Staffing and Quality of Care in Hospitals
Nursing is a critical factor in determining the quality of care in hospitals and the nature of patient outcomes. Twenty-four hour nursing care is one of the distinctive hallmarks of inpatient care in hospitals. Historically, hospitals have been at the core of the U.S. health care system, and nursing services are central to the provision of hospital care. They have also functioned as the traditional place of work for nursing personnel and especially for registered nurses (RN). Nursing personnel comprise the largest proportion of patient care givers in a hospital. Nursing care in hospitals takes on added importance today because increase in acuity of patients requires intensive nursing care.
In recent years, the nursing profession has been especially concerned about the nature of the transformation taking place in the health care sector. Reports of hiring freezes and layoffs of RNs in hospitals have led to increasing apprehension among them and their supporting organizations about the potential threat to the quality of patient care in hospitals as well as their physical and economic well-being. RNs have expressed concerns that hospitals are implementing a variety of nursing care delivery systems involving major staff substitutions, reducing the proportion of RNs to other nursing personnel by replacing them with lesser-trained (and at times untrained), and lower-salaried, personnel at a time when the increasing complexity of hospital inpatient caseloads calls for more skilled nursing care.
At the same time, the aggregate quantity of RNs is at a high level, creating uncertainties about job security. Much health care is moving to ambulatory settings, the community, and the home through home health services. The nursing profession also has concerns about the training needs to accommodate these
shifts in work settings. With respect to the hospital setting, a rapidly changing health care environment, continuing pressures to contain costs, and the rising levels of severity of illness and comorbidity of inpatients all make it imperative for hospitals to explore innovative ways to redesign delivery of care without compromising quality.
Throughout the decade of the 1980s, hospital expansion, scientific advances, and technological development led to the use of an increasing number of nursing personnel, particularly the RN. As discussed in Chapter 4, employment of RNs in hospitals has increased steadily for the past several decades. In 1993, RN employment in hospitals continued to increase, but the rate of growth over the previous year showed a slight decline for the first time in many years (AHA, 1995b). However, a comparison of first-quarter 1994 data with preliminary data for the first quarter of 1995, shows that while total hospital employment was down; RN employment increased by 3.5 percent, and licensed practical nurse (LPN) employment declined by 1.2 percent (see Table 4.3). These figures may represent a 1-year artifact or an indication of an underlying shift in the health care delivery system.
Information about trends in employment levels of RNs and other nursing personnel needs to be understood in the context of the changing health care system, as elaborated in Chapters 3 and 4. In particular, hospital inpatient lengths of stay continued to decline, along with inpatient days; admissions increased in 1995 after remaining relatively level in 1994. The increasing acuity of patients requiring intense nursing care, the large increase in hospital outpatient services, and the relative increase in beds dedicated to intensive care units also may account for at least part of the continued increase in hospital employment of RNs.
In sum, although the committee heard reports of widespread layoffs of RNs and other nursing personnel, national statistics suggest that in the aggregate these employment losses appear to have been more than offset by hires. (This generalization does not hold for licensed practical nurses [LPN], whose employment by hospitals has been declining for some years.) The continued growth in RN employment appears to run counter to many assertions the committee heard from nurses during site visits, testimony and numerous written and oral communications throughout the study. Aggregate trends, of course, obscure local and regional variations that respond to local market conditions and other factors, and anecdotal information cannot be discounted totally as it often is a warning indicator of changes that are not yet reflected in national statistics.
This chapter examines the relationship of staffing patterns of nursing personnel in hospitals and quality of patient care. The chapter begins with a discussion of the restructuring of hospital care and the changing roles of nursing personnel in hospitals. It then provides a brief overview of the elements of quality of care, measurement issues, and the status of quality in hospitals. Next, it proceeds to assess whether there is any reliable evidence linking nurse staffing to the quality
of patient care in hospitals. The chapter ends with a brief overview of legislative and regulatory requirements for hospital quality assurance.
Restructuring In The Hospitals
As the average length of stay for patients decreased and subsequently as the number of staffed beds also declined, hospitals began redesigning their systems of care, scheduling practices, and approaches to the care of hospitalized patients, in order to accommodate this decreased need.
Increasing numbers of hospitals are restructuring their organization, staffing, and services. Redesign and reengineering have become principal strategies of the 1990s for many institutions and systems. Although redesign initiatives are undertaken for a variety of reasons, more than half of the efforts are driven by the need to reduce operating costs, and have focused on the transformation of work processes and the redesign of roles and jobs. Staff reductions or changes in labor mix are at times implemented without attention to the organizational changes that might facilitate the possibility of better outcomes with fewer, more appropriately trained and used staff, while at the same time focusing on improved patient outcomes (VHA, 1995).
