This report is based on a systematic review of the literature on non-pharmacological interventions for post-discharge care that are currently in use. The approach and criteria used in this technology assessment were agreed upon by consensus of the EPC, CMS, and AHRQ. Several published systematic reviews on this topic indicate that a large number of randomized controlled trials (RCTs) are available. Therefore, we focus only on RCTs in this report.
We searched Medline, Medline In-Process & Other Non-Indexed Citations, the Cochrane Library, CINHAL, and Ovid HealthStar from 1990 through July 2007 to identify articles relevant to the key question. We limited the start date of the search to 1990; the time point associated with rapid advances in the medical management of HF, and changing healthcare needs of patients and changes in practice patterns. We reviewed reference lists of published systematic reviews on the same topic, selected narrative reviews, and retrieved primary articles for potentially relevant articles. We also searched ClinicalTrials.gov for registered but not yet published studies and contacted authors of these trials. We combined search terms for discharge plan, case management, and heart failure, and limited the search to English-language RCTs in adult humans. (See Appendix A for complete search strategy.)
We included all comparisons of interventions for post-discharge care that reported at least readmission and/or mortality outcomes among patients with HF. We assessed titles and/or abstracts of citations identified in literature searches. A low threshold was used to retrieve articles for evaluation. Full-text articles of potentially relevant abstracts were retrieved and reviewed according to the criteria. Results published only in abstract form were not included in our reviews because these reports have generally not been peer-reviewed and therefore lack adequate information to assess the validity of the data.
The population of interest for this report is adults with diagnosed HF. Although the Medicare-eligible populations are age 65 years and above, in consultation with CMS, we used a lower threshold for limiting studies with respect to the age cut-off. Studies were accepted if the mean age of the population was at least 50 years. We also included studies that had subjects who were recruited during the index hospitalization, at the time of discharge from the index hospitalization, or at the outpatient HF clinic. In general, the studies excluded patients discharged to settings other than home (i.e., nursing home or long-term facility) because of the difficulty in followup. We excluded studies with less than 10 patients per arm and studies of secondary analysis of an RCT. When there were multiple publications of the same study, we used the publication that reported the largest number of subjects. We excluded two studies that were included in previous systematic reviews (one had an incorrect citation,26 and the other could not be retrieved.27)
The site of initiation of an intervention was the key feature of interest among the settings. In the first setting, the intervention(s) was initiated while patients were still in the hospital prior to discharge. In the second setting, the intervention started for recently discharged patients in the outpatient clinic. In the third setting, the intervention was initiated with patients who had been attending a HF management clinic, and at least a proportion of those have had a recent hospitalization within 3 months of study recruitment. However, in this report, the patients recruited from the third setting were analyzed separately.
We accepted descriptions of the components of interventions as they were reported in published articles. Interventions were implemented in hospitals, outpatient settings or in patients' homes. Studies of interventions with any duration were included.
Education on symptoms and disease management
Patients received a simple explanation of the pathophysiology of HF, symptoms, and treatment of HF and were offered advice about when to seek expert help.
Instruction on self management behaviors
Instruction to increase the self-care behavior in patients and includes assessment of self-care ability. Instruction occurs about behaviors to be taken in relation to early signs of worsening HF. The goal is for the patient to understand the role of fluid retention in worsening symptoms and to seek care early. HF diaries or notebooks are sometimes provided to aid self-management.
Diet advice, sodium restriction
Patients were advised to restrict their sodium intake to 1.5 to 2 grams per day, and instructed on how this level could be achieved. The intervention may involve a dietician's visit and an individualized sodium restricted diet or a list of dietary recommendations.
The name of each drug and its purpose, dosage, frequency, and significant adverse effects, of specific HF drugs is reviewed by a care provider, and HF drugs (e.g. beta blockers) up titrated slowly as tolerated. It may involve a written medication schedule for complicated medication regimens and increase a patients' adherence to prescribed medications.
Review of education goals with patients on subsequent clinic visits or telephone calls was designed to increase the self-care behavior in patients and included an assessment of self-care ability.
This includes recommendations about moderate aerobic exercise training, which can result in improved exercise duration, less fatigue, faster pace of activities, and improved general well-being.
Instructions are given to the patients on weight monitoring that may include daily or regular monitoring, and provide HF failure diaries or notebooks to document weight for self-management.
A care provider telephoned the patients in weekly or biweekly intervals for the proactive telephone followup of patients at home after discharge from the hospital. This may involve a review of symptoms and weight and/or questions, which follow a computerized care plan. Support may be provided by a non-physician, usually a nurse, but could also involve a dietician or pharmacist.
Increased clinic visits
Increased clinic visits were additional clinic visits to assess or in response to a change in clinical stability.
A member of the multidisciplinary HF team (usually a HF specialist nurse, a pharmacist or dietician) visited the patient in his/her home to assess clinical stability and implement care to correct any variation from care plan.
Social and psychological support
The level of family or career support that is available to the patient was assessed and recommendations were provided to increase support where necessary. Both forms of support are expected to help patients deal better with their social function (e.g., in their role in the family) and, as a result, influence the psychological state (e.g., anxiety or depression) of the patient as he/she copes with the disease.
