Hospital readmission after hip fracture

Transcription

1 DOI /s TRAUMA SURGERY Hospital readmission after hip fracture Stephen L. Kates Caleb Behrend Daniel A. Mendelson Peter Cram Susan M. Friedman Received: 20 July 2014 Ó Springer-Verlag Berlin Heidelberg 2014 Abstract Introduction Readmission to the hospital following a hip fracture is common, often involves an adverse event, and strains an already overburdened health care system. Objectives To assess the rate of 30-day readmission to the hospital after discharge for care of hip fracture. A secondary objective was measurement of the 30-day mortality rate for those patients readmitted versus those patients not readmitted to the hospital after discharge. Materials and methods Study design was a retrospective review of registry data comparing readmitted patients to those not readmitted after hip fracture. Setting was a university affiliated level 3 trauma center. Participants: 1,081 patients aged 65 and older. Measurements: rate of readmission, rate of mortality, predictors of readmission. Results 129 patients (11.9 %) were readmitted to the hospital within 30 days of their initial discharge date. The primary causes of readmission were surgical in nature for S. L. Kates (&) Department of Orthopaedics and Rehabilitation, University of Rochester, 601 Elmwood Ave, Box 665, Rochester, NY 14642, USA Stephen_kates@urmc.rochester.edu C. Behrend Carilion Clinic Orthopaedics, 3 Riverside Circle, Roanoke, VA 24016, USA D. A. Mendelson S. M. Friedman (&) Division of Geriatrics, Department of Medicine, University of Rochester, 1000 South Ave., Rochester, NY 14620, USA Susan_friedman@urmc.rochester.edu P. Cram Division of General Internal Medicine, University of Toronto, 200 Elizabeth Street, Eaton North, 14th Floor, Rm 216, Toronto, ON M5G 2C4, Canada 24/129 (18.6 %) patients and 105/129 (81.4 %) were readmitted for medical or other reasons. Twenty-four (18.6 %) patients who were readmitted died during readmission. The one-year mortality rate for patients readmitted within 30 days was 56.2 vs. a 21.8 % 1-year mortality rate for those patients not readmitted (p \ ). Independent predictors of readmission were age [85 (OR = 1.52; p = 0.03), time to surgery [24 h (OR = 1.50; p = 0.05), Charlson score C4 (OR = 1.70; p = 0.04), delirium (OR = 1.65; p = 0.01), dementia (OR = 1.61; p = 0.01), history of arrhythmia with pacemaker placement (OR = 1.75; p = 0.02), and presence of a pre-op arrhythmia (OR = 1.62; p = 0.02). Conclusion Readmission after hip fracture is harmful and undesirable 18.6 % of readmitted patients died during their readmission and the average length of stay was 8.7 days. Approximately one of every six readmissions was identified as potentially preventable with interventions. Keywords Readmission Hip fracture Geriatric fracture Mortality Complication Introduction Hospital readmission following hip fracture is a frequent and serious sentinel event that may be avoidable or indicate a gap in care [1]. Although nearly all patients with hip fracture are admitted to acute care, there is considerable regional variation in readmission rates according to recent published data [1]. Hip fracture is the leading orthopaedic discharge diagnosis associated with 30-day readmission [1]. Readmission to the hospital following a hip fracture is common, often involves an adverse event, and strains an already overburdened health care system [2 5]. US

2 national readmission rates following hip fracture have remained essentially unchanged from 2004 (14.3 %) to 2009 (14.5 %) [1]. Most hip fractures occur in patients C65 years of age. The population of older adults is rapidly increasing as the United States (US) population ages. The number of hip fractures worldwide is predicted to more than double by 2050, increasing from 39 million to 89 million [6 8]. Other studies have predicted an increase of 51 % by 2025 [9]. With an estimated 330,000 hip fractures occurring yearly in the US, the growth in hip fractures poses serious economic and logistical challenges [10]. The hip fracture population is frail, with multiple comorbidities, and, as such, is at high risk for hospital readmission and other adverse outcomes. The present study examined the 30-day hospital readmission rates of 1,081 patients with a native, low-energy proximal femur fracture treated surgically at a level 3 trauma program over a 65-month period. The Geriatric Fracture Center Model of care was employed at our center [11 13]. Details of this model have been previously published [11 13]. We examined causes of readmission and patient outcomes after readmission. The length of stay and complications experienced by the readmitted patients were compared to hip fracture patients who were not readmitted. Our hypothesis was that the GFC care