MMHA6700 IT Transforming Health Care Paper

JONA Volume 42, Number 5, pp 256-265 Copyright B 2012 Wolters Kluwer Health | Lippincott Williams & Wilkins THE JOURNAL OF NURSING ADMINISTRATION Destination Bedside Using Research Findings to Visualize Optimal Unit Layouts and Health Information Technology in Support of Bedside Care Nicholas Watkins, PhD Mary Kennedy, MS, RN Nelson Lee Michael O’Neill, PhD Erin Peavey, MArch Maria DuCharme, MS, RN, ME-BC Cynthia Padula, PhD, RN This study explored the impact of unit design and healthcare information technology (HIT) on nursing workflow and patient-centered care (PCC). Healthcare information technology and unit layoutYrelated predictors of nursing workflow and PCC were measured during a 3-phase study involving questionnaires and work sampling methods. Stepwise multiple linear regressions demonstrated several HIT and unit layoutY related factors that impact nursing workflow and PCC. that should guide healthcare reform. A sixth and ultimate goal, patient-centered care (PCC), is at the heart of the other aims. Patient-centered care has been characterized as ‘‘provision of care that is respectful and responsive to patient preferences and needs, ensuring that patient values guide clinical decisions.’’2 Patient-centered care is being pursued in a variety of ways through advancements in healthcare information technology (HIT), electronic health records (EHRs), and inpatient unit layouts. Emphasizing the importance of such interventions, the American Recovery and Reinvestment Act of 2009 allocated approximately $27 billion to facilitate the adoption of the EHR and related components using a meaningful-use approach.3 The literature indicates that HIT and related technologies can help prevent medical errors.4 The overall aim of the present study was to explore relationships between HIT and unit layout and their impact on nursing workflow and PCC. Decisions on where to place HIT solutions on an inpatient unit and in relation to the patient room may have implications for nursing workflow and PCC. Unfortunately, there is a scarcity of empirical research that demonstrates optimal relationships among HIT, EHR, and unit layouts. Research on the relationships among these is timely, given the growing prevalence of units with private patient rooms and with decentralized nursing station options.5 For over a decade, the Institute of Medicine has been an advocate for change in the healthcare industry.1,2 It has indicated that timeliness, efficiency, equitability, safety, and effectiveness of care are quality aims Author Affiliations: Director of Research (Dr Watkins), Research Specialist (Ms Peavey), Hellmuth, Obata, and Kassabaum, New York; Independent Consultant (Ms Kennedy), Barrington, Rhode Island; Chief Executive Officer (Mr Lee), Rapidmodeling, Cincinnati, Ohio; Senior Director Workplace Research (Dr O’Neill), Knoll, Inc, New York; Chief Nursing Officer (Ms DuCharme), Research Specialist (Dr Padula), The Miriam Hospital, Providence, Rhode Island. This research was funded through in-kind matching donations and collaboration among the Miriam Hospital; Hellmuth, Obata, and Kassabaum; and faculty from the Rhode Island School of Nursing. The views expressed in this article do not necessarily reflect the position or policy of Miriam Hospital, HOK, or Rhode Island School of Nursing. The authors declare no conflict of interest. Correspondence: Dr Watkins, Hellmuth, Obata, and Kassabaum (HOK), 1065 Ave of the Americas, 6th floor, New York, NY 10018 (nick.watkins@hok.com). Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.nursingresearchonline.com). DOI: 10.1097/NNA.0b013e3182480918 256 Unit Layout and Clinical Information Systems Healthcare information technology and unit layout are dormant or latent conditions that may indirectly JONA Vol. 42, No. 5 May 2012 Copyright © 2012 Lippincott Williams & Wilkins. MMHA6700 IT Transforming Health Care Paper

 

