Abstract
Improving patient satisfaction is a major focus of hospitals. Patient satisfaction could be driven by patient perception of hospital team communication with their primary care physician (PCP). A retrospective mixed methods approach was used to characterize the relationship between patient satisfaction and patient perception of hospital team–PCP communication. Data were obtained through general medicine inpatient and postdischarge interviews, oversampling “vulnerable elders,” and a faxed PCP survey. Among 1044 patients and their PCPs, 22.3% of PCPs were not aware of their patient’s hospitalization. Among PCPs who reported that communication did not occur, half (49.2%) of their patients thought communication had occurred, implying a lack of patient awareness of discontinuity of care and possibly impeding safety. Patients who perceived that communication occurred were more satisfied with care (70.0% vs 53.1%, P < .001). Therefore, hospitals could potentially improve patient safety and satisfaction by seizing a missed opportunity to improve patient awareness of communication.
Improving patient experience and satisfaction is a major focus of hospitals. In 2006, the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) Survey was implemented nationally by the Centers for Medicare & Medicaid Services and the Agency for Healthcare Research and Quality to publicly report patients’ perspectives on inpatient care. 1 Among its 21 substantive questions, the HCAHPS survey requests 2 global ratings: an overall hospital rating and willingness to recommend. These global ratings are largely perceived as reflective of patient satisfaction, a subject that has been brought to the forefront of hospitals’ attention in recent years, in part because of the publication of data such as the patient care experience items included in the HCAHPS survey. This has led many hospitals to identify improving patient satisfaction as one of their key quality improvement opportunities. One factor that could drive satisfaction is patient perception of continuity of care and communication between their hospital team and primary care physician (PCP).
Hospitalists play a vital role in PCP communication and care transitions. In 2006, the Society of Hospital Medicine (SHM) published a list of core competencies for hospitalists, including in its transitions of care section the ability to strive to personally communicate with every receiving or referring physician during care transitions and expeditiously inform PCPs about significant changes in patient clinical status. 2 Hospitalists also are charged with keeping patients informed. In 2007, six societies released a consensus statement at the Transitions of Care Consensus Conference highlighting that patients should be informed of who is in charge of their care during transitions. 3
Ensuring hospital team–PCP communication is particularly important for older patients who have multiple physicians and medical conditions. Older patients are vulnerable to breakdowns in care during transitions, such as at discharge.4-6 In recognition of this, the Assessing Care of Vulnerable Elders (ACOVE) project developed quality indicators unique to frail elders, including one relevant to discharge.7,8 A higher performance on ACOVE quality indicators was found to be associated with lower likelihood of death one year after discharge. 8 The importance of a coordinated discharge process for older patients is further highlighted by the development of a discharge checklist specifically targeted for elderly patients. 9
Unfortunately, communication between inpatient and outpatient physicians is infrequent.10-12 A multicenter trial found that direct communication between PCPs and hospital teams is reported in less than one quarter (23%) of cases and that less than half (42%) of PCPs report receiving a discharge summary within 2 weeks. 12 Perhaps more concerning was the authors’ finding that nearly one quarter (23%) of PCPs were not aware of their patient’s hospitalization. Many hospital physicians do not attempt direct communication with PCPs, with a recent study finding that more than half (54.4%) of hospital providers reported no attempt at direct communication, with the most common reason for lack of attempt cited as the providers’ belief that the discharge summary is adequate. 11 However, discharge summaries often arrive late (with availability at the first postdischarge visit of 12% to 34%) and lack key information, such as test results pending at discharge (65%) and discharge medications (2% to 40%). 10 Advances in shared electronic medical records have not solved this problem. Although a recent study found that 71% of PCPs within a large academic medical center reported receiving automated discharge notifications, only 39% of PCPs felt that the automated notifications plus a discharge summary were sufficient communication for a safe patient transition from hospital to community. 13 Rather, PCPs preferred receiving an e-mail or telephone communication regarding patients with complex hospitalizations, yet only 31% of PCPs reported receiving such personalized communication.13,14
Physicians are not the only stakeholders who would prefer improved communication between inpatient and outpatient teams. Many patients feel that their hospital teams have an obligation to communicate with their PCP, and when interviewed express beliefs that “there are two doctors who are attending me and they should have communication with each other.” 15 Furthermore, patients believe that communication between their physicians is important for both their care in the hospital and their care following discharge. 16 Additionally, nearly three fourths (73%) of interviewed patients expressed increased trust in physicians whom they have known for a long time and nearly two thirds (65%) of patients believed that a doctor who has known them for a long time can provide better care. 16 Thus, although much of the medical community’s focus on communication has centered on the discharge process, many patients feel that they would receive better care if their physicians spoke at admission or during their hospitalization.
