Readmission Risk Factors after Hospital Discharge Among the Elderly

Abstract

Hospital readmission rates among the elderly are attracting increasing attention. Readmission is costly, especially as proposed new guidelines could deny reimbursement for readmissions. Identifying key factors at discharge that can serve as prognostic indicators for readmission is an important step toward developing and targeting interventions to reduce hospital readmissions rates. Published literature has listed predominantly demographic, clinical, and health care utilization characteristics to describe the factors that put the elderly at risk. However, additional factors are proposed that include social, clinical, individual-level, environmental, and system-level factors. Multimodal interventions have been tested and some reduction in readmissions has been shown. Whether these additional factors might lead to a further reduction remains unclear. In addition to possible factors at discharge, factors identified after the patient has been discharged also must be identified and addressed. The patient safety literature characterizes factors that put the elderly at risk for adverse drug events, which function as antecedent factors for readmission and likely include the environmental and system-level factors. Synthesizing these factors from the readmission and patient safety literature provides the basis to develop a more comprehensive conceptual framework to identify research gaps aimed at reducing hospital readmissions among the elderly. (Population Health Management 2012;15:338–351)

Introduction

H ospital readmission is a growing topic in health care reform and hospitals are taking aim at reducing readmission rates and associated costs. The cost to Medicare of unplanned rehospitalizations in 2004 was $17.4 billion with 2.3 million, or 20%, of Medicare beneficiaries readmitted to the hospital within 30 days.¹ “Readmission” or rehospitalization is defined as a return to the hospital shortly after discharge from a recent hospital stay. Readmission rates track how often patients are readmitted to the hospital shortly after being discharged and are commonly used as a measure of hospital care.

Prospective and retrospective studies have reported a wide range of readmission rates. Although readmission rates vary in published literature, the elderly (60 years of age and older) consistently have the highest rate of hospital readmission compared to other age groups.^2
–4 Published studies have associated readmission rates primarily with patient demographics, chronic conditions, comorbidity, and utilization factors. However, Medicare patients have many more reasons for readmission to hospitals. Adverse events (AEs), an injury resulting from medical management rather than the underlying disease, can occur before or during hospitalization, at discharge, or at home. Moreover, the specific reasons for readmission caused by AEs may be the result of patient-related factors, provider factors, or health system-related factors. Medication errors are a major contributor to AEs within and outside the hospital setting that lead to increased health care utilization (ie, clinic visit, emergency department [ED] visit, hospitalization).

According to the 2008 report to Congress, medication errors after discharge are not uncommon and contribute to readmissions.⁵ Older adults (65 years of age or older) are twice as likely as others to seek care in EDs for adverse drug events (ADEs; more than 177,000 emergency visits each year) and are nearly 7 times more likely to be hospitalized after an emergency visit.⁶ However, providers may not recognize a patient's symptoms to be the result of an underlying ADE or may fail to include ADEs as a reason for readmission or as one of the diagnoses. Consequently, identification of ADEs is subject to underreporting and misclassification; therefore, it is less recognized in the readmission literature.

There are similarities in those risk factors associated with medication errors and with hospital readmissions among the elderly in published patient safety and readmission literatures. Furthermore, the risk factors for medication errors also contribute as antecedents to other causes of readmission. Integrating this information from the readmission and the patient safety literature may prove useful to identify risk factor patterns in readmissions among the elderly. Therefore, the purpose of this article is to summarize the similarities and patterns in the literature on the factors that put the elderly at risk for readmission and at risk for ADEs. This will provide a starting point to identify gaps and areas in need of more research.

A literature search was undertaken, using PubMed and the Cochrane Database, for articles describing the incidence of and contributors to readmission after discharge and of ADEs among the elderly. To summarize the factors that put the elderly at risk for readmission, articles that characterized or identified the incidence, contributors, and causes of hospital readmission or ADEs were selected. These included review articles, epidemiological articles, meta-analyses, and clinical investigations. Additionally, prevalence of prescription use and associated populations were obtained from published Web sites such as the Slone Epidemiology Center and the National Health and Nutrition Examination Survey Web site.

This article is organized from a patient-centered focus and addresses the factors shown in Figure 1.

FIG. 1.

Conceptual diagram depicting the factors affecting readmission in the elderly.

Incidence of Readmissions and Adverse Drug Events

Adverse drug event rates

In a study conducted between 1990 and 1993, ADEs led to more than 7000 deaths, 1.5 million injuries, and 700,000 ED visits per year, which translated to loss of life and more than $77 billion in avoidable health care costs.⁷ Data from US health care facilities reported that approximately 3.5 million patients were treated for ADEs in ambulatory care settings.⁸ Estimates based on extrapolations and meta-analyses indicate that up to 1.5 million patients are treated for ADEs in EDs,⁹ and approximately 1.5 million patients are hospitalized each year for ADEs.¹⁰ More recent population-level estimates of outpatient ADEs using the National Center for Health Statistics found that the rate of ADEs is rising, with more than 4.3 million persons seeking medical care for an AE annually.¹¹ The incidence of ADEs requiring medical treatment increased substantially between 1995 and 2005; in particular, patients ages 65 years and older had an incidence of ADE visits as high as 1 in 20 persons seeking related medical care.¹¹

Although untoward drug effects concern all segments of society, ADEs are especially prevalent among the elderly.^9,12,13 The estimated annual population rate of ADEs requiring hospitalization was nearly 7 times the rate for persons younger than 65 years of age.⁶ Patients aged 65 years or older accounted for 37.0% of estimated unintentional injury visits requiring hospitalization and 48.9% of estimated ADE visits requiring hospitalization.⁶ Patients aged 65 years or older were more than twice as likely to be treated for ADEs in EDs, and more of these visits were made by women.⁶

Readmission rates

Readmission rates are reported mostly at 30-, 60-, or 90-day intervals after discharge. Thirty-day readmission rates ranged from 12% to 19.6%; 90-day readmission rates ranged from 23% to 34%, and admission rates related to AEs ranged from 13% to 49% (Table 1).^{1,3,4,6,14

–18}

Table 1.

