Causes of Hospital Readmissions at the Community Level

Reducing inpatient hospital readmissions has been an important component of efforts to improve outcomes and reduce health care costs. This study focused on evaluation of the clinical causes of hospital readmissions of adult medical/surgical patients within 30 days between October 2015 and September 2016. It was based on the principal diagnoses of readmissions, a definition that is used throughout the health care industry in the United States. The study focused on adult medicine and adult surgery readmissions in Syracuse, New York, a small metropolitan area, during a twelve month period. It included almost 4000 individual readmissions. The study data demonstrated that only about 22 percent of inpatient readmissions were for the same diagnoses as the initial admissions that preceded them. The study data also indicated that another 20 percent of hospital readmissions involved a diagnosis different from that of the initial admission but in the same body system. Most importantly, the study demonstrated that a consistent majority of inpatient readmissions were caused by diagnoses in different body systems than the initial. The data suggested that efforts to address the causes of hospital readmissions should be based on management of a broad range of adult medicine conditions, rather than individual diagnoses.


Introduction
Historically, improving the efficiency of health care has been a major concern in the United States.This subject has gained increased attention with efforts by the new administration in Washington to contain the costs of Medicaid and Medi-care.Both of these major payers could be addressed with limitations in federal funding [1] [2] [3].
In recent years, a recognition has developed that improvements in the efficiency and outcomes of care are linked.This recognition has stimulated the development of efforts to reduce adverse outcomes and related costs.Because of their high costs, the reduction of hospital readmissions has become an important focus of these efforts [4] [5] [6].
These initiatives have included the development of studies of factors related to readmissions.Researchers have developed models of readmissions that can predict these outcomes.These studies have produced much useful information, but have failed to develop models that can predict readmissions with a high degree of accuracy [7] [8].
These studies have substantially increased understanding of hospital inpatient readmissions.They have been supported by the development of computer tools that address these outcomes [9].Through all of this research, one of the most important aspects of this subject has been the clinical causes of hospital readmissions.These causes involve broad ranges of diagnoses rather than individual conditions.Definition and understanding of them appear to be essential to progress in improving these outcomes [10].

Population
This study evaluated causes of inpatient readmissions within 30 days of the initial admission in the metropolitan area of Syracuse, New York.This area includes three large acute care facilities, Crouse Hospital (19,478 inpatient discharges excluding well newborns, 2016), St. Joseph's Hospital Health Center (25,101 discharges, 2016), and the State University of New York Upstate University Hospital (29,427 discharges, 2016).These hospitals have provided primary and secondary acute care to an immediate service area with a population of approximately 600,000 and tertiary services to the 11 county Central New York Health Service Area with a population of 1,400,000.
Historically, the Syracuse hospitals have maintained a relatively low inpatient admission rate, despite demographics that may contribute to increased readmission rates among disadvantaged populations.This rate has been comparable to those of metropolitan areas such as Rochester and Albany, New York that have higher managed care penetration [11].
The Syracuse hospitals have worked cooperatively to improve the efficiency and outcomes of care in the community through the Hospital Executive Council.These efforts have included the reduction of hospital lengths of stay and readmissions, as well as the development of subacute and complex care programs aimed at supporting efficient transitions of care for difficult to place patients.
The readmissions program has been carried out in cooperation with 3M Health Information Systems.

Method
This study evaluated clinical causes of inpatient readmissions in the hospitals of Syracuse, New York during a twelve month period.Economic and social determinants of health as potential causes of readmissions were not examined.It included more than 4000 individual adult medicine and adult surgery readmissions, 82.8 percent of the total, in the area's three acute hospitals [11].It was based on simple descriptive statistics.
The study was carried out using patient specific data from each of the hospitals by the Hospital Executive Council.These data were obtained through Business Associate Agreements with each of the hospitals.The Council functions as a mechanism for the development of multihospital studies in the Syracuse metropolitan area.
Hospital readmissions were identified using the Potentially Preventable Readmissions system developed by 3M Health Information Systems.This software uses hospital administrative data to identify readmissions within 30 days of the initial admission and a number of clinical and demographic indicators for each patient.
Readmissions were identified for adult medicine and adult surgery patients using the All Patients Refined Diagnosis Related Group System (APR DRG).
This system identifies the hospital service of each inpatient based on the principal diagnosis, secondary diagnoses, principal procedure, and other clinical and demographic characteristics.