Although many patients and physicians believe that improved communication will result in improved care, the results of previous studies on the impact of communication on patient outcomes is mixed. A recent meta-analysis of interventions to improve patient discharge from hospital to community found that the majority of interventions had statistically significant effects in favor of the intervention group, such as reducing hospital use, improving continuity of care, and improving patient satisfaction after discharge. 17 However, many of the interventions studied are not aimed directly at improving hospital team–PCP communication, and the authors acknowledge that the heterogeneity and complexity of the interventions limit the conclusions that can be drawn. Among research specifically examining inpatient–outpatient communication, a prior study found that frail older patients were more likely to report problems after discharge when their PCP was unaware of their hospitalization. 15 Despite this, PCP reports of communication with the hospital team have not been linked to health outcomes such as patient deaths, readmissions, or emergency department (ED) visits within 30 days of discharge.11,12
Although the focus of prior research has centered on communication from the physician’s perspective, few studies examine patient perception of communication. The current study aims to characterize the relationship between patient satisfaction and patient perception of hospital team–PCP communication, focusing on the experience of older patients.
Methods
Study Design
Patients for this study were recruited between 2003 and 2013 using the University of Chicago Hospitalist Study, an ongoing study that interviews hospitalized patients regarding quality of care. 18
The researchers asked to consent every 10th hospitalized general medicine patient. Additionally, the researchers oversampled frail elders by asking to consent all “vulnerable elder” patients as defined by the Vulnerable Elders Survey, a validated tool that identifies patients based on age, self-rated health, and functional limitation. 19 Patients who lacked a PCP and patients whose named PCP denied caring for them were excluded. The University of Chicago Institutional Review Board approved this study and participants provided written informed consent.
Hospitalist Inpatient Interview
Trained research assistants approached patients to complete an in-person inpatient interview, beginning with the telephone version of the Mini-Mental Status Examination (MMSE). 20 For patients who scored 17 or below on this 22-point instrument, a proxy was asked to complete the patient interviews. Patients answered questions on demographics, self-reported health, and prior year hospitalizations. Patients cared for by a hospitalist were included in the overall analysis and analyzed separately. General medicine patients are either admitted to the teaching service staffed by residents who are supervised by general medicine or hospitalist attendings or to the nonteaching service staffed by hospitalists with nurse practitioners. Hospitalists on the nonteaching service also served as hospitalist attendings on the teaching service. Patients are admitted to the nonteaching service when teaching teams are “capped,” and there is no specific triage criteria for patients who are sent to the nonteaching service. Given this structure, only patients who were cared for by the nonteaching service were included in the subgroup of patients cared for by a hospitalist.
One Month Postdischarge Telephone Interview
Research assistants, who were blinded to PCP knowledge of patient admission, conducted telephone interviews with patients one month after discharge. To elicit perception of communication, patients were asked: “To the best of your knowledge, during your hospital stay, did your regular doctor communicate with the doctors caring for you in the hospital?” Patients were asked 2 questions to elicit satisfaction: “Overall, how would you rate the care you received at the hospital?” (hereafter referred to as overall satisfaction) and “How would you rate the overall coordination and teamwork between your regular outpatient doctor and the doctors who cared for you during your hospital stay?” (hereafter referred to as satisfaction with inpatient–outpatient coordination). Patients answered using a 5-point Likert-type scale with the options excellent, very good, good, fair, poor, and don’t know. Additionally, patients were asked if they had experienced acute care events since discharge, including hospitalizations and ED or urgent care visits, 21 and if they were aware of tests pending at discharge.
