Abstract
Background
Children may benefit greatly in terms of safety and care coordination from the information sharing promised by health information exchange (HIE). While information exchange capability is a required feature of the certified electronic health record, we known little regarding how this technology is used in general and for pediatric patients specifically.
Methods
Using data from an operational HIE effort in central Texas, we examined the factors associated with actual system usage. The clinical and demographic characteristics of pediatric ED encounters (n = 179,445) were linked to the HIE system user logs. Based on the patterns of HIE system screens accessed by users, we classified each encounter as: no system usage, basic system usage, or novel system usage. Using crossed random effects logistic regression, we modeled the factors associated with basic and novel system usage.
Results
Users accessed the system for 8.7% of encounters. Increasing patient comorbidity was associated with a 5% higher odds of basic usage and 15% higher odds for novel usage. The odds of basic system usage were lower in the face of time constraints and for patients who had not been to that location in the previous 12 months.
Conclusions
HIE systems may be a source to fulfill users' information needs about complex patients. However, time constraints may be a barrier to usage. In addition, results suggest HIE is more likely to be useful to pediatric patients visiting ED repeatedly. This study helps fill an existing gap in the study of technological applications in the care of children and improves knowledge about how HIE systems are utilized.
Background
Health information exchange (HIE), the process of electronically sharing identified, patient-level information between different organizations,[1] is a potentially transformative solution to problems of cost,[2] timeliness,[3] patient-centeredness,[4] safety,[3] and efficiency [5] that plague the healthcare system. Furthermore, children and adolescents may especially benefit from broad and easy information sharing. First, HIE has the ability to better support the care and detection of vaccine preventable conditions by incorporating immunization histories and linking to both local public health agencies and schools [4,6,7]. Second, minors constitute a substantial proportion of emergency department (ED) visits in the US,[8] with infants having the highest rates of ED visits [9]. The care delivered in the ED setting may benefit the most from improved information sharing [10,11]. Additionally, medication errors can be particularly dangerous for children; HIE improves communication and may prevent such mistakes [12,13]. Lastly, because HIE improves coordination among providers,[14] these information system can support providers in their provision of a medical home for all patients in general and children with special healthcare needs in particular.
Current federal policy dramatically advances the prospect for widespread HIE. The Health Information Technology for Economic & Clinical Health (HITECH) Act, part of the American Recovery & Reinvestment Act, identified information exchange capability and connectivity as a required feature of certified electronic health records (EHRs). To eligible for any EHR incentive payments, providers must now demonstrate Meaningful Use, which includes testing of HIE capabilities [15]. Despite the high level of support for HIE, we know very little about providers' motivations to use HIE systems or the effectiveness of HIE,[16-18] beyond the fact that these information systems are predominately accessed by a minority of providers [19] and for a minority of patients [20].
This paper aims to address the knowledge gap in HIE utilization regarding treatment of children. Previous researchers have argued for examinations of information technology for children separately from adult populations due to the particular vulnerabilities and unique needs of children [21,22]. In this examination, we address the question, what factors indicative of an information need or value of using HIE are associated with HIE usage? Both patient and encounter characteristics can change healthcare professionals' need for additional information. Factors such as patient complexity [23] or recent utilization [24,25] increase the uncertainty associated with delivering care and could prompt use of an HIE system [26]. Conversely, some encounters have little to do with the patient's previous utilization or are relatively uniformed by information created in other organizations. Likewise, the value of seeking potentially useful information may be lessened by other factors such as time constraints [27-29]. This study examines the factors associated with actual HIE usage during children's ED encounters.
Methods
The Integrated Care Collaborative (ICC) of Central Texas is a fully functional HIE facilitating effort established in 1997 encompassing Austin, Texas area safety-net providers. The ICC exists as a separate nonprofit entity with 24 member organizations including: hospital systems, clinics, and governmental agencies. The study sample includes all ED encounters among patients less than 18 years old between 1/1/2006 and 6/30/2009 included in the ICC's master patient index/clinical data repository, I-Care. I-Care is a centralized database containing electronic patient level demographic and clinical information. ICC member organizations contribute patient level electronic data to I-Care on medically indigent patients. In turn, authorized users at each location may access data from I-Care through a secured website. Authorized users vary by location, but can include physicians, nurses, and/or administrative staff. Parents or guardians provide consent for minors to be included in the information exchange and this study only included consenting patients. We also excluded emergency encounters occurring at facilities before the hospital employed an authorized user of the I-Care system. The final dataset included 179,445 encounters from 11 emergency departments.