model will result in similar hospital readmission rates following hip fracture treatment compared to national rates. Our primary outcome measurement was the rate of 30-day readmission to the hospital after discharge. A secondary outcome measurement was the 30-day mortality rate for those patients readmitted versus those patients not readmitted to the hospital after discharge. We also hypothesized that readmissions would reflect the underlying comorbidities and frailty of the hip fracture population, and that most readmissions would be for medical rather than surgical diagnoses. We sought to identify patient factors and potentially preventable causes of readmission that might be able to direct targeted efforts to reduce hospital readmission in this at-risk population. Materials and methods Setting Patients aged C65 admitted to a university affiliated level 3 trauma center between 4/30/2005 and 9/30/2010 with a unilateral native low-energy hip fracture were identified from our fracture registry and included for analysis; details of the registry have been published previously [11]. Patients with peri-prosthetic fractures, pathologic fractures, bilateral injuries and high-energy mechanisms were excluded. Patients who were not managed surgically were also excluded. Patients who died during their initial hospitalization were excluded as they could not be readmitted. The inclusion and exclusion criteria were designed to examine the frequently encountered low-energy hip fracture that accounts for the majority of cases seen in most centers. The Geriatric Fracture Center (GFC) program uses a co-managed, patient-centered standardized protocol driven approach to hip fracture care [11, 12]. This model offers comprehensive care to older patients who frequently are burdened by multiple comorbidities. Each patient is seen daily by a geriatrician and an orthopaedic surgeon with regular communication and shared responsibility for their care. Data Data were collected for all patients using retrospective chart reviews completed by a member of the research team which was subsequently entered into the registry as part of a hospital quality management initiative. Data collected included demographic information, comorbidities (as measured in the Charlson comorbidity index [14]), baseline mobility (as measured by the Parker mobility scale [15]), surgical management, in-hospital complications, as well as the occurrence and cause for hospital readmission within 30 days of discharge. Readmissions were identified in several complementary ways. Readmissions within our own University Healthcare System, which includes two hospitals, were identified using our administrative billing records software. Beginning in 2008, we began conducting routine post-discharge followup telephone calls to all patients or their family members/ caregivers to assess post-hospital outcomes including readmission; however, our call system was not implemented for the first 2.5 years of our study and thus it is nearly certain that some readmissions were missed. For all patients who were readmitted, three of the clinician authors reviewed the entire available medical record separately and recorded the primary diagnosis responsible for readmission and whether the admission was possibly preventable or not. Discrepancies were resolved through discussion. If an unclear admitting diagnosis was recorded upon readmission (such as fever or pain), the primary diagnosis was determined by examining the medical record from the readmission stay. Analysis We used descriptive statistics to describe the demographic characteristics and comorbidity of our fracture cohort. We used similar statistics to describe hospital length of stay (LOS) and readmission rates. Second, for patients who were readmitted, we evaluated the primary cause of

3 readmission; the primary cause for readmission was categorized by disease category and specific event with specific subgroups to provide additional detail. Third, we used bivariate methods (Chi-square test) to compare patients who were and were not readmitted. Finally, we used logistic regression models to examine the association between patients readmitted and those patients not readmitted while controlling for differences in a patient s age, gender, preexisting medical conditions, in hospital complications, ambulatory status, type of residence, and time to surgery as predictors of a readmission event. Statistical analysis was performed using SPSS Ò Statistics 20 (IBM Ò Chicago, Illinois) software to calculate descriptive statistics, cumulative risk incidence, and mortality rates. Uncertainty is presented as 95 % confidence intervals, with statistical significance being reached when p \ This study was approved by the university research subjects review board. Results