A Mixed-Methods Research Framework for Healthcare Process Improvement

Journal of Pediatric Nursing (2015) xx, xxx–xxx

A Mixed-Methods Research Framework for Healthcare Process Improvement1 Nathaniel D. Bastian ABD, MS, MEng a,b,c,⁎, David Munoz PhD, MEng a , Marta Ventura MS a a

Center for Integrated Healthcare Delivery Systems, Department of Industrial & Manufacturing Engineering, The Pennsylvania State University, 362 Leonhard Building, University Park, PA b Center for Healthcare Innovation, Education and Research, Department of Health Organization Management, Texas Tech University, 703 Flint Avenue, Lubbock, TX c Center for Health and Humanitarian Systems, Stewart School of Industrial & Systems Engineering, Georgia Institute of Technology, 755 Ferst Drive, NW, Atlanta, GA Received 4 September 2015; accepted 6 September 2015

Key words: Mixed-methods research; Healthcare management; Pediatric care; Quality engineering; Process improvement; Workflow assessment

Purpose The healthcare system in the United States is spiraling out of control due to ever-increasing costs without significant improvements in quality, access to care, satisfaction, and efficiency. Efficient workflow is paramount to improving healthcare value while maintaining the utmost standards of patient care and provider satisfaction in high stress environments. This article provides healthcare managers and quality engineers with a practical healthcare process improvement framework to assess, measure and improve clinical workflow processes. Design and Methods: The proposed mixed-methods research framework integrates qualitative and quantitative tools to foster the improvement of processes and workflow in a systematic way. The framework consists of three distinct phases: 1) stakeholder analysis, 2a) survey design, 2b) time-motion study, and 3) process improvement. Results: The proposed framework is applied to the pediatric intensive care unit of the Penn State Hershey Children's Hospital. The implementation of this methodology led to identification and categorization of different workflow tasks and activities into both value-added and non-value added in an effort to provide more valuable and higher quality patient care. Conclusions: Based upon the lessons learned from the case study, the three-phase methodology provides a better, broader, leaner, and holistic assessment of clinical workflow. The proposed framework can be implemented in various healthcare settings to support continuous improvement efforts in which complexity is a daily element that impacts workflow. Practical Implications: We proffer a general methodology for process improvement in a healthcare setting, providing decision makers and stakeholders with a useful framework to help their organizations improve efficiency. Published by Elsevier Inc.

THE HEALTHCARE SYSTEM in the United States is spiraling out of control due to ever-increasing costs without significant improvements in quality, access to care, satisfaction,

and efficiency. In 2010, healthcare expenditures grew 3.9%, reaching $2.6 trillion (Martin et al., 2012). In order to transform the current healthcare system into one that is high quality,

1 This paper is based upon work partially supported by the National Science Foundation under Grant No. IIP-1361509. In addition, the first author was supported by the National Science Foundation under Grant No. DGE1255832. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation and Pennsylvania State University. ⁎ Corresponding author: Nathaniel D. Bastian, ABD, MS, MEng. E-mail address: [email protected].

http://dx.doi.org/10.1016/j.pedn.2015.09.003 0882-5963/Published by Elsevier Inc.

2 patient-centered, and efficient, case studies can help drive process improvement efforts supported by evidence-based practices. The purpose of this article is to provide healthcare managers and quality engineers with a mixed-methods research framework for healthcare process improvement. Mixedmethods approaches use a systematic integration of qualitative and quantitative methods with the ultimate objective of developing a better and deeper understanding of a phenomenon (Chen, 2006; Creswell and Plano Clark, 2007). Thus, the use of both quantitative and qualitative tools becomes useful to foster the improvement of processes and workflow in a systematic way. Among the various initiatives for reforming healthcare financing, value-based purchasing (VBP) is a strategy that is oftentimes used to measure, report, and reward excellence in healthcare delivery. VBP involves the actions of coalitions, employer purchasers, public sector purchasers, health plans, and individual consumers in making decisions that take into consideration access, price, quality, efficiency, and alignment of incentives. VBP promotes the quality and the value of healthcare services with a clear return for every dollar spent. It does so by eliminating healthcare errors, adopting evidence-based care standards and protocols, and re-engineering hospital operations and processes (Scanlon et al., 2002). An example of the way VBP seeks to reward hospitals for improving the quality of care is by distributing Medicare payments in a way so that the hospitals with higher performance in terms of quality receive a greater proportion of the payment than do the lower performing hospitals. To implement VBP or similar provider reimbursement mechanisms, there must be health system re-engineering that improves clinical processes and operations by increasing the value of healthcare delivered to patients. In an effort to help re-engineer a more valuable healthcare delivery system across the United States, healthcare managers and quality engineers can employ a mix of methods from industrial and systems engineering, management science, and information technology arenas to overcome the current barriers facing the successful implementation of process improvement in hospital settings. Upon presentation of our mixed-methods framework for healthcare process improvement, we demonstrate its usefulness using a case study where we assess clinical workflow in the pediatric intensive care unit (PICU) at the Penn State Hershey Children's Hospital (PSHCH, 2014). Efficient workflow is paramount to improving value while maintaining the utmost standards of patient care and provider satisfaction in high stress environments, like the PICU. There are significant organizational factors within healthcare operations such as staffing, teamwork, patient volume, pressures of work, information technology, and structure that can impact patient outcomes and clinician satisfaction. Moreover, the dynamic environment of intensive care requires clinicians to change tasks many times when performing patient care activities. Hence, a key component of process improvement is better understanding of the various elements comprising workflow. Specifically, we must learn and incorporate the voice of the customer, uncover root causes to workflow inefficiencies, conduct task identification and categorization, and then assess and analyze the clinical workflow