Readmission Rates

Source	Setting/Study Population	Design	Number of Patients	Readmission/Adverse Event-Related Admission^ Rates*
Jencks et al, 2009¹	National sample	Retrospective: Medicare claims data from 2003–2004	11,855,702	20% (30-day) 34% (90-day)
Wier et al, 2011⁴	15 selected states with 42% of the total US population	Retrospective: HCUP data on all-cause readmissions in 2008.	8.5 million	19% (30-day)
Joynt et al, 2011¹⁴	Patients discharged from US hospitals between 2006–2008	Retrospective: National Medicare data	3.1 million	13% (30-day)
Allaudeen et al, 2011³	550-bed tertiary care academic medical center. Mean age: 59.6 years	Retrospective: observational study of an administrative database	6805	17% (30-day)
Smith et al, 2000¹⁵	Nine VA medical centers	Retrospective: secondary analysis from RCT	1378	23% (90-day)
Mudge et al, 2011¹⁶	Australian teaching hospital, 2006–2007	Prospective cohort study.	142	32%–40% (6-month) 39% (90-day)
Bero et al, 1991¹⁷	Medicare patients in a community hospital	Retrospective: secondary analysis from RCT	706	^*35% (90-day)
Budnitz et al, 2006⁶	63 US hospitals participating in the National	Retrospective NEISS-AIP data	701,547	^*49%
	Electronic Injury Surveillance System–
	All Injury Program (NEISS-AIP), 2004–2005
Forster et al, 2003¹⁸	A tertiary care US academic hospital.	Prospective cohort study	400	^*19%

HCUP, Healthcare Cost and Utilization Project; RCT, randomized controlled trial; VA, Veterans Administration.

Factors that Contribute to Readmission in the Elderly

Risk factors for hospital readmission include sociodemographic and clinical characteristics, individual patient-level factors, environmental trends in outpatient care, and system-level factors.

Sociodemographic factors

Sociodemographic factors that contribute to hospital readmission encompass age, sex, socioeconomic status, education, social support/resources, race/ethnicity, insurance, financial, and access to/availability of services. Some studies have associated readmission and ADEs among the elderly with race, sex, and insurance type. Among the elderly, African American race and Medicaid as payer status were associated with readmission, and drug-related events were higher among women (Table 2).^3,9,14,19 However, one study with a small sample size (n=142) that included patients aged 50 years and older with previous hospitalization did not detect a significant influence of age and sex on readmission.¹⁶

Table 2.

Sociodemographics

Source	Setting/Study Population	Study Design	Number of Patients	Sociodemographic Factors	Readmission/Adverse Event^ Rates (95% Confidence Interval)*
Allauden et al, 2011³	550-bed tertiary care, academic medical center. Mean age: 59.6 years	Retrospective: observational study of an administrative database	6805	African American	43%
				Medicaid	15%
Joynt et al, 2011¹⁴	Patients discharged from US hospitals between 2006–2008	Retrospective: National Medicare data	3.1 million	African American	13% higher odds
				Minority serving hospitals vs. non-minority serving	23% higher odds
Hafner et al, 2002⁹	University-affiliated tertiary care setting with emergency department (ED) visit between March 1 and May 31, 1997,	Retrospective: ED charts: matched control for: age for disposition, survival, severity, payer, sex, race, age, number of drugs, and total cost	434	Female	^*Odds ratio:1.48
					(1.01 to 2.16)
Forster et al, 2004¹⁹	A multisite Canadian teaching hospital, medicine service discharged patients	Prospective	328	Female	^*70%

Clinical characteristics

Health conditions and clinical characteristics (eg, type of diagnosis and chronic illnesses, presence of comorbidities, duration of disease, age-related physiological changes) affecting pharmacokinetics and pharmacodynamics also are associated with readmissions. Overall, the most common diagnoses associated with readmission were heart failure, depression, chronic obstructive pulmonary disease, coagulopathy, and diabetes (Table 3).^{1,3,9,15,16,19,20} In a meta-analysis by Soeken and colleagues, patients with chronic illnesses had a mean readmission rate of 34%.²¹ Similarly, patients with 5 or more medically comorbid conditions had more than twice the likelihood of an unplanned readmission within 30 days than patients without those conditions ^16,20

Table 3.

Clinical Characteristics

Source	Setting/Study Population	Study Design	Number of Patients	Clinical Factors	Readmission/^Adverse Drug Event Related to Admission Rates (CI 95%)*
Jencks et al, 2009¹	National sample	Retrospective: Medicare claims data from 2003–2004	11,855,702	Heart failure	27%
				Psychoses	25%
				COPD	23%
				Pneumonia	20%
				GI conditions	19%
Allaudeen et al, 2011³	550-bed tertiary care academic medical center. Mean age: 59.6 years	Retrospective: observational study of an administrative database	6805	Inpatient use of:
				Narcotics	OR: 1.33
					(1.16–1.53)
				Corticosteroids	OR: 1.24
					(1.09–1.42)
				• Congestive heart failure	OR: 1.30
					(1.09–1.56)
				• Renal disease	OR: 1.19
					(1.05–1.36)
				• Cancer (without metastasis)	OR: 1.95
					(1.54–2.47)
				• Weight loss	OR: 1.26
					(1.09–1.47)
				• Iron deficiency anemia	OR: 1.27
					(1.13–1.44)
Smith et al, 2000¹⁵	Nine VA medical centers-Medical Service	Retrospective: secondary analysis from RCT	1378	CHF	23%–35%
				COPD	28%–38%
				DM	15%–26%
				Higher blood urea nitrogen	1.01
					(1.004, 1.020)
				COPD	1.33
					(1.166, 1.524)
Mudge et al, 2011¹⁶	Australian teaching hospital acute medical wards, February 2006-February 2007	Prospective cohort study	142	CHF	69%
				DM	67%
				Chronic lung Dz.	63%
				Chronic Dz diagnosis	3.4
					(1.3–9.3)
				Underweight (BMI<18.5)	12.7
					(2.3–70.7)
				Depressive symptoms	3.0
					(1.3–6.8)
Forster et al, 2004¹⁹	A multisite Canadian teaching hospital, medicine service discharged patients	Prospective	328	Chronic conditions:
				A-Fib
				DM- type 2	^*2.0
					(0.9–3.6)
				Acute conditions:	^*1.9
					(1.0–3.6)
				Pneumonia	^*1.9
					(1.0–3.6)
				CHF	^*1.3
					(0.6–2.8)
				Acute renal failure	^*1.3
					(0.6–2.9)
Hafner et al, 2002⁹	University-affiliated tertiary care	Retrospective: ED charts: matched control for: age for disposition, survival, severity, payer, sex, race, age, number of drugs, and total cost	434	Common Dx:
	ED visits between March 1 and May 31, 1997			Hypoglycemia	^*40%
				Coagulopathy/hemorrhage	^*35%
				Rash	^*25%
				Common Medications:
				Insulin	^*40%
				Warfarin	^*20%
				Furosemide	^*11%
Marcantonio et al, 1999²¹	Medicare managed care plan patients admitted to an academic center	Matched case-control study from a clinical information system	308	5 or more medical comorbidities	OR: 2.6
					(1.5–4.7)
				Depression hx.	OR: 3.2
					(1.4–7.9)
				Lack of documented patient or family education	OR: 2.3
					(1.2–4.5)
				New medical problem	46%
				Relapse	44%
				Complication of initial treatment	27%
				ADR	9%
				Caregiver or facility problem	4%

ADR, adverse drug reaction; A-Fib, atrial fibrillation; BMI, body mass index; CHF, congestive heart failure; CI, confidence interval; COPD, chronic obstructive pulmonary disease; DM, diabetes mellitus; Dx, diagnosis; Dz, disease; GI, gastrointestinal; hx, history; OR, odds ratio; RCT, randomized controlled trial; VA, Veterans Administration.

