ORIGINAL ARTICLE
Predicting Factors and Risk Stratification for Return Visits to the Emergency Department Within 72 Hours in Pediatric Patients Sheng-Feng Sung, MD,* Kang Ernest Liu, PhD,† Solomon Chih-Cheng Chen, MD, PhD,‡ Chia-Lun Lo, MS,§ Kuei-Chih Lin, MS,k and Ya-Han Hu, PhD§ Objectives: A return visit (RV) to the emergency department (ED) is usually used as a quality indicator for EDs. A thorough comprehension of factors affecting RVs is beneficial to enhancing the quality of emergency care. We performed this study to identify pediatric patients at high risk of RVs using readily available characteristics during an ED visit. Methods: We retrospectively collected data of pediatric patients visiting 6 branches of an urban hospital during 2007. Potential variables were analyzed using a multivariable logistic regression analysis to determine factors associated with RVs and a classification and regression tree technique to identify high-risk groups. Results: Of the 35,435 visits from which patients were discharged home, 2291 (6.47%) visits incurred an RV within 72 hours. On multivariable analysis, younger age, weekday visits, diagnoses belonging to the category of symptoms, signs, and ill-defined conditions, and being seen by a female physician were associated with a higher probability of RVs. Children younger than 6.5 years who visited on weekdays or between midnight and 8:00 AM on weekends or holidays had the highest probability of returning to the ED within 72 hours. Conclusions: Our study reexamined several important factors that could affect RVs of pediatric patients to the ED and identified high-risk groups of RVs. Further intervention studies or qualitative research could be targeted on these at-risk groups. Key Words: classification and regression tree, emergencies, health services, return visits (Pediatr Emer Care 2015;31: 819–824)
T
aiwan's National Health Insurance has been well acknowledged for its many merits, including low cost. Consequently, most of the emergency departments (EDs) in Taiwan are overloaded with patients.1 The large number of patients using the ED services makes the administrative management of emergency medicine increasingly important to assure a high quality of health care with respect to suitable cost controls. The frequency of return visits (RVs) to EDs is usually used as an indicator for quality of care2,3 and is currently included in the Taiwan Clinical Performance Indicators to measure quality improvement. High RV rates From the *Department of Internal Medicine, Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi City; †Department of Economics, National Chung Cheng University, Chiayi County; ‡Departments of Pediatrics and Medical Research, Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi City; §Department of Information Management and Institute of Healthcare Information Management, National Chung Cheng University, Chiayi County; and kInformation Technology Office, Taipei City Hospital, Taipei, Taiwan. Sheng-Feng Sung and Kang Ernest Liu contributed equally to this work. Disclosure: The authors declare no conflict of interest. Reprints: Ya-Han Hu, PhD, Department of Information Management and Institute of Healthcare Information Management, National Chung Cheng University, 168 University Rd, Min-Hsiung, Chiayi County 62102, Taiwan (e‐mail: [email protected]). Supported by National Science Council of the Republic of China under the Grants NSC 102-2410-H-194-104-MY2 and NSC 102-2410-H-194-087. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0749-5161
may be attributed to improper treatment at the initial visit, insufficient education at discharge, system failures, inadequate access to alternative primary care services, or even missed identification of subtle signs of disease processes.3,4 Some RVs may be avoidable or unnecessary and thus are costly both in financial terms and in the burden they place on the health care system. To improve the quality of emergency care, a thorough understanding of the underlying factors of RVs to the ED is necessary. Factors related to the illness, the patient, the physician, the health care system, or the ED visit may induce RVs.5–13 It is difficult to implement any specific intervention before we understand how these factors are related to RVs to the ED. Although factors affecting RVs within 72 hours in adults have been reported in Taiwan,14,15 the characteristics associated with RVs in children are not fully understood. Therefore, we conducted this study to examine the association between characteristics readily available during an ED visit and RVs within 72 hours in pediatric patients. We also sought to identify high-risk subgroups for RVs that may serve as a potential target for interventions.
