicians assessed their patients daily. Nurses weaned salbutamol for children in the intervention group, according to the steps outlined in the pathway (Fig 1). The pathway included detailed information on when to contact a physician in the event of a patient’s acute deterioration.
Control Group
Children randomized to the control group received CHEO’s standard medical care for asthma. Physicians weaned patients’ salbutamol when called by the child’s nurse or when they deemed it necessary.
Measurements
Baseline information collected included child’s age, sex, prematurity history, concurrent diagnosis of bacterial pneumonia or viral illness based on medical record coding of the chart, number of asthma-related ED visits and hospital admissions in the previous 12 months, history of ICU admission or mechanical ventilation for asthma, and home asthma therapy before admission. Household income, parents’ education level, and child’s exposure to tobacco and pets were collected. Information about illness severity included PRAM score on arrival to the ED, oxygen-dependence on admission, administration of steroids and/or magnesium sulfate in the ED, and frequency of salbutamol administration upon admission and study enrollment. Satisfaction questionnaires were requested from all participants before hospital discharge as well as from all residents, physicians, and nurses working on the units.
Information about adverse events was recorded, including physician notification for deteriorating patients, ICU team consultation, hospital readmission, and visits to the ED within 2 weeks of discharge. Adverse events were reviewed by an independent Data Safety Monitoring Board composed of a pediatric respirologist and a statistician at CHEO, as well as a pediatrician at another pediatric tertiary care center. An RA blinded to group allocation phoned participants 2 weeks after discharge to collect information on asthma-related health care utilization after hospital discharge.
Qualitative interviews explored stakeholders’ acceptability of the pathway by eliciting nurses’, residents’, and physicians’ feedback as well as parents’ feedback on the care experience. Semistructured phone interviews were conducted with a subgroup of parents from the intervention group ∼2 weeks after hospital discharge. Qualitative interviews were conducted with involved health care professionals at the end of the study period (semistructured interviews for nurses and physicians, focus group for residents).
Outcomes
The primary outcome, LOS in hours, was calculated from randomization date and time to discharge date and time. Discharge date and time is recorded by the patient nurse when the patient’s family is handed discharge papers and prescriptions, and discharge order has been processed. Secondary outcomes included number of salbutamol treatments each patient received in hospital; number of children transferred to the ICU; number of children seeking medical attention for asthma-related issues in the 2 weeks after hospital discharge; nursing and physician satisfaction with the pathway, and parents’ satisfaction with hospital care collected via satisfaction questionnaires.
Sample Size
A 12-hour difference was deemed a minimally clinically important difference. This was believed to be obtainable based on a previous study,7 which found a 13-hour difference in LOS with an SD of ∼22 hours. A trial sample size of 206 was incorrectly registered on clinicaltrials.gov before beginning the study, due to a calculation error. The correct sample size for the trial was determined, using the same inputs from our protocol. Assuming a type I error of 5% and an SD of 22 hours, a sample size of 108 (54 in each group) provided 80% power to detect a 12-hour difference in LOS.
Quantitative Analysis
The primary analysis followed the intention-to-treat principle, with LOS analyses using a nonparametric, Mann-Whitney U test to compare groups. After the conduct of the trial and during initial data analysis, a new statistical analysis was proposed by a consulting statistician to address the nonnormal distribution of LOS and to facilitate adjustment for time until randomization occurred. A zero-truncated negative binomial regression was used to account for the identified overdispersion in LOS, now treated as count data. The zero-truncated negative binomial regression was also used to compare counts of salbutamol treatments groups. A χ2 test was used to compare the number of children seeking medical attention for asthma-related issues postdischarge. A Mann-Whitney U test was used to compare satisfaction among nurses, physicians, and patients.
Qualitative Analysis
A conventional qualitative content analysis12 was used to analyze the interview and focus group transcripts. The principal investigator and 2 qualitative experts independently read the transcripts multiple times to obtain a sense of the whole and develop their own preliminary coding structures.13 Next, they met to develop consensus on 1 coding structure. The qualitative experts then collaboratively reanalyzed the transcripts by hand, using the agreed upon coding structure and identified exemplars of each overarching theme for reporting purposes.
