Article Text
PDF
Abstract
Background Lack of consistent and standardised handoffs is a leading cause of patient harm. With increased census in our hospital medicine (HM) service, failure to handoff using a standardised method has the potential to cause significant patient harm. We used a quality improvement methodology to standardise an existing and validated handoff tool within our HM team to improve handoff communication among providers and improve patient safety.
Methods A quality improvement team was charged with studying handoff communication among HM teams and between day and night shift providers at a tertiary oncology hospital. Multiple plan-do-study-act cycles were conducted, and process flow maps, root cause analysis and an affinity diagram were developed based on feedback from the HM team. The quality improvement team developed a plan to implement I-PASS (Illness severity, Patient summary, Action list, Situation awareness and contingency plan, and Synthesis by receiver) as the standardised handoff tool to be used among the providers in HM at the end of shift and for handoff to the nocturnal covering service. Rates of I-PASS use were collected before and after several educational interventions to encourage use of I-PASS and were displayed in a control chart. After the I-PASS interventions, HM providers were surveyed twice to evaluate the secondary outcomes: the tool’s impact on workflow, perceptions of patient safety, ease of use and satisfaction with I-PASS. Survey results were compared using Fisher exact tests.
Results The HM team’s rate of use of I-PASS handoffs increased from 23% to 72%, an improvement of 68%. By the end of the quality improvement project, I-PASS use had increased to 90%. No significant differences were detected in the reported duration of handoffs after I-PASS implementation (on average <5 min per patient, p=0.205). Provider perceptions of handoff quality, efficiency, communication errors and the I-PASS tool’s effectiveness were satisfactory.
Conclusion We used a quality improvement methodology to encourage the HM team’s adoption of a validated handoff tool. Adherence to the standardised handoff tool significantly improved workflows and facilitated communication between the day and night shift teams.
- Patient Handoff
- Patient safety
- Quality improvement methodologies
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Healthcare organisations need help standardising handoffs to comply with accrediting bodies’ regulations. I-PASS (Illness severity, Patient summary, Action list, Situation awareness and contingency plan, and Synthesis by receiver) is a validated methodology that reduces medical errors and patient harm.
WHAT THIS STUDY ADDS
Developing easily accessible, diverse training materials, mandatory training with individualised feedback, leveraging the electronic health record and using performance incentives resulted in unprecedented compliance rates in our department.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Implementing these strategies at other institutions can result in high compliance rates and provider satisfaction.
Introduction
Problem description
Communication failures are among the most common root causes of medical errors that harm patients.1 Handoffs occur daily, each time patient care responsibility is transferred from one provider to another (eg, shift changes or transporting a patient from one location to another in the hospital). Thus, lack of standardised handoff methods between providers increases the potential for communication failure-related harm.1 Indeed, many studies investigating the causes of medical errors have found ineffective handoff communication to be a primary contributing factor.2 Thus, the Accreditation Council for Graduate Medical Education requires sponsoring institutions to have a structured handoff process for all residents to reduce the risk of communication-related errors.3 In fiscal year (FY) 2017, communication/coordination was the number one root cause of safety events in the human factor category in our institution. Additionally, multiple safety root cause analyses conducted had handoff as one of the contributing factors. Lastly, in the 2018 Culture of Safety Survey, our department’s handoff favourability score was 27%.
Available knowledge
The hospital medicine (HM) team consists of advanced practice providers (APPs), residents or fellows, and attending physicians. Their handoff training varies, as do their handoff practices. In many cases, pertinent information is not always communicated to receiving providers, often leading to delays in patient care.4 Furthermore, considering the high patient censuses and time constraints, hospitalists look for ways to decrease the time spent on handoffs while ensuring that communication is accessible for everyone at any time.5
Rationale
The Society of Hospital Medicine Healthcare Quality & Patient Safety recommends the use of a structured handoff tool for handoff communication without being specific about which tool to use.4 Our team selected I-PASS (Illness severity, Patient summary, Action list, Situation awareness and contingency plan, and Synthesis by receiver) as the standardised handoff tool to improve communication and patient care. According to a study of I-PASS, implementing this standardised handoff tool decreased medical errors and preventable adverse events with no change in the time required to handoff information.6
Specific aim
We aimed to standardise handoff communication among the HM team and with the nocturnal covering service by increasing I-PASS utilisation from a baseline of 23% to 70% by August 2019. Handoff duration and provider satisfaction were secondary measures to assess handoff efficiency and quality. Our sustainability aim was to increase adherence to 90%.
