Christopher P. Landrigan, MD, MPH

HEALTH LITERACY

SHARED DECISION MAKING

FAMILY CENTERED DISCHARGES

HEALTH LITERACY

SHARED DECISION MAKING

FAMILY CENTERED DISCHARGES

HEALTH LITERACY

SHARED DECISION MAKING

FAMILY CENTERED DISCHARGES

Alarm fatigue is not a new issue for hospitals. In a commentary written over 3 decades ago, Kerr and Hayes described what they saw as an alarming issue developing in intensive care units.[1] Recently multiple organizations, including The Joint Commission and the Emergency Care Research Institute have called out alarm fatigue as a patient safety problem,[2, 3, 4] and organizations such as the American Academy of Pediatrics and the American Heart Association are backing away from recommendations for continuous monitoring.[5, 6] Hospitals are in a scramble to set up alarm committees and address alarms locally as recommended by The Joint Commission.[2] In this issue of the Journal of Hospital Medicine, Paine and colleagues set out to review the small but growing body of literature addressing physiologic monitor alarms and interventions that have tried to address alarm fatigue.[7]

After searching through 4629 titles, the authors found 32 articles addressing their key questions: What proportion of alarms are actionable? What is the relationship between clinicians' alarm exposure and response time? Which interventions are effective for reducing alarm rates? The majority of studies identified were observational, with only 8 studies addressing interventions to reduce alarms. Many of the identified studies occurred in units taking care of adults, though 10 descriptive studies and 1 intervention study occurred in pediatric settings. Perhaps the most concerning finding of all, though not surprising to those who work in the hospital setting, was that somewhere between <1% and 26% of alarms across all studies were considered actionable. Although only specifically addressed in 2 studies, the issue of alarm fatigue (i.e., more alarms leading to slower and sometimes absent clinician response) was supported in both, with nurses having slower responses when exposed to a higher numbers of alarms.[8, 9]

The authors note several limitations of their work, one of which is the modest body of literature on the topic. Although several interventions, including widening alarm parameters, increasing alarm delays, and using disposable leads or daily lead changes, have early evidence of success in safely reducing unnecessary alarms, the heterogeneity of this literature precluded a meta‐analysis. Further, the lack of standard definitions and the variety of methods of determining alarm validity make comparison across studies challenging. For this reason, the authors note that they did not distinguish nuisance alarms (i.e., alarms that accurately reflect the patient condition but do not require any intervention) from invalid alarms (i.e., alarms that do not correctly reflect the patient condition). This is relevant because it is likely that interventions to reduce invalid alarms (e.g., frequent lead changes) may be distinct from those that will successfully address nuisance alarms (e.g., widening alarm limits). It is also important to note that although patient safety is of paramount importance, there were other negative consequences of alarms that the authors did not address in this systemic review. Moreover, although avoiding unrecognized deterioration should be a primary goal of any program to reduce alarm fatigue, death remains uncommon compared to the number of patients, families, and healthcare workers exposed to high numbers of alarms during hospitalization. The high number of nonactionable alarms suggests that part of the burden of this problem may lie in more difficult to quantify outcomes such as sleep quality,[10, 11, 12] patient and parent quality of life during hospitalization,[13, 14] and interrupted tasks and cognitive work of healthcare providers.[15]

Paine and colleagues' review has some certain and some less certain implications for the future of alarm research. First, there is an imminent need for researchers and improvers to develop a consensus around terminology and metrics. We need to agree on what is and is not an actionable alarm, and we need valid and sensitive metrics to better understand the consequences of not monitoring a patient who should be on monitors. Second, hospitals addressing alarm fatigue need benchmarks. As hospitals rush to comply with The Joint Commission National Patient Safety Goals for alarm management,[2] it is safe to say that our goal should not be zero alarms, but how low do you go? What can we consider a safe number of alarms in our hospitals? Smart alarms hold tremendous potential to improve the sensitivity and positive predictive value of alarms. However, their ultimate success is dependent on engineers in industry to develop the technology as well as researchers in the hospital setting to validate the technology's performance in clinical care. Additionally, hospitals need to know which interventions are most effective to implement and how to reliably implement these in daily practice. What seems less certain is what type of research is best suited to address this need. The authors recommend randomized trials as an immediate next step, and certainly trials are the gold standard in determining efficacy. However, trials may overstate effectiveness as complex bundled interventions play out in complex and dynamic hospital systems. Quasiexperimental study designs, including time series and step‐wedge designs, would allow for further scientific discovery, such as which interventions are most effective in certain patient populations, while describing reliable implementation of effective methods that lead to lower alarms rates. In both classical randomized controlled trials and quasiexperiments, factorial designs[16, 17] could give us a better understanding of both the comparative effect and any interaction between interventions.

Alarm fatigue is a widespread problem that has negative effects for patients, families, nurses, and physicians. This review demonstrates that the great majority of alarms do not help clinicians and likely contribute to alarm fatigue. The opportunity to improve care is unquestionably vast, and attention from The Joint Commission and the lay press ensures change will occur. What is critical now is for hospitalists, intensivists, nurses, researchers, and hospital administrators to find the right combination of scientific discovery, thoughtful collaboration with industry, and quality improvement that will inform the literature on which interventions worked, how, and in what setting, and ultimately lead to safer (and quieter) hospitals.

Alarm fatigue is not a new issue for hospitals. In a commentary written over 3 decades ago, Kerr and Hayes described what they saw as an alarming issue developing in intensive care units.[1] Recently multiple organizations, including The Joint Commission and the Emergency Care Research Institute have called out alarm fatigue as a patient safety problem,[2, 3, 4] and organizations such as the American Academy of Pediatrics and the American Heart Association are backing away from recommendations for continuous monitoring.[5, 6] Hospitals are in a scramble to set up alarm committees and address alarms locally as recommended by The Joint Commission.[2] In this issue of the Journal of Hospital Medicine, Paine and colleagues set out to review the small but growing body of literature addressing physiologic monitor alarms and interventions that have tried to address alarm fatigue.[7]

After searching through 4629 titles, the authors found 32 articles addressing their key questions: What proportion of alarms are actionable? What is the relationship between clinicians' alarm exposure and response time? Which interventions are effective for reducing alarm rates? The majority of studies identified were observational, with only 8 studies addressing interventions to reduce alarms. Many of the identified studies occurred in units taking care of adults, though 10 descriptive studies and 1 intervention study occurred in pediatric settings. Perhaps the most concerning finding of all, though not surprising to those who work in the hospital setting, was that somewhere between <1% and 26% of alarms across all studies were considered actionable. Although only specifically addressed in 2 studies, the issue of alarm fatigue (i.e., more alarms leading to slower and sometimes absent clinician response) was supported in both, with nurses having slower responses when exposed to a higher numbers of alarms.[8, 9]

The authors note several limitations of their work, one of which is the modest body of literature on the topic. Although several interventions, including widening alarm parameters, increasing alarm delays, and using disposable leads or daily lead changes, have early evidence of success in safely reducing unnecessary alarms, the heterogeneity of this literature precluded a meta‐analysis. Further, the lack of standard definitions and the variety of methods of determining alarm validity make comparison across studies challenging. For this reason, the authors note that they did not distinguish nuisance alarms (i.e., alarms that accurately reflect the patient condition but do not require any intervention) from invalid alarms (i.e., alarms that do not correctly reflect the patient condition). This is relevant because it is likely that interventions to reduce invalid alarms (e.g., frequent lead changes) may be distinct from those that will successfully address nuisance alarms (e.g., widening alarm limits). It is also important to note that although patient safety is of paramount importance, there were other negative consequences of alarms that the authors did not address in this systemic review. Moreover, although avoiding unrecognized deterioration should be a primary goal of any program to reduce alarm fatigue, death remains uncommon compared to the number of patients, families, and healthcare workers exposed to high numbers of alarms during hospitalization. The high number of nonactionable alarms suggests that part of the burden of this problem may lie in more difficult to quantify outcomes such as sleep quality,[10, 11, 12] patient and parent quality of life during hospitalization,[13, 14] and interrupted tasks and cognitive work of healthcare providers.[15]

Paine and colleagues' review has some certain and some less certain implications for the future of alarm research. First, there is an imminent need for researchers and improvers to develop a consensus around terminology and metrics. We need to agree on what is and is not an actionable alarm, and we need valid and sensitive metrics to better understand the consequences of not monitoring a patient who should be on monitors. Second, hospitals addressing alarm fatigue need benchmarks. As hospitals rush to comply with The Joint Commission National Patient Safety Goals for alarm management,[2] it is safe to say that our goal should not be zero alarms, but how low do you go? What can we consider a safe number of alarms in our hospitals? Smart alarms hold tremendous potential to improve the sensitivity and positive predictive value of alarms. However, their ultimate success is dependent on engineers in industry to develop the technology as well as researchers in the hospital setting to validate the technology's performance in clinical care. Additionally, hospitals need to know which interventions are most effective to implement and how to reliably implement these in daily practice. What seems less certain is what type of research is best suited to address this need. The authors recommend randomized trials as an immediate next step, and certainly trials are the gold standard in determining efficacy. However, trials may overstate effectiveness as complex bundled interventions play out in complex and dynamic hospital systems. Quasiexperimental study designs, including time series and step‐wedge designs, would allow for further scientific discovery, such as which interventions are most effective in certain patient populations, while describing reliable implementation of effective methods that lead to lower alarms rates. In both classical randomized controlled trials and quasiexperiments, factorial designs[16, 17] could give us a better understanding of both the comparative effect and any interaction between interventions.

Alarm fatigue is a widespread problem that has negative effects for patients, families, nurses, and physicians. This review demonstrates that the great majority of alarms do not help clinicians and likely contribute to alarm fatigue. The opportunity to improve care is unquestionably vast, and attention from The Joint Commission and the lay press ensures change will occur. What is critical now is for hospitalists, intensivists, nurses, researchers, and hospital administrators to find the right combination of scientific discovery, thoughtful collaboration with industry, and quality improvement that will inform the literature on which interventions worked, how, and in what setting, and ultimately lead to safer (and quieter) hospitals.

Alarm fatigue is not a new issue for hospitals. In a commentary written over 3 decades ago, Kerr and Hayes described what they saw as an alarming issue developing in intensive care units.[1] Recently multiple organizations, including The Joint Commission and the Emergency Care Research Institute have called out alarm fatigue as a patient safety problem,[2, 3, 4] and organizations such as the American Academy of Pediatrics and the American Heart Association are backing away from recommendations for continuous monitoring.[5, 6] Hospitals are in a scramble to set up alarm committees and address alarms locally as recommended by The Joint Commission.[2] In this issue of the Journal of Hospital Medicine, Paine and colleagues set out to review the small but growing body of literature addressing physiologic monitor alarms and interventions that have tried to address alarm fatigue.[7]

After searching through 4629 titles, the authors found 32 articles addressing their key questions: What proportion of alarms are actionable? What is the relationship between clinicians' alarm exposure and response time? Which interventions are effective for reducing alarm rates? The majority of studies identified were observational, with only 8 studies addressing interventions to reduce alarms. Many of the identified studies occurred in units taking care of adults, though 10 descriptive studies and 1 intervention study occurred in pediatric settings. Perhaps the most concerning finding of all, though not surprising to those who work in the hospital setting, was that somewhere between <1% and 26% of alarms across all studies were considered actionable. Although only specifically addressed in 2 studies, the issue of alarm fatigue (i.e., more alarms leading to slower and sometimes absent clinician response) was supported in both, with nurses having slower responses when exposed to a higher numbers of alarms.[8, 9]

The authors note several limitations of their work, one of which is the modest body of literature on the topic. Although several interventions, including widening alarm parameters, increasing alarm delays, and using disposable leads or daily lead changes, have early evidence of success in safely reducing unnecessary alarms, the heterogeneity of this literature precluded a meta‐analysis. Further, the lack of standard definitions and the variety of methods of determining alarm validity make comparison across studies challenging. For this reason, the authors note that they did not distinguish nuisance alarms (i.e., alarms that accurately reflect the patient condition but do not require any intervention) from invalid alarms (i.e., alarms that do not correctly reflect the patient condition). This is relevant because it is likely that interventions to reduce invalid alarms (e.g., frequent lead changes) may be distinct from those that will successfully address nuisance alarms (e.g., widening alarm limits). It is also important to note that although patient safety is of paramount importance, there were other negative consequences of alarms that the authors did not address in this systemic review. Moreover, although avoiding unrecognized deterioration should be a primary goal of any program to reduce alarm fatigue, death remains uncommon compared to the number of patients, families, and healthcare workers exposed to high numbers of alarms during hospitalization. The high number of nonactionable alarms suggests that part of the burden of this problem may lie in more difficult to quantify outcomes such as sleep quality,[10, 11, 12] patient and parent quality of life during hospitalization,[13, 14] and interrupted tasks and cognitive work of healthcare providers.[15]

Paine and colleagues' review has some certain and some less certain implications for the future of alarm research. First, there is an imminent need for researchers and improvers to develop a consensus around terminology and metrics. We need to agree on what is and is not an actionable alarm, and we need valid and sensitive metrics to better understand the consequences of not monitoring a patient who should be on monitors. Second, hospitals addressing alarm fatigue need benchmarks. As hospitals rush to comply with The Joint Commission National Patient Safety Goals for alarm management,[2] it is safe to say that our goal should not be zero alarms, but how low do you go? What can we consider a safe number of alarms in our hospitals? Smart alarms hold tremendous potential to improve the sensitivity and positive predictive value of alarms. However, their ultimate success is dependent on engineers in industry to develop the technology as well as researchers in the hospital setting to validate the technology's performance in clinical care. Additionally, hospitals need to know which interventions are most effective to implement and how to reliably implement these in daily practice. What seems less certain is what type of research is best suited to address this need. The authors recommend randomized trials as an immediate next step, and certainly trials are the gold standard in determining efficacy. However, trials may overstate effectiveness as complex bundled interventions play out in complex and dynamic hospital systems. Quasiexperimental study designs, including time series and step‐wedge designs, would allow for further scientific discovery, such as which interventions are most effective in certain patient populations, while describing reliable implementation of effective methods that lead to lower alarms rates. In both classical randomized controlled trials and quasiexperiments, factorial designs[16, 17] could give us a better understanding of both the comparative effect and any interaction between interventions.

Alarm fatigue is a widespread problem that has negative effects for patients, families, nurses, and physicians. This review demonstrates that the great majority of alarms do not help clinicians and likely contribute to alarm fatigue. The opportunity to improve care is unquestionably vast, and attention from The Joint Commission and the lay press ensures change will occur. What is critical now is for hospitalists, intensivists, nurses, researchers, and hospital administrators to find the right combination of scientific discovery, thoughtful collaboration with industry, and quality improvement that will inform the literature on which interventions worked, how, and in what setting, and ultimately lead to safer (and quieter) hospitals.

Handoffs among hospital providers are highly error prone and can result in serious morbidity and mortality. Best practices for verbal handoffs have been described[1, 2, 3, 4] and include conducting verbal handoffs face to face, providing opportunities for questions, having the receiver perform a readback, as well as specific content recommendations including action items. Far less research has focused on best practices for printed handoff documents,[5, 6] despite the routine use of written handoff tools as a reference by on‐call physicians.[7, 8] Erroneous or outdated information on the written handoff can mislead on‐call providers, potentially leading to serious medical errors.