A cross-sectional study of administrative data from 2002 to 2005 containing a maximum of 20 million Medicare and commercially insured patients for a 1-year period evaluated patients for suspected ADEs. This study found that circulatory, endocrine, nutritional, and metabolic systems ranked among the top 5 coexisting health problems in hospitalized patients.²² Other studies have associated outpatient drug-related harm with patients older than 65 years of age who take 3 or more medications or specific medications such as digoxin, warfarin, and insulin.^12,23

Medication class

Medication class also was associated with ADEs in the patient safety literature (Table 4).^{11,12,24
–26} Overall, the most common categories of drugs associated with ADEs were antimicrobial agents, diabetic agents, and cardiovascular agents. However, none of the ADEs associated with antimicrobials resulted in hospitalization.²⁵

Table 4.

Medication Class

Source	Setting/Study Population	Study Design	Number of Patients	Medication Class	Adverse Drug Event Rates
Gurwitz et al, 2003¹²	Multispecialty group practice with Medicare enrollees	Prospective cohort study between 1999–2000	27,617	Cardiovascular medications	24.5%
				Diuretics	22.15
				Non-opioid analgesics	15.4%
				Hypoglycemics	10.95
				Anticoagulants	10.2%
Bourgeois et al, 2010¹¹	National Center for Health Statistics Data in the United States; visits between 1995 and 2005	Retrospective	ADE-related visits were 3,544,908	Cardiotonic glycosides (Digoxin)	Hospital readmission
					12.8%; CI: 8.6–18.6%
				Anticoagulants	12.8%; CI: 8.6–18.6%
				Anticonvulsants	6.5%; CI: 3.7–11.2
Forster et al, 2005²⁷	Urban academic health sciences center	Secondary analysis of a prospective cohort study	400	Anti-infective	31%
				Corticosteroids	16%
				Cardiovascular	16%
				Analgesic/narcotics	11%
				Anticoagulants	9%
				Antiepileptic	4%
Budnitz et al, 2005²⁴	National Electronic Injury Surveillance System–All Injury Program hospital EDs between July to September 2002	Retrospective	598		Hospitalization
				Cardiovascular	23.1%
				Diabetic	17.3%
				Anticoagulants	15.4%
Budnitz et al, 2007²⁶	Nationally representative, National Electronic Injury Surveillance System–Cooperative	Retrospective (using Beer's criteria)	177,504 ED visits		ED Visit (%)
	Adverse Drug Event Surveillance System, 2004–2005; National			Warfarin	17.3 (12.7–21.9)
	Ambulatory Medical Care Survey, 2004; and National Hospital Ambulatory			Insulin	13.0 (9.4–16.6)
	Medical Care Survey, 2004.			Aspirin	5.7 (3.3–8.2)
				Clopidogrel	4.7 (1.5–7.9)
				Digoxin	3.2 (1.6–4.7)
				Metformin	2.3 (1.4–3.2)

ADE, adverse drug event; CI, confidence interval; ED, emergency department.

The Beers criterion is a commonly used measure of medication appropriateness for older patients.²⁷ A study was conducted to identify inappropriate use of medications in older adults and to estimate the risk for ED visits for AEs using Beers criteria.²⁶ More than half of the estimated ED visits were related to use of anticholinergics or antihistamines, nitrofurantoin, or propoxyphene, which are among the 41 medications or medication classes always considered to be potentially inappropriate. Nine of the 10 most commonly implicated medications for adverse events belonged to 3 classes: oral anticoagulant or antiplatelet agents, antidiabetic agents, and narrow therapeutic index agents. Together, these 3 classes of medications accounted for an estimated 47.5% (95% confidence interval [CI], 40.2%–54.8%) of all ED visits for ADEs among older adults.²⁶

Number of medications

The patient safety literature has identified that polypharmacy is a significant risk factor for ADE-related visits (Table 5).^9,11,24,28 Overall, the higher the number of drugs prescribed, the greater the risk of an ADE. Studies examining drug categories and patient characteristics among outpatient ADEs that lead to ED visits or hospitalizations found that the majority of outpatient harm occurred in patients older than 65 years of age who take 3 or more medications; specific drugs such as digoxin, warfarin, and insulin were implicated in the majority of instances of outpatient drug-related harm.^9,12,23 One study found that patients were prescribed a large number of medications and the more drugs prescribed, the greater the risk of an ADE; however, the risk increase was not linear, as there was a dramatic increase once patients were prescribed more than 11 medications.²⁴

Table 5.

Number of Medications

Source	Setting/Study Population	Study Design	Number of Patients	Medication Factors	Adverse Drug Event Related to Admission Rates (95% CI)
Bourgeois et al, 2010¹¹	National Center for Health Statistics Data in the United States; visits between 1995 and 2005	Retrospective	ADE-related ED visitswere 3,544,908	5 or more drugsCardiotonic glycosides (ie, digoxin)AnticoagulantsAnticonvulsantsAntineoplastics	OR: 1.88(1.56–2.24)13%(9%–19%)11%(8%–11%)7%(4%–11%)7%(4%–11%)
Hafner et al, 2002⁹	University-affiliated tertiary care	Retrospective: ED charts: matched control for: age for disposition, survival, severity, payer, sex, race, age, number of drugs, and total cost	434		Hospitalization:
	ED occurring between March 1 and May 31, 1997,			4 or more drugs	OR: 2.2(1.7–2.8)
Sarkar et al, 2011²⁸	National Ambulatory Medical Care Survey and the National Hospital and Ambulatory Medical Care Survey. From 2005 to 2007	Retrospective; secondary data analysis	13.5 million ADE-related visits		Ambulatory visits
			(4.5 million ADE visits per year)	6-8 meds	OR: 3.83
					(2.20–6.65)
Forster et al, 2005²⁷	Urban academic health sciences center	Secondary analysis of a prospective cohort study	400	12 or more medications at discharge	Risk of ADE:
					18%
					(11%–27%)

ADE, adverse drug event; CI, confidence interval; ED, emergency department; OR, odds ratio.