There were 1,081 patients who met the inclusion criteria as summarized in Table 1. The mean age was 85.1 ± 8.4 years. Seventy-six percent of the population was female and 94.8 % were Caucasian. Prior to fracture 48.9 % of patients were living at home, 37.6 % in a skilled nursing facility, and 13.5 % from assisted living facility. Dementia was present in 47.4 %. The average time from hospital admission to surgery was 25 h, 35 min, with a hospital length of stay of 4.6 ± 2.3 days. 129 patients (11.9 %) were readmitted to the hospital within 30 days of their initial discharge date. The primary causes of readmission were surgical in nature for 24/129 (18.6 %) patients and 105/129 (81.4 %) were readmitted for medical or other reasons. Surgical causes included: fixation failure, re-fracture, new fracture, dislocation, hematoma and wound complications. There were a total of 24 other medical diagnoses associated with 105 medical readmissions. Individual diagnoses are listed in Table 2. Each readmission was assigned 1 primary root cause for readmission. Many of the patients had secondary contributing factors that were not directly analyzed. The average length of stay for patients readmitted for surgical reasons was 6.3 days. Patients readmitted with medical diagnoses had an average length of stay of 8.8 days. Pulmonary problems (primarily pneumonia) were the most common medical reason for readmission at 27.1 % (35 patients). A full list of readmission diagnoses is shown in Table 2. Twenty patients (15.6 %) were felt to have readmissions that were potentially preventable. Preventable causes of Table 1 Characteristics of patients Total, n 1,081 Gender, % Female 76.0 Age Mean ± SD 85.1 ± 8.4 Race, % Caucasian 94.8 Hispanic 1.3 Black 1.2 Asian 2.1 Native American 0.2 Other 0.3 Pre-fracture residence, % Community 48.9 Skilled nursing facility 37.6 Assisted living facility 13.5 Charlson score 3.1 ± 2.1 Mean ± SD Dementia, % 47.4 LOS, days 4.6 ± 2.3 LOS at readmit, days 8.7 ± 18.6 Parker Mobility score 3.8 ± 3.2 Readmission rate, % 11.9 Re-operation rate, % 0.82 readmission include constipation, some surgical failures, C. difficile infection, superficial wound infection, pressure ulcers, some cases of urinary tract infection and some cases of congestive heart failure. Twenty-four (18.6 %) patients who were readmitted died during readmission. The 1-year mortality rate for patients readmitted within 30 days was 56.2 vs. a 21.8 % 1-year mortality rate for those patients not readmitted (p \ ). The risk of readmission according to baseline characteristics is presented in Table 3. Older patients and men were more likely to be readmitted within 30 days. Patients with more functional dependence and those with dementia at baseline had a trend toward higher risk of readmission. Independent predictors of readmission were age [85 (OR = 1.52; p = 0.03), time to surgery [24 h (OR = 1.50; p = 0.05), Charlson score C4 (OR = 1.70; p = 0.04), delirium (OR = 1.65; p = 0.01), dementia (OR = 1.61; p = 0.01), history of arrhythmia with pacemaker placement (OR = 1.75; p = 0.02), and presence of a pre-op arrhythmia (OR = 1.62; p = 0.02). A complete listing of odds ratios for readmission is represented in Table 4. Age category, occurrence of an in-hospital complication, the presence of arrhythmia with our without

4 Table 2 Causes for readmission by organ system Organ system Subtotal Total % Pulmonary Pneumonia Respiratory failure 6 Chronic obstructive 2 Gastrointestinal Gastrointestinal bleed Small bowel obstruction 3 Fecal Impaction* 3 C. difficile infection* 6 Ileus 2 Failure to thrive 2 Neurologic Stroke Delirium 2 Seizure 2 Intracranial hemorrhage 1 Cardiovascular Congestive heart failure* Atrial fibrillation 7 Myocardial infarction 2 Musculoskeletal Re-fracture Failure of fixation* 3 New site fracture 7 Deep wound infection 3 Superficial wound infection* 2 Dislocation joint 2 Pressure ulcer* 3 Hematoma 1 Genitourinary Urinary infection* Urosepsis 2 Urinary retention 1 Acute renal failure 3 Electrolyte abnormality 2 Hematologic Anemia Pulmonary emboli or deep venous 3 thrombosis Other Total * Diagnoses that include potentially avoidable readmissions pacemaker placement, and the presence of cognitive impairment were used to estimate subgroups who were at high risk of readmission. Patients over the age of 85 with cognitive impairment or a history of arrhythmia with or without pacemaker had a 22 % readmission rate. If patients were at risk for hospital-acquired pneumonia [16] the readmission rate was 28.4 %. Measured risk factors for hospital-acquired pneumonia include male gender, impaired mobility, history of chronic pulmonary disease, history of reflux disease, or prior history of pneumonia [16]. Additional risk factors dysphagia or poor oral hygiene were