Letters RESEARCH LETTER the Agency of Healthcare Research and Quality–sponsored Healthcare Cost and Utilization Project State Inpatient Data- Association Between Elements of Electronic base. Hospital health information technology adoption data Health Record Systems and the Weekend Effect were assessed using the Healthcare Information and Manage- in Urgent General Surgery ment Systems Society Analytics Database provided through the Temporal disparities in care are increasingly being recog- Dorenfest Institute for Health Information Technology, Re- nized as important determinants of health outcomes. These search, and Education for 2011. Hospital data were obtained disparities are characterized by differences in outcomes based from the American Hospital Association Annual Survey Data- on the time when care is delivered and include several well- base. Included were all patients admitted on a weekend who studied phenomena including the “weekend effect.”1,2 underwent 1 of 3 representative urgent or emergent general sur- Like many disparities in health care, there is potential for gery procedures (acute cholecystectomy, acute appendec- temporal disparities to be overcome.3,4 Previous study has tomy, or acute hernia repair). shown that fully implemented electronic health record (EHR) The primary outcome was the weekend effect, defined as systems can help hospitals overcome the weekend effect for a greater observed than expected hospital length of stay at patients undergoing urgent general surgical procedures.5 We the patient level. Expected length of stay was calculated evaluated how specific components of EHR systems were as- using a mixed-effects linear regression model fit on data from sociated with the weekend effect. patients with weekday admissions who otherwise met the inclusion criteria. Fixed effects included age, sex, race/ Methods | This was a retrospective, cross-sectional review of ethnicity, Charlson Comorbidity Index score, and procedure data from Florida in 2011. Patient-level data were obtained from type. A random effect was included for the hospital. Patients Figure. Unadjusted and Adjusteda Odds of Overcoming Weekend Effect (WE) by Elements of Electronic Health Record (EHR) Systems Odds Ratio Decrease Increase Element (95% CI) Odds of WE Odds of WE EHR status Unadjusted 0.86 (0.50-1.15) Adjusted 0.75 (0.34-1.65) EHR connectivity Unadjusted 0.80 (0.68-0.94) Adjusted 0.84 (0.66 - 1.07) EHR in OR Unadjusted 0.76 (0.60-0.97) Adjusted 0.81 (0.56-1.16) Electronic OR scheduling Unadjusted 0.75 (0.58-0.97) CDSS indicates clinical decision Adjusted 0.67 (0.46-0.97) support system; CPOE, computerized Electronic medication reconciliation physician order entry; and Unadjusted 0.91 (0.78-1.06) OR, operating room. Adjusted 0.86 (0.67-1.10) a Adjusted odds based on Electronic medication administration mixed-effects logistic regression; Unadjusted 0.86 (0.70-1.06) fixed effects include sex, primary Adjusted 0.83 (0.58-1.18) payer, deficiency anemia, CPOE congestive heart failure, Unadjusted 0.85 (0.73-0.99) coagulopathy, electrolyte disorder, Adjusted 0.90 (0.70-1.14) neurologic disorder, weight loss, CDSS hospital characteristics (Commission on Cancer approval, Unadjusted 0.90 (0.61-1.35) teaching hospital, and medical Adjusted 0.77 (0.44-1.36) school affiliation), hospital size Bed management system (total hospital beds, adjusted Unadjusted 0.69 (0.57-0.82) average daily census, and inpatient Adjusted 0.65 (0.50-0.84) surgical operations), and hospital resources (total Medicare 0 0.5 1.0 1.5 2.0 discharges and full-time registered Odds Ratio (95% CI) nurses) and the random effect was hospital. 602 JAMA Surgery June 2017 Volume 152, Number 6 (Reprinted) jamasurgery.com © 2017 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 Letters were considered to have the weekend effect if they were Corresponding Author: Paul C. Kuo, MD, MS, MBA, Department of Surgery, admitted on the weekend and their observed length of stay Loyola University Medical Center, 2160 S First Ave, Maywood, IL 60153 ([email protected]). was greater than the upper bound of the 95% CI of the Published Online: March 29, 2017. doi:10.1001/jamasurg.2017.0015 model-derived expected length of stay. Multilevel models Author Contributions: Dr Kothari had full access to all of the data in the study including hospital covariates, such as structural characteris- and takes responsibility for the integrity of the data and the accuracy of the data tics, hospital size, and hospital resources, were used to deter- analysis. mine associations between EHR elements and the weekend Concept and design: Kothari, Zapf, Gupta, Kuo. effect. Additional hospital-level confounders including Acquisition, analysis, or