Objective To compare patients’ and providers’ views on contributors to 30-day hospital readmissions.
Design Analysis of a qualitative interview survey between 18 May–30 June 2015.
Setting Interviews were conducted during the 30-day readmission hospitalisation at a single tertiary care academic hospital.
Participants We conducted 178 interviews of readmitted patients.
Measures We queried opinions of what factors patients believed contributed to their rehospitalisation and compared this with the perspective of the index admission provider. The primary outcome was the view that the readmission was preventable. A review by a RN (nurse) case manager also provided an assessment based on patient report, provider report and chart review.
Results Patients were more likely to view a readmission as preventable compared with physicians (p<0.0001). Patients identified system issues (defined as factors controlled by the hospital discharge process) as contributors to their readmission in 58% (103/178) of cases while providers identified system issues as the contributor to a patients’ readmission in 2% (2/101) of cases. Patients with poor functional status were more likely to feel the cause of their readmission was due to system issues than patients with better functional status (p=0.03). A RN case manager review determined that in 48% (86/178) of cases the system had some amount of contribution to a patient’s readmission. There was no significant difference in belief that the readmission was preventable between the RN case manager and the patient (p=0.47).
Conclusions Readmitted patients often feel that the hospital system contributed to their readmission. Providers did not recognise patient and RN case manager identified issues as contributors to hospital readmissions.
- quality improvement
- audit and feedback
- healthcare quality improvement
- hospital medicine
- patient-centred care
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
- quality improvement
- audit and feedback
- healthcare quality improvement
- hospital medicine
- patient-centred care
In 2011, Medicare estimated the readmission rates for those enrolled in Medicare part A (age >64) within 30 days of initial hospital discharge at 19%.1 This resulted in roughly 3.3 million readmissions in 2011 alone and an additional $41.3 billion of hospital costs.2 To incentivise cost control, in 2012, the Medicare Hospital Readmissions Reduction Program (HRRP) instituted financial penalties for hospitals not meeting prespecified readmissions rates.3–5 It is reported that 78% of all hospitals received some amount of HRRP penalty in 2015 totalling $428 million in withheld payments to hospitals.4 Furthermore, large teaching hospitals appear to have a higher risk of being penalised under HRRP.6
The increased financial pressure on hospitals in recent years has led to a flurry of clinical trials on readmission prevention. Data now exist on everything from prediction algorithms, to closer primary care provider (PCP) follow-up, to discharge planning services, to medication interventions.7–10 Ongoing research both in the hospital and in the community settings seek to keep discharged patients out of the hospital.9 11 This combination of research effort and financial pressure3 culminated in a decrease in readmission rates from 19% in 2011, to 18.5% in 2012, to 17.5% in 2013 and 17.8% in 2016.1 3 12 This is a notable reduction in readmission rates, but still likely represents the tip of the iceberg since previous studies have hypothesised that 26%–76% of readmissions are avoidable.13 14 The question remains, how do we prevent the remaining avoidable readmissions and what factors are contributing?
Only a handful of studies have asked patients their viewpoint on the cause of readmission.15 16 Recently, Auerbach et al showed that communication, readiness for discharge and supporting patient self-management contribute to preventable readmissions.14 Our goal was to compare readmitted patients with their discharging physicians regarding assumed causes of their readmission. We then sought to compare these responses to those of an RN case manager who reviewed the chart and interview responses.
We used data collected from qualitative interviews of randomly selected patients over 18 years of age readmitted within 30 days of discharge from Stanford Hospital, a 475-bed suburban teaching hospital, between 18 May and 30 June 2015. The survey was conducted on behalf of Stanford Hospital administrators as part of a quality improvement initiative.