In their most basic form, printed handoff documents list patients for whom a provider is responsible. Typically, they also contain demographic information, reason for hospital admission, and a task list for each patient. They may also contain more detailed information on patient history, hospital course, and/or care plan, and may vary among specialties.[9] They come in various forms, ranging from index cards with handwritten notes, to word‐processor or spreadsheet documents, to printed documents that are autopopulated from the electronic health record (EHR).[2] Importantly, printed handoff documents supplement the verbal handoff by allowing receivers to follow along as patients are presented. The concurrent use of written and verbal handoffs may improve retention of clinical information as compared with either alone.[10, 11]

The Joint Commission requires an institutional approach to patient handoffs.[12] The requirements state that handoff communication solutions should take a standardized form, but they do not provide details regarding what data elements should be included in printed or verbal handoffs. Accreditation Council for Graduate Medical Education Common Program Requirements likewise require that residents must become competent in patient handoffs[13] but do not provide specific details or measurement tools. Absent widely accepted guidelines, decisions regarding which elements to include in printed handoff documents are currently made at an individual or institutional level.

The I‐PASS study is a federally funded multi‐institutional project that demonstrated a decrease in medical errors and preventable adverse events after implementation of a standardized resident handoff bundle.[14, 15] The I‐PASS Study Group developed a bundle of handoff interventions, beginning with a handoff and teamwork training program (based in part on TeamSTEPPS [Team Strategies and Tools to Enhance Performance and Patient Safety]),[16] a novel verbal mnemonic, I‐PASS (Illness Severity, Patient Summary, Action List, Situation Awareness and Contingency Planning, and Synthesis by Receiver),[17] and changes to the verbal handoff process, in addition to several other elements.

We hypothesized that developing a standardized printed handoff template would reinforce the handoff training and enhance the value of the verbal handoff process changes. Given the paucity of data on best printed handoff practices, however, we first conducted a needs assessment to identify which data elements were currently contained in printed handoffs across sites, and to allow an expert panel to make recommendations for best practices.

METHODS

I‐PASS Study sites included 9 pediatric residency programs at academic medical centers from across North America. Programs were identified through professional networks and invited to participate. The nonintensive care unit hospitalist services at these medical centers are primarily staffed by residents and medical students with attending supervision. At 1 site, nurse practitioners also participate in care. Additional details about study sites can be found in the study descriptions previously published.[14, 15] All sites received local institutional review board approval.

We began by inviting members of the I‐PASS Education Executive Committee (EEC)[14] to build a collective, comprehensive list of possible data elements for printed handoff documents. This committee included pediatric residency program directors, pediatric hospitalists, education researchers, health services researchers, and patient safety experts. We obtained sample handoff documents from pediatric hospitalist services at each of 9 institutions in the United States and Canada (with protected health information redacted). We reviewed these sample handoff documents to characterize their format and to determine what discrete data elements appeared in each site's printed handoff document. Presence or absence of each data element across sites was tabulated. We also queried sites to determine the feasibility of including elements that were not presently included.

Subsequently, I‐PASS site investigators led structured group interviews at participating sites to gather additional information about handoff practices at each site. These structured group interviews included diverse representation from residents, faculty, and residency program leadership, as well as hospitalists and medical students, to ensure the comprehensive acquisition of information regarding site‐specific characteristics. Each group provided answers to a standardized set of open‐ended questions that addressed current practices, handoff education, simulation use, team structure, and the nature of current written handoff tools, if applicable, at each site. One member of the structured group interview served as a scribe and created a document that summarized the content of the structured group interview meeting and answers to the standardized questions.

RESULTS

Nine programs (100%) provided data on the structure and contents of their printed handoff documents. We found wide variation in structure across the 9 sites. Three sites used a word‐processorbased document that required manual entry of all data elements. The other 6 institutions had a direct link with the EHR to enable autopopulation of between 10 and 20 elements on the printed handoff document.

The content of written handoff documents, as well as the sources of data included in them (present or future), likewise varied substantially across sites (Table 1). Only 4 data elements (name, age, weight, and a list of medications) were universally included at all 9 sites. Among the 6 institutions that linked the printed handoff to the EHR, there was also substantial variation in which elements were autoimported. Only 7 elements were universally autoimported at these 6 sites: patient name, medical record number, room number, weight, date of birth, age, and date of admission. Two elements from the original brainstorming were not presently included in any sites' documents (emergency contact and primary language).

Nine institutions (100%) conducted structured group interviews, ranging in size from 4 to 27 individuals with a median of 5 participants. The documents containing information from each site were provided to the authors. The authors then tabulated categorical information, reviewed narrative comments to understand current institutional practices, and conducted simple content analysis to identify areas of concordance or discordance, particularly with respect to data elements and EHR usage. Based on the results of the printed handoff document review and structured group interviews, with additional perspectives provided by the I‐PASS EEC, the expert panel came to consensus on a list of 23 elements that should be included in printed handoff documents, including 15 essential data elements and 8 additional recommended elements (Table 2).

Evidence ratings[19] of these elements are included. Several elements are classified as I‐B (effective, nonrandomized studies) based on either studies of individual elements, or greater than 1 study of bundled elements that could reasonably be extrapolated. These include Illness severity,[20, 21] patient summary,[21, 22] action items[21, 22] (to do lists), situation awareness and contingency plan,[21, 22] and medications[22, 23, 24] with attention to importing from the EHR. Medications entered as free text were classified as IIa‐C because of risk and potential significance of errors; in particular there was concern that transcription errors, errors of omission, or errors of commission could potentially lead to patient harms. The remaining essential elements are classified as I‐C (effective, expert consensus). Of note, date of birth was specifically included as a patient identifier, distinct from age, which was felt to be useful as a descriptor (often within a one‐liner or as part of the patient summary).

The 8 recommended elements were elements for which there was not unanimous agreement on inclusion, but the majority of the panel felt they should be included. These elements were classified as IIa‐C, with 1 exception. Code status generated significant controversy among the group. After extensive discussion among the group and consideration of safety, supervision, educational, and pediatric‐specific considerations, all members of the group agreed on the categorization as a recommended element; it is classified as IIb‐C.

All members of the group agreed that data elements should be directly imported from the EHR whenever possible. Finally, members agreed that the elements that make up the I‐PASS mnemonic (illness severity, patient summary, action items, situation awareness/contingency planning) should be listed in that order whenever possible. A sample I‐PASS‐compliant printed handoff document is shown Figure 1.

Figure 1

DISCUSSION

We identified substantial variability in the structure and content of printed handoff documents used by 9 pediatric hospitalist teaching services, reflective of a lack of standardization. We found that institutional printed handoff documents shared some demographic elements (eg, name, room, medical record number) but also varied in clinical content (eg, vital signs, lab tests, code status). Our expert panel developed a list of 15 essential and 8 recommended data elements for printed handoff documents. Although this is a large number of fields, the majority of the essential fields were already included by most sites, and many are basic demographic identifiers. Illness severity is the 1 essential field that was not routinely included; however, including this type of overview is consistently recommended[2, 4] and supported by evidence,[20, 21] and contributes to building a shared mental model.[16] We recommend the categories of stable/watcher/unstable.[17]

Several prior single‐center studies have found that introducing a printed handoff document can lead to improvements in workflow, communication, and patient safety. In an early study, Petersen et al.[25] showed an association between use of a computerized sign‐out program and reduced odds of preventable adverse events during periods of cross‐coverage. Wayne et al.[26] reported fewer perceived inaccuracies in handoff documents as well as improved clarity at the time of transfer, supporting the role for standardization. Van Eaton et al.[27] demonstrated rapid uptake and desirability of a computerized handoff document, which combined autoimportation of information from an EHR with resident‐entered patient details, reflecting the importance of both data sources. In addition, they demonstrated improvements in both the rounding and sign‐out processes.[28]

Two studies specifically reported the increased use of specific fields after implementation. Payne et al. implemented a Web‐based handoff tool and documented significant increases in the number of handoffs containing problem lists, medication lists, and code status, accompanied by perceived improvements in quality of handoffs and fewer near‐miss events.[24] Starmer et al. found that introduction of a resident handoff bundle that included a printed handoff tool led to reduction in medical errors and adverse events.[22] The study group using the tool populated 11 data elements more often after implementation, and introduction of this printed handoff tool in particular was associated with reductions in written handoff miscommunications. Neither of these studies included subanalysis to indicate which data elements may have been most important.

In contrast to previous single‐institution studies, our recommendations for a printed handoff template come from evaluations of tools and discussions with front line providers across 9 institutions. We had substantial overlap with data elements recommended by Van Eaton et al.[27] However, there were several areas in which we did not have overlap with published templates including weight, ins/outs, primary language, emergency contact information, or primary care provider. Other published handoff tools have been highly specialized (eg, for cardiac intensive care) or included many fewer data elements than our group felt were essential. These differences may reflect the unique aspects of caring for pediatric patients (eg, need for weights) and the absence of defined protocols for many pediatric conditions. In addition, the level of detail needed for contingency planning may vary between teaching and nonteaching services.

Resident physicians may provide valuable information in the development of standardized handoff documents. Clark et al.,[29] at Virginia Mason University, utilized resident‐driven continuous quality improvement processes including real‐time feedback to implement an electronic template. They found that engagement of both senior leaders and front‐line users was an important component of their success in uptake. Our study utilized residents as essential members of structured group interviews to ensure that front‐line users' needs were represented as recommendations for a printed handoff tool template were developed.

As previously described,[17] our study group had identified several key data elements that should be included in verbal handoffs: illness severity, a patient summary, a discrete action list, situation awareness/contingency planning, and a synthesis by receiver. With consideration of the multivoting results as well as known best practices,[1, 4, 12] the expert panel for this study agreed that each of these elements should also be highlighted in the printed template to ensure consistency between the printed document and the verbal handoff, and to have each reinforce the other. On the printed handoff tool, the final S in the I‐PASS mnemonic (synthesis by receiver) cannot be prepopulated, but considering the importance of this step,[16, 30, 31, 32] it should be printed as synthesis by receiver to serve as a text‐reminder to both givers and receivers.

The panel also felt, however, that the printed handoff document should provide additional background information not routinely included in a verbal handoff. It should serve as a reference tool both at the time of verbal handoff and throughout the day and night, and therefore should include more comprehensive information than is necessary or appropriate to convey during the verbal handoff. We identified 10 data elements that are essential in a printed handoff document in addition to the I‐PASS elements (Table 2).

Patient demographic data elements, as well as team assignments and attending physician, were uniformly supported for inclusion. The medication list was viewed as essential; however, the panel also recognized the potential for medical errors due to inaccuracies in the medication list. In particular, there was concern that including all fields of a medication order (drug, dose, route, frequency) would result in handoffs containing a high proportion of inaccurate information, particularly for complex patients whose medication regimens may vary over the course of hospitalization. Therefore, the panel agreed that if medication lists were entered manually, then only the medication name should be included as they did not wish to perpetuate inaccurate or potentially harmful information. If medication lists were autoimported from an EHR, then they should include drug name, dose, route, and frequency if possible.

In the I‐PASS study,[15] all institutions implemented printed handoff documents that included fields for the essential data elements. After implementation, there was a significant increase in completion of all essential fields. Although there is limited evidence to support any individual data element, increased usage of these elements was associated with the overall study finding of decreased rates of medical errors and preventable adverse events.

EHRs have the potential to help standardize printed handoff documents[5, 6, 33, 34, 35]; all participants in our study agreed that printed handoff documents should ideally be linked with the EHR and should autoimport data wherever appropriate. Manually populated (eg, word processor‐ or spreadsheet‐based) handoff tools have important limitations, particularly related to the potential for typographical errors as well as accidental omission of data fields, and lead to unnecessary duplication of work (eg, re‐entering data already included in a progress note) that can waste providers' time. It was also acknowledged that word processor‐ or spreadsheet‐based documents may have flexibility that is lacking in EHR‐based handoff documents. For example, formatting can more easily be adjusted to increase the number of patients per printed page. As technology advances, printed documents may be phased out in favor of EHR‐based on‐screen reports, which by their nature would be more accurate due to real‐time autoupdates.

In making recommendations about essential versus recommended items for inclusion in the printed handoff template, the only data element that generated controversy among our experts was code status. Some felt that it should be included as an essential element, whereas others did not. We believe that this was unique to our practice in pediatric hospital ward settings, as codes in most pediatric ward settings are rare. Among the concerns expressed with including code status for all patients were that residents might assume patients were full‐code without verifying. The potential inaccuracy created by this might have severe implications. Alternatively, residents might feel obligated to have code discussions with all patients regardless of severity of illness, which may be inappropriate in a pediatric population. Several educators expressed concerns about trainees having unsupervised code‐status conversations with families of pediatric patients. Conversely, although codes are rare in pediatric ward settings, concerns were raised that not including code status could be problematic during these rare but critically important events. Other fields, such as weight, might have less relevance for an adult population in which emergency drug doses are standardized.

CONCLUSIONS

In this article we describe the results of a needs assessment that informed expert consensus‐based recommendations for data elements to include in a printed handoff document. We recommend that pediatric programs include the elements identified as part of a standardized written handoff tool. Although many of these elements are also applicable to other specialties, future work should be conducted to adapt the printed handoff document elements described here for use in other specialties and settings. Future studies should work to validate the importance of these elements, studying the manner in which their inclusion affects the quality of written handoffs, and ultimately patient safety.

Handoffs among hospital providers are highly error prone and can result in serious morbidity and mortality. Best practices for verbal handoffs have been described[1, 2, 3, 4] and include conducting verbal handoffs face to face, providing opportunities for questions, having the receiver perform a readback, as well as specific content recommendations including action items. Far less research has focused on best practices for printed handoff documents,[5, 6] despite the routine use of written handoff tools as a reference by on‐call physicians.[7, 8] Erroneous or outdated information on the written handoff can mislead on‐call providers, potentially leading to serious medical errors.

In their most basic form, printed handoff documents list patients for whom a provider is responsible. Typically, they also contain demographic information, reason for hospital admission, and a task list for each patient. They may also contain more detailed information on patient history, hospital course, and/or care plan, and may vary among specialties.[9] They come in various forms, ranging from index cards with handwritten notes, to word‐processor or spreadsheet documents, to printed documents that are autopopulated from the electronic health record (EHR).[2] Importantly, printed handoff documents supplement the verbal handoff by allowing receivers to follow along as patients are presented. The concurrent use of written and verbal handoffs may improve retention of clinical information as compared with either alone.[10, 11]

The Joint Commission requires an institutional approach to patient handoffs.[12] The requirements state that handoff communication solutions should take a standardized form, but they do not provide details regarding what data elements should be included in printed or verbal handoffs. Accreditation Council for Graduate Medical Education Common Program Requirements likewise require that residents must become competent in patient handoffs[13] but do not provide specific details or measurement tools. Absent widely accepted guidelines, decisions regarding which elements to include in printed handoff documents are currently made at an individual or institutional level.