Unauthorized reproduction of this article is prohibited. affect nursing workflow and care quality issues related to adverse events.6,7 Existing research on latent conditions indicates that facility design and equipment and supply failures may be leading contributors to poor patient safety and care delivery inefficiencies.8 Given these circumstances, a deeper understanding of the challenges and opportunities associated with and between latent conditions such as unit layout and HIT is warranted. Although the parts of an inpatient unit remain fairly constant, the ideal configuration of patient rooms, nursing stations, corridors, and support core spaces remains a source of debate. A descriptive study including 81 medical-surgical inpatient units concluded that a unit consisting of spokes of patient rooms was an inferior configuration due to low patient visibility and long travel distances.9 Research comparing radial to rectangular units found that nurses on radial units walk less and spend more time performing patient care activities.10,11 A retrospective study of an intensive care unit found severely ill patients admitted to rooms not visible from the main nursing station experienced statistically significant higher mortality rates than those patients admitted to rooms in view.12 Decentralized nursing stations are immediately inside or outside a patient’s room, often with windows for direct patient observation. Accommodations may include medication and supply storage, a hand-washing facility, work surfaces for charting, a computer, and telecommunication devices.13 Decentralized nursing stations may contribute to decreases in patient falls.14 In hospitals, adverse drug events are frequent and a common result of medication errors.2 Fortythree percent of medication errors may be due to workplace distractions during preparation and dispensing.15 During medication preparation, major sources of interruptions include other nurses and searches for missing medications and equipment.16,17 One highly controlled study found that patient rooms with locked medication cabinets had a statistically significant lower medication error rate compared with medication carts.18 A corroborating study indicated improved efficiencies and fewer interruptions after implementing decentralized medication cabinets at the bedside.19 Multiple variables can influence the efficiency and accuracy of information entered into an HIT application. For example, exploratory research of mobile computer technologies or computers-on-wheels versus stationary computers indicated that nurses document more at a computer-on-wheels than a stationary computer during the first hour of data collection.20 However, the computers-on-wheels were noted by nurses as being clumsy, difficult to push, and inoperable at key locations. Other research has found that bedside computer terminals are associated with a 24.5% decrease in overall nurse documentation time per shift.21 Advancements in HIT and unit layout inspire questions on how best to orchestrate HIT with unit layout, nursing workflow, and PCC. There is very little literature specific to relationships between HIT and unit layouts. HIT is advancing at a rate exceeding hospital administrators’ ability to adopt and implement it. Furthermore, there is scarcity of research that demonstrates what unit layouts and system configurations are optimal for PCC. Methods Research Approach and Phases The present study was conducted in 3 phases using mixed methods. During mixed methods research, qualitative research involving interviews and focus group exercises can complement quantitative research involving questionnaires and behavior observation.22 One method can compensate for the limitations of the other. The 1st phase of the study used questionnaires to detect relationships between nursing workflow and patient experience. The RNs’ scores on a questionnaire were correlated with their patients’ scores on a different questionnaire. Using this approach, the researchers could examine the associations among the unit layout, HIT use, care delivery, nursing workflow, and patient outcomes. Aspects of unit layout, HIT use, care delivery, and nursing workflow that had negative consequences for patient outcomes were not considered patient-centered. Patients who participated in phase 1 of the study were assigned to rooms based on room availability. Therefore, differences among patients were controlled by a natural instance of random assignment. The 2nd phase was a work sampling investigation during which the RNs’ walking distances, space utilization, HIT use, and frequency of patient care at bedside were monitored. MMHA6700 IT Transforming Health Care Paper

 