The labor intensity of nursing services in hospitals cannot be disputed when one considers the fact that the average nursing department's full-time-equivalent (FTE) personnel represent around 40 percent of the overall hospital FTE personnel and around 30 percent of the average annual hospital budget (Witt Associates, 1990). This means that the nursing department represents the largest single department within the institution. In light of the efforts by hospitals to meet the multiple demands that are reshaping their future and requiring them to reduce costs, the nursing department can be a major area for cost reduction efforts simply because of the size of the budget. Brannon (1994, p. 3) notes that in "response to greater market competition and pressures to contain cost, community hospitals not only transformed themselves into diversified health care organizations, [but] corporate managers reorganized the work of hospital workers to contain labor costs and increase productivity. Nursing was at the center of these changes." However, because it is assumed that the changing health care system needs to balance costs with maintaining or improving quality of care, assuring that both the right number and the right kind of nursing resources are available becomes essential for a coordinated and cost-effective health care system.
Concurrent with the efforts to restructure hospital services has been the development of total quality management and what is often referred to as patient-centered and patient-focused care. The typical patient in a traditionally-organized hospital may interact with as many as 60 staff in one 4-day hospital stay (Lathrop, 1992). Hence, efforts for redesigning hospital care have been focused on the integration of many hospital services in an effort to provide more patient-centered or -focused care. Both these efforts—total or continuous quality man-
agement and patient-centered care have led to major work design and organizational change in several hospitals. These innovative, team approaches may also involve case managers,1 the development of critical pathways for managing patients most efficiently during a hospital stay, and other steps that, collectively, lead to restructuring in the hospital.
The restructuring of hospital inpatient services is but one part of the larger restructuring efforts of the care delivery system related in large part to managed care and the development of integrated delivery systems. While much anecdotal information is available about these changes, objective data are not available to determine how widespread these changes are and whether or not this redesign accomplishes its dual goals of increasing patient-centered care and cost reductions.
Because of the resource intensity of hospital nursing services, restructuring, work redesign, and cost reduction efforts have a direct impact on the nursing workforce. It is not surprising, therefore, that the restructuring of hospitals and redesign of nursing services are among the most pressing issues for the nursing profession and ultimately for the future of health care delivery in hospitals. Staffing to provide safe, effective, and therapeutic patient care is a challenge for nurse administrators under any circumstances, and substantial changes are occurring in the organization and delivery of hospital care.
Changing Roles and Responsibilities of RNs
While the RN in the hospital remains in the pivotal position for coordinating care in hospitals, sometimes as a case manager, the position of the nurse assistant (NA) has been changing. In some institutions it is being upgraded, with NAs assuming, under the direction of the RN, increasing responsibility for more direct care activities than in the past. This results in an increasing level of management and supervisory skills being required of RNs. In some hospitals the redesign of the nurse assistant role has occurred in conjunction with the redesign of other support activities such as dietary, housekeeping, and transportation services. The integration of these functions is viewed as one way to have fewer people interacting with the patient, while also providing the potential for cost savings.
These redesign efforts have led to changes in the patterns developed over the
past two decades for organizing the delivery of nursing care in hospitals. The challenge today is for care givers and patients to think about the continuum of care needed rather than simply the event of hospitalization. The emphasis is no longer on the inpatient hospital care of a patient, but rather to view the event of hospitalization as one event in the illness continuum. Changing the emphasis to a continuum of care requires hospital nursing services to develop new structures and practices. Foremost among these changes is to help nursing and other health care givers to learn how to plan for patient care before the patient is admitted to the hospital, as well as for care needed after discharge from the hospital.
This change also calls for more flexibility in staffing patterns and time schedules that will focus on the needs of the patient. Flexibility in scheduling was a major issue of the late 1980s when hospitals had nursing shortages. A change to 12-hour schedules with fewer days worked by nursing staff became a standard in many hospitals in an effort to increase nurse satisfaction and thereby decreasing the turnover of nursing staff. Today's work redesign appears to be changing some of those schedules back to the more traditional 8-hour day and 40-hour week, reducing some costs in doing so and providing the potential for more stability in staffing systems that provide opportunity for some nursing staff to practice across the boundaries of inpatient and ambulatory or community nursing care.
