Effects of a QI Intervention on Pediatric Asthma Treatment using Patient Outcomes and Workflow in an Emergency Department

Abstract

Objective:

Evaluate a nurse-initiated quality improvement (QI) intervention aimed at enhancing asthma treatment in a pediatric emergency department (ED), utilizing outcomes and workflow.

Methods:

We evaluated the impact of QI interventions for pediatric patients presenting to the ED with asthma with pre-post analysis. A pediatric asthma score (PAS) of >8 indicated moderate to severe asthma. This secondary analysis of the electronic health record (EHR), evaluated on (1) patient outcomes (time to clinical treatment, ED length of stay [EDLOS], admissions and discharges home) (2) clinical workflow.

Results:

We compared 886 visits occurring between 01/01/2015 and 09/27/2015 (pre-implementation period) with 752 visits between 01/01/2016 and 09/27/2016 (post-implementation). Time to first documentation of PAS was decreased post-intervention (p<.001) by 30 minutes (75 ±57 to 39 ±54 min.). There were significant decreases in time to treatment with both steroid and bronchodilator administration (both p<.001). EDLOS did not significantly change. Based on acuity level, those discharged home from the ED with high acuity (PAS score ≥8), had a significant decrease in time to initial PAS, steroid and bronchodilator use and EDLOS. Of those with high acuity who were admitted to the hospital, there was a difference pre- to post-implementation, in time to first PAS (p<.05), but not to treatment. Workflow visualization provided additional insights and detailed (task level) comparisons of the timing of ED activities.

Conclusions:

Nurse-initiated ED interventions, can significantly improve the timeliness of pediatric asthma evaluation and treatment. Examining workflow along with the outcomes, can better inform QI evaluations and clinical management.

1. INTRODUCTION

Systematic interventions can improve the quality and safety of care. However, interventions should be carefully evaluated to understand how and why these interventions improve quality and safety. Such an understanding can provide useful feedback for identifying adaptations to potential interventions¹. Although improving patient and organizational outcomes are usually the main target of health care interventions, a good understanding of process measures (i.e., workflow), can provide insight into how the intervention can be revised for optimal effectiveness².

Workflow is a useful concept in examining delivery of care³. Understanding workflow, allows for identifying quality and patient safety concerns, and informs interventions to prevent adverse events^4-6. With few exceptions², current studies are limited to describing workflow; however, linking workflow to patient outcomes can significantly increase our understanding of how to improve delivery of care. The purpose of this study was to evaluate a bundle of nurse-initiated quality improvement (QI) interventions aimed at asthma treatment in a pediatric emergency department (ED).

Asthma exacerbations are among the most common reasons for hospital utilization by children⁷, with 692,000 ED visits and 79,000 hospitalizations annually in the US⁸. Timely intervention in this patient group is critical. Workflow studies have improved asthma treatment in EDs by; (1) demonstrating the need for flexibility², (2) identifying factors associated with the timing of administration of first bronchodilator⁹, and (3) demonstrating that the earlier sequence placement of corticosteroid administration is associated with improved outcomes¹⁰.

The use of the electronic health record (EHR) system as a data source, has the advantage of examining patient outcomes and clinical workflow, by leveraging large sample sizes and sophisticated quantitative and visual analytics^2,11-15. This study is a preliminary attempt to use both patient outcomes and workflow to better inform QI evaluation. This link is critical to better design systematic interventions to improve safety and quality of care of ED patients, and to target, modify and adapt interventions to local contexts.

2. METHODS

2.1. Study Design

This was an observational study, designed to assess the impact of a bundle of nurse-initiated QI interventions for pediatric patients presenting to the ED with moderate to severe asthma exacerbations. This secondary analysis of EHR data focused on; (1) patient outcomes (time to clinical treatment, ED length of stay [EDLOS], admissions and discharges home) (2) workflow to reveal sequence and patterns of activities (i.e., temporal relationships among patient-care related activities). Consistent with the patient-oriented workflow approach^16,17, each workflow instance represented a patient-care episode.

Analysis of this limited dataset was approved as exempt by Colorado Multiple Institution Review Board (COMIRB).

2.2. Setting

The study took place at an academic, tertiary care pediatric ED in the US mountain region, with over 75,000 annual visits. The ED uses an asthma care pathway that is regularly updated to reflect the latest recommendations of the National Asthma Education and Prevention Program.