METHODS Patients This retrospective cohort study was performed in the Taipei City Hospital, which successfully integrated 10 municipal hospitals into 1 unit in 2005 to provide comprehensive health care needs to 2.6 million residents in Taipei. Records of pediatric patients from 6 major branches, including Zhongxing, Renai, Heping, Yangming, Zhongxiao, and Fuyou, were used in this study. The Fuyou branch is specialized as a pediatric hospital, where pediatricians or pediatric emergency medicine physicians provide emergency care, whereas the remaining 5 branches are general hospitals, where only some patients are seen by adult emergency medicine physicians. During the study period, there were 8 medical centers and 12 regional hospitals in Taipei. All of the study hospitals were regional hospitals. The Taipei City Hospital Institutional Review Board approved this study. We reviewed the hospital's administrative database and identified all visits made by patients aged 18 years or younger to the EDs of the 6 branches between January 1 and December 31 in 2007. The patients who left without being seen, left against medical advice, had out-of-hospital cardiac arrest, were admitted to the hospital, or were transferred to another hospital for admission were excluded from the analysis. In addition, the records with unknown disposition and miscellaneous missing data, such as unknown time of registration or disposition, unrecorded triage level, and missing date of birth, were also dropped out. The remaining visits comprised the study sample (index visits). Potential variables affecting the rate of RVs were collected and were grouped into the following 4 categories: those related to the patient, those related to the hospital, those related to the
Pediatric Emergency Care • Volume 31, Number 12, December 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.pec-online.com
819
Pediatric Emergency Care • Volume 31, Number 12, December 2015
Sung et al
physician, and those related to the ED visit. Age, sex, and whether the patient has chronic diseases, for example, diabetes mellitus, hypertension, asthma, endocrine disorders, hypertensive cardiovascular diseases, and hypertensive renal diseases, were considered as potential patient-related factors. Whether the hospital is general or pediatric specialized was treated as a characteristic of the hospital. Variables related to the physician included age, sex, and whether the treating physician is a board-certified pediatrician. As for the characteristics of ED visits, we collected the level of triage, the time of registration in the ED and time of disposition from the ED, as well as the primary discharge diagnosis based on codes from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). Nursing triage levels were categorized as emergent, urgent, or nonurgent. From the date and time of registration, we created 3 other variables. We classified seasons of visits as spring (February–April), summer (May– July), fall (August–October), and winter (November–January). The date of the visit was categorized into weekdays (Monday– Friday) and weekends (Saturday–Sunday). Public holidays in
which physician offices are typically closed were grouped with weekends. The time of the visit was used to classify visits into day shift (8:00–15:59), evening shift (16:00–23:59), and night shift (00:00–7:59). The length of stay in the ED was calculated as the time from registration to the time of disposition. Discharge diagnoses were categorized into 17 different groups according to the ICD-9-CM, and these groups were reclassified into the following 7 categories: diseases of the respiratory system (ICD-9-CM, 460–519); diseases of the digestive system (520–579); infectious and parasitic diseases (001–139); diseases of the nervous system and sense organs (320–389); diseases of the skin and subcutaneous tissue (680–709); symptoms, signs, and ill-defined conditions (780–799); and a miscellaneous category encompassing the remaining diseases.
