BACKGROUND AND OBJECTIVES:

Self-harm among adolescents is a common problem, resulting in large numbers of patients admitted for medical stabilization after a suicide attempt. Because of limited mental health resources, these high-risk patients remain in inpatient settings once medically stabilized until psychiatric placement can be arranged. During this time, patients are at risk for safety events, including self-harm and elopement. Using quality improvement (QI) methodology, we aimed to reduce the frequency of significant safety events (SSEs) in this population by targeting modifiable risk factors and standardizing care.

METHODS:

This was a QI study conducted at a medium-sized academic center. Key interventions included the development of the Pediatric Behavioral Health Safety Protocol, standardization of the patient safety search, and implementation of a daily Safety Huddle. Process measures were selected as metrics of use and adherence to the newly developed protocol. The rate of SSEs per 100 patient days was the primary outcome measure.

RESULTS:

There were 224 patients included in our study: 53 in the preimplementation and 171 in the postimplementation groups. Use of the Pediatric Behavioral Health Safety Protocol increased to 91.8% after implementation. The rate of SSEs per 100 patient days decreased from an average of 2.7 events per 100 patient days in the preimplementation period to 0.17 events per 100 patient days in the postimplementation period.

CONCLUSIONS:

The use of QI methodology to improve safety for adolescents admitted after a suicide attempt led to a substantial and sustainable reduction in the rate of SSEs at our institution.

Rates of adolescent mental health disorders have reached epidemic proportions in the United Sates, with ∼20% of adolescents meeting diagnostic criteria for a mental health disorder.1,2  Depression, suicide, and suicidal ideation represent a significant proportion of these diagnoses and predispose adolescents to self-harm.35  Although much of the management of these patients occurs in the outpatient setting, roles for the emergency department (ED) and inpatient environments are increasing.69  In a recent study, nearly 10% of pediatric admissions nationwide had a primary mental illness as the principal diagnosis, with depression being the most common.10  Researchers suggest that 33% to 50% of pediatric patients who require inpatient psychiatric treatment who present to the ED for care are boarded in the ED or on general inpatient units while awaiting psychiatric placement.1113  Additionally, the majority of suicides and acts of violence among hospitalized patients occur during the first week of hospitalization, a time frame relevant to the boarding population.1416  Within designated adult psychiatric inpatient facilities, safety protocols and guidelines have been described in the literature.17  For pediatric medical units, however, studies pertaining to psychiatric patient safety are lacking. Therefore, instituting protocols to keep this growing population safe and developing measures to determine success are imperative.18,19 

After experiencing a cluster of significant safety events (SSEs) (defined as elopement, self-harm, or harm of others) among psychiatric boarders on the pediatric medical floors at our institution, a multidisciplinary, interprofessional team formed to address this problem. The global aim of this work group was to improve the safety of care provided to adolescents admitted after a suicide attempt. The specific aim in this quality improvement (QI) project was to reduce the frequency of SSEs per 100 patient days in the target population by 50% within 1 year of protocol implementation.

From January 2017 to June 2018 we conducted a QI project at our 150-bed tertiary-care pediatric academic medical center located within the Pacific Northwest. We applied our interventions to all patients admitted for medical stabilization after a suicide attempt but analyzed only those who were admitted to 2 specific study units: the PICU and the pediatric acute-care medical unit (PACM). Patients admitted or transferred to the PACM were placed on a general hospitalist service with pediatric psychiatry serving as consultants. We excluded patients who were admitted to the surgical care unit or had pediatric psychiatry team involvement for reasons other than suicide attempt.

During the preintervention period, there was no standardized approach to the care of adolescents admitted after a suicide attempt. These patients were admitted to the PACM or PICU, and a full-time patient safety attendant (PSA) or sitter, with the level of training of a certified nursing assistant, was placed within the patient’s room. There were no formal agreements on many important aspects of care (eg, walks off the medical unit, visitors, access to personal cell phones, or other belongings). Questions regarding these issues were left to the discretion of attending hospitalists or psychiatry consultants, sometimes with conflicting opinions.