The I‐PASS study is a federally funded multi‐institutional project that demonstrated a decrease in medical errors and preventable adverse events after implementation of a standardized resident handoff bundle.[14, 15] The I‐PASS Study Group developed a bundle of handoff interventions, beginning with a handoff and teamwork training program (based in part on TeamSTEPPS [Team Strategies and Tools to Enhance Performance and Patient Safety]),[16] a novel verbal mnemonic, I‐PASS (Illness Severity, Patient Summary, Action List, Situation Awareness and Contingency Planning, and Synthesis by Receiver),[17] and changes to the verbal handoff process, in addition to several other elements.

We hypothesized that developing a standardized printed handoff template would reinforce the handoff training and enhance the value of the verbal handoff process changes. Given the paucity of data on best printed handoff practices, however, we first conducted a needs assessment to identify which data elements were currently contained in printed handoffs across sites, and to allow an expert panel to make recommendations for best practices.

METHODS

I‐PASS Study sites included 9 pediatric residency programs at academic medical centers from across North America. Programs were identified through professional networks and invited to participate. The nonintensive care unit hospitalist services at these medical centers are primarily staffed by residents and medical students with attending supervision. At 1 site, nurse practitioners also participate in care. Additional details about study sites can be found in the study descriptions previously published.[14, 15] All sites received local institutional review board approval.

We began by inviting members of the I‐PASS Education Executive Committee (EEC)[14] to build a collective, comprehensive list of possible data elements for printed handoff documents. This committee included pediatric residency program directors, pediatric hospitalists, education researchers, health services researchers, and patient safety experts. We obtained sample handoff documents from pediatric hospitalist services at each of 9 institutions in the United States and Canada (with protected health information redacted). We reviewed these sample handoff documents to characterize their format and to determine what discrete data elements appeared in each site's printed handoff document. Presence or absence of each data element across sites was tabulated. We also queried sites to determine the feasibility of including elements that were not presently included.

Subsequently, I‐PASS site investigators led structured group interviews at participating sites to gather additional information about handoff practices at each site. These structured group interviews included diverse representation from residents, faculty, and residency program leadership, as well as hospitalists and medical students, to ensure the comprehensive acquisition of information regarding site‐specific characteristics. Each group provided answers to a standardized set of open‐ended questions that addressed current practices, handoff education, simulation use, team structure, and the nature of current written handoff tools, if applicable, at each site. One member of the structured group interview served as a scribe and created a document that summarized the content of the structured group interview meeting and answers to the standardized questions.

RESULTS

Nine programs (100%) provided data on the structure and contents of their printed handoff documents. We found wide variation in structure across the 9 sites. Three sites used a word‐processorbased document that required manual entry of all data elements. The other 6 institutions had a direct link with the EHR to enable autopopulation of between 10 and 20 elements on the printed handoff document.

The content of written handoff documents, as well as the sources of data included in them (present or future), likewise varied substantially across sites (Table 1). Only 4 data elements (name, age, weight, and a list of medications) were universally included at all 9 sites. Among the 6 institutions that linked the printed handoff to the EHR, there was also substantial variation in which elements were autoimported. Only 7 elements were universally autoimported at these 6 sites: patient name, medical record number, room number, weight, date of birth, age, and date of admission. Two elements from the original brainstorming were not presently included in any sites' documents (emergency contact and primary language).

Nine institutions (100%) conducted structured group interviews, ranging in size from 4 to 27 individuals with a median of 5 participants. The documents containing information from each site were provided to the authors. The authors then tabulated categorical information, reviewed narrative comments to understand current institutional practices, and conducted simple content analysis to identify areas of concordance or discordance, particularly with respect to data elements and EHR usage. Based on the results of the printed handoff document review and structured group interviews, with additional perspectives provided by the I‐PASS EEC, the expert panel came to consensus on a list of 23 elements that should be included in printed handoff documents, including 15 essential data elements and 8 additional recommended elements (Table 2).

Evidence ratings[19] of these elements are included. Several elements are classified as I‐B (effective, nonrandomized studies) based on either studies of individual elements, or greater than 1 study of bundled elements that could reasonably be extrapolated. These include Illness severity,[20, 21] patient summary,[21, 22] action items[21, 22] (to do lists), situation awareness and contingency plan,[21, 22] and medications[22, 23, 24] with attention to importing from the EHR. Medications entered as free text were classified as IIa‐C because of risk and potential significance of errors; in particular there was concern that transcription errors, errors of omission, or errors of commission could potentially lead to patient harms. The remaining essential elements are classified as I‐C (effective, expert consensus). Of note, date of birth was specifically included as a patient identifier, distinct from age, which was felt to be useful as a descriptor (often within a one‐liner or as part of the patient summary).

The 8 recommended elements were elements for which there was not unanimous agreement on inclusion, but the majority of the panel felt they should be included. These elements were classified as IIa‐C, with 1 exception. Code status generated significant controversy among the group. After extensive discussion among the group and consideration of safety, supervision, educational, and pediatric‐specific considerations, all members of the group agreed on the categorization as a recommended element; it is classified as IIb‐C.

All members of the group agreed that data elements should be directly imported from the EHR whenever possible. Finally, members agreed that the elements that make up the I‐PASS mnemonic (illness severity, patient summary, action items, situation awareness/contingency planning) should be listed in that order whenever possible. A sample I‐PASS‐compliant printed handoff document is shown Figure 1.

Figure 1

DISCUSSION

We identified substantial variability in the structure and content of printed handoff documents used by 9 pediatric hospitalist teaching services, reflective of a lack of standardization. We found that institutional printed handoff documents shared some demographic elements (eg, name, room, medical record number) but also varied in clinical content (eg, vital signs, lab tests, code status). Our expert panel developed a list of 15 essential and 8 recommended data elements for printed handoff documents. Although this is a large number of fields, the majority of the essential fields were already included by most sites, and many are basic demographic identifiers. Illness severity is the 1 essential field that was not routinely included; however, including this type of overview is consistently recommended[2, 4] and supported by evidence,[20, 21] and contributes to building a shared mental model.[16] We recommend the categories of stable/watcher/unstable.[17]

Several prior single‐center studies have found that introducing a printed handoff document can lead to improvements in workflow, communication, and patient safety. In an early study, Petersen et al.[25] showed an association between use of a computerized sign‐out program and reduced odds of preventable adverse events during periods of cross‐coverage. Wayne et al.[26] reported fewer perceived inaccuracies in handoff documents as well as improved clarity at the time of transfer, supporting the role for standardization. Van Eaton et al.[27] demonstrated rapid uptake and desirability of a computerized handoff document, which combined autoimportation of information from an EHR with resident‐entered patient details, reflecting the importance of both data sources. In addition, they demonstrated improvements in both the rounding and sign‐out processes.[28]

Two studies specifically reported the increased use of specific fields after implementation. Payne et al. implemented a Web‐based handoff tool and documented significant increases in the number of handoffs containing problem lists, medication lists, and code status, accompanied by perceived improvements in quality of handoffs and fewer near‐miss events.[24] Starmer et al. found that introduction of a resident handoff bundle that included a printed handoff tool led to reduction in medical errors and adverse events.[22] The study group using the tool populated 11 data elements more often after implementation, and introduction of this printed handoff tool in particular was associated with reductions in written handoff miscommunications. Neither of these studies included subanalysis to indicate which data elements may have been most important.

In contrast to previous single‐institution studies, our recommendations for a printed handoff template come from evaluations of tools and discussions with front line providers across 9 institutions. We had substantial overlap with data elements recommended by Van Eaton et al.[27] However, there were several areas in which we did not have overlap with published templates including weight, ins/outs, primary language, emergency contact information, or primary care provider. Other published handoff tools have been highly specialized (eg, for cardiac intensive care) or included many fewer data elements than our group felt were essential. These differences may reflect the unique aspects of caring for pediatric patients (eg, need for weights) and the absence of defined protocols for many pediatric conditions. In addition, the level of detail needed for contingency planning may vary between teaching and nonteaching services.

Resident physicians may provide valuable information in the development of standardized handoff documents. Clark et al.,[29] at Virginia Mason University, utilized resident‐driven continuous quality improvement processes including real‐time feedback to implement an electronic template. They found that engagement of both senior leaders and front‐line users was an important component of their success in uptake. Our study utilized residents as essential members of structured group interviews to ensure that front‐line users' needs were represented as recommendations for a printed handoff tool template were developed.

As previously described,[17] our study group had identified several key data elements that should be included in verbal handoffs: illness severity, a patient summary, a discrete action list, situation awareness/contingency planning, and a synthesis by receiver. With consideration of the multivoting results as well as known best practices,[1, 4, 12] the expert panel for this study agreed that each of these elements should also be highlighted in the printed template to ensure consistency between the printed document and the verbal handoff, and to have each reinforce the other. On the printed handoff tool, the final S in the I‐PASS mnemonic (synthesis by receiver) cannot be prepopulated, but considering the importance of this step,[16, 30, 31, 32] it should be printed as synthesis by receiver to serve as a text‐reminder to both givers and receivers.

The panel also felt, however, that the printed handoff document should provide additional background information not routinely included in a verbal handoff. It should serve as a reference tool both at the time of verbal handoff and throughout the day and night, and therefore should include more comprehensive information than is necessary or appropriate to convey during the verbal handoff. We identified 10 data elements that are essential in a printed handoff document in addition to the I‐PASS elements (Table 2).

Patient demographic data elements, as well as team assignments and attending physician, were uniformly supported for inclusion. The medication list was viewed as essential; however, the panel also recognized the potential for medical errors due to inaccuracies in the medication list. In particular, there was concern that including all fields of a medication order (drug, dose, route, frequency) would result in handoffs containing a high proportion of inaccurate information, particularly for complex patients whose medication regimens may vary over the course of hospitalization. Therefore, the panel agreed that if medication lists were entered manually, then only the medication name should be included as they did not wish to perpetuate inaccurate or potentially harmful information. If medication lists were autoimported from an EHR, then they should include drug name, dose, route, and frequency if possible.

In the I‐PASS study,[15] all institutions implemented printed handoff documents that included fields for the essential data elements. After implementation, there was a significant increase in completion of all essential fields. Although there is limited evidence to support any individual data element, increased usage of these elements was associated with the overall study finding of decreased rates of medical errors and preventable adverse events.

EHRs have the potential to help standardize printed handoff documents[5, 6, 33, 34, 35]; all participants in our study agreed that printed handoff documents should ideally be linked with the EHR and should autoimport data wherever appropriate. Manually populated (eg, word processor‐ or spreadsheet‐based) handoff tools have important limitations, particularly related to the potential for typographical errors as well as accidental omission of data fields, and lead to unnecessary duplication of work (eg, re‐entering data already included in a progress note) that can waste providers' time. It was also acknowledged that word processor‐ or spreadsheet‐based documents may have flexibility that is lacking in EHR‐based handoff documents. For example, formatting can more easily be adjusted to increase the number of patients per printed page. As technology advances, printed documents may be phased out in favor of EHR‐based on‐screen reports, which by their nature would be more accurate due to real‐time autoupdates.

In making recommendations about essential versus recommended items for inclusion in the printed handoff template, the only data element that generated controversy among our experts was code status. Some felt that it should be included as an essential element, whereas others did not. We believe that this was unique to our practice in pediatric hospital ward settings, as codes in most pediatric ward settings are rare. Among the concerns expressed with including code status for all patients were that residents might assume patients were full‐code without verifying. The potential inaccuracy created by this might have severe implications. Alternatively, residents might feel obligated to have code discussions with all patients regardless of severity of illness, which may be inappropriate in a pediatric population. Several educators expressed concerns about trainees having unsupervised code‐status conversations with families of pediatric patients. Conversely, although codes are rare in pediatric ward settings, concerns were raised that not including code status could be problematic during these rare but critically important events. Other fields, such as weight, might have less relevance for an adult population in which emergency drug doses are standardized.

CONCLUSIONS

In this article we describe the results of a needs assessment that informed expert consensus‐based recommendations for data elements to include in a printed handoff document. We recommend that pediatric programs include the elements identified as part of a standardized written handoff tool. Although many of these elements are also applicable to other specialties, future work should be conducted to adapt the printed handoff document elements described here for use in other specialties and settings. Future studies should work to validate the importance of these elements, studying the manner in which their inclusion affects the quality of written handoffs, and ultimately patient safety.

Handoffs among hospital providers are highly error prone and can result in serious morbidity and mortality. Best practices for verbal handoffs have been described[1, 2, 3, 4] and include conducting verbal handoffs face to face, providing opportunities for questions, having the receiver perform a readback, as well as specific content recommendations including action items. Far less research has focused on best practices for printed handoff documents,[5, 6] despite the routine use of written handoff tools as a reference by on‐call physicians.[7, 8] Erroneous or outdated information on the written handoff can mislead on‐call providers, potentially leading to serious medical errors.

In their most basic form, printed handoff documents list patients for whom a provider is responsible. Typically, they also contain demographic information, reason for hospital admission, and a task list for each patient. They may also contain more detailed information on patient history, hospital course, and/or care plan, and may vary among specialties.[9] They come in various forms, ranging from index cards with handwritten notes, to word‐processor or spreadsheet documents, to printed documents that are autopopulated from the electronic health record (EHR).[2] Importantly, printed handoff documents supplement the verbal handoff by allowing receivers to follow along as patients are presented. The concurrent use of written and verbal handoffs may improve retention of clinical information as compared with either alone.[10, 11]

The Joint Commission requires an institutional approach to patient handoffs.[12] The requirements state that handoff communication solutions should take a standardized form, but they do not provide details regarding what data elements should be included in printed or verbal handoffs. Accreditation Council for Graduate Medical Education Common Program Requirements likewise require that residents must become competent in patient handoffs[13] but do not provide specific details or measurement tools. Absent widely accepted guidelines, decisions regarding which elements to include in printed handoff documents are currently made at an individual or institutional level.

The I‐PASS study is a federally funded multi‐institutional project that demonstrated a decrease in medical errors and preventable adverse events after implementation of a standardized resident handoff bundle.[14, 15] The I‐PASS Study Group developed a bundle of handoff interventions, beginning with a handoff and teamwork training program (based in part on TeamSTEPPS [Team Strategies and Tools to Enhance Performance and Patient Safety]),[16] a novel verbal mnemonic, I‐PASS (Illness Severity, Patient Summary, Action List, Situation Awareness and Contingency Planning, and Synthesis by Receiver),[17] and changes to the verbal handoff process, in addition to several other elements.

We hypothesized that developing a standardized printed handoff template would reinforce the handoff training and enhance the value of the verbal handoff process changes. Given the paucity of data on best printed handoff practices, however, we first conducted a needs assessment to identify which data elements were currently contained in printed handoffs across sites, and to allow an expert panel to make recommendations for best practices.

METHODS

I‐PASS Study sites included 9 pediatric residency programs at academic medical centers from across North America. Programs were identified through professional networks and invited to participate. The nonintensive care unit hospitalist services at these medical centers are primarily staffed by residents and medical students with attending supervision. At 1 site, nurse practitioners also participate in care. Additional details about study sites can be found in the study descriptions previously published.[14, 15] All sites received local institutional review board approval.