Primary Care Physician Surveys
Following discharge, patients’ self-identified PCPs were faxed a survey to elicit knowledge of admission. PCPs aware of hospitalization were asked about communication with the hospital team, including discharge summaries, and whether they discussed tests pending at discharge.
Data Analysis
Quantitative data were analyzed using Stata 12.0 software (StataCorp LP, College Station, Texas). Descriptive statistics were used to tabulate the communication outcome of percentage of patients and PCPs reporting hospital team–PCP communication and the 2 patient satisfaction outcomes, with satisfaction defined as excellent or very good. Patients who answered don’t know to the communication question and PCPs unaware of hospitalization were categorized as not reporting communication. The nonparametric Kruskal–Wallis test was used to test the association between patient satisfaction and perception of communication. The χ2 test was used to test the association between acute care events and communication.
Results
Among 2045 eligible patients who completed the postdischarge interview and allowed the researchers to contact their PCP, more than half (1044 [51.1%]) of their PCPs completed the PCP survey (Figure 1). These 1044 patients and their PCPs are included in the sample, which had a mean age of 72 ± 17 years. Table 1 depicts the characteristics of the sample. More than two thirds (68.4%) were deemed “frail elders.” The percentages of females and African Americans in the sample are representative of patients seen at University of Chicago Medicine (UCM). Based on the MMSE screening tool, 19.8% completed the surveys via proxy. Patients who used a proxy were older (mean age 81 ± 12 years vs 70 ± 16 years, P < .001) and less likely to live in their own house (65.6% vs 82.6%, P < .001).
Enrollment methods.
Patient Characteristics (Total N = 1044).
Abbreviations: PCP, primary care physician; SD, standard deviation; VES-13, Vulnerable Elders Survey.
n ranged from 845 to 1042 for starred questions.
More than one third of PCPs (390 [37.4%]) did not communicate with the hospital team, including 233 PCPs (22.3%) who were not aware of their patient’s hospitalization. Among these 390 PCPs’ patients, approximately half (198 [50.8%]) did not believe that communication between their physicians occurred, while the other half (192 [49.2%]) did believe that their hospital teams and PCPs communicated. Thus, when a breakdown in communication occurred, only half of patients were aware that their care teams did not communicate (Figure 2).
Communication outcomes.
Two thirds of PCPs (654 [62.6%]) did communicate with the hospital team. Among their patients, only two thirds (442 [67.6%]) were aware of communication, while nearly one third (212 [32.4%]) were not aware of communication between their PCP and hospital team. Thus, even when communication did occur among patients’ physicians, patients underestimated that communication (Figure 2).
PCPs of frail elders were more likely to be aware of hospitalization (80.5% vs 71.3%, P = .001) and to report that communication occurred (64.7% vs 56.0%, P = .009). Despite this, vulnerable elder patients were not more likely to perceive communication (60.4% vs 62.5%, P = .517). (Data not shown.)
Among 239 PCPs who reported communicating with the hospital team about tests pending at discharge, 153 were told there were pending tests; however, only 41 (26.8%) of these patients were aware of their pending tests. By contrast, 23.3% (20) of patients believed they had tests pending when their PCPs were told there were none. Thus, patients were poor historians at relaying whether or not they had tests pending at discharge. (Data not shown.)
The researchers analyzed 2 satisfaction metrics: overall satisfaction and satisfaction with inpatient–outpatient coordination. Although 63.3% of patients were satisfied with overall care, fewer (49.6%) were satisfied with inpatient–outpatient coordination. Although patients who perceived communication were more satisfied overall (70.0% vs 53.1%, P < .001), there was no difference in satisfaction among patients whose PCPs did or did not report communication (63.2% vs 63.6%, P = .194; Figure 3).
Relationship between patient perception of communication and overall patient satisfaction.