Readmissions were defined as an unplanned admission to Stanford Hospital on any inpatient service within 30 days of date of discharge from Stanford Hospital. Admissions were considered planned if there were preadmission orders for surgery or chemotherapy. Any diagnosis was eligible for inclusion except child birth. Patients with multiple readmissions during the study period had each readmission included as a separate event. Patients with different index and readmission services were included as were patients with readmissions for causes unrelated to their index diagnosis. Eligible participants were identified by a daily update sent out by hospital administration flagging all 30-day hospital readmissions. Not all readmissions on the daily list were interviewed due to time constraints. Interviewers selected patients randomly off the daily list without preset criteria. The purpose of random selection was to allow interviewers to see as many readmitted patients as possible without limiting their work flow. Since this study was conducted as a qualitative hospital survey, response rates were not collected.
Eligible participants were interviewed in person during their readmission visit. Interviews lasted approximately 30 min, and all answers were typed and directly uploaded into a secure computer database after the interview. Patients had to either have ability to complete the interview themselves or have a family member or caregiver who could act as their proxy. If neither patient nor caregiver could complete the interview, the patient was excluded. Language interpreters, either in person or over the phone, were used for all non-English speaking participants. All interviews, chart reviews and provider contact were conducted by two trained RN case managers.
Interviews were performed using a standardised questionnaire created from a combination of published data plus questions added specific to Stanford Hospital. Patients were asked 11 questions with various formats including yes/no, multiple-choice and free text. For the purposes of this study, only question number three which focused on perceived contributors to the readmission was analysed (figure 1). This was a multiple-choice question with the option for ‘other’ to write-in an answer.
Following the patient interview, the attending physician at the time of discharge from the patient’s index admission based on chart review was contacted either by phone, pager or email for feedback by the RN case manager. Each was given the patient’s readmission diagnosis and was asked to review the chart and provide feedback on what factors could have contributed to the patient’s readmission. They were asked a single multiple-choice question regarding what they believed contributed to the readmission (figure 1).
Review by RN case manager
All interviewed patients’ medical records underwent a manual chart review by the same RN case managers who conducted the patient and provider interviews. An audit tool based on previously published data for chart review17–19 plus additions unique to Stanford Hospital electronic medical record was used to perform a standardised chart review. The manual chart review obtained: follow-up appointment timing and attendance, postdischarge service utilisation and timing, that is, skilled nursing facility (SNF) or home health services (HHS), and admission/readmission diagnosis. The RN case manager then compiled the chart review with the patient and provider interviews and formed their own conclusion if anything could have been done different by the healthcare system to help prevent the readmission. The RN case managers could provide more than one response if they concluded multiple patient and/or system factors affected the readmission. Thus, their perspectives total greater than 100%.
Electronic chart review
The electronic chart review used a research database known as Stanford Translational Research Integrated Database Environment to evaluate carefully selected variables from the electronic medical record and uploads them into Excel. We evaluated potential risk factors for readmissions based on previously published data.20–23 The variables we included were: need for durable medical equipment at discharge, functional status at discharge based on RN bedside assessment, PCP follow-up, follow-up phone call after discharge and SNF/HHS utilisation. These variables were then compared with the patient’s perspective of contributors to readmission to see if any areas correlated with patient perceived risk for readmission.
Descriptive statistics were used for all demographic data described as medians with IQRs, total counts and per cent. When looking at larger trends, responses were grouped into ‘system’ and ‘patient’ and ‘no modifiable cause’ for ease of statistical analysis (figure 1). The primary outcome was the view that the readmission was preventable. The patient perspective was compared with risk factors for readmission with χ2 analysis. For the question of whether anything could have been done to present readmission, we used McNemar’s test to compare beliefs that the readmission was preventable between providers and patients and between RN case managers and patients.
In situations where the RN case managers identified more than one contributor to readmission for an individual patient within separate categories (ie, patient factors and systems issues) each value was counted separately. When RN case managers identified more than one contributor to readmission within the same category (ie, two systems issues), this was counted as a unit when summarising totals for each category (ie, total systems issues identified by RN case managers). A two-sided p value <0.05 was considered statistically significant. Statistical analyses were performed using Excel spreadsheet.