not documented in the database and could not be examined. Patients who had a complication during their initial hospitalization had a readmission rate of 31.1 %. Discussion Hospital readmission after hip fracture occurs frequently and is associated with prolonged hospitalization and high in-hospital and 1-year mortality rates. Readmission is unwelcome from the patient s viewpoint as it typically represents a delay in their recovery process and may be harmful to their wellbeing. Diagnoses leading to readmission are often serious medical and surgical complications. Readmission is certainly problematic from the physician s standpoint as it may represent a failure of the initial efforts at caring for the patient. As we hypothesized, over 81 % of patients who are readmitted have primary medical diagnoses as the cause for readmission. This reflects the frail nature of hip fracture patients, who have a high prevalence of medical comorbidities and are at risk for multiple complications resulting from a major and urgent surgery. It also reinforces the need for careful and multi-faceted attention to detail in the index admission. Khan et al. [17] reached similar conclusions in their UK-based series. There has been surprisingly little written on the subject of readmission following hip fracture although it is recognized as a problem by physicians, hospitals, and the US government. French et al. [4] described a 30-day readmission rate of 18.3 % using claims data from 41,331 US veterans aged C65 years with a hip fracture. The readmitted patients in that study had a 1-year mortality rate of 48.5 % compared with a 24.9 % mortality rate in veterans who were not readmitted [4]. Additionally, French et al. identified specific comorbidities that were risk factors for readmission using logistical regression analysis. Fluid and electrolyte disorders, renal disease, cardiac arrhythmia, congestive heart failure, and chronic pulmonary disease were associated with an increased risk of 30-day readmission [4]. Bookvar et al. [3] described a prospective analysis of 562 hip fracture patients aged C50 years. They noted a 14.2 % 1-month readmission rate. Of these readmissions, 11 % were readmitted for surgical causes and 89 % were readmitted for medical reasons, similar to our findings. Only 12 % of their patients were admitted initially from a

5 Table 3 The risk of readmission according to baseline characteristics In this study population, age and gender were statistically significant predictors of readmission Characteristic Total Readmitted Rate (%) Significance 1, Age C Gender Male Female Residence Community Assisted living facility Skilled nursing facility Pre-operative Parker Mobility Score High (9) Medium (5 8) Low (0 4) Pre-operative function Independent Partial dependence Dependent Charlson score Low (0 1) Medium (2 3) High (4 or more) Dementia Yes No nursing home. They also describe a worsened prognosis for readmitted patients. Readmitted patients in their series were found to have an increased risk of mortality, impaired gait and residence in a nursing home 6 months following fracture [3]. Jencks et al. [5] published a 17.9 % 30-day readmission rate after hip fracture surgery (DRG 210) and cited pneumonia and CHF as being the two most frequent causes of readmission. This population-based study used Medicare claims data from [5]. The Dartmouth Atlas report describes some interesting trends in hospital readmissions after hip fracture during the past decade [1]. The US national readmission rates following hip fracture were 14.3 % in 2004 and 14.5 % in 2009 [1]. For New York State hospitals, the rate in 2004 was 14.5 % and increased to 15.3 % in 2009 [1]. Possible reasons for the slight increase in readmissions include shorter lengths of stay, increased age of the patients and increased burden of comorbidity accompanying these patients. In the Dartmouth Atlas report, 89.8 % of hip fracture patients were discharged to a nursing home. Khan et al. found an 11.8 % readmission rate in their 28-day discharge series of 467 patients. Pneumonia was the most common cause of readmission in their series. Their hospital length of stay was typically [20 days [17]. Our study shows a slightly lower than expected readmission rate using the GFC model of care (11.9 vs % NY State rate) when compared with the previous reports [1, 3 5]. This is despite a patient population that is at higher risk for adverse outcomes in several ways. The patients treated in the GFC were more likely to be admitted for their index hospitalization from a facility (51.1 vs. 12 % reported in Bookvar s series). Nursing home residence prior to fracture has previously been shown to be a risk factor for mortality and poor outcomes after hip fracture [11]. The length of stay is considerably shorter for the GFC