In 2016 we experienced a cluster of SSEs defined as elopement, harm to self, or harm to others, with 2 occurring within 1 calendar month. On the basis of root cause analysis (RCA) of these events, we noted a lack of consistency in both the approach to care of this patient population and the communication between team members. We formed a multidisciplinary team that included pediatric hospitalists, pediatric psychiatrists, nursing leadership, floor nurses, and social workers to develop interventions aimed at improving the safety of care provided for adolescents admitted after a suicide attempt. On the basis of results from the RCAs, we identified modifiable risk factors common to all reviewed SSEs and used these to develop 3 key drivers to promote patient safety: standardized approach to care, consistent performance of a safety search, and development of a shared mental model among care team members (Fig 1). Over the following year, using plan-do-study-act methodology, we developed and implemented 3 improvement interventions: the Pediatric Behavioral Health Safety Protocol (PBHSP) with electronic medical record (EMR) order set, a standardized safety search with EMR documentation, and a standardized communication process, or Safety Huddle, between care team members.

FIGURE 1

Key driver diagram for behavioral health safety QI project.

FIGURE 1

Key driver diagram for behavioral health safety QI project.

Close modal

The improvement team created the PBHSP treatment protocol for the care of patients admitted after a suicide attempt (see Supplemental Fig 4). Given the lack of literature to guide the care of this population in the medical inpatient setting, our protocol was based on SSE RCAs, expert opinion, and care team members’ experience.

First, we discovered that virtually none of the patients admitted to the inpatient setting had a behavioral hold order placed or consent for mental health treatment signed and scanned into the medical record. Should a patient elope, this behavioral hold order is necessary for public safety personnel (state-certified police officers employed by the institution to provide campus security) to physically return the patient to the unit. We added this behavioral hold order to the protocol and order set as well as the requirement for having a guardian sign a consent for mental health treatment and scanning this consent into the chart. We standardized the process of documenting a patient’s description in the medical chart and reporting the description to public safety on admission. The handoff tool used by our PSAs and nurses was updated (see Supplemental Fig 5), reinforcing a shared mental model and understanding of individual patient risk factors. Additional nursing responsibilities included performing a full safety search of patient belongings on arrival to the units, documentation of patient belongings, and removal of items that could pose harm to the patient. We also clearly delineated what patients would be allowed to do during their admission: expectation of attending school, not allowing walks off the medical unit or to unsecured courtyards, limiting visitors, and not allowing patients to wear their own shoes and/or clothing or to have access to personal cell phones and tablets because of concerns regarding social media use. We created an EMR order set that included the order for the behavioral health protocol as well as all the elements included therein and links to consent forms and institutional safety search policies.

Through our RCA reviews it came to light that patients and their belongings were not routinely being searched for items that could cause harm. In addition to including the performance of a safety search in the PBHSP, we defined the standard work for performing the search as well as documentation in the EMR. In accordance with our institutional policy, a safety search of all patient belongings and the patients themselves was to be conducted by 2 nurses within 2 hours of arrival to the medical inpatient unit. Information on patient belongings, including objects found and whether items were stored in locked cabinets in patient rooms or sent home with family members, was documented in the EMR. Throughout the rollout phase of the protocol, weekly real-time chart audits and nurse interviews as well as monthly run chart reviews were conducted. We learned from nurse interviews that accessing the correct flowsheet rows was a barrier to completion of documentation. After learning this, EMR improvement requests were made to streamline the safety search documentation process.

It became clear during our case review process that team communication was not optimal. Lack of a shared mental model regarding severity of illness, potential for behavioral escalation, and contingency planning should a patient escalate contributed to inconsistencies and delays in behavioral interventions. To address this issue, we implemented the Safety Huddle: a standardized communication huddle involving the medical and psychiatric care teams as well as social work, nursing, and public safety. In the initial iteration of the PBHSP, Safety Huddles were to occur on an as-needed basis when behavioral or safety concerns arose. During our weekly process audits, nurse and/or provider interviews, and run chart review, we found that the Safety Huddles were not occurring at the appropriate time. On the basis of this learning, in June of 2017, we adjusted the expectation of when Safety Huddles should occur: for all patients within 24 hours of arrival to the PACM and for those with ongoing behavioral concerns. In addition to adding a standardized Safety Huddle for all patients on admission, a standard script and dot phrase was created to guide discussion on behavioral triggers, de-escalation strategies, and plans for intervention in case of behavioral escalation (see Supplemental Fig 6).