It is this latter challenge, along with the demands for increased efficiency within a standard of good quality of care that, in part, has led many hospitals to implement the concept of care teams. These care teams are generally interdisciplinary in nature, bringing members of the appropriate disciplines together from all areas, ambulatory, inpatient, and community for example, to collectively develop a plan of care that will optimally benefit the patient, meet a pre-established set of standards, and use as few resources as possible in carrying out the plan of care. Such care teams often use practice guidelines, sometimes referred to as care maps or critical pathways, to determine the plan of care and progress of the patient along a time line established for the continuum of care. These guidelines are similar to decision trees and require a team of care givers who engage in high levels of effective communication and have the knowledge and skills required to enter into collaborative planning and evaluation. As managed care becomes more the norm, expedient decision making and good judgment will be increasingly more important for all health care providers and the use of interdisciplinary approaches such as care teams also will become increasingly the norm in the hospital sector (Sovie, 1995).
In order to accomplish the work of care teams effectively and efficiently, case managers are often used. This system of organizing care relies on this manager to integrate in-depth clinical knowledge, community resources, and financial and organizational requirements with patient needs and the institutional goals of providing high quality, cost effective care. While the needs of the patient should determine who will be the case manager, in acute care settings this role is
most often performed by an RN, frequently one who has been prepared with education beyond the basic program of nursing. In most hospitals the nursing case manager is the person who spans the boundaries of inpatient, ambulatory, and community nursing (Girard, 1994).
In summary, the dynamics of staffing and scheduling in hospitals, always more complex than one would expect, have taken on even greater complexities as care giving becomes much more interdisciplinary in nature and care givers are required to consider more than the event of illness presented, that is, the current hospitalization or the current outpatient visit. The continuum of care, sharing of information across the system, and the increasing involvement of patients and families in their own care giving requires that staffing be considered in its broadest definition.
Evolving Roles in Advanced Practice Nursing
The committee takes note of the growing trend toward complexity of illness and sophisticated care management of patients in an illness episode that includes the event of hospitalization but is not limited to it. The care planning and managing begins before the patient is admitted to the hospital and continues beyond the hospitalization to discharge planning and management of care needed after discharge from the hospital. Leading and managing the organizational transformations described above require talents or training that not all RNs now have. For the evolving hospital, the committee believes that it will be imperative for these management and leadership skills to be fostered through various educational programs. The committee believes that more advanced, or more broadly trained, RNs will be needed in the future. Such training is essentially like that now provided for RNs who receive certification as, for example, advanced practice nurses (i.e., clinical nurse specialists, nurse practitioners, nurse midwives, and nurse anesthetists).
Clinical nurse specialists can be found in every specialty area of nursing. In each of these areas clinical nurse specialists function as practitioners, educators, case managers, consultants, researchers, and administrators; in the mental health arena, they may also serve as psychotherapists. They play a critical role in the ongoing clinical management of caseloads of patients. Nurse practitioners manage patients with acute and chronic conditions. They frequently have responsibility for managing patients with illness such as diabetes or hypertension. They also are responsible for the ongoing primary care of a group of healthy individuals.
The value of such clinical nurse specialists, in terms of both patient care and economic factors have been studied over the past 20 to 25 years. In particular, a number of randomized clinical trials have been conducted.
One set of studies was directed at testing the effectiveness of programs conducted by clinical nurse specialists in caring for hospitalized elderly patients,
especially in comprehensive discharge planning. Outcomes such as length of stay, number of, or length of time before rehospitalization, and costs, as well as functional status, were all better among those patients whose care was coordinated and implemented by clinical nurse specialists. Neidlinger and colleagues (1987) found the use of a comprehensive discharge planning protocol implemented by a clinical nurse specialist saved an average of $60 per patient day more than their control group. In a follow-up to this study Kennedy and colleagues (1987) found that for the same control and experimental groups the experimental treatment group's average length of stay was reduced by 2 days, and the length of time before hospital readmission increased by 11 days.
A pilot study by Naylor (1990) had similar results. She found that there was a significant difference between the two groups in frequency of hospital readmissions. Later, in a randomized clinical trial, Naylor and colleagues (1994) found that from initial discharge to 6 weeks after discharge, patients in the intervention group managed by clinical nurse specialists had fewer number of hospital readmissions, fewer total days of rehospitalization, lower readmission charges, and lower charges for health care after discharge. Functional status was the focus of a study by Wanich and colleagues (1992). These researchers found that in their clinical trial patients in the intervention group (those whose care was coordinated by the clinical nurse specialist) were more likely to improve in functional status than those who did not who did not receive such care. These same patients were less likely to deteriorate on measures of functional status during their hospital stay. These outcomes may also help to reduce length of hospital stay and decrease costs, although cost was not measured as an outcome in this study.
