Medical and socioenvironmental predictors of hospital readmission in patients with congestive heart failure☆

Article Outline

• Abstract
• Methods
  • Study patients
  • Demographic and medical information
  • Socioenvironmental information
  • Follow-up information
  • Statistical analysis
• Results
  • Patient characteristics
  • Rates of death and readmission
  • Bivariate association between variables and readmission
  • Independent predictors of CHF hospital readmission
• Discussion
  • Predicting factors for CHF readmission
  • Implications
  • Limitations
• Acknowledgements
• References
• Copyright

Abstract 

Background Patients with chronic congestive heart failure (CHF) require frequent rehospitalization because of the exacerbation of CHF. It is of clinical importance to determine predicting factors for readmission to reduce this likelihood. Previous studies have focused primarily on the demographic and medical characteristics in selected subsets of patients. Therefore, within a broad cohort of consecutively hospitalized patients, we sought to identify not only demographic and medical predictors but also socioenvironmental factors associated with readmission. Methods We assessed demographic (age, sex), medical (etiology of CHF, New York Heart Association functional class, left ventricular ejection fraction, previous admission for CHF, length of hospital stay, comorbidity, and medications), and socioenvironmental variables (occupation, financial resources, living alone, and follow-up visits) in 230 patients discharged with a diagnosis of CHF and recorded hospital readmission. Results Within 1 year after discharge, 81 patients (35%) were readmitted. Five variables, including poor follow-up visits (odds ratio [OR] 4.9, 95% CI 2.0-11.8), previous admission for CHF (OR 3.3, 95% CI 1.8-6.1), no occupation (OR 2.6, 95% CI 1.2-5.5), longer hospital stay (OR 3.2, 95% CI 1.2-8.5), and hypertension (OR 2.0, 95% CI 1.1-3.7), were identified as significant independent predictors for readmission by multivariate logistic regression analysis. Conclusions Our independent predictors of readmission support the importance of medical and socioenvironmental factors in the deterioration of CHF. Therefore interventions to decrease readmission should also target social management in all hospitalized patients. (Am Heart J 2001;142:e7.)

Patients with congestive heart failure (CHF) are frequently readmitted to the hospital because of exacerbation of their symptoms. The 3- to 6-month readmission rate has been reported to be as high as 30% to 50%.¹, ², ³, ⁴

Previous studies examined risk factors for readmission to identify patients at high risk. They found associations between advanced age, prior hospital admission, the severity of illness, and medical comorbidity with hospital readmission.², ⁵, ⁶, ⁷ However, previous studies have focused primarily on such characteristics as age, sex, or severity of illness. Very few studies have surveyed participants about occupation, financial resources, home living situation, social support, and follow-up visits. In addition, most previous studies have been performed among the selected subsets of CHF patients,⁴, ⁵ and a few studies used unselected patients.⁶

Accordingly, this study aimed to identify factors, on the basis of not only the demographic and medical characteristics of the patients but also the socioenvironmental variables, that would predict risk of readmission after discharge in patients consecutively hospitalized with CHF. To address this question, we developed a database for 230 patients discharged with a principal diagnosis of CHF at 5 teaching hospitals in Fukuoka, Japan, and the survival and hospital readmission were followed up. In our study patients, the rate of hospital readmission for an exacerbation of CHF was as high as 40% within 1 year after discharge.⁸ Identification of high-risk patients can help physicians and nurses target their efforts to reduce the risk for readmissions not only within the hospital but also after discharge.

Back to Article Outline

Methods 

Study patients 

The study institutions included 5 cardiology units (1 university hospital and 4 nearby hospitals) serving as primary, secondary, and tertiary referral medical centers for cardiovascular patients in Fukuoka, Japan, which has 1.3 million residents. There are several other hospitals with cardiology wards that care for CHF patients, and thus 5 hospitals enrolled in this study did not capture all CHF patients in the study area. The institution medical records identified all patients discharged with a principal diagnosis of CHF between January 1, 1997, and December 31, 1997. First, potential cases of CHF were identified through the discharge diagnosis, which indicated patients who had a diagnosis of CHF. Once these patients had been identified, the complete medical records of each candidate case were reviewed carefully. The validity of diagnosis of CHF was ascertained by use of the Framingham criteria.⁹ A diagnosis of CHF was established by the simultaneous presence of at least 2 major criteria or 1 major criterion in conjunction with 2 minor criteria (Table I).

