BACKGROUND/RATIONALE:
Nationally, the 30-day all-cause readmission rate after a hospital discharge from a stay for heart failure continues to be over 20% despite many initiatives to improve the care of patients with heart failure. Reducing readmissions is an important strategic goal of the Veterans Health Administration (VHA) and the Health Services Research and Development chronic heart failure quality enhancement research initiative (CHF QUERI). Readmission rates are being tracked. However, few VHA medical centers have tools to identify these particular readmissions as they occur and collect timely information about factors that contributed to each readmission and how it might have been prevented.

OBJECTIVE(S):
The primary objective of this one-year project is to develop and distribute a Veterans Health Information Systems and Technology Architecture (VistA) computer program that will alert local hospital staff to each 30-readmission of a patient with heart failure and provide a patient interview/chart abstract form to systematically review the care received prior to each readmission.

METHODS:
The project began on October 1, 2011. An expert programmer at the Minneapolis VA Medical Center wrote, pre-tested and documented a class III VistA program according to Office of Information and Technology procedures and standards. The program reviews all local admissions entered into VistA records during the previous day and alerts selected facility staff via VistA mailman about any admissions that were categorized as readmissions within 30 days of a previous discharge from the same facility that listed an ICD-9 diagnosis code for heart failure as the principal diagnosis.

Forms to systematically collect information about the care received prior to the readmission by interviewing the patient and reviewing the medical record were developed by reviewing the literature and local health care providers and reviewed by particiapting sites.

Voluntary sites were recruited via a web-based presentation and electronic mail sent to the national CHF QUERI network of heart failure care providers and managers in November 2011. Subsequently 18 sites expressed an interest in using the tools to systematically assess their readmissions. However, 5 sites later decided they didn't have sufficient staffs/time. Two other sites were not able to get local approval, and 2 sites did not respond to follow-up emails. Thus, 9 sites continued to be interested in using the readmission assessment tools.

The VistA program is currently installed and running at 6 sites (first site in January, most recent site in May 2012). One site had difficulty getting local help to install the VistA program and has been using facility utilization management reports as an alternative method to identify their readmissions.

Of the 7 sites who have been identifying their readmissions (6 using the project VistA software), 4 have shared their collected information about 80 readmissions (8 to 38 per site). This is about the number of readmissions expected given heart failure admission and readmission rates and the period of observation.

FINDINGS/RESULTS:
The following summary is based on 21 readmission forms completed by one of the most active sites that had 29 readmission alerts within approximately 6 months.

Chart Reviews
1. Most (76%) of the 21 readmissions were from a private residence and were not planned or scheduled (95%).

2. There were a variety of reasons for the readmissions, and more than one medical reason was noted for 38%. Eight of the 9 (89%) readmissions attributed to symptoms of heart failure had systolic dysfunction. None of the 12 readmissions of patients with heart failure and normal systolic function were attributed to worsening heart failure. Most commonly, they were due to some type of infection including 2 for cellulitis and 1 case of pneumonia. Other medical reasons included renal insufficiency, chest pain, low blood sugar, leg pain, a poorly fit leg brace and a bronchial mass.

3. Most of the evidence-based guideline recommendations for treatment of patients with heart failure pertain to the subgroup of 9 subjects with systolic dysfunction. Whether a patient met the criteria for use of the following recommended therapies or had a valid contraindication was not ascertained.
a. 100% were being treated with a loop diuretic
b. 50% were being treated with an angiotensin converting enzyme inhibitor or receptor blocker, and 38% of the others were being treated with an alternative hydralazine/nitrate combination
c. 100% were treated with a beta-blocker
d. 12% were being treated with an aldosterone receptor antagonist
e. 12 % had a cardiac resynchronization device
f. 63% were in home care, tele-health care or case management programs

4. Three (14%) of the 21 readmissions were related to a previously prescribed medication (unnecessary discontinuation of an angiotensin converting enzyme inhibitor outside the VA, C. difficile infection subsequent to antibiotic use, and a problem obtaining a prescription for a replacement antibiotic). None of the admissions were related to previously placed devices or procedures. Remarkably, 4 of 19 (21%; two missing data) may have been related to social or family issues including veterans refusing social services, palliative or hospice care and home caretakers refusing to follow recommended diets (restricted sodium, for dysphagia from a previous stroke).

