Trio of studies tackles trends affecting pediatric residents
AT PEDIATRIC HOSPITAL MEDICINE 2013
NEW ORLEANS – Inpatient staffing in the wake of tighter residency work hours, avoiding premature diagnostic closure, and hospitalist perceptions of resident effects on quality of care and costs were tackled in a trio of oral abstracts presented at Pediatric Hospital Medicine 2013.
Work-rules effects
A Web-based survey of 152 pediatric residency programs in the United States revealed marked variation in how programs have responded after the Accreditation Council for Graduate Medical Education mandated new residency work hour restrictions in July 2011.
The most common changes were the addition of a resident night-float system and an increase in the number of attending physicians present at night, primarily in the form of pediatric hospitalists and pediatric intensivists, said Dr. Jennifer Oshimura, with Indiana University, Riley Hospital for Children, Indianapolis.
The number of programs using a night-float system increased from 43% to 71% after the new residency work hours, while the number of programs utilizing resident admission caps increased only slightly, from 12% to 15%.
In all, 23% of programs increased night in-house attending coverage following the work-hour restrictions.
Of these, 63% of programs increased their night-time attending coverage by adding pediatric hospitalists, 37% added pediatric intensivists, and 26% added neonatologists, Dr. Oshimura said.
Overall, the number of programs with pediatric hospitalist attendings in-house 24/7 rose from 16% to 20%, with 41% of programs planning to add this kind of coverage within the next 5 years.
Finally, only 12% of residency programs had no in-house attending coverage at night of any kind. This compares with 22% of programs in a previous study conducted by the group in 2010, before the work hour restrictions were in place, she noted.
The anonymous survey was sent to 198 pediatric residency training programs in the United States. The response rate was 77% (152/198), representing 7,828 pediatric residents or 79% of all pediatric residents in the United States.
The programs varied "drastically" in terms of size (average, 51.5 residents; range, 6-168) and average daily patient census per ward intern during the daytime (7.4 patients; range, 2-30), Dr. Oshimura remarked.
Premature closure
Results of a randomized trial show that simulated patient encounters can be used to help teach residents to avoid premature closure, or the tendency to accept a diagnosis before it is fully verified. The phenomenon prevents consideration of alternative diagnoses and is a common source of error in the care of inpatients.
"Premature closure can be summarized with the axiom: ‘When the diagnosis is made, the thinking stops,’ " Dr. Lauren Nassetta said at the meeting, sponsored by the Society of Hospital Medicine, the American Academy of Pediatrics, and the Academic Pediatric Association.
Dr. Nassetta and her colleagues at the University of Alabama at Birmingham randomly assigned 61 residents on inpatient pediatrics rotations to receive a text page with a brief description of a patient that contained either an incorrect diagnosis (stem A) or a symptom (stem B). The remainder of the case was carefully scripted and presented identically. Each of the 36 teams received 2 cases, and the cases were paired such that residents who received stem A for case 1 received stem B for case 2, and vice versa.
The cases involved common pediatric presentations for unusual reasons, such as a 14-month-old who presents with wheezing, secondary to congestive heart failure. For this case, the incorrect diagnosis was given as asthma.
The residents had 10 minutes to evaluate (history, physical exam, laboratory, and radiology) and treat simulated patients. High-fidelity mannequins were used to provide responses to the residents, while investigators played the role of nurse and family member.
Residents obtained significantly more elements in the evaluation when they received the symptom stem versus the incorrect diagnosis stem (63% vs. 55%; P less than .05), Dr. Nassetta said.
Residents given the symptom stem made the correct diagnosis 86% of the time, compared with only 69% of the time when given the incorrect diagnosis. The difference did not reach statistical significance at the group level (P = .16), but was significant on an individual participant level.
Still, the finding resulted in a relative risk of 2.2, "meaning that having the incorrect diagnosis made it twice as likely to miss the correct diagnosis," she said.
When data from individual participants were used in a multivariable logistic regression model, the adjusted odds ratio was 14.9 for making a correct diagnosis with the symptom stem vs. the incorrect diagnosis. The model adjusted for a variety of variables, including prior simulation time, case difficulty, level of training, program type, and confidence in independent care of patients.