We began by inviting members of the I‐PASS Education Executive Committee (EEC)[14] to build a collective, comprehensive list of possible data elements for printed handoff documents. This committee included pediatric residency program directors, pediatric hospitalists, education researchers, health services researchers, and patient safety experts. We obtained sample handoff documents from pediatric hospitalist services at each of 9 institutions in the United States and Canada (with protected health information redacted). We reviewed these sample handoff documents to characterize their format and to determine what discrete data elements appeared in each site's printed handoff document. Presence or absence of each data element across sites was tabulated. We also queried sites to determine the feasibility of including elements that were not presently included.

Subsequently, I‐PASS site investigators led structured group interviews at participating sites to gather additional information about handoff practices at each site. These structured group interviews included diverse representation from residents, faculty, and residency program leadership, as well as hospitalists and medical students, to ensure the comprehensive acquisition of information regarding site‐specific characteristics. Each group provided answers to a standardized set of open‐ended questions that addressed current practices, handoff education, simulation use, team structure, and the nature of current written handoff tools, if applicable, at each site. One member of the structured group interview served as a scribe and created a document that summarized the content of the structured group interview meeting and answers to the standardized questions.

RESULTS

Nine programs (100%) provided data on the structure and contents of their printed handoff documents. We found wide variation in structure across the 9 sites. Three sites used a word‐processorbased document that required manual entry of all data elements. The other 6 institutions had a direct link with the EHR to enable autopopulation of between 10 and 20 elements on the printed handoff document.

The content of written handoff documents, as well as the sources of data included in them (present or future), likewise varied substantially across sites (Table 1). Only 4 data elements (name, age, weight, and a list of medications) were universally included at all 9 sites. Among the 6 institutions that linked the printed handoff to the EHR, there was also substantial variation in which elements were autoimported. Only 7 elements were universally autoimported at these 6 sites: patient name, medical record number, room number, weight, date of birth, age, and date of admission. Two elements from the original brainstorming were not presently included in any sites' documents (emergency contact and primary language).

Nine institutions (100%) conducted structured group interviews, ranging in size from 4 to 27 individuals with a median of 5 participants. The documents containing information from each site were provided to the authors. The authors then tabulated categorical information, reviewed narrative comments to understand current institutional practices, and conducted simple content analysis to identify areas of concordance or discordance, particularly with respect to data elements and EHR usage. Based on the results of the printed handoff document review and structured group interviews, with additional perspectives provided by the I‐PASS EEC, the expert panel came to consensus on a list of 23 elements that should be included in printed handoff documents, including 15 essential data elements and 8 additional recommended elements (Table 2).

Evidence ratings[19] of these elements are included. Several elements are classified as I‐B (effective, nonrandomized studies) based on either studies of individual elements, or greater than 1 study of bundled elements that could reasonably be extrapolated. These include Illness severity,[20, 21] patient summary,[21, 22] action items[21, 22] (to do lists), situation awareness and contingency plan,[21, 22] and medications[22, 23, 24] with attention to importing from the EHR. Medications entered as free text were classified as IIa‐C because of risk and potential significance of errors; in particular there was concern that transcription errors, errors of omission, or errors of commission could potentially lead to patient harms. The remaining essential elements are classified as I‐C (effective, expert consensus). Of note, date of birth was specifically included as a patient identifier, distinct from age, which was felt to be useful as a descriptor (often within a one‐liner or as part of the patient summary).

The 8 recommended elements were elements for which there was not unanimous agreement on inclusion, but the majority of the panel felt they should be included. These elements were classified as IIa‐C, with 1 exception. Code status generated significant controversy among the group. After extensive discussion among the group and consideration of safety, supervision, educational, and pediatric‐specific considerations, all members of the group agreed on the categorization as a recommended element; it is classified as IIb‐C.

All members of the group agreed that data elements should be directly imported from the EHR whenever possible. Finally, members agreed that the elements that make up the I‐PASS mnemonic (illness severity, patient summary, action items, situation awareness/contingency planning) should be listed in that order whenever possible. A sample I‐PASS‐compliant printed handoff document is shown Figure 1.

Figure 1

DISCUSSION

We identified substantial variability in the structure and content of printed handoff documents used by 9 pediatric hospitalist teaching services, reflective of a lack of standardization. We found that institutional printed handoff documents shared some demographic elements (eg, name, room, medical record number) but also varied in clinical content (eg, vital signs, lab tests, code status). Our expert panel developed a list of 15 essential and 8 recommended data elements for printed handoff documents. Although this is a large number of fields, the majority of the essential fields were already included by most sites, and many are basic demographic identifiers. Illness severity is the 1 essential field that was not routinely included; however, including this type of overview is consistently recommended[2, 4] and supported by evidence,[20, 21] and contributes to building a shared mental model.[16] We recommend the categories of stable/watcher/unstable.[17]

Several prior single‐center studies have found that introducing a printed handoff document can lead to improvements in workflow, communication, and patient safety. In an early study, Petersen et al.[25] showed an association between use of a computerized sign‐out program and reduced odds of preventable adverse events during periods of cross‐coverage. Wayne et al.[26] reported fewer perceived inaccuracies in handoff documents as well as improved clarity at the time of transfer, supporting the role for standardization. Van Eaton et al.[27] demonstrated rapid uptake and desirability of a computerized handoff document, which combined autoimportation of information from an EHR with resident‐entered patient details, reflecting the importance of both data sources. In addition, they demonstrated improvements in both the rounding and sign‐out processes.[28]

Two studies specifically reported the increased use of specific fields after implementation. Payne et al. implemented a Web‐based handoff tool and documented significant increases in the number of handoffs containing problem lists, medication lists, and code status, accompanied by perceived improvements in quality of handoffs and fewer near‐miss events.[24] Starmer et al. found that introduction of a resident handoff bundle that included a printed handoff tool led to reduction in medical errors and adverse events.[22] The study group using the tool populated 11 data elements more often after implementation, and introduction of this printed handoff tool in particular was associated with reductions in written handoff miscommunications. Neither of these studies included subanalysis to indicate which data elements may have been most important.

In contrast to previous single‐institution studies, our recommendations for a printed handoff template come from evaluations of tools and discussions with front line providers across 9 institutions. We had substantial overlap with data elements recommended by Van Eaton et al.[27] However, there were several areas in which we did not have overlap with published templates including weight, ins/outs, primary language, emergency contact information, or primary care provider. Other published handoff tools have been highly specialized (eg, for cardiac intensive care) or included many fewer data elements than our group felt were essential. These differences may reflect the unique aspects of caring for pediatric patients (eg, need for weights) and the absence of defined protocols for many pediatric conditions. In addition, the level of detail needed for contingency planning may vary between teaching and nonteaching services.

Resident physicians may provide valuable information in the development of standardized handoff documents. Clark et al.,[29] at Virginia Mason University, utilized resident‐driven continuous quality improvement processes including real‐time feedback to implement an electronic template. They found that engagement of both senior leaders and front‐line users was an important component of their success in uptake. Our study utilized residents as essential members of structured group interviews to ensure that front‐line users' needs were represented as recommendations for a printed handoff tool template were developed.

As previously described,[17] our study group had identified several key data elements that should be included in verbal handoffs: illness severity, a patient summary, a discrete action list, situation awareness/contingency planning, and a synthesis by receiver. With consideration of the multivoting results as well as known best practices,[1, 4, 12] the expert panel for this study agreed that each of these elements should also be highlighted in the printed template to ensure consistency between the printed document and the verbal handoff, and to have each reinforce the other. On the printed handoff tool, the final S in the I‐PASS mnemonic (synthesis by receiver) cannot be prepopulated, but considering the importance of this step,[16, 30, 31, 32] it should be printed as synthesis by receiver to serve as a text‐reminder to both givers and receivers.

The panel also felt, however, that the printed handoff document should provide additional background information not routinely included in a verbal handoff. It should serve as a reference tool both at the time of verbal handoff and throughout the day and night, and therefore should include more comprehensive information than is necessary or appropriate to convey during the verbal handoff. We identified 10 data elements that are essential in a printed handoff document in addition to the I‐PASS elements (Table 2).

Patient demographic data elements, as well as team assignments and attending physician, were uniformly supported for inclusion. The medication list was viewed as essential; however, the panel also recognized the potential for medical errors due to inaccuracies in the medication list. In particular, there was concern that including all fields of a medication order (drug, dose, route, frequency) would result in handoffs containing a high proportion of inaccurate information, particularly for complex patients whose medication regimens may vary over the course of hospitalization. Therefore, the panel agreed that if medication lists were entered manually, then only the medication name should be included as they did not wish to perpetuate inaccurate or potentially harmful information. If medication lists were autoimported from an EHR, then they should include drug name, dose, route, and frequency if possible.

In the I‐PASS study,[15] all institutions implemented printed handoff documents that included fields for the essential data elements. After implementation, there was a significant increase in completion of all essential fields. Although there is limited evidence to support any individual data element, increased usage of these elements was associated with the overall study finding of decreased rates of medical errors and preventable adverse events.

EHRs have the potential to help standardize printed handoff documents[5, 6, 33, 34, 35]; all participants in our study agreed that printed handoff documents should ideally be linked with the EHR and should autoimport data wherever appropriate. Manually populated (eg, word processor‐ or spreadsheet‐based) handoff tools have important limitations, particularly related to the potential for typographical errors as well as accidental omission of data fields, and lead to unnecessary duplication of work (eg, re‐entering data already included in a progress note) that can waste providers' time. It was also acknowledged that word processor‐ or spreadsheet‐based documents may have flexibility that is lacking in EHR‐based handoff documents. For example, formatting can more easily be adjusted to increase the number of patients per printed page. As technology advances, printed documents may be phased out in favor of EHR‐based on‐screen reports, which by their nature would be more accurate due to real‐time autoupdates.

In making recommendations about essential versus recommended items for inclusion in the printed handoff template, the only data element that generated controversy among our experts was code status. Some felt that it should be included as an essential element, whereas others did not. We believe that this was unique to our practice in pediatric hospital ward settings, as codes in most pediatric ward settings are rare. Among the concerns expressed with including code status for all patients were that residents might assume patients were full‐code without verifying. The potential inaccuracy created by this might have severe implications. Alternatively, residents might feel obligated to have code discussions with all patients regardless of severity of illness, which may be inappropriate in a pediatric population. Several educators expressed concerns about trainees having unsupervised code‐status conversations with families of pediatric patients. Conversely, although codes are rare in pediatric ward settings, concerns were raised that not including code status could be problematic during these rare but critically important events. Other fields, such as weight, might have less relevance for an adult population in which emergency drug doses are standardized.

CONCLUSIONS

In this article we describe the results of a needs assessment that informed expert consensus‐based recommendations for data elements to include in a printed handoff document. We recommend that pediatric programs include the elements identified as part of a standardized written handoff tool. Although many of these elements are also applicable to other specialties, future work should be conducted to adapt the printed handoff document elements described here for use in other specialties and settings. Future studies should work to validate the importance of these elements, studying the manner in which their inclusion affects the quality of written handoffs, and ultimately patient safety.

Despite 15 years of national and local investment in improving the safety of hospital care, patient safety remains a leading problem in both adult and pediatric hospitals. A 2010 study found that 180,000 Medicare beneficiaries likely die each year due to harm suffered as a result of medical care,[1] a death toll surpassed only by deaths due to cardiovascular disease and cancer. Even though initial efforts in the field have shown great promise for stemming the tide of healthcare‐associated infections,[2] surgical errors,[3] handoff failures,[4] and errors in the care of adults hospitalized for myocardial infarction and congestive heart failure,[5] much work remains to be done.[6] The root causes of many adverse events are poorly understood and unaddressed. Resultant tragedies remain all too common.

In the current issue of the Journal of Hospital Medicine, Bonafide and colleagues report the results of an innovative observational pilot study designed to assess the role of an inadequately addressed root cause of serious errors: alarm fatigue.[7] Alarm fatigue is the phenomenon of desensitization to alarms, particularly in the context of excessive false alarms. In a videotaped observational assessment of nurse response times to 5070 alarms on a pediatric ward and intensive care unit (ICU), the authors found that nurses responded significantly more slowly as the number of nonactionable alarms in the preceding 2 hours increased. Although a substantial majority of these alarms were technically valid (ie, representing true deviations of vital signs outside of the normal range rather than sensor or equipment problems), the vast majority required no action to be takenapproximately 7 out of 8 in the ICU and an astonishing 99 out of 100 on the ward.

As any hospitalist, intensivist, or nurse knows well, alarms are rampant throughout hospitals. It is impossible to walk down any hallway on a busy hospital wardnever mind an ICUwithout seeing a flashing light or 2 above a doorway, and hearing the incessant beeping of oxygen saturation and cardiovascular/respiratory monitors, a thousand bits of technology forever crying wolf. The problem, of course, is that sometimes there really is a wolf, but it is hard to take the risk seriously when the false alarms happen not just twice before a true threat materializes, as in Aesop's fable, but 7 times in the ICU, or worse, 99 times in the setting where most hospitalists practice. Moreover, even when the threat is real, in most cases it is caught in time one way or another, and no lasting harm results.

So why not simply shut off the unremitting noise? In 1987, outside of Baltimore, Amtrak experienced what at the time was the deadliest rail crash in its history after 1 of its passenger trains collided with a Conrail freight train. A major root cause of the crash was that the crew on the freight train had placed duct tape over an annoying automated signal alarm.[8, 9] Tragically, on this particular day, the suppressed alarm was all too relevant. Identifying the real alarm, however, can be nearly impossible when it sounds the same as 100 irritating sounds constantly emanating from the environment. It is the challenge of identifying the needle in the haystack, after you have developed an allergy to the hay.

What then to do? More research like that conducted by Bonafide and colleagues is needed to better understand how healthcare providers respond to the onslaught of alarms they encounter, and to inform refinement of these systems. Understanding how alarm fatigue plays out in the context of different clinical settings, with different workloads, varying levels of distraction, and different rates of true and false‐positive alarms will be critical. Furthermore, understanding how individuals' physiologic fatigue, circadian misalignment, mood, stress, and cognitive state may play into alarm response is likewise essential, if we are to design appropriate alarm systems that function effectively in the busy 24‐hour environment of healthcare. Ongoing work suggests that smart alarms, using algorithms that integrate data from multiple vital sign readings over time, may reduce the frequency of false alarms and better identify clinically significant events.[10] Replacing existing range‐limit monitors with these types of smart alarms has the potential to greatly improve both the sensitivity and specificity of hospital alarms, but further work in this area is needed.

Ultimately, if we can better separate out the signal, we will be better poised to respond to the true emergencies that arise that are currently obscured by the ever‐present noise. Better trust in the alarm systems we have would help all of us focus our energies on the problems that matter most. Doing so, we could better care for our patients, and better identify the system failures that cause them harm in our hospitals.

Disclosures: Dr. Landrigan is supported in part by the Children's Hospital Association for his work as an Executive Council Member of the Pediatric Research in Inpatient Settings network. Dr. Landrigan serves as a consultant to Virgin Pulse regarding sleep, safety, and health. In addition, Dr. Landrigan has received monetary awards, honoraria, and travel reimbursement from multiple academic and professional organizations for delivering lectures on sleep deprivation, physician performance, handoffs, and patient safety, and has served as an expert witness in cases regarding patient safety.