Methods
Context
This quality improvement (QI) study was conducted from November 2018 to October 2021 in the HM service in a tertiary oncology-focused teaching hospital in a large metropolitan medical centre. The QI initiative encompassed four plan-do-study-act (PDSA) cycles. The HM team is responsible for an average of 130 patients daily, spanning patients with solid tumours, those in remission and those with no active cancer diagnosis. The HM service has eight teams, each consisting of an APP or three to four residents and a pharmacist working with an attending physician. Each team sees 15–17 patients daily from 07:00 to 17:00. An afternoon admitting team consisting of an attending physician and one or two residents works from 15:00 to 23:00 Monday through Friday to admit patients to the HM service. A nocturnal covering service provides episodic and emergent care after hours and handles new admissions. This team usually consists of two APPs covering all HM patients and other solid tumour patients (average 200 patients/night), with one or two attendings available as needed on weekdays and weekends from 18:00 to 06:00. Internal medicine residents already used I-PASS to hand off their patients and were not included in the QI study. Before this QI initiative, physicians, APPs and pharmacists did not participate in I-PASS training.
The hospital layout includes two main buildings dedicated to patient care, with floors dedicated to specific tumour types, medical and surgical services, the intensive care unit (ICU) and a bone marrow transplant unit. HM patients are usually spread across various floors throughout these two buildings. Thus, HM patients are rarely located in a single place where providers can see them all at once.
This project was approved by the institution’s Quality Improvement Advisory Board and was part of the Clinical Safety & Effectiveness (CS&E) programme sponsored by our institution’s Office of Performance Improvement.
Needs assessment
In 2016, the institution integrated I-PASS into the electronic health record (EHR) system (Epic/OneConnect, Epic Systems, Verona, Wisconsin, USA). It is a specific handoff communication tool in the EHR that can be accessed within the patient chart (figure 1). I-PASS became readily accessible and viewable by practitioners in all specialties. Alterations were made to meet the hospital’s specific needs directly in Epic. Before the introduction of I-PASS in HM, handoff between providers occurred every Sunday evening for attendings and on varying days for APPs (schedules differed from physician schedules) via email. The APP/attending dyads would use encrypted institutional email communication to securely send patient information and updates to incoming providers. These email handoffs were not standardised, with variable content and length, and were only available at the end of provider shift. For more acutely ill patients, the primary HM team would hand off to the nocturnal covering service using a web-based information management system, Microsoft SharePoint, to which they uploaded patient information and specific items for the incoming team to follow-up on or address. However, this SharePoint site was challenging to find on the hospital intranet, and many providers were unaware that it existed. If a handoff were sent to the night team, they would send an email following up in the morning on the patients who were handed off. Only certain members of the HM team consistently received this email because of failure to update the group email with new providers. Additionally, verbal handoffs were challenging, as the night team’s shift did not overlap with that of the day team.
Completed I-PASS in the EHR. EHR, electronic health record; I-PASS, Illness severity, Patient summary, Action list, Situation awareness and contingency plan, and Synthesis by receiver.
Between November 2018 and January 2019, the topic of handoffs and I-PASS was introduced by the QI team in multiple APP and faculty meetings to review cases in which the use of a standardised handoff could have prevented delays in care and communication errors. A fishbone diagram was created illustrating providers’ barriers to using I-PASS for daily handoffs (figure 2). The barriers identified were similar for faculty and APPs. Based on these findings, a priority impact matrix was created to prioritise strategies to improve I-PASS utilisation and identify opportunities for tool improvement (figure 3). Our strategy consisted of creating dot phrases in the EHR to reduce manual input, development of readily available training materials, education sessions with at-the-elbow support, comprehensive and personalised feedback, email reminders and inclusion of I-PASS training for new hires and trainees. These strategies were implemented in several PDSA cycles.
Fishbone diagram showing reasons providers did not use I-PASS. I-PASS, Illness severity, Patient summary, Action list, Situation awareness and contingency plan, and Synthesis by receiver.
Priority impact matrix. The circled interventions are considered the most feasible and most efficacious. I-PASS, Illness severity, Patient summary, Action list, Situation awareness and contingency plan, and Synthesis by receiver.
