Abstract
Objective
The purpose of this study was to review inpatients undergoing tracheostomies at a tertiary care pediatric hospital in a 24-month period and to identify the indications, comorbidities, hospital course, patient complexity, and predischarge planning for tracheostomy care. The goal was to analyze these factors to highlight potential areas for improvement.
Study Design
Case series with chart review.
Setting
Tertiary care pediatric hospital.
Subjects
Ninety-five inpatients at Boston Children’s Hospital requiring a primary or revision tracheostomy during the 24-month period encompassing 2010 to 2011.
Methods
Inpatients undergoing tracheostomy during the study period were identified using 2 different databases: the Boston Children’s Hospital Department of Otolaryngology and Communication Enhancement database and institution-specific information from the Child Health Corporation of America’s Pediatric Health Information System (PHIS). We extracted the specified metrics from the inpatient charts.
Results
Patients undergoing tracheostomy are complex, with an average of 3.4 comorbidities and 13.6 services involved in their care. The tracheostomy was mentioned in 97.9% of physician and 69.5% of nurse discharge notes, and 42.5% of physician discharge notes contained a plan or appointment for follow-up. Of the patients, 33.7% were discharged home (27.3% of the nonanatomic group and 52.4% of the anatomic group). Overall, 8.4% of tracheostomy patients died before discharge.
Conclusion
The complexity of pediatric tracheostomy patients presents challenges and opportunities for optimizing quality of care for these children. Future directions include the introduction and assessment of multidisciplinary tracheostomy care teams, tracheostomy nurse specialists, and tracheostomy care plans in the pediatric setting.
Pediatric tracheostomy is a valuable procedure in children with severe respiratory compromise or airway obstruction. Over the past 4 to 5 decades, the indications for tracheostomy in children have changed significantly.1-6 In the past, acute infections such as diphtheria, croup, and epiglottitis were the leading causes of airway compromise leading to pediatric tracheostomy. These indications have become less common with the introduction of vaccines against Haemophilus influenzae and Corynebacterium diphtheriae and with the increased use of endotracheal intubation. With continued improvements in neonatal and pediatric intensive care units, premature infants and children with comorbidities are surviving longer. Therefore, tracheostomy is now performed most often in children who have upper airway anomalies or need prolonged mechanical ventilation due to respiratory failure caused by underlying conditions such as hypotonia or bronchopulmonary dysplasia.
Pediatric tracheostomy is generally carried out in children with multiple chronic comorbid conditions requiring complex care.7,8 This means that a growing array of providers is involved in the care of children with tracheostomies, including otorhinolaryngologists, pulmonologists, cardiologists, neurologists, primary care physicians, nurses, therapists, and equipment specialists. At the same time, however, parents and health care providers have identified disorganized care, poor communication, and information mismanagement as impediments to optimal care for pediatric tracheostomy patients. 9 Thus, interprovider coordination of inpatient care is an area with great potential for quality improvement. This is especially the case during the transition from inpatient to rehabilitation or home care for these patients, since past research has shown that such transitions, while challenging, are a time when interventions can do much to improve quality and safety.10,11
We performed a retrospective case review of the hospitalizations surrounding all tracheostomies performed at a tertiary care pediatric hospital during 2010 to 2011. We counted the number of services involved in the care of these patients and assessed the indication for tracheostomy, length of stay, and other patient characteristics.
Methods
This study was conducted at a tertiary care pediatric hospital. Approval was obtained from the Children’s Hospital Electronic Research Portal (CHeRP) institutional review board. All tracheostomies performed at the hospital between January 2010 and December 2011 were identified using the Department of Otorhinolaryngology (ORL) and Communication Enhancement’s records and the Child Health Corporation of America’s Pediatric Health Information System (PHIS). Ninety-five tracheostomies were identified for review. Of these, 5 were revision tracheostomies; in 4 of these, the original tracheostomy was performed at an outside hospital or outside the study period, and in the fifth, both the original and the revision tracheostomy were included in the case series for review.
A retrospective chart review was performed of the inpatient hospitalizations during which each of the 95 tracheostomies took place. Data were collected in various categories, including demographics (age, sex, and need for interpreter), source of admission (emergency department, elective, interhospital transfer), length of stay, mechanical ventilation status, and timing of ORL consult, bronchoscopy, and tracheostomy. More detailed information was gathered on patient complexity and the discharge transition, including number of services (both physician-led services [eg, cardiology, neurology] and non–physician-led services [eg, speech and language pathology, equipment specialists]) involved during the hospital stay, number of comorbidities, discharge disposition, and discharge summary contents.