Utilization

Readmissions also are characterized by health care utilization factors such as visits to the ED, hospitalization, and ambulatory primary and specialty care visits (Table 6). ^8,9,15,28 Patients are readmitted for a variety of reasons that include therapeutic management, diagnostic evaluation, rehabilitation, prevention, palliation, research, or cosmetic interventions. After considering the type of disease and the severity level of the patient's condition, researchers found readmission rates varied substantially by hospital and by geographic area.¹ Although some studies predicted the risk of readmission based on previous hospital admissions or ED utilization, a study by Dormann and colleagues²⁹ found that ADEs predicted further readmissions but a lack of ADEs did not preclude readmissions. In this study, ADEs caused hospitalizations in 6.2% of first admissions, and only 4.2% of hospitalizations were readmissions.²⁹

Table 6.

Utilization

Source	Setting/Study Population	Study Design	Number of Patients	Utilization Factors	Readmission/^Adverse Drug Event Related to Admission Rates*
Smith et al, 2000¹⁵	Nine VA medical centers- Medical Service	Retrospective: secondary analysis from RCT	1378	Last 6 months:
				ED visits	RR:1.13
					(1.04–1.24)
				Hospitalization	RR:1.85
					(1.43–2.41)
Hafner et al, 2002⁹	University-affiliated tertiary care ED visits between March 1 and May 31, 1997,	Retrospective: ED charts: matched control for: age for disposition, survival, severity, payer, sex, race, age, number of drugs, and total cost	434	Hospitalization	^*31%
Zhan et al, 2005⁸	Nationally representative sample of visits	Retrospective:	311,000 to 535,000 visits/year		^*Per 1,000 population per year
	to physician offices, hospital outpatient departments, and emergency departments	NAMCS and NHAMCS public use data from 1995 to 2001		Ambulatory care	11.5
				Hospitalization	2.8
Sarkar et al, 2011²⁸	National Ambulatory Medical Care Survey and the National Hospital and Ambulatory Medical Care Survey. From 2005 to 2007	Retrospective; secondary data analysis	13.5 million ADE-related visits (4.5 million ADE visits per year)	↑ ADEs with primary care visits compared to specialty care visits	^*OR: 2.22
					(1.70–2.89)
				Hospitalization	9%
				F/u visit for ADE	22%

ADE, adverse drug event; ED, emergency department; F/u, follow-up; NAMCS, National Ambulatory Medical Care Survey; NHAMCS, National Hospital and Ambulatory Medical Care Survey; OR, odds ratio; RCT, randomized controlled trial; RR, risk ratio.

Patient- and System-Level Factors in Outpatient Care

At the individual level, risk factors or antecedents for readmission include inadequate self-care behavior skills, disease knowledge, health literacy, and communication, patient adherence, patient choice (intentional and unintentional), frailty and assistance with activities of daily living, and follow-up of health care appointments. Similarly, system-level factors such as inadequate communication due to lack of accuracy (eg, medication instructions specifying “twice a day” versus “after breakfast and after dinner”), interpretation, completeness, and timeliness of information place the patient at risk for readmission. Inadequate communication can occur between the patient and the health care worker or among health care providers. Similarly, the availability and accessibility of resources at a health care setting also influence patient readmissions.

Based on the Congressional Research Service readmission report, there is evidence that avoidable medical errors occur while patients are in the hospital setting, and that these errors can cause AEs following discharge that result in readmissions for some Medicare beneficiaries.⁵ According to the Agency for Healthcare Research and Quality Patient Safety Network, systematic problems in care transitions are at the root of most AEs that arise after discharge.³⁰ For example, discontinuity between inpatient and outpatient providers is common, and studies have shown that traditional communication systems (eg, the dictated discharge summary) generally fail to reach outpatient providers in a timely fashion and often lack essential information. Some studies have implicated assessment and communication of unresolved problems at the time of discharge, inadequate patient education regarding medications and other therapies, and failure to monitor drug therapies and overall condition after discharge as contributing to readmissions.¹⁸ Moreover, patients frequently receive new medications or have medications changed during hospitalizations and lack of medication reconciliation results in the potential for inadvertent medication discrepancies and ADEs, particularly for patients with low health literacy, or those prescribed high-risk or complex medication regimens.³⁰

Failure to monitor and implement appropriate drug monitoring was an especially common cause of preventable and ameliorable ADEs.²⁴ A study to identify drug-related problems during and after hospitalization found that preventable ADEs resulted from lack of medication access, nonadherence, and inadequate drug monitoring after discharge.³¹

A study evaluating the economic impact of adverse drug reactions (ADRs) on recurrent hospitalizations in an internal medicine setting (median patient age of 57 years) found that 44.3% of ADRs were preventable.²⁹ When considering admissions and readmissions, 11% (>973 days) of all treatment days were judged to be preventable.²⁹ Additionally, the occurrence and numbers of ADRs per admission were found to prolong the hospitalization period significantly (P<0.001). Twenty percent of 9107 treatment days were caused by in-house (1130 days) and community-acquired ADRs (669 days).²⁹

Medication discrepancy

Medication discrepancy is defined as a patient-associated or system-associated discrepancy (Table 7).³² Medication discrepancies between, before, and after discharge often are associated with ADEs. In a study among community-dwelling adults aged 65 years and older, hospital readmission rates among patients with identified medication discrepancies was significantly higher (14.3%) than that among patients with no identified medication discrepancies (6.1%; P=0.04).³² Overall, 50.8% of identified contributing factors for discrepancies were categorized as patient-associated factors, and 49.2% were categorized as system-associated factors. Five medication classes accounted for 50% of all identified medication discrepancies: anticoagulants (13%), diuretics (10%), angiotensin-converting enzyme inhibitors (10%), lipid-lowering agents (10%), and proton pump inhibitors (7%).³² The variables significantly associated with patients who experienced medication discrepancies were the number of medications taken (odds ratio [OR], 1.13; 95% CI, 1.04–1.23) and the presence of congestive heart failure (OR, 2.10; 95% CI, 1.09–4.03). However, the number of medications taken was not associated with readmission rates (P=0.71).³²

Table 7.