N.D. Bastian et al. processes based on these tasks. These steps combined with quality improvement and data visualization tools can help to effectively drive process improvement efforts in a healthcare organization. Healthcare managers and quality engineers, also referred as continuous performance improvement (CPI) consultants, are the main facilitators of process improvement in hospital settings (Weed, 2010). Understanding the interrelationships impacting and shaping system behavior can help drive process improvement efforts in healthcare operations. These efforts seek to understand systems, their sub-components, and their relationships to each other, while believing that the understanding of the interrelations and interactions among its elements deeply define the system and its behavior (Adam and de Savigny, 2012). With a better understanding of the system and its elements (i.e., clinicians, patients, processes, information, etc.), healthcare managers and quality engineers can more easily identify the leverage points leading to desired process improvement outcomes. The Lean Six Sigma methodology strives to eliminate waste of physical resources, time, effort, and talent, while assuring quality in production and processes. Therefore, the implementation of lean principles in a hospital setting can help maximize healthcare service value while minimizing waste. The identification and characterization of both value-added and non-value-added tasks helps to effectively assess workflow and focus healthcare process improvement efforts. Value-added tasks are those activities that transform material, information, or people into something that the customer cares about (e.g., diagnosis, treatment, care plan, etc.), while non-value-added tasks consume resources but do not add value (e.g., searching for supplies, staff waiting, re-work, redundant paperwork, etc.) to the process. There are numerous methods for process improvement using approaches such as value stream mapping, process flow mapping, system dynamics modeling, statistical process control, social network analysis, and simulation. Given the variety of existing qualitative and quantitative methods, the application of a mixed-methods approach proves useful to face the challenges for process and organizational improvement.

Literature Review The application of process improvement and organizational change management methods in a hospital setting is not new, especially given the rapid growth of healthcare expenditures in the United States. With the onset of VBP financial incentive programs, which reimburse providers based on quality outcomes achieved rather than volume of health services delivered, health system re-design and re-engineering efforts have emerged to improve both value and quality of care as well as the efficiency of healthcare operations and processes. Methods that incorporate continuous improvement practices into organizational change can help guide and drive healthcare process improvement efforts. For example, Jimmerson et al. (2005) used a lean thinking approach for re-designing work and processes within hospital operations to facilitate problem-solving activities.

Mixed-Methods Research Framework for Healthcare Process Improvement These efforts, for instance, involve value stream mapping to distinguish between value-added and non-value-added steps, which provides a high-level view of the processes to improve process improvement efforts and problem-solving. More recently, Holden (2011) used a lean thinking approach for re-designing workflow processes in 15 emergency departments (ED) in the United States, Australia, and Canada, to reduce problems related to crowding, delays, cost, and patient safety. The study revealed numerous ED process changes involving separate patient streams, as well as structural changes such as new technologies, communications systems, staffing changes, and reorganization of the physical space. Even more recent research to facilitate learning and behaviors in hospitals during the early stages of lean implementation indicated that when lean was implemented properly, it was a transformational experience for those individuals who were directly involved (Mazur et al., 2012). Benneyan et al. (2003) also demonstrated how to use statistical process control charts to measure variation in healthcare processes to help determine whether changes were making a real, positive difference in outcomes. In addition to the integration of lean principles, Walley et al. (2006) investigated how hospital performance measurement systems influence decision-making processes. This research specifically used statistical process control charts and time-series analyses to monitor process improvement events. An additional study, by Borycki et al. (2006), described a simulation-based methodological framework to evaluate the effects of healthcare information systems on clinician workflow in the performance of both routine and complex clinical tasks in a laboratory setting. Furthermore, healthcare information technology and data standardization can also support effective process improvement. Zheng et al. (2010) presented a set of new analytical methods to find latent regularities embedded in clinical work processes and workflow. Recently, Smith et al. (2012) also discussed the benefits of and barriers to data standardization across the healthcare industry and how to support progress toward standardizing data in an improved healthcare supply chain. In terms of organizational change for process improvement, Callender and Grasman (2010) assessed material management in the healthcare sector by highlighting the areas of improvement, identifying barriers for implementing supply chain management practices, and analyzing material management best practices. Finally, Schell and Kuntz (2013) discussed the challenges and opportunities facing healthcare managers and clinical nurse leaders who are partnering in the complex, multi-faceted healthcare system to improve patient outcomes by implementing change.

3

result, this paper addresses a gap in the literature in that no previous studies have described. This study specifically proposes a mixed-methods research framework for healthcare process improvement, which is accompanied by a case study that demonstrates the application of the proposed method. We extend the existing healthcare management body of knowledge by proffering a practical methodology that provides a better, broader, leaner, and holistic assessment necessary to measure and improve clinical workflow processes in a hospital setting. This research will help practicing healthcare managers and quality engineers integrate continuous process improvement practices into healthcare facilities. Further, the guidance provided, supports the identification and analysis of the main elements of the system and how key workflow tasks can be improved to provide valuable and quality patient care.