Table I. Framingham criteria for CHF

Our preliminary study has ascertained that the necessary information on patient symptoms and signs is recorded in the majority of medical records. Data recorded on the abstraction form included (1) demographics, (2) symptoms and signs, (3) etiology, (4) examinations performed at the time of index admission including electrocardiogram, chest roentgenogram, and echocardiograms, and (4) discharge medications. Because this study focused on readmission, patients were excluded when they died during the index admission or were transferred to another hospital. Patients who were readmitted to the hospital during the study period were included only by the index admission. The research protocol was approved by the hospitals' institutional review board. Informed consent was attained from the study patients.

Demographic and medical information 

On the basis of clinical judgment and previous literature,², ⁴, ⁵, ⁶, ⁷ we selected the following demographic and medical factors as potential predictors for readmission. Demographic variables included age and sex. Medical variables included etiology, atrial fibrillation, New York Heart Association (NYHA) functional class on admission, left ventricular (LV) ejection fraction (EF) by echocardiography, prior admission for CHF, length of hospital stay, types of hospitals (primary or secondary and tertiary), comorbidity (hypertension, diabetes mellitus, renal failure, and stroke), and discharge medications. Inadequate education about the discharge plan was evidenced by “failure to provide a focused, comprehensive, multidisciplinary discharge plan for the patient and/or family before discharge.” Discharge plan included the knowledge of CHF and related self-care, the assessment of the need and subsequent plan for outpatient follow-up, and the patient use of health and social services.

Socioenvironmental information 

Socioenvironmental factors included occupation, financial resources, marital status, family caregiver, professional support (health care professionals or providers), and follow-up visits. “Financial resources” were considered to be present when the patients had financial support for physician visits and medications. For the patients who had financial support, a follow-up question inquired about the types of sources including private income or social security benefits. “Professional support” was identified as weekly or biweekly home visits by health care professionals or providers, usually nurses, during which patients conditions were assessed, medications were reviewed, and any necessary changes were made. “Follow-up visits” were evaluated by the number of outpatient follow-ups at regularly scheduled clinical visits per month.

For the medical record review, a data abstraction form was developed. Demographic and medical information was collected from the medical and nursing records. Socioenvironmental information was collected first from the nursing records and outpatient medical records and further confirmed by the interview directly from the patient or family.

Follow-up information 

From October 1999 to December 1999, the status of all patients was surveyed by outpatient medical records and by mail or telephone contact. The mean follow-up period was 2.4 years (886 ± 116 days). The following information was obtained: (1) patient survival, (2) cause of death, and (3) hospital readmission because of an exacerbation of CHF that required more than continuation of the usual therapy. We attempted to contact all patients to ascertain whether they might be admitted to other hospitals. Sudden cardiac death was defined as death occurring instantaneously within 1 hour of a change in symptoms or unexpectedly during sleep.

Statistical analysis 

All data are expressed as means ± SD. Survival and readmission rate during the follow-up were derived by the methods of Kaplan and Meier. We performed univariate analysis to compare the patient characteristics between readmitted and nonreadmitted patients by t test for continuous variables and the χ² test for discrete variables. To investigate the independent association between variables and readmission, we used all variables significantly (P < .05) identified by the univariate analysis as candidate variables and entered in a multiple logistic regression model. An odds ratio for readmission was calculated for all independent variables along with 95% CIs. Further, the association between numbers of predictors and the risk of CHF-related readmission was assessed with the use of the Cochran-Armitage test for trend.

Back to Article Outline

Results 

Patient characteristics 

Among 3654 patients admitted to 5 participating hospitals during the study period, 263 patients were discharged with a principal diagnosis of CHF. Twenty-six patients (10%) had medical records that did not meet Framingham criteria. Seven patients (2.7%) refused to participate in the study. Thus the remaining 230 patients were included in the current study.