5. Personnel who reviewed the readmissions made suggestions about how 7 (33%) the readmission might have been prevented by better or enhanced home/self care in 5 cases or better medical care/care coordination in 2 cases.

6. In the judgment of case reviewers, 6 (67%) of the 9 readmissions due to worsening symptoms of heart failure and 9 (75%) of the other readmissions could not have been handled without the readmission. Three (14%) died during the readmission.

Patient Interviews
1. For varying reasons, 4 (19%) of the 21 readmitted patients were not interviewed, and 5 (23%) were interviewed by telephone after being discharged. Thus, only 12 (58%) were interviewed as inpatients when recall would be better and not influenced by the care given during and after the readmission.

2. The 7 interviewed patients who were readmitted for worsening symptoms of heart failure reported they were aware of their worsening condition for a minimum of 3 days (median 5 days) before being readmitted, and 4 (57%) talked to a doctor or nurse before coming to the hospital. None of the 7 reported problems calling or getting to see a VA doctor or nurse. However, five (71%) said they needed more help taking care of themselves at home although none said their health care providers asked them to do more than they were able to do to care for themselves, or felt confused about what they were supposed to do.

Among the 10 that were readmitted for reasons other than worsening heart failure and interviewed, 4 (40%) reported they needed more help from a doctor or nurse during the week before the readmission, and that they had problems calling or getting to see a VA doctor or nurse. Five (50%) felt they were asked to do more than they were able to do to take care of themselves at home.

3. Of the 7 that were readmitted for symptoms of heart failure and interviewed, 6 (86%) reported they continued to take their medications as prescribed.

4. Two (29%) readmitted for symptoms of heart failure had questions or concerns about what they should eat or drink, however 4 (57%) reportedly ate more salty foods. Three (43%) ate at restaurants more often than usual before being admitted. When asked if there was anything they could have done to prevent this admission, all 3 indicated they should have followed the recommended diet.

5. When asked if there was anything the VA could have done to avoid the readmission, 5 (50%) of those that were readmitted for reasons other than worsening symptoms of heart failure said there was. Four felt the medical problem wasn't adequately addressed during the previous admission, and 1 had a problem getting home care when their regular provider became unavailable. Of those that were readmitted for signs and symptoms of heart failure, none said the VA could have done something to avoid the readmission. One said a non-VA health care stopped their medication (after they excessively increased the dose leading to worsening renal function).

Conclusions:

The VistA software can be installed by VA medical centers throughout the nation, will identify most readmissions that occur within 30 days after a veteran has been discharged alive from an admission due to heart failure, and will alert local personnel to readmissions in a timely manner. Cooperation from facility administrators and Information Resources Management is needed to make this software widely available.

Busy health care providers were able to interview patients and review charts to collect information on most readmissions at their hospital. However, the lack of staff time to regularly check for readmission alerts and follow-up to collect timely information was a major limiting factor. Ideally, facilities would have resources and be held accountable for reviewing their readmissions. However, this effort needs to be proven worthwhile.

Review of the modest amount of information that was collected has not identified any commonly occurring reasons for readmissions. A variety of problems with home or self-care seemed to contribute to a substantial number of the readmissions for heart failure at the site summarized herein. Improving other apsects of care might have more impact on readmissions that were not due to heart failure. Other sites indicated that timely identification and review of readmissions may help them prevent readmissions in individual cases.

IMPACT:
Whether routine use of this or similar readmission identification and review processes can reduce readmissions remains to be determined. A controlled clinical trial would be the best way to answer this question.

---

External Links for this Project

---

PUBLICATIONS:

Journal Articles

1. Anand IS, Win S, Rector TS, Cohn JN, Taylor AL. Effect of fixed-dose combination of isosorbide dinitrate and hydralazine on all hospitalizations and on 30-day readmission rates in patients with heart failure: results from the African-American Heart Failure Trial. Circulation. Heart failure. 2014 Sep 1; 7(5):759-65. [view]

Journal Other

1. Rector TS. Use of qualitative methods to improve the quality of cardiovascular care. (Commentary). Circulation. Cardiovascular Quality and Outcomes. 2013 Jul 31; 6(2):171-177. [view]

DRA: Health Systems
DRE: Technology Development and Assessment
Keywords: none
MeSH Terms: none