Intervention
PDSA cycle 1 (November 2018–February 2019)
With leadership support and buy-in from most providers, the QI study team conducted educational sessions during departmental meetings. Two members of the study team conducted one-to-one session with each member of the HM team at least once to provide hands-on instruction on how to use I-PASS. The training included providing the definitions of each element of I-PASS and a slide presentation to show step-by-step how to open, fill out and save I-PASS in Epic each shift. The trainer also opened I-PASS in Epic to conduct a mock patient handoff using I-PASS and offer scenarios where I-PASS improved the providers’ workflow in Epic. The handoff tool was customised to suit HM needs by creating a built-in smart phrase to include a ‘to do for HM team’ and ‘to do for the nocturnal team’ to assign the responsibilities better. The training slide presentation and step-by-step guide were added to the shared drive for all attendings and residents rotating through the HM service.
Additionally, the team created pocket cards that outlined the parts of I-PASS that were expected to be completed each shift, how they appeared on Epic and a description of what information to include in each section. Quick tips were also included on the cards showing the steps to access the I-PASS handoff in Epic. Reminder cards were placed in working room computers to provide further visual cues. The providers agreed that a verbal handoff is encouraged in addition to the electronic handoff, especially for higher acuity (watcher and unstable) patients. During this cycle, we identified problems with compliance and variability in the handoff content, often failing to follow the I-PASS framework.
The intervention team consisted of one attending, one APP in HM and two APPs in the nocturnal covering service that provided education to respective teams during the first PDSA cycle.
PDSA cycle 2 (March 2019–August 2019)
We addressed the barriers identified in PDSA cycle 1. Monthly I-PASS handoff utilisation rates were extracted electronically and analysed preintervention (January 2018–October 2018) and postintervention (November 2019–August 2019). To reinforce the routine, use of the I-PASS tool in the EHR, a member of the study team sent daily emails for 3 months to remind providers to update I-PASS. Emails were later limited to one every 1–2 weeks and to specific providers who did not use I-PASS daily.
Additionally, examples of optimal and suboptimal handoffs were sent to individual providers to assist with standardising handoff content. During this cycle, periodic training refreshers were provided during staff meetings and progress updates. The interventions and the results of the first two PDSA cycles were presented at the end of the CS&E education programme, and participants were granted certification of completion.
PDSA cycle 3 (September 2019–November 2020)
The study team worked with the biostatistics department to design a postintervention survey. After multiple iterations, the finalised survey included questions about handoff quality, efficiency, avoidable adverse events and provider satisfaction with the new I-PASS handoff tool. The survey was administered in November 2019 and 1 year later. Both the primary and nocturnal covering teams participated in this survey. We added some features to the tool based on the survey feedback, such as a reminder if the handoff was not updated and smart phrases to autopopulate information from the handoff into the patient’s progress notes. Also, to avoid handoff emails altogether, we were able to integrate I-PASS into Haiku. Haiku (Epic Systems, Verona, Wisconsin, USA), an app available to all providers that allows them to access patient information using their smartphones. This app helped the day team check for any overnight events documented by the nocturnal team before the start of the day shift. Additionally, with the Quality Education team’s help, an I-PASS training module for providers was created within the institution’s learning centre. This training is now required during orientation for new hires with clinical duties.
PDSA cycle 4 (December 2020–October 2021)
The analytics team created a real-time dashboard to ensure sustainability in monitoring the I-PASS utilisation rates. The dashboard provided the ability to analyse by provider and monitor I-PASS compliance by component. These I-PASS utilisation metrics were presented monthly at HM meetings. Additionally, another survey was administered 1 year after the first survey. Most recently, the I-PASS utilisation rate was included as a team-based incentive metric for attendings in FY2021 to increase compliance to more than 90%.
Study of interventions
The study team frequently met before and after each intervention to continually refine the interventions and monitor I-PASS utilisation throughout the four PDSA cycles.