2.3. Sample

We included all ED visits from pre-implementation (01/01/2015 to 09/27/2015) to post-implementation (01/01/2016 to 09/27/2016) for patients aged 5-17 years who met the following criteria:

1. visits with an Emergency Severity Index-ESI of 2-4. The ESI level is a triage system, with the level assigned by the nurse, indicating urgency in being seen in the ED with “1” requiring immediate attention (extremely urgent) and “5”, attention in 30 minutes or more (least urgent).

2. ICD-9 or ICD-10 code associated with the visit if their primary code for the visit was for asthma (ICD-9 code 493.XX, ICD-10 code J45) or other diagnostic code for the visit was asthma

   AND

3. the patient was administered a bronchodilator medication during the ED visit and had a primary diagnostic code for an acute upper or lower respiratory tract illness.

2.4. Description of the QI Intervention

We implemented in 2015, a bundle of nurse-initiated interventions aimed at decreasing time to treatment in patients with moderate to severe asthma exacerbation (Pediatric Asthma Severity Score-PAS ≥8). Our primary aim was to decrease the time to administer oral steroids after arrival in the ED. The following interventions were implemented:

a). Rapid identification of patients with asthma

We developed a new EHR process to identify those with asthma-related complaints. For each patient seen in the ED, a triage nurse clicked a Yes/No button in the triage navigator to answer the following question: “Does this patient have a history of asthma, wheezing and recurrent albuterol use?”.

b). Using the PAS score to determine asthma severity

If a triage nurse answered yes to the above question, a “pop-up” screen would prompt the triage nurse to enter a PAS score. A PAS score calculator was made available in the EHR and all nurses received a badge attachment with a PAS score table to facilitate ease of reference. PAS posters were also available in triage rooms.

The Pediatric Asthma Severity Score (PAS), indicates severity of an asthma exacerbation, to assist in determining treatment, hospitalization or discharge home. The PAS is based on an algorithm of level of oxygen, respiratory muscle use, ability to speak in full sentences etc. Scores can range from 5-25 with low scores indicating least severity^18,19. In our institution, a PAS of 5-7 relates to mild asthma exacerbation, 8-11 to moderate and 12-15 severe asthma exacerbation. PAS is recorded throughout the ED stay, and guides the medical team to determine next steps in treatment per the local care pathway.

c). Implementation of nursing standing orders

If the PAS score was ≥8, a triage nurse was then prompted to initiate a set of standing orders. Standing orders included a dose of oral steroid (dexamethasone) and bronchodilator treatment. (The local asthma care pathway was updated to recommend dexamethasone as a preferred corticosteroid (steroid) for acute exacerbations.)

The QI team used several strategies to implement this project, including recruiting local champions to conduct extensive provider and nursing team education, and making changes to nursing orientation. Audit and feedback strategies were used to identify ongoing barriers. Process improvement team members reviewed charts of patients meeting study inclusion criteria who had delayed steroid administration (>1 hour from arrival). If rapid steroid administration was not identified (via chart review), the provider(s) and nurse(s) involved in care of that particular patient, were contacted to explore barriers to implementation.

2.5. Data Collection and Curation

The research informatics department at the study setting extracted the data from the institution’s EHR system and created a limited dataset for ED visits. Variables collected included: Encounter ID, Primary Diagnosis by ICD-9 or ICD- 10 codes, Arrival Mode (walk-in or ambulance), Clinic (main ED or Satellite Clinics 1-4), Acuity Level (ESI, 1–5, with 1 representing most severe and 5 least severe), Event Name, Medications, Event Time, Length of stay in the ED (EDLOS), Asthma scores (PAS) with time stamp, and Disposition.

After integrity checks, secondary data cleaning and curation was performed to select patients who: 1) were 5-17 years; 2) had an ED visit from 01/01/2015 to 09/27/2016; 3) had an ICD code as described earlier; 4) presented with an acuity level (ESI) of 2-4; and 5) were administered a bronchodilator during the ED visit.