6 Table 4 Independent predictors of readmission Factor Odds 95 % confidence Significance Age [ Female Assisted living Skilled nursing Time to surgery [24 h Parker Mobility Score Medium (5 8) Low (0 4) Activities of daily living Partial or complete disability Charlson score Medium (2 3) High (4 or more) Past medical history Pacemaker GERD Diabetes Dementia Cardiac disease Deep venous thrombosis Alcoholism Tobacco Congestive heart failure In-hospital complication Delirium Hematoma Urinary tract infection Pre-operative arrhythmia patients (4.6 days) than the previous series. It has been hypothesized that lower length of stay may increase risk of readmission, as patients are discharged sicker and quicker. Our series shows 18.6 % of readmitted patients died during their second hospitalization. The average length of stay during readmission was 8.7 days compared with 4.6 days for the original stay. Furthermore, the 1-year mortality of patients who are readmitted is 56 vs. 22 % for those who are not. These statistics demonstrate the serious impact that readmissions have on patients prognosis. Khan et al. from the UK showed similarly elevated rates of mortality for patients readmitted after hip fracture (41.8 vs %). This UK study of 467 patients also showed the most common causes of 28-day readmission to be pneumonia, dehydration and renal dysfunction [17]. The Dartmouth Atlas report on readmissions highlights the considerable variation seen in readmission rates seen among both community and academic medical centers [1]. The specific causes for this variation are not clear. There has been no improvement in readmission rates over the past decade from 2004 to 2009 [1]. Many causes for readmission have been described including communication issues, problems with medication reconciliation, lack of satisfactory follow-up care, and defects in the original inpatient care [1, 18]. Many hospital readmissions are likely preventable in nature [1, 2]. The GFC model of care only addresses the inpatient care phase [12], but incorporates several elements into standardized practice that may improve transitions and reduce readmissions, including standardized best practices to improve early mobilization, reduce delirium, and minimize respiratory and bowel issues; an interdisciplinary discharge summary; and careful medication reconciliation. There have been no interventions extending past the inpatient stay. The authors assessment was that 15.6 % of the readmissions were likely of a preventable nature. Preventable causes of readmission include constipation, some surgical failures, C. difficile infections, superficial wound infection, pressure ulcers, some cases of urinary tract infection and some cases of congestive heart failure [19]. There are

7 likely some cases that could be avoided with improved communication with receiving providers at the time of the discharge hand-off [20]. Some surgical failures are preventable by proper implant position and accurate fracture reduction [19]. Many pressure ulcers are also preventable by early surgery, and aggressive prevention programs [21]. One surprising cause for readmission was constipation. One of the patients readmitted for this condition died during the readmission. This should be preventable in most cases with education, an aggressive bowel regimen and careful attention to the problem [22]. Clostridium difficile infections are preventable in many cases with proper hand hygiene, appropriate antibiotic use and environmental disinfection [23]. Some cases of congestive heart failure caused readmission can likely be avoided by attention to proper fluid balance while in the hospital for the original admission [24]. Appropriate diuresis during the original admission may be able to avoid some of these readmissions. There have been several successful methods published for reduction of readmissions following medical hospitalization including early follow-up care with the primary care physician, the Coleman discharge coaching model and the Naylor model [20, 25, 26]. There have been no previously published successful methods for reduction of readmission following hip fractures. Additional efforts at improving the discharge process, communication and postdischarge follow-up may reduce the readmission rates. There are no published risk indices for prediction of readmission following discharge after hip fracture. Giusti et al. have identified comorbidities and poor functional status at rehabilitation as predictive factors for hospital readmission following hip fracture [27]. French has indicated that patients with comorbidities of fluid and electrolyte disturbances, renal failure, cardiac arrhythmia, COPD and CHF, are more likely to be readmitted after hip fracture [4]. Since these are common