Exploring Innovation and Quality Improvement in Health Care Micro-Systems: A Cross-Case Analysis (2000)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 17 TABLE 2 Micro-System Descriptions—Continued Hospital Unit Micro-Systems (n = 9) 1. We are a geriatric unit in a large medical center. 2. We are a Level III Intensive Care Nursery caring for intermediate and critically ill new- borns. It is staffed by a multidisciplinary team of neonatalogists, residents, NNPs, nurses, respiratory therapists, and others. 3. We are an Emergency Department with 10 docs, a slew of nurses, and other people. 4. We are a cardiothoracic surgical unit. 5. The Critical Care micro-system consists of 36 beds divided into the 12 bed Shock- Trauma-Respiratory ICU, the 16 bed Medical-Surgical ICU, and the eight bed Respiratory Special Care Unit. All are open ICUs. The hospital is a academic referral center for a 400 mile radius and a Level 1 Trauma Center. The system integrates the activities of five full time hospital employed academic critical care medicine (CCM) physicians along with six private practice pulmonary/ CCM physician with about 90 private staff physicians who admit and care for this population including the active Level 1 trauma and the neurosurgi- cal services. 6. Critical Care Services: MICU (10 beds), SICU (14 beds), CCU (10 beds (total = 34 beds), NICU, EC, and Critical Care Transport teams. 225 physicians, all specialties and subspe- cialties 7. We do only [one or two surgical procedures]. We have 11 surgeons, eight assistants. The entire staff is about 75. 8. We are five surgeons doing cardiothoracic surgery. Private practice. three partners, two associates. We work at the hospital with 12 mid-level PAs and NPs who were hired by the hospital. We have four secretarial office staff 9. We are a MICU and SICU. We have an open ICU—any physician with admitting privi- leges can admit to the ICU. As noted, not every respondent was asked every question because some questions were not relevant or were optional. Table 3 summarizes the interview completion rate. For each question, the table shows the number of sites asked the question and the completion rate for that question (calculated as the number of sites asked divided by 43). This is a very con- servative rate, however, because in responding to the open-ended questions, some respon- dents formulated a response to a later question before the interviewer had an opportunity to ask it. When this occurred, the information was recorded as part of the original question rather than breaking apart responses and inserting them into later sections. Responses that answered a later questions were frequent and are not reflected in this table. However, all re- sponses were incorporated into the final analysis. 

18 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS TABLE 3 Question Completion Rate Sites % Completion Interview Question Asked (sites asked / 43) Level of Performance 43 100% Success 42 98% Measures 28 65% Patient 37 86% Clinician 28 65% Culture 23 53% Professional 10 23% How long 22 51% Patient and Clinician Experience 41 95% New Patient 28 65% Scheduling 15 35% Risk assessment 17 40% Pt information 23 53% Unusual problems 24 56% Waits and delays 25 58% Incentives 9 21% Community 16 37% Information and IT 34 79% Improvement 40 93% Specific projects 28 65% Evidence of success 4 9% Barriers 26 60% Awareness of results 2 5% Funded projects 5 12% Leadership training 6 14% Expert systems 25 58% Clinical evidence 12 28% Best practices 15 35% Information sharing 6 14% Error and patient safety 21 49% What happens 21 49% Culture 3 7% Procedures 3 7% Sources of error 6 14% Leadership 42 98% Macro-system helps 19 44% Macro-system is toxic 17 40% Ideal financial structures 15 35% Replication 30 70% Barriers 23 53% 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 19 A contact summary sheet (top of Appendix D) was used to summarize each interview (Miles and Huberman 1994). Its purpose was to capture the primary issues that emerged from the interview shortly after it was completed. The contact summary sheet prompted the inter- viewer to think about the main issues that emerged during the interview and to identify ver- batim comments that illustrated them. This step was helpful in the transition from transcrib- ing notes to coding data because it engaged thinking about the analysis throughout the interview process and while the interview was fresh in mind, instead of waiting until the completion of all the interviews to begin analysis. Stage 3: Data Analysis Case-Level Summary Analysis of Health Care Micro-Systems Q.S.R. NUD*IST was selected as best suited for managing and organizing the data which comprised multiple cases but a single source from each case. Q.S.R. NUD*IST® 4.0 (Non numerical Unstructured Data Indexing Searching and Theorizing) is a multi-functional software system for the development, support and management of qualitative data analysis. Because this research was exploratory, it was important to be able to code and make coding revisions as the analysis progressed. This software facilitated the coding, sorting, and refin- ing of categories by creating logs of the changes that were made and allowing custom searching and retrieval of text. Transcribed interviews were entered as data to form display matrices. These can be thought of as meta-matrices, or master charts used to assemble multiple cases in a standard format.25 The objective is to include all the case-level data in one matrix before summarizing, refining, and further reducing the data. The matrices are considered to be “partially ordered” because very little order is imposed on the display of the data. The completed meta-matrices are the first look at the cross-case data. The creation of the matrices required identifying variables that were thought to be relevant to the study. To avoid imposing a rigid framework on the data early in the analysis, initially the interview questions were used as the relevant variables. For example, because each interview is coded by interview question, it is possible to find all the micro-system re- sponses to Question I.6. “If I were a patient at _____ , how would I experience the care dif- ferently?” Although the questions from the interview served as the initial relevant variables, additional variables emerged as the study progressed. Looking at the data by interview question is useful, but still represents raw data. For this reason, coding was then used to assign descriptive codes to each phrase, sentence, or groups of words that represent common concepts. This is called “first level coding.”26 Table 4 lists the variables that emerged from the transcripts and that were used for the first level coding of the interview data. 

20 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS TABLE 4 Micro-System Variables Variable Working Definition 1. Investment in Improvement An effort to ensure that improvement is part of the work of the micro-system. 2. Alignment of Roles and Training The match between a health professionals’ educational training, certification, etc. and their work. 3. Constancy of Purpose Integration of the aim throughout the micro-system. 4. Values A set of beliefs that guide the work of the micro-system 5. Organizational Support Ways the macro-system facilitates the work of the micro- system. 6. Multidisciplinary Team The existence and recognition of the team approach to care. 7. Community Connection Micro-system is a resource to the community/community is a resource to the micro-system. 8. Micro-system Measures Variables high-performing micro-systems are monitoring (or think are important to monitor). 9. Use of Information and Information is key, technology can be very helpful. Information Technology 10. Barriers Challenges and constraints to the work of the micro- system. 11. Resources for Replication Necessary elements to design and implement a similar micro-system. 12. Evidence of the Micro-System An indication that the site is a micro-system. 13. Improvement Example Examples of improvement projects made within the mi- cro-systems 14. Leadership Importance of leadership to the work of the micro-system Cross-Case Analysis Cross-case analysis involved searching each interview for examples. This was an it- erative process because themes emerged and evolved throughout the coding. As the analysis continued, the variables listed in Table 4 were refined—some were grouped into categories, and some were dropped because they did not rise to the status of a theme that could characterize the micro-system. For example, “barriers” was a common idea found throughout the interviews, but barriers are not a characteristic of micro-systems. How the micro-systems deal with barriers, perhaps through an investment in improvement or use of information and information technology, did appear to be characteristic of the micro-systems interviewed, however and were coded in this way. 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 21 Eight themes emerged that became a framework for thinking about characteristics of high performing micro-systems. Because the framework emerged during the analysis, it was necessary to return to the data and search each interview again and again to ensure that each interview was correctly coded. RESULTS Leaders of 43 micro-systems responded to questions grouped into five topics: level of performance, patient experience, information and information technology, improvement, and leadership. Analysis is presented in two parts. The first analysis (I) is a case-level summary of each of the five topics with samples of verbatim responses. The fifth and last topic in- cludes, in addition, some cross-cutting issues that respondents emphasized as they described requirements for replication of their accomplishments, what we call “lessons” for replication. The second part of the analysis (II) was a cross-case analysis based on eight themes that appeared to be associated with high performing micro-systems. Those themes are: inte- gration of information, measurement, interdependence of care team, supportiveness of the larger system, constancy of purpose, connection to community, investment in improvement, and alignment of roles and training. I. Case-Level Summaries by Topic Level of Performance To determine the level of performance of the micro-system, the first part of the inter- view asked respondents what their micro-system does very well and how they know this; that is, what data are being collected about performance. The majority of micro-systems (70 per- cent) identified taking care of specific types of patients (e.g., the frail elderly) or providing a specific type of care (e.g., women’s reproductive care or diabetes care) as what they do espe- cially well. Other areas that were identified are working as a team (14 percent), using informa- tion technology (12 percent), conducting research (7 percent), educating and training providers and staff (5 percent), improving access to care (5 percent), and communicating (1 percent). The connection between what the micro-system does very well and how micro- systems knew they did was so not so clear. Forty-nine percent of the micro-systems inter- viewed mentioned measuring their success by looking at clinical outcomes or some defined set of measures that includes clinical, functional, and financial indicators. Seven percent of the micro-systems cited measuring their micro-system performance against guidelines or protocols. For example, one micro-system tracks which protocols are being used, by how many physicians, and what percent of time. Forty-four percent of the micro-systems men- tioned measuring patient satisfaction, and seven percent of the micro-systems identified pro- vider satisfaction as an important indicator. Nine percent of the micro-systems identified benchmarking as a specific method for comparing their outcomes to others. However, one micro-system leader viewed benchmark- ing as potentially problematic: 