A total of 3601 patients were discharged during the study time period from 18 May to 30 June 2015. Of these patients (469/3601), 13% returned as readmissions within 30 days of their last discharge. A total of (182/469) 39% were interviewed regarding what they believed contributed to their 30-day hospital readmission. Seventeen of these patients were readmitted twice during the 44-day window interviews were conducted and data from both readmissions were included. Four of the patients enrolled had their interview data excluded due to inability to complete the interview but demographic data were collected as they did complete enrolment. The patient population was evenly split between male and female (table 1). They had a median of four total admissions per person over the preceding 5 years and represented a mix of surgical, medical and transplant service admissions. The most common reasons for readmission were pain and infection management (table 1).
Of the 182 patients enrolled, we received 101 (57%) responses from the original attending physicians. The providers who responded were (31/101) 30% hospitalists and (70/101) 70% specialists.
Patient, provider, RN case managers perspectives on readmission
We analysed the sum of how often patients, providers and the RN case managers felt a modifiable system issue contributed to readmission, the patient contributed to readmission or nothing could have prevented readmission. A total of (103/178) 58% of patients believed a modifiable system issue contributed to their readmission. The most commonly mentioned systems issue was being discharged too early which (37/178) 21% of patients expressed (table 2). Patients were not directly queried if their own compliance or personal issues contributed to the readmission. Providers felt that no modifiable cause for readmission could be identified (72/101) 71% of the time. Patient psychosocial or adherence factors comprised another (28/101) 28% and only (2/101) 2% of the time was a system issue faulted as a contributor to readmissions (table 2). RN case managers found no modifiable cause in (73/178) 41% of the readmissions. However, unlike the providers, they identified systems issues in (86/178) 48% of readmissions and patient adherence issues in another (35/178) 20% of readmissions (table 2).
Providers and patients agreed that the admission was preventable in 22 and non-preventable in 30. In 49 cases, the providers disagreed (in 42 patients felt the admission was preventable and in 7 the providers felt it was preventable, p<0.0001). However, when specific reasons for reason for readmission were examined, we found that providers rarely agreed with patient’s perspective. Only (2/101) 2% of cases did providers agree with patients that the system contributed to a readmission. Providers only shared the patient’s perspective when both parties felt nothing could have been done different to prevent the readmission which accounted for (30/101) 30% of cases (table 3).
Unlike the providers, the RN case managers agreed with the patient perspective in (145/178) 81% of readmissions. In (81/178) 45% of cases, the RN case managers agreed with the patient that the system contributed to the readmission and another (64/178) 36% both parties agreed that nothing could have been done different to prevent the readmission (table 3). There was no significant difference (p=0.47) between patient and RN case managers views that the readmission was preventable.
Patient risk factors
Poor functional status was associated with patients perceiving system issues as having contributed to their readmission by χ2 testing (p=0.03) (table 4).
We are the first to our knowledge to report the patient perspective of readmissions, provider perspective of readmissions and in-depth chart review by an RN case manager. This study found a marked divergence in perspective between providers, patients and RN case managers on contributing factors to hospital 30-day readmissions. Patients often felt on discharge more could have been done to prevent readmission. Providers, on the other hand, most frequently said nothing could have prevented the readmission or that patient-specific factors led to the readmission. Our RN case managers found areas for hospital improvement in methods of discharge for nearly half of readmissions and often agreed with patient perspectives.
We found that patients frequently identified opportunities for the medical system to improve discharge timing, follow-up, home health and skilled services. This confirms what previous studies have demonstrated as contributors to readmissions.24 Interestingly, patients identified items that theoretically could have been addressed before discharge. Most commonly, patients cited lack of readiness for discharge which 21% of patients expressed as the main reason they were readmitted. Furthermore, the patient perception of hospital process contribution to readmission was correlated with their functional status at the time of discharge from the index admission. Patients with poorer functional status at the time of discharge more often perceived the hospital as having contributed to their readmission. Even in this small population physically limiting needs seemed linked to the patients’ perceived ability to stay out of the hospital. This is consistent with prior publications citing functional status as a risk factor for readmission.22 25
We found that providers did not share the patient’s opinions of discharge. Providers nearly always said no factors could be modified to prevent the readmission or pointed to patient-specific factors as cause for readmission. However, review by a RN case manager found in 49% of readmissions the hospital system had some amount of opportunity to improve the discharge process. The RN case managers more often agreed with the patient’s perspective of readmission than the provider’s.