Patient and System-Associated Medication Discrepancies

Patient-Associated Discrepancies	System-Associated Discrepancies
• Adverse drug effects	• Prescribed medication with known allergies/intolerances
• Drug intolerance	• Conflicting information from different informational sources
• Failure to fill prescription	• Confusion between brand and generic names
• Unnecessary prescription	• Incomplete discharge instructions
• Money/financial barriers	• Inaccurate, duplicate, or illegible transcription
• Intentional nonadherence	• Incorrect dosage, quantity, or label
• Unintentional nonadherence	• Unrecognized cognitive impairment
• Deficit in managing medications	• No caregiver or unrecognized need for assistance
	• Unrecognized sight or dexterity limitations

Prescriptions

The risk of an ADE increased as the number of medications prescribed increased.²⁴ The top 5 most commonly prescribed classes of medications at the time of discharge were cardiovascular agents (1.2 prescriptions per patient), nutrient agents (including electrolyte and vitamin supplements [1.1 prescriptions per patient]), gastrointestinal agents (0.9 prescriptions per patient), respiratory agents (0.7 prescriptions per patient), and anti-infective agents (0.7 prescriptions per patient). ADEs per prescription was highest for corticosteroids, anticoagulants, antibiotics, analgesics, and cardiovascular medications.²⁴

Outpatient Trends and Environmental Factors

A number of trends place the elderly at high risk for readmission. With an aging US population, there is an increased movement toward outpatient care and medication usage. Additionally, the numbers of patients with chronic diseases is increasing, the development and availability of potent prescription medications is rising, the transition of prescription medications to over-the-counter (OTC) availability is growing, discounted generics and mail service are readily available, and policy changes affecting Medicare drug coverage benefits have increased uncertainty in the availability and affordability of medications.²⁶ Subsequently, there is growing concern that these trends will impact safe medication use and put the elderly at risk for hospital admission.

Outpatient prescription use

Prescription medicines are the most frequently used therapeutic intervention in the outpatient setting. Nearly 2.5 billion prescriptions were dispensed by US pharmacies in 1998,³³ and the percent increase in primary care office visits that involved the initiation or continuation of medication therapy increased from 61.0% to 72.6% from 1995 to 2005.³⁴ Although the age breakdown is not evident in these reports, in 2011 the Intercontinental Marketing Services (IMS) Institute for Healthcare Informatics reported that because of the convenience and availability of discounted generics, chain drugstores followed by mail service were increasingly chosen by patients to fill their prescriptions (54% of all prescriptions or 2.2 billion) in 2010. Hence, consumption of medicines may have led to fewer doctor office visits, which were down 4.2% in 2010. The top 5 most expensive drug classes in 2010 were: oncologics ($22 B), respiratory agents ($19.3 B), lipid regulators ($18.7 B), antidiabetic agents ($16.9 B), and antipsychotics ($16.1 B). Absolute spending growth gains were highest for antidiabetic agents, antipsychotics, respiratory agents, HIV antivirals, and autoimmune disease agents.³⁵

Increase in prescription use

Between the periods 1999–2000 and 2007–2008, the percentage of Americans using at least 1, 2 or more, and 5 or more prescription medications in the past month increased from 44% to 48%, from 25% to 31%, and from 6% to 11%, respectively. Prescription drug use increased with age between 1988 and 1994.³⁶ More than 88% of Americans 60 years of age and older used at least 1 prescription. Further, more than 76% used at least 2 or more prescription drugs, and 36% consumed 5 or more prescription drugs. Additionally, women were more likely to use prescription drugs than men, and the non-Hispanic white population had the highest prescription drug use in contrast to the Mexican American population, who had the lowest.³⁶ Those who were without a regular place for health care, without health insurance, or without a prescription drug benefit had less prescription drug use compared with those who had these benefits. The most commonly used types of prescription drug were cholesterol-lowering drugs, diuretics, and beta-blockers, which usually are used to treat cholesterol, high blood pressure, and heart problems, respectively.^36,37

The use of OTC medications has increased because of the transition of potent prescription medications to OTC status and the resulting increased availability. Although the overall prevalence of medication use has not changed from 1998, polypharmacy has increased since 2000 from 6.3% to 12% for the use of at least 5 prescription medications, and from 23% to 29% for the use of 5 or more prescription medications.³⁸ Approximately 82% of US adults take at least 1 medication (ie, prescription or nonprescription drug, vitamin/mineral, herbal/natural supplement) and 29% take 5 or more medications. Americans aged 65 and older are the biggest consumers of medications; 17% to 19% of patients in this age group take at least 10 or more medications. The most commonly used drug among prescription and nonprescription drugs is acetaminophen, followed by 2 cholesterol-lowering drugs (atorvastatin and simvastatin).³⁸

Out of the 41% of US adults who use a vitamin product in a given week, 63% of consumers are older women (≥65 years). Twenty-two percent of US adults use herbals/natural supplements, and 32% of prescription drug users also take an herbal/natural supplement.³⁸

In addition to trends such as increased outpatient medication use among the elderly, the elderly are at increased odds for poor self-management and at risk for hospital readmission resulting from age-related physiological changes and co-existing illnesses, a higher prevalence of cognitive and functional impairment with increasing age, increasing numbers of chronic diseases, and geographic and functional isolation (ie, older adults living in the community compared to nursing home residents).

Other factors that lead to readmissions in the elderly include health care- or patient-related factors. Health care-related factors include: inadequate information and communication by hospital discharge planners to patients, caregivers, and/or post-acute care providers,^39,40 inadequate follow-up care from post-acute and long-term care providers,⁴¹ variation in hospital bed supply,⁴⁷ and medical errors in a hospital that may occur during an initial admission and result in illness, injury, or harm to a patient.⁴² Patient-related factors include: poor patient adherence, insufficient use of the supportive capacity of family caregivers,^43,44 and deterioration of a patient's clinical condition.¹

Interventions to Decrease Readmissions

Although there is no consensus on how to decrease readmissions, there is some evidence that comprehensive, multimodal interventions may be more effective at preventing readmissions than targeting individual components of the discharge process. Disease management programs that target specific disease conditions have shown potential to decrease morbidity and mortality.⁴⁵ However, more research on the utility of these programs is needed to compare the effect of different aspects of disease management programs on different populations. Moreover, disease management programs target specific diseases such as diabetes or heart failure, even though patients face concurrent comorbidities; this adds to the complexity of achieving success with such programs.

Targeted interventions to decrease readmissions have included medication reconciliation,³² pharmacists counseling at discharge with a follow-up telephone call,³¹ and documentation of adherence with any or all of the 6 required discharge instructions.⁴⁶ A meta-analysis reviewed 12 randomized controlled intervention studies published from 1980 through 1990 to determine the efficacy of planned interventions to reduce readmissions. These interventions included home health teaching, in-hospital teaching, geriatric consultation, geriatric special care unit, home health visits, pharmacological counseling, structural discharge interview, home visitation, and comprehensive discharge planning for elderly. In 8 of the 12 studies the readmission rates were lower than for the control group.²¹

Similarly, a review of the literature on interventional studies to reduce readmissions also found a 12%–75% reduction in readmissions in emergency visits. Interventions in these studies were varied (Table 8).² Although multimodal interventions are effective, their utility in terms of cost and resources is unknown. There is a lack of research related to identifying and targeting which interventions are most effective for those patients at highest risk for readmissions.