Despite 15 years of national and local investment in improving the safety of hospital care, patient safety remains a leading problem in both adult and pediatric hospitals. A 2010 study found that 180,000 Medicare beneficiaries likely die each year due to harm suffered as a result of medical care,[1] a death toll surpassed only by deaths due to cardiovascular disease and cancer. Even though initial efforts in the field have shown great promise for stemming the tide of healthcare‐associated infections,[2] surgical errors,[3] handoff failures,[4] and errors in the care of adults hospitalized for myocardial infarction and congestive heart failure,[5] much work remains to be done.[6] The root causes of many adverse events are poorly understood and unaddressed. Resultant tragedies remain all too common.

In the current issue of the Journal of Hospital Medicine, Bonafide and colleagues report the results of an innovative observational pilot study designed to assess the role of an inadequately addressed root cause of serious errors: alarm fatigue.[7] Alarm fatigue is the phenomenon of desensitization to alarms, particularly in the context of excessive false alarms. In a videotaped observational assessment of nurse response times to 5070 alarms on a pediatric ward and intensive care unit (ICU), the authors found that nurses responded significantly more slowly as the number of nonactionable alarms in the preceding 2 hours increased. Although a substantial majority of these alarms were technically valid (ie, representing true deviations of vital signs outside of the normal range rather than sensor or equipment problems), the vast majority required no action to be takenapproximately 7 out of 8 in the ICU and an astonishing 99 out of 100 on the ward.

As any hospitalist, intensivist, or nurse knows well, alarms are rampant throughout hospitals. It is impossible to walk down any hallway on a busy hospital wardnever mind an ICUwithout seeing a flashing light or 2 above a doorway, and hearing the incessant beeping of oxygen saturation and cardiovascular/respiratory monitors, a thousand bits of technology forever crying wolf. The problem, of course, is that sometimes there really is a wolf, but it is hard to take the risk seriously when the false alarms happen not just twice before a true threat materializes, as in Aesop's fable, but 7 times in the ICU, or worse, 99 times in the setting where most hospitalists practice. Moreover, even when the threat is real, in most cases it is caught in time one way or another, and no lasting harm results.

So why not simply shut off the unremitting noise? In 1987, outside of Baltimore, Amtrak experienced what at the time was the deadliest rail crash in its history after 1 of its passenger trains collided with a Conrail freight train. A major root cause of the crash was that the crew on the freight train had placed duct tape over an annoying automated signal alarm.[8, 9] Tragically, on this particular day, the suppressed alarm was all too relevant. Identifying the real alarm, however, can be nearly impossible when it sounds the same as 100 irritating sounds constantly emanating from the environment. It is the challenge of identifying the needle in the haystack, after you have developed an allergy to the hay.

What then to do? More research like that conducted by Bonafide and colleagues is needed to better understand how healthcare providers respond to the onslaught of alarms they encounter, and to inform refinement of these systems. Understanding how alarm fatigue plays out in the context of different clinical settings, with different workloads, varying levels of distraction, and different rates of true and false‐positive alarms will be critical. Furthermore, understanding how individuals' physiologic fatigue, circadian misalignment, mood, stress, and cognitive state may play into alarm response is likewise essential, if we are to design appropriate alarm systems that function effectively in the busy 24‐hour environment of healthcare. Ongoing work suggests that smart alarms, using algorithms that integrate data from multiple vital sign readings over time, may reduce the frequency of false alarms and better identify clinically significant events.[10] Replacing existing range‐limit monitors with these types of smart alarms has the potential to greatly improve both the sensitivity and specificity of hospital alarms, but further work in this area is needed.

Ultimately, if we can better separate out the signal, we will be better poised to respond to the true emergencies that arise that are currently obscured by the ever‐present noise. Better trust in the alarm systems we have would help all of us focus our energies on the problems that matter most. Doing so, we could better care for our patients, and better identify the system failures that cause them harm in our hospitals.

Disclosures: Dr. Landrigan is supported in part by the Children's Hospital Association for his work as an Executive Council Member of the Pediatric Research in Inpatient Settings network. Dr. Landrigan serves as a consultant to Virgin Pulse regarding sleep, safety, and health. In addition, Dr. Landrigan has received monetary awards, honoraria, and travel reimbursement from multiple academic and professional organizations for delivering lectures on sleep deprivation, physician performance, handoffs, and patient safety, and has served as an expert witness in cases regarding patient safety.

Despite 15 years of national and local investment in improving the safety of hospital care, patient safety remains a leading problem in both adult and pediatric hospitals. A 2010 study found that 180,000 Medicare beneficiaries likely die each year due to harm suffered as a result of medical care,[1] a death toll surpassed only by deaths due to cardiovascular disease and cancer. Even though initial efforts in the field have shown great promise for stemming the tide of healthcare‐associated infections,[2] surgical errors,[3] handoff failures,[4] and errors in the care of adults hospitalized for myocardial infarction and congestive heart failure,[5] much work remains to be done.[6] The root causes of many adverse events are poorly understood and unaddressed. Resultant tragedies remain all too common.

In the current issue of the Journal of Hospital Medicine, Bonafide and colleagues report the results of an innovative observational pilot study designed to assess the role of an inadequately addressed root cause of serious errors: alarm fatigue.[7] Alarm fatigue is the phenomenon of desensitization to alarms, particularly in the context of excessive false alarms. In a videotaped observational assessment of nurse response times to 5070 alarms on a pediatric ward and intensive care unit (ICU), the authors found that nurses responded significantly more slowly as the number of nonactionable alarms in the preceding 2 hours increased. Although a substantial majority of these alarms were technically valid (ie, representing true deviations of vital signs outside of the normal range rather than sensor or equipment problems), the vast majority required no action to be takenapproximately 7 out of 8 in the ICU and an astonishing 99 out of 100 on the ward.

As any hospitalist, intensivist, or nurse knows well, alarms are rampant throughout hospitals. It is impossible to walk down any hallway on a busy hospital wardnever mind an ICUwithout seeing a flashing light or 2 above a doorway, and hearing the incessant beeping of oxygen saturation and cardiovascular/respiratory monitors, a thousand bits of technology forever crying wolf. The problem, of course, is that sometimes there really is a wolf, but it is hard to take the risk seriously when the false alarms happen not just twice before a true threat materializes, as in Aesop's fable, but 7 times in the ICU, or worse, 99 times in the setting where most hospitalists practice. Moreover, even when the threat is real, in most cases it is caught in time one way or another, and no lasting harm results.

So why not simply shut off the unremitting noise? In 1987, outside of Baltimore, Amtrak experienced what at the time was the deadliest rail crash in its history after 1 of its passenger trains collided with a Conrail freight train. A major root cause of the crash was that the crew on the freight train had placed duct tape over an annoying automated signal alarm.[8, 9] Tragically, on this particular day, the suppressed alarm was all too relevant. Identifying the real alarm, however, can be nearly impossible when it sounds the same as 100 irritating sounds constantly emanating from the environment. It is the challenge of identifying the needle in the haystack, after you have developed an allergy to the hay.

What then to do? More research like that conducted by Bonafide and colleagues is needed to better understand how healthcare providers respond to the onslaught of alarms they encounter, and to inform refinement of these systems. Understanding how alarm fatigue plays out in the context of different clinical settings, with different workloads, varying levels of distraction, and different rates of true and false‐positive alarms will be critical. Furthermore, understanding how individuals' physiologic fatigue, circadian misalignment, mood, stress, and cognitive state may play into alarm response is likewise essential, if we are to design appropriate alarm systems that function effectively in the busy 24‐hour environment of healthcare. Ongoing work suggests that smart alarms, using algorithms that integrate data from multiple vital sign readings over time, may reduce the frequency of false alarms and better identify clinically significant events.[10] Replacing existing range‐limit monitors with these types of smart alarms has the potential to greatly improve both the sensitivity and specificity of hospital alarms, but further work in this area is needed.

Ultimately, if we can better separate out the signal, we will be better poised to respond to the true emergencies that arise that are currently obscured by the ever‐present noise. Better trust in the alarm systems we have would help all of us focus our energies on the problems that matter most. Doing so, we could better care for our patients, and better identify the system failures that cause them harm in our hospitals.

Disclosures: Dr. Landrigan is supported in part by the Children's Hospital Association for his work as an Executive Council Member of the Pediatric Research in Inpatient Settings network. Dr. Landrigan serves as a consultant to Virgin Pulse regarding sleep, safety, and health. In addition, Dr. Landrigan has received monetary awards, honoraria, and travel reimbursement from multiple academic and professional organizations for delivering lectures on sleep deprivation, physician performance, handoffs, and patient safety, and has served as an expert witness in cases regarding patient safety.

Hospital communication failures are a leading cause of serious errors and adverse events in the United States.[1, 2, 3, 4] With the implementation of duty‐hour restrictions for resident physicians,[5] there has been particular focus on the transfer of information during handoffs at change of shift.[6, 7] Many residency programs have sought to improve the processes of written and verbal handoffs through various initiatives, including: (1) automated linkage of handoff forms to electronic medical records (EMRs)[8, 9, 10]; (2) introduction of oral communication curricula, handoff simulation, or mnemonics[11, 12, 13]; and (3) faculty oversight of housestaff handoffs.[14, 15] Underlying each initiative has been the assumption that improving written and verbal handoff processes will ensure the availability of optimal patient information for on‐call housestaff. There has been little investigation, however, into what clinical questions are actually being asked of on‐call trainees, as well as what sources of information they are using to provide answers.

The aim of our study was to examine the extent to which written and verbal handoffs are utilized by pediatric trainees to derive answers to questions posed during overnight shifts. We also sought to describe both the frequency and types of on‐call questions being asked of trainees. Our primary outcome was trainee use of written handoffs to answer on‐call questions. Secondary outcomes included trainee use of verbal handoffs, as well as their use of alternative information resources to answer on‐call questions, including other clinical staff (ie, attending physicians, senior residents, nursing staff), patients and their families, the medical record, or the Internet. We then examined a variety of trainee, patient, and question characteristics to assess potential predictors of written and verbal handoff use.

METHODS

Institutional approval was granted to prospectively observe pediatric interns at the start of their overnight on‐call shifts on 2 inpatient wards at Boston Children's Hospital during 3 winter months (November through January). Our study was conducted during the postintervention period of a larger study that was designed to examine the effectiveness of a new resident handoff bundle on resident workflow and patient safety.[13] Interns rotating on study ward 1 used a structured, nonautomated tool (Microsoft Word version 2003; Microsoft Corp., Redmond, WA). Interns on study ward 2 used a handoff tool that was developed at the study hospital for use with the hospital's EMR, Cerner PowerChart version 2007.17 (Cerner Corp., Kansas City, MO). Interns on both wards received training on specific communication strategies, including verbal and written handoff processes.[13]

For our study, we recorded all questions being asked of on‐call interns by patients, parents, or other family members, as well as nurses or other clinical providers after completion of their evening handoff. We then directly observed all information resources used to derive answers to any questions asked pertaining to patients discussed in the evening handoff. We excluded any questions about new patient admissions or transfers, as well as nonpatient‐related questions.

Both study wards were staffed by separate day and night housestaff teams, who worked shifts of 12 to 14 hours in duration and had similar nursing schedules. The day team consisted of 3 interns and 1 senior resident per ward. The night team consisted of 1 intern on each ward, supervised by a senior resident covering both wards. Each day intern rotated for 1 week (Sunday through Thursday) during their month‐long ward rotation as part of the night team. We considered any intern on either of the 2 study wards to be eligible for enrollment in this study. Written consent was obtained from all participants.

The night intern received a verbal and written handoff at the shift change (usually performed between 5 and 7_pm) from 1 of the departing day interns prior to the start of the observation period. This handoff was conducted face‐to‐face in a ward conference room typically with the on‐call night intern and supervising resident receiving the handoff together from the departing day intern/senior resident.

RESULTS

Twenty‐eight observation nights (equivalent to 77 hours and 6 minutes of total direct observation time), consisting of 13 sessions on study ward 1 and 15 sessions on study ward 2, were completed. A total of 15 first‐year pediatric interns (5 male, 33%; 10 female, 66.7%), with a median age of 27.5 years (interquartile range [IQR]: 2629 years) participated. Interns on the 2 study wards were comparable with regard to trainee week of service (P=0.43) and consecutive night of call at the time of observation (P=0.45). Each intern was observed for a mean of 2 sessions (range, 13 sessions), with a mean observation time per session of approximately 2 hours and 45 minutes ( 23 minutes).

Answers

Of the 260 recorded questions, 90% (233) had documented answers. For the 10% (27) of questions with undocumented answers, 21 were observed to be verbally deferred by the intern to the day team or another care provider (ie, other physician or nurse), and almost half (42.9% [9]) involved general care‐plan questions; the remainder involved medication (4), diet (2), diagnostic testing (5), or vital sign (1) questions. An additional 6 questions went unanswered during the observation period, and it is unknown if or when they were answered.

Of the answered questions, 90% (209) of answers were provided by trainees within 5 minutes and 9% (21) within 1.5 hours. In all, interns reported using 1 information resource to provide answers for 61% (142) of questions, at least 2 resources for 33% (76) questions, and 3 resources for 6% (15) questions.

Across both study wards, interns reported using information provided in written or verbal handoffs to answer 32.6% of questions. Interns reported using the written handoff, either alone or in combination with other information resources, to provide answers for 7.3% (17) of questions; verbal handoff, either alone or in combination with another resource (excluding written handoff), was reported as a resource for 25.3% (59) of questions. Of note, interns were directly observed to look at the written handoff when answering 21% (49) of questions.

A variety of other resources, including general medical/clinical knowledge, the EMR, and parents or other resources, were used to answer the remaining 67.4% (157) of questions. Intern general medical knowledge (ie, reports of simply knowing the answer to the question in their head[s]) was used to provide answers for 53.2% (124) of questions asked.

Unadjusted univariate regression analyses assessing predictors of written and verbal handoff use are shown in Figure 1. Multivariate logistic regression analyses showed that both dietary questions (odds ratio [OR]: 3.64, 95% confidence interval [CI]: 1.518.76; P=0.004) and interns' consecutive call night (OR: 0.29, 95% CI: 0.090.93; P=0.04) remained significant predictors of written handoff use. After adjusting for risk factors identified above, no differences in written handoff use were seen between the 2 wards.

Figure 1

Multivariate logistic regression for predictors of the verbal handoff use showed that questions regarding patients with longer lengths of stay (OR: 1.97, 95% CI: 1.023.8; P=0.04), those regarding general care plans (OR: 2.07, 95% CI: 1.133.78; P=0.02), as well as those asked by clinical staff (OR: 1.95, 95 CI: 1.043.66; P=0.04), remained significant predictors of reported verbal handoff use.