We derived the dependent variable representing type of usage from the I-Care system log files. Log files provide an objective and recommended [17,30] measure of system usage unbiased by subject recall [31]. The I-Care interface is an EpicWeb proprietary software system where authorized users navigate through several different web pages or screens containing demographics, prior utilization history, contact information, payer history, medication orders, prior diagnoses and other information. As part of the Health Insurance Portability and Accountability Act compliance, I-Care generates electronic log files in order to document the user's activities including: patient viewed, date accessed, time accessed, and screen(s) viewed. Through the logged date and time, we could follow the sequence of screens viewed by each user for a given patient on a given date. The entire sample included 77 different patterns of screen views in the associated log file. A single pattern accounted for 82% of sessions; this pattern consisted of an end user identifying a patient on a selection screen and then viewing a single screen containing a summary of recent encounters. We classified this type of session as basic usage. All other session patterns were classified as novel usage. A novel usage session consisted of any user session that included additional screen views (such as medications, a demographic summary, or detailed encounter records) beyond the initial patient selection screen and summary of recent encounters screen accessed in a basic usage scenario. A patient encounter in an ED could result in three usage outcomes: 1) no usage, 2) basic usage, and 3) novel usage. We linked user sessions to encounters based on patient identifier, date, user's work location, and place of encounter. Because ED encounters can occur late at night, we allowed for linkages up to 3 AM the next day.
We considered three factors as indicative of uncertainty that creates an information need: comorbidity, prior utilization, and unfamiliarity with the patient. The number of unique disease categories for each encounter measured comorbidity. Disease categories were defined by the Agency for Healthcare Research & Quality's (AHRQ) Clinical Classifications Software applied to all reported ICD-9 diagnosis codes [32]. For prior utilization, we determined the total number of ED encounters, inpatient hospitalizations, and primary care clinic visits at ICC member facilities in the 12 months prior to the encounter date. We did not include previous visits to the same ED in these counts. Following existing definitions of encounter frequency, we divided ED and primary care visits into 0 encounters, infrequent users (1 to 3), and frequent users (4 or more) [33-35]. Due to small cell counts, we could only consider hospitalization in the previous 12 months in a binary fashion. Finally, patients unfamiliar to a specific ED were marked by the absence of any encounters at that same ED in the previous 12 months. Because we excluded visits at the same facility from the measure of past ED encounters, we avoided collinearity for this measure of patient unfamiliarity.
To measure potential time constraints, we created a binary variable to classify the encounter date at that ED as busy or not busy. For each ED, we divided the total number of encounters on a date by the ED's previous year's average number of daily encounters for that same day of the week and month. A busier than average day existed when this ratio was greater than one.
We categorized the primary diagnosis and payer to help describe the sample. First, AHRQ's Chronic Condition Indicator and Body Systems definitions categorized the primary diagnosis as a chronic condition and assigned the primary diagnosis into 18 indicators roughly analogous to major diagnostic categories [36]. We selected factors influencing health status and all c