22 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS We measure success against ourselves. We try very hard not to measure against benchmarks. Benchmarks can limit you. Sometimes the benchmarking in and of itself becomes the goal. We do 1400 hearts a year. We should be the benchmark. Success to us is any incremental thing that makes us better than yesterday. . . . It is a mistake to benchmark pieces of your process against multiple other pieces of process. . . . Just keep working on little projects to improve what you are doing. Finally, nine percent of the micro-systems interviewed acknowledged that measuring and collecting data is difficult work. Other people use surveys and other ways to benchmark. We just do it seat-of- the-pants. We figure that we will get feedback. We don’t use any modern tech- niques to measure anything. It’s very expensive. We don’t have extra capital to invest in recreational data collection to prove how we are doing to some- one else when we know how we are doing. When thinking about the micro-system concept, a common question is, “How do we recognize a micro-system? Is it just another word for a team?” In consideration of these questions, respondents were asked to describe how a patient would experience care differ- ently in their micro-system. Similarly, respondents were asked how a clinician would experi- ence the micro-system differently from another micro-system that treats similar patients. Respondents reported most frequently that patients would perceive care differently be- cause of the level of information that the micro-system gives the patient. Respondents men- tioned, for example, making welcome calls to new enrollees, sending information about the services provided, and making sure the patient has a copy of the physician’s notes at the con- clusion of the appointment. One respondent noted that there are “no barriers to information.” Data on the measures we are monitoring are displayed on a wall—patients can see what the micro-system is working on. The level of information may include an increased use of information technology. Some micro-systems are communicating with patients by e-mail and referring patients to web sites for patient education. Other differences in the patient experience were a team approach to care and the focus on building a relationship with the patient and family. When asked how clinicians would experience the micro-system differently, one re- spondent said, “the clinical part is not that different—it’s the technology and the teams.” Other respondents indicated that technology has an increasingly significant role. However, one per- son articulated the importance of not confusing information with information technology: Frankly, all this stuff about how information systems are what has been hold- ing us backthat’s [nonsense]. Everyone is just waiting for some kind of cure-all information technology system instead of figuring out how to track things themselves. 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 23 Other differences that respondents pointed out as being different for clinicians in their micro-systems were standardization cross training of staff, and the infusion of improvement into daily work. To understand the culture of the micro-system, we asked respondents to describe the day-to-day work environment of their micro-system. Most comments discussed the impact of a team approach to care. There has been a radical change since we introduced teams. You can see it even where they hang out. Before the docs were together, the nurses together, etc. But now the team hangs out with the team. At the morning meetings, you may see the medical assistants providing the leadership. The medical director calls it the “fast break”—three people on the floor and anybody can finish the play. Other aspects of the culture mentioned by micro-system leaders were the freedom to make decisions regarding own work, an increased level of communication, and a commit- ment to improve. Respondents were asked whether their micro-system had made specific efforts to support professional ethics, encourage peer feedback or develop the skills of its members. Answers range from micro-systems that admitted “we haven’t done much” to one micro- system that has a full-time person who is responsible for organizing and leading sessions on the issues involved in working in teams. Other sites acknowledged the importance of this type of training, but lacked a systematic way of doing it, We try to do this through the course of our activities. But we don’t do it con- scientiously. It’s kind of on-the-job training. The final (optional) question in the “level of performance” section asked how long the micro-system had been working the way the respondent described it. Answers ranged from one year to “since 1945.” Of the respondents who were asked this question (n = 22), three sites reported more than ten years (16 years, 22 years, and 55 years). All the others re- ported less than ten years. Patient Experience In the second section of the interview, respondents were asked to describe the pa- tient’s experience in the micro-system. Specific questions asked about a new patient’s expe- rience, scheduling, risk assessment, referral, waits and delays, and patient education. These questions were designed to elicit information about innovations in delivery of care. Six of the sites have moved to an “open access” model, where patients are given an appointment to be seen the same day they call if they wish to come in that day. We assure that a patient can be seen that day if they can be seen by five; oth- erwise, the next day. That is not a big problem because phone calls to be seen 

24 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS that day drop way off in the afternoon. The primary focus has to be: “We are here for you.” Another comment from a micro-system with open access shows that they feel they have developed an approach that works well. In the old system, variation in quality was caused when patients went else- where to be seen (for example, an urgent care center) or when they gave up trying to be seen. Now the variation in quality is based on the doctors. In the first generation of open access people carve out slots based on predicted ur- gent care demand. But you need to move beyond this and dispel the myth of "needs vs. wants where wants are seen as unjustified demands. . . . The way to manage demand is over time, not with a call to a nurse. Other sites continue to carve out slots for urgent, same-day appointments but other- wise schedule future appointments. This does not appear to eliminate barriers to access and, as the following comment suggests, may not be the best solution for providers, patients, or the health care system in general. We have quick access, but not open access. We take care of anyone who just walks in, but we don’t advertise that. We try to triage based on urgency. Next available appointment slots may be a month out. The extenders have more open slots. The older, established MDs have a longer wait time for next available ap- pointment. We maintain 10 percent open slots for same day appointments. Other innovations in organizing and delivering care include building time into the daily work for teams to communicate, present cases, and learn from each other. Building in mechanisms for communication seems to be key to managing referrals, as well, and informa- tion technology can facilitate this communication. We started as a multi-specialty group. If I pick up a phone I can connect di- rectly to a specialist. This makes the transfer of care smooth. The Epic system generates referrals for non-urgent referrals. My notes go with the referral. It's the same method for getting information back to me. We are also connected via e-mail, and we do a fair amount of communicating this way. Many micro-system have specifically addressed the need to reduce waiting time and delays. They described improvements such as standard stocking of rooms, pulling up infor- mation about the patient before the visit, and adding a patient-flow facilitator to the team. We asked respondents how their patients get information about their health condition. Predominately, they reported that patient education is conveyed during one-on-one interac- tion with providers, via printed materials, videos, and classes. There appears to be an in- creasing level of comfort with technology and the integration of technology into patient edu- cation. Everett Rogers’s framework for studying the adoption of innovation can be applied to this phenomenon.27 Rogers’s findings from decades of research in the diffusion of innovation 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 25 demonstrate that the rate of adoption over time follows an S-shaped curve. During the early stages of an innovation, such as use of computer technology in providing patient education, there are relatively few adopters, but eventually more and more groups adopt it. Figure 1 shows Rogers’s model for diffusion of innovation overlaid with three examples from the mi- cro-system interviews regarding the current use of e-mail for patient education in these mi- cro-systems. Respondents were also asked about incentives that reward management and staff for meeting and exceeding patient expectations. Responses fall into three categories, (1) no in- centives, (2) incentives, and (3) misaligned incentives. The first is no incentive: The only reward is the knowledge that you are providing good personal care for each patient. Another micro-system discussed an Independent Development Plan (IDP) that recog- nizes successful efforts to improve with a raise in salary. We just started this year. Next year it will be mandatory to meet your IDP to get a raise. For example, one group wanted to improve patient satisfaction in their team. One team wanted to decrease supply costs. They cut supply costs by 28 percent. Two respondents mentioned incentives that appear to be misaligned; that is, the in- centives do not promote the functioning of the micro-system either because the incentive is not connected to the work of the micro-system or because the incentive is not given to all the people working in the micro-system. There are only incentives for high-level administrators to meet HEDIS meas- ures. Nothing filters down. If at the end of a quarter, there are savings from the unit, the money is split one third to the facility, one third to the health plan, and one third to the phy- sicians. The final question in the patient experience section asked respondents how the micro- system seeks input from the community or keeps the community aware of what the micro- system is doing. The micro-systems responded that they interact with the community at two levels—acting as a resource for the patient population and acting as a resource for other cli- nicians and health care providers by providing education and setting the standard of care in the community. Information and Information Technology Forty-nine percent of the sites included in this study indicated that patient records are paper based; 39 percent indicated that patient records and financial systems are computer based but separate; and 12 percent indicated that patient records and financial systems are to 