Measures
The process measure was I-PASS utilisation. I-PASS was available for all providers in the EHR long before the intervention period. Thus, electronic data were available to determine preintervention and postintervention utilisation rates. We defined ‘complete handoff’ as the presence of content in three sections—illness severity, patient summary and action list—regardless of the date or time entered. The goal was to encourage daily handoffs between primary and nocturnal covering service; for this reason, we also determined the daily utilisation rate. Provider perceptions of I-PASS handoff were measured via survey 10 months after implementation and a year after. These surveys assessed provider satisfaction with the tool, perceptions of patient safety, ease of use and the tool’s effect on communication between providers. Ongoing assessments have been tailored to allow feedback during monthly meetings on I-PASS issues and to repeat education as needed. The nocturnal covering service team also sent email feedback regarding issues with day-team handoff and patients’ acuity labels.
Analysis
The I-PASS utilisation rate was calculated by dividing the number of handoff days (weekdays and weekends) by the number of patient days. A patient day is the census day calculated from midnight to midnight. Handoffs by non-providers and discharge handoff days were excluded from the data set. Monthly average utilisation was displayed using a statistical process control (SPC) P-chart created in Minitab (Minitab, LLC, Herndon, Virginia, USA). The Mann-Whitney test was used to compare preintervention and postintervention I-PASS utilisation rates. Survey results were analysed using SAS statistical software (SAS Institute, Cary, North Carolina, USA). Statistical comparisons between the results of the two surveys were performed using the Fisher exact test. Statistical significance was defined as a p value less than 0.05.
Ethical
The institutional Quality Improvement Board approved this QI study. The intervention was implemented to advance the ethical principle of benevolence by improving patient safety. It encouraged autonomous behaviour within the confines of the structured I-PASS elements. No patients were directly involved, and the care provided to patients was not altered from the standard of care. There are no potential conflicts of interest.
Results
After two PDSA cycles, the SPC chart showed a special cause variation in which the centreline shifted from a preintervention mean of 23% to a postintervention mean of 72% (68% improvement, p<0.001), achieving our first aim. PDSA cycle 3 addressed opportunities for improvement reported by providers in the first survey and led to the creation of strategies to sustain and further improve I-PASS compliance. Special cause variation was also observed during PDSA cycles 3 and 4; the mean I-PASS utilisation rate increased from 70% to 80% to 90% (figure 4). The control limits are narrow because the P chart was not designed for a large sample size. In this case, values beyond the control limits indicated that I-PASS utilisation rates were steadily improving.
Statistical process control P chart for I-PASS utilisation. I-PASS, Illness severity, Patient summary, Action list, Situation awareness and contingency plan, and Synthesis by receiver; LCL, lower control limit; PDSA, plan-do-study-act; UCL, upper control limit.
Initially, we did not have a system to automatically monitor I-PASS compliance by component. In PDSA cycle 2, the providers received periodic feedback on what information should be entered in each I-PASS component and reminders to update the handoff as needed. After automating the data collection and creating the dashboard, we observed that the action list was the most updated component daily (figure 5). We could not measure illness severity because it is a discrete field in Epic. However, illness severity and patient summary are done the first day of admission and updated if patient acuity changes. Another component added to our I-PASS tool was the overnight events section, used by the nocturnal covering team to document interventions during their shift. The situation awareness component is used only when the team wishes to alert the upcoming provider of a foreseeable event and provide guidance on how to proceed to minimise delays in care.
Use of I-PASS components. I-PASS, Illness severity, Patient summary, Action list, Situation awareness and contingency plan, and Synthesis by receiver.
Regarding our Culture of Safety Survey, in 2018, the handoff favourability’s score specific to the statement, ‘Important patient care information is often lost during shift changes’, was 24%. In the same survey in 2022, the favourability score was 64%—a 166% improvement.
HM survey analysis
A total of 27 participants took the first survey, which was conducted from 4 November to 27 November 2019. All 27 participants completed the survey and answered ‘yes’ to the question “Are you using I-PASS?” A total of 25 participants took the second survey, which was conducted from 21 September 2020 to 7 October 2020, and 23 of them completed the survey. Among the 23 who completed the survey, all answered ‘yes’ to the question “Are you using I-PASS?”
For the question ‘It takes on average __________ minute (s) to fill out the I-PASS handoff Per PATIENT’, in survey 1, there were 18 (3 APPs, 11 MDs and 4 PharmDs) picked a time range ‘<5 min’, and 6 (3 APPs and 3 MDs) ‘5–8 min’ and 3 (1 APP and 2 MD) ‘9–12 min’, no one selected ‘>15 min’. In survey 2, there were 20 (4 APP, 15 MD and 1 PharmD) picked a time range ‘< 5 min’ and 3 (2 APP and 1 MD) ‘5–8 min’. No statistical significance was detected on this time question between the two surveys (p=0.205).