Proposed Mixed-Methods Research Framework Continuous improvement practices are at the core of our proposed healthcare process improvement framework. This framework embodies operational thinking, which involves understanding the physics of operations and the cause-andeffect relationships within the processes of a system with a goal to decrease or eliminate non-value-added time. Moreover, the proposed framework incorporates lean thinking, which seeks to maximize service value while minimizing waste. Lean thinking helps identify and categorize both value-added and non-value-added tasks to effectively assess workflow and processes in a given system. In designing a process improvement framework, these principles and practices provide rich guidelines to develop a mixed-methods approach in which both qualitative and quantitative tools are used to explore not only the elements of a system but also how they interact. As a result, our proposed mixed-methods framework for healthcare process improvement consists of three distinct phases: 1) stakeholder analysis, 2a) survey design, 2b) time-motion study, and 3) process improvement. These complementary approaches provide a holistic understanding of the complex workflow for process improvement in healthcare settings while incorporating lean principles for waste elimination. Figure 1 presents an overview of our mixed-methods framework for healthcare process improvement, while Figure 2 provides a detailed flow chart of the sub-steps included in each phase of the proposed methodology. For a more detailed description of each sub-component depicted in Figure 2, please refer to the supplementary material available online.

Motivation and Purpose

Case Study of Mixed-Methods Research Framework Implementation

Although healthcare process improvement methods are not new, this paper is unique in that we provide a general mixed-methods approach for replication in other healthcare settings, as well as highlight lessons learned from a real-world implementation of the proposed method. As a

In order to demonstrate the usefulness of our proposed mixed-methods framework for healthcare process improvement, we describe its implementation using a case study and highlight several lessons learned. In this case study, we specifically assess the current state of workflow in the PICU

4

N.D. Bastian et al.

Figure 1

Mixed-methods research framework for healthcare process improvement.

at the Penn State Hershey Children's Hospital (PSHCH). The PSHCH, a component of the Penn State Milton S. Hershey Medical Center, serves an area in South Central Pennsylvania with more than one million children, caring for more than 125,000 patients each year (PSHCH, 2014). In 2013, it was ranked by the U.S. News and World Report as one of the nation's best children's hospitals for five specialties: cancer, urology, orthopedics, neurology, and neurosurgery (U.S. News and World Report, 2013). The PICU is composed by a multidisciplinary team led by a pediatric intensivist. This team is composed of nursing staff, pharmacists, respiratory therapists, social workers, and other medical subspecialties.

Figure 2

The rest of this section summarizes the main case study results and lessons learned.

Phase 1: Stakeholder Analysis Unstructured Interviews To narrow the scope of the problem at-hand while gaining an understanding of the most pressing clinical workflow issues at the PICU to effectively capture the voice of the customer, we held two separate meetings during the initial visit, one with two physicians (attending physician and pediatric intensivist) and the other with two nurses (nurse

Process flow chart of methodology.

Mixed-Methods Research Framework for Healthcare Process Improvement manager and clinical nurse specialist). Note that we had separate meetings to avoid frictions between physicians and nurses, while exploring some areas of inefficiencies and dissatisfaction. During the unstructured interviews, we provided some guidance to the respondents as an exercise to explore in detail the latent issues. Based on the key PICU stakeholders' comments during the interviews, the discussion was invaluable because it provided them with the possibility of mentally “walking through” various processes and workflow elements that they did not previously consider important. Observational Study After conducting interviews with the key stakeholders, we observed the activities and tasks conducted by both physicians and nurses in the PICU. Although the emphasis of the observational study was to examine the issues previously identified, we were also able to identify other causes of clinical workflow inefficiencies. More specifically, we observed how repetitive and non-repetitive tasks were conducted, the interactions among clinicians, the interactions between clinicians and technology, sequences of tasks, and so on. Insights Gained From the Stakeholder Analysis Phase The unstructured interviews and the observational study were both extremely helpful for both the research team and the healthcare stakeholders. These complementary approaches served to effectively capture the voice of the customer, define the current situation, and understand the scope of the problem. The issues encountered were classified into three main groups: daily PICU operations (inefficient procedures and activities, documentation management, shift changes, information flow, and medication management), PICU layout (potential issues with the family-centered layout including long distances, lack of visibility, centralized nurse stations, and location of med rooms), and PICU IT system (lack of system integration, lack of flexibility, system response time, and usability).

Phase 2a: Survey Design We designed and deployed a customized survey to the physicians and nurses of the PICU. The survey (displayed in the Appendix A) was focused on the workflow efficiency and clinician satisfaction issues encountered in the stakeholder analysis phase. The survey was composed of 10 questions including both qualitative and quantitative questions. The main components of the survey included perception of time spent on different work-related activities (and the value of those activities), information flow, and information technology. The qualitative section included an open-ended question for the owners of the processes (physicians and nurses) to provide recommendations for improvement based on their own best practices. We used an online survey platform to distribute and collect the responses. The head physicians of the PICU were in charge of distributing the survey to all the other physicians and nurses of the PICU. We obtained 20 responses in total, including