26 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 100% Late Adopter "We have printed materials for patients. We do very little with e-mail and are uncomfortable with it. We are afraid of missing something." Percent of Adoptions Early Adopter "We use health coaches to work with clinicians and patients. The smoking cessation program will test using e-mail interaction with patients." Innovator "The elderly are the most rapidly increasing computer users. I consider responding to e-mail part of my call-time." 0% Time FIGURE 1 The Diffusion of Innovation and the Use of E-Mail for Patient Education. some extent or entirely integrated. Computer-based records tended to be concentrated in hos- pital units. The majority (58 percent) of the sites interviewed were either linked or had access to pa- tient data from other units, such as laboratories, pharmacies, or the emergency department. Some micro-systems used computer-based information systems to generate reports about their practice (n = 15), to support real-time patient care (n = 12), or to support clinical deci- sions (n = 6). Only one micro-system indicated that the clinical information system includes direct data input by patients (patients use a touch pad to answer relevant questions and report on their health status), You would be given a touchpad computer when you come in for your visit for filling out all the intake information. Your picture would be taken digitally. All this would happen, and the doctor would see it, before you see the doctor. The doctor would explain your responses—e.g., what the SF-36 score means. Improvement The micro-systems provided rich examples of improvement projects. Respondents commented about what they had done to redesign services and to improve the quality of care and how they knew that these efforts were successful. Projects range from improving clinical 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 27 care (such as improving diabetes or asthma care), to improving administrative or service as- pects of care (such as scheduling or reducing delays). When asked about the context for quality improvement efforts, respondents frequently cited time and financial constraints and lack of organizational support for improvement. The amount of change in staff is huge. Staff changes are as frequent as every month. Second, building our team and dealing with the administration that deals with 20 physicians has also been tough. . . . One hand, they say “work as a team,” and on the other hand, they don’t let the team meet or work to- gether. . . . The last barrier is still having a paper-based medical record. This is the primary source of information. There is definitely a lag time before all the information is there. One site noted the need for improvement teams to be anchored in the micro-system. We did something wrong the first time. We created an ad hoc team to lower infection rates. They brought the [suggestions for] change back to the unit. The unit didn’t want to make the changes. The improvement team was “off- line.” . . . Our goal is to make a unit that creates improvements. Respondents were asked if the micro-system uses any guidelines, protocols, or expert systems to help clinicians get up-to-date information. Most micro-systems have guidelines and protocols in place and cited many instances of their adaptation, use, and development. However, most reported difficulty in integrating the guidelines and protocols into the daily work of the micro-system. There are a lot of guidelines in most institutions, but the way they are imple- mented destroys their usefulness. For example, the diabetes guidelines are 40 pages. As a physician, I look at them and decide on the two to three most im- portant things that should be done, and I work on getting those done consis- tently. Work on the others later. Even this is very hard to implement consis- tently. A few of the respondents mentioned formal benchmarking arrangements with other organizations. Overall, among the sites interviewed, there appears to be a lack of a formal mechanism for learning about best practices and for sharing new information. Patient Safety. With the recent publication of the Institute of Medicine’s report, To Err is Human,28 national attention has been focused on medical errors and patient safety. One part of the micro-systems interview asked participants to describe what happens in their micro- system when someone makes an error. It also asked about the major sources of error or harm, the extent to which there is a blame free culture, and procedures that had been implemented to improve patient safety. Medication errors and follow-up of abnormal lab results were the most frequently mentioned sources of error. Several respondents described formal mecha- 

28 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS nisms that were in place. For those without a formal mechanism, talking about errors ap- peared to be more difficult. It’s hard to talk about “error” because it is culturally not acceptable for fear of litigation. But we try. Several respondents recognize the importance of a systems approach to reducing errors. If something bad happens, it seems to me then the system has set the person up for failure. When you gather the data, it almost never is what it seems to be. We had a patient who wasn’t doing well. The physician ordered lidocaine. The nurse gave the patient a whole amp of epinephrine. We all thought “how stu- pid.” But when we started looking at the medications they were beside each other in almost identical boxes. Still, she shouldn’t have made the mistake, but you could see how it could happen the way we had things set up. The system can be an advocate. It can be a reminder that a mammogram needs to be done, that there is a system in place to make sure it happens, that things go well. A system can also empower the medical assistant to insist that a patient be seen, even if it means clashing with a provider. Leadership and Management: Lessons for Replication The last section of the interview was a set of questions related to leadership and repli- cation, including: the role of the macro-system, financial structures for payment and rewards that would be ideal for improving the quality of care, key factors to successthe key lessons for others who would like to replicate what had been done, the major barriers to replicating this elsewhere, and how barriers had been overcome. Most micro-systems function within a larger system, or “macro-system.” It is possible that micro-systems are successful only in certain organizational environments. On the other hand, a micro-system could fail because of its organizational environment. Respondents were asked to provide examples of helpful and toxic ways the macro-system affected the care they provided. Respondents provided examples of supportive macro-systems—that is, supportive in providing resources or in creating the environment or culture for the micro-system’s work. However, the tension between the micro- and macro-system was evident in other responses. They have been very supportive in terms of wanting to do cutting edge work. The priority for the system is patient care. They identified areas where CQI teams were needed. They supported us financially too. They have paid close attention to the results. They have identified ___ care as an area where they want a center of excellence. It is a priority of the system. The administration is a barrier. Sometimes I wish that they would just open the door and get out of the way. Respondents were asked to comment on what they would consider to be an ideal fi- nancial structure for improving the quality of care. Among the respondents asked this ques- 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 29 tion, a common response was to have some sort of capitated system, as suggested in the fol- lowing comment. To encourage improvement, you need a structure that makes you responsible for a defined population—some sort of capitated system. In a couple of sec- tions here, the payment scheme is fee-for-service—this makes people less in- volved in the team. The incentive is to maximize their own profits. This hurts improvement efforts. This section of the interview was of special interest to the IOM Committee on Quality of Health Care. Because it has been difficult to deploy improvements either within or across organizations,29,30 we were particularly interested in what micro-system respondents would tell us about how to replicate what they were accomplishing. Many of those interviewed expressed clear ideas about how they were reorganizing practices, their principles for doing so, and their commitment to an ongoing process. Respon- dents described their early limited successes or outright failures. We heard what had and had not been successful, If you can have those three things in place before you startthe right team, the senior leader support, and the financial issues resolvedyou can repli- cate what we have done. What we are doing is not undoable in other places. In many cases it’s just common sense. It is helpful to have a clear sense of goals, a philosophy of the service. Line everything else up with that. Funding must be aligned somehow to make the model possible. It is helpful to have some leaders who are in the micro-system all the time working on the administrative and organizational support of the model of care. We get visitors a lot. . . . They are interested in how everyone involved understands the goal of care, the high level of communication. Pro- ductivity expectations, but paid on salaries, are helpful for improvement. Plus recognition for those working on improvements. There isn’t a hierarchy of how much opinions are valued. Everyone’s opinions are valued. The meetings and care plans are done for a thought out reason. It isn’t by accident that this is how we got here. It would help to have supervision from someone who has done the model. Our model has been replicated. Mentoring has helped. There needs to be a connection over time—someone to talk to about difficulties and barriers as they occur. Talk it through with someone who has been there. It’s hard to set up a model just by reading about it. 

30 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS Respondent answers about requirements for replication could be grouped into six “lessons for replication”: 1. Executive and governance-level support for innovation and improvement efforts, 2. Strong, focused, and sustained clinical leadership, 3. Collaboratively functioning multidisciplinary clinical teams, 4. Explicit attention to the development of systems of care, 5. Good information systems that made measurement of their performance possible, and 6. A focus on the needs of patients. Sections below explain each of these factors and include tables with illustrative ver- batim comments. Executive and Governance-Level Support for Innovation and Improvement Efforts. Most micro-systems function within a larger system, or “macro-system.” In this study, the macro-system was generally a hospital or health plan. Micro-system leaders repeatedly cited the support of senior executive management as a sine qua non to their ability to succeed (Ta- ble 5). Respondents cited support of their efforts at innovation as critical, whether by setting direction, demonstrating interest and attending to the results, by providing financial and ad- ministrative resources, or the “space” for innovation despite sometimes strong external finan- cial pressure. Support included a willingness to set aside time for the micro-system leaders and members to work on improvement rather than their having to carve it out of other clinical responsibilities. In such micro-systems, the aim(s) of the micro-system are consistent with the aims of the larger “macro” system. Strong, Focused, and Sustained Clinical Leadership. A second area that emerged very clearly as leaders articulated how their work might be replicated was the need for strong clini- cal leadership at the micro-system level combined with a clear sense of purpose or aim that guides the work (Table 6). Respondents cited a high level of energy, focus, credibility, com- mitment, patience, and inclusiveness as important to success. Where aims are clear, they are communicated across micro-system boundaries, “When I walk in a room, everyone thinks, ‘diabetes.’ ” Collaboratively Functioning Multidisciplinary Clinical Teams. Respondents cited the importance of collaborative work both for clinical care and for improvement efforts (Table 7). They emphasized the need for quality improvement work to be based within the team, the need to recognize the contributions that all members of the group could make, with various members taking leadership roles for specific improvement activities. They also described new or expanded roles and the need for coaching and training new members of the micro- system in their work relationships. Explicit Attention to the Development of Systems of Care. Respondents often cited their micro-system’s investment in improving their systems of care, and they identified such work as being as critical to their micro-system as one-on-one patient care (Table 8). This investment 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 31 TABLE 5 Importance of Executive and Governance-level Support for Innovation and Improvement Efforts Type of Micro-System Comments Spine center They provided space, money, people, and a chance to make my vision a reality. Breast care center They have been very supportive of cutting edge work. They iden- tified areas where CQI teams were needed. They supported us financially, too. They have paid close attention to the results. They have identified breast care as an area where they want a center of excellence. It is a priority of the system. Advance care planning team We had the commitment from top administrators—the presidents from four systems set up the task force. The task force was to talk about ways to collaborate to improve healthcare. We set as a goal that at least 50 percent of adults in our community would have an advance care plan before a crisis and that the program we imple- mented to do this would be accepted by the community. The en- dorsement from the administrators made the task force much easier…. I met very little resistance. My organization, in particu- lar, put a lot of importance on this and asked me to put a lot of time in it. I wasn’t just asked to work it in to my other responsi- bilities. Ophthalmic center We can make changes quickly and are free to make investments and commit resources to change. We recently created a manage- ment services division here. We help other clinics and care sites to do marketing, quality improvement in patient flow, etc. The larger organization provided us with some resources to allow us to do this. Endoscopy unit The top leadership support must be thereif the CEO is directly obstructing you, just pack your bags and leave. Emergency department The hospital system has shown great effort in helping us out with patient restraint protocols. Restraint management has been an area where they have excelled and this has made the ER a safe place to work. They are also helping us out in quality end-of-life issues and identifying how cultural differences necessitate indi- vidualized care. Hospital cardiac care service The VP of Medical Staff has worked with the physicians. The Chief of Staff was supportive of disease management. The 12 chiefs work closely with our department. If they are given num- bers they don’t like, the VP of Medical Staff will not let them get by with that. They have to work to improve it. Newborn intensive care Someone at the leadership level has to be committed to good quality. You must keep the stimulus there to be the best. Leader- ship must think of ways to encourage, support change, and think of ways to change. Continued 

