Would you admit your mother to the residency service? Introducing the JCHIMP resident safety column.

There remain tremendous opportunities to improve the stability and safety of American health care. Within this context, residents and residency programs face two essential questions: how to reduce the risk to patients resulting from resident inexperience, and how to change our programs to create the safer physician of the future? The spread of side-by-side teaching and non-teaching services creates a natural setting to study these questions and improve both services. When asked the question, "Would you admit your mother to the resident service?", many of us respond, "It depends". We are focusing this column on helping programs answer this question definitively in the positive, share potential best practices, and underscore community hospital's contribution to our understanding of patient safety.

It is 8 p.m. on a Saturday in late July. You are just home from the gym, when you notice the voice mail light blinking on your phone. It is a call from your local community hospital emergency department (ED). You are not on call, and they neither use your paging service nor leave a patient name. They want you to call back as soon as possible. It feels ominously personal.

You return the call. A friend picks up the extension and explains that your mother is alright, but that she had developed cough and chills that afternoon. After being unable to reach you, she came to the ED. She is now showing signs of an early lobar infiltrate; she is non-toxic, but probably needs to be admitted. He wants to know if you have a preference for admitting her to the teaching or non-teaching service.

Would you admit your mother to the residents? It is a bothersome dilemma for a community hospital residency program director. The residents at our hospital are quite good. But when it really matters, would our residents provide an equal or better service than an experienced hospitalist? And in what way could they be better: better choice of antibiotics, closer monitoring, effective warmth and reassurance, less chance of error, a lower risk of death? Traditionally, our answer for all these would have been the return question – ‘Who's on call for both services?’

In the chaotic world of medicine in which I trained in, before the 2000 Institute of Medicine report, ‘To Err is Human’ (1), the best answer to uncertainty in medicine was to place one's faith in an individual physician. Those of us privy to a clinical reputation and real outcomes know that all doctors are not equal and that having the right doctor on the spot might make all the difference. I know which of my intern and resident pairings will do a better job. I also know which hospitalists seem to have a higher rate of patients upgrading to intensive care or ending up on quality review. Just picking the best admission service is only the beginning. There are many opportunities for error after admission. With shorter shifts and more sign-outs on both services, how do I ensure that when my mother needs the ‘right doctor’, she has him or her? How do I know that the ‘right doctor’ is getting a really good sign-out – can I trust that they will know my Mom gets red man syndrome when vancomycin is pushed too fast and that she will get delirious with zolpidem and that her left arm is a much better site for an IV line? The usual clinician family member will then conclude that it does not matter which service admits their mother because they will dedicate themself to being the shadow attending, nurse, pharmacist, or transport tech, thereby ensuring safety for every task.

We have started this column to explore questions like this, and to share best practices in teaching the science of safety to our trainees. Community hospital trainee programs, JCHIMP's core audience, have a special role to play in improving patient safety. It may not be initially obvious to the reader that this is true. However, in the safety literature, several themes have emerged as keys to decreasing patient harm. These themes are standardization of work around evidence-based medicine, reliability of the team in producing the same outcome, and a culture of safety where everyone searches for defects and responds to them rapidly (2, 3). Significant progress has been made identifying best practices that reduce the impact of frequent problems such as line infections, falls, and medication reconciliation (4). More general techniques around team building (5), communication, and disciplined quality improvement have also been widely applied. These advances have had less success when focused on the work of physicians, especially internal medicine physicians. Techniques for preventing diagnostic errors (6), effectively working with diverse clinical teams on different units, and preventing less well-defined or uncommon errors have not been developed. Smaller, less complex institutions and training programs have a potential for cohesiveness that places these goals within reach, more quickly than may be possible in more complex university programs. Moreover, training programs themselves have the potential to progress more rapidly than their surrounding institutions.

House staff teams are designed to expect errors. We already measure and correct performance. In comparison to experienced attendings or specialists, residents have receptivity and flexibility several orders of magnitude greater. Moreover, once properly trained, they can sometimes reinforce and sustain change peer to peer. Besides being malleable, the residents are superb observers of the system. Working in every department and on every shift, and having to respond to every system breakdown or error, they notice every one. If asked, they have superb insight and creativity on how to improve the system.

So why do we fear July? Why am I concerned about admitting my mother to their care? There are many reasons to believe that the residents might be dangerous. Intern skills are immature, they miscommunicate, and they do not know what they do not know. They are confronted with novel challenges and perform complex multistep tasks, both of which increase error rates (7). Their supervising residents are more experienced clinically but novices in delegation and management. Attendings rely on data which is filtered through these less reliable interpreters, or if personally cross-checking, with a time delay. The attendings suffer from the same heuristic errors and lapses as all physicians.

Although supervision of residents has improved over the past decade, there is tremendous opportunity to study error epidemiology and improve systems of care. We can better understand the contribution of dysfunctional systems that predispose them to error. We can move from curricula about safety and begin to change the way residents work.