Table 8.

Interventions to Reduce Readmissions

• Domiciliary aftercare for 2 weeks after discharge

• Discharge planning

• Home care assistance

• Team assessment with recommendations to attending physicians

• Interventions to counteract passivity and enhance active patient involvement

• Case managers

• Educational material mailed to discharged patients followed by telephone call to resolve unmet needs

• Review of warning signs and barriers to keeping appointments

• Assessment of the appropriateness of care

• Follow-up home visits by a geriatric team (physician, nurse, physiotherapist)

• Postdischarge follow-up by a nurse and a primary care physician,

• Comprehensive geriatric assessment before and after discharge

Methods and Measurement of Adverse Drug Event-Related Readmission and Prevention

Although numerous conditions contribute to readmissions, heart failure remains the most common primary discharge diagnosis among Medicare patients.⁴⁰ However, even with heart failure, readmission rates and associated factors have varied across institutions because of differences in the spectrum of patients, disease severity, and comorbid conditions. Additionally, documentation of diagnoses such as a major hip or knee surgery may not take into account patient falls as the underlying cause of hospitalization. Similarly, the patient safety literature underestimates ADEs as the cause of readmission, and differences in institutional reporting make aggregation and analyses of disparate data challenging when estimating the proportion of readmissions resulting from medication errors. To date, a variety of data sources and study designs have been used to examine the incidence, causes, and factors related to hospital readmissions in both the readmission and medication safety literature.

Data sources

A majority of researchers have used administrative data from large national databases. These data have been analyzed to describe the factors and patterns associated with readmission, to characterize population estimates on annual readmission incidence, to identify the risk factors of outpatient ADEs, and to examine the factors that underlie better US hospital performance on current discharge metrics. These large data sets included claims data,^1,22 data from the National Ambulatory Medical Care Survey, the National Hospital and Ambulatory Medical Care Survey,^11,28 the Hospital Quality Alliance data set,⁴⁷ and the National Electronic Injury Surveillance System–All Injury Program.²⁵

Other studies have used either single or multiple data sources such as administrative databases, electronic medical records, medical charts, patient reports, telephone interviews, and administrative incident reports from the group's affiliated pharmacies to examine contributing factors associated with readmission or ADEs.^3,9,18,32,48

In studies that examined home-based interventions, geriatric care managers gathered data when conducting home-based assessments to determine the prevalence of the use of OTC drugs, dietary supplements, Part D-excluded medications, and potentially inappropriate medications by homebound older adults.⁴⁹ An intervention study also used patient self-reports, case summaries, medication lists at admission and discharge, the hospital discharge summary, any available outpatient visit notes, discharge summaries from ED visits or hospital readmissions, and laboratory test results to identify drug-related problems during and after hospitalization.³¹

Use of a large national survey has the strength to allow for reliable national estimates; however, each visit has limited data. For example, a determination of whether the primary or underlying reason for readmission is symptom management, progression of disease, inadequate self-management, or attribution of an ADE to a specific medication or treatment is not contained in these data sets. Data obtained from databases also rely on documentation and voluntary reporting of ADEs by the treating physician, which is probably less sensitive than research studies involving chart review by specially trained pharmacists or physicians, computer-generated signals, patient interviews, or combination approaches to identify undiagnosed and undocumented ADEs. Moreover, when using estimates to calculate population rates for ADE visits, multiple ADE visits by the same individual are not accounted for, and only ADEs that led to health care utilization are captured, suggesting that ambulatory ADEs may be underestimated.

Readmission studies also have relied on medical records and have largely focused on sociodemographic and clinical characteristics. However, contributors to hospital readmission that occur in the patient's environment, such as inadequate monitoring of gradual onset of signs and symptoms, improper or unsafe medication use that may not lead to an ADE, asymptomatic errors, or errors with the potential to cause harm, remain undetected or unreported. Consequently, patients face many challenges after discharge (Table 9)⁵⁰ and data on such barriers are not easily collected or available.

Table 9.

Patient Challenges After Discharge

• Formidable costs

• Lack of insurance coverage affecting medication procurement and administration

• Brief patient visits with inadequate education and inadequate communication resulting from fragmentation of care across multiple providers, facilities, and pharmacies

• Inadequate patient knowledge

• Ineffective health care behaviors

• Interaction problems encountered by patients or providers not readily detected or reported

Understanding the nature and effect of such factors from the patient's perspective would help prioritize intervention targets for improving outpatient care after discharge to reduce readmissions. Thus, the magnitude of the problem regarding self-management and the severity of the consequences that lead to readmission in the patient's environment are unknown. With no structured mechanism to collect and monitor outpatient characteristics of self-management, it seems logical that most errors or problems that can be mitigated go unreported unless they are serious enough to bring the patient to a clinical setting.

Study design and methods

With readmission and ADE studies, most approaches included prospective cohort, retrospective cohort, and secondary analysis study designs. Few studies had preplanned intervention, or randomized trial study designs. Case-control studies, which assess the outcomes of interventions purely in terms of reduction in admissions among a cohort of older people without any reference to a control group, often assume that patients identified to be at high risk on the basis of their previous admissions would continue to be at high risk of admission in the absence of the intervention. An alternative approach would be to use 2 separate control groups, one for direct age comparison and the other for all additional comparisons. With case summaries, adjudication for the contributors of readmission may rely heavily on medical record or a clinical expert. This may be challenging if adjudicators are aware that the experts had the advantage of follow-up and consequently place unequal weights on the data source from the experts over the medical records. Therefore, improved methods of systematic screening for contributors of readmission should be evaluated and prospective methods for predictors of readmission should be developed.

Conclusion

Medicare patients have higher readmission rates after discharge from the hospital than other age groups. Some of the characteristics common among those readmitted are female sex, more comorbidities, chronic conditions, more ADEs, and utilization of ambulatory clinic care and ED services. Among the elderly, drug-related events were more numerous among women and were associated with polypharmacy, increased comorbidity, chronic conditions involving the circulatory system, and conditions such as cardiac heart failure, diabetes, and atrial fibrillation. ADEs also were associated with higher numbers of ED and primary care visits. Admission and readmission rates were at 37%, of which readmission rates were 4.2%. Medication discrepancies before and after discharge were attributed mostly to anticoagulants and diuretics. Similarly, cardiovascular drugs followed by diuretics and non-opioid analgesics were the medication class of drugs responsible for most ADEs.

Patients with heart failure have the highest readmission rate. Other conditions such as psychoses, chronic obstructive pulmonary disease, diabetes, and atrial fibrillation also are common. Other contributors to readmission and ADEs include polypharmacy and, in particular, cardiovascular medications followed by diuretics, non-opioid analgesics, hypoglycemics, and anticoagulants.