DISCUSSION

In light of the recent changes in duty hours implemented in July 2011, many pediatric training programs are having trainees work in day and night shifts.[18] Pediatric resident physicians frequently answer questions that pertain to patients handed off between day and night shifts. We found that on average, information provided in the verbal and written handoff was used almost once per hour. Housestaff in our study generally based their answers on information found in 1 or 2 resources, with almost one‐third of all questions involving some use of the written or verbal handoff. Prior research has documented widespread problems with resident handoff practices across programs and a high rate of medical errors due to miscommunications.[3, 4, 19, 20] Given how often information contained within the handoff was used as interns went about their nightly tasks, it is not difficult to understand how errors or omissions in the handoff process may potentially translate into frequent problems in direct patient care.

Trainees reported using written handoff tools to provide answers for 7.3% of questions. As we had suspected, they relied less frequently on their written handoffs as they completed more consecutive call nights. Interestingly, however, even when housestaff did not report using the written handoff, they were observed quite often to look at it before providing an answer. One explanation for this discrepancy between trainee reports and our observations is that the written handoff may serve as a memory tool, even if housestaff do not directly attribute their answers to its content. Our study also found that answers to questions concerning patients' diet and fluids were more likely to be ascribed to information contained in the written handoff. This finding supports the potential value of automated written handoff tools that are linked to the EMR, which can best ensure accuracy of this type of information.

Housestaff in our study also reported using information received during the verbal handoff to answer 1 out of every 4 on‐call questions. Although we did not specifically rate or monitor the quality of verbal handoffs, prior research has demonstrated that resident verbal handoff is often plagued with incomplete and inaccurate data.[3, 4, 19, 21] One investigation found that pediatric interns were prone to overestimating the effectiveness of their verbal handoffs, even as they failed to convey urgent information to their peers.[19] In light of such prior work, our finding that interns frequently rely on the verbal transfer of information supports specific residency training program handoff initiatives that target verbal exchanges.[11, 22, 23]

Although information obtained in the handoff was frequently required by on‐call housestaff, our study found that two‐thirds of all questions were answered using other resources, most often general medical or clinical knowledge. Clearly, background knowledge and experience is fundamental to trainees' ability to perform their jobs. Such reliance on general knowledge for problem solving may not be unique to interns. One recent observational study of senior pediatric cardiac subspecialists reported a high frequency of reliance on their own clinical experience, instinct, or prior training in making clinical decisions.[24] Further investigation may be useful to parse out the exact types of clinical knowledge being used, and may have important implications for how training programs plan for overnight supervision.[25, 26, 27]

Our study has several limitations. First, it was beyond the scope of this study to link housestaff answers to patient outcomes or medical errors. Given the frequency with which the handoff, a known source of vulnerability to medical error, was used by on‐call housestaff, our study suggests that future research evaluating the relationship between questions asked of on‐call housestaff, the answers provided, and downstream patient safety incidents may be merited. Second, our study was conducted in a single pediatric residency program with 1 physician observer midway through the first year of training and only in the early evening hours. This limits the generalizability of our findings, as the use of handoffs to answer on‐call questions may be different at other stages of the training process, within other specialties, or even at different times of the day. We also began our observations after the handoff had taken place; future studies may want to assess how variations in written and verbal handoff processes affect their use. As a final limitation, we note that although collecting information in real time using a direct observational method eliminated the problem of recall bias, there may have been attribution bias.

The results of our study demonstrate that on‐call pediatric housestaff are frequently asked a variety of clinical questions posed by hospital staff, patients, and their families. We found that trainees are apt to rely both on handoff information and other resources to provide answers. By better understanding what resources on‐call housestaff are accessing to answer questions overnight, we may be able to better target interventions needed to improve the availability of patient information, as well as the usefulness of written and verbal handoff tools.[11, 22, 23]

Hospital communication failures are a leading cause of serious errors and adverse events in the United States.[1, 2, 3, 4] With the implementation of duty‐hour restrictions for resident physicians,[5] there has been particular focus on the transfer of information during handoffs at change of shift.[6, 7] Many residency programs have sought to improve the processes of written and verbal handoffs through various initiatives, including: (1) automated linkage of handoff forms to electronic medical records (EMRs)[8, 9, 10]; (2) introduction of oral communication curricula, handoff simulation, or mnemonics[11, 12, 13]; and (3) faculty oversight of housestaff handoffs.[14, 15] Underlying each initiative has been the assumption that improving written and verbal handoff processes will ensure the availability of optimal patient information for on‐call housestaff. There has been little investigation, however, into what clinical questions are actually being asked of on‐call trainees, as well as what sources of information they are using to provide answers.

The aim of our study was to examine the extent to which written and verbal handoffs are utilized by pediatric trainees to derive answers to questions posed during overnight shifts. We also sought to describe both the frequency and types of on‐call questions being asked of trainees. Our primary outcome was trainee use of written handoffs to answer on‐call questions. Secondary outcomes included trainee use of verbal handoffs, as well as their use of alternative information resources to answer on‐call questions, including other clinical staff (ie, attending physicians, senior residents, nursing staff), patients and their families, the medical record, or the Internet. We then examined a variety of trainee, patient, and question characteristics to assess potential predictors of written and verbal handoff use.

METHODS

Institutional approval was granted to prospectively observe pediatric interns at the start of their overnight on‐call shifts on 2 inpatient wards at Boston Children's Hospital during 3 winter months (November through January). Our study was conducted during the postintervention period of a larger study that was designed to examine the effectiveness of a new resident handoff bundle on resident workflow and patient safety.[13] Interns rotating on study ward 1 used a structured, nonautomated tool (Microsoft Word version 2003; Microsoft Corp., Redmond, WA). Interns on study ward 2 used a handoff tool that was developed at the study hospital for use with the hospital's EMR, Cerner PowerChart version 2007.17 (Cerner Corp., Kansas City, MO). Interns on both wards received training on specific communication strategies, including verbal and written handoff processes.[13]

For our study, we recorded all questions being asked of on‐call interns by patients, parents, or other family members, as well as nurses or other clinical providers after completion of their evening handoff. We then directly observed all information resources used to derive answers to any questions asked pertaining to patients discussed in the evening handoff. We excluded any questions about new patient admissions or transfers, as well as nonpatient‐related questions.

Both study wards were staffed by separate day and night housestaff teams, who worked shifts of 12 to 14 hours in duration and had similar nursing schedules. The day team consisted of 3 interns and 1 senior resident per ward. The night team consisted of 1 intern on each ward, supervised by a senior resident covering both wards. Each day intern rotated for 1 week (Sunday through Thursday) during their month‐long ward rotation as part of the night team. We considered any intern on either of the 2 study wards to be eligible for enrollment in this study. Written consent was obtained from all participants.

The night intern received a verbal and written handoff at the shift change (usually performed between 5 and 7_pm) from 1 of the departing day interns prior to the start of the observation period. This handoff was conducted face‐to‐face in a ward conference room typically with the on‐call night intern and supervising resident receiving the handoff together from the departing day intern/senior resident.

RESULTS

Twenty‐eight observation nights (equivalent to 77 hours and 6 minutes of total direct observation time), consisting of 13 sessions on study ward 1 and 15 sessions on study ward 2, were completed. A total of 15 first‐year pediatric interns (5 male, 33%; 10 female, 66.7%), with a median age of 27.5 years (interquartile range [IQR]: 2629 years) participated. Interns on the 2 study wards were comparable with regard to trainee week of service (P=0.43) and consecutive night of call at the time of observation (P=0.45). Each intern was observed for a mean of 2 sessions (range, 13 sessions), with a mean observation time per session of approximately 2 hours and 45 minutes ( 23 minutes).

Answers

Of the 260 recorded questions, 90% (233) had documented answers. For the 10% (27) of questions with undocumented answers, 21 were observed to be verbally deferred by the intern to the day team or another care provider (ie, other physician or nurse), and almost half (42.9% [9]) involved general care‐plan questions; the remainder involved medication (4), diet (2), diagnostic testing (5), or vital sign (1) questions. An additional 6 questions went unanswered during the observation period, and it is unknown if or when they were answered.

Of the answered questions, 90% (209) of answers were provided by trainees within 5 minutes and 9% (21) within 1.5 hours. In all, interns reported using 1 information resource to provide answers for 61% (142) of questions, at least 2 resources for 33% (76) questions, and 3 resources for 6% (15) questions.

Across both study wards, interns reported using information provided in written or verbal handoffs to answer 32.6% of questions. Interns reported using the written handoff, either alone or in combination with other information resources, to provide answers for 7.3% (17) of questions; verbal handoff, either alone or in combination with another resource (excluding written handoff), was reported as a resource for 25.3% (59) of questions. Of note, interns were directly observed to look at the written handoff when answering 21% (49) of questions.

A variety of other resources, including general medical/clinical knowledge, the EMR, and parents or other resources, were used to answer the remaining 67.4% (157) of questions. Intern general medical knowledge (ie, reports of simply knowing the answer to the question in their head[s]) was used to provide answers for 53.2% (124) of questions asked.

Unadjusted univariate regression analyses assessing predictors of written and verbal handoff use are shown in Figure 1. Multivariate logistic regression analyses showed that both dietary questions (odds ratio [OR]: 3.64, 95% confidence interval [CI]: 1.518.76; P=0.004) and interns' consecutive call night (OR: 0.29, 95% CI: 0.090.93; P=0.04) remained significant predictors of written handoff use. After adjusting for risk factors identified above, no differences in written handoff use were seen between the 2 wards.

Figure 1

Multivariate logistic regression for predictors of the verbal handoff use showed that questions regarding patients with longer lengths of stay (OR: 1.97, 95% CI: 1.023.8; P=0.04), those regarding general care plans (OR: 2.07, 95% CI: 1.133.78; P=0.02), as well as those asked by clinical staff (OR: 1.95, 95 CI: 1.043.66; P=0.04), remained significant predictors of reported verbal handoff use.

DISCUSSION

In light of the recent changes in duty hours implemented in July 2011, many pediatric training programs are having trainees work in day and night shifts.[18] Pediatric resident physicians frequently answer questions that pertain to patients handed off between day and night shifts. We found that on average, information provided in the verbal and written handoff was used almost once per hour. Housestaff in our study generally based their answers on information found in 1 or 2 resources, with almost one‐third of all questions involving some use of the written or verbal handoff. Prior research has documented widespread problems with resident handoff practices across programs and a high rate of medical errors due to miscommunications.[3, 4, 19, 20] Given how often information contained within the handoff was used as interns went about their nightly tasks, it is not difficult to understand how errors or omissions in the handoff process may potentially translate into frequent problems in direct patient care.

Trainees reported using written handoff tools to provide answers for 7.3% of questions. As we had suspected, they relied less frequently on their written handoffs as they completed more consecutive call nights. Interestingly, however, even when housestaff did not report using the written handoff, they were observed quite often to look at it before providing an answer. One explanation for this discrepancy between trainee reports and our observations is that the written handoff may serve as a memory tool, even if housestaff do not directly attribute their answers to its content. Our study also found that answers to questions concerning patients' diet and fluids were more likely to be ascribed to information contained in the written handoff. This finding supports the potential value of automated written handoff tools that are linked to the EMR, which can best ensure accuracy of this type of information.

Housestaff in our study also reported using information received during the verbal handoff to answer 1 out of every 4 on‐call questions. Although we did not specifically rate or monitor the quality of verbal handoffs, prior research has demonstrated that resident verbal handoff is often plagued with incomplete and inaccurate data.[3, 4, 19, 21] One investigation found that pediatric interns were prone to overestimating the effectiveness of their verbal handoffs, even as they failed to convey urgent information to their peers.[19] In light of such prior work, our finding that interns frequently rely on the verbal transfer of information supports specific residency training program handoff initiatives that target verbal exchanges.[11, 22, 23]

Although information obtained in the handoff was frequently required by on‐call housestaff, our study found that two‐thirds of all questions were answered using other resources, most often general medical or clinical knowledge. Clearly, background knowledge and experience is fundamental to trainees' ability to perform their jobs. Such reliance on general knowledge for problem solving may not be unique to interns. One recent observational study of senior pediatric cardiac subspecialists reported a high frequency of reliance on their own clinical experience, instinct, or prior training in making clinical decisions.[24] Further investigation may be useful to parse out the exact types of clinical knowledge being used, and may have important implications for how training programs plan for overnight supervision.[25, 26, 27]

Our study has several limitations. First, it was beyond the scope of this study to link housestaff answers to patient outcomes or medical errors. Given the frequency with which the handoff, a known source of vulnerability to medical error, was used by on‐call housestaff, our study suggests that future research evaluating the relationship between questions asked of on‐call housestaff, the answers provided, and downstream patient safety incidents may be merited. Second, our study was conducted in a single pediatric residency program with 1 physician observer midway through the first year of training and only in the early evening hours. This limits the generalizability of our findings, as the use of handoffs to answer on‐call questions may be different at other stages of the training process, within other specialties, or even at different times of the day. We also began our observations after the handoff had taken place; future studies may want to assess how variations in written and verbal handoff processes affect their use. As a final limitation, we note that although collecting information in real time using a direct observational method eliminated the problem of recall bias, there may have been attribution bias.

The results of our study demonstrate that on‐call pediatric housestaff are frequently asked a variety of clinical questions posed by hospital staff, patients, and their families. We found that trainees are apt to rely both on handoff information and other resources to provide answers. By better understanding what resources on‐call housestaff are accessing to answer questions overnight, we may be able to better target interventions needed to improve the availability of patient information, as well as the usefulness of written and verbal handoff tools.[11, 22, 23]

Hospital communication failures are a leading cause of serious errors and adverse events in the United States.[1, 2, 3, 4] With the implementation of duty‐hour restrictions for resident physicians,[5] there has been particular focus on the transfer of information during handoffs at change of shift.[6, 7] Many residency programs have sought to improve the processes of written and verbal handoffs through various initiatives, including: (1) automated linkage of handoff forms to electronic medical records (EMRs)[8, 9, 10]; (2) introduction of oral communication curricula, handoff simulation, or mnemonics[11, 12, 13]; and (3) faculty oversight of housestaff handoffs.[14, 15] Underlying each initiative has been the assumption that improving written and verbal handoff processes will ensure the availability of optimal patient information for on‐call housestaff. There has been little investigation, however, into what clinical questions are actually being asked of on‐call trainees, as well as what sources of information they are using to provide answers.

The aim of our study was to examine the extent to which written and verbal handoffs are utilized by pediatric trainees to derive answers to questions posed during overnight shifts. We also sought to describe both the frequency and types of on‐call questions being asked of trainees. Our primary outcome was trainee use of written handoffs to answer on‐call questions. Secondary outcomes included trainee use of verbal handoffs, as well as their use of alternative information resources to answer on‐call questions, including other clinical staff (ie, attending physicians, senior residents, nursing staff), patients and their families, the medical record, or the Internet. We then examined a variety of trainee, patient, and question characteristics to assess potential predictors of written and verbal handoff use.