Results
Participants
During the study period, there were 604 eligible children; 113 were enrolled, with 56 allocated to the nurse-driven pathway and 57 to standard care (Fig 2). Primary outcome data were available for all patients. There was 1 protocol deviation; 1 participant allocated to the intervention group was changed to standard care by the attending physician because she felt more comfortable with a physician-led wean. Table 1 presents study participants baseline demographic and clinical characteristics.
Consolidated Standards of Reporting Trials flow diagram. q, every; prn, as needed.
Demographics and Clinical Characteristics
LOS
Median LOS was 49 hours (21–243 hours) in the control group, and 47 hours (22–188 hours) in the intervention group (P = .11, Mann-Whitney U test). In the post hoc analysis, the zero-truncated negative binomial model resulted in an 18% relative reduction in mean LOS, 0.82 (95% confidence interval [CI] 0.68–0.99). There was no difference in time to randomization between groups. Figure 3 illustrates histograms for LOS.
Histogram of length of stay in control and intervention groups.
Secondary Outcomes
Median number of salbutamol doses was 15 in the control group and 13 in the intervention group (P = .19, Mann-Whitney U test). Results from the zero-truncated negative showed a nonstatistically significant 23% decrease in the intervention group as compared with the control group, 0.77 (95% CI 0.60–1.02). No children were transferred to the ICU in either group for the study duration. In the 2 weeks after hospital discharge, 2 of the 37 patients reached by phone in the control group had unplanned medical visits for asthma-related issues, whereas 2 of 41 in the intervention group did; risk ratio 0.9 (95% CI 0.13–6.09).
There was no difference in paired comparisons of nurses’ satisfaction before and after pathway implementation on the satisfaction questions (n = 15). Independent comparisons of nurses’ satisfaction before (n = 50) and after (n = 10) pathway implementation showed increased satisfaction after pathway implementation with regard to the quality of care nurses felt they provided (P = .025). Independent comparisons of physicians’ and residents’ satisfaction before (n = 29) and after (n = 19) pathway implementation showed increased satisfaction after pathway implementation with regard to teaching parents to deal more effectively with their child’s asthma (P = .04). There was no difference in paired comparisons of residents’ satisfaction before and after pathway implementation on any of the satisfaction questions (n = 6) (Supplemental Figure 4). Families’ satisfaction was very high in both the control (n = 20) and intervention (n = 24) groups with no significant difference between them.
Qualitative Interviews
Three nurses, 2 physicians, and 4 parents participated in the semistructured interviews. Five residents participated in the focus group. Nine themes were illustrated by the participant groups (Table 2).
Qualitative Analysis Themes and Exemplar Quotations
Discussion
The results of our adjusted analysis support a modest reduction in LOS for children hospitalized with asthma and treated with our nurse-driven clinical pathway, compared with children receiving standard care. Our study is novel in that it is the first to use a pathway to standardize care and allow nurses to wean salbutamol without a physician’s assessment. Our findings are consistent with the literature supporting the use of clinical pathways in the treatment of asthma.5,6,14–16 A randomized controlled study published in 2000 showed a 13-hour decrease in LOS in children admitted with asthma who were on a nurse-driven clinical pathway.7 This pathway, however, still required a physician’s order to wean children’s medications, once assessed by nurses.
A review of adverse events during the study period revealed no increased adverse events in either group of patients. In addition, there was no significant difference in the number of children having unplanned medical visits for asthma in the 2 weeks after hospital discharge, suggesting that our nurse-driven pathway is safe. This is not surprising, given that protocols and standardized order sets have been shown to improve patient safety and clinical outcomes through standardization of care.17–19
We noted a trend toward a decrease in the number of salbutamol treatments in children treated under the pathway compared with standard management; this may be partly attributable to the reduced LOS. Although a formal cost-minimization analysis was not performed, it is likely that the decrease in the number of administered salbutamol doses would result in lower hospitalization costs, as demonstrated in previous retrospective studies examining the impact of asthma pathways on medication use.7,15 A decrease in the number of administered salbutamol doses also has the potential to improve patient safety by diminishing the risk of toxicity symptoms (eg, agitation, hypokalemia, cardiac rhythm disturbances) associated with salbutamol therapy. Finally, patients reported high satisfaction with the care received and felt that their needs were met while in the hospital. Satisfaction with the pathway was high for all groups, including health care professionals.