With an aging US population, and trends such as increasing numbers of patients with chronic diseases, the development of potent prescription medications, the transition to and availability of prescription medications as OTC, and changes to Medicare drug coverage benefits, the rise in readmission rates is a growing concern.

A majority of factors that contribute to readmissions occur after discharge and few studies have examined patient-centered factors as patients receive care from multiple providers, decipher complex pharmaceutical names with look-alike and sound-alike medications, manage their medication regimen and dosage changes without assurance of fully understanding the change, and have intermittent or infrequent communications with their providers about their problems. Additionally, patients value other intangibles (Table 10), which influence their motivation to comply with therapeutic medication regimens and self-management recommendations.

Table 10.

Intangibles Valued by Patients

• Choice of treatment options

• Trust in their provider and plan for medical care

• Lifestyle preference

• Ease of medication regimen

• Ease of access to providers to procure medications and to obtain answers to their questions and concerns

• Freedom from troublesome medication side effects that are difficult to characterize but cause related burden and distress that affect their overall well-being and quality of life, even if health care providers value only the health care outcomes in terms of health-related quality of life and extended survival

Strategies to improve monitoring must take into account the difficulties the elderly face. Such difficulties include their social setting, frailty and increased comorbidity, risk of developing ADEs from physiological changes affecting the pharmacokinetics and pharmacodynamics of many drugs, requiring assistance with activities of daily living, difficulty attending follow-up clinics that require enhanced communication with community care providers, and poor coordination of home-care services, hospital care, hospital-based follow-up clinics, and early telephone contact.

Readmission studies have used predominantly administrative and survey data and medical charts that limit the predictors of readmission to demographic and clinical variables, ADEs, reported and documented symptoms, and therapeutic lab values. Factors such as health literacy, depression, heart failure knowledge, self-care behaviors, cognition, social support, communication between providers, and adequacy of follow-up care have not been explored sufficiently and may contribute to readmissions.

Structured mechanisms are needed to monitor, collect, and measure outpatient characteristics of self-management after discharge, as are study designs that consider the etiology of why and how patients respond before, during, and after readmission. Similarly, a model that examines the antecedents of unintentional or unavoidable contributors to readmission and the development of strategies to decrease frequency and severity has the potential to guide the methodology. Hence, multiple strategies are required to identify the predictors of readmission.

Potentially preventable readmissions may require system changes that focus on 4 areas: evaluating patients at the time of discharge; teaching patients about drug therapies, side effects, and what to do if specific problems develop; improving monitoring of therapies; and improving monitoring of patients' overall conditions. Similarly, the development of measures to monitor these interventions also is needed.

Currently no study has developed models to predict a patient's risk for readmission or to compare readmission rates among different settings. Most studies looking at readmissions have associated patient demographics, clinical characteristics, and healthcare utilization with readmissions. Additionally, studies vary in methods used for case and outcome identification, use disparate and multiple data sources, and examine readmissions with varying follow-up periods (ie, 30 days, 60 days, 90 days). Although these characteristics may be important predictors of readmission, few variables were identified consistently. Examining intervening or mediating contributors to hospital admission from the patient safety and readmission literature will provide a model to guide interventions and future research. We have proposed a conceptual framework that supports a multivariate model to develop a more comprehensive proposal, with the objective of developing a risk profile or score to enhance the probability of early identification of those at greatest risk for readmission and those who might benefit from minimal versus more extensive interventions

Future research that utilizes rigorously designed methods to identify high-risk patients and targets interventions to reduce the burden of readmissions is likely to achieve the greatest impact. The costs of potentially preventable hospital admissions are considerable. Therefore, patient interventions to prevent hospital admissions may be cost-effective or even cost saving.

Footnotes

Author Disclosure Statement

Drs. Robinson, Howie-Esquivel, and Vjahov disclosed no conflicts of interest.

References

Jencks

, Williams

, Coleman

. Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med, 2009; 360:1418–1428.

Benbassat

, Taragin

. Hospital readmissions as a measure of quality of health care: Advantages and limitations. Arch Intern Med, 2000; 160:1074–1081.

Allaudeen

, Vidyarthi

, Maselli

, Auerbach

. Redefining readmission risk factors for general medicine patients. J Hosp Med, 2011; 6:54–60.

Wier

, Barrett

, Steiner

, Jiang

. All-Cause Readmissions by Payer and Age, 2008. HCUP Statistical Brief #115. Rockville, MD: Agency for Healthcare Research and Quality, 2011.

Stone

, Hoffman

. Report to Congress: Medicare Hospital Readmissions: Issues, Policy Options and PPACA. Washington, DC: Congressional Research Service, 2010; 39.

Budnitz

, Pollock

, Weidenbach

, Mendelsohn

, Schroeder

, Annest

. National surveillance of emergency department visits for outpatient adverse drug events. JAMA, 2006; 296:1858–1866.

Classen

, Pestotnik

, Evans

, Lloyd

, Burke

. Adverse drug events in hospitalized patients. Excess length of stay, extra costs, and attributable mortality. JAMA, 1997; 277:301–306.

Zhan

, Arispe

, Kelley

et al. Ambulatory care visits for treating adverse drug effects in the United States, 1995-2001. Jt Comm J Qual Patient Saf, 2005; 31:372–378.

Hafner

Jr , Belknap

, Squillante

, Bucheit

. Adverse drug events in emergency department patients. Ann Emerg Med, 2002; 39:258–267.

10.

Lazarou

, Pomeranz

, Corey

. Incidence of adverse drug reactions in hospitalized patients: A meta-analysis of prospective studies. JAMA, 1998; 279:1200–1205.

11.

Bourgeois

, Shannon

, Valim

, Mandl

. Adverse drug events in the outpatient setting: An 11-year national analysis. Pharmacoepidemiol Drug Saf, 2010; 19:901–910.

12.

Gurwitz

, Field

, Harrold

et al. Incidence and preventability of adverse drug events among older persons in the ambulatory setting. JAMA, 2003; 289:1107–1116.

13.

Bedell

, Jabbour

, Goldberg

et al. Discrepancies in the use of medications: Their extent and predictors in an outpatient practice. Arch Intern Med, 2000; 160:2129–2134.

14.

Joynt

, Orav

, Jha

. Thirty-day readmission rates for Medicare beneficiaries by race and site of care. JAMA, 2011; 305:675–681.

15.

Smith

, Giobbie-Hurder

, Weinberger

et al. Predicting non-elective hospital readmissions: A multi-site study. Department of Veterans Affairs Cooperative Study Group on Primary Care and Readmissions. J Clin Epidemiol, 2000; 53:1113–1118.

16.