Oncology clinical nurse specialists have also been shown to improve patient outcomes. McCorkle and colleagues (1989), for example, conducted a randomized clinical trial of lung cancer patients. The study demonstrated that lung cancer patients receiving care from specialized oncology clinical nurse specialists experienced less distress, less dependence, fewer rehospitalizations, and shorter lengths of stay than did patients cared for without intervention from these advanced trained personnel. According to Russell (1989) the cost of care for patients undergoing a modified radical mastectomy who were followed by an oncology clinical nurse specialist was significantly lower than for those not so followed. Their average length of stay was 3.4 days, while that of the control group was 6.7 days. The costs of hospitalization averaged a difference of $1,668.43 per patient.
Another set of studies involved low birthweight infants and families who received care and consultation from clinical nurse specialists. Outcomes such as length of stay were better and costs were lower among study participants who were in the group using such specialist nurses (Brooten et al., 1986; 1988; Damato et al., 1993).
The ability of clinical nurse specialists to function in a number of different roles and their ability to work independently to solve problems and be patient
advocates as well as integral members of a health care team have been cited as a reason for improved outcomes and cost savings that they help to bring about. Nurses in this role are an important part of the total patient care picture across settings and as such are essential to improved patient outcomes.
Although the discussion to this point has drawn on the use of advanced practice nurses in discharge planning and working with patients in the home after hospitalization, the place of such personnel in the entire continuum of care for which hospitals are responsible needs to be understood. For one thing, all but the smallest hospitals operate outpatient clinics of various sorts (including those that deal with non-urgent problems of patients who present to emergency departments). Moreover, as the U.S. health care system restructures itself, many larger hospitals and academic health centers are becoming the center of integrated health delivery systems that vertically integrate providers from small physician office practices through multispecialty groups through a variety of other aspects of health care. With this trend, plus the growing phenomenon of delegation and substitution of responsibilities from physicians to nurses (i.e., nurse practitioners or other types of advanced practice nurses), it is clear that the role of advanced practice nurses is now and will continue to expand.
Based on this type of information, combined with what was learned from testimony, site visits, and the professional expertise and experience of its members, the committee concludes that high-quality, cost-effective care for certain types of patients, particularly those with complicated or serious clinical conditions, will be fostered by the use of such advanced nurse specialists. The committee believes that increased use of advanced practice nurses would improve the cost-effectiveness of our health care systems and facilities. That is to say, changing the mix of nursing personnel involved in caring for patients with increasingly complex management problems may yield both improved outcomes and lower costs.
RECOMMENDATION 5-1: The committee recommends that hospitals expand the use of registered nurses with advanced practice preparation and skills to provide clinical leadership and cost-effective patient care, particularly for patients with complex management problems.
Advanced practice nurses are typically classified in at least one of four ways, and their educational training and duties differ accordingly. Clinical nurse specialists typically are master's degree-trained RNs; some may also have PhDs. Their clinical specialties can include oncology, neonatology and, or, pediatrics, mental health, adult health, women's health, geriatrics, and AIDS. They commonly work in clinical settings and provide primary care; case management services; psychotherapy; and a variety of organizational, administrative, and leadership services as well. Nurse practitioners are usually prepared at the master's degree level and also certified in a specialty area of practice, such as pediatrics,
family practice, or primary care. Their usual responsibilities include managing clinical care; they conduct physical examinations, track medical histories, make diagnoses, treat minor illnesses and injuries, and perform an array of counseling and educational tasks. Nurse practitioners may also, in some circumstances, order and interpret diagnosis tests and prescribe medications. Certified nurse midwives are RNs who have graduated from a nurse midwifery program accredited by the American College of Nurse-Midwives (ACNM) and are certified as a nurse-midwife by the ACNM; some may have taken a master's program offered by a school of nursing or a school of public health. They provide prenatal and gynecological care, deliver babies in a variety of settings (hospitals, birthing centers, or homes), and render postpartum care. Finally, certified RN anesthetists have a bachelor of science in nursing and 2 to 3 years of additional education and training in anesthesiology, often at the master's level. They, too, have a rigorous certification process, managed through programs approved by the American Association of Nurse Anesthetists. Particularly in rural areas, these nurse specialists may administer the majority of anesthesia or anesthetics in health care settings today.