All members of the multidisciplinary care team are involved in the care of HF patients. The followup visits include consultation with a cardiologist, HF specialist nurse, physiotherapist, dietician, and psychologist.
The acceptable comparators of interest included those defined as usual care, routine care, or standard care. Usual care is generally not structured and consists of instruction on discharge medications by the unit nurse or possibly specialist HF nurse, and includes information on the next appointment with the cardiologist or primary care physician. There were many variations in the description of the usual care among the studies and in general, usual care was not well defined. We also included studies that compared one intervention versus another type of intervention.
We restricted our evaluation to outcomes deemed clinically important and of greatest interest to CMS. Outcomes of interest included:
- Readmission, all causes
- Mortality, all causes
- Length of hospital stay
- Quality of life, using any validated quality of life measures or any measures of symptom relief
- Other outcomes such as costs or combined endpoint of mortality and hospitalization
Items extracted included study year, country where the trial was conducted, setting, funding source, study design, eligibility criteria, patient characteristics, components of intervention, and outcomes (see Appendix B for a sample data extraction form). We recorded the method of randomization, allocation concealment, blinding, and whether results were reported on an intention-to-treat basis. Details regarding the duration of HF, history of myocardial disease or other cardiovascular disease, angiotensin converting enzyme (ACE) inhibitor and beta-blocker use, and the intervention were also extracted. For each outcome of interest, baseline, followup, and change from baseline data were extracted, including information on statistical significance. For most outcomes, only data from the last reported time point was included.
We assessed the methodological quality of studies based on predefined criteria. We used a 3-category (A, B, and C) grading system to denote the methodological quality of each study (defined below). This generic scheme defines a grading system that is applicable within a study design. For randomized trials, we mainly considered the methods used for randomization, allocation concealment, and blinding, as well as the use of intention-to-treat analysis and the report of dropout rate. We also considered the extent to which valid primary outcomes were described, as well as whether information was clearly reported. Studies were not rejected due to poor quality.
Category A studies have the least bias, and their results are considered valid. A study that adheres mostly to the commonly held concepts of high quality includes the following: adequate random allocation method; blinded evaluation; clear descriptions of the population, setting, interventions, and comparison groups; appropriate measurement of outcomes; appropriate statistical and analytic methods and reporting; no reporting errors; less than 20 percent dropout; clear reporting of dropouts; and no other obvious bias.
Category B studies are susceptible to some bias, but not a sufficient amount to invalidate the results. They do not meet all the criteria in category A because they have some deficiencies, but they have not deficiencies that are likely to cause major bias. The study may be missing information, making it difficult to assess limitations and potential problems.
Category C studies have significant bias that may invalidate the results. These studies have serious errors in design, analysis, or reporting; large amounts of missing information; or major discrepancies in reporting.
Applicability addresses the relevance of a given study to the target population of interest. Every study applies certain eligibility criteria when selecting study subjects. Most of these criteria are explicitly stated (e.g., disease status, age, comorbidities). However, some criteria may be implicit or the recruited study population was affected by unintentional biases, such as those related to location (e.g., multi-center vs. single-center, intensive care vs. all inpatients). The populations and the interventions that are of interest specified by the key questions, as opposed to those of interest to the study investigators, dictate the applicability of a study.
We categorized study populations into 1 of 3 levels of applicability that are defined as follows:
Wide Sample is representative of the Medicare population in relevant settings. Patients' age (older adult), gender, spectrum of disease severity and type, etc. are representative of the population of interest.
Moderate Sample is an important subgroup of the population of interest. It is possibly limited to a narrow or young age range, lower New York Heart Association (NYHA) score, type of disease, gender etc.
Narrow Sample represents only a narrow, atypical subgroup of the population of interest.
Data Synthesis and Meta-analyses
For studies with binary outcomes, we calculated the relative risk and 95% confidence interval. For studies with continuous outcomes, we calculated the mean difference and standard error of the mean difference.
There was significant heterogeneity among studies, with regard to intervention components, intervention duration, or other factors. In the presence of significant clinical heterogeneity among studies of any interventions compared with usual care, we did not perform meta-analyses to obtain a single effect size estimate for the clinical outcomes of interest.
However, certain individual interventions, specifically, education on symptoms and disease management, instruction on self management, dietary advice (including sodium restriction), medication review, exercise recommendations and weight monitoring have become usual care. Thus, interventions consisting of increased access to providers — telephone support, clinic visits, home visits, home telemonitor, and multidisciplinary discharge care — were deemed interventions of interest. To assess the effectiveness of intervention compared with usual care on readmission, we performed meta-analyses by grouping studies that utilized similar interventions for post-discharge care.
We used DerSimonian and Laird's random effects model for all meta-analyses, which generally provides a wider confidence interval (more conservative) in the presence of heterogeneity across studies.28 We tested for heterogeneity using Cochran's Q (statistically significant at P <0.05) and assessed its extent with I2, which evaluates the proportion of between study variability that is attributed to heterogeneity rather than chance.29
Subgroup analyses were performed to examine the impact of study quality, followup duration, site of initiation of interventions (inpatient versus outpatient), the effects of duration of intervention (<3 months, 3 to 6 months, >6 months), country where the study was conducted (US versus non US), and severity of HF on the rate of readmission.