บทคัดย่อพื้นหลังเด็กอาจได้รับประโยชน์อย่างมากในแง่ของความปลอดภัย และดูแลการประสานงานจากข้อมูลใช้ร่วมกันตามสัญญา โดยการแลกเปลี่ยนข้อมูลสุขภาพ (HIE) ในขณะที่ความสามารถในการแลกเปลี่ยนข้อมูลเป็นคุณลักษณะที่ต้องการระเบียนสุขภาพรับรองอิเล็กทรอนิกส์ เรารู้จักน้อยเกี่ยวกับว่าเทคโนโลยีนี้จะใช้ทั่วไป และผู้ป่วยเด็กโดยเฉพาะวิธีการเราใช้ข้อมูลจากความพยายาม HIE ปฏิบัติในเท็กซัสกลาง ตรวจสอบปัจจัยที่เกี่ยวข้องกับการใช้งานระบบจริง พบลักษณะทางคลินิก และประชากรของเด็กนักเรียน (n = 179,445) การเชื่อมโยงการบันทึกผู้ใช้ระบบ HIE ขึ้นอยู่กับรูปแบบของหน้าจอระบบ HIE เข้าถึงได้ โดยผู้ใช้ เราจัดพบละเป็น: ไม่การใช้งานระบบ ระบบพื้นฐานการใช้งาน หรือการใช้งานระบบนวนิยาย ใช้ข้ามสุ่มผลการถดถอยโลจิสติก เราสร้างแบบจำลองปัจจัยที่สัมพันธ์กับการใช้งานระบบพื้นฐาน และนวนิยายผลลัพธ์ผู้ใช้เข้าถึงระบบ 8.7% ของผลงาน เพิ่มผู้ป่วย comorbidity ที่สัมพันธ์กับราคาสูงกว่า 5% ของพื้นฐานการใช้งานและ 15% สูงกว่าราคาใช้นวนิยาย ราคาของการใช้งานระบบพื้นฐานล่าง หน้าข้อจำกัดเวลา และผู้ป่วยที่ไม่ได้อยู่ที่ใน 12 เดือน ได้บทสรุประบบ HIE อาจมาเพื่อตอบสนองความต้องการข้อมูลของผู้ใช้เกี่ยวกับผู้ป่วยที่ซับซ้อน อย่างไรก็ตาม ข้อจำกัดของเวลาอาจจะเป็นอุปสรรคต่อการใช้งาน นอกจากนี้ ผลแนะนำ HIE มีแนวโน้มที่จะเป็นประโยชน์กับผู้ป่วยเด็กที่เยี่ยมชม ED ซ้ำ ๆ การศึกษานี้ช่วยเติมช่องว่างอยู่ในการศึกษาการใช้เทคโนโลยีในการดูแลเด็ก และช่วยเพิ่มความรู้เกี่ยวกับวิธีใช้ระบบ HIEพื้นหลังระบุกระบวนการทางอิเล็กทรอนิกส์ร่วมกันแลกเปลี่ยนข้อมูลในสุขภาพ (HIE), ข้อมูลผู้ป่วยระดับระหว่างองค์กรต่าง ๆ, [1] เป็นการเปลี่ยนแปลงอาจแก้ไขปัญหาต้นทุน เที่ยงตรง [2], [3] ผู้ป่วย centeredness [4] ความ ปลอดภัย, [3] และมีประสิทธิภาพ [5] ที่เกิดภัยพิบัติระบบสุขภาพ นอกจากนี้ เด็กและวัยรุ่นอาจโดยเฉพาะอย่างยิ่งได้รับประโยชน์จากข้อมูลกว้าง และง่าย ก่อน HIE มีความสามารถในการสนับสนุนการดูแลและตรวจสอบเงื่อนไข preventable วัคซีน โดยการรับวัคซีนหากเพจ และการเชื่อมโยงกับหน่วยงานสาธารณสุขท้องถิ่นและโรงเรียน [4,6,7] สอง เยาวเป็นสัดส่วนพบของแผนกฉุกเฉิน (ED) ในสหรัฐอเมริกา, [8] มีทารกที่มีราคาสูงสุดของเอ็ด [9] ดูแลจัดส่งนักเรียนตั้งอาจได้รับประโยชน์สูงสุดจากการปรับปรุงข้อมูลร่วมกัน [10,11] นอกจากนี้ ยาข้อผิดพลาดสามารถเป็นอันตรายอย่างยิ่งสำหรับเด็ก HIE ช่วยสื่อสาร และอาจป้องกันความผิดพลาดดังกล่าว [12,13] สุดท้ายนี้ เนื่องจาก HIE ช่วยประสานงานระหว่างผู้ให้บริการ, [14] ข้อมูลที่ระบบสนับสนุนผู้ให้บริการในการจัดบ้านทางการแพทย์สำหรับผู้ป่วยทั้งหมดโดยทั่วไปและเด็กดูแลสุขภาพพิเศษเหล่านี้ ต้องการเฉพาะนโยบายปัจจุบันรัฐบาลกลางล่วงหน้าโอกาสสำหรับ HIE แพร่หลายอย่างมาก เทคโนโลยีสารสนเทศสุขภาพเศรษฐกิจและพระราชบัญญัติสุขภาพคลินิก (ไฮเทค) ส่วนหนึ่งของการกู้คืนอเมริกันและ Reinvestment บัญญัติ ระบุความสามารถในการแลกเปลี่ยนข้อมูลและการเชื่อมต่อเป็นคุณลักษณะที่ต้องการของสุขภาพผ่านการรับรองอิเล็กทรอนิกส์ (EHRs) การมีสิทธิสำหรับการชำระเงินจูงใจใด ๆ EHR ผู้ให้บริการต้องตอนนี้เห็นถึงใช้ความหมาย ซึ่งรวมถึงการทดสอบความสามารถ HIE [15] แม้ มีคุณภาพสนับสนุน HIE เรารู้น้อยมากเกี่ยวกับโต่งของผู้ให้บริการจะใช้ระบบ HIE หรือประสิทธิภาพของ HIE, [16-18] นอกเหนือจากข้อเท็จจริงระบบข้อมูลเหล่านี้เข้าถึงชนกลุ่มน้อย และชนกลุ่มน้อยของผู้ป่วย [20] [19] ผู้ให้บริการ predominatelyThis paper aims to address the knowledge gap in HIE utilization regarding treatment of children. Previous researchers have argued for examinations of information technology for children separately from adult populations due to the particular vulnerabilities and unique needs of children [21,22]. In this examination, we address the question, what factors indicative of an information need or value of using HIE are associated with HIE usage? Both patient and encounter characteristics can change healthcare professionals' need for additional information. Factors such as patient complexity [23] or recent utilization [24,25] increase the uncertainty associated with delivering care and could prompt use of an HIE system [26]. Conversely, some encounters have little to do with the patient's previous utilization or are relatively uniformed by information created in other organizations. Likewise, the value of seeking potentially useful information may be lessened by other factors such as time constraints [27-29]. This study examines the factors associated with actual HIE usage during children's ED encounters.MethodsThe Integrated Care Collaborative (ICC) of Central Texas is a fully functional HIE facilitating effort established in 1997 encompassing Austin, Texas area safety-net providers. The ICC exists as a separate nonprofit entity with 24 member organizations including: hospital systems, clinics, and governmental agencies. The study sample includes all ED encounters among patients less than 18 years old between 1/1/2006 and 6/30/2009 included in the ICC's master patient index/clinical data repository, I-Care. I-Care is a centralized database containing electronic patient level demographic and clinical information. ICC member organizations contribute patient level electronic data to I-Care on medically indigent patients. In turn, authorized users at each location may access data from I-Care