Clearly, well-trained advanced practice nurses can function in a number of different roles. They can work independently to solve patient care problems, serve as patient advocates, and be integral members of a health care team. Advanced practice nurses can improve the cost-effectiveness of health care systems and facilities because changing the mix of personnel involved in caring for patients with complex management problems may yield better outcomes, lower costs, or both. The committee concludes that the way should be clearer for such personnel to be used in both inpatient and outpatient settings and for them to be able to take up leadership positions and act independently.
One obstacle, however, to accomplishing the changes advocated in this section lies in the differing ways in which states recognize advanced practice nurses, chiefly in terms of the breadth of independent authority (e.g., diagnosing, prescribing, and dispensing of medical therapeutic agents or controlled substances) (Pearson, 1995; Ray and Hardin, 1995). Some state boards of nursing have not yet recognized the expanded responsibilities that such personnel can and should discharge. To address this problem, the committee believes that all states should recognize nurses in advanced practice in their nurse-practice acts and delineate the qualifications and scope of practice of these nurses.
Ancillary Nursing Personnel
Today, almost all hospitals in the United States use some kind of ancillary nursing personnel. As stated in Chapter 4, this group of personnel includes nurses aides or assistants (NA), some of whom may be certified, as well as a variety of other ancillary personnel. By definition, they have less formal education and training than RNs or LPNs; on average, when hired they may also have
less exposure to, and time or experience in, the inpatient setting. Their education, however, does not stop after the basic training. Many serve for several years and learn from physicians, RNs, and LPNs to perform tasks that once were not done by NAs and to be responsible for specific aspects of clinical care.2
As already noted, far less information about employment trends is available on this group of the nursing workforce than on the more traditional nursing categories. However, the transformation of the hospital care delivery system is clearly going to involve these types of personnel at least in the near future. The use of NAs and other ancillary nursing personnel to assist RNs with patient care is reported to have increased in recent years. In most instances, NAs and other ancillary nursing personnel are used in simple bedside care or as unit assistants (e.g., changing dressings, taking vital signs such as blood pressure and temperature); in some instances, they are being used to assist RNs in total bedside care or in other duties such as telemonitoring, lifting teams, electrocardiography, or physical therapy (Barter et al., 1994). In other cases, tasks performed by these types of personnel may overlap with those of other support units, such as dietary, housekeeping, or transportation services.
Krapohl and Larson (1995) describe the evolution of nursing delivery systems in hospitals, from team nursing (with all the variations such as clinical, nonclinical, and integrated nursing models) to primary nursing models. The various "patient-focused" team models that some hospitals are implementing incorporate less skilled nursing personnel to varying extent; the models themselves vary according to the specific needs of different hospitals (or hospital systems). The authors also reviewed the literature regarding the use and evaluation of ancillary nursing personnel in hospitals and found no strong evidence to confirm that these nursing personnel improve (or reduce) quality or increase (or decrease) nurse or patient satisfaction. A review of studies of primary nursing also do not conclusively show the superiority of primary nursing models over various team nursing models. They conclude, basically, that although nursing care has been provided in hospitals since the hospital's beginning, no single delivery system has emerged as ideal. The authors also note the methodological and design weaknesses of the studies reviewed.
Nevertheless, some informal information about these new team approaches is encouraging. The committee learned, for instance, about variations of the "partners-in-practice" program pioneered in the 1980s, which linked NAs and other ancillary nursing personnel with an RN (Manthey, 1988, 1992). At the site visit in Oregon, committee members and staff were able to observe and interact
with care teams in which RNs assumed a very close working relationship with the other care partners. Nursing personnel at all levels worked together at all times; thus, RNs were able to assess the knowledge and clinical capabilities of each member of the team and, where necessary, step in to supervise, and then teach, the less-well-prepared nursing personnel. Other hospitals have implemented similar systems, according to information made available to the committee.
The underlying message of the literature review and the observational information gathered by the committee is that, in the hospital sector, issues of training and competency of non-RN staff remain critical. No national standards exist for minimum training or certification of ancillary nursing personnel employed by hospitals (unlike, as discussed in Chapter 6, for NAs in the nursing home sector); thus, they vary widely in educational attainments and in their training for simple nursing or quasi-nursing tasks. Furthermore, no accepted mechanism exists either to measure competency or to certify in some fashion that ancillary nursing personnel have attained at least a basic or rudimentary mastery of needed skills. Hospitals vary widely in the levels of training they provide to these personnel. Barter and colleagues (1994) found that 99 percent of the hospitals in California reported less than 120 hours of on-the-job training for newly hired ancillary nursing personnel. Only 20 percent of the hospitals required a high school diploma. The majority of hospitals (59 percent) provided less than 20 hours of classroom instruction and 88 percent provided 40 hours or less of instruction time. RNs and their supporting organizations have expressed much concern that NAs and other ancillary nursing personnel are being given various nursing-related tasks in hospitals in the absence of competency requirements. The committee is greatly concerned about this lack and the potential for adverse impact on patient care.
