Using the California Office of Statewide Health Planning and Development (OSHPD) files for 2007, we examined the relationship between certain patient and hospital characteristics and 7-day bounce-back admission following discharge from California EDs. An admission following ED discharge may have important implications regarding the ED care and follow-up of patients after discharge. We found older white males and patients with a disposition of eloped or having left against medical advice (AMA) especially at risk for a bounce-back admission. We also found that use of Medicaid or Medicare was associated with higher rates of bounce-back admission. We identified a number of primary discharge diagnoses associated with bounce-back admission and found that hospitals more likely to discharge patients that get admitted within a week of discharge either have for-profit status or a teaching affiliation. Interestingly, we found that a majority of bounce-back admission diagnoses mirror the diagnosis of the index ED visit.
To our knowledge our study is the first large scale evaluation of 7-day bounce-back admissions in the United States involving adults age 18 and older. Existing literature conducted in the US provides an important foundation and reports a bounce-back admission rate of 0.06% to 1.3%, but is limited by sample size, measurement bias, prolonged follow-up time, and focus on a population not generalizable to the entire US.8–10,12,14,32
Studies conducted in foreign countries are limited by differences in health care systems, patient population, and sample size.11,17,22
We report a bounce-back admission rate of 2.6%, suggesting that more than 1 of every 50 patients discharged from an ED require an admission within 7 days. We found that the patients at highest risk for a bounce-back admission are the most vulnerable and include the very old (age 80+ OR 2.82, 95% CI 2.76–2.9), individuals with Medicare insurance (OR1.53, 95%CI 1.50–1.55), or the underinsured (California’s Medicaid program) (OR1.42, 95%CI 1.40–1.45). Previous studies have suggested that older patients with poor physical functioning often recover poorly following an ED visit.8,12,16,33
Older patients also have a higher baseline risk due to an increased comorbidity burden. This combined with our findings suggests that when evaluating older patients who are less able to care for themselves, special attention should be paid to their ability to recover and their support mechanisms following the ED visit.
A patients’ insurance is not a direct measure of the need for services; nonetheless, we found that despite controlling for age and diagnosis, those with Medicare or Medicaid insurance had a higher likelihood of bounce-back admissions. For Medicare patients, this could represent unmeasured co-morbidities or be attributed to non-elderly individuals with Medicare such as persons with renal disease or a disability. The association between Medicaid use and bounce-back admission may be attributed to the increasing dependence of Medicaid patients on the ED for care34,35
and the possibility that these patients are not receiving the needed follow-up care they require following discharge due to a limited availability of willing providers that use Medicaid. Our findings strengthen the notion that a key component of the ED evaluation of vulnerable patients is an assessment of their access to care both prior to and following the ED visit.
We confirm the notion that patients who leave against medical advice or elope are at high-risk for worse outcomes following the ED visit. Contrary to popular belief, these patients are not less ill than other ED patients but are found to leave the ED as a result of other factors such as insurance status or social reasons.36–39
Although not in the control of the ED provider, we recommend that when encountering a patient who may leave prior to completion of the ED visit, all attempts to prevent the patient from leaving be made.
We discovered that hospitals most likely to discharge patients that require bounce-back admissions are for-profit (OR 1.2, 95% CI 1.1–1.3) or have a teaching affiliation (OR 1.2, 95% CI 1.0–1.3). This phenomenon can be attributed to several factors. Historically, for-profit hospitals have been found to rely heavily on a private patient payer mix. 40,41
Since the implementation of the Emergency Medical Treatment And Labor Act in 1986 mandating the emergent evaluation of all patients, it is possible that for-profit hospitals are stabilizing patients with emergencies, but due to financial strains, administrative pressures, and variable availability of consultants, are prematurely discharging these patients. Teaching hospitals often see patients with more complex medical needs, such as transplant recipients,42
that require more frequent admissions and are staffed by trainees, resulting in variability of diagnostic accuracy and admission practices.
We identify the top discharge diagnoses that place patients at greater risk of bounce-back admission. Although the diagnoses of chronic and end stage renal disease (CRD PP 6.18%, ESRD PP 5.57%), congestive heart failure (OR 2.5, 95% CI 2.3–2.6), and blood disease disorders (OR 2.4, 95% CI 2.2–2.6), which includes anemia, sickle cell disease, coagulation defects, and diseases of white blood cells, describe chronic conditions that often require regular encounters, the need for a short-term bounce-back admission could indicate a more concerning disease process or a limitation of appropriate follow-up care following the ED visit. An especially concerning diagnosis is chronic or end stage renal disease. Chronic renal disease includes a diagnosis of nephritis, nephrosis, renal sclerosis, acute renal failure, and chronic renal failure, not end stage. Both include conditions that may seem stable but in reality harbor more devastating disease processes. Emergency physicians evaluating patients with advanced stages of renal disease should pay particular attention to the whole patient and the potential role renal disease could play in their presenting chief complaint.
Our findings suggest that bounce-back admission diagnoses often mirror the index ED visit diagnosis. This suggests that patients visited the ED with a complaint that could have been incompletely managed either during the visit or shortly following discharge. A similarly concerning finding was that mental illness, which includes the diagnosis of substance abuse, psychosis, dementia, and developmental delay was not only a common bounce-back admission diagnosis but also common as a subsequent admission diagnosis for a large number of prior ED visit encounters. The literature has found that patients with these conditions have a greater tendency to return to hospitals and EDs for a variety of reasons.43,44
Patients with a mental illness are a vulnerable population at risk for being misdiagnosed due to errors in communication and estimation of risk as well as lacking the ability to receive proper follow-up.
A surprising finding was that symptomatic diagnoses such as chest pain, headache, or syncope were not associated with bounce-back admission. This is confirmed in a previous study our group conducted evaluating poor outcomes after discharge from the ED.24
A possible explanation is that ED physicians assessing these patients exercise a greater amount of caution and only discharge the patients who appear to be at lowest risk for poor outcomes.
Bounce-back admissions may be an indicator of incomplete ED or follow-up care and have important policy and quality improvement implications. Our study found that a majority of bounce-back admissions have the same diagnosis as the index ED visit and could reflect care given during the index ED visit. We identify important patient and hospital characteristics associated with bounce-back admissions within 7 days of ED discharge. Our findings suggest that quality improvement efforts focus on high-risk individuals, such as the old or patients with renal disease, and that the disposition plan of patients include consideration of vulnerable individuals.