Outcome Measures The main outcome of interest was an RV within 72 hours of an index visit. Time of visit was determined as the time of registration in the ED. An index visit was classified to with RV group if it
TABLE 1. Characteristics of the Study Patients Characteristic Patient Age, mean (SD), y Female Chronic disease ED visit LOS, mean (SD), h Weekend or holiday Time of day Day Evening Night Season Winter Spring Summer Fall Triage level Nonurgent Urgent Emergent Diagnostic category Respiratory system Digestive system Infectious diseases Nervous system Skin diseases Symptoms/signs/ill-defined Miscellaneous Hospital General hospital Physician Age, mean (SD), y Female Pediatrician
Total (n = 35,435)
With RV (n = 2291)
Without RV (n = 33,144)
P
5.1 (3.7) 15,999 (45.2) 1771 (5.0)
4.0 (3.0) 989 (43.2) 119 (5.2)
5.1 (3.7) 15,010 (45.3) 1652 (5.0)
Predicting Factors and Risk Stratification for Return Visits to the Emergency Department Within 72 Hours in Pediatric Patients Sheng-Feng Sung, MD,* Kang Ernest Liu, PhD,† Solomon Chih-Cheng Chen, MD, PhD,‡ Chia-Lun Lo, MS,§ Kuei-Chih Lin, MS,k and Ya-Han Hu, PhD§ Objectives: A return visit (RV) to the emergency department (ED) is usually used as a quality indicator for EDs. A thorough comprehension of factors affecting RVs is beneficial to enhancing the quality of emergency care. We performed this study to identify pediatric patients at high risk of RVs using readily available characteristics during an ED visit. Methods: We retrospectively collected data of pediatric patients visiting 6 branches of an urban hospital during 2007. Potential variables were analyzed using a multivariable logistic regression analysis to determine factors associated with RVs and a classification and regression tree technique to identify high-risk groups. Results: Of the 35,435 visits from which patients were discharged home, 2291 (6.47%) visits incurred an RV within 72 hours. On multivariable analysis, younger age, weekday visits, diagnoses belonging to the category of symptoms, signs, and ill-defined conditions, and being seen by a female physician were associated with a higher probability of RVs. Children younger than 6.5 years who visited on weekdays or between midnight and 8:00 AM on weekends or holidays had the highest probability of returning to the ED within 72 hours. Conclusions: Our study reexamined several important factors that could affect RVs of pediatric patients to the ED and identified high-risk groups of RVs. Further intervention studies or qualitative research could be targeted on these at-risk groups. Key Words: classification and regression tree, emergencies, health services, return visits (Pediatr Emer Care 2015;31: 819–824)
T
aiwan's National Health Insurance has been well acknowledged for its many merits, including low cost. Consequently, most of the emergency departments (EDs) in Taiwan are overloaded with patients.1 The large number of patients using the ED services makes the administrative management of emergency medicine increasingly important to assure a high quality of health care with respect to suitable cost controls. The frequency of return visits (RVs) to EDs is usually used as an indicator for quality of care2,3 and is currently included in the Taiwan Clinical Performance Indicators to measure quality improvement. High RV rates From the *Department of Internal Medicine, Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi City; †Department of Economics, National Chung Cheng University, Chiayi County; ‡Departments of Pediatrics and Medical Research, Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi City; §Department of Information Management and Institute of Healthcare Information Management, National Chung Cheng University, Chiayi County; and kInformation Technology Office, Taipei City Hospital, Taipei, Taiwan. Sheng-Feng Sung and Kang Ernest Liu contributed equally to this work. Disclosure: The authors declare no conflict of interest. Reprints: Ya-Han Hu, PhD, Department of Information Management and Institute of Healthcare Information Management, National Chung Cheng University, 168 University Rd, Min-Hsiung, Chiayi County 62102, Taiwan (e‐mail: [email protected]). Supported by National Science Council of the Republic of China under the Grants NSC 102-2410-H-194-104-MY2 and NSC 102-2410-H-194-087. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0749-5161
may be attributed to improper treatment at the initial visit, insufficient education at discharge, system failures, inadequate access to alternative primary care services, or even missed identification of subtle signs of disease processes.3,4 Some RVs may be avoidable or unnecessary and thus are costly both in financial terms and in the burden they place on the health care system. To improve the quality of emergency care, a thorough understanding of the underlying factors of RVs to the ED is necessary. Factors related to the illness, the patient, the physician, the health care system, or the ED visit may induce RVs.5–13 It is difficult to implement any specific intervention before we understand how these factors are related to RVs to the ED. Although factors affecting RVs within 72 hours in adults have been reported in Taiwan,14,15 the characteristics associated with RVs in children are not fully understood. Therefore, we conducted this study to examine the association between characteristics readily available during an ED visit and RVs within 72 hours in pediatric patients. We also sought to identify high-risk subgroups for RVs that may serve as a potential target for interventions.