RECOMMENDATION 5-2: The committee recommends that hospitals have documented evidence that ancillary nursing personnel are competent and that such personnel are tested and certified by an appropriate entity for this competence. The committee further recommends that the training for ancillary nursing personnel working in hospitals be structured and enriched by including training of the following types: appropriate clinical care of the aged and disabled; occupational health and safety measures; culturally sensitive care; and appropriate management of conflict.
The committee believes that hospitals should take the lead in ensuring the competence of, and provision of appropriate training to, all direct care personnel employed by them, including ancillary nursing personnel. The committee does not believe that the first course of action should be enforcement by law or regulation at the federal, state, or municipal level. It does caution however, that if real quality-of-care problems were to emerge in hospitals that could be related to negligence by hospitals in ensuring competence, then the public might be ex-
pected to clamor for the enactment and enforcement of more stringent, external, regulation. Such rules would then protect the public from problems that hospitals themselves should have guarded against.
Hospitals are in a better position than nursing homes to assure the competence of NAs because NAs are not the predominant care givers in hospitals that they are in nursing homes. Hospital NAs are more likely to work in teams with other care givers and to have more direct supervision from the RN, who is more immediately available than is usually the case in nursing homes.
Finally, culturally sensitive care will become increasingly important in the years ahead. As noted in Chapter 2, the population, and therefore the patient population, is not only aging but also is becoming more racially and ethnically diverse. Thus, increasingly, care givers and care receivers may come from different cultural backgrounds. The imperative for cultural sensitivity is obvious.
Involving Personnel in Planning for Change
The changes briefly described above are appealing conceptually, and time will tell if they are effective and practical as the hospital sector reinvents itself. In the short term, however, these shifts in the way hospitals do business, and the way they organize to conduct their business, are causing notable disruptions and misgivings among the nursing staff. From the frequency and intensity of the commentaries that the committee heard during this study, RNs are concerned about both the employment ramifications and more importantly the professional implications of the organizational changes that are occurring; they believe that these changes may lead to undesirable and unanticipated effects on quality of care.
In response to pressures to contain costs and improve quality of care, which may or may not be related to the downward trend in inpatient hospital use, hospitals are restructuring services, units, and activities. These efforts are generally oriented toward increasing productivity or efficiency and/or reducing operating costs. As stated above, redesign efforts often involve the integration and coordination of work across departmental lines, which may also lead to elimination of positions, layoffs, redefinition of positions, and realignment of supervisory lines.
Restructuring of inpatient services in hospitals, accompanied by a changing mix of nursing personnel, is an inevitable consequence of the demands by society, through the payers of care, to control the costs of health services. Downsizing of the patient care workforce in inpatient hospital settings will continue, at least in the near future. Nursing personnel will not be immune from such downsizing. "Reengineering" of patient care processes, including changes in skill mix, will also continue, at least for the short term.
Overall, the sense of disquiet about the future, especially among RNs, was palpable, in part because of the unpredictability of the effects of these changes
and in part because of the seeming lack of input and control that many nurses felt about the changes being made. The committee heard from nurses who had lost their positions in hospitals about management decisions for downsizing that had been made without any staff involvement—a phenomenon that adds to the feelings of threat and uncertainty for many hospital-based RNs. At the same time, the committee had the benefit of learning about other hospitals where management involved staff in substantial ways in reaching solutions about how the necessary staff restructuring ought to take place. At one of its public hearings, for instance, the committee heard from witnesses about the beneficial results of using free federal mediating services. On site visits, committee members visited some hospitals where change had been successfully implemented through well-conceived planning and implementation processes that involved both nursing administrators and staff nurses.
In the committee's view, the harmful and demoralizing effects of these changes on the nursing staff can be mitigated, if not forestalled altogether, with more recognition on the part of the hospital industry that involvement of nursing personnel from the outset in the redesign efforts is critical.
RECOMMENDATION 5-3: The committee recommends that hospital leaders involve nursing personnel (RNs, LPNs, and NAs) who are directly affected by organizational redesign and staffing reconfiguration in the process of planning and implementing such changes.