Results
The initial component of the study focused on identification of the clinical causes of a full range of readmissions for adult medicine and adult surgery in the Syracuse hospitals.The remaining readmissions were produced by obstetrics, pediatrics, and mental health services.Related data are summarized in Table 1  Syracuse hospitals, a majority of these readmissions involved a combination of infectious diseases and adult medicine conditions such as respiratory, circulato-ry, and digestive disorders.During the twelve month period of the study, these four Major Diagnostic Categories were associated with 60.4 to 63.6 percent of readmissions that occurred outside the MDC of the initial admission.
The study data indicated that infectious diseases, MDC 18, were the single largest source of readmissions outside the original Major Diagnostic Category for each hospital and time period.Within this Major Diagnostic Category, sepsis was the major cause of readmissions.The data suggested that these diagnoses were related to readmissions and other utilization issues such as extended stays.
The study data also demonstrated that large numbers of readmissions that occurred outside the Major Diagnostic Category on the initial admission involved anatomical areas such as the respiratory, circulatory, and digestive systems that have been sources of most adult medicine hospital inpatient admissions and readmissions.Most of these conditions were present on the initial admissions as secondary diagnoses.
This portion of the study demonstrated that the largest percentages of medical and surgical readmissions involved more than one body system.The causes of these admissions were identified within the definition of adult medicine.

Discussion
The reduction of hospital inpatient readmissions has been a major focus of efforts to improve patient outcomes and reduce health care costs.Significant attention has been applied nationally to the management of chronic conditions such as heart failure, chronic obstructive pulmonary disease, and diabetes in hospitals in efforts to reduce recidivism.The complicated nature of readmissions suggests, however, that efforts to address them need to be accompanied by careful evaluation of a wide range of clinical causes rather than individual diagnoses.
This study focused on evaluation of the clinical causes of hospital readmissions of adult medical-surgical patients within 30 days of the admissions that preceded them.It was based on the principal diagnoses of readmissions.The principal diagnosis is defined as the condition that was responsible for each hospital admission.This is a definition that is used throughout the health care industry in the United States.
The study focused on adult medicine and adult surgery readmissions in the hospitals that comprised the acute care system of a small metropolitan area during a twelve month period.It included almost 4000 individual readmissions, approximately 83 percent of the community total.
The study data demonstrated that, at the aggregate and hospital specific levels, only about 22 percent of inpatient readmissions were for the same diagnosis as the initial admission that preceded them.The quarterly ranges of each of the three hospitals were consistent with this rate.The data suggested that the most direct relationship between initial admissions and readmissions existed for a minority of all adult medicine and adult surgery readmissions in the community.
The study data indicated that another 20 percent of hospital readmissions in the population involved a diagnosis different than that of the initial admission,

For
purposes of this study, the clinical cause of each inpatient readmission was identified as the All Patients Refined Diagnosis Related Group.This indicator is based on the principal diagnosis or principal procedure of the inpatient stay.The principal diagnosis is the condition, which in the opinion of the discharge abstractor, was the principal cause of the admission.The analysis also included Major Diagnostic Categories (MDCs) which are collections of APR DRGs by anatomical areas.The first component of the analysis focused on clinical causes of inpatient adult medicine and adult surgery readmissions for the period October 2015-September 2016.This was the latest time interval for which complete data were available.Each of the readmissions was identified as one of three categories.The first was a return in the same APR DRG as the initial admission, such as a Chronic Obstructive Pulmonary Disease (COPD) patient returning for COPD.The second was a return in the same MDC as the initial admission but a different APR DRG, such as a COPD patient returning for pneumonia.The third category was a return in a different MDC than the initial admission, such as a COPD patient returning for a digestive disorder.Numbers of readmissions were identified for each of the three categories by hospital and total.The data were generated for adult medicine and adult surgery readmissions for October-December 2015, January-March 2016, April-June 2016, and July-September 2016.Differences in numbers of readmissions and percen-tages of the adult medicine and adult surgery total among the three categories and for the combined population were identified.The second component of the analysis focused on clinical causes of inpatient readmissions in the Syracuse hospitals for the third category, those who returned with a Major Diagnostic Category outside the MDC of the Initial Admission.These data were aggregated for each of the three month periods by MDC for the combined hospitals.Differences in numbers of readmissions by MDC were compared.
Source: Hospital Executive Council.

Table 2 .
Potentially preventable readmissions within 30 days, readmissions to different APR Major Diagnostic Category medical/surgical patients-all Payors, Syracuse hospitals.
Source: Hospital Executive Council.