The number of comorbidities carried by a patient was determined by counting the number of systems affected either by a separate disease process or as sequelae of the primary disease process. Fifteen total systems were used in this classification: neurological (nondevelopmental); neurological (developmental); cardiovascular; respiratory (lung/pleural/parenchymal); respiratory (airway); respiratory (other); nonairway head, ears, eyes, nose, and throat (HEENT); musculoskeletal; gastrointestinal; renal or genitourinary; endocrine; constitutional (including intrauterine growth restriction and failure to thrive); psychiatric; hematologic or other infectious; and nonsolid cancer.
Chart reviews were performed by 2 authors (C.L. and S.S.), with frequent communication carried out to maximize consistency of data collection. Interrater reliability was assessed for 8 hospitalizations reviewed by both authors, and the κ or weighted κ statistics for all fields of data collected indicated a high degree of interrater reliability. P values for differences between patient groups were calculated using the Pearson χ2 test for categorical variables and independent 2-sample t test for continuous variables.
Statistical and interrater reliability analysis was carried out using Stata/IC software version 12.1 (StataCorp, College Station, Texas), and results were tabulated using Microsoft Excel (Microsoft Corp, Redmond, Washington).
Results
The interrater reliability for the raters was κ = 0.78 (P < .001; 95% confidence interval [CI], 0.6-0.96), indicating substantial to excellent agreement between the 2 raters. Tables 1 to 6 report all the metrics that were extracted from patient charts for the 95 hospitalizations, and Figure 1 depicts the time course of these patients’ admissions.
Patient Demographics.
Abbreviation: n/a, not applicable.
P < .01.
Source of Admission.
P < .01.
Overall P value across all sources of admission.
Hospital Time Course.
Abbreviations: ICU, intensive care unit; ORL, otorhinolaryngology.
P < .05.
P < .01.
Ventilation.
Abbreviation: n/a, not applicable.
P < .05.
Overall P value across all ventilator statuses at admission.
Overall P value across all ventilator statuses at time of tracheostomy.
Discharge.
Abbreviation: n/a, not applicable.
Percentage calculated with deceased patients omitted.
P < .05.
Overall P value across all discharge dispositions.
Complexity.
P < .05.
P < .01.
Days of admission are marked on the x-axis. Patient categories are marked on the y-axis. The day of tracheostomy insertion is marked as “day 0.” The figure illustrates the number of days between admission and otorhinolaryngology (ORL) consult and also the number of days from tracheostomy insertion to discharge.
All tracheostomies reviewed were open procedures performed by the otorhinolaryngology service. Anatomic indications for tracheostomy, which include vocal cord paralysis, airway stenosis, and other anatomic abnormalities, accounted for only 22% of tracheostomies performed. Nonanatomic indications, which include complications of prematurity and other causes of acute and chronic respiratory failure, were much more common, accounting for 81% of tracheostomies performed (3 patients had both anatomic and nonanatomic indications).
The mean patient age at tracheostomy was 5.2 years (median, 1.0 years); 47.4% of patients were male. Nonanatomic indications were associated with significantly older mean patient age than were anatomic indications (6.2 vs 0.7 years, P = .0074) ( Table 1 ). In total, 59.6% of all patients undergoing tracheostomy were transferred from outside institutions (65.8% for the nonanatomic group vs 42.9% for the anatomic group, P = .053), and 17% were admitted from the emergency department (19.7% nonanatomic vs 4.8% anatomic, P = .111). Of the patients, 23.4% were elective admissions, with this route of admission significantly more common in the anatomic indication group (14.5% nonanatomic vs 52.4% anatomic, P < .001) ( Table 2 ).
The mean length of stay was 87.8 days, with tracheostomy occurring an average of 42.2 days after admission. Length of stay was significantly shorter for the anatomic group than for the nonanatomic group (52.5 vs 97.3 days, P = .0488), as was time from admission to tracheostomy (22.2 vs 48.4 days, P = .0797), although the latter fell just short of statistical significance. There was no significant difference between anatomic and nonanatomic groups in time from ORL consultation to tracheostomy (20.3 vs 19.1 days, P = .8191), time from last bronchoscopy to tracheostomy (6.2 vs 12.7 days, P = .3361), and time from tracheostomy to discharge (30.3 vs 48.9 days, P = .1734) ( Table 3 ). Of note, 74% of patients in the nonanatomic group had their first and only bronchoscopy on the same day as tracheostomy compared with 53% of patients in the anatomic group. In total, 57.5% of the hospital stay in the anatomic group was posttracheostomy vs 50.3% in the nonanatomic group ( Figure 1 ).