interpretation of data: Kothari, Brownlee, Blackwell, Markossian, Kuo. weekend-specific staffing, resource allocation, surgeon staff- Drafting of the manuscript: Kothari, Kuo. ing, and models of surgical care (ie, acute care surgery pro- Critical revision of the manuscript for important intellectual content: All authors. grams) could not be assessed owing to limitations of the data Statistical analysis: Kothari, Blackwell. set. This study was deemed exempt from institutional review Obtained funding: Kuo. Administrative, technical, or material support: Kothari, Zapf, Gupta, Kuo. board approval by the Loyola University Medical Center insti- Supervision: Gupta, Kuo. tutional review board based on the use of publicly available, Conflict of Interest Disclosures: None reported. deidentified patient data. All statistical analyses were per- Funding/Support: This study was supported by National Institutes of Health formed using Stata version 13 (StataCorp LP). grant T32 GM08750-16. Role of the Funder/Sponsor: The funder had a role in the design and conduct Results | There were 2979 patients who met the criteria for in- of the study but not in the collection, management, analysis, and interpretation clusion. Of these, 2033 did not have the weekend effect and of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. 946 had the weekend effect. Patients without the weekend ef- 1. Gallego B, Magrabi F, Concha OP, Wang Y, Coiera E. Insights into temporal fect were more likely to be exposed to high-speed EHR con- patterns of hospital patient safety from routinely collected electronic data. nectivity (62.6% vs 57.3%, P = .006), EHR in the operating room Health Inf Sci Syst. 2015;3(Suppl 1 HISA Big Data in Biomedicine and Healthcare (90.0% vs 87.3%, P = .03), electronic operating room sched- 2013 Con):S2. uling (91.6% vs 89.1%, P = .03), computerized physician or- 2. Becker DJ. Do hospitals provide lower quality care on weekends? Health Serv der entry (45.0% vs 41.0%, P = .04), and electronic bed man- Res. 2007;42(4):1589-1612. agement systems (79.9% vs 73.2%, P < .001). 3. Blecker S, Goldfeld K, Park H, et al. Impact of an intervention to improve weekend hospital care at an academic medical center: an observational study. Specific components of EHR systems decreased the odds J Gen Intern Med. 2015;30(11):1657-1664. of having the weekend effect. These included electronic op- 4. Selker HP, Beshansky JR, Pauker SG, Kassirer JP. The epidemiology of delays erating room scheduling (adjusted odds, 0.67; 95% CI, 0.46- in a teaching hospital: the development and use of a tool that detects 0.97; P = .03) and electronic bed management systems (ad- unnecessary hospital days. Med Care. 1989;27(2):112-129. justed odds, 0.65; 95% CI, 0.50-0.84; P = .001) (Figure). 5. Kothari AN, Zapf MA, Blackwell RH, et al. Components of hospital perioperative infrastructure can overcome the weekend effect in urgent general surgery procedures. Ann Surg. 2015;262(4):683-691. Discussion | Implementation of EHR systems is one mecha- nism to help hospitals combat an important temporal dis- parity of care, the weekend effect, for patients undergoing US Emergency Department Encounters for Law urgent general surgical procedures. Specific components of Enforcement–Associated Injury, 2006-2012 EHR systems, including electronic operating room schedul- Deaths of civilians in contact with police have recently gained ing and electronic bed management systems, are most national public and policy attention. While journalists track strongly associated with decreasing the odds of the week- police-involved deaths,1 epidemiologic data are incomplete,2,3 end effect. Together, these data illustrate an important rela- and trends in nonfatal injuries, which far outnumber deaths, are tionship that may exist between EHR systems and temporal poorly understood. The International Classification of Diseases, care disparities in nonelective surgical settings. Ninth Revision, Clinical Modification, includes external cause-of- injury codes identifying injuries owing to contact with law en- forcement (E970-E978). Using these codes, prior studies have Anai N. Kothari, MD identified 715 118 nonfatal injuries, 3958 hospitalizations, and Sarah A. Brownlee, BA 3156 deaths between 2003 and 2011 from US Centers for Disease Robert H. Blackwell, MD Control and Prevention data and the Nationwide Inpatient Matthew A. C. Zapf, BA Sample,4 and 55 400 fatal and nonfatal injuries in 2012 from the Talar Markossian, PhD Vital Statistics mortality census, Nationwide Inpatient and Emer- Gopal N. Gupta, MD gency Department Samples, and journalists’ reports.5 In this Paul C. Kuo, MD, MS, MBA study,we used a nationally representative database to determine