The RED intervention decreased hospital utilization (combined emergency department visits and readmissions) within 30 days of discharge by about 30% among patients on a general medical service of an urban, academic medical center. More intervention group participants reported seeing their PCP for follow-up within 30 days and reported higher levels of preparedness for discharge. In addition, the intervention was successful in reducing hospital utilization among participants who frequently used hospital services. These data support implementation of a comprehensive program for hospital discharge among similar hospitals.
Our intervention includes patient-centered education, comprehensive discharge planning, and postdischarge reinforcement and is practical and easily applied to general medical patients. The RED intervention has 3 core elements: the DA, the AHCP, and the follow-up telephone call by those of the pharmacist. Because these elements were bundled, we could not clearly determine the degree that each part contributed to the effects demonstrated. No previous studies have evaluated this trio of interventions together, although the roles of the DA and the pharmacist build on previous literature (12
). For example, peridischarge nursing support services have been shown to improve discharge for patients with heart failure (14
). Coleman and colleagues (12
) used a nurse “transition coach” to demonstrate reduced readmissions at 30 and 90 days among elderly patients. Naylor and coworkers (13
) found that nurse specialists involved during and after discharge also effectively reduced acute readmissions.
Several studies have analyzed pharmacist interventions. Dudas and colleagues (15
) randomly assigned patients to receive a telephone call by a pharmacist after discharge and demonstrated fewer emergency department visits. Schnipper and coworkers (16
) used pharmacist counseling before and after discharge and showed reductions in preventable adverse drug events and medication-related readmissions and emergency department visits. Al-Rashed and colleagues (41
) found that predischarge pharmacist-based counseling for elderly patients followed by a postdischarge home visit resulted in fewer unplanned primary care visits and fewer readmissions.
The techniques used to teach the AHCP, its content, and its format (for example, pictures, color, and large font) were informed by the literature on limited health literacy (42
). Overall, the intervention improved patient comprehension of key elements of self-care: 30 days after discharge, intervention participants were better able to identify their primary diagnosis and reported better understanding of their diagnosis, medications, and appointments. The content, format, and teaching of discharge preparation tools deserve further attention because few studies have assessed the effect of patient education on subsequent hospital utilization.
Because intervention group participants were more likely to report seeing their PCPs after discharge and we transmitted discharge information to PCPs promptly after discharge, the intervention optimized the chance that PCPs could identify and address outstanding issues. In addition, the pharmacist follow-up telephone call identified any problems that a patient was having after discharge and relayed those issues to the PCP. Previous studies have suggested that improved access to community-based follow-up alone may not be enough to reduce hospital readmissions (18
). We provide evidence that when combined with other elements of RED, improving PCP follow-up may help reduce hospital utilization.
Implementing this discharge intervention required about 1.5 hours of nursing time and 30 minutes of pharmacist time per participant. Because some of the DA activities were redundant with those of existing hospital personnel, implementation of the RED intervention using existing hospital staff would require less time per patient. Also, because information was manually entered to create each AHCP, hospital information technology solutions could be developed to make this process more efficient. Despite this, we demonstrated hospital utilization cost savings averaging $412 per discharge. These figures do not include the cost of the intervention, which involved 0.5 full-time equivalent for a nurse and 0.15 full-time equivalent for a clinical pharmacist. If adopted broadly, this intervention could produce substantial effects on health care financing (44
). However, an important challenge for programs like RED is that health providers, who are best situated to implement such a program, may have no financial incentive to do so. Hospitals serving capitation-based patient populations may benefit financially from reducing unneeded rehospitalization. Under the fee-for-service scheme, the payer will benefit even after paying the full cost of the intervention. Hospitals will also benefit from decreasing the rehospitalization rate as an important quality-improvement target, and investment in strategies proven to work will be attractive to payers. The National Quality Forum is reviewing new metrics of quality care surrounding readmission rates (45
), and programs like RED may be used to improve health care organizations' quality ratings.
Our study has limitations. Because of staffing limitations, we were only able to enroll 2 to 3 participants per day, and we could not enroll participants on some weekends and holidays. Because of the nature of our urban, underserved patient population and exclusion of patients coming from nursing homes, the study sample was younger and had fewer comorbid conditions than those in other studies; thus, our results may not be generalizable to all patient groups. Also, we relied on participant self-report for outcomes that we could not gather from EMRs, notably data on PCP follow-up and visits at hospitals other than Boston Medical Center. Previous studies have suggested that patient reports of emergency department and hospital use correlate well with electronic records from 6 months to 1 year (46
). Ritter and colleagues (48
) demonstrated that patients tended to underreport outpatient visits over 6 months compared with electronic charts and found no demographic or health-related predictors of underreporting. In our case, recall bias should be expected to be nondifferential because our study was randomized, we reached both study groups equally, and outcome assessors were blinded to study assignment. We assumed that study participants not reached by telephone for an outcome assessment were alive for 30 days after the index discharge, and we relied on hospital EMRs to gather primary outcomes. Therefore, we did not capture deaths or hospital utilizations at institutions other than Boston Medical Center for this limited number of participants. For the cost analysis, we could not determine a generalizable cost for the intervention because costs vary widely by institution and location. Similarly, we could not estimate the downstream cost implications of avoided emergency department visits and readmissions. Still, we present the actual costs for 3 important types of directly related medical utilization. The cost of hospital utilization and outpatient visits also cannot be easily generalized. Our goal is to provide the direct comparison that can be made for these key costs between study groups, and we observed a 33.9% reduction in these costs.
In summary, the RED program successfully reduced hospital utilization, improved patient self-perceived preparation for discharge, and increased PCP follow-up. In 2007, the National Quality Forum Consensus Standards Maintenance committee identified hospital discharge as a critical area for improvement. The resulting National Quality Forum “Safe Practice” was based largely on the principles of the RED program (49
). Our study provides data supporting the implementation of the discharge standards promoted by the National Quality Forum.