2.6. Data Analysis

Data analysis consisted of two steps: (1) descriptive/inferential statistics and (2) visualization. For descriptive statistics, we calculated the time from ED arrival to documentation of; PAS scores, administration of asthma rescue medications (e.g., albuterol, levalbuterol, terbutaline, epinephrine) and corticosteroids (oral or IV) during the ED visit and ED disposition (length of stay in the ED). We also looked at these events by those admitted to the hospital or discharged home. We used predominantly independent t tests to determine if differences existed between the mean times pre and post implementation across six groups of patients: 1) an ED disposition of “discharge” and a first PAS≥8; 2) an ED disposition of “discharge” with a first PAS <8; 3) an ED disposition of “Discharge” with no PAS recorded; 4) an ED disposition of “admit” whose first PAS≥8; 5) an ED disposition of “Admit” whose first PAS <8; and 6) an ED disposition of “admit” with no PAS recorded during the ED stay. We visualized clinical workflow using EventFlow, which combines methods to search interactively, whereby temporal data-patterns, and aggregated data-pattern summaries are provided^20,21.

Common events associated with an asthma exacerbation were identified, involving discrete EHR data (with few missing values), and were relevant to the assessment of care delivery and outcomes. Visualization included the following events with colors representing different events in Figures 1, 2a, 2b 3a and 3b:

Figure 1.

Example of Eventflow diagram displaying three events.

Figure 2.

Pre- (upper figure) and post- (lower figure) intervention workflow for discharged patients whose first PAS≥8

Figure 3.

Pre- (upper figure) and post- (lower figure) intervention workflow for admitted patients whose first PAS < 8.

1. Green: ED arrival

2. Blue: Asthma rescue medication administration

3. Yellow: Corticosteroid administration

4. Red: Asthma assessment that yields a PAS

5. Brown: Patient disposition- (admission or discharge).

Figure 1 is an illustration of a typical EventFlow for three events: ED arrival in green, PAS assessment in red and corticosteroid administration in yellow. The x axis represents time, in this and all figures, the total time represented is 17 hours noted in the right upper-hand corner of the figure. All patient encounters started with ED arrival, represented by the green color band. The first 97 cases on the y axis are represented with the contiguous red bar, indicating that assessment of the PAS score was documented. To the right of the red bar, steroid administration (yellow bar) occurred at irregular intervals (noted by sporadic yellow bars). This entry occurred within minutes to an hour or more after assessment and documentation of the PAS. In most cases, corticosteroid administration occurs after a patient has been assessed, however in some cases, steroids are administered without evidence of documentation of an asthma assessment.

Using EventFlow, we developed data visualizations of pre- and post- intervention workflow for patients for two groups: (1) patients presenting to the ED with moderate to severe asthma (PAS ≥8) who were discharged home (Figures 2a and 2b) and (2) patients presenting with less severe asthma (PAS <8) who were admitted to hospital (Figures 3a and 3b). Eventflow aggregates workflow events to highlight patterns related to the sequence of activities.

3. RESULTS

3.1. Description of the sample

We identified 886 unique visits occurring between 01/01/2015 and 09/27/2015 (pre-implementation period) and 752 unique visits between 01/01/2016 and 09/27/2016 (post-implementation period). The mean age of subjects was 9 yrs. ±3 with the majority (>60%) males. There was no difference in race or ethnicity between periods (Table 1).

Table 1.

Description of the patient sample

	Pre- implementation N=886	Post- implementation N=752	p
Mean Age (SD)	9.0 (3.4)	9.2 (3.4)	.559
Gender (Female N, %)	333 (37.6)	276 (36.7)	.720
Race (N, %)
White	314 (35.4)	308 (41.0)
Black	288 (32.5)	221 (29.4)	.104
Other	267 (30.1)	214 (28.5)
Unknown	17 (1.9)	9 (1.2)
Ethnicity (Hispanic, N, %)	435 (49.1)	383 (50.9)	.488
ESI scores (N, %)
ESI-2	153 (17.3)	159 (21.1)	.290
ESI-3	469 (52.9)	407 (54.1)
ESI-4	264 (29.8)	186 (24.7)
Insurance
Government (%)	81.9	83.4	.286
Private (%)	15.1	15.3
Self-Pay (%)	2.9	1.3

3.2. Impact of the intervention on outcomes

Because timeliness of treatment is critical in the ED, we evaluated if a difference existed pre- to post-implementation between: time from arrival to documentation of the first: PAS score (“Time to first PAS”); first steroid administration (“Time to first Steroid”); and administration of first rescue medication (“Time to Bronchodilator”) (Table 2). We also looked at Length of Stay in the ED (EDLOS), and whether discharged home or admitted to the hospital.