32 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS TABLE 5 Importance of Executive and Governance-level Support for Innovation and Improvement Efforts—Continued Type of Micro-System Comments Primary care There has to be a high degree of commitmentfrom administra- tion and also from someone willing to do the work, collect the data. There has to be a commitment of resources, both human and financial. Diabetic management Senior management support is critical because it consumes sys- program tem resources. At various times they have pushed back and said we were just doing individual quality improvement projects. We prevailed in saying that this is system-wide disease management, not just individual quality improvement projects. If you can have  three things in place before you startthe right team, the senior  leader support, and the financial issues resolvedyou can repli- cate what we have done. Treatment of severe mental They have a sense overall of an organization trying to learn, de- illness velop, and improve. They provide training for managers that places a high value on communication. If changes are made they are well advertised within the group. There is some interaction between micro-systems. We know what is going on in the other micro-systems. comes in the form of resources, such as time, money, and training, but also as an investment in creating the culture of the micro-system. For example, a respondent from a neonatal inten- sive care unit said, “We charged the entire operating structure of the unit with improvement.” Various approaches were described, including taking people “off-line” to focus on their proc- esses of care, standardizing techniques and protocols and measuring their effect, trying small scale (“rapid-cycle”) changes using small samples, and applying engineering concepts from other industries, such as continuous flow concepts for scheduling and care. Good Information Systems. We were interested in the extent to which electronic medical records, knowledge servers, decision support tools, continuous electronic patient-clinician communication, and consumer informatics had been incorporated into the work of the micro- systems. Respondents were asked both about clinical information systems to support individ- ual care (Table 9a) and systems to provide information about their performance to use as a basis of improvement (Table 9b). Although some micro-systems reported use of information technologies to support individual patient care—most community-based practices as well as much of clinical practice within hospitals did not have integrated data systems, knowledge servers, or decision support tools to use for real-time clinical practice nor for improving the quality of care for their patient population. 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 33 TABLE 6 Importance of Strong, Focused, and Sustained Clinical Leadership Type of Micro-System Comments Treatment of severe It is helpful to have a clear sense of goals, a philosophy of the service. mental illness Line everything else up with that. Integrated geriatric care You have to have energetic and powerful leadership that believes that this is the right thing to do. They really have to be willing to take this on as a missionunderstand and embrace it. First you have to believe in it. Then, you have to be committeda commitment to follow it through to the end, not just to get started. Hospital cardiac service You need to have the leadership in placehave the vision, be able to articulate it, and have the passion to carry it through. You also have to have a high level of credibility. Geriatric care, large Dedication, hard work, and patience to organize, implement, and stay medical center committed is vital. Neonatal intensive care Our culture was, “Of course babies get infections. They are not well to unit begin with.” But other sites saw an infection as a failure, not an enti- tlement. All the way to the bedside the unit knew that infection was a failure. That philosophy has to permeate the organization. Diabetic management What we do well is communicate the importance of diabetes care up group to the senior leaders of the organization; across to other providers’ and out, to the community. We are advocates for our own work. Whenever I walk into a room, people think “diabetes.” Primary care Our principle is that all of today’s work is done today. We have adopted the principle: If you call today, we will see you. If your own doctor is here, she’ll see you. Advance care planning The focus of this micro-system is improving advance care planning. team This is a joint effort of two [competitive] health systems. We assist and encourage adults to do advance care planning and then make sure written plans are available and followed. Intensive care unit An RN and I work as a team, almost one person. She has a unique ability to communicate with people like I’ve never seen before. She makes people enthusiastic and is able to interrelate with everyone. My strength is my credibility. Breast care center To replicate this model you need . . . agreement among whoever is involved that these are our common goals, processes, roles; a shared visionwe will need to change the system to get there Continued 

34 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS TABLE 6 Importance of Strong, Focused, and Sustained Clinical Leadership—Continued Type of Micro-System Comments Ophthalmic center Care givers need to be diligent, make fast changes. Many gains are only achievable with a leap of faith. Sometimes, it’s a lonely feeling to believe in quality improvement, but you try and make a difference by being persistent. A leader has to accept the insecurity and ambiguity that goes with the job. It takes guts to lead. Cardiac thoracic surgery You need to have good leadership. Without MDs as part of the leader- practice ship, you aren’t going to get anywhere. Quality improvement can’t be directed from administration. It has to start with the first step. For us it was agreeing to show up at the OR on time. Then we decided to work on something else. The biggest barrier is the first step. Spine center When things are successful it is because someone had a vision. I’ve watched what has happened to the program I started somewhere else. The longer I’ve been away, the more it has fallen apart. Computers can continue to work the same way, but people aren’t computers. They won’t work the same way once you walk away from them. You have to look for the person with the fire in their eyes. A lot of people want what we have here. You can provide the tools but only a handful of people will be able to do anything with it. I try to become unimpor- tantgive people the tools that will enable them. You have to enable the people around you to be successful. Some of the people will take it and make it better, but if you aren’t continuously improving it won’t work. Newborn intensive care Taking care of sick babiesthe quality of clinical care. The neonatol- ogy group has a commitment to being a resource to the region. We have a commitment to the health of a population. This is crucial to our success. As a resource, we provide education and review the quality of care for the whole region. A Focus on the Needs of Patients. Respondents often prefaced their comments about how their micro-system worked with descriptions about their aims in meeting a variety of patients needs (treating the “whole patient,” ensuring their dignity, the timeliness of services, attending to symptoms such as pain and to suffering, and making sure they have the informa- tion they need (Table 10). 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 35 TABLE 7 Importance of Collaboratively Functioning Multidisciplinary Clinical Teams Type of Micro-System Comments Primary care Other industries train and use people based on developing collabo- rative relationships. This is a particular problem for medicine and its fierce socialization process. It requires recognition, training, and a management philosophy. Geriatric unit I have a bias to the team approach. I am “content” oriented as op- posed to “process” oriented. The latter deals with who is in charge and who gets to speak, etc. The former depends on a team of profes- sional people who have various experiences and expertise. They re- spect each other and their opinion. Primary care The receptionist talks them through the systems of the office. They are trained to follow through specific areas of care such as screen- ing, childhood immunization, and antenatal care, so they have one person to contact. They have become expert in their areas. Intensive care unit If the Respiratory Therapist notes an abnormal lab value, she is com- fortable not just taking a blood sample and reporting it, but manag- ing it. The technicians are caregivers. Expectations have changed. They [adjust] therapy to within physiological parameters. They are cross-trained so that they can take on nursing tasks, for example, starting IVs when needed. When fully trained and confident, they may tell an admitting doc that a patient is not ready to have a venti- lator tube removed. Cardiac care unit We developed multidisciplinary roundseveryone involved in car- ing for the patient. The major value is having everyone communicate directly with one another. Each person knows they may be asked about the patients and has to be prepared. Geriatric unit in medical It is impossible for one individual to take care of an elderly person. center Older and frail people have many health needs that can only be met by a group of dedicated individuals. Ophthalmic center We believe strongly in team care. Staff satisfaction is very impor- tant. Everyone is not equal, but everyone is important and has a dif- ferent responsibility. I try to make sure that the clinicians know that working here requires a balance of getting to do what you want to do and of doing things as part of a team. Primary care We emphasize training medical assistants to a much higher level than most expect. We use two NPs extensively. Medical Assistants are trained in using technology, standardized triage functions, train- ing patients in self-management. We are trying to “push the enve- lope” and rely less on credentialing and more on continually devel- oping new skills. Continued 