5

four from physicians and 16 from nurses, which represented a response rate of 75% for physicians and 45% for nurses. As argued by the head physicians and nurse managers (key healthcare stakeholders) in the PICU, one of the main advantages of designing a customized survey was that detailed insights could be identified, and, thus, targeted solutions could be provided to overcome them. Perceived Task-Time Allocation The perception of time spent on different workflow related activities obtained through survey responses was divided into five categories: communication, delivery, documentation, transit, and supervision/teaching. In Figure 3, box plots are shown to highlight the differences (and variability) in the perceived time spent on each of the categories broken down by physician, nurse, and both. Please note that Q1 represents the 25th percentile and Q3 represents the 75th percentile. As indicated in Figure 3, nurses perceived to spend a much greater amount of time than physicians in communicating with patients at bedside, with other healthcare providers and family, and with others on non-work related activities. However, there is much greater variability and skewness in the nurses' responses compared to those of the physicians. In terms of delivery, which includes the time spent in delivering medications and blood products, the box plot for the nurses indicates a relatively low variability within the central box (2nd to 3rd quartiles); however, data are skewed and the maximum value is about six times the mean. From this result, future investigations can be proposed to determine the causes of the variability and high maximum values obtained. Both physicians and nurses perceived to spend a considerable amount of time conducting documentation, such as entering patient data and viewing orders/other health care provider notes/vital signs/physical assessments, etc. These results were explored in greater detail to determine whether factors such as level of training, mentoring, interface usability, or others have an impact on the variability observed for both, physicians and nurses. Regarding transit time, nurses were the only ones involved in the time spent obtaining medications from pyxis, blood products from the blood bank, and equipment necessary for patient care. As transit is considered a non-value added task, the clinical activities that require transit can be analyzed to propose mechanisms to reduce it. Therefore, the remaining time on other activities that add value to the provision of care can be reallocated. Lastly, both physicians and nurses were perceived to spend a significant amount of time supervising and teaching other healthcare providers. Overall, these survey results indicate that physicians were perceived to spend more time (on average) documenting and supervising/teaching, while the nurses were perceived to spend more time (on average) on communication, delivery, and transit. These findings helped us identify and categorize the key tasks pertinent to improving clinical workflow processes. Moreover, the variability encountered in the dimensions analyzed in this case study serve as the baseline for future improvements and

6

N.D. Bastian et al.

Figure 3

Box plots of perceived task-time allocation.

efforts to develop a standardized provision of care in which variability is an undesirable component. Information Flow The IT and electronic health record (EHR) systems at the PICU both played an important role in nearly every task performed by both nurses and physicians; yet, it caused dissatisfaction to nurses and physicians with respect to documenting the patient's health because it was not very user-friendly. Survey responses indicated that 75% of the physicians and 50% of nurses agree that the information flow needed improvement. Another cause of inefficiency described through the survey was that IT systems did not provide enough support for communication and coordination. Additional issues identified related directly to the mechanisms to inform and communicate between physicians and nurses. Some survey respondents believed that although exchange of information can be made directly through the IT system, verbal communication could not be replaced. Based on these results, we identified various areas of opportunity to improve information flow, communication, and coordination. Establishing clear communication procedures as well as an appropriate utilization of electronic devices was a critical part of improving workflow efficiency (Munoz et al., 2014). Information Technology As expected from the initial site visit, the survey results indicated that clinical staff was not satisfied with the IT system interface or with its functionality. The IT categories and their IT factors associated can be seen in Table 1. For the system design category, we found that doctors were fairly dissatisfied with each one of the sub-categories included, because the three factors in this category (i.e., login issues, password issues, and integration issues) all received

an average rating of 1.25 on a 5-point scale (very dissatisfied to very satisfied). On the other hand, nurses were relatively neutral when asked about system design because the average rating for login issues was 3.44, for password issues was 3.06, and for integration issues was 2.81. For the second category, entering data, physicians were again more dissatisfied than nurses regarding the three factors in this category. For physicians, the averages for the electronic patient notes, entering orders, and entering patient data were 2.25, 2.25, and 3, respectively. For nurses, the averages for the electronic patient notes, entering orders, and entering patient data were 3.19, 3.36, and 2.93, respectively. Although the averages for this category were slightly better; they were still below 4 (satisfied) on the 1–5 Likert scale. The last IT category, viewing data, was composed of three factors: viewing patient data, viewing lab values and viewing radiographic studies. The averages for those three categories according to physicians were 2.50, 3.50, and 3.33 respectively, and for nurses 2.94, 3.31 and 3.36. In general, no meaningful differences were found between physicians' Table 1

IT categories and factors.

IT category

Factors

Code

System design

Login Password Integration

LOG PASS INT

Entering data

Electronic patient notes Entering orders Entering patient data

NOTE ORD ENT

Viewing data

Viewing patient data Viewing lab values Viewing radiographic studies

VIEW LAB RAD

Mixed-Methods Research Framework for Healthcare Process Improvement Table 2 Correlation coefficient intensities, adaptation of Salkin (2007).

and nurses' level of satisfaction in this category. On average, the viewing data category had a better perception according to physicians when compared with the other two categories explored in this research. From an open-ended question related to the IT system included in the survey, we received valuable comments and recommendations to improve the IT interface design. The elimination of multiple-clicks appears to be a main concern for clinicians; this non-value-added task could be reduced or eliminated with a friendlier interface design. Elimination of the automatic shutdown of PowerChart – a multi-entity electronic medical record (EMR) that includes functionalities such as entering clinical data, completing orders, and generating automated patients' reports – was also suggested. Thus, a balance between the benefit of assuring patient confidentiality and workflow efficiency must be considered. Non-value-added activities such as having to login multiple times during the day and having to remember multiple passwords were also important causes of dissatisfaction with the IT system. Another common cause mentioned frequently by survey respondents was that the system froze too often; thus, valuable time was being lost every day causing not only clinician dissatisfaction but also reducing clinician performance. All the identified IT issues provided the basic starting point to focus efforts of IT designers when improving clinicians' satisfaction in the use of IT systems. A more

Table 3

Correlation matrix IT.