Patients who perceived communication were much more satisfied with inpatient–outpatient coordination than patients who did not believe that communication occurred (64.4% vs 26.5%, P < .001); patients whose PCPs reported communication also were more satisfied with inpatient–outpatient coordination than those whose PCPs did not (51.2% vs 46.9%, P = .033; Figure 4).
Relationship between patient perception of communication and patient satisfaction with inpatient–outpatient coordination.
PCPs of patients cared for by a hospitalist on the nonteaching service were less likely to report communication with the hospital team (46.4% vs 65.1%, P < .001); however, patient perception of communication did not differ between those cared for by a hospitalist and those not cared for by a hospitalist (61.6% vs 60.6%, P = .823). Despite the discrepancy, patients cared for by a hospitalist were equally satisfied with overall care and inpatient–outpatient coordination. (Data not shown.)
Neither patient nor PCP report of communication was associated with acute care events within 30 days after discharge: 13.9% of patients who reported communication experienced an acute care event, compared with 11.7% who did not (P = .308); 12.4% of patients whose PCP reported communication experienced an event, compared with 14.1% whose PCPs did not (P = .425). (Data not shown.)
There were no differences in satisfaction between vulnerable elders and nonvulnerable elders. Likewise, no differences in communication or satisfaction were found by whether a proxy was used. PCPs not located at the study hospital were less often aware of their patient’s hospitalization (69.1% vs 81.5%, P < .001) and less likely to report communication (37.8% vs 71.5%, P < .001). PCP report of communication ranges from 58.8% to 73.3% between 2003 to 2010, then falls to 36.7% in 2011 and 45.3% in 2012. (Data not shown.)
Discussion
More than one fifth of PCPs were unaware of hospitalization and many additional PCPs reported no communication with the hospital team. This finding confirms prior research demonstrating infrequent inpatient–outpatient communication.10-12 It is particularly concerning that PCPs reported communication less often when their patient was cared for by a hospitalist on the nonteaching service. This infrequent communication persists despite recognition of the importance of continuity of care from multiple professional societies and inclusion on SHM’s list of core competencies for hospitalists.2,3
One explanation for why communication occurs less frequently for patients on the hospitalist nonteaching service is that PCP communication is not part of the workflow. Residents assume responsibility for the majority of PCP communications for patients on the hospitalist teaching service. Because the attendings who cover the hospitalist services also serve as attendings on the teaching service, they may not be accustomed to contacting PCPs on the nonteaching service, which lacks residents. Furthermore, because nonteaching hospitalists have more handoffs (shift and service changes) than the teaching service, it is possible that contacting the PCP is lost in the handoff either at admission from night hospitalist to day team, or during weekly service changes or coverage periods such as weekends. Interestingly, PCP-reported communication decreases sharply in 2011, the year that resident duty hours restrictions tightened, perhaps implying that missed PCP communications are one consequence of increased resident handoffs and busier days. Because hospital team–PCP communication is critical, the researchers recommend that attendings maintain accountability for its occurrence and not rely on residents to ensure that communication occurs. Furthermore, efforts to incorporate essential PCP communication into the workflow for all members of the patient’s health care team and to offer transitions of care training for all providers should be undertaken.
Perhaps in recognition of the criticality of continuity of care for older patients, PCPs of vulnerable elders had more often communicated with the hospital team and were more often aware of hospitalization. Despite this, vulnerable elders were equally likely to report communication, highlighting the gap between communication and its perception by patients.
The study design is distinctive in that it measures patient awareness of communication, a topic that is important because of its potential impact on patient satisfaction and because of implications for patient safety. Regarding the latter, the researchers believe that it is important that patients understand the information that has been transferred to their PCP so that in the event of infrequent communication, the patient has the opportunity to take a more active role in relaying key elements of their hospitalization to their PCP. In addition to improving patient safety, this approach allows patients to feel more empowered by taking a proactive role in physician communication. 11 Thus, although it is worrying that this study found that patients both underestimated communication when it occurred and overestimated communication when it did not, the overestimation of communication is most concerning for patient safety because it implies a lack of patient awareness that discontinuity of care occurred. Therefore, the researchers recommend that hospital physicians inform patients when communication occurs, and when it does not occur they should help patients understand key information to share with their PCP. The teach-back method is one proven practice to increase the likelihood that patients and physicians are in agreement; it may be useful to apply teach-back with patients at discharge.