Demographic data are presented with counts and/or percentages and grouped by pre- and postimplementation groups in reference to the PBHSP go-live date of January 27, 2017 (Table 1). A χ2 test or independent sample t test was used to assess differences in group characteristics. SPSS version 25 (IBM SPSS Statistics, IBM Corporation, Armonk, NY) was used for analysis.

TABLE 1

Study Population (N = 224)

CharacteristicPreimplementation (n = 53)Postimplementation (n = 171)Pa
Age y, average (SD) 15.1 (1.7) 15.0 (1.5) .72 
Sex, n (%)   .49 
 Female 43 (81) 131 (77)  
 Male 10 (19) 40 (23)  
Admission unit, n (%)   .73 
 PICU 24 (45) 82 (48)  
 Medical unit 29 (55) 89 (52)  
Race and/or ethnicity, n (%)   .03 
 Non-Hispanic white 33 (62) 120 (70)  
 Non-Hispanic African American 1 (2) 5 (3)  
 Hispanic 6 (11) 30 (18)  
 Other 13 (25) 16 (9)  
Payer, n (%)   .67 
 Government 27 (51) 77 (45)  
 Private 25 (47) 92 (54)  
 Other 1 (2) 2 (1)  
LOS, d, mean (SD) 3.1 (2.7) 3.3 (2.6) .62 
Disposition, n (%)   .83 
 Home 33 (62) 114 (67)  
 Psychiatric hospital 18 (34) 52 (30)  
 Other 2 (4) 5 (3)  
CharacteristicPreimplementation (n = 53)Postimplementation (n = 171)Pa
Age y, average (SD) 15.1 (1.7) 15.0 (1.5) .72 
Sex, n (%)   .49 
 Female 43 (81) 131 (77)  
 Male 10 (19) 40 (23)  
Admission unit, n (%)   .73 
 PICU 24 (45) 82 (48)  
 Medical unit 29 (55) 89 (52)  
Race and/or ethnicity, n (%)   .03 
 Non-Hispanic white 33 (62) 120 (70)  
 Non-Hispanic African American 1 (2) 5 (3)  
 Hispanic 6 (11) 30 (18)  
 Other 13 (25) 16 (9)  
Payer, n (%)   .67 
 Government 27 (51) 77 (45)  
 Private 25 (47) 92 (54)  
 Other 1 (2) 2 (1)  
LOS, d, mean (SD) 3.1 (2.7) 3.3 (2.6) .62 
Disposition, n (%)   .83 
 Home 33 (62) 114 (67)  
 Psychiatric hospital 18 (34) 52 (30)  
 Other 2 (4) 5 (3)  

LOS, length of stay.

a

t tests and χ2 tests were used for continuous and categorical variables.

Baseline rate of SSEs per 100 patient days was collected via retrospective chart review from June 2016 to December 2016. An SSE was determined to have occurred if documentation in the medical chart reflected either patient elopement from the medical unit, inflicting self-harm, or harm of others. Monthly chart reviews were conducted during the postimplementation period (January 2017 to June 2018) for tracking process and outcome measures. Review and discussion of run charts was conducted with nursing leadership on a weekly basis and reviewed monthly at interprofessional improvement team meetings. Microsoft Excel was used for data tracking and management. Length of stay and demographics data were pulled directly from the EMR.

The main outcome measure was the rate of SSEs per 100 patient days in the target population. Process measures were selected as metrics of use and adherence to the PBHSP. Process measures included percent of patients with the order for the PBHSP placed on admission, percent of patients with a signed consent for mental health treatment scanned into the medical chart, and percent of patients with a safety search documented within 2 hours of admission. Balancing measures included transfer from the PACM to PICU because of a need for higher level of care for behavioral interventions and whether episodes of behavioral escalation were triggered by protocol restrictions. We compared measures after the implementation of the PBHSP (January 27, 2017 to June 30, 2018) with the same measurements during the baseline period (June 1, 2016 to January 26, 2017). To ensure completeness of data, patients were identified through review of a comprehensive database maintained by the pediatric psychiatry service and screened for inclusion criteria. The number of patients per month ranged from 2 to 16, with a median of 9. There were no additional hospital-wide interventions targeting patients admitted after a suicide attempt or behavioral escalation during the study period.