The demographic and clinical characteristics are summarized in Tables II and III, part of which has been reported in our recent study.⁸ Many patients (38%) had a prior history of hospital admission for CHF. The length of hospitalization was an average of 34 days. The most common comorbidity was hypertension (41% of all patients). At discharge, diuretics were administered to 77% of the patients, angiotensin-converting enzyme (ACE) inhibitors to 54%, digitalis to 52%, and β-blockers to 18%. ACE inhibitors were prescribed in 74% of patients with an EF <40%. At discharge, 65% of patients or families had received education on discharge planning.

Table II. Demographic characteristics between readmitted and nonreadmitted patients

NS, Not significant.

Table III. Medical characteristics between readmitted and nonreadmitted patients

NS, Not significant.

The socioenvironmental characteristics of the study patients are summarized in Table IV. Twenty-eight percent of patients had an occupation. The majority of patients (70%) were supported mainly by social security benefits and 29% by private income. Twenty-one percent lived alone and did not have anyone at home who could take care of them. Most of the patients (80%) received nonprofessional support from the family. Only 13% had professional support. After discharge, almost all the patients were followed up medically on the regular basis, which, however, ranged from once a week to once per 3 months.

Table IV. Socioenvironmental characteristics between readmitted and nonreadmitted patients

NS, Not significant.

Rates of death and readmission 

During the follow-up period of 2.4 ± 0.3 years, 42 patients (18%) died. Seventy-four percent of all-cause deaths were of cardiovascular origin. Twenty-nine percent of cardiovascular deaths were identified as being caused by sudden cardiac death. The 6-month, 1-year, and 2-year mortality rates were 6.1%, 8.3%, and 16.5%, respectively (Figure 1).

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1. 

  Cumulative rate of readmission, cardiac death, and noncardiac death in all patients with heart failure after discharge.

In contrast to a relatively lower mortality rate, 81 patients (35%) had hospital readmission because of an exacerbation of CHF within 1 year of hospital discharge (Figure 1). Ninety-three patients (40%) had readmission during the follow-up period. The number of hospital readmissions ranged from 1 to 5 (1.5 ± 0.8).

Bivariate association between variables and readmission 

Table II, Table III, Table IV show the comparison of the demographic, medical, and socioenvironmental variables between readmitted and nonreadmitted patients. Patients who were readmitted did not differ in age and sex from nonreadmitted patients.

The incidence of ischemic heart disease was slightly higher in readmitted patients than in nonreadmitted patients (38% vs 33%), but this difference did not reach statistical significance. NYHA functional class and echocardiographic EF were also comparable between the 2 groups. Fifty-seven percent of readmitted patients had a prior history of hospitalization because of CHF before the index admission, as opposed to 28% of nonreadmitted patients (P < .001). The length of hospitalization was longer in patients with readmission. In addition, patients with a history of hypertension were more likely to be readmitted (49% vs 36%, P < .05). There were no significant differences between readmitted and nonreadmitted patients with regard to the use of specific classes of cardiac medications prescribed on discharge including diuretics, ACE inhibitors, β-blockers, and digitalis. Therefore these medications did not reduce CHF-related readmission in our study population. Discharge factors such as patient or family education were not associated with readmission.

Three socioenvironmental variables were significantly associated with readmission: occupation, professional support, and follow-up visits.

Independent predictors of CHF hospital readmission 

We then determined independent predictors of hospital readmission. Of the 6 variables significantly associated with readmission by the bivariate analysis, 5 factors were significantly (P < .05) associated with readmission by logistic regression analysis (Table V).

Table V. Independent predictors for hospital readmission

Significant clinical predictors included prior admission for CHF and longer hospital stay. The comorbidity that was independently predictive of CHF readmission was a history of hypertension.

Socioenvironmental factors including poor follow-up and no occupation were also significant independent predictors. Lack of professional support tended to be a significant predictor, which, however, was not statistically significant in the multiple logistic regression model (P = .077). These results were not altered even when the model for predictors of CHF readmission was run again with the addition of variables shown in previous studies to be predictors (male, comorbidity such as diabetes, hypertension, renal failure, and stroke, and EF <40%).

Table VI shows the CHF readmission rate according to the number of predictors in our study patients.