7

efficient and user-friendly IT system design will have an impact on workflow efficiency and clinician satisfaction. Another interesting finding from the survey results was the study of the correlation among the IT factors. The results obtained from the correlation study served to inform IT designers and engineering managers to find common patterns that might lead to dissatisfaction in different IT dimensions. In order to study this, we calculated the Pearson product–moment correlation coefficients and created a correlation matrix to present the correlations among each of the IT factors. The interpretation of the intermediate values of correlation coefficients can be obtained from the Salkin scale (Salkin, 2007) and are presented in Table 2. Additionally, the correlation matrix is shown in Table 3. As emphasized in the proposed framework, understanding the relationship among different elements is important to comprehend the behavior of the system as a whole. A correlation analysis is helpful for this purpose. An overall strong, positive correlation can be seen among the factors in system design category (login, password, and integration). A correlation coefficient of 0.836 was calculated between the satisfaction reported for login and password. In addition, a strong positive correlation was found for integration and login (0.732), and integration and password (0.675). At least moderate positive correlations were found among the factors composing the category of entering data. Entering notes was moderately positive correlated to entering orders (0.485) and entering patient data (0.496). A strong positive association was found between entering orders and entering patient data (0.708). In the third category, viewing data, at least moderate associations were found among the three factors. Viewing patient data was moderately positive associated to both, viewing lab values (0.471) and viewing radiographic studies (0.555). Additionally, a strong, positive correlation was found between viewing lab values and viewing radiographic studies (0.745). Other interesting correlations were found among inter-category factors. For instance, system integration was found to be strongly

8 positive correlated to entering notes (0.790) and entering patient orders (0.603). Another strong association was found between viewing patient data and entering patient data. It is also worth to be mentioned that 72% (26 out of 36) of the correlations among every item were at least moderately correlated (i.e., a correlation coefficient greater than 0.4). This correlation analysis was useful to understand that the components of an IT system cannot be isolated from the others; as a consequence, most of them will have commonalities or patterns that affect users' satisfaction. These results can be used to inform IT designers to identify these commonalities and impact correlated IT dimensions. Insights Gained From the Survey Design Phase The customized survey used in this research was extremely valuable to further capture the voice of the customer in a structured way, discover root causes to the problem, and better understand the areas needing improvement through workflow task identification and categorization. Including all the key stakeholders was necessary as they each had different goals, and, therefore, different needs. In addition, by working with the key stakeholders we obtained various best practices and ideas about potential recommendations to improve workflow efficiency. Finally, we were able to identify what the perceived critical to quality (CTQ) elements to measure work performance were. Those CTQ elements included the satisfaction of patients' families, patient safety, efficient use of resources and time, as well as job satisfaction. Most of these CTQ elements were aligned with the national markers that measure work performance such as the virtual PICU (standardize mortality rates, length of stay, severity of illness, and readmission rates), hospital-associated infections, unplanned extubations, and job satisfaction/turnover (AHRQ, 2014).

Phase 2b: Time-Motion Study Since a time-motion study can reveal various interactions among the elements of a system, as well as how different processes and tasks shape the behavior of the clinical workflow and provision of care, we used this technique to estimate the times spent on each activity in the PICU. As mentioned previously, a time-motion study is a direct and continuous observation of a task, using a timekeeping device (e.g., stopwatch, handheld computer, etc.) to record the time it takes to accomplish a task. A time-motion study is often used if there are repetitive work cycles of short to long duration, where a wide variety of dissimilar work is performed and process control elements form part of a cycle. A comprehensive time-motion study consists of a study goal setting, experimental design, time data collection, data analysis, and reporting. The data obtained through our time-motion study served as a baseline for our improvement efforts. Given our detailed understanding of the current state of activities in the PICU, we could then proceed with proposing an improved process and workflow methods to eliminate non-value-added activities. The workflow task categorization is shown in Table 4.

N.D. Bastian et al. Table 4

Workflow task categorization.

ID

ID name

Sub ID

Sub ID name

1

Patient monitoring

A B C

General care Patient communication Family communication

2

Collaboration

D E

Rounds General interaction

3

Medication

F G

Preparation Administration

4

Documentation

H I

Electronic Paper-based

5

Transit

J K

Medicine storage Other

6

Supervision

L

Supervising others

7

Miscellaneous

M N O

Non-work communication Breaks Other

Time-Motion Study Template and Report We developed a template in Excel, as depicted in Figure 4, to facilitate the data collection for our time-motion study. The user is only required to enter the activity ID in the appropriate cell, and then all other cells in the template will be updated automatically (i.e., time stamp, duration, and ID name). In addition, more activities can be added to the template at a later time. This fact makes the time-motion study template flexible enough to be used in a higher or lower level of abstraction, or even be used in a different healthcare setting. As illustrated in Figure 5, another important feature of this template for our time-motion study is that the report is automatically generated. It depicts the percentage and time spent on each activity through both pie and bar charts. An additional benefit of the tool is that it can be easily adapted to build reports across multiple days and/or across multiple PICU personnel. Because the healthcare stakeholders involved in this project were asked to collect the data for our time-motion study, this template was extremely helpful as it provided a user-friendly platform for data collection. Moreover, they saw that the tool could be used as an academic-learning experience for new staff joining the PICU to understand the clinical workflow activities involved in the provision of patient care (Munoz et al., 2014). Time-Motion Study Sample A second site visit to the PICU facility was conducted with the main objective to test the time-motion study template and provide training about its use to the PICU clinicians. In addition, one nurse and one physician were shadowed during an eight-hour shift. During these observations the time-motion study template was employed by the clinicians to collect data about time spent on daily workflow activities. Although the specific results are not presented here, the time-motion study was invaluable in that it enabled