It is important that PCPs and patients are aware of tests pending at discharge. This study found that patients were inaccurate in their reports of pending tests, underscoring the need for increased hospital team–PCP communication so that crucial information such as this is promptly followed up. When miscommunication results in a failure to follow up on tests pending, it is likely to be wasteful of health care resources and can potentially lead to adverse patient outcomes.22-24
Improved communication between inpatient and outpatient physicians as part of the discharge process has been identified by professional organizations as a critical focus of improvement for patient safety in order to reduce adverse events associated with the transition from hospital to community, such as medication errors and lack of follow-up on actionable tests pending at discharge. When surveyed, both hospitalists and PCPs identified improved communication and more clearly defined accountability as critical components to successful care coordination during the transition from hospital to home. 25 In addition to physicians, patients also recognize the need for improved coordination of care. The current study found that fewer patients reported satisfaction with inpatient–outpatient coordination than with overall satisfaction, thus highlighting that even patients satisfied with their care view coordination among their physicians as a weakness and perceive the need for improvement.
Beyond advancing patient safety, improving communication may increase satisfaction. This study found that patient perception of communication was positively associated with the satisfaction metrics and that many patients underestimated communication. Therefore, hospitals could potentially increase patient satisfaction for work they are already doing by improving patient education.
This study has several limitations. Most important, the results may be limited by a responder bias among PCPs to the survey and the PCP response rate of 51.1%. Of note, nonresponder PCPs were more likely to have male patients (32.5% vs 28.4%, P = .043) and African American patients (77.2% vs 73.2%, P = .042) and less likely to be located at UCM (60.4% vs 70.9%, P < .001). Given that PCPs not located at UCM were less often aware of hospitalization and less likely to report communication, this study’s results may overestimate the extent of communication that occurred. Second, generalizability is limited because this study was completed at a single institution. The study patient population is composed of largely underserved African American patients and may differ from other populations in their experiences with the health care system. Third, the researchers cannot determine whether the relationship between patient satisfaction and perception of communication is causal. Fourth, the study results may be affected by the tendency for patients who are generally more pleased with their care to respond more positively across all survey metrics. However, even if this phenomenon is influencing the results, it is notable that patients responded differently between the 2 satisfaction metrics.
In conclusion, this study found that many PCPs were not aware of their patient’s hospitalization and that PCPs of patients cared for by a hospitalist on the nonteaching service were less likely to report communication. Patients underestimated hospital team–PCP communication when it occurred and overestimated communication when it did not, implying a lack of patient awareness that discontinuity of care occurred that may impede patient safety. Furthermore, patients’ perceptions of communication are positively associated with satisfaction. This highlights the need for increased hospital team–PCP communication and suggests that hospitals could improve patient safety and satisfaction by seizing a missed opportunity to improve patient awareness of communication.
Footnotes
Authors’ Note
The study was presented at the Society of Hospital Medicine National Meeting, March 2014, Las Vegas, NV; and at the American Geriatric Society Presidential Poster Session, May 2014, Orlando, FL.
Declaration of Conflicting Interests
The authors declared no conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received the following financial support for the research, authorship, and/or publication of this article: National Institute on Aging K23AG033763 (V. Arora, Principal Investigator [PI]); National Institute of Aging K24AG031326 (D. Meltzer, PI); AHRQ R01-HS10597 Multicenter Trial Hospitalists (D. Meltzer, PI); NIGMS R01 GM075292 TEACH Research (D. Meltzer, PI); Agency for Healthcare Research and Quality U18 HS016967 Hospital Medicine and Economics CERT (D. Meltzer, PI). The sponsors had no role in study design; collection, analysis, and interpretation of data; writing of the report; or the decision to submit the report for publication.