The committee found impressive the testimonies and descriptions of these collaborative "redesign" efforts that involved all levels of nursing personnel in the restructuring process as illustrated above. The rationale for inclusion of nursing personnel in hospital restructuring efforts relates to several factors: such involvement brings to the table the professional knowledge and experience needed in developing such changes, staff commitment to the decisions made, and may affect the likelihood of success and of improvement in the quality of care. It is not simply to make the affected nursing staff feel better. Change is likely to fail if a top-down approach is imposed on hospital nursing staff.
Furthermore, the health care sector is moving rapidly to adopt principles of continuous quality improvement and total quality management as means for addressing issues in quality of care, for advancing the state of the art of quality measurement and management, and for promoting continuous progress in health care processes and patient outcomes. These newer quality assurance and improvement techniques rely heavily on input from multiple segments of a health organization's personnel and departments; that is, they do not deal with quality issues that relate to only a single department, in part because most problems in health institutions and facilities are systemic rather than traceable to single events, people, or units. Logic alone would dictate, therefore, that as an organization wishes to reinvent its structure and systems, it ought to adopt these same prin-
ciples of involving individuals from across the departmental and personnel spectrum.
Tracking the Effects of Change
Although available national statistics on hospital employment do not show reductions in levels of nursing staff at the national level, the media frequently report on staff layoffs in hospitals. Anecdotal information abounds, and ad hoc inquiries are conducted by unions, nurse associations, magazines (of their membership and subscribers), and other organizations. Unfortunately, the very low response rates of many of these inquiries and the deficiencies in the design of surveys and of questions do not permit consumers or policymakers to derive objective measures of the links between staffing patterns and processes and outcomes of care, and to draw valid conclusions.
As a consequence of declining trends in inpatient hospital use, some hospitals have been reducing the number of operating beds and reducing nursing positions by attrition or layoffs. Other hospitals are closing, and some are converting beds to long-term care and other services. Some RNs, LPNs, and NAs have been laid off or redeployed from acute care units to other services, programs, or settings. These types of downsizing and consequent restructuring efforts necessarily affect employment of nursing personnel and will continue as long as hospitals face low-use patterns. This turbulence in the health care delivery system and the resultant unstable situation fuel the concern that large decreases in RN staffing in hospitals are both occurring and leading to decrements in patient care and to threats to the health and well-being of nursing personnel.
As stated earlier, throughout the period of the committee's study changes were occurring in hospitals in the use of RNs and in the ratio of RNs to other nursing personnel in the organization of the delivery of patient care. Many of them intimated that such changes potentially will diminish the quality of care provided but the committee was unable to find evidence of a decline in the quality of hospital care because of any changes in staffing. Lacking reliable measures and data, no one is in a position to draw valid conclusions. The amount of testimony provided, however, and the depth of concern cited, was sufficient to lead the committee to believe that this is an area that requires on-going monitoring and research in order to ensure that the responsibility for providing safe, effective, quality, and cost effective care is fulfilled within the health care system.
The committee finds that lack of reliable and valid data on the magnitude and distribution of temporary or permanent unemployment, reassignments of existing nursing staff, and similar changes in the structure of nursing employment opportunities greatly hampers efforts at understanding the problem and planning for the future. Answers are needed to numerous questions, such as: What happens to nursing staff after they are laid off? Are they employed in another hospital or reemployed at the same hospital? Do they move to outpatient, community, or
long-term care settings? Do they return to school for retraining or for more advanced nurse training? Do they leave nursing altogether for another occupation?
Information on RN employment patterns is necessary. Research is needed on whether career paths of RNs will change markedly over the next 5 to 10 years. These changes can have implications for career choices, curriculum design, structure of occupational ladders, and perhaps, quality of care. Among the many questions that warrant attention are the implications of restructuring for career choices, the structure of occupational ladders, and both entry and midcareer curriculum design.
RECOMMENDATION 5-4: The committee recommends that hospital management monitor and evaluate the effects of changes in organizational redesign and reconfiguration of nursing personnel on patient outcomes, on patient satisfaction, and on nursing personnel themselves.
More detailed data also are needed on the employment patterns of NAs and LPNs over time. Throughout its deliberations the committee focused largely on RNs not only because they form the largest proportion of nursing personnel in this country and because their professional associations are the most well organized, but also because of the paucity of comparably detailed data on NAs and LPNs. For that reason, hospitals should not concentrate their monitoring and evaluation solely on the relationships between RN staffing and quality of care or on work-related illness and injury. Rather, hospitals should focus their monitoring and evaluation efforts of the restructuring and redesign of staffing on the entire spectrum of their nursing personnel.