Table 2.

Effects of QI intervention on time to documentation of asthma severity, treatment, duration of ED visit and disposition from ED.

Outcome		Pre-Implementation		Post-Implementation		p	Cohen's d
		N	Mean ± SD	N	Mean ± SD
Length of time	Time to First PAS	485	75.10 ± 57.19	519	39.26 ± 54.08	<.001	0.65
	Time to First Bronchodilator	804	98.31 ± 72.10	682	79.65 ± 63.78	<.001	0.27
	Time to First Steroid	670	108.31 ± 74.74	668	81.67 ± 61.22	<.001	0.39
	Length of stay in ED	886	286.96 ± 138.93	752	280.88 ± 148.27	.195	0.22
Disposition	Discharged to Home	725	18.20	626	17.00	.543	NA
	Admitted	161	81.80	128	83.00		NA

Post-implementation, the time to first documentation of asthma acuity (PAS) was decreased by >30 minutes (75.1 ±57 pre to 39.3 ±54 post), p<.001. There was also a significant decrease in time to treatment with both steroids and bronchodilator administration (p<.001). There was no difference between groups based on acuity, in LOS in the ED or whether discharged home or admitted.

We observed a significant decrease post implementation in average length of time (in minutes) to various key activities in asthma treatment: Cohen’s effect size confirms the p value and shows (in bold) if the post interventions were significantly reduced (Table 2). We then assessed differences pre- and post-implementation for discharges home vs hospital admissions. Overall, there was a similar percentage of those admitted to the hospital pre- (18%) to post-implementation (17%), p=ns. We then looked at those discharged home vs admitted, to determine if differences existed in the main outcomes, based on those identified as high and low patient acuity (Table 3).

Table 3.

Effects of the intervention on duration of ED visit, initiation of steroid, &/or bronchodilator use and first PAS assessment.

Discharged to Home								Admitted
		Pre		Post		p	Cohen’s d	Pre		Post		p	Cohen’s d
		n	Mean (± SD)*	n	Mean (±SD)*			n	Mean (±SD)*	n	Mean (±SD)*
Time to first PAS (min)	≥8	148	72 ± 61	205	30 ± 49	<.001	0.76	98	52 ± 42	94	36 ± 50	0.019	0.34
	<8	216	88 ± 58	203	49 ± 50	<.001	0.68	23	70 ± 50	17	45 ± 52	0.13	0.5
	No PAS	0	N/A	0	N/A	N/A	N/A	0	N/A	0	N/A	NA	NA
Time to First Steroid (min)	≥8	143	88 ± 64	209	56 ± 49	<.001	0.58	83	80 ± 53	92	70 ± 62	0.235	0.18
	<8	193	98 ± 58	208	89 ± 54	0.129	0.15	20	141 ±139	17	154 ± 102	0.765	0.1
	No PAS	208	138 ± 85	124	111 ± 62	<.001	0.35	23	124 ± 58	8	74 ± 65	0.052	0.83
Time to First Bronchodilator(min)	≥8	150	72 ± 60	212	55 ± 42	0.002	0.35	99	51 ± 40	97	48 ± 43	0.595	0.08
	<8	227	88 ± 57	222	97 ± 65	0.154	0.14	23	77 ± 45	16	99 ± 42	0.31	0.5
	No PAS	271	138 ± 77	123	111 ± 76	0.002	0.34	34	123 ± 99	12	111 ± 100	0.72	0.12
Length of Stay in the ED (min)	≥8	150	338 ± 125	219	304 ± 129	0.014	0.26	100	373 ± 123	98	392 ± 137	0.31	0.14
	<8	229	279 ± 125	242	248 ± 150	0.013	0.23	23	463 ± 165	17	465 ± 144	0.964	0.01
	No PAS	346	225 ± 118	163	215 ± 117	0.353	0.09	38	365 ± 155	13	323 ± 120	0.38	0.28

^*numbers in minutes

Discharges home based on acuity.