METHODS

Institutional approval was granted to prospectively observe pediatric interns at the start of their overnight on‐call shifts on 2 inpatient wards at Boston Children's Hospital during 3 winter months (November through January). Our study was conducted during the postintervention period of a larger study that was designed to examine the effectiveness of a new resident handoff bundle on resident workflow and patient safety.[13] Interns rotating on study ward 1 used a structured, nonautomated tool (Microsoft Word version 2003; Microsoft Corp., Redmond, WA). Interns on study ward 2 used a handoff tool that was developed at the study hospital for use with the hospital's EMR, Cerner PowerChart version 2007.17 (Cerner Corp., Kansas City, MO). Interns on both wards received training on specific communication strategies, including verbal and written handoff processes.[13]

For our study, we recorded all questions being asked of on‐call interns by patients, parents, or other family members, as well as nurses or other clinical providers after completion of their evening handoff. We then directly observed all information resources used to derive answers to any questions asked pertaining to patients discussed in the evening handoff. We excluded any questions about new patient admissions or transfers, as well as nonpatient‐related questions.

Both study wards were staffed by separate day and night housestaff teams, who worked shifts of 12 to 14 hours in duration and had similar nursing schedules. The day team consisted of 3 interns and 1 senior resident per ward. The night team consisted of 1 intern on each ward, supervised by a senior resident covering both wards. Each day intern rotated for 1 week (Sunday through Thursday) during their month‐long ward rotation as part of the night team. We considered any intern on either of the 2 study wards to be eligible for enrollment in this study. Written consent was obtained from all participants.

The night intern received a verbal and written handoff at the shift change (usually performed between 5 and 7_pm) from 1 of the departing day interns prior to the start of the observation period. This handoff was conducted face‐to‐face in a ward conference room typically with the on‐call night intern and supervising resident receiving the handoff together from the departing day intern/senior resident.

RESULTS

Twenty‐eight observation nights (equivalent to 77 hours and 6 minutes of total direct observation time), consisting of 13 sessions on study ward 1 and 15 sessions on study ward 2, were completed. A total of 15 first‐year pediatric interns (5 male, 33%; 10 female, 66.7%), with a median age of 27.5 years (interquartile range [IQR]: 2629 years) participated. Interns on the 2 study wards were comparable with regard to trainee week of service (P=0.43) and consecutive night of call at the time of observation (P=0.45). Each intern was observed for a mean of 2 sessions (range, 13 sessions), with a mean observation time per session of approximately 2 hours and 45 minutes ( 23 minutes).

Answers

Of the 260 recorded questions, 90% (233) had documented answers. For the 10% (27) of questions with undocumented answers, 21 were observed to be verbally deferred by the intern to the day team or another care provider (ie, other physician or nurse), and almost half (42.9% [9]) involved general care‐plan questions; the remainder involved medication (4), diet (2), diagnostic testing (5), or vital sign (1) questions. An additional 6 questions went unanswered during the observation period, and it is unknown if or when they were answered.

Of the answered questions, 90% (209) of answers were provided by trainees within 5 minutes and 9% (21) within 1.5 hours. In all, interns reported using 1 information resource to provide answers for 61% (142) of questions, at least 2 resources for 33% (76) questions, and 3 resources for 6% (15) questions.

Across both study wards, interns reported using information provided in written or verbal handoffs to answer 32.6% of questions. Interns reported using the written handoff, either alone or in combination with other information resources, to provide answers for 7.3% (17) of questions; verbal handoff, either alone or in combination with another resource (excluding written handoff), was reported as a resource for 25.3% (59) of questions. Of note, interns were directly observed to look at the written handoff when answering 21% (49) of questions.

A variety of other resources, including general medical/clinical knowledge, the EMR, and parents or other resources, were used to answer the remaining 67.4% (157) of questions. Intern general medical knowledge (ie, reports of simply knowing the answer to the question in their head[s]) was used to provide answers for 53.2% (124) of questions asked.

Unadjusted univariate regression analyses assessing predictors of written and verbal handoff use are shown in Figure 1. Multivariate logistic regression analyses showed that both dietary questions (odds ratio [OR]: 3.64, 95% confidence interval [CI]: 1.518.76; P=0.004) and interns' consecutive call night (OR: 0.29, 95% CI: 0.090.93; P=0.04) remained significant predictors of written handoff use. After adjusting for risk factors identified above, no differences in written handoff use were seen between the 2 wards.

Figure 1

Multivariate logistic regression for predictors of the verbal handoff use showed that questions regarding patients with longer lengths of stay (OR: 1.97, 95% CI: 1.023.8; P=0.04), those regarding general care plans (OR: 2.07, 95% CI: 1.133.78; P=0.02), as well as those asked by clinical staff (OR: 1.95, 95 CI: 1.043.66; P=0.04), remained significant predictors of reported verbal handoff use.

DISCUSSION

In light of the recent changes in duty hours implemented in July 2011, many pediatric training programs are having trainees work in day and night shifts.[18] Pediatric resident physicians frequently answer questions that pertain to patients handed off between day and night shifts. We found that on average, information provided in the verbal and written handoff was used almost once per hour. Housestaff in our study generally based their answers on information found in 1 or 2 resources, with almost one‐third of all questions involving some use of the written or verbal handoff. Prior research has documented widespread problems with resident handoff practices across programs and a high rate of medical errors due to miscommunications.[3, 4, 19, 20] Given how often information contained within the handoff was used as interns went about their nightly tasks, it is not difficult to understand how errors or omissions in the handoff process may potentially translate into frequent problems in direct patient care.

Trainees reported using written handoff tools to provide answers for 7.3% of questions. As we had suspected, they relied less frequently on their written handoffs as they completed more consecutive call nights. Interestingly, however, even when housestaff did not report using the written handoff, they were observed quite often to look at it before providing an answer. One explanation for this discrepancy between trainee reports and our observations is that the written handoff may serve as a memory tool, even if housestaff do not directly attribute their answers to its content. Our study also found that answers to questions concerning patients' diet and fluids were more likely to be ascribed to information contained in the written handoff. This finding supports the potential value of automated written handoff tools that are linked to the EMR, which can best ensure accuracy of this type of information.

Housestaff in our study also reported using information received during the verbal handoff to answer 1 out of every 4 on‐call questions. Although we did not specifically rate or monitor the quality of verbal handoffs, prior research has demonstrated that resident verbal handoff is often plagued with incomplete and inaccurate data.[3, 4, 19, 21] One investigation found that pediatric interns were prone to overestimating the effectiveness of their verbal handoffs, even as they failed to convey urgent information to their peers.[19] In light of such prior work, our finding that interns frequently rely on the verbal transfer of information supports specific residency training program handoff initiatives that target verbal exchanges.[11, 22, 23]

Although information obtained in the handoff was frequently required by on‐call housestaff, our study found that two‐thirds of all questions were answered using other resources, most often general medical or clinical knowledge. Clearly, background knowledge and experience is fundamental to trainees' ability to perform their jobs. Such reliance on general knowledge for problem solving may not be unique to interns. One recent observational study of senior pediatric cardiac subspecialists reported a high frequency of reliance on their own clinical experience, instinct, or prior training in making clinical decisions.[24] Further investigation may be useful to parse out the exact types of clinical knowledge being used, and may have important implications for how training programs plan for overnight supervision.[25, 26, 27]

Our study has several limitations. First, it was beyond the scope of this study to link housestaff answers to patient outcomes or medical errors. Given the frequency with which the handoff, a known source of vulnerability to medical error, was used by on‐call housestaff, our study suggests that future research evaluating the relationship between questions asked of on‐call housestaff, the answers provided, and downstream patient safety incidents may be merited. Second, our study was conducted in a single pediatric residency program with 1 physician observer midway through the first year of training and only in the early evening hours. This limits the generalizability of our findings, as the use of handoffs to answer on‐call questions may be different at other stages of the training process, within other specialties, or even at different times of the day. We also began our observations after the handoff had taken place; future studies may want to assess how variations in written and verbal handoff processes affect their use. As a final limitation, we note that although collecting information in real time using a direct observational method eliminated the problem of recall bias, there may have been attribution bias.

The results of our study demonstrate that on‐call pediatric housestaff are frequently asked a variety of clinical questions posed by hospital staff, patients, and their families. We found that trainees are apt to rely both on handoff information and other resources to provide answers. By better understanding what resources on‐call housestaff are accessing to answer questions overnight, we may be able to better target interventions needed to improve the availability of patient information, as well as the usefulness of written and verbal handoff tools.[11, 22, 23]

Since the term hospitalist was coined in 1996,1 the field of hospital medicine has grown exponentially. Hospitalists are caring for increasing numbers of adultsincluding Medicare beneficiaries in hospitals across the United States.2 Pediatric hospital medicine has grown in parallel. By 1998, 50% of pediatric department chairs across the US and Canada had implemented hospitalist programs, with another 27% reporting they were soon to do so.3 A bit more than a decade later, pediatric hospitalists can be found in nearly every major academic medical center, and in a large proportion of community hospitals throughout the US and Canada.

In the past several years, major advances have begun to occur in the manner in which hospital medicine research is conducted. In this article, we will describe the manner in which pediatric hospital medicine research has advanced over the past several years, culminating in the conduct of several large multicenter research projects through the Pediatric Research in Inpatient Settings (PRIS) Network. We believe that lessons learned in the development of PRIS could help foster the growth of other current and future networks of hospitalist researchers, and lay the groundwork for national improvement efforts.

HOSPITAL MEDICINE RESEARCH: GROWTH AND DEVELOPMENT

In 2001, a small group of thought leaders in pediatric hospital medicine (see Acknowledgements) conceived the notion of starting a hospitalist research network, which they named the Pediatric Research in Inpatient Settings (PRIS) Network.4 PRIS was modeled in part after a successful pediatric primary care network.5 Since hospitalists in institutions across the country were being tasked to improve the care of hospitalized patients, and to lead diverse quality and safety initiatives, why not create a network to facilitate identification of high priority problems and evidence‐based approaches to them, and coordinate improvement efforts? The ambitious goal of the fledgling network was to conduct transformative research into inpatient healthcare delivery and discover both condition‐dependent and condition‐independent processes of care that were linked to patient outcomes.

PRIS began as (and remains) an open research networkfrom the outset, any hospitalist could join. The notion of this network, even in its earliest stages, was sufficiently appealing to professional societies that the Society of Hospital Medicine (SHM), the Academic Pediatric Association (APA), and the American Academy of Pediatrics (AAP) agreed to cosponsor the network, fostering its early growth. The community of pediatric hospitalists was tremendously supportive as well; over 300 hospitalists initially signed up to participate. Initial studies were generated through surveys of members, through which variability in systemic organization and variation in the management of clinical conditions and systems‐based issues across inpatient settings was identified and quantified.68

In the 2000s, as PRIS grew as a network, the research capacity of individuals within the field also grew. An increasing number of hospitalists began dedicating their academic careers to pursuing rigorous methodological training and conducting pediatric hospital medicine research. A series of studies began to emerge analyzing data from large administrative datasets that described the variation in hospital care (but lack clinical results and clinical outcomes outside of the hospital setting), such as the Pediatric Health Information Systems (PHIS) database operated by the Children's Hospital Association (formerly known as the Child Health Corporation of America).913 Pediatric hospital medicine fellowships began to appear,14 and over time, a cohort of hospitalist investigators with sufficient independence to mentor others arose.

THE REDESIGN OF PRIS

In 2009, a Pediatric Hospital Medicine Roundtable of 22 international leaders was convened under the guidance of SHM, APA, and AAP.15 This initiative, roughly a decade after the inception of the field, was critical to bringing pediatric hospitalist research and PRIS to the next level. It was recognized in that meeting that while PRIS had made a good start, it would not be possible to grow the network to the point of conducting top quality multicenter studies without the active involvement of a larger number of rigorously trained hospitalist researchers. To stimulate the network's growth, the existing PRIS Steering Committeea diverse group of clinical, educational, administrative, and research leaders in the fieldfacilitated the transfer of leadership to a new Executive Council led entirely by trained researchers (see Table 1), with the support of the APA. The Executive Council subsequently developed a series of standard operating procedures (see Table 2) that have created a transparent process to deal with important, but often difficult, academic issues that networks face.

DEVELOPMENT OF MULTICENTER RESEARCH PROJECTS

The redesign of PRIS did not alter its objective: to build the evidence base regarding the optimal inpatient management of children. Evidence on how best to care for many pediatric conditions remains lacking, largely due to the facts that: a) death, the most definitive and readily measured of outcomes, is rare in pediatric hospitals; b) many pediatric conditions are relatively uncommon in any single hospital; and c) few validated, well‐developed metrics of inpatient pediatric quality exist.

As PRIS sought to launch multicenter studies of inpatient care quality, it continued to receive strong support from the APA, SHM, and AAP, and gained the support of a new partner, the Children's Hospital Association, which is comprised of a large group of children's hospitals across Canada and the US. The membership of PRIS grew to involve over 600 pediatric hospitalists from more than 75 hospitals.4 With a core group of funded hospitalist investigators, and strong support from partner organizations, the network sought and received funding for 3 major studies that are currently underway. Release of the federal government's Affordable Care Act and Comparative Effectiveness Research portfolio stimulated much of this work, stimulating the network to reach out to existing and new stakeholders and successfully compete for several multicenter studies.

RELEVANCE TO OTHER NETWORKS

We believe that the story of PRIS' development, current studies, and future plans has relevance to other adult, as well as pediatric, hospital medicine networks (see Table 3). As in pediatrics, a growing group of midcareer adult hospital medicine investigators has emerged, with proven track records in attracting federal funding and conducting research germane to our field. Some have previously worked together on large‐scale multisite studies.2023 A core group have come together to form the HOspital MEdicine Reengineering Network (HOMERUN).24 HOMERUN has recently secured funding from the Association of American Medical Colleges (AAMC) for a project that is linking clinical data from several hospitals to a centralized database, a project analogous to PHIS+, and will allow for Comparative Effectiveness Research studies that have more accurate case ascertainment (by using clinical data to build cohorts) and ensuring additional power by securing a larger number of cases. Defining which clinical questions to address first will help establish this new entity as a leader in hospital medicine research. Attracting stakeholder involvement will help make these endeavors successful. In recent months, PRIS and HOMERUN jointly collaborated on the submission of a large Centers for Medicare and Medicaid Innovation (CMMI) proposal to extend the work of I‐PASS to include several internal medicine and additional pediatric resident and hospitalist care settings. Future collaborations between networks may help foster more rapid advances in care.

Another pediatric hospitalist network has also emerged in the past few years, with a focus on quality improvement across inpatient pediatric settings, the Value in Pediatrics (VIP) Network.25 Although still early in its development, VIP has already successfully engaged in national quality improvement work regarding benchmarking care provided for children with bronchiolitis, reducing patient identification (ID) band errors, and improving discharge communications. VIP recently became part of the AAP's Quality Improvement Innovation Network (QuINN) group through which it is receiving infrastructure support.

As they develop, hospital medicine research and improvement networks will seek to systematically design and rigorously execute multicenter projects that provide answers to those clinical questions which practicing hospitalists face on a daily basis. As they do so, mentoring of both junior investigators and novice investigators will be necessary for the longevity of networks. To foster junior investigators, PRIS has undertaken a series of workshops presented at various national conferences, in addition to working with junior investigators directly on its currently funded studies.