Mixed-Methods Research Framework for Healthcare Process Improvement

Figure 4

9

Time-motion study template.

proper task sequencing for process flow mapping that was used to assess and analyze workflow efficiency in the PICU. Insights Gained From the Time-Motion Study Phase The time-motion study template and report serves as a process improvement mechanism to assess the appropriateness and the time decomposition of workflow processes. In addition, as this approach includes the sequence of the activities, fragmentation of workflow can be further investigated to identify critical activities. This provides a holistic view of the interactions among the various processes in the provision of care. The PICU leaders also mentioned that the developed time-motion study template could be used for academic-learning experiences to help new PICU staff to understand the main activities and workflow involved in the provision of care.

Phase 3: Process Improvement The previous phases helped to understand the clinical workflow and processes at the PICU and also identify the main areas of opportunity. Tools such as RCA can be used to properly organize those areas of opportunity to be addressed. This tool aims to identify root causes of problems as opposed to merely addressing their consequences or symptoms. In this section, we provide an initial description of how this tool was used to guide process improvement by identifying root causes and potential tailored solutions to remove sources of inefficiencies. Based on the previous phases of the proposed methodology, we identified four main categories impacting clinical efficiency

and satisfaction: communication, layout, procedures and standards, and health information technology (HIT). As depicted in Figure 6, we performed a bit of root cause analysis using a fishbone diagram to explore into greater detail the potential causes for issues within the HIT category. The HIT category was further decomposed into four groups that might impact clinical efficiency and satisfaction: interface design, systems integration, mental workload, and systems. This type of fishbone diagram served as the baseline or road map for generating recommendations and courses of action for process improvement. In this sense, a similar procedure was conducted using a fishbone diagram but for recommendations (Figure 7) rather than root causes. In order to identify these tailored recommendations, techniques such as revision of best practices in literature reviews, brainstorming, or expert focus groups could be used. For these last two techniques, we recommend that both internal stakeholders and external evaluators participate. Moreover, techniques such as FMEA can be incorporated into this discussion to prioritize courses of action that provide the best value given the limited resources to be assigned to cover the identified recommendations. It must be understood that the recommendations stated should be analyzed accordingly to evaluate whether or not they can be implemented and what is the level of resources needed in such a case. Additionally, it should be noticed that most of the recommendations provided require the collaboration with other areas of the hospital. In the example that was shown, the IT team should be highly involved in the implementation of the recommendations that are viable.

10

N.D. Bastian et al. REPORT TIME-MOTION STUDY PICU Penn State Hershey Children's Hospital Date: Department: Position: ID 1 2 3 4 5 6 7

Saturday, February 9th Pediatric Intensive Care Unit

ID Name

Duration

Frequency

Avg Duration

11.3 34.0 8.1 26.4 13.4 5.8 2.0

12 7 6 17 6 5 1

56.7 291.0 80.8 93.0 133.8 70.0 117.0

Patient Monitoring Collaboration Medication Documentation Transit Supervision Miscellaneous

Duration %

Min Cycle

Max Cycle

6% 2% Patient Monitoring

11%

Collaboration

13%

Medication

34%

Documentation Transit

26%

Supervision

8%

Miscellaneous

Duration 40.0 35.0 30.0 25.0 20.0 15.0 10.0 5.0 0.0

Patient Monitoring Collaboration Medication Documentation Transit Supervision Miscellaneous

Figure 5

Time-motion study template automated report.

Summary of Case Study Lessons Learned The involvement of the healthcare stakeholders in the implementation of the mixed-methods approach was critical to fully capture the voice of the customer, understand the PICU workflow and processes, and discover the root causes of inefficiencies (including waste and variability). From an implementation perspective, engaging the owners of the process in the problem development was helpful to clearly define (and re-define) the objectives and scope of the process improvement efforts. In addition, the healthcare stakeholders served as central communication channels to explain the objectives of the healthcare process improvement effort to

the rest of the PICU team. This communication was helpful in terms of encouraging the personnel to respond to the survey. The data obtained from the survey were particularly helpful to determine areas of inefficiency and dissatisfaction in addition to identifying and categorizing the clinical workflow tasks. The unstructured interviews and time-motion study sample were invaluable for the healthcare stakeholders to walk through the processes involved in the provision of care. These two complementary approaches were not only useful for us to frame the problem, but also for the PICU leaders to gain a holistic view of their workflow. Moreover, the results

Mixed-Methods Research Framework for Healthcare Process Improvement MENTAL WORKLOAD

11

INTERFACE DESIGN COMMUNICATION Non-Customizable screens

Unnecessary steps Unorganized screens High depth menus

PROCEDURES AND STANDARDS

No user-friendly design Many screens Amount of Info to skip Unintuitive steps

Redundancy

HIT Freezes too frequently

Clinical Efficiency and Satisfaction

Lack of system integration Non-Integrated password

Non-timely IT support Printing capabilities

Duplicated capabilities

Frequent log-outs

LAYOUT

Lack of multiple screens

SYSTEM

INTEGRATION

Figure 6

Root cause analysis for HIT issues identified.

obtained served as the baseline to investigate best practices and the implementation of lean principles to reduce waste in terms of time, motion, and over-processing.