36 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS TABLE 7 Importance of Collaboratively Functioning Multidisciplinary Clinical Teams— Continued Type of Micro-System Comments Primary care [The doctors] are worried about managing clinical conditions. They work under pressure and stress and try to find a way to control it. The myth is that they can control it with highly specified systems that raise barriers. They all claim that “my patients are sicker.” I re- ply, “Give me your sickest patientsthose with congestive heart failure, the ones on coumadin, patients with diabetes, hypertension, the old, sick people, anyone who seems to require more than the av- erage resources and time.” When they ask why I would say this, I reply, “Because I will enlist help, resourcesclinical pathways, care managers.” We provide these resources to the practice and should never charge [or penalize] the doctors for this help. The doctors have not learned yet how to enhance the team with other kinds of provid- ershealth education, behavioral medicine, physical therapy, phar- macy. Integrated geriatric care It has to be collaborative in nature. You have to find the people with the clinical competencies, but then train them. You have to train people to work in a different way. This is the only place you see true integration of acute and long term care. In the first three months af- ter hiring people we provide in-services on team work, resolving conflicts, working together. Cardiac care unit I’ve already mentioned the importance of support from high, senior management. It is critical. Second, is support of the nursing staff. They drive this, they are the core group who are there 24 hours a day. They are crucial to making change. Third, the doctors must be willing to give up some of their autonomy and to be a part of a team. You can’t bring someone in from outside to do this. It has to be someone who is there and well respected. Hospice The nurses aides are members of the team. Include them, listen to them. Treatment of severe mental It is helpful to have some leaders who are in the micro-system all the illness time working on the administrative and organizational support of the model of care. Everyone involved understands the goal of the care, and there is a high level of communication. Productivity expecta- tions, salary, plus recognition of those working on improvements are helpful. There isn’t a hierarchy of how much opinions are val- uedeveryone’s opinions are valued. The meetings and care plans are done for a thought out reason. It isn’t by accident that this is how we got here. 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 37 TABLE 8 Importance of Explicit Attention to the Development of Systems of Care Type of Micro-System Comments Behavioral health care In a given week we are spending about 100 person-hours on [the work of] teams. People are being paid to spend their time doing this, not just during their lunch hour. Someone said, “You have to assume you’ll be around here five years from now. Do you want to be doing things the same way?” Most of us don’t. This requires a new attitude that results in understanding that industries must in- vest in change in these micro-systems. You have to tolerate pull- ing people off-line to work on it. This is a radically new way of thinking in medicine which traditionally views any sort of meeting as a waste of time… [and that] the only useful time is spent seeing patients. I think that unless you spend time considering how to deliver care better, much of that time seeing patients is wasted. Primary care for underserved, The system can be an advocate. It can be a reminder that a mam- minority population mogram needs to be done, that there is a system in place to make sure it happens, that things go well. Primary care We need to train MDs in systems. They must have a sense of ac- countability and a sense of the patient-doctor relationship. Hospital endoscopy unit I try to help people understand that we can “work smarter.” You can feel rotten about how you are practicing. I tell them, “You are rightand it’s going to get worse.” But change is possible. We don’t need a billion-dollar solution. We need a billion $1 solu- tions. You have to create the will to change. There’s the will to change, then execution. Behavioral health Our philosophy is, Just Do It! A credible change agent is neces- sary. A change agent seems to be most effective if he/she is like the people he/she is trying to change. For some settings this means being a physician leader, but not an administrator. Find a partner to work with. They will push you and point out where you need to go. Allow the teams to do the work. Empower them to make change, spend money if necessary. Primary care for underserved It’s an incredible relief to try small changes on a small scale. It’s minority population so simple it’s brilliant. My time is dedicated to this. A dedicated person keeps everyone connected. The team makes use of the strengths of the individual team members. We had been managing indigent diabetic patients for years and didn’t think we could do any better. The providers believed that these people are so hard. But the patients responded to the changes we made. You have to craft something that is doable. You have to look for the simplicity in complex things. Continued 

38 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS TABLE 8 Importance of Explicit Attention to the Development of Systems of Care— Continued Type of Micro-System Comments Small hospital specializing in Although there are small, minor differences, every surgeon who a few procedures joins the staff, regardless of seniority, starts by assisting, then being assisted in 150 cases before being left on his own. If we are not completely confident he has mastered the technique, supervision is extended another 100 cases. The secret of success is in everyone using the same technique. The total cost is 50–60 percent of a gen- eral hospital. It decreases complications and is more cost effective. Women and newborn service It takes a major commitment to do what we are trying to do. It is very expensive. But once someone has done this, and there is a model out there of data driven quality improvement, the cost of replication will decrease. We have commercial vendors involved in some of our projects who will develop and sell these techniques. So, we are just one success story away. Emergency department There has been a process of radical reengineering around customer voice. There has been process improvement and rigorous cycle time analysis. The outcomes we measure include cost, quality of life, patient satisfaction. The quality of life is important not only for patients, but also for providers. We are able to show through our fast track program for less urgent patients that total time from beginning to end has dropped from 92 to 47 minutes. Cycle time between the arrival of a patient to a doctor seeing that patient has dropped from 32 to 18 minutes. The “decision to admit” on the floors of the hospital has dropped in cycle time from 3.5 hours to 1 hour. We have also reduced phar- macy cycle time. We have bedside registration. Each room receives a portable com- puter rolled in on a cart. Computer orders for lab and pharmacy are made from the bedside. In terms of clinical data, we have re- duced time for getting a lab result from 66 to 16 minutes. The reengineering approach included forming a task force. We needed a baseline measurement of how we were doing. We com- pared this to a registry which included state norms, hospital norms, etc. We then used a clever theoretical construct created by the NIH which centers around subintervals. We borrowed the 4Ds concept: “door, data, decision, delivery.” We introduced multi- processing or “parallel processing.” We looked at the four sub-intervals to see where we could improve care. Using parallel processing, we have empowered and educated our nurses. If “data” are needed for a “decision” to be made, a nurse can go ahead and order an EKG, CBC, or chest x-ray. We have done a similar thing with antibiotic prescription and care for patients with pneumonia. Continued 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 39 l TABLE 8 Importance of Explicit Attention to the Development of Systems of Care— Continued Type of Micro-System Comments Emergency department— Our pain management program is trying to increase the team’s Continued sensitivity to the quality of care at the end of life. We have also focused on stroke management, noise management in the hospital, etc. We have embraced the concept of “real time tracking.” We have developed a radar screen that has 8 simultaneous processes con- tinuously monitored. Each process is depicted in 15 minutes cut of data for the last 4 hours. We get information on the census in the ER, the status of the patients, the x-ray cycle, etc. We know where in the process not only the patient is, but where the system is. Each process measured is summarized on the screen by graphs. All we have to do to obtain data is touch the screen. The graphs are equipped with goal lines, not control lines, but goal lines that are based on some sort of customer satisfaction, e.g., people don’t like to wait to long, etc. The micro-system is less like a chassis and more like the human body. The key word to describe a micro-system is homeostasis. A micro-system is always changing and adapting, just like the human body. The most exciting thing I can tell you is that we have identi- fied the “pathophysiology” of a micro-system. It is powerful and yet very predictable. Think about two downstream processes, x- ray cycle time and getting patients to the floor. If the downstream graphs go out of control, there are predictable changes in the sys- tem. Occupancy in the ER goes up, the number of new patients in the ER goes down. The number of free beds in the ER goes down, and the cycle time between arrival to a bed goes up for a patient. Eventually, every measurement goes up. What is the intervention? A series of algorithms built into people’s behavior. When we ob- tain three consecutive 15 minute intervals going in the wrong way, we realize that something needs to be done. Other micro-systems use a 1-size-fits-all approach with monthly quality improvement meetings or something similar. We try to intervene early. For x-ray cycle time, we dropped from 72 to 23 minutes. We reengineered processes so that the ER docs see x-rays first, that old x-rays are quickly taken away, that twice as many x-rays and techs are present in the ER. When we are very busy, our x-ray techs call up other techs off duty at home and tell them to come in. They do it automatically, without asking man- agement. We say, “There are three ways of responding and reengineering: a bad way, a good way, and a world-class way.” 