The surveys also asked participants to rate, on a 3-point Likert-type scale, their perceptions of improvement in several areas with the use of I-PASS (table 1). No significant differences between the survey responses were detected on these questions (agree: p=0.26, disagree: p=1.00, neutral: p=0.19).
Responses to the question "Do you think the use of I-PASS has improved?” for Hospital Medicine.
A significant difference was detected in the survey 1 and survey 2 responses to the question “How often did the nocturnal team follow up on your action list?” (p=0.0063). The feedback received from the first survey and the subsequent interventions in PDSA cycles 3 and 4 improved the reported frequency of follow-up on action items handed off by the day team (table 2). Additionally, we added a question to survey 2: ‘How easy is it to navigate the handoff tool in Epic?’ Of the 23 participants, 83% (n=19) found it very easy to navigate, and 17% (n=4) were neutral. Furthermore, most participants reported satisfaction with the one-on-one training, the PowerPoint presentation training and the email feedback.
Responses to the question ‘How often did the nocturnal team follow up on your action list?’
Night team survey analysis
A total of 15 members of the nocturnal covering team participated in the first night team survey, which was administered from 7 January to 24 January 2020. Twenty people participated in the second night team survey, which was conducted from 21 September to 6 October 2020. In both surveys, all participants answered ‘yes’ to the question “Are you using I-PASS?” Thirteen of the participants took the survey both times. Two participants in survey 1 did not participate in survey 2.
Both surveys asked the night team to estimate the time (per patient) spent understanding the current patient’s clinical condition before and after adopting I-PASS. The results indicated that without a handoff (‘before’), 14 participants reported taking 5–15 min to understand the current patient’s clinical condition for each patient. When using I-PASS (‘after’), 17 participants indicated that it took them 10 min or less to do the same job. For example, three participants reported taking 5–10 min before the implementation of I-PASS but spent under 5 min when using I-PASS. Another four participants reported taking 10–15 min before I-PASS and being able to provide handoffs in less than 5 min by using I-PASS.
No participants reported spending more than 15 min per patient to review the current patient’s clinical condition required when using I-PASS, including five participants who indicated that they had been spending more than 15 min before I-PASS (four spent 5–10 min using I-PASS and one spent 10–15 min). A paired t-test comparing the ‘before’ and ‘after’ data for these items showed strong significance (survey 1: p=0.000736 and survey 2: p=4.14×10–5). A t-test comparing the difference in time spent on the current patient’s clinical condition (ie, time spent before I-PASS minus time spent after I-PASS) between the two surveys also showed significant differences (p=0.00488). These findings demonstrate that using I-PASS saved the care team’s time (table 3).
Responses to the question "How long does it take on average to gather sufficient information to attend to patients need before and after using I-PASS during night shift?”
For the question regarding perceptions of improvement in several areas with the use of I-PASS the quality of the handoff, efficiency and safety, no statistical significance was detected on these questions between the surveys (agree (p=0.26), disagree (p=1), neutral (p=0.19)) (table 4).
Responses to the question "Do you think the use of I-PASS has improved?” for the Night team.
In addition, no significant differences were detected in how often the action list filled out by the HM team actually reflected the patient’s nighttime needs (p=0.80) (table 5).
Responses to ‘How often does the action list filled out by GIM hospitalist team actually reflect the patient’s needs at night?’
Discussion
Summary
The goal of this QI project in the HM service was to optimise communication between teams, minimise gaps in patient care and provide rapid identification and early intervention in the care of potentially unstable oncology patients. These goals were aligned with those of the institution and the Joint Commission to return patients to cancer care from inpatient hospitalisation safely and without delay.
Before this project, HM providers performed handoffs via emails that were not standardised, often lengthy and inconsistent in content. A SharePoint site outside the institution’s EHR was used for handoff between the primary and the nocturnal covering services. This was used inconsistently and for some patients at the discretion of the primary team based on acuity level and anticipated needs. While the I-PASS handoff tool is built into the institution’s EHR, it was underused. Before our interventions, the utilisation rate of I-PASS in the HM service was only 23%.