Chart review data were used to calculate process and outcome measures. Process measures were plotted over time in run charts.20  The outcome measure (rate of SSEs per 100 patient days) was monitored by a statistical process control (SPC) u-chart. Run and SPC charts were developed with Microsoft Excel QI macros.

The Oregon Health and Science University Institutional Review Board approved this study. No interventions involved comparison of therapies, and subjects were not randomly assigned. All charts were accessed by a subset of improvement team members, and no personal health information was shared outside of the organization.

A total of 224 patients were reviewed: 53 from the preimplementation period and 171 postprotocol implementation. All patients included in the study were admitted for suicide attempt. The characteristics of patients during the different time periods are presented in Table 1. Both groups were similar for all variables, with the exception of race (P = .03).

After protocol implementation, 91.9% of patients had the order for the PBHSP placed on admission. A shift (8 consecutive points greater than the mean) first occurred in February of 2017 (Fig 2A). The percentage of patients with a signed, scanned consent for mental health treatment in the EMR increased from 4.7% to 84.9% after go live of the protocol, and a shift first occurred in January of 2017 (Fig 2B). The rate of safety search documentation did not change significantly after the implementation of our improvement interventions (Fig 2C). Our second improvement intervention focusing on EMR improvement for nursing safety search documentation was not fully implemented because of institutional availability of informaticist time.

FIGURE 2

Run charts reflecting process measures. A, Percent of patients with PBHSP ordered at time of admission. B, Percent of patients with completed consent for mental health treatment scanned into the chart. C, Percent of patients with completed safety search documentation.

FIGURE 2

Run charts reflecting process measures. A, Percent of patients with PBHSP ordered at time of admission. B, Percent of patients with completed consent for mental health treatment scanned into the chart. C, Percent of patients with completed safety search documentation.

Close modal

After implementation of the PBHSP, the frequency of SSEs per 100 patient days decreased significantly from an average of 2.7 to 0.17 events per 100 patient days, with narrowing of the control limits (Fig 3). There were no transfers to the PICU because of a need for a higher level of care for behavioral interventions and no recorded behavioral escalations in the study group triggered by protocol restrictions.

FIGURE 3

Number of SSEs per 100 patient days: SPC u-chart reflecting the rate of SSEs per 100 patient days. LCL, lower control limit; UCL, upper control limit.

FIGURE 3

Number of SSEs per 100 patient days: SPC u-chart reflecting the rate of SSEs per 100 patient days. LCL, lower control limit; UCL, upper control limit.

Close modal

The use of QI methodology to improve safety for adolescents admitted after a suicide attempt led to both a substantial and sustainable reduction in the rate of SSEs at our institution. At the conclusion of the study, we had gone more than 1 year between reported SSEs in our target population. Our 3 improvement interventions led to the routine use of the PBHSP and enhanced documentation of consent for mental health treatment. There have been many publications that describe the growing impact of pediatric mental health disorders and its implications for children’s hospitals and health care delivery systems6  as well as the challenges associated with treating this patient population in the acute-care setting.13  One of our study’s strengths is that we are among the first to demonstrate improved patient safety outcomes for this high-risk population through efforts to standardize care.

During the preimplementation period, our approach to the care of adolescents admitted after a suicide attempt was variable. This lack of a standard approach led to significant variability in care and a lack of a shared understanding of risk factors among care team members. Through our RCA of SSEs before the development of our behavioral health protocol, it became clear that our care teams did not adequately anticipate which patients would be at risk for SSEs or behavioral escalation. Standardizing admission procedures and management practices, including conducting safety searches and Safety Huddles, led to a systematic approach to all patients admitted after a suicide attempt regardless of perceived level of risk. As demonstrated in other settings,21,22  we believe standardizing care contributed significantly to our improved safety outcomes.

The process measure that our interventions did not significantly impact was the percent of patients with a completed safety search within 2 hours of admission (Fig 2C). The safety search is a chaotic process, and our efforts to standardize this work only made the system slightly less disordered. One possible cause for this is the participation of the bedside nurse in the Safety Huddle itself (see Supplemental Fig 6). The bedside nurse attends the Safety Huddle and has an important role in bringing up concerns with the care team. Confirmation that a safety search has been completed is also within the script of the Safety Huddle. Although the nurse participating in the Safety Huddle may not be the admitting nurse, the Safety Huddle process helped to increase nursing engagement and call out the importance of the safety search.