What is possible if we fully engage the principles of a safe culture? What if we take advantage of our teams to create oases of reliable care? What if our residents are essential to successfully transforming our institutions? Can we overcome the inherent risk of a resident's inexperience by leveraging and expanding our systems of teamwork? Can our grief over needing more sign-outs become opportunities to improve care through thoughtful cross-checking?

Some early data are encouraging. Our hospital, which is pretty typical of many small community hospitals, has been experimenting with these goals. The residents have found over the past 10 years that they have reduced mortality for patients admitted to their floor service by 70% (0.91–0.37%) (see Chart 1) and that resident mortality has trended lower than hospital mortality. There is no clear specific intervention resulting in this change. We still have all the usual trainee struggles. Neither is there a clear explanatory system change which applies only to the residents or decrease in patient severity on their service. As a retrospective cohort study, full accounting for potential confounders is difficult. Potential contributors to the change include orientation simulations of typical safety events and an orientation focus on learning safe behaviors. A dedicated senior resident daily collects and investigates safety events. Interns and residents devote a weekly conference to reviewing events for opportunities to improve practices and communication. They are finding 15–20 events per week. The residents have initiated multiple iterative changes to their practices to anticipate and reduce common errors (see Chart 2). Our central lesson is that a comprehensive safety program which engages residents benefits patients, hospital systems, and resident education. Imagine what might be possible with continued effort.

Chart 1

Timeline for residency safety project.

Chart 2

Comparison of resident and hospital mortality.

JCHIMP is dedicated to providing community hospital practitioners an opportunity to contribute to the medical and scientific discussion, and improvement of graduate medical education. We have an opportunity and responsibility to explore work that many universities will find much more challenging. This column will help emphasize community program work in improving safety. I look forward to hearing from readers about what is working or not working well in their hospitals. I also look forward to hearing that mothers everywhere are preferentially being admitted to teaching services.

Date	Intervention (italics are by residents)	Description
Mar-05	Electronic sign-out tool	Server-based structured sign-out tool designed to emphasize safety issues.
Mar-06	Weekly safety rounds	45 min morning report dedicated to reviewing safety events. Key features: anonymous, focused on event prevention and management, communication strategies. Cash prize for resident with most number of safety reports. Regular lectures on safety science.
Oct-07	Paper transfer orders	In response to multiple errors at transfer, paper orders standardized to force physicians to review existing orders.
Jun-08	Orientation safety simulations	Residents and nurses present OSCE like scenarios at end of intern orientation. Scenarios chosen to emphasize cases where an intern must act in the first 10 min while awaiting arrival of support. e.g.: chest pain, shock, hyperkalemia, mental status change, shortness of breath, agitated patient, HIPAA breach.
Oct-08	RRT: Rapid response team	Implementation of house wide rapid response team for any significant vital sign change or concern raised by staff member, patient or family member.
Jan-09	Major hand-washing campaign	Nursing leaders, executives, physicians sign public pledge to wash hands.
Oct-09	EMR	Institution begins implementation of electronic documentation.
Dec-09	Safety event electronic reporting	Institution implements Maryland safe patient electronic tool for tracking safety events.
Feb-10	Standard sign-out with I-SAFE	Interactive- Situation- Active issues- Follow-up tasks- Expected events to watch for. Sign-out process redesigned to emphasize standardization, protected time, and interactive process.
Jun-10	Rounding checklist	Development of a specific checklist for team to use during intern presentations. Checklist managed by team safety officer for the day (pre-call intern).
Sep-10	Procedure checklist	Electronic checklist created for all invasive procedures – designed to have observing intern watch for safety events. List includes internet link to video of correct procedure.
Oct-10	CPOE	Computer physician order entry roll-out begins.
Mar-11	Medication reconciliation standardization	Redesign of admission reconciliation to distinguish high-risk patients, acquire at least 3 prescription lists for high-risk patients, and indicate in admission note when the medication reconciliation was complete. Development of our CPOE ‘10 commandments’ – a list of the 10 most common errors created by CPOE.
Sep-11	Rigorous event investigation	Institutional standardization of investigation of safety events with root cause analysis for all serious events, apparent cause analysis for significant near misses, task force committees to follow frequent events. Significant progress made with line sepsis, catheter associated urinary tract infections, and falls.
Mar-12	Safety assessment in admission note	Medication reconciliation and safety risk assessment built into admission note, progress note and sign-out.
Jan-13	Team-based analysis	Use of cognitive error tools by rounding team members to cross-check thinking of primary intern: using a pathophysiologic assessment, a worst-case scenario assessment, and a tool designed to uncover treatment dilemmas.
Jun-13	ICU order read-back	In response to multiple CPOE errors, ICU interns began trial of having nurses read-back orders entered into the computer to verify identity, correct dosing, and communicate plan.
Sep-13	Order review	Completion of order review implemented. This is a check box on the admission note showing that the admitting team completed a cross check of intern orders looking for med reconciliation errors, or mistakes in diagnostic and treatment plan.