40 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS TABLE 9A Importance of Good Information Systems for Individual Patient Care Type of Micro-System Comments Diabetic management I can show diabetics a graph of their HgA1-C and comment on how it has dropped along with their weight which is graphed on the same screen. I can also refer them to web sites, for example, if they are interested in alternative care, acupuncture, asthma management. One thing I have been concerned about is how to communicate using the computer without losing contact [while you put information into the computer]. By having the medical assistant enter the information, I can invite them to tell the whole story, and I can listen, so it actually increases communication. Primary Care When a patient comes to the office with a new problem (say headache), he/she would be handled in a standardized way. He or she is given an extensive questionnaire using the Knowledge Coupler®. The medical assistant takes them through all the steps. When I come in, almost everything has been done, but the patient is invited to tell his or her story again. I don’t need to take a lot of notes but can embellish on what is there. I can listen. We can then go over the options for care, looking at the screen together. I share the degree of uncertainty I feel. The patient leaves with a copy of my note. At that point, all the work is done. There is no dictation to be done, and I have had time to deal with the problem. I also explain to the patient that we will need a more comprehen- sive database that includes information about their health habits, family history. The patient returns for this, and we create a prob- lem list. The important thing about this whole process is that it is standardized. I use the same rational approach for each patient and don’t prematurely reach conclusions or forget to ask or rec- ord some things. I’ve been using what is close to a paperless EMR since 1993. We continually look for ways to use technology to help us become more sophisticated and integrated. . . The electronic medical rec- ord (EMR) does drug-drug interaction alerts. When the patient leaves the office, he/she gets a printout of their medication list. We try to make maximum use of information technology. We are trying to create as paperless an office as possible. We have sum- maries of patient records (problems, current medications, consult records) that can be called up on laptops for remote access. 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 41 TABLE 9B Importance of Good Information Systems for Improving Care Type of Micro-System Comments Emergency Department We have developed a radar screen that has eight simultaneous pro- cesses continuously monitored. Each process is depicted in 15 min- utes cut of data for the last four hours. We know where in the proc- ess not only the patient is, but where the system is. Each process measured is summarized on the screen by graphs. All we have to do to obtain data is touch the screen. When we obtain three consecu- tive 15 minute intervals going in the wrong way, we realize that something needs to be done. Spine center We use a value compass. We can query a database at any time for individual patients, but also for all patients we serve. We are also hooked up to 26 other centers. We can look at data by the point of service or longitudinally. We measure functional status, health status, work measures, treatment, who you have seen (type of pro- vider), age, sex, height, weight, SF36, satisfaction, clinical comor- bidities, smoking, cost of lost work over time. Primary care The development of an instrument panel of measures has been very important, then feeding this back to the staff has really stimulated our thinking. Emergency department We can track process length through our real time “flight simulator” system. By touching the screen, we instantly know such things as arrival to bed, bed to nurse, arrival to doctor, aggregated cycle times. Cardiac care unit Databases are importantyou have to make that investment. We communicate regularly and give progress reports. You have to make sure that you keep the data concurrent. Then the internal re- sources must be in placethe statisticians, the people who are working with the data. An electronic medical record would cut down on the need for some of the databases that we’ve bought. OB-Gyn practice Work with providers who are very interested in evidence-based outcomes, look at results, and apply results using continous quality improvement. You have to look at the most important outcomes. Analyze something important to you and important to patients, too. Primary care Information management has been the lubricant to improvement. I think that is key to our success. We have a seamless flow of com- munication. Our information system has allowed us to move through many barriers. Breast care center You need population-based clinical information systems with data. We were lucky that there was enough money in our health care system back then to create an automated clinical data system. In 1985 , the leadership put in the personnel, the money, some vision, and created something good. Continued 

42 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS TABLE 9B Importance of Good Information Systems for Improving Care—Continued Type of Micro-System Comments Critical care unit The bottom line is risk-adjusted mortality. The centerpiece of this is participation in Project IMPACT, a national database supported by the Society for Critical Care Medicine. It uses 3 predictive models of mortality: MPM, APACHE-II, and Simple Acute Physiology Scores. All three use physiologic parameters to generate scores of likely mortality. We do a quarterly download to compare them- selves over time and to other, similar institutions. The database produces 4-quadrant scatter grams of their patients with predicted mortality on one axis and resource consumption on the other. When we began we included 100 percent of patients. Now we are satisfied with the internal validity and track a 50 percent random sample. We track mortality, admission and discharge rates, LOS, readmit to ICU and reintubation rates. With the pressure to move patients out of the CCU, this helps us know if changes that affect efficiency are affecting quality of care. Although admissions are up and the LOS down significantly, our reintubation rate is very low. Thus [we know that] increased through put is not adversely affecting patients. Cardiac care unit We use two data bases: one of the Society of Thoracic Surgeons and another of the American College of Cardiac Interventions. We used to look at care case-by-case. Now we look at the aggregate data to determine where there is room for improvement based on the benchmark. We look at a group of cases and identify patterns. For example, we looked at emergent patients with CABG [coro- nary artery bypass grafts] following angioplasty. We abstracted the charts and created a verbal summary by practitioner. We also dis- cussed the results in the M&M [morbidity and mortality rounds]. We collect data on which protocols are being used, by how many physicians, and what percentage of time. We also collect data on outcomes, such as how well we are able to control glucose levels. We give quarterly reports to the Chiefs of Surgery Cardiology. Our rates have improved dramatically. CABG has decreased 50 per- cent; PTCA [percutaneous transluminal coronary angioplastyl complications have decreased by 75 percent; return to the operat- ing room following CABG has decreased by 50 percent. We do a utilization reports and variance reports. We can drill down to look at financial data. Breast care center We have created a generic model regardless of the specific disease. 1) assess the population, 2) stratify the riskwho do we focus on first, 3) assess the individual, 4) set goals and develop a care plan, 5) deliver and coordinate care, and 6) monitor and evaluate care. For each of these steps we have had to identify the roles of the care team. We have found that the roles of the care team may have to change. Continued 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 43 TABLE 9B Importance of Good Information Systems for Improving Care—Continued Type of Micro-System Comments Newborn intensive care How do we do against some comparison? We participate in a re- gional network. There are 300 participantseveryone contributes data. We can compare how we do with very low birth weight (<1,500 g) babies. We can compare our outcomes to similar insti- tutions such as other level III nurseries in a teaching institution, and outcomes are adjusted. Primary care We have an information technology project team. It includes re- ceptionists, nurses, and others who are involved. Formerly we tracked morbidity and mortality in detail using 10-year’s accumu- lated statistics. This was condition-specific—for example, neuro- logical conditions, cardiovascular disease. These data were col- lected automatically for the national database and allowed comparisons among practices. 

44 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS TABLE 10 Importance of a Focus on the Needs of Patients Type of Micro-System Comments Primary care In health care, what we sell is a relationship. But then [when patients want to be seen] we put up a barrier in the form of “we think you’ll get better if you just wait.” If they come in for what we think is an “inap- propriate” appointment, so what? First, they’ll find a way to get in anyway. Second, it destroys the relationship. Third, it is an opportunity to do other things—preventive care, to explain how they might handle the problem themselves the next time, and an invitation to them to call me. Build systems around what people want, and you can’t lose. . . . Patients want a relationship. They want someone whom they can trust. When you try to “manage demand” you teach them not to trust you. Hospice I’ll tell you what is critical: that the CEO focuses on patient needs and expectations. That is fundamental to what is important to methat the focus be on the individuala complex person. You try to do the best you can for them. It seems odd to say, but that is what is fun. The rest is just dials. We did focus groups with families and learned that four key things are important (1) the organization and delivery of care; (2) shared medical decisionmaking; (3) treating each person as an individ- ual; and (4) attending to those who care for and love the dying person. Diabetic management You have to educate the patient, then let them work through the proc- program ess. When one person is truly present to another person, something happens. I remember being with a patient one day and connecting with that person. I thought, “Take off your shoes you are sitting on holy ground.” If you are not there for the patient, you might as well be a technician. Primary care We take seriously the whole patient. We see our role as primary care. A problem isn’t solved until the patient agrees that it is. Newborn intensive care The preconception of NICUs is highly technical and that families aren’t (NICU) part of it. We want to astound themfull participation of familiesno barriers to access, no barriers to information. Diabetic management Patients are treated with dignity. We’ve changed the mindset. We’ve made them realize that they are in charge. Traditionally, a patient would come in, the doctor would say, “You need to lose 50 pounds and have a blood sugar level of 110.” The patient would leave, feel at fault, and a wall would go up. Now I tell people, “No one can ever fool you about your diabetes again.” Heart failure manage- We treat the patient as a whole. We look at more than just the cardio- ment team vascular problem. When a person comes in, all organ systems are checked. We talk to the patients about psychological and social support. We carry many patients to end-of-life care. We are with them until hos- pice care and sometimes even beyond hospice. We tell them about du- rable power of attorney, medications, shopping, eating less saturated Continued 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 45 TABLE 10 Importance of a Focus on the Needs of Patients—Continued Type of Micro-System Comments Heart failure manage- fat, increasing activity, the importance of family, independence, etc. We ment team—Continued do all of this during the first visit. We always put things in writing or print it out for them. We highlight key words and phrases, like what an ACE inhibitor is supposed to do. We explain what is happening to them and what the medications will do in “laymen’s” terms. We are in touch with all the patients. The secretary calls all the patients at least once a month. We ask basic questions during this phone call. When patients come in, they get a sheet of questions we want them to answer as well as their recommendations on how we can improve the care we give them. We ask them questions like, “Do you understand your medica- tions?” or “Do you feel like you are in control?” or “Are you comfort- able with what is happening in your life?” or “Are you a source of hap- piness to your friends?” Breast care center In 1990, a group of clinicians met to improve diagnostics of breast screening. At that time it took about a month for follow-up of abnormal results on a mammogram. . We identified “sleepless nights” as what we wanted to improve. We started streamlining the process. We got to- gether primary care, radiology, and surgery. We had physicians and nurses from different areas. The team decreased the process to a few dayswe went from 2 to 4 weeks (from abnormal test result) to 3–7 days, on average. Pain management At first, the nurses didn’t want the pain scales in the room because they service thought that it would be worse for the patient if we brought it to their attention, but we know that just isn’t the case. We graph pain on the vital sign sheet just below temperature. We have a place on the vital sign sheet to document pain and whether the pain management is effec- tive. But really you have to listen to patients, to have a conversation with the patient about what level of pain is acceptable. A post-surgery patient should be able to breathe deeply and get up and walk and do more for themselves each day. A terminally-ill patient should be able to eat and visit with people. When a person has pain that is a 5 or more we have to talk with them to understand what that means. The nurse is learning and the patient is learning too that this is not about how much pain can you stand. Primary care We measure success from the patient perspective as the match rate the likelihood that a patient see his/her own doctor versus a teammate, an NP, or is diverted to an ER. When we began it was 47 percent. Now it is 75 percent. Given that the average doctor is only in the center 72 percent of the time, this is terrific. We decided to rebuild the system based on what patients want. We learned: 1. Patients want to choose their own primary care doctor. 2. They want access to that doctor. 3. They want to be treated with dignity and respect-which means not having to wait all day. Continued 