This phase identified whether specific features (ie, unit layout, HIT, and care delivery characteristics) contributed to the frequency and quality of patient care at bedside. In the 3rd phase of the study, nurses participated in design charrettes where they discussed results from phases 1 and 2. A charrette is a common brainstorming and ideation technique used in architecture to generate visual solutions to a design challenge.23 This phase enabled the RNs to create sketches of unit and patient room layout with HIT solutions. JONA Vol. 42, No. 5 May 2012 Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 257 Setting The study site was an academically affiliated teaching hospital located in an urban area of the United States. The hospital has 247 adult inpatient beds with 7 medical-surgical inpatient units, all included in the study. All units had computers-on-wheels and balanced headwalls with identical locations for gasses, outlets, and call buttons on both sides of the patient beds (see Table, Supplemental Digital Content 1; http://links.lww.com/JONA/A68) as well as computerized physician order entry and bar-code medication administration. Sample Phase 1 consisted of 109 patients and 89 RNs from the day shift. On average, a participating patient was 62 years of age. On average, a nurse in phase 1 was 35 (SD, 11.19) years old, had 7 (SD, 8.47) years of experience, and cared for 5 (SD, 0.82) patients. On average, a patient in phase 1 was 63 (SD, 15.52) years old. In total, there were 45 pairs of staff and patients. Phase 2 consisted of 111 day- and eveningshift nurses. Of these nurses, 29 used a personal digital assistant (PDA) device and pedometer and completed the questionnaire; 48 only used a PDA, and 34 only filled out a questionnaire. On average, a nurse in phase 2 was 33 (SD, 8.48) years old, had 5.5 (SD, 7.06) years of experience, and cared for 5 (SD, 0.86) patients. Phase 3 consisted of 40 nurses who took part in the charrettes. Measures Three self-designed questionnaires were developed for the study. There is a dearth of valid and reliable instruments in the literature specifically assessing the effect of design features on patients and clinicians. A comprehensive literature review of empirical research of inpatient unit settings was conducted to identify the items for inclusion in the questionnaires. Two rounds of focus groups and work sessions with nursing administration and staff from the participating hospitals were held to refine the questionnaires. To establish content validity, the 1st round involved content experts consisting of 4 nurses, 2 healthcare design architects, and 1 environmental psychologist specializing in research of healthcare design. The 2nd round involved the content experts and additional nursing staff including unit managers. Nursing staff and managers piloted the questionnaires to determine appropriate length and provide additional feedback. Suggestions and recommendations from the content experts and staff were incorporated into the questionnaires. Questions were reversed and scored to prevent response bias by participants. 258 The 1st questionnaire developed was the Patient and Staff Experience Questionnaire: RN Portion (PSEQ-RNP). The initial version of the PSEQ-RNP consisted of 117 questions. The final version of the PSEQ-RNP consisted of 94 questions. Various subsections addressed characteristics of nurses and their care delivery activities: demographic information and years of work experience, focus and concentration, nurses’ perception of patients’ cognitive state, aspects of patients and patient care, activities and modes of communication, and space, equipment, and supplies in use for the patients. The 2nd questionnaire developed was the End of Shift Questionnaire (ESQ). The initial version of the ESQ consisted of 164 questions. The final version of the ESQ resulted in 143 questions. Subsections addressed demographic information and years of work experience, focus and concentration, the care continuum, adverse event occurrence, activities and modes of communication, employee lounge use, team cohesion, and anxiety. MMHA6700 IT Transforming Health Care Paper

 

Also, the subsections addressed space, equipment, and supplies in use for the patients. A question on the ESQ asked the nurses to record how many steps were taken during a shift. The 3rd questionnaire was the Patient Experience Survey (PES). The initial version of the PES consisted of 118 questions. The final version of the PES consisted of 93 questions. Subsections address characteristics of patients, the patient experience, and patient outcomes: demographics, current room and areas around it, happenings in the room, frequency and intensity of reported pain, experiences with visitors and staff, and experiences with a roommate. During the 1st phase of the study, participating nurses were asked to fill out the PSEQ-RNP. Patients of participating nurses were also administered the PES. Principal components analysis with varimax rotation was performed on the PES to identify relevant components for statistical analyses. To identify relationships among HIT use, unit layout, and patient safety, the PES component bed-to-bathroom transfer was included. Items that were included in the component (! = .90) were ‘‘ease of walking on my own to the bathroom,’’ ‘‘ease of performing activities inside the bathroom,’’ and ‘‘ease of getting in and out of bed to go to the bathroom.’’ Nurses who participated in phase 2 were asked to complete the ESQ toward the end of each shift and carry a PDA and a pedometer throughout the entire shift. Prior research had utilized the combination of techniques.24 To convert steps recorded on pedometers to travel distances, a standard conversion rate of 2.6 ft per step was used.6 Distance and PDA work sampling data were divided by a nurse’s shift duration so that data were normalized for statistical analyses. JONA Vol. 42, No. 5 May 2012 Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Procedure Prior to data collection, the study was approved by the participating institution’s human subjects review board. Following approval, the researchers walked the 7 nursing units while taking an inventory of their features including nursing stations, sinks, supplies, equipment, and patient bathroom locations. During the first phase, the PSEQ-RNP and PES were administered approximately at the same time toward the end of each nurse’s shift. Every day for 2 weeks, a research team member arrived on a medicalsurgical unit during the day shift and asked the nurses which patients were alert and oriented and capable of filling out a questionnaire. Nurses and their patients who participated in the study participated only once. The PSEQ-RNP was available to the nurses through a SurveyMonkey software link sent via an e-mail or through an intranet-based home page accessible at computers-on-wheels or computers at nursing stations. The research team member reviewed the contents of an informational letter with each nurse. A consenting nurse was asked to complete the PSEQRNP toward the end of his/her shift. The nurse’s eligible patients were approached by the research team member with an informational letter. A consenting patient was asked to fill out a hard copy of the PES. Patients filled out the questionnaire independently of the research team member. Once finished with the questionnaire, the patient placed it in a sealed envelope. During the 2nd phase, RNs took part in a work sampling study during which each nurse wore a pedometer and carried a PDA device that rang randomly 30 times per 12-hour shift. A participating nurse wore a pedometer and carried a PDA device for only 1 shift. When initially prompted for an ID, the nurses selected a designation that did not reflect their true identity. Each additional query asked about the location and activity of the RN. At the end of their shifts, the nurses filled out the ESQ. MMHA6700 IT Transforming Health Care Paper