Although not directly assessed, we feel our interventions had a positive impact on the people caring for adolescents after a suicide attempt and the systems within which they work. This is supported by a recent Cochrane Review that found that improvements in interprofessional collaboration can result in better patient care.23  After our interventions, there is now less confusion among care team members with regards to what patients are allowed to do, and all team members are empowered to bring forth concerns during Safety Huddles. Nursing leadership has taken an active role in the maintenance of this work, and process improvements on the Safety Huddle and the safety search are the focus of unit-driven QI work. The Safety Huddle concept has spread to the pediatric ED, where the pediatric psychiatry service is now conducting daily Safety Huddles for patients boarding with behavioral and/or mental health complaints.

Although we showed positive results toward improving safety outcomes for adolescents admitted after a suicide attempt, several important limitations should be mentioned. The study was conducted in a single-center, medium-sized academic children’s hospital, thus potentially limiting its generalizability. The race and/or ethnicity characteristics between the pre- and postimplementation groups were statistically significant. We suspect this was due to the small sample size of our preimplementation group and higher percentage of Hispanic patients in the postimplementation group. Because of data acquisition constraints, 6 months of preimplementation data were collected, limiting the accuracy of determining baseline rates for process and outcome measures. Additionally, improvement interventions were variably implemented between the PICU and medical unit because of competing priorities. This may have been a contributing factor to the lack of improvement in safety search documentation noted in our run chart (Fig 2C). Finally, although systems for reporting SSEs exist within our institution, it is unclear if every SSE is reported. Building better tools for monitoring SSEs within this and other areas of patient safety remains a common challenge to hospital systems.

Through the development and implementation of a standardized approach to the care of adolescents hospitalized after a suicide attempt, we demonstrated a decrease in SSEs for this high-risk population. After the study period, unit-based and nurse-driven process improvement work is ongoing. We believe our protocol could be implemented at other hospitals that treat this patient population, resulting in safer, more equitable care. Additional studies that identify risk factors for behavioral escalation and safety events as well as the use of targeted interventions are warranted.

We thank the resident and nursing staff on the inpatient units at Doernbecher Children’s Hospital, 9N RN leadership for their ongoing work and dedication to improvement efforts, Jose Rodriguez for help with data collection, and Kyle Johnson, MD, and Rebecca Marshall, MD, for ongoing support and engagement in the care of patients with psychiatric care needs.

Dr Valazquez-Campbell assisted with data collection, contributed to the initial draft of the manuscript, and reviewed and revised the manuscript; Dr Austin assisted with data analysis and reviewed and revised the manuscript; Dr Noelck designed and conceptualized the study, drafted the initial manuscript, and reviewed and revised the manuscript; and all authors approved the final manuscript as submitted.

FUNDING: No external funding.