Interventions for this QI project included:
Education of providers on using I-PASS handoff through educational materials, online and at point of care stations, daily email reminders and teaching sessions for individuals and the group.
Collection of provider feedback by doing frequent chart reviews, performance updates and weekly emails.
Collaboration with the Internal Medicine Chair, nocturnal and Epic reporting teams.
Optimisation of the I-PASS tool in Epic with the creation of smart phrases to simplify its use and integration of I-PASS in Haiku.
The key findings of our study are that the intervention surpassed our primary aim of achieving an I-PASS utilisation rate of 70% and that our sustainability aim of 90%. The postintervention surveys showed that most providers agreed that using I-PASS improved communication and helped make patient care safer. Reassuringly, we also found that I-PASS handoffs did not prolong the times used for handoff; surveys showed that most providers took less than 5 min per patient to complete I-PASS. Additionally, feedback from providers showed that they were very satisfied with the use of the I-PASS handoff, especially compared with emails. Finally, the nocturnal covering service reported that it took less time to triage and understand patients’ clinical condition with I-PASS than with the previous system.
A particular strength of this QI project was the large number of providers who participated. Also, because HM service has the largest volume of patients in the hospital, changes instituted based on this QI project will affect a substantial number of providers and patients. Finally, including the nocturnal covering service in the surveys broadened the applicability of the study’s findings, as the night team provides coverage for most of the patients in the hospital after hours.
Interpretations and limitations
Our interventions, addressing resistance, identifying barriers to change and conducting education effectively achieved a change in handoff practices. Our experience stands in contrast with the results reported previously, perhaps because of the robustness of our implementation campaign and the support it received from the institutional administration.7
In postimplementation surveys, providers reported that their satisfaction with the I-PASS handoff tool improved after the QI interventions. Also, the I-PASS tool’s utilisation rate increased from 23% to 72%. Although the project’s primary aim was to increase the I-PASS utilisation rate to 70%, the actual increase to 90% showed the commitment and engagement of providers and department leaders in changing the culture and integrating I-PASS handoff into everyone’s day-to-day workflows.
The limitation of our study was the lack of resources to conduct time motion analysis and direct observations. We relied heavily on the provider’s perception to collect the time it takes to complete I-PASS and hand off quality via a survey. Another limitation we had, which most of the studies face, was to correlate our process measure to patient outcomes.8
The monetary costs of the project were minimal, mainly the cost of creating visual guides. However, one trade-off in cost was the time spent educating the providers. Several meetings, presentations and one-on-one training were needed to help providers understand how to use the I-PASS handoff consistently and be more comfortable using it.
Conclusion
This QI project may be used as a model for a larger-scale intervention to improve hospital-wide handoff standardisation in our institution. We are also currently considering a plan for sustaining these results. An institutional task force for I-PASS handoff has been created, with representation from most inpatient services. Meetings and education will continue. The institution is working on an I-PASS dashboard that will display the number of I-PASS handoffs by department. An online I-PASS handoff learning module is available and is mandatory for new trainees and new hires in the HM department.
Inadequate handoff practices and communication failures can lead to uncertainty, work duplication and, potentially, patient harm.9 The Joint Commission recognises I-PASS as an effective tool that helps communicate patient-specific information between individual caregivers or teams of caregivers.10 This QI project was important because it improved the I-PASS standardised handoff tool utilisation rate from 23% to 90% in the HM department. The following steps are to adapt the I-PASS handoff tool for implementation across the institution. We would also like to determine if there is a relationship between the rate of I-PASS use and ICU transfers and safety reports at our institution.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
Ethics approval
This project was reviewed and approved by the Quality Improvement Advisory Board at UT MD Anderson Cancer Center.
Acknowledgments
We thank the providers of Hospital Medicine, Sanjay Shete and Robert Yu from Biostatistics Department, and Amy Ninetto and Kate Krause from the Research Medical Library at UT MD Anderson Cancer Center.
Footnotes
MCFV and MAA contributed equally.
Contributors MCFV, MS and MAA designed the study. MAA performed data collection and data management. DB provided guidance throughout execution of the project. All authors (MCFV, MAA, MS, MG, LD, AM, ZN and DB) contributed to program planning, data interpretation and manuscript writing for this project. MCFV and MAA are responsible for the overall content as the guarantor.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.