At the time of tracheostomy, 72.6% of patients were intubated and ventilated, 16.8% were on noninvasive positive pressure ventilation (NIPPV), and the remaining 10.5% were not intubated or on any form of ventilation. Patients in the anatomic group were more likely to be on no form of ventilation at the time of tracheostomy (23.8% vs 6.5%, P = .018), although there was no difference between anatomic and nonanatomic groups in the proportion of patients intubated (66.7% vs 75.3%, P = .389) or on NIPPV (9.5% vs 18.2%, P = .357) at the time of tracheostomy ( Table 4 ).
Overall, 33.7% of all patients were discharged home and 8.4% of patients died during the hospitalization. Discharge to home was significantly more likely to occur in the anatomic group (52.4% anatomic vs 27.3% nonanatomic, P = .023), and death during the hospitalization was more likely in the nonanatomic group, although this difference did not reach statistical significance (10.4% nonanatomic vs 0% anatomic, P = .127). All mortalities were related to the child’s underlying medical illnesses and as a complication of tracheostomy. Documentation of tracheostomy and preparation for transition of tracheostomy care varied in completeness, with 97.9% of physician discharge notes mentioning the tracheostomy and 69.5% of nurse discharge notes mentioning the tracheostomy. Omitting patients who died during the hospitalization, it was estimated that 42.5% of physician discharge notes contained a plan or appointment for follow-up tracheostomy care. A similar percentage of physician discharge notes mentioned the tracheostomy in the anatomic and nonanatomic groups (100% vs 97.4%, P = .46); this is also true for nursing discharge notes (81.0% vs 64.9%, P = .25). The percentage of physician discharge notes containing a plan or appointment for follow-up tracheostomy care was significantly greater in the anatomic group than in the nonanatomic group (61.9% vs 34.8%, P = .021) ( Table 5 ).
Of the 93 patients whose discharge records contained physician reference to the tracheotomy, we were able to identify the physician’s specialty in 83 cases. The 12 specialties represented, in descending frequency, were critical care (25), neonatology (20), cardiology (9), cardiac surgery (6), hematology/oncology (6), otorhinolaryngology (4), general pediatrics (4), general surgery (3), pulmonary (2), neurosurgery (2), neurology (1), and palliative care (1). There were no statistically significant differences in physician specialty between the anatomic and nonanatomic groups, with the exception that the physician leaving the discharge note was more likely to be an otorhinolaryngologist in the anatomic group (4/18 vs 0/65, P < .001).
Finally, each patient had an average of 3.4 comorbidities and 13.6 services involved in his or her care during the hospital stay. This included 9.3 physician-led services and 4.3 non–physician-led services. Patients with nonanatomic indications had more services involved during their hospital stay (14.5 vs 11.3, P = .0224), including more physician-led services (10.1 vs 7.0, P = .0054), than those with anatomic indications ( Table 6 ).
Discussion
Indications for tracheostomy in children have changed over the past few decades. This study supports the finding that these patients increasingly require tracheostomy for long-term mechanical ventilation or management of multiple underlying conditions; furthermore, these patients are often cared for by large and diverse groups of providers. As pediatric tracheostomy patients and their care teams have become more complex, disorganized care and poor communication have continued to be obstacles to optimal patient outcomes. Improved coordination and care transitions will be key to improving the quality of care for these patients.
The ORL consultation was sought earlier in the hospital admission of the anatomic group (mean 22.2 days into admission vs 48.4 days), although this different fell just short of statistical significance (P = .0797). This may be because these patients’ primary reason for admission is airway related, and a tracheostomy is hoped to expedite clinical improvement and discharge. As patients in the nonanatomic category have multifactorial and potentially modifiable reasons for prolonged intubation, they may be observed longer before deciding on tracheostomy. There was no statistically significant difference between nonanatomic and anatomic patients in the time from ORL consultation to tracheostomy (19.1 vs 20.3 days, P = .8191) or from last bronchoscopy to tracheostomy (12.7 vs 6.2 days, P = .3361). A relatively constant interval between ORL consultation or bronchoscopy and tracheostomy would be anticipated regardless of indication as parents are counseled, the patient’s medical status is optimized, operating room time is booked, and plans are made to share general anesthesia with other surgical services caring for the patient. The overall mean length of stay was quite long, at 87.8 days, with an average of 45.6 days from tracheostomy to discharge; this period may be prolonged to allow time for home equipment to be acquired and family members to be trained. A significantly higher proportion of patients in the anatomic group were discharged home (52.4% vs 27.3%, P = .023), which may reflect the overall more complex patients in the nonanatomic group.