Mudge

, Kasper

, Clair

et al. Recurrent readmissions in medical patients: A prospective study. J Hosp Med, 2011; 6:61–67.

17.

Bero

, Lipton

, Bird

. Characterization of geriatric drug-related hospital readmissions. Med Care, 1991; 29:989–1003.

18.

Forster

, Murff

, Peterson

, Gandhi

, Bates

. The incidence and severity of adverse events affecting patients after discharge from the hospital. Ann Intern Med, 2003; 138:161–167.

19.

Forster

, Clark

, Menard

et al. Adverse events among medical patients after discharge from hospital. CMAJ, 2004; 170:345–349.

20.

Marcantonio

, McKean

, Goldfinger

, Kleefield

, Yurkofsky

, Brennan

. Factors associated with unplanned hospital readmission among patients 65 years of age and older in a Medicare managed care plan. Am J Med, 1999; 107:13–17.

21.

Soeken

, Prescott

, Herron

, Creasia

. Predictors of hospital readmission. A meta-analysis. Eval Health Prof, 1991; 14:262–281.

22.

Kane-Gill

, Van Den Bos

, Handler

. Adverse drug reactions in hospital and ambulatory care settings identified using a large administrative database. Ann Pharmacother, 2010; 44:983–993.

23.

Thomsen

, Winterstein

, Sondergaard

, Haugbolle

, Melander

. Systematic review of the incidence and characteristics of preventable adverse drug events in ambulatory care. Ann Pharmacother, 2007; 41:1411–1426.

24.

Forster

, Murff

, Peterson

, Gandhi

, Bates

. Adverse drug events occurring following hospital discharge. J Gen Intern Med, 2005; 20:317–323.

25.

Budnitz

, Pollock

, Mendelsohn

, Weidenbach

, McDonald

, Annest

. Emergency department visits for outpatient adverse drug events: Demonstration for a national surveillance system. Ann Emerg Med, 2005; 45:197–206.

26.

Budnitz

, Shehab

, Kegler

, Richards

. Medication use leading to emergency department visits for adverse drug events in older adults. Ann Intern Med, 2007; 147:755–765.

27.

Fick

, Cooper

, Wade

, Waller

, Maclean

, Beers

. Updating the Beers criteria for potentially inappropriate medication use in older adults: Results of a US consensus panel of experts. Arch Intern Med, 2003; 163:2716–2724.

28.

Sarkar

, Lopez

, Maselli

, Gonzales

. Adverse drug events in U.S. adult ambulatory medical care. Health Serv Res, 2011; 46:1517–1533.

29.

Dormann

, Neubert

, Criegee-Rieck

et al. Readmissions and adverse drug reactions in internal medicine: The economic impact. J Intern Med, 2004; 255:653–663.

30.

United States Department of Health & Human Services. Patient Safety Primer: Adverse Events after Hospital Discharge. http://www.psnet.ahrq.gov/printviewPrimer.aspx?primerID=11. 2011 July 24.

31.

Schnipper

, Kirwin

, Cotugno

et al. Role of pharmacist counseling in preventing adverse drug events after hospitalization. Arch Intern Med, 2006; 166:565–571.

32.

Coleman

, Smith

, Raha

, Min

. Posthospital medication discrepancies: Prevalence and contributing factors. Arch Intern Med, 2005; 165:1842–1847.

33.

Institute of Medicine. To Err Is Human: Building a Safer Health System. Washington, DC: National Academy Press, 2000.

34.

Middleton

, Hing

, Jianmin

. National Hospital Ambulatory Medical Care Survey: 2005 Outpatient Department Summary. Hyattsville, MD: National Center for Health Statistics, 2007; 36.

35.

IMS Institute for Healthcare Informatics. The Use of Medicines in the United States: Review of 2010. Parsippany, NJ: IMS Health and the IMS Institute for Healthcare Informatics, 2011.

36.

National Health Nutrition Examination Survey. Patterns of prescription drug use in the United States, 1988–94. www.cdc.gov/nchs/data/nhanes/databriefs/preuse.pdfSimilar. 2011 June 27.

37.

, Dillon

, Burt

. Prescription Drug Use Continues to Increase: U.S. Prescription Drug Data for 2007-2008. Washingon, DC: National Center for Health Statistics, 2010.

38.

Slone Epidemiology Center. Patterns of medication use in the United States. http://www.bu.edu/slone/SloneSurvey/AnnualRpt/SloneSurveyReport2004.pdf. 2011 June 27.

39.

Kripalani

, LeFevre

, Phillips

, Williams

, Basaviah

, Baker

. Deficits in communication and information transfer between hospital-based and primary care physicians: Implications for patient safety and continuity of care. JAMA, 2007; 297:831–841.

40.

Medicare Payment Advisory Commission. Report to Congress: Promoting Greater Efficiency in Medicare. Washington, DC: Medicare Payment Advisory Commission, 2007.

41.

Sin

, Tu

. Are elderly patients with obstructive airway disease being prematurely discharged? Am J Respir Crit Care Med, 2000; 161:1513–1517.

42.

Institute of Medicine. Identifying and Preventing Medication Errors. Washington, DC: National Academies Press, 2007.

43.

Arbaje

, Wolff

, Yu

, Powe

, Anderson

, Boult

. Postdischarge environmental and socioeconomic factors and the likelihood of early hospital readmission among community-dwelling Medicare beneficiaries. Gerontologist, 2008; 48:495–504.

44.

Coleman

, Parry

, Chalmers

, Min

. The care transitions intervention: Results of a randomized controlled trial. Arch Intern Med, 2006; 166:1822–1828.

45.

Gohler

, Januzzi

, Worrell

et al. A systematic meta-analysis of the efficacy and heterogeneity of disease management programs in congestive heart failure. J Card Fail, 2006; 12:554–567.

46.

VanSuch

, Naessens

, Stroebel

, Huddleston

, Williams

. Effect of discharge instructions on readmission of hospitalised patients with heart failure: Do all of the Joint Commission on Accreditation of Healthcare Organizations heart failure core measures reflect better care? Qual Saf Health Care, 2006; 15:414–417.

47.

Jha

, Orav

, Epstein

. Public reporting of discharge planning and rates of readmissions. N Engl J Med, 2009; 361:2637–2645.

48.

Field

, Gurwitz

, Harrold

et al. Risk factors for adverse drug events among older adults in the ambulatory setting. J Am Geriatr Soc, 2004; 52:1349–1354.

49.

Golden

, Qiu

, Roos

. Medication assessments by care managers reveal potential safety issues in homebound older adults. Ann Pharmacother, 2011; 45:492–498.

50.

Gandhi

, Weingart

, Borus

et al. Adverse drug events in ambulatory care. N Engl J Med, 2003; 348:1556–1564.