Conclusions and Implications This research proposes a mixed-methods framework for process improvement in a healthcare setting and demonstrates its use for practicing healthcare managers and quality engineers through a case study. In this application of our proposed approach, we assessed, measured, and analyzed the

MENTAL WORKLOAD

clinical workflow in the PICU of the Penn State Hershey Children's Hospital. In particular, the implementation of this methodology led to identification and categorization of different workflow tasks and activities into both value-added and non-value added in an effort to provide more valuable and higher quality patient care. We found that the use of continuous improvement practices was valuable in terms of guiding the development of the mixed-methods framework as a way to develop a holistic understanding of the details regarding the workflow in this system.

INTERFACE DESIGN

Conduct mental workload assessment that includes an assessment of clicks, screens needed for human-computer interaction tasks Assess current processes or steps needed for critical activities involving HIT

Conduct usability assessment to develop a user-friendly interface Assess redundancies Investigate what data is useful in PICU medical reports Allow customizable reports

Redesign processes to improve intuitiveness

Put in place a user and IT designer team to improve interface usability

Assessment of current hardware capacity

Evaluate potential integration of current systems Investigate viability of Develop a specific protocol for implementing a unique HIT assistance login system (biometric Redesign automated log-outs fingerprint or smartcard) to save information and Improve security / password retain working information administration for a unique in a faster manner password or code Develop mechanism to load Integrate electronic devices for only needed data data entry

SYSTEM

INTEGRATION

Figure 7

Root cause analysis for HIT recommendations.

Improve Clinical Efficiency and Satisfaction through HIT recommendations

12 The mixed-methods research approach described in this study could be implemented in various healthcare settings to support process improvement efforts in which complexity is a daily element that impacts workflow. This methodology provides the decision makers and stakeholders in healthcare with a useful framework to help their organizations improve efficiency by better understanding of the complexity of the workflow. The tools employed in our proposed approach will help healthcare managers and quality engineers to recognize which activities do not add value to the organization. Further, those managers who are already implementing lean principles in a healthcare organization can gain insights about how physicians and nurses spend their time. In addition, healthcare management practitioners who are just starting to investigate how lean thinking can be applied in their organizations can benefit from the proposed methodology, since it will help them gain a better understanding of the activities and processes that do not support efficiency or satisfaction. Understanding the elements that comprise workflow to gain a better, holistic understanding is essential to support a lean thinking transformation in the organization, which can increase the likelihood of a lean project to succeed (Mazur et al., 2012). Furthermore, using a mixed-methods approach may provide long-term benefits for an organization, such as gaining a better understanding of how processes actually work, which can be useful to support future lean projects. In addition, the case study incorporated the participation of all the nurses and physicians at the healthcare unit. It is valuable for the healthcare managers as well as for the clinician staff, so they can realize that better outcomes can be obtained if changes on certain procedures or tasks are made. Another advantage of the proposed methodology is that it can be used by practitioners at a minimal cost. As with most improvement efforts, the main investment needed to implement the proposed approach is the time for employees to study the process and implement improvement. This cost, however, is minimal in comparison with the benefits that can be obtained in terms of improved workflow efficiency and clinician satisfaction. Assessing the impact in terms of workflow efficiency and user satisfaction with healthcare information technology is a key element incorporated in this methodology. Healthcare managers will also be able to identify the main interface usability issues that impact efficiency and satisfaction. These two factors are critical to ensure a successful implementation and adoption of the EHR. From the proposed methodology, usability issues can arise and can serve as a base to collaborate with the EHR designer to create a more user-friendly interface to positively impact workflow outcomes. Based upon our initial lessons learned from the case study, we have already started working on phase 3 using quality improvement and data visualization principles to better understand workflow fragmentation and further study the central activities that play a critical role on clinical

N.D. Bastian et al. workflow. In particular, we have started investigating the clinician workflow complexities for a deeper assessment and visualization of workflow inefficiencies to better understand relationships and connections between tasks and root causes of inefficiencies and dissatisfaction. In future work, we will conduct another time-motion study at the PICU to collect additional data necessary to further implement process improvement initiatives. In addition, we will include staff registered nurses as key stakeholders at a PICU, since they would add significant, valuable insight to the workflow and processes given that they typically know the workarounds and process failures as they work through these on a daily basis.

Appendix A. PSHCH PICU Workflow Assessment Survey 1. What is your position at the pediatric intensive care unit (physician, nurse, resident, or nurse practitioner)? 2. How many years have you worked in the PICU at Penn State Hershey Children's Hospital? 3. On a typical day, which of the following tasks do you normally complete? How much time (in minutes) do you spend conducting each of these tasks? TIME (Minutes)

TASK Patient Care: Time spent directly with patient at the bedside (not including documentation) Collaboration or Communication: Time spent communicating with peers, other health care providers, or family Non-Work Related Communication: Time spent verbally or non-verbally (e.g. texting) communicating with others on non-work related activities Medication Delivery: Time spent from when the medication is physically in the room to time the medication infusion begins Blood Products Delivery: Time spent from when the blood product is physically in the room to time the blood product infusion begins Documentation Entry: Time spent to enter patient data (Vital signs, Physical Assessments, IandOs, Medications or Blood Products given), notes, and orders Documentation Viewing: Time spent viewing orders, other health care provider notes, vital signs, physical assessments, IandOs, medications administered Transit (Medication): Time spent obtaining medications from pyxis Transit (Blood Products): Time spent obtaining blood products from blood bank Transit (Non-Medication): Time spent obtaining equipment necessary for patient care Supervision: Time spent supervising other health care providers Teaching: Time spent teaching others

4. Are there other time-consuming tasks not listed above? If so, please list the task and approximate time you spend performing that task on a typical work day. 5. Of the tasks that you complete on a typical day, which (if any) of them seem unnecessary (i.e. not essential for your work)? 6. On a typical day at the PICU, how would you rate the flow of information? Needs Improvement

Satisfactory

Excellent

7. If you selected "Needs Improvement" or "Satisfactory" in the previous question, what are practical ways to improve the flow of information? 8. Based upon your use of IT, how satisfied are you with the current IT system?