CONCLUSION

Hospitalists' engagement in research and quality improvement networks builds upon their already successful engagement in clinical care, education, and quality improvement at a local level. A research and improvement mission that is tightly coupled with the day‐to‐day needs of these other important hospitalist activities creates a synergy with the potential to lead to transformative advances in patient care. If hospitalists can discover how best to deliver care, train the next generation of providers, and work to implement needed improvements at a local and national level, they will have an unprecedented opportunity to improve the care and health of children and adults.

Since the term hospitalist was coined in 1996,1 the field of hospital medicine has grown exponentially. Hospitalists are caring for increasing numbers of adultsincluding Medicare beneficiaries in hospitals across the United States.2 Pediatric hospital medicine has grown in parallel. By 1998, 50% of pediatric department chairs across the US and Canada had implemented hospitalist programs, with another 27% reporting they were soon to do so.3 A bit more than a decade later, pediatric hospitalists can be found in nearly every major academic medical center, and in a large proportion of community hospitals throughout the US and Canada.

In the past several years, major advances have begun to occur in the manner in which hospital medicine research is conducted. In this article, we will describe the manner in which pediatric hospital medicine research has advanced over the past several years, culminating in the conduct of several large multicenter research projects through the Pediatric Research in Inpatient Settings (PRIS) Network. We believe that lessons learned in the development of PRIS could help foster the growth of other current and future networks of hospitalist researchers, and lay the groundwork for national improvement efforts.

HOSPITAL MEDICINE RESEARCH: GROWTH AND DEVELOPMENT

In 2001, a small group of thought leaders in pediatric hospital medicine (see Acknowledgements) conceived the notion of starting a hospitalist research network, which they named the Pediatric Research in Inpatient Settings (PRIS) Network.4 PRIS was modeled in part after a successful pediatric primary care network.5 Since hospitalists in institutions across the country were being tasked to improve the care of hospitalized patients, and to lead diverse quality and safety initiatives, why not create a network to facilitate identification of high priority problems and evidence‐based approaches to them, and coordinate improvement efforts? The ambitious goal of the fledgling network was to conduct transformative research into inpatient healthcare delivery and discover both condition‐dependent and condition‐independent processes of care that were linked to patient outcomes.

PRIS began as (and remains) an open research networkfrom the outset, any hospitalist could join. The notion of this network, even in its earliest stages, was sufficiently appealing to professional societies that the Society of Hospital Medicine (SHM), the Academic Pediatric Association (APA), and the American Academy of Pediatrics (AAP) agreed to cosponsor the network, fostering its early growth. The community of pediatric hospitalists was tremendously supportive as well; over 300 hospitalists initially signed up to participate. Initial studies were generated through surveys of members, through which variability in systemic organization and variation in the management of clinical conditions and systems‐based issues across inpatient settings was identified and quantified.68

In the 2000s, as PRIS grew as a network, the research capacity of individuals within the field also grew. An increasing number of hospitalists began dedicating their academic careers to pursuing rigorous methodological training and conducting pediatric hospital medicine research. A series of studies began to emerge analyzing data from large administrative datasets that described the variation in hospital care (but lack clinical results and clinical outcomes outside of the hospital setting), such as the Pediatric Health Information Systems (PHIS) database operated by the Children's Hospital Association (formerly known as the Child Health Corporation of America).913 Pediatric hospital medicine fellowships began to appear,14 and over time, a cohort of hospitalist investigators with sufficient independence to mentor others arose.

THE REDESIGN OF PRIS

In 2009, a Pediatric Hospital Medicine Roundtable of 22 international leaders was convened under the guidance of SHM, APA, and AAP.15 This initiative, roughly a decade after the inception of the field, was critical to bringing pediatric hospitalist research and PRIS to the next level. It was recognized in that meeting that while PRIS had made a good start, it would not be possible to grow the network to the point of conducting top quality multicenter studies without the active involvement of a larger number of rigorously trained hospitalist researchers. To stimulate the network's growth, the existing PRIS Steering Committeea diverse group of clinical, educational, administrative, and research leaders in the fieldfacilitated the transfer of leadership to a new Executive Council led entirely by trained researchers (see Table 1), with the support of the APA. The Executive Council subsequently developed a series of standard operating procedures (see Table 2) that have created a transparent process to deal with important, but often difficult, academic issues that networks face.

DEVELOPMENT OF MULTICENTER RESEARCH PROJECTS

The redesign of PRIS did not alter its objective: to build the evidence base regarding the optimal inpatient management of children. Evidence on how best to care for many pediatric conditions remains lacking, largely due to the facts that: a) death, the most definitive and readily measured of outcomes, is rare in pediatric hospitals; b) many pediatric conditions are relatively uncommon in any single hospital; and c) few validated, well‐developed metrics of inpatient pediatric quality exist.

As PRIS sought to launch multicenter studies of inpatient care quality, it continued to receive strong support from the APA, SHM, and AAP, and gained the support of a new partner, the Children's Hospital Association, which is comprised of a large group of children's hospitals across Canada and the US. The membership of PRIS grew to involve over 600 pediatric hospitalists from more than 75 hospitals.4 With a core group of funded hospitalist investigators, and strong support from partner organizations, the network sought and received funding for 3 major studies that are currently underway. Release of the federal government's Affordable Care Act and Comparative Effectiveness Research portfolio stimulated much of this work, stimulating the network to reach out to existing and new stakeholders and successfully compete for several multicenter studies.

RELEVANCE TO OTHER NETWORKS

We believe that the story of PRIS' development, current studies, and future plans has relevance to other adult, as well as pediatric, hospital medicine networks (see Table 3). As in pediatrics, a growing group of midcareer adult hospital medicine investigators has emerged, with proven track records in attracting federal funding and conducting research germane to our field. Some have previously worked together on large‐scale multisite studies.2023 A core group have come together to form the HOspital MEdicine Reengineering Network (HOMERUN).24 HOMERUN has recently secured funding from the Association of American Medical Colleges (AAMC) for a project that is linking clinical data from several hospitals to a centralized database, a project analogous to PHIS+, and will allow for Comparative Effectiveness Research studies that have more accurate case ascertainment (by using clinical data to build cohorts) and ensuring additional power by securing a larger number of cases. Defining which clinical questions to address first will help establish this new entity as a leader in hospital medicine research. Attracting stakeholder involvement will help make these endeavors successful. In recent months, PRIS and HOMERUN jointly collaborated on the submission of a large Centers for Medicare and Medicaid Innovation (CMMI) proposal to extend the work of I‐PASS to include several internal medicine and additional pediatric resident and hospitalist care settings. Future collaborations between networks may help foster more rapid advances in care.

Another pediatric hospitalist network has also emerged in the past few years, with a focus on quality improvement across inpatient pediatric settings, the Value in Pediatrics (VIP) Network.25 Although still early in its development, VIP has already successfully engaged in national quality improvement work regarding benchmarking care provided for children with bronchiolitis, reducing patient identification (ID) band errors, and improving discharge communications. VIP recently became part of the AAP's Quality Improvement Innovation Network (QuINN) group through which it is receiving infrastructure support.

As they develop, hospital medicine research and improvement networks will seek to systematically design and rigorously execute multicenter projects that provide answers to those clinical questions which practicing hospitalists face on a daily basis. As they do so, mentoring of both junior investigators and novice investigators will be necessary for the longevity of networks. To foster junior investigators, PRIS has undertaken a series of workshops presented at various national conferences, in addition to working with junior investigators directly on its currently funded studies.

CONCLUSION

Hospitalists' engagement in research and quality improvement networks builds upon their already successful engagement in clinical care, education, and quality improvement at a local level. A research and improvement mission that is tightly coupled with the day‐to‐day needs of these other important hospitalist activities creates a synergy with the potential to lead to transformative advances in patient care. If hospitalists can discover how best to deliver care, train the next generation of providers, and work to implement needed improvements at a local and national level, they will have an unprecedented opportunity to improve the care and health of children and adults.

Since the term hospitalist was coined in 1996,1 the field of hospital medicine has grown exponentially. Hospitalists are caring for increasing numbers of adultsincluding Medicare beneficiaries in hospitals across the United States.2 Pediatric hospital medicine has grown in parallel. By 1998, 50% of pediatric department chairs across the US and Canada had implemented hospitalist programs, with another 27% reporting they were soon to do so.3 A bit more than a decade later, pediatric hospitalists can be found in nearly every major academic medical center, and in a large proportion of community hospitals throughout the US and Canada.

In the past several years, major advances have begun to occur in the manner in which hospital medicine research is conducted. In this article, we will describe the manner in which pediatric hospital medicine research has advanced over the past several years, culminating in the conduct of several large multicenter research projects through the Pediatric Research in Inpatient Settings (PRIS) Network. We believe that lessons learned in the development of PRIS could help foster the growth of other current and future networks of hospitalist researchers, and lay the groundwork for national improvement efforts.

HOSPITAL MEDICINE RESEARCH: GROWTH AND DEVELOPMENT

In 2001, a small group of thought leaders in pediatric hospital medicine (see Acknowledgements) conceived the notion of starting a hospitalist research network, which they named the Pediatric Research in Inpatient Settings (PRIS) Network.4 PRIS was modeled in part after a successful pediatric primary care network.5 Since hospitalists in institutions across the country were being tasked to improve the care of hospitalized patients, and to lead diverse quality and safety initiatives, why not create a network to facilitate identification of high priority problems and evidence‐based approaches to them, and coordinate improvement efforts? The ambitious goal of the fledgling network was to conduct transformative research into inpatient healthcare delivery and discover both condition‐dependent and condition‐independent processes of care that were linked to patient outcomes.

PRIS began as (and remains) an open research networkfrom the outset, any hospitalist could join. The notion of this network, even in its earliest stages, was sufficiently appealing to professional societies that the Society of Hospital Medicine (SHM), the Academic Pediatric Association (APA), and the American Academy of Pediatrics (AAP) agreed to cosponsor the network, fostering its early growth. The community of pediatric hospitalists was tremendously supportive as well; over 300 hospitalists initially signed up to participate. Initial studies were generated through surveys of members, through which variability in systemic organization and variation in the management of clinical conditions and systems‐based issues across inpatient settings was identified and quantified.68

In the 2000s, as PRIS grew as a network, the research capacity of individuals within the field also grew. An increasing number of hospitalists began dedicating their academic careers to pursuing rigorous methodological training and conducting pediatric hospital medicine research. A series of studies began to emerge analyzing data from large administrative datasets that described the variation in hospital care (but lack clinical results and clinical outcomes outside of the hospital setting), such as the Pediatric Health Information Systems (PHIS) database operated by the Children's Hospital Association (formerly known as the Child Health Corporation of America).913 Pediatric hospital medicine fellowships began to appear,14 and over time, a cohort of hospitalist investigators with sufficient independence to mentor others arose.

THE REDESIGN OF PRIS

In 2009, a Pediatric Hospital Medicine Roundtable of 22 international leaders was convened under the guidance of SHM, APA, and AAP.15 This initiative, roughly a decade after the inception of the field, was critical to bringing pediatric hospitalist research and PRIS to the next level. It was recognized in that meeting that while PRIS had made a good start, it would not be possible to grow the network to the point of conducting top quality multicenter studies without the active involvement of a larger number of rigorously trained hospitalist researchers. To stimulate the network's growth, the existing PRIS Steering Committeea diverse group of clinical, educational, administrative, and research leaders in the fieldfacilitated the transfer of leadership to a new Executive Council led entirely by trained researchers (see Table 1), with the support of the APA. The Executive Council subsequently developed a series of standard operating procedures (see Table 2) that have created a transparent process to deal with important, but often difficult, academic issues that networks face.

DEVELOPMENT OF MULTICENTER RESEARCH PROJECTS

The redesign of PRIS did not alter its objective: to build the evidence base regarding the optimal inpatient management of children. Evidence on how best to care for many pediatric conditions remains lacking, largely due to the facts that: a) death, the most definitive and readily measured of outcomes, is rare in pediatric hospitals; b) many pediatric conditions are relatively uncommon in any single hospital; and c) few validated, well‐developed metrics of inpatient pediatric quality exist.

As PRIS sought to launch multicenter studies of inpatient care quality, it continued to receive strong support from the APA, SHM, and AAP, and gained the support of a new partner, the Children's Hospital Association, which is comprised of a large group of children's hospitals across Canada and the US. The membership of PRIS grew to involve over 600 pediatric hospitalists from more than 75 hospitals.4 With a core group of funded hospitalist investigators, and strong support from partner organizations, the network sought and received funding for 3 major studies that are currently underway. Release of the federal government's Affordable Care Act and Comparative Effectiveness Research portfolio stimulated much of this work, stimulating the network to reach out to existing and new stakeholders and successfully compete for several multicenter studies.

RELEVANCE TO OTHER NETWORKS

We believe that the story of PRIS' development, current studies, and future plans has relevance to other adult, as well as pediatric, hospital medicine networks (see Table 3). As in pediatrics, a growing group of midcareer adult hospital medicine investigators has emerged, with proven track records in attracting federal funding and conducting research germane to our field. Some have previously worked together on large‐scale multisite studies.2023 A core group have come together to form the HOspital MEdicine Reengineering Network (HOMERUN).24 HOMERUN has recently secured funding from the Association of American Medical Colleges (AAMC) for a project that is linking clinical data from several hospitals to a centralized database, a project analogous to PHIS+, and will allow for Comparative Effectiveness Research studies that have more accurate case ascertainment (by using clinical data to build cohorts) and ensuring additional power by securing a larger number of cases. Defining which clinical questions to address first will help establish this new entity as a leader in hospital medicine research. Attracting stakeholder involvement will help make these endeavors successful. In recent months, PRIS and HOMERUN jointly collaborated on the submission of a large Centers for Medicare and Medicaid Innovation (CMMI) proposal to extend the work of I‐PASS to include several internal medicine and additional pediatric resident and hospitalist care settings. Future collaborations between networks may help foster more rapid advances in care.

Another pediatric hospitalist network has also emerged in the past few years, with a focus on quality improvement across inpatient pediatric settings, the Value in Pediatrics (VIP) Network.25 Although still early in its development, VIP has already successfully engaged in national quality improvement work regarding benchmarking care provided for children with bronchiolitis, reducing patient identification (ID) band errors, and improving discharge communications. VIP recently became part of the AAP's Quality Improvement Innovation Network (QuINN) group through which it is receiving infrastructure support.

As they develop, hospital medicine research and improvement networks will seek to systematically design and rigorously execute multicenter projects that provide answers to those clinical questions which practicing hospitalists face on a daily basis. As they do so, mentoring of both junior investigators and novice investigators will be necessary for the longevity of networks. To foster junior investigators, PRIS has undertaken a series of workshops presented at various national conferences, in addition to working with junior investigators directly on its currently funded studies.

CONCLUSION

Hospitalists' engagement in research and quality improvement networks builds upon their already successful engagement in clinical care, education, and quality improvement at a local level. A research and improvement mission that is tightly coupled with the day‐to‐day needs of these other important hospitalist activities creates a synergy with the potential to lead to transformative advances in patient care. If hospitalists can discover how best to deliver care, train the next generation of providers, and work to implement needed improvements at a local and national level, they will have an unprecedented opportunity to improve the care and health of children and adults.