Research Letters FREE

Patients in teaching hospitals may be confused regarding who is involved in their care and by their respective roles in their care. Physicians in training may contribute to this problem through the failure to introduce themselves and their role to patients.¹ Resident duty-hour limits and the use of nonteaching hospitalists may also contribute to this problem owing to increased care fragmentation. In addition to system factors, certain patients may be less likely to identify their hospital physician because of various socioeconomic and demographic factors. Regardless of the cause, patients who are unable to identify their physicians may be less informed about their hospitalization and not know to whom to direct questions, impairing their ability to make informed decisions both during and after hospitalization. To empower patients in this setting, several organizations advocate that hospitalized patients stay informed about the technical aspects of their care, such as who is in charge of their care.^2- 4

Despite these concerns, little is known regarding the effects of a variety of system-level and patient-level factors on hospitalized patients' ability to identify their inpatient physicians. In addition, the impact of this knowledge on patient satisfaction is not known.

We conducted a prospective longitudinal survey of adult patients admitted to the inpatient general medicine service of the University of Chicago from July 2005 to October 2006. Patients were assigned to care by a resident teaching service, composed of 1 attending physician, 1 resident physician, 1 intern or subintern (fourth-year medical student), and up to 1 third-year medical student.⁵ After the resident teaching team capped at 10 admissions or midnight (whichever came first), patients were admitted by a night-float resident. The next morning, the patients admitted by the night-float resident were handed off to either a nonteaching hospitalist day service or the resident teaching service admitting that day. The institutional review board of the University of Chicago approved the study protocol, and written consent was obtained from all participants. Patients who were cognitively impaired, defined as a score of less than 17 of 22 on a modified mini-mental status examination, were excluded.⁶

During an inpatient interview, hospitalized general medicine patients⁵ were asked to rate their understanding of the roles of physicians caring for them in the hospital using a 5-point Likert scale (1, poor understanding, to 5, excellent understanding). Patients were then asked to name the physicians on their inpatient team by stating, “As best you can, please name the physicians and trainees on the general medicine team caring for you in the hospital.” Patients' responses were then transcribed as spoken by research assistants. Patients were also asked to report their race and educational level. Age, use of a night-float physician, teaching service admission, emergency department admission, and length of stay variables were obtained by abstracting medical records. Patient satisfaction was obtained using questions from the Picker-Commonwealth survey (eg, overall satisfaction, satisfaction with care received by doctors, trust in their physicians) obtained through 1-month postdischarge telephone surveys.^7- 8

All persons named by patients were checked against the patients' medical record (outpatient electronic record and inpatient paper record) to identify the named person as one of their inpatient physicians or trainees (attending physician, resident, intern, or medical student), primary care physician, subspecialist, nurse, or other person. Descriptive statistics were used to summarize patients’ ability to identify their inpatient physicians. χ² Tests and multivariate logistic regression analysis were used to test the effect of demographic factors (eg, race and age) and system factors (eg, admitted by float team, teaching service) on patient ability to identify their inpatient physicians. The relationship between the ability to identify inpatient physician and patient satisfaction was also explored. Because most patients are satisfied with their care, analyses were undertaken using dichotomized satisfaction variables denoting low levels of satisfaction.

Between July 2005 and October 2006, 2807 patients (64%) were enrolled in this study. Table 1 describes the demographic characteristics of patients. Of 2807 patients, 2110 (75%) were not able to name anyone when asked to identify an inpatient physician in charge of their care. For the 697 patients who responded with at least 1 name, only 281 (40%) were able to correctly identify 1 inpatient physician on their team. Attending physicians and interns were more frequently identified than residents. For those patients who did not correctly identify a physician in charge of their care, patients often named specialists (n = 258 [37%]) or their primary care physician (n = 52 [7%]). Interestingly, of the 2807 patients, the majority (n = 1582; 56%) of patients rated their understanding of the roles of the physicians on their team as “very good” or “excellent.” Those patients who expressed a greater understanding of physician roles were more likely to name a physician and correctly identify an inpatient physician (65% of patients understood role vs 55% of patients who did not understand role; P = .005).

