Patient Concerns at Hospital Admission

Patient-centered care, communication, and patient satisfaction are increasingly important metrics of hospital care.¹ A key aspect of patient-centered communication is eliciting and addressing patients' concerns.² Outpatient research indicates that when concerns are addressed, patients report higher satisfaction.³ To our knowledge, hospitalized patients' concerns have not been studied.

We surveyed patients before and after the attending physician hospital admission encounter to describe the number and topics of concerns and the degree to which physicians addressed them. The study was conducted between August 2008 and March 2009 on the general medical services at 2 hospitals within a university system where attending physicians care for patients with and without trainees. Participants were attending hospitalist physicians and patients admitted under their care who were able to give informed consent and communicate verbally in English. Eligible patients were approached before meeting the physician and, if they agreed to participate, asked to list “all of the problems and concerns you want to talk with the doctor about today.”⁴ During the encounter, the study coordinator waited outside the patient's room to measure encounter length. After the encounter, patients rated how well each of their pre-encounter concerns was addressed: “not at all,” “somewhat,” “mostly,” or “completely.”

In quantitative content analysis,⁵ we iteratively developed a code book to describe themes within the topics of patients' concerns. The final code book included 11 conceptual categories. Two-coder agreement on a 20% sample of concerns was 92% on at least 1 category and 79% on all categories. We assessed associations among whether concerns were addressed, number of pre-encounter concerns, and encounter length using logistic regression (Stata version 11; StataCorp, College Station, Texas). The institutional review board at the University of California, San Francisco, approved the study; participants gave written informed consent.

We enrolled 109 patients (consent rate, 65%; mean [SD] age, 54 [19] years; 44% male; 6% Hispanic; 8% Asian; and 8% African American) of 30 physicians (consent rate, 91%; mean [SD] age, 35 [5] years; 43% male; 3% Hispanic; and 30% Asian).

Patients listed a median of 2 pre-encounter concerns (range, 0-10). Ninety-five (87%) patients listed at least 1 concern. While 77 (71%) reported multiple concerns, only 30 (28%) reported more than 3. Concerns related to patients' hospital care and the ongoing care of hospital admission diagnoses (Table). Most frequently, concerns regarded treatments, including medications, procedures, therapies, and adverse effects; diagnoses, including known diagnoses and desire to obtain a diagnosis or cause of illness; and logistics, including facilities, communication, and coordination of care.

Eighty-five patients completed the post-encounter survey, of which 76 listed at least 1 pre-encounter concern. Thirty-eight (50%) reported at least 1 “somewhat” or “not at all” addressed concern. Only 27 patients (36%) reported all concerns “completely” addressed. Many patients with few pre-encounter concerns reported a “somewhat” or “not at all” addressed concern: 27% of patients with 1 concern, 49% of patients with 2 to 3, and 68% of patients with more than 3. In an unadjusted model, patients with more than 3 pre-encounter concerns were more likely to report a “somewhat” or “not at all” addressed concern (odds ratio, 5.9; 95% confidence interval, 1.1-32.6) compared with patients with 1; an adjusted model revealed similar results. Mean encounter length was 21 minutes (range, 3-68 minutes). Reporting a “somewhat” or “not at all” addressed concern was not associated with encounter length in unadjusted or adjusted analyses.

As when making outpatient visits, patients being admitted to the hospital have preexisting concerns that they hope physicians will address. Exploring concerns is an opportunity for physicians to ensure understanding of and adherence to the care plan. Unaddressed concerns may lead to lower satisfaction as well as lower quality medical care because clinically relevant symptoms, questions, or treatment barriers are not disclosed.^3,4

Yet, similar to outpatient studies,⁴ many patients' few relevant concerns were incompletely addressed. That encounters were not longer for patients who reported their concerns addressed suggests that how physicians and patients communicate influences addressing of concerns. In outpatient studies physicians infrequently elicit all of patients' concerns and interrupt patients before they finish describing them.^4,6,7 Methods of addressing concerns include prompting patients to identify concerns before an encounter⁸ and physicians eliciting all concerns and agenda setting at the beginning of encounters.^2,6

There were several limitations: we did not account for the effect of other health care providers on patients' responses, studied only 1 encounter, and studied a small number of patients at 2 hospitals within the same academic system. Sources of bias include (1) increased concern disclosure because of pre-encounter listing, (2) physicians exhibiting best communication behaviors secondary to study involvement, and (3) administering fewer post-encounter surveys on physicians' busiest days as a result of relying on physicians to notify us of encounters' occurrence. We would expect these factors to bias toward overestimating the frequency of concerns being addressed.

In conclusion, research, education, and quality improvement efforts should focus on eliciting and addressing hospitalized patients' concerns.

Correspondence: Dr Anderson, Division of Hospital Medicine, Department of Medicine, UCSF, 521 Parnassus Ave, Box 0903, San Francisco, CA 94143 (anderson.wg@gmail.com).

Author Contributions: Dr Anderson had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Anderson and Auerbach. Acquisition of data: Winters. Analysis and interpretation of data: Anderson, Winters, and Auerbach. Drafting of the manuscript: Anderson and Winters. Critical revision of the manuscript for important intellectual content: Anderson and Auerbach. Obtained funding: Anderson. Administrative, technical, and material support: Winters. Study supervision: Auerbach.

Funding/Support: Dr Anderson was funded by the National Palliative Care Research Center and the UCSF Clinical and Translational Science Institute Career Development Program, National Institutes of Health grant 5 KL2 RR024130-04. Project costs were funded by a grant from the UCSF Academic Senate.

Additional Contributions: We thank the patients and physicians who generously donated their time to participate. Eric Vittinghoff, PhD, MPH, UCSF Department of Epidemiology and Biostatistics, provided guidance in the selection and execution of statistical analyses.