comorbidities in the elderly hip fracture population, these risk factors offer limited predictive power in the clinical setting. Our study shows two potentially modifiable risk factors, namely time to surgery \24 h and delirium. It is possible that a delay in surgery is a proxy for unstable medical conditions at admission, which may in turn increase risk of readmission. In this study, we did not distinguish between delays due to medical conditions vs. other causes. Another possibly modifiable risk factor is delirium which can be prevented in some patients with a multimodal approach. Our program uses several modalities to reduce the incidence of delirium [13], but it may be possible to reduce this rate further. The non-modifiable risk factors for readmission include age [85, pre-operative arrhythmia, dementia, history of prior pacemaker placement, and Charlson comorbidity score C4. These risk factors should alert the clinicians caring for the hip fracture patient that the patient is at elevated risk for readmission. Given the high risk of mortality in patients who are readmitted, patients with these characteristics may benefit from discussions about goals of care during their index admission. Limitations of this study There are several important limitations of this study. The sample size of 1,081 with 129 readmissions is certainly not large enough to generalize these results. Larger, multicenter studies may be useful to determine if these results can be generalized or translated to other centers. This study evaluates risk factors for and incidence of readmission in a Geriatric Fracture Center model of care. Although this model is being adopted at other centers as a way of optimizing care and outcomes, it is not available in most centers. The data are generalizable to patients in this model of care. Another limitation is the retrospective nature of the data collection which may not fully capture all readmissions or adverse events. As a countermeasure, we have tried to capture all the 30-day readmissions by reviewing medical records, hospital information systems and calling the patients or families following discharge. The patients included in this study may not be representative of the populations seen at many centers. Over half of our patient population was admitted from a nursing home or assisted (residential care) living home, whereas most published studies describe % of hip fracture patients come from a home living setting. Additionally, there was no control group or pre-implementation data available for our program. We were also unable to access detailed hospitalization records for the 6 patients readmitted to regional hospitals. This is a limitation inherent to the US healthcare system where medical data are typically not shared between regional hospitals. Conclusion The GFC model of care has shown a slightly lower than expected readmission rate (11.9 vs %) following hip fracture in this series of 1,081 consecutive patients. Patients were readmitted most frequently for medical complications following their original hospital stay. The most common reasons for readmission include pulmonary, musculoskeletal, gastrointestinal and cardiovascular complications. Readmission is harmful and undesirable 18.6 % of readmitted patients died during their readmission and the average LOS was 8.7 days. The 1-year mortality rate for

8 patients readmitted within 30 days was 56.2 vs. a 21.8 % 1-year mortality rate for those patients not readmitted (p \ ). Approximately one of every six readmissions was identified as preventable with further reduction possible with early surgery, delirium avoidance, improved hand-offs at transfer, better attention to bowel regimens and careful attention to medical problems present at the time of discharge. Some hospital system changes may be able to reduce Clostridium difficile-related readmissions. It is likely that a multifactorial approach to readmission reduction will be required given the broad range of conditions resulting in readmission. Development of a model to predict those patients at highest risk for readmission may be possible with a much larger cohort of patients. Acknowledgments The authors would like to acknowledge assistance provided by Mary Sears, LPN for data management and collection. There was no direct support provided for this study. Partial program support was provided by an unrestricted grant from Synthes Spine paid to the hospital. Conflict of interest The hospital program has received institutional support from Synthes Spine to assist with data collection. No direct support of any kind was received for this study. References 1. Goodman DC, Fisher ES, Chang CH (2011) After hospitalization: a Dartmouth Atlas report on post-acute care for Medicare beneficiaries. Dartmouth, Hanover 