In univariate analysis, both patient factors and system-level factors were associated with decreased ability to identify an inpatient physician. Patients who were African American, had an educational level below that of a high school graduate, were older, and were unmarried were less likely to identify their inpatient physician. In addition, patients admitted by a night-float team, admitted through the emergency department, cared for by a nonteaching hospitalist service, or not cared for by a University of Chicago primary care physician were less likely to identify an inpatient physician (Table 2). In multivariate analysis, controlling for other predictors and time trends, several of these relationships persisted for both patient factors (age, race, and education) and hospital stay factors (nonteaching hospitalist service, admission from emergency department, primary care physician not at University of Chicago).

Of the 2807 patients interviewed, 1901 (68%) completed a follow-up interview 1 month after discharge regarding their patient satisfaction. Overall satisfaction with physicians was high (64% very satisfied). Interestingly, in both univariate and multivariate analyses, patients who could correctly name their inpatient physician were more likely to be unsatisfied with the care of their physicians and to be less trusting of their physicians.

Only a small fraction of patients could successfully identify at least 1 physician or trainee on their inpatient team in charge of their hospital care. Both patient factors (race, age, and lower educational level) and system factors (nonteaching service and admitted by a float team) were associated with a diminished likelihood of correctly identifying an inpatient physician. In addition, those patients who identified their physicians correctly were less satisfied with their care by physicians. While this may at first appear counterintuitive, this finding may represent the intention to file a grievance or potentially pursue legal action after the hospitalization. In other words, patients may be more likely to note their physician’s name if they are concerned or skeptical regarding their care.

This study highlights the variety of reasons that patients lack understanding of who is in charge of their care and the importance of interventions to target certain patient populations or take into account the system of care. For example, in today's teaching hospitals, there are more handoffs, making it less likely that a patient would interact with the same physician for an extended period. This suggests the importance of interventions for patients who have multiple handovers during their admission, especially with the increasing use of night-float and nonteaching hospitalist services in academic centers. In addition, interventions can be designed and implemented specifically for those patients who are particularly vulnerable, such as elderly patients or those without a high degree of education.

At least 1 study has suggested that placing physician photographs in the patients' rooms may help them to identify physicians.⁹ At the same time, patients should be encouraged to take specific measures to stay informed about who is involved in their care and what the responsibility level of each caregiver is. Going 1 step further, the Lewis Blackman Hospital Patient Safety Act,¹⁰ recently enacted in South Carolina, mandates that all hospital personnel wear appropriate name tags that identify their name and their role to patients and that patients are provided with education on how to immediately contact the attending physician in charge of their care.

There are several limitations of this study. This single institution study may not be generalizable to other settings, such as community hospitals. Loss to follow-up raises concerns of selection bias. Fortunately, patients who were lost to follow-up did not significantly differ by their ability to correctly identify their inpatient physician. Patients lost to follow-up were, however, more likely to be African American, male, hospitalized longer, hospitalized in the last year, and without a primary care physician at the University of Chicago, highlighting the difficulty in following up this group of patients. Third, patients may not have understood the questions or misinterpreted the question to refer to their primary care physician. Because of this, during pilot testing of earlier questions, we added the phrase “caring for you in the hospital.”

Despite these limitations, the majority of hospitalized patients are unable to name someone in charge of their care. This suggests that academic hospitals should focus on improving the ability of patients to understand the names and roles of their inpatient physicians.

Correspondence: Dr Arora, Department of Medicine, University of Chicago, 5841 S Maryland Ave, MC 2007, AMB W216, Chicago, IL 60637 (varora@medicine.bsd.uchicago.edu).

Author Contributions: The authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Arora and Meltzer. Acquisition of data: Arora, Gangireddy, Mehrotra, Ginde, and Meltzer. Analysis and interpretation of data: Arora, Gangireddy, Mehrotra, Ginde, and Meltzer. Drafting of the manuscript: Arora and Gangireddy. Critical revision of the manuscript for important intellectual content: Arora, Gangireddy, Mehrotra, Ginde, and Meltzer. Statistical analysis: Arora, Gangireddy, and Meltzer. Obtained funding: Arora and Meltzer. Administrative, technical, and material support: Arora, Mehrotra, Ginde, and Meltzer. Study supervision: Arora and Meltzer.

Financial Disclosure: None reported.

Funding/Support: This study received funding from the University of Chicago Hospitals, the National Institutes of General Medical Sciences, and the Donald W. Reynolds Foundation.

Additional Contributions: Jennifer Higa, BA, and Korry Schwanz, BA, assisted in research.