Table VI. Readmission and death by number of predictors

A greater number of risk factors was associated with a higher risk of adverse outcome (readmission and death). Patients with none or one of the risk predictors had a risk of readmission of 12%, whereas patients with 4 or more had a CHF readmission rate of 77% (P < .01 for trend). Importantly, these predictors were also associated with the risk for death (P < .01 for trend). Thus these predictors can provide a stratification of risk for readmission for CHF as well as death.

Back to Article Outline

Discussion 

The major findings of the current study are (1) the readmission rate of patients with CHF is high (35% within 1 year after discharge), (2) patients with a prior history of hospitalization with CHF, hypertension, and longer hospital stay at index admission are at increased risk for readmission, and (3) socioenvironmental factors including poor follow-up and no occupation were also independent predictors associated with CHF readmission. Of the 40 potential predictors, only 5 were found to be significantly associated with risk of CHF readmission in the final multivariate model. These predictors, however, provided a significant gradient in risk for CHF readmission.

Predicting factors for CHF readmission 

The current study showed that approximately 40% of patients with CHF were readmitted during the follow-up period of 2.4 years. This rate is comparable with the reported values in the previous studies.¹, ², ³, ⁴ Several previous studies have identified the predictors associated with an increased likelihood of being readmitted because of exacerbation of CHF symptoms, which included advanced age, prior hospital admission, length of hospital stay, severity of illness, and medical comorbidities.¹, ², ⁴, ⁶ Krumholz et al⁵ identified predictors of readmission in unselected, consecutively hospitalized Medicare patients with CHF, including male sex, prior admission within 6 months of the index admission, comorbidity, and length of stay more than 7 days.

In the current study, prior history of admission for CHF and longer hospital stay on the index admission were significant independent predictors of readmission. Longer hospital stay has been shown to be a risk factor for readmission in several previous studies.², ⁴, ⁵ The longer hospitalization is an indicator of severity or complexity of illness. However, the severity of CHF (NYHA functional class and LV EF by echocardiography) as well as other medical comorbidities including diabetes mellitus, renal failure, and stroke were not different between readmitted and nonreadmitted patients in our study population. Consistent with previous studies,⁴, ¹⁰ hypertension was also one of the medical correlates for readmission. Hypertension may participate in the deterioration of CHF by increasing afterload of patients. In contrast, age, sex, and disease severity were not predictive of readmission in our study patients.

The current study demonstrated that, in addition to the demographic and medical risk factors, the socioenvironmental factors related to postdischarge patient support and care could be independent predictors for readmission. The socioenvironmental factors are poorly documented in the medical records, so that their importance may likely be underestimated. These factors should be assessed to identify the high-risk patients.

The current study has demonstrated that poor follow-up is a strong predictor for CHF readmission. Patients with few numbers of follow-up visits had a 5-fold odds increase in the risk for CHF readmission. This association could not be explained by disease severity on admission or the presence of comorbidities because there were no significant differences in these variables between readmitted and nonreadmitted groups. It is conceivable that regular visits to clinics can improve the compliance of CHF patients with CHF, especially that to the medical treatment.¹¹, ¹² The patients who had occupations were less likely to be readmitted. This result may be related to the higher physical activity of these patients. Further, an interaction between medical and socioenvironmental variables could not be excluded. In fact, there was a relationship between older age (>70 years) and no occupation (P < .01).

The patients provided with postdischarge professional support had a lower likelihood of readmission although it was not a significant predictor by the logistic regression model. Poor professional support may worsen clinical condition through lack of needed help and services as well as advice. Although approximately 80% of patients had family members who were responsible for the posthospital patient care, the high readmission rate of the patients with family caregivers and the lack of significance in the family support between readmitted and nonreadmitted patients may result from the family caregiver's inability to provide sufficient care for the patient. This might be due to the lack of appropriate home-based care by the family member.¹³, ¹⁴ Therefore professional support is more effective in reducing the readmission rate than is nonprofessional support by family caregivers. There were no significant differences in the demographic and clinical variables between patients who received professional support and those who did not, except for sex. Men received professional support less frequently compared with women (P = .046). The reasons the patients had no professional support might be inadequate information concerning the availability of support or the patient's unwillingness to receive the support.