Overall, we found that only 42.5% of physician discharge notes contained a plan or appointment for tracheostomy follow-up, suggesting an area of potential improvement in the care of these patients. Notably, in the 22% of patients with anatomic indications for tracheostomy, a markedly higher percentage of physician discharge notes contained a plan or appointment for tracheostomy follow-up compared with patients with nonanatomic indications (61.9% vs 34.8%, P = .021). This may be because these patients were less complex overall, with fewer services involved in their care compared with patients with nonanatomic indications (11.3 vs 14.5, P = .0224), a shorter mean length of stay (52.5 vs 97.3 days, P = .0488), and a lower likelihood of being in the ICU at the time of tracheostomy (76.2% vs 97.4%, P = .003). It may also be that the anatomic airway obstruction, and therefore the tracheostomy, was more central to these patients’ presentation and care compared with other patients undergoing tracheostomy. However, it is also possible that providers caring for children with anatomic indications for tracheostomy have developed better approaches to or attitudes toward the transitioning of these patients to home or rehabilitation settings, in which case these approaches should be identified and disseminated.
Another way in which care coordination and transitions could be improved for children undergoing tracheostomies is through multidisciplinary teams (MDTs), which can include physicians, nurses, therapists, and equipment specialists, among others. Tracheostomy MDTs, whose role begins perioperatively and whose aim is to improve tracheostomy care and education, have been implemented at various adult institutions and have been shown to help prevent complications, lower care costs, decrease ICU and hospital stay, improve speaking valve use, and shorten time to decannulation.12-20 Speed et al 21 performed a meta-analysis of 7 of these studies and concluded that overall, the evidence was poor, but MDTs did seem to decrease tracheostomy time and improve speaking valve use. While it is not possible to directly infer that MDTs would be successful in the pediatric setting, it does suggest that their value should be investigated.
Tracheostomy nurse specialists have been shown to decrease complications, ICU readmissions, and length of hospital stay. 22 The expertise of tracheostomy nurse specialists is far-reaching, allowing education of general ward nurses and families. This may be a more streamlined and cost-effective option in optimizing patient care.
At present, there are no published reports on tracheostomy care plans in the literature, and so this would be a good future direction moving forward from this study.
Our study has several limitations. First, all cases reviewed were performed at a single tertiary care pediatric hospital. Some findings may be specific to the hospital’s patient care and recordkeeping practices and may be difficult to generalize to other hospitals or care settings. Limiting our analysis to one institution, however, did allow us to collect information not available in large administrative databases. These include the number of services involved in inpatient care and the percentage of discharge notes containing a plan for tracheostomy follow-up care. It also allowed us to make comparisons that are more reliable across indication categories in variables such as length of stay and mortality. Second, the fact that chart reviews were performed by 2 of the authors means that some interrater variation may exist in the data collected. The authors collaborated closely to minimize this variation, and interrater reliability analysis showed a high degree of concordance in the information extracted. Third, because the study institution is a tertiary referral hospital and receives patients from across the country and the world, follow-up was often organized not at the study institution itself but rather closer to the patient’s home after discussion with his or her local otorhinolaryngologist. There was thus no way to reliably determine to what degree mentioning the tracheostomy or a plan for tracheostomy follow-up in the discharge note correlated with actual follow-up care. Finally, the institution studied is an academic, tertiary care pediatric hospital, and our findings may not be readily generalized to nonteaching, nonpediatric, or community settings in which pediatric tracheostomy may also be performed. Past studies have shown, however, that pediatric patients undergoing tracheostomy at children’s hospitals and hospitals with greater tracheostomy case volume are at lower risk of inpatient mortality. 5 We believe this makes institutions such as the one studied ideal sites for the development and dissemination of interventions aimed at improving pediatric tracheostomy care.
Conclusion
The complexity of pediatric tracheostomy patients presents both challenges and opportunities for optimizing their quality of care. Our study describes the large number of providers involved in the care of children receiving tracheostomies at a tertiary care pediatric hospital, as well as the documentation of tracheostomies in the discharge notes. We also identified variation in these parameters among patients with different indications for tracheostomy, indicating both the need and the potential for interventions to improve quality of care for these children. Future directions include multi-institutional analyses of care coordination and transitions for pediatric tracheostomy patients as well as testing the effectiveness of interventions such as multidisciplinary tracheostomy care teams, tracheostomy nurse specialists, and tracheostomy care plans in a pediatric care setting.
Author Contributions
Disclosures
Footnotes
Acknowledgements
The authors thank the entire otorhinolaryngology faculty who participated in the care of these patients, particularly Dr Reza Rahbar, who has created the tracheostomy care protocols our service uses.
No sponsorships or competing interests have been disclosed for this article.