METHODS Patients This retrospective cohort study was performed in the Taipei City Hospital, which successfully integrated 10 municipal hospitals into 1 unit in 2005 to provide comprehensive health care needs to 2.6 million residents in Taipei. Records of pediatric patients from 6 major branches, including Zhongxing, Renai, Heping, Yangming, Zhongxiao, and Fuyou, were used in this study. The Fuyou branch is specialized as a pediatric hospital, where pediatricians or pediatric emergency medicine physicians provide emergency care, whereas the remaining 5 branches are general hospitals, where only some patients are seen by adult emergency medicine physicians. During the study period, there were 8 medical centers and 12 regional hospitals in Taipei. All of the study hospitals were regional hospitals. The Taipei City Hospital Institutional Review Board approved this study. We reviewed the hospital's administrative database and identified all visits made by patients aged 18 years or younger to the EDs of the 6 branches between January 1 and December 31 in 2007. The patients who left without being seen, left against medical advice, had out-of-hospital cardiac arrest, were admitted to the hospital, or were transferred to another hospital for admission were excluded from the analysis. In addition, the records with unknown disposition and miscellaneous missing data, such as unknown time of registration or disposition, unrecorded triage level, and missing date of birth, were also dropped out. The remaining visits comprised the study sample (index visits). Potential variables affecting the rate of RVs were collected and were grouped into the following 4 categories: those related to the patient, those related to the hospital, those related to the
Pediatric Emergency Care • Volume 31, Number 12, December 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.pec-online.com
819
Pediatric Emergency Care • Volume 31, Number 12, December 2015
Sung et al
physician, and those related to the ED visit. Age, sex, and whether the patient has chronic diseases, for example, diabetes mellitus, hypertension, asthma, endocrine disorders, hypertensive cardiovascular diseases, and hypertensive renal diseases, were considered as potential patient-related factors. Whether the hospital is general or pediatric specialized was treated as a characteristic of the hospital. Variables related to the physician included age, sex, and whether the treating physician is a board-certified pediatrician. As for the characteristics of ED visits, we collected the level of triage, the time of registration in the ED and time of disposition from the ED, as well as the primary discharge diagnosis based on codes from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). Nursing triage levels were categorized as emergent, urgent, or nonurgent. From the date and time of registration, we created 3 other variables. We classified seasons of visits as spring (February–April), summer (May– July), fall (August–October), and winter (November–January). The date of the visit was categorized into weekdays (Monday– Friday) and weekends (Saturday–Sunday). Public holidays in
which physician offices are typically closed were grouped with weekends. The time of the visit was used to classify visits into day shift (8:00–15:59), evening shift (16:00–23:59), and night shift (00:00–7:59). The length of stay in the ED was calculated as the time from registration to the time of disposition. Discharge diagnoses were categorized into 17 different groups according to the ICD-9-CM, and these groups were reclassified into the following 7 categories: diseases of the respiratory system (ICD-9-CM, 460–519); diseases of the digestive system (520–579); infectious and parasitic diseases (001–139); diseases of the nervous system and sense organs (320–389); diseases of the skin and subcutaneous tissue (680–709); symptoms, signs, and ill-defined conditions (780–799); and a miscellaneous category encompassing the remaining diseases.
Outcome Measures The main outcome of interest was an RV within 72 hours of an index visit. Time of visit was determined as the time of registration in the ED. An index visit was classified to with RV group if it
TABLE 1. Characteristics of the Study Patients Characteristic Patient Age, mean (SD), y Female Chronic disease ED visit LOS, mean (SD), h Weekend or holiday Time of day Day Evening Night Season Winter Spring Summer Fall Triage level Nonurgent Urgent Emergent Diagnostic category Respiratory system Digestive system Infectious diseases Nervous system Skin diseases Symptoms/signs/ill-defined Miscellaneous Hospital General hospital Physician Age, mean (SD), y Female Pediatrician
Total (n = 35,435)
With RV (n = 2291)
Without RV (n = 33,144)
P
5.1 (3.7) 15,999 (45.2) 1771 (5.0)
4.0 (3.0) 989 (43.2) 119 (5.2)
5.1 (3.7) 15,010 (45.3) 1652 (5.0)