Our reviewers’ frequent agreement and providers’ disagreement with the patient perspective is an interesting discrepancy. RN reviewers and physicians were working in the same hospital, at the same time, with access to the same patient chart information, and reached very different conclusions. There are several hypotheses as to what caused this difference. Potentially, physician do not easily recognise their own systems or personal limitations in the discharge process of an individual patient. It could suggest that patient-specific psychosocial-economic factors leading to readmission risks that are known to treating providers are not easily picked up from chart review or scripted interview. It could suggest that fundamentally physicians place more onus on patients to self-advocate for care while the RN case managers emphasise the system providing support. Finally, there could be study biassing factors that lead to physician responses. Perhaps providers did not remember individual patients and chose no modifiable action as the simplest answer when queried. Or providers felt system issues were not modifiable and thus chose that as the correct answer.
Overall, the gap between providers’ and patients’ perspective of discharge is a potential contributing root cause to preventable readmissions, and we should incorporate more patient engagement into the discharge process. Possibilities to improve provider patient engagement might include provider training around discharge processing, improved patient education, more robust support services such as case management and social work. The need for physical therapy and poor functional status also correlated strongly with patient perceived readmission risk and serves as another marker to bring provider attention to high-risk patients for readmission.
This study has several limitations. The data collection was part of an internal institutional review and questions of providers, and patients were not equivalent though both were asked if the admission was preventable. We only have provider perspectives for a subset of the patient population. Furthermore, the questions directed towards patients did not provide adequate opportunity for them to self-identify if their own actions might have contributed to readmission. This study included a small sample size from individual institution and does not represent a population at large. The discharge practices, services and patient populations practised at this particular institution may not be applicable across other institutions. This study is retrospective and cannot surmise the opinions of patients or providers at the time of discharge from their index admission. Further research with prospective data comparing readmitted with non-readmitted patient opinions at the time of their index admission discharge would be beneficial.
Conclusions and future directions
This study sheds light on the dramatic gap between perceptions of patients and that of their discharging provider on causes of readmissions. Furthermore, it shows that patients with poorer functional status more often perceive the hospital as contributing to their readmission. It is a novel insight into readmitted patient’s perceptions and potential gaps in provider communication. Further research into the generalisability of these findings at other institutions is needed.
Everyone involved in the hospital readmission patient perspective study. Stanford Department of Internal Medicine for funding this study. A special thank you to Olgica Grujic, Jamie Chang, Ashley Chan for data collection and patient interviews. Another special thanks to Tim Morrison and Stephanie Bowen for development of interview and data collection tools.
Patient consent for publication Not required.
Contributors AS and LS: participated in conception and design, analysis and interpretation of data, drafting the manuscript, critical revision of the manuscript, final approval of the manuscript. PAH: participated in analysis and interpretation of the data, critical revision of the manuscript, final approval of the manuscript. GK: participated in analysis and interpretation of data, critical revision of the manuscript, final approval of the manuscript. JC: participated in analysis and interpretation of data, critical revision of the manuscript, final approval of the manuscript. JH: participated in critical revision of the manuscript, final approval of the manuscript.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Departmental funding provided by Stanford Department of Internal Medicine.
Competing interests None declared.
Ethics approval Quality improvement exempt by Stanford University IRB.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Data is available upon request for interview data results. Email email@example.com
Presented at Presented as a poster at Society of Hospital Medicine National Conference in San Diego, 7 March 2016.