 

On the ESQ, the nurses reported their IDs and the number of steps recorded by their pedometers. During the 3rd phase, 40 nurses took part in charrettes. Nurses were presented with the research findings and provided a time for review and comment in a focus group format. The nurses were then asked to provide details of an ideal nursing unit and patient room to a sketch artist who helped visualize their ideas. Results Phase 1: Patient-Centered During analyses of phase 1, stepwise multiple linear regression analyses and Pearson pairwise correla- tions were performed to explore relationships among care delivery, unit layout, HIT use, and the patient experience. Data from the PES and PSEQ-RNP were used in the analyses. Predictors of Pain Intensity A stepwise multiple linear regression was performed to identify variables that predicted patients’ reported pain intensity. Data from the PES and PSEQ-RNP were used in the analysis. Variables excluded from and included in the final regression model are shown in Table 1. The patient bed-to-bathroom transfer and nurses’ reported frequency of documentation at the main nursing station accounted for 16% of the variance in patient pain intensity. The harder it was for patients to perform bed-to-bathroom transfers as reported by the patients, the greater the pain reported by the patients. The more often the nurses reported documentation at the main nursing station, the greater the pain reported by the patients. Predictors of Patients’ Near-Falls Stepwise multiple linear regression was performed to identify variables that predicted the patients’ reported frequency of catching himself/herself from a fall or a near-fall. Data from the PES and PSEQRNP were used in the analysis. Variables excluded from and included in the final regression model are shown in Table 2. In total, patient bed-to-bathroom transfers and nurses’ reported frequency of documentation at the patients’ bedsides accounted for 12% of the variance for near-falls. The patient bed-to-bathroom transfer was a predictor of more reported near-falls. However, the nurses’ reported frequency of documentation at the patients’ bedsides predicted fewer instances of near-falls. Distractions and Computers-on-Wheels Pearson pairwise correlations revealed that nurses who reported frequent use of computer-on-wheels reported more documentation in the hallway (r = 0.76, P G .01). Conversely, there was a smaller but statistically significant correlation between nurses’ reported computer-on-wheels use and documentation at the patient bedside (r = 0.28, P G .05). Sixty-three nurses reported experiencing distractions while preparing medications. Nurses reported the most distractions while preparing medications at the computers-onwheels. Figure 1 shows the frequencies of nurses’ reported distractions at various locations on the units while preparing medications. Phase 2: Nurse Work Sampling During analyses of phase 2 data, stepwise multiple linear regression analyses were used to examine relationships among nurses’ characteristics, communication JONA Vol. 42, No. 5 May 2012 Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction o…