46 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS TABLE 10 Importance of a Focus on the Needs of Patients—Continued Type of Micro-System Comments Primary care— Patients don’t want to go to an urgent care clinic. When they are sick Continued they want to see “my doctor.” The reason they needed to go to the ur- gent care clinic was because of the backlog of appointments which made it increasingly unlikely they would see their doctor. I call this the “urgent care death spiral.” The only way to solve this is to get rid of the backlog. We knew that the number of patients seen daily were about the same as those calling to be seen. But of those who called, some were seen the same day as an urgent visit, and those who could wait were given routine appointments. Now we adopted the principle: If you call today, we will see you. If your own doctor is here, she’ll see you. We closed the urgent care clinics and distributed the urgent care doctors to the various offices. We decided it was a big mistake to divide people into the streams: well, acute, and chronic because: a) the patient doesn’t see him/herself that waythey divide themselves by their doctor; b) wellness, acute illness, and chronic care are dynamic-needs. All three exist at various times and often simultaneously; c) it is a waste of time to try to get them into the right category, and we don’t get it right anyway; d) it increases work in the system because all the urgent care clinic does is acute care and they have to make another appointment for chronic and wellness care; and e) it turns nurses and appointment staff into antagonists of patients who have to fit into the correct category of urgent or not be allowed to come in. In the old system, we had a steady state-constant input and output and a “lake of waiting” in the middle. This system eliminated the “lake.” We used a carrot and stick approach: the carrot: you get to take care of your own patients. The stick: You have to take care of your own pa- tients. The number of patient visits went down by 8 percent. Using this system all our preventive care numbers went up—pneumovax, pap smears, mammograms. 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 47 II. Cross-Case Analysis—Themes Related to Effective Micro-System Performance Using cross-case analysis, we identified several common themes that appeared re- peatedly during the interviews. Each micro-system can be arrayed on each scale along a spectrum of “low” to “high.” Table 11 lists each factor and the percentage of micro-systems for which that theme was raised during the interview. In the table, the themes are arranged from the highest to the lowest percentage of micro-systems in which they appeared. Integration of Information Micro-systems vary on how well information from a variety of sources is integrated into the daily work of the micro-system and the role that technology plays in integrating in- formation among clinician, patients, and other members of the micro-system. Some micro- systems have developed advanced systems, but providing useful, timely, and accurate infor- mation is a huge task for micro-systems, and even in these high performing micro-systems, the potential of information technology has not, for the most part, yet been tapped. If you were a patient you would experience care differently here compared to the care you might receive elsewhere. You would be given a touchpad com- puter when you come in for your visit for filling out all the intake information. Your picture would be taken digitally. All this would happen, and I would see it, before you see me. I would explain what your responses mean. Most of the information is there; you have to find a way to harness it. Really all that is needed is a simple system to get back information quickly. Comput- ers, lines, high tech come to mind but it doesn’t have to be that way. Talking is a way to communicate too. Information technology doesn’t have to be an elaborate system. Table 12 provides several verbatim responses from the interviews that illustrate low and high levels of integration of information. Measurement Effective micro-systems measure their own performance and use that information to modify the care of individual patients and their processes of care. Part of the work of the mi- cro-system is the development of a set of measures that are appropriate for its own goals. Although all the micro-system are measuring some outcomes, many lack measures that are useful for their daily work. As one Table 10 respondent said, “At the local level I don’t get the measures that I need and the measures that I get at the regional level aren’t at the level I need.” It may be that recognition of the need and type of useful measures and finding ways to gather those data are important for high performance. Table 13 provides examples of low and high levels of measurement. 

48 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS TABLE 11 Summary of Micro-System Framework Integration of information Low High Information free environment Information is key, technology may be very helpful “We don’t have control over the information that we “I can show diabetics a graph of their HgA1c and need.” comment on how it has dropped along with their weight which is graphed on the same screen.” Measurement Low High Absence of a set of useful measures Micro-system routinely measures processes and outcomes, feeds data back to providers, makes changes based on data “We have data on demograhics and length of stay, “We have developed a radar screen that has eight si- however, we don’t have data on outcomes of care.” multaneous processes continuously monitored.” Interdependence of care team Low High Providers and staff function as individuals, Care provided by a multidisciplinary team, No clear way of sharing information or communicating Information is key to the relationship “Often physicians have difficulty working with “We developed multidisciplinary rounds—everyone non-physician providers, giving them the control.” involved in caring for the patient.” Supportiveness of the larger system Low High Larger organization’s actions Micro-system views larger organization as helpful perceived as “toxic” to the micro-system “If we have to practice like the rest of the system, “They have identified breast care as an area where we feel that we’ll be practicing ‘mediocre’ care.” they want a center of excellence. It is a priority of the system.” Constancy of purpose Low High Lack of a clear, consistent aim Integration of the aim throughout the micro-system “The original aim was that we would practice the “Those other sites saw an infection as a failure, not best medicine we could, understanding that we entitlement. All the way to the bedside the unit knew couldn’t be as financially successful. Now some of that infection was a failure. The philosophy has to the physicians are compromising for the financial permeate the organization.” aspects.” Connection to community Low High No clear connection to community Micro-system is a resource to the community, beyond current patient population community is a resource to the micro-system “The only way we get information about the com- “I invite the peer support groups that are in the com- munity is from the managed care organization.” munity to educate the residents.” Investment in improvement Low High Training, resources not available Resources made available for improvement (training, $$, time) “We don’t know how to improve the system. We “The Quality Council’s goal will be to provide guid- have closets full of good ideas but don’t know how ance and facilitation. ‘Yes, that project meets our over- to implement them.” all goals, what resources do you need?’” Alignment of role and training Low High Health professionals not expected to work within Health professionals expected to work the limits of their education, certification(overqualified) at the upper limits of education, training “I want to be more involved in the care process.” “When fully trained and confident they may tell an admitting doc that a patient is not ready to have a ven- tilator tube removed.” 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 49 TABLE 12 Micro-System Examples of Integration of Information Integration of information Low High Information free environment Information is key, Technology may be helpful “We don’t have control over the information that “I can show diabetics a graph of their HgA1-C and we need. We need to be able to define who our comment on how it has dropped along with their weight panels are — we can’t do that ourselves. Control of which is graphed on the same screen. I can also refer information is a barrier. Change will be more rapid them to web sites, for example, if they are interested in in the teams as we have more control over the in- alternative care, acupuncture, asthma management. One formation.” thing I have been concerned about is how to communi- cate using the computer without losing contact [while you put information into the computer]. By having the medical assistant enter the information, I can invite them to tell the whole story, and I can listen, so it actually increases communication.” “If you aren’t going to have the same nurse work- “The team that takes care of patients is a working ing with the patient then you have to have better group that meets daily for 45-60 minutes. We discuss communication. Patients get the best care when you the status of all the patients and we brainstorm treat- have health care workers who communicate very ments as well as discharge planning there. All patients well and collaborate very well. One of the biggest are listed on this blackboard that is used to organize problems I see is physicians not talking to each information on the care process for each of the pa- other. Also, so many nurses work part-time, vary- tients.” ing shifts. We struggle with getting them to com- municate. It’s hard to get them to put equal empha- sis on communicating, documenting, teaching and the physical tasks that need to be done before the end of the shift. You don’t get the same negative feedback from your coworkers if you aren’t teach- ing the patient as you do if you leave some of the physical tasks undone at the end of the shift. A nurse will prioritize and get every thing done before the end of the shift, but they don’t look at the pa- tient’s care plan and do the teaching that needs to be done before discharge.” “At 7 p.m. one evening a person was giving care to a “Sharing information with patients is the biggest safe- patient in a hospital who was receiving cancer treat- guard (against medical error). The electronic medical ment. The patient wanted an advance directive—if record (EMR) does drug-drug interaction alerts. When my heart stops, I don’t want CPR. The person told the patient leaves the office, he/she gets a printout of the nurse at the unit desk about this request and their medication list. Once in a while a patient will call asked that the nurse please tell the doctor. The doctor later and say, ‘I was looking over the list, and I am not never heard this. At 6 am the next morning, the pa- taking x anymore, but Dr. So and So has put me on y.’ tient had a cardiac arrest and a code was called. 20 It takes all of us. Another safeguard is that the system minutes into a code the request was seen in the pa- we use forces me to consider all the possibilities. For tient’s record that the patient didn’t want this to hap- example, if a patient comes in with headaches and pen. We saw that there was not a clear responsibility vomiting, it has a structured sequence that makes you to report the request to the nurse, to report to the consider the causes, including cerebral hemorrhage.” MD. The physician always decides whether an order will be written or whether to go talk to the patient before writing the order. The system worked a lot of the time, but it wasn’t consistent.” 