2. Gorodeski EZ, Starling RC, Blackstone EH (2010) Are all readmissions bad readmissions? N Engl J Med 363(3): doi: /nejmc Boockvar KS, Halm EA, Litke A, Silberzweig SB, McLaughlin M, Penrod JD, Magaziner J, Koval K, Strauss E, Siu AL (2003) Hospital readmissions after hospital discharge for hip fracture: surgical and nonsurgical causes and effect on outcomes. J Am Geriatr Soc 51(3): (pii:jgs51115) 4. French DD, Bass E, Bradham DD, Campbell RR, Rubenstein LZ (2008) Rehospitalization after hip fracture: predictors and prognosis from a national veterans study. J Am Geriatr Soc 56(4): doi: /j x 5. Jencks SF, Williams MV, Coleman EA (2009) Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med 360(14): doi: /nejmsa Braithwaite RS, Col NF, Wong JB (2003) Estimating hip fracture morbidity, mortality and costs. J Am Geriatr Soc 51(3): Brauer CA, Coca-Perraillon M, Cutler DM, Rosen AB (2009) Incidence and mortality of hip fractures in the United States. JAMA 302(14): doi: /jama Melton LJ 3rd (1990) Hip fracture incidence and survival among members of a California medical care program. Clin Orthop Relat Res 256: Burge R, Dawson-Hughes B, Solomon DH, Wong JB, King A, Tosteson A (2007) Incidence and economic burden of osteoporosis-related fractures in the United States, J Bone Miner Res 22(3): doi: /jbmr Barrett M, Wilson E, Whalen D (2010) 2007 Nationwide Inpatient Sample Comparison Report. AHRQ, Rockville, MD, 9 Sept Kates SL, Mendelson DA, Friedman SM (2010) Co-managed care for fragility hip fractures (Rochester model). Osteoporos Int 21(Suppl 4):S621 S625. doi: /s Friedman SM, Mendelson DA, Kates SL, McCann RM (2008) Geriatric co-management of proximal femur fractures: total quality management and protocol-driven care result in better outcomes for a frail patient population. J Am Geriatr Soc 56(7): doi: /j x 13. Friedman SM, Mendelson DA, Bingham KW, Kates SL (2009) Impact of a comanaged Geriatric Fracture Center on short-term hip fracture outcomes. Arch Intern Med 169(18): doi: /archinternmed Charlson ME, Pompei P, Ales KL, MacKenzie CR (1987) A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 40(5): Parker MJ, Palmer CR (1993) A new mobility score for predicting mortality after hip fracture. J Bone Joint Surg Br 75(5): Kollef MH, Shorr A, Tabak YP, Gupta V, Liu LZ, Johannes RS (2005) Epidemiology and outcomes of health-care-associated pneumonia: results from a large US database of culture-positive pneumonia. Chest 128(6): doi: /chest Khan MA, Hossain FS, Dashti Z, Muthukumar N (2012) Causes and predictors of early re-admission after surgery for a fracture of the hip. J Bone Joint Surg Br 94(5): doi: / x.94b Bell CM, Brener SS, Gunraj N, Huo C, Bierman AS, Scales DC, Bajcar J, Zwarenstein M, Urbach DR (2011) Association of ICU or hospital admission with unintentional discontinuation of medications for chronic diseases. JAMA 306(8): doi: /jama Baumgaertner MR, Curtin SL, Lindskog DM, Keggi JM (1995) The value of the tip-apex distance in predicting failure of fixation of peritrochanteric fractures of the hip. J Bone Joint Surg Am 77(7): Peikes D, Chen A, Schore J, Brown R (2009) Effects of care coordination on hospitalization, quality of care, and health care expenditures among Medicare beneficiaries: 15 randomized trials. JAMA 301(6): doi: /jama Lyder CH, Wang Y, Metersky M, Curry M, Kliman R, Verzier NR, Hunt DR (2012) Hospital-acquired pressure ulcers: results from the national Medicare Patient Safety Monitoring System study. J Am Geriatr Soc 60(9): doi: /j x 22. Pieper B, Sieggreen M, Freeland B, Kulwicki P, Frattaroli M, Sidor D, Palleschi MT, Burns J, Bednarski D, Garretson B (2006) Discharge information needs of patients after surgery. J Wound Ostomy Cont Nurs 33(3): (pii: ) 23. Surawicz CM, Brandt LJ, Binion DG, Ananthakrishnan AN, Curry SR, Gilligan PH, McFarland LV, Mellow M, Zuckerbraun BS (2013) Guidelines for diagnosis, treatment, and prevention of clostridium difficile infections. Am J Gastroenterol. doi: / ajg ajg Retrum JH, Boggs J, Hersh A, Wright L, Main DS, Magid DJ, Allen LA (2013) Patient-Identified Factors Related to Heart Failure Readmissions. Circ Cardiovasc Qual Outcomes. doi: /CIRCOUTCOMES Naylor MD, Brooten DA, Campbell RL, Maislin G, McCauley KM, Schwartz JS (2004) Transitional care of older adults hospitalized with heart failure: a randomized, controlled trial. J Am Geriatr Soc 52(5): doi: /j xJGS52202

9 26. Coleman EA, Parry C, Chalmers S, Min SJ (2006) The care transitions intervention: results of a randomized controlled trial. Arch Intern Med 166(17): doi: /archinte Giusti A, Barone A, Razzano M, Pizzonia M, Oliveri M, Pioli G (2008) Predictors of hospital readmission in a cohort of 236 elderly discharged after surgical repair of hip fracture: one-year follow-up. Aging Clin Exp Res 20(3): (pii:4665)