Mixed-Methods Research Framework for Healthcare Process Improvement Very Dissatisfied

Dissatisfied Neutral Satisfied

13 Very Satisfied

NonApplicable

Computer Login Computer Passwords Integration Between Systems Entering Electronic Patient Notes Computer Order Entry Entering Patient Data Viewing Patient Data Viewing Laboratory Values Viewing Radiographic Studies

9. Based upon your responses to question #8, in what ways would you improve the IT system at the PICU? 10. How do you measure work performance? What metrics could be applied to show that you are doing a good job?

Appendix B. Supplementary Material Supplementary material to this article can be found online at http://dx.doi.org/10.1016/j.pedn.2015.09.003.

References Adam, Taghreed, & de Savigny, Don (2012). Systems thinking for strengthening health systems in LMICs: Need for a paradigm shift. Health Policy and Planning, 27. Agency for Healthcare Research and Quality (AHRQ) (2014). AHRQ Quality Indicators Toolkit for Hospitals. U.S. Department of Health & Human Services (available from http://www.ahrq.gov/professionals/ systems/hospital/qitoolkit/). Benneyan, J. C., Lloyd, R. C., & Plsek, P. E. (2003). Statistical process control as a tool for research and healthcare improvement. Quality Safety Health Care, 12, 458–464. Borycki, Elizabeth M., Kushniruk, Andre W., Kuwata, Shige, & Kannry, Joseph (2006). Use of simulation approaches in the study of clinical workflow. AMIA 2006 Symposium Proceedings (pp. 61–65). Callender, Carlos, & Grasman, Scott E. (2010). Barriers and best practices for material management in the healthcare sector. Engineering Management Journal, 22, 11–19. Chen, Huey T. (2006). A theory-driven evaluation perspective on mixed methods research. Research in the Schools, 13, 75–83. Creswell, John W., & Plano Clark, Vicki L. (2007). Designing and conducting mixed methods research. Thousand Oaks, Calif: SAGE Publications. Holden, Richard J. (2011). Lean thinking in emergency departments: A critical review. Annals of Emergency Medicine, 57, 265–278. Jimmerson, Cindy, Weber, Dorothy, & Sobek, Durward K. (2005). Reducing waste and errors: Piloting lean principles at intermountain healthcare. Journal on Quality and Patient Safety, 31, 249–257.

Martin, Anne, Lassman, David, Washington, Benjamin, & Catlin, Aaron (2012). Growth in US health spending remained slow in 2010: Health shared of gross domestic product was unchanged from 2009. Health Affairs, 31, 208–219. Mazur, Lukasz, McCreery, John, & Rothenberg, Lori (2012). Facilitating lean learning and behaviors in hospitals during the early stages of lean implementation. Engineering Management Journal, 24, 11–22. Munoz, David, Bastian, Nathaniel, & Ventura, Marta (2014). A workflow assessment for a pediatric intensive care unit: A mixed-methods approach”. Proceedings of the 2014 Industrial and Systems Engineering Research Conference (pp. 1–10). Penn State Hershey Children’s Hospital (PSHCH) (2014). retrieved August 29, 2014, available from http://www.pennstatehershey.org/web/childrens/home. Salkin, Neil J. (2007). Encyclopedia of measurement and statistics, 3-volume set. California, USA: SAGE Publications. Scanlon, Dennis, Chernew, Michael, & Doty, Hilary (2002). Evaluating the impact of value-based purchasing: A guide for purchaser. AHRQ, Publication, No. 02–0029. Schell, William, & Kuntz, Sandra (2013). Driving change from the middle: An exploration of the complementary roles and leadership behaviors of clinical nurse leaders and engineers in healthcare process improvement. Engineering Management Journal, 25, 33–43. Smith, Brian K., Nachtmann, Heather, & Pohl, Edward A. (2012). Improving healthcare supply chain processes via data standardization. Engineering Management Journal, 24, 3–10. U.S. News and World Report (2013). Best children's hospitals 2013–14. retrieved January 20, 2014, available from http://health.usnews.com/ best-hospitals/pediatric-rankings Walley, Paul, Silvester, Kate, & Mountford, Shaun (2006). Health-care process improvement decisions: A systems perspective. International Journal of Health Care Quality Assurance, 19, 93–104. Weed, Julie (2010). Factory efficiency comes to the hospital. The New York Times (retrieved July 15, 2015, available from http://www.nytimes.com/ 2010/07/11/business/11seattle.html?_r=0). Zheng, Kai, Haftel, Hilary M., Hirschl, Ronald B., O’Reilly, Michael, & Hanauer, David A. (2010). Quantifying the impact of health IT implementations on clinical workflow: A new methodological perspective. Journal of the American Medical Information Association, 17, 454–461.