The current study indicates that socioenvironmental factors are of prognostic importance in patients with CHF. These results are consistent with those of the previous study, that social network is useful in reducing readmission in cardiac patients.¹ Chin and Goldman⁶ have identified single marital status as an independent correlate of readmission, which may be a proxy for poor social support. Social factors have also been shown to be important predictors of morbidity and mortality in patients with coronary artery disease.¹⁵, ¹⁶ The importance of social support has been confirmed by a recent study: the absence of emotional support is a strong, independent predictor of cardiovascular events (death and readmission) among elderly patients hospitalized with CHF.¹⁷ Further, it is also supported that multidisciplinary interventions with psychosocial support components for patients with CHF could reduce the readmission rate within 90 days of discharge from 42.1% to 28.9%.¹⁰

Implications 

The importance of socioeconomic factors for hospitalization for chronic HF in this study has important clinical implications. These findings offer a means of better understanding the mechanisms underlying exacerbation of chronic HF and indicate that readmission is caused by the interplay between medical and social factors in the patient, family, and caregivers. Further, they indicate that the establishment of postdischarge social support and medical follow-up is clearly an important part of effective discharge planning and education. Interventions should be performed for the patients themselves and for family members during hospitalization to prevent readmission, such as providing adequate patient and family support (professional support and caregiver) and establishing regular follow-up by the medical providers. The effectiveness of such interventions has been supported by the reduction of readmission in patients followed up at HF clinics.¹⁸ In addition, a home-based intervention has the potential to decrease the rate of readmission, prolong survival, and improve quality of life in patients with CHF.¹⁹ Therefore home-based patient management should also be incorporated into the care for patients with CHF.

Limitations 

There are several limitations to be acknowledged in this study. First, data collection was based on retrospective medical record review, which may have reduced the ability to identify predictors for readmission. Some potentially important factors are poorly documented in the medical records, especially those related to socioenvironmental variables. Therefore we collected the data not only from the medical record but also from the postdischarge interviews with the patient and family. Second, this study was performed among a relatively small number of patients hospitalized at 5 hospitals, and caution must be used in generalizing our findings before they are validated in other settings. Third, although we identified 5 predictors independently associated with hospital readmission, the effectiveness of such interventions in high-risk patients, not only during hospitalization but also after discharge, focusing on the socioenvironmental factors, should be tested by an appropriately designed trial. Multidisciplinary approaches described by Rich et al¹⁰ appear to provide an important opportunity to improve the outcomes of these patients. Our results demonstrate the widespread need for these programs in all patients hospitalized with CHF. Fourth, although the current study focused primarily on the HF-related readmission, there are other important outcomes, especially in elderly individuals with severe CHF, including functional status and quality of life. Further, the all-cause hospital readmission rate should also be investigated, particularly in elderly patients with CHF. Fifth, the long hospital stay in our study patients might be related to the differences in the health care system in Japan because the demographic and clinical variables are similar between our Japanese patients and those reported from the United States. Caution must be used in generalizing our current results to different cohorts of patients.

In conclusion, patients hospitalized with CHF have a high risk for readmission after discharge. Patients with a history of hospitalization as a result of CHF, longer hospital stay, and a history of hypertension are at increased risk for readmission, and our data suggest that socioeconomic factors, including poor follow-up visits, poor professional support, and no occupation, are also potentially important predictors. Therefore a systematic CHF patient management system that coordinates care in the hospital, outpatient, and home settings is clearly needed to reduce the morbidity and mortality of patients with CHF and thus lower the overall costs for the treatment of these patients.

Back to Article Outline

Acknowledgements 

Participating investigators constituting the study hospitals are as follows; Samon Koyanagi, MD (National Kyushu Medical Center Hospital); Tetsuji Inou, MD, and Masami Matsuyama, RN (Fukuoka Red Cross Hospital); Yuji Maruoka, MD (Hamanomachi Hospital); Yusuke Yamanoto, MD, and Koji Todaka, MD (Saiseikai Fukuoka General Hospital). This study could not have been carried out without the help, cooperation, and support of the cardiologists in the study hospitals. We also thank the administrators for allowing us to obtain the data and for providing secretarial help.