50 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS Interdependence of the Care Team Members One element of a micro-system is the key playersthe practitioners and staff who work together on a daily basis, but like the other features examined, the interdependence of the care team varied across micro-systems. Table 14 provides examples of interdependence of members of the care team. In sites with a high degree of interdependence, the existence and recognition of the importance of the team approach to care was evident in the interviews. We developed multidisciplinary roundseveryone involved in caring for the patient. The major value is having everyone communicate directly with one another. Each person knows they may be asked about the patient and has to be prepared. We believe strongly that in team care, staff satisfaction is very important. Eve- ryone is not equal, but everyone is important and has a different responsibil- ity. I try to make sure that the clinicians know that working here requires a balance of getting to do what you want to do and of doing things as part of a team. Supportiveness of the Larger System Supportiveness of the larger (“macro”) system overlaps with other factors. In high performing micro-systems, the aim(s) of the micro-system is consistent with the aim(s) of the larger system. The interviews made clear that the larger system demonstrates that improve- ment is a priority by making the necessary resources available to the micro-system. Even though there is overlap with other factors, it is important to recognize the importance of the larger system for the success of the micro-system. Table 15 provides examples of such sup- portiveness by the larger. We can make changes quickly and are free to make investments and commit resources to change. We recently created a management services division here. We help other clinics and care sites to do marketing, quality improve- ment in patient flow, etc. This is our entrepreneurial spirit. The larger organi- zation provided us with some resources to allow us to do this. The hospital system has shown great effort in helping us out with patient re- straint protocols. Restraint management has been an area where they have excelled and this has made the ER a safe place to work. They are also helping us out in quality end-of-life issues and how cultural differences of people ne- cessitate individualized care. Lack of support and shared aims was cited even by some high-performing micro- systems as an impediment to the micro-system effectiveness. It is a mixed message. The organization talks about team care but then sub- verts their vision. They put in a centralized phone system with a nurse in charge of scheduling appointments. Well, she has no way of knowing whether Doctor X and Y are on the same team. 

INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS 51 TABLE 13 Micro-System Examples of Measurement Measurement Low High Absence of a set of useful measures Micro-system routinely measures processes and outcomes, feeds data back to providers, makes changes based on data “I think we are deficient in measuring. We are “We have developed a radar screen that has eight measuring the more global outcomes.” simultaneous processes continuously monitored. Each process is depicted in 15 minutes cut of data for the last four hours. We know where in the pro- cess not only the patient is, but where the system is. Each process measured is summarized on the screen by graphs. All we have to do to obtain data is touch the screen. When we obtain three con- secutive 15 minute intervals going in the wrong way, we realize that something needs to be done.” “When it comes to collecting raw data, we have “We use a value compass. We can query a data- found it to be difficult. We have data on demo- base at any time for individual patients, but also graphics, and length of stay, however we don’t for all patients we serve. We are also hooked up to have data on outcomes of care. This will come 26 other centers. We can look at data by the point soon in the future.” of service or longitudinally. We measure func- tional status, health status, work measures, treat- ment, who you have seen (type of provider), age, sex, height, weight, SF36, satisfaction, clinical comorbidities, smoking, cost of lost work over time.” “Other people use surveys and other ways to “We track our endpoints extensively and have benchmark. We just do it seat-of-the-pants. We been able to do 3-yr follow-up of 75-85% of pa- figure that we will get feedback. We don’t use any tients. We have an annual banquet in January and modern techniques to measure anything. It’s very invite all former patients to come. 80% of those expensive. We don’t have extra capital to invest in whose surgery was in the last 2 years come to this recreational data collection to prove how we are banquet. We book a large hotel, and they are our doing to someone else when we know how we are guests. It is social but also an opportunity to do a doing.” follow-up check. We have 15 doctors doing ex- ams. 700-800 people generally come. There is a lot of camaraderie among patients.” “Every physician says they practice excellent “The development of an instrument panel of medicine, but you have to look at some other pa- measures has been very important, then feeding rameters. We look at HEDIS and NCQA. It’s this back to the staff has really stimulated our hard to look at other outcomes—no one knows thinking.” how to do that.” “There was a problem with how to track it [data “We can track process length through our real time about meeting open access goals]. There were ‘flight simulator’ system. By touching the screen, problems because the physicians weren’t getting we instantly know such things as arrival to bed, feedback on time about how they were doing bed to nurse, arrival to doctor aggregated cycle working down the backlog and meeting open ac- times.” cess goals. Then the MDs wouldn’t get the incen- tive because they hadn’t met the goals.” 

52 INNOVATION AND QUALITY IMPROVEMENT IN MICRO-SYSTEMS TABLE 14 Micro-System Examples of Interdependence of Care Team Interdependence of care team Low High Providers and staff Care provided by a function as individuals multidisciplinary team No clear way of sharing information or communicating Information is key to the relationship “Often physicians have difficulty working with “We developed multidisciplinary rounds—every- non-physician providers, giving them the control. one involved in caring for the patient. The major Some physicians don’t do well sharing responsi- value is having everyone communicate directly bility for patient care like this.” with one another. Each person knows they may be asked about the patients and has to be prepared.” “It’s always hard when we get new clinicians. “It is impossible for one individual to take care of They aren’t used to working with parapro- an elderly person. Older and frail people have fessionals in the community. We try to illustrate many health needs that can only be met by a what works. Doctors focus on what they do in the group of dedicated individuals.” exam room but that’s not enough.” “Finally, not all doctors like the interdisciplinary “There are just the three of us. We work very well philosophy. They like to do whatever they want” together. M. is in charge of the office, I am in charge of the patients, and Dr. D. is the physician champion. He holds the key to resources and new patients.” “We created a phone center to handle problems “We believe strongly that in team care, staff satis- with phone access. We have six people answering faction is very important. Everyone is not equal, phones. I saw it as decentralization and didn’t like but everyone is important and has a different re- that idea for the micro-system concept. My phone sponsibility. I try to make sure that the clinicians nurse knows my patients—she knows when a pa- know that working here requires a balance of get- tient really needs 20 minutes instead of 10. This ting to do what you want to do and of doing has been borne out with the phone center and it is things as part of a team.” still hard to get through [on the phone].” Constancy of Purpose A theme that emerged clearly in these interviews was the importance of “constancy of purpose,” or aim that guides the work of the micro-system. As Table 16 illustrates, where aims are clear, they are communicated across micro-system boundaries. In contrast, lack of a clear consistent aim may be destructive of the micro-system and, ultimately, of patient care. One respondent discussed the change in aims that his micro-system had undergone. The thing that distinguished those places that are achieving excellence is the organizational culture. Our culture [used to be], “Of course babies get infec- tions. They are not well to begin with.” But other sites saw an infection as a failure, not an entitlement. All the way to the bedside the unit knew that infec- tion was a failure. That philosophy has to permeate the organization. What we do well is communicate the importance of diabetes care up to the senior leaders of the organization; across to other providers’ and out, to the community. We are advocates for our own work. Whenever I walk into a room, people think “diabetes.”