through a secured website. Authorized users vary by location, but can include physicians, nurses, and/or administrative staff. Parents or guardians provide consent for minors to be included in the information exchange and this study only included consenting patients. We also excluded emergency encounters occurring at facilities before the hospital employed an authorized user of the I-Care system. The final dataset included 179,445 encounters from 11 emergency departments.We derived the dependent variable representing type of usage from the I-Care system log files. Log files provide an objective and recommended [17,30] measure of system usage unbiased by subject recall [31]. The I-Care interface is an EpicWeb proprietary software system where authorized users navigate through several different web pages or screens containing demographics, prior utilization history, contact information, payer history, medication orders, prior diagnoses and other information. As part of the Health Insurance Portability and Accountability Act compliance, I-Care generates electronic log files in order to document the user's activities including: patient viewed, date accessed, time accessed, and screen(s) viewed. Through the logged date and time, we could follow the sequence of screens viewed by each user for a given patient on a given date. The entire sample included 77 different patterns of screen views in the associated log file. A single pattern accounted for 82% of sessions; this pattern consisted of an end user identifying a patient on a selection screen and then viewing a single screen containing a summary of recent encounters. We classified this type of session as basic usage. All other session patterns were classified as novel usage. A novel usage session consisted of any user session that included additional screen views (such as medications, a demographic summary, or detailed encounter records) beyond the initial patient selection screen and summary of recent encounters screen accessed in a basic usage scenario. A patient encounter in an ED could result in three usage outcomes: 1) no usage, 2) basic usage, and 3) novel usage. We linked user sessions to encounters based on patient identifier, date, user's work location, and place of encounter. Because ED encounters can occur late at night, we allowed for linkages up to 3 AM the next day.We considered three factors as indicative of uncertainty that creates an information need: comorbidity, prior utilization, and unfamiliarity with the patient. The number of unique disease categories for each encounter measured comorbidity. Disease categories were defined by the Agency for Healthcare Research & Quality's (AHRQ) Clinical Classifications Software applied to all reported ICD-9 diagnosis codes [32]. For prior utilization, we determined the total number of ED encounters, inpatient hospitalizations, and primary care clinic visits at ICC member facilities in the 12 months prior to the encounter date. We did not include previous visits to the same ED in these counts. Following existing definitions of encounter frequency, we divided ED and primary care visits into 0 encounters, infrequent users (1 to 3), and frequent users (4 or more) [33-35]. Due to small cell counts, we could only consider hospitalization in the previous 12 months in a binary fashion. Finally, patients unfamiliar to a specific ED were marked by the absence of any encounters at that same ED in the previous 12 months. Because we excluded visits at the same facility from the measure of past ED encounters, we avoided collinearity for this measure of patient unfamiliarity.To measure potential time constraints, we created a binary variable to classify the encounter date at that ED as busy or not busy. For each ED, we divided the total number of encounters on a date by the ED's previous year's average number of daily encounters for that same day of the week and month. A busier than average day existed when this ratio was greater than one.We categorized the primary diagnosis and payer to help describe the sample. First, AHRQ's Chronic Condition Indicator and Body Systems definitions categorized the primary diagnosis as a chronic condition and assigned the primary diagnosis into 18 indicators roughly analogous to major diagnostic categories [36]. We selected factors influencing health status and all c
การแปล กรุณารอสักครู่..