Back to Article Outline

References 

1. Berkman B, Dumas S, Gastfriend J, et al.  Predicting hospital readmission of elderly cardiac patients. Health Soc Work. 1987;12:221–228
   • View In Article
   • MEDLINE
2. Gooding J, Jette AM. Hospital readmissions among the elderly. J Am Geriatr Soc. 1985;33:595–601
   • View In Article
   • MEDLINE
3. Rich MW, Freedland KE. Effect of DRGs on three-month readmission rate of geriatric patients with congestive heart failure. Am J Public Health. 1988;78:680–682
   • View In Article
   • MEDLINE
   • CrossRef
4. Vinson JM, Rich MW, Sperry JC, et al.  Early readmission of elderly patients with congestive heart failure. J Am Geriatr Soc. 1990;38:1290–1295
   • View In Article
   • MEDLINE
5. Krumholz HM, Parent EM, Tu N, et al.  Readmission after hospitalization for congestive heart failure among Medicare beneficiaries. Arch Intern Med. 1997;157:99–104
   • View In Article
   • MEDLINE
6. Chin MH, Goldman L. Correlates of early hospital readmission or death in patients with congestive heart failure. Am J Cardiol. 1997;79:1640–1644
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (215 KB)
   • CrossRef
7. Krumholz HM, Chen YT, Wang Y, et al.  Predictors of readmission among elderly survivors of admission with heart failure. Am Heart J. 2000;139:72–77
   • View In Article
   • Abstract
   • Full-Text PDF (502 KB)
   • CrossRef
8. Tsuchihashi M, Tsutsui H, Kodama K, et al.  Clinical characteristics and prognosis of consecutively hospitalized patients with congestive heart failure: a study in Fukuoka, Japan. Jpn Circ J. 2000;64:953–959
   • View In Article
   • MEDLINE
   • CrossRef
9. Ho KK, Anderson KM, Kannel WB, et al.  Survival after the onset of congestive heart failure in Framingham Heart Study subjects. Circulation. 1993;88:107–115
   • View In Article
   • MEDLINE
10. Rich MW, Beckham V, Wittenberg C, et al.  A multidisciplinary intervention to prevent the readmission of elderly patients with congestive heart failure. N Engl J Med. 1995;333:1190–1195
    • View In Article
    • MEDLINE
    • CrossRef
11. Rich MW, Gray DB, Beckham V, et al.  Effect of a multidisciplinary intervention on medication compliance in elderly patients with congestive heart failure. Am J Med. 1996;101:270–276
    • View In Article
    • Abstract
    • Full-Text PDF (735 KB)
    • CrossRef
12. Weinberger M, Oddone EZ, Henderson WG. Does increased access to primary care reduce hospital readmissions? Veterans Affairs Cooperative Study Group on Primary Care and Hospital Readmission. N Engl J Med. 1996;334:1441–1447
    • View In Article
    • MEDLINE
    • CrossRef
13. McMurray JJ, Stewart S. Nurse led, multidisciplinary intervention in chronic heart failure. Heart. 1998;80:430–431
    • View In Article
    • MEDLINE
14. Grady KL, Dracup K, Kennedy G, et al.  Team management of patients with heart failure: a statement for healthcare professionals from the cardiovascular nursing council of the American Heart Association. Circulation. 2000;102:2443–2456
    • View In Article
15. Berkman LF, Leo-Summers L, Horwitz RI. Emotional support and survival after myocardial infarction: a prospective, population-based study of the elderly. Ann Intern Med. 1992;117:1003–1009
    • View In Article
    • MEDLINE
16. Bucher HC. Social support and prognosis following first myocardial infarction. J Gen Intern Med. 1994;9:409–417
    • View In Article
    • MEDLINE
    • CrossRef
17. Krumholz HM, Butler J, Miller J, et al.  Prognostic importance of emotional support for elderly patients hospitalized with heart failure. Circulation. 1998;97:958–964
    • View In Article
    • MEDLINE
18. Erhardt L, Cline C. Heart failure clinics: a possible means of improving care. Heart. 1998;80:428–429
    • View In Article
    • MEDLINE
19. Stewart S, Marley JE, Horowitz JD. Effects of a multidisciplinary, home-based intervention on unplanned readmissions and survival among patients with chronic congestive heart failure: a randomised controlled study. Lancet. 1999;354:1077–1083
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (92 KB)
    • CrossRef