The majority of the qualitative findings were positive. Participants believed that the pathway led to more efficient care, and that nurses were ideally positioned to carry out the tasks outlined by the pathway as they know their patients well and spend significant time with them. Some concerns were raised by participants regarding the educational experience of residents and the importance of continuing to teach residents how to assess asthmatic children. Participants also felt that education sessions would be essential to maintain their competency in the use of the pathway, especially before the onset of each year’s “peak season” for asthma hospitalizations.
Our study has some limitations. Recruitment was more difficult than anticipated, partly related to the exclusion of the 18- to 24-month-old group in the first 3 years of the study. Although we reached our target recalculated sample of 54 children per group, it took 18 months longer than anticipated. This is unlikely to have significantly affected our results because the standard of care for asthma remained unchanged for the entire study duration. However, nursing turnover was higher than anticipated when planning the study, leading to more difficulties in training nurses to have detailed familiarity with the pathway. This could have negatively influenced our findings because untrained nurses would likely be more reluctant to wean patients’ medications without a physician’s order. Also, some patients could not be enrolled for various reasons, including readiness for discharge or having already weaned off their medication to the minimal frequency of administration on the day of screening. Because attending physicians ultimately authorized the research team to approach patients, 34 children were not recruited because the treating physician was concerned they were too unwell. These concerns arose from physician’s individual comfort with the study and not on predefined criteria as we aimed to approach all eligible patients. This may have potentially led to favoring healthier children and leaving out the sickest. However, the PRAM scores on arrival to the ED (8 and 9 for the control and intervention groups, respectively) suggest that our participant group’s illness severity was representative of children typically admitted with asthma exacerbation.9 In addition, despite the randomization nature of the study, there could have been a difference in disease severity between groups not captured by PRAM score on arrival. Twenty patients in the control group and 15 in the intervention group could not be reached at the 2-week follow-up phone call to determine whether medical care for asthma-related reasons was sought; this secondary outcome should therefore be interpreted with caution. Although the PRAM score has not been validated for inpatient use, it has been validated as an objective measure of respiratory status in children with asthma that is sensitive to clinical change.9 Furthermore, our average LOS was about 48 hours, which is similar to the time frame in which the PRAM score has been validated in the ED setting. Several inpatient pediatric asthma scores have been developed6,20; however, their sensitivity to clinical changes has not been validated. Moreover, the PRAM score was in widespread use in our Pediatric ED before this study, and most residents and many nurses were already familiar with it. The increasing use of the PRAM score on inpatient units during the course of the study could have led to contamination of patients in the standard care arm, whose caregivers may have consciously or subconsciously used the PRAM score to objectively evaluate them. This could potentially have lessened the effect size of the intervention. Because of the nature of the intervention, participants and members of the health care team could not be blinded. It is possible that physicians may have been tempted to delay hospital discharge for patients randomized to the nursing pathway because they may have been concerned that treatments were weaned too quickly. This did not transpire from the qualitative interviews, and, if anything, would have contributed to diminish the effect of the intervention. Finally, although saturation was reached, the qualitative exploration was conducted with a small number of participants. It is unknown whether there were any differences between those who participated in the qualitative component and those who did not. It is likely that those who participated where more vocal and interested in the topic area. Hence, they may have expressed different views than those who did not participate.
Conclusions
Our nurse-driven clinical pathway increased efficiency by providing modest evidence for reduced LOS, without compromising patient safety. By using an asthma clinical pathway, we ensured standardization of care, allowing utilization of nursing resources and decreasing physician burden, resulting in high patient, nurse, and physician satisfaction. Moving forward with pathway implementation, it will be important to provide adequate training to all involved nurses and to ensure that residents continue to learn to assess children with asthma because this is an important skill for pediatricians.
Acknowledgments
We thank Ms Kristina Rohde for her input in the design of the qualitative component of the study, Dr Sarah Pilon for her involvement with the day-to-day functioning of the study, and Dr Amy Plint for reviewing and editing the manuscript.
Footnotes
This trial has been registered at www.clinicaltrials.gov (identifier NCT NCT02037841).
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: All phases of this study were funded by the Alternate Funding Plan Innovation Fund, Ontario Academic Health Science Centres.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
References
- Copyright © 2017 by the American Academy of Pediatrics