Among those with high acuity (based on initial assessment in the ED), there were significant reductions in time to first assessment, steroid and bronchodilator administration and EDLOS among those discharged home post intervention. Those assessed with high acuity, experienced a reduction in time by over 30 minutes (72 ±61 to 30 ±49) p<.001 and steroid administration by 30 minutes (88 ±64 to 56 ±49) post-implementation, p<.001. Those with high acuity, discharged home, also had significant reductions in time with bronchodilator administration (p= .002) and length of stay-LOS (p= .01). For those with low acuity, a significant reduction was seen in time to first assessment and EDLOS, but not steroid or bronchodilator administration (Table 3).

Hospital admissions based on acuity

Among those admitted with high acuity, there was a significant reduction in time to first assessment (p=.019), but no difference in time to first steroid administration (80 ±53 to 70 ±62 min., p=ns), bronchodilator or EDLOS. For those less acutely ill (PAS <8), there was no difference in time to steroid, bronchodilator administration or EDLOS (Table 3).

Lack of entry of PAS Score on presentation to the ED

PAS scores were not entered for many patients, prior to treatment. The lack of documentation was improved prior to steroid administration and bronchodilator in the post-intervention period. Pre-intervention, 401 cases were lacking documentation (45%of cases), while post-intervention, this number was reduced to 133 (31%) of cases (p<.001). For patients without PAS documentation, steroid and bronchodilator administration times improved for those discharged home post intervention.

Impact on admissions and discharges

In evaluating for differences in discharges home and hospital admissions, there was a significant decrease in patients admitted (n=219, 31%) post- as opposed to pre-implementation (n=150, 40%) p<.05). This occurred among those with high acuity, but not low acuity. (Table 4).

Table 4.

Effects of QI intervention on admission rates based on severity of asthma

PAS group	Pre			Post
	Discharged to Home	Admitted	Admitted Rate in group	Discharged to Home	Admitted	Admitted Rate in group	p*
	n (%)	n (%)	%	N (%)	N (%)	%
≥8	150 (16.9)	100 (11.3)	40.00	219 (29.1)	98 (13.0)	30.9	.025
< 8	229 (25.8)	23 (2.6)	9.1	242 (32.1)	17 (2.3)	6.6	.180
Without a PAS	346 (39.0)	38 (4.3)	9.90	163 (21.7)	13 (1.7)	7.4	.214
Total	725 (81.8)	161 (18.2)	18.2	624 (83.0)	128 (17.0)	17.0	.559

3.4. Workflows

Figure 2a depicts pre- (n=150) and Figure 2b post- (n=219) intervention workflow for discharged patients of high acuity. Each colored band represents an activity while the gray area between bands, shows average delays (i e. no captured activity) between activities. Both workflow visualizations in Figures 2a and 2b start with “Patient arrived in ED,” represented by the green color band. The second activity can be a PAS assessment (red color band), Corticosteroid administration (yellow color band) or bronchodilator administration (blue color band). In Figure 2a, these activities do not occur early, on admission to the ED. The PAS is assessed first in 82 of the 150 cases, however, in cases 83-150 (45% of cases), assessment does not occur until documentation that medication administration has been occurred. In cases when the PAS was documented , documentation occurs on average, 51 minutes after the patient arrived. Whereas post-intervention, this time was reduced to 30 minutes. One can visualize the timeframe with activities, with post-intervention gaps smaller than pre-intervention.

In the lower figure, the second activity is PAS for 196 out of 219 (89.5%) of cases. After PAS, Corticosteroid administration or bronchodilator administration or another PAS can occur. PAS assessment does also quicken steroid administration in this diagram.

Upper and lower diagrams in Figure 2 can be compared using the gray spaces between bars which shows the average duration between the activities. Patients go through first PAS in lower figure (2b) is earlier than any activity patient go through after arrival in upper figure (2a). Moreover, first Corticosteroid administration or bronchodilator administration is earlier in lower figure (2b) compared to upper figure (2a).

Figure 3 illustrates pre- (upper diagram, n=23 on y axis) and post- (lower diagram, n=17, on y axis) intervention workflow for admitted patients with first PAS<8 (i.e., less acute for asthma). For this patient group, some outcomes are worse in post-implementation (Table 3). During pre- implementation, the majority of the cases have PAS assessment after “Patient arrived in ED” almost immediately followed by bronchodilator administration. When PAS assessment did not occur after arrival, first Corticosteroid administration or bronchodilator administration occurred around the same time when PAS assessment occurred after arrival.

In the lower figure (3b), the majority of cases have the PAS assessment as the second activity. However, there are delays in activities. Asthma-related activities following a low PAS score (less acuity), were delayed when they did not during pre-implementation.