1
Schaffer
D
,
Fisher
P
,
Dulcan
MK
, et al
.
The NIMH Diagnostic Interview Schedule for Children Version 2.3 (DISC-2.3): description, acceptability, prevalence rates, and performance in the MECA study. Methods for the epidemiology of child and adolescent mental disorders study
.
J Am Acad Child Adolesc Psychiatry
.
1996
;
35
(
7
):
865
877
2
Perou
R
,
Bitsko
RH
,
Blumberg
SJ
, et al
;
Centers for Disease Control and Prevention (CDC)
.
Mental health surveillance among children–United States, 2005-2011
.
MMWR Suppl
.
2013
;
62
(
2
):
1
35
3
Olfson
M
,
Druss
BG
,
Marcus
SC
.
Trends in mental health care among children and adolescents
.
N Engl J Med
.
2015
;
372
(
21
):
2029
2038
4
Mojtabai
R
,
Olfson
M
,
Han
B
.
National trends in the prevalence and treatment of depression in adolescents and young adults
.
Pediatrics
.
2016
;
138
(
6
):
e20161878
5
Kann
L
,
Kinchen
S
,
Shanklin
SL
, et al
;
Centers for Disease Control and Prevention (CDC)
.
Youth risk behavior surveillance–United States, 2013 [published correction appears in MMWR Morb Wkly Rep. 2014;63(26):576]
.
MMWR Suppl
.
2014
;
63
(
4
):
1
168
6
Plemmons
G
,
Hall
M
,
Doupnik
S
, et al
.
Hospitalization for suicide ideation or attempt: 2008-2015
.
Pediatrics
.
2018
;
141
(
6
):
e20172426
7
Gallagher
KAS
,
Bujoreanu
IS
,
Cheung
P
, et al
.
Psychiatric boarding in the pediatric inpatient medical setting: a retrospective analysis
.
Hosp Pediatr
.
2017
;
7
(
8
):
444
450
8
Zima
BT
,
Rodean
J
,
Hall
M
,
Bardach
NS
,
Coker
TR
,
Berry
JG
.
Psychiatric disorders and trends in resource use in pediatric hospitals
.
Pediatrics
.
2016
;
138
(
5
):
e20160909
9
Alakeson
V
,
Pande
N
,
Ludwig
M
.
A plan to reduce emergency room ‘boarding’ of psychiatric patients
.
Health Aff (Millwood)
.
2010
;
29
(
9
):
1637
1642
10
Bardach
NS
,
Coker
TR
,
Zima
BT
, et al
.
Common and costly hospitalizations for pediatric mental health disorders
.
Pediatrics
.
2014
;
133
(
4
):
602
609
11
Mansbach
JM
,
Wharff
E
,
Austin
SB
,
Ginnis
K
,
Woods
ER
.
Which psychiatric patients board on the medical service?
Pediatrics
.
2003
;
111
(
6, pt 1
).
12
Wharff
EA
,
Ginnis
KB
,
Ross
AM
,
Blood
EA
.
Predictors of psychiatric boarding in the pediatric emergency department: implications for emergency care
.
Pediatr Emerg Care
.
2011
;
27
(
6
):
483
489
13
Claudius
I
,
Donofrio
JJ
,
Lam
CN
,
Santillanes
G
.
Impact of boarding pediatric psychiatric patients on a medical ward
.
Hosp Pediatr
.
2014
;
4
(
3
):
125
132
14
Martelli
C
,
Awad
H
,
Hardy
P
.
In-patients suicide: epidemiology and prevention [in French]
.
Encephale
.
2010
;
36
(suppl 2):
D83
D91
15
Madsen
T
,
Agerbo
E
,
Mortensen
PB
,
Nordentoft
M
.
Predictors of psychiatric inpatient suicide: a national prospective register-based study
.
J Clin Psychiatry
.
2012
;
73
(
2
):
144
151
16
Shain
B
;
Committee on Adolescence
.
Suicide and suicide attempts in adolescents
.
Pediatrics
.
2016
;
138
(
1
):
e20161420
17
De Santis
ML
,
Myrick
H
,
Lamis
DA
,
Pelic
CP
,
Rhue
C
,
York
J
.
Suicide-specific safety in the inpatient psychiatric unit [published correction appears in Issues Ment Health Nurs. 2015;36(4):315]
.
Issues Ment Health Nurs
.
2015
;
36
(
3
):
190
199
18
Goetz
SB
,
Taylor-Trujillo
A
.
A change in culture: violence prevention in an acute behavioral health setting
.
J Am Psychiatr Nurses Assoc
.
2012
;
18
(
2
):
96
103
19
Bardach
NS
,
Burkhart
Q
,
Richardson
LP
, et al
.
Hospital-based quality measures for pediatric mental health care
.
Pediatrics
.
2018
;
141
(
6
):
e20173554
20
Tague
NR
.
The Quality Tool Box
. 2nd ed.
Milwaukee, WI
:
ASQ Quality Press
;
2005
21
Rozich
JD
,
Howard
RJ
,
Justeson
JM
,
Macken
PD
,
Lindsay
ME
,
Resar
RK
.
Standardization as a mechanism to improve safety in health care
.
Jt Comm J Qual Saf
.
2004
;
30
(
1
):
5
14
22
Lavelle
J
,
Schast
A
,
Keren
R
.
Standardizing care processes and improving quality using pathways and continuous quality improvement
.
Curr Treat Options Pediatr
.
2015
;
1
(
4
):
347
358
23
Reeves
S
,
Pelone
F
,
Harrison
R
,
Goldman
J
,
Zwarenstein
M
.
Interprofessional collaboration to improve professional practice and healthcare outcomes
.
Cochrane Database Syst Rev
.
2017
;
6
:
CD000072

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

Supplementary data