4. DISCUSSION

This study evaluated nurse-initiated ED interventions including EHR driven standardized asthma severity assessment and early administration of asthma medications (bronchodilators) and/or steroids based upon nursing standing orders. Nurse driven interventions in the ED have been associated with expedited care and improved outcomes²². In our study, implementing consistent use of asthma severity scoring during nursing assessment was associated with faster delivery of steroids and bronchodilators. Early steroid delivery has been previously shown to decrease EDLOS and admission rates²³.We did not show a difference in EDLOS or admission rates post-intervention. The QI intervention, initiated by nurses were to administer oral steroids. It is likely that the slower release of steroid by the oral, vs systemic route, affected discharge.

While the intervention was to improve workflow for discharged to home patients with moderate to severe asthma exacerbation (PAS≥8), we also evaluated its effect on other patient groups. Among patients discharged to home, EDLOS decreased both in our target acute asthma group (PAS≥8) and low PAS <8 group, but not in the “no PAS” group, even though these patients had decreased time to both asthma medications. While “Time to First Steroid” improved in PAS≥8 and “no PAS” groups only. Only PAS≥8 group experienced decreased admission rate which could be explained by the more expedited steroid delivery patients experienced as a result of the intervention. It is possible that patients outside of our target group partially benefited from the new workflows, resulting in expedited disposition and faster medication delivery for some patients.

In our target group (PAS≥8), “Time to First Steroid” was significantly shorter in the “Discharged Home” group than the “Admitted” group, suggesting that those who received steroids earlier due to the new workflow were more likely to be discharged^23,24. In our study, documentation of asthma severity scores significantly improved following the intervention. Understanding the severity of asthma by this score, should result in more rapid administration of medications. In this intervention, we demonstrated more rapid delivery of steroids and bronchodilators for patients who were discharged home. We were not however able to demonstrate this same earlier delivery, when patients with initial lower acuity or higher acuity were admitted.

This study involved workflow data as well as outcome data. Our analysis shows that not every patient group (low vs high acuity) gets the same benefit from this intervention. Workflow studies can deliver useful insights by providing task level analysis. These insights can be useful for (1) informing and (2) evaluating a QI study or the current care delivery and plans for new QI studies.

An example for informing the design of a QI study, In Figure 2a, the task level analysis provides an initial insight that early PAS leads to early steroid or asthma medication even during pre-implementation. As an example of evaluating, Figure 3b (post-implementation) shows that there is a delay between PAS assessment and treatment. Interestingly, Figure 3a (pre-implementation) shows no delay between PAS assessment and treatment. A follow-up QI study could be performed to further examine this unintended consequence. Additionally, Figures 2a, 2b, 3a and 3b also shows that there is a lot of variability at task level in asthma treatment.

EHR data (e.g., patient chart or audit data) can be a useful data source to evaluate QI interventions ²⁵ because of access to large sample size. In this study large sample size allowed us to create relatively homogenous patient groups and use advanced quantitative methodologies such as visual analytics. Making EHR data more available to QI teams by developing procedures on how to extract data and deidentify them utilizing audit data and patient charts, would facilitate QI efforts. Likewise, utilization of EHR for these types of studies require improvement in data structure for ease of utilization beyond payer use of health records.

Those with no documentation of a PAS score, should be a concern. While the lack of a score does not necessarily mean a score (or assessment of acuity) was not assessed. It may simply be a function of the nurse failing to document the score in order to begin administering urgent medications (i e. bronchodilators, steroids), placing IV’s etc. We were encouraged to see an increase in the number of scores documented, post-intervention.

This study has its limitations. The PAS score utilized in this study has only moderate interrater reliability²⁶ and is one of many scores used to stratify asthma patients by illness severity. Also, the PAS score was established on admission to the ED. This may account for subjects of “lower acuity” requiring hospital admission.

Secondly, we recognize the payor-driven nature of EHR data, which limits the capacity to capture clinical determinations for treatment that might be better captured in notes and through detailed patient chart audit. Moreover, EHR measures are behavior-oriented and do not include conceptual measures such as time to treatment. Lastly, while we note the limitations using a retrospective naturalistic study, we also see this as a “real-life” opportunity for health care improvement.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author, MO, upon reasonable request.