The federal agencies concerned with health workforce data and research have a major role to play. Thus, in this regard the committee supports the missions of agencies such as the Division of Nursing of the Bureau of Health Professions (in the Health Resources and Services Administration), the National Institute of Nursing Research (in the National Institutes of Health), and the National Center for Health Statistics (in the Centers for Disease Control and Prevention) all in the Department of Health and Human Services. The need for timely and relevant data that is amenable to integration across systems is urgent. In the committee's view, these agencies should work productively together and in collaboration with private organizations to develop databases containing information that will shed light on workforce issues and on the relationships of staffing of nursing personnel and care processes and patient outcomes.
Measuring Quality Of Care In Hospitals
The legislative mandate to this Institute of Medicine (IOM) committee asks it to examine structural issues—the number of nurses and the mix of types of
nurses—but it also asks the committee to focus on outcome issues in terms of nurses themselves as well as the quality of patient care that could involve either processes, or patient outcomes, or both.
The committee looked at quality broadly beyond nursing inputs in terms of the overall quality of care received by the patient in the hospital, and examined the relationship between structural variables and both processes and outcomes of care. In this effort, it was guided by the IOM's definition of quality of care: ''… the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge" (IOM, 1990, vol. I, p. 21).
The IOM definition does not contend that quality can or should be defined in terms of available resources. Excluding resource constraints in the definition provides the opportunity for Quality assurance (QA) and Quality Improvement (QI) systems to distinguish quality-of-care problems from problems arising from resource availability. Quality measurement methods and QA/QI approaches should be able to (1) identify how and to what degree resource constraints affect structure, process, and outcome elements of health care; (2) identify the agent(s) that are responsible for the constraints and have the authority to address the problems they may be causing; and (3) perhaps conduct corrective actions and monitor progress in improving care.
How one measures or improves the quality of health care is linked closely to the assumptions made about what constitutes quality of care in the first place. This multidimensional definition of quality of care is compatible with the perspective that patients, consumers, providers, payers, and public entities all have interests in the quality of care rendered by health care institutions and personnel. Having an understanding of the important dimensions of quality of care is a key initial step toward developing measurement and intervention approaches and implementing QA/QI strategies. Standards and indicators of performance must be closely linked to the operational concepts used in defining quality. Care is then assessed or measured against criteria or benchmarks to determine whether standards are met and where lie opportunities for improvement, regardless of whether specific standards have been met or exceeded.
Elements of Quality of Care
The most fundamental conceptual framework in the area of quality of care was articulated three decades ago by Avedis Donabedian (1966); his formulation is based on a triad of structural factors, process-of-care variables, and outcomes or end results of that care. These elements are briefly discussed below to provide some context for discussions in this and the next chapter.
Structural criteria are, in effect, proxy measures of quality of care—for an entire organization such as a health plan, hospital, or nursing home, or for an individual health care clinician such as a nurse or physician. Structural criteria may involve variables such as the numbers of various kinds of staff (such as RNs or LPNs or NAs), staff-to-patient or staff-to-bed ratios, the training and supervision expected of or given to staff, the patient record system, the procedures for infection control, building code requirements, and the quality of the physical plant and equipment. These elements may all reasonably be thought to affect the processes of care and, hence, the subsequent health and functional outcomes of patients in hospitals or residents in nursing homes.
Structural measures focus on the presumed capacity of people and entities to deliver adequate or high quality care, but they do not measure the care itself. Accordingly, deficiencies in structural measures cannot be confidently used as direct evidence either of poor care or poor outcomes, but failings in this area, whether perceived or real, certainly can be considered indicative of problems until proven otherwise.
The principal problem with structural variables is that little empirical evidence has been generated linking structural variables directly with good (or bad) processes or (especially) outcomes of care. Thus, the question of whether problems with the levels of nurse staffing or the mix of competencies within a nursing staff may be associated with poor care and risk of patient harms is a reasonable one, but the probable lack of explicit information on any association between such structural variables and the larger issues of interest must be clearly understood from the outset.
Process of Care
The process of care encompasses what is done to, with, and for the patient or health care consumer. Process criteria pertain to the appropriate and correct performance of specific "technical" procedures and services; they also involve interpersonal skills and attitudes, such as those of compassionate communication. Process indicators are very broad, and over the years they have constituted the most widely used set of measures of quality of care. They can include such elements of good quality care as promoting the participation of the patient or resident in the selection of care management strategies and honoring privacy of information and personal space.
Outcomes of Care
Outcome measures are typically considered to be the end results of health care in terms of biologic, psychologic, and functioning variables. They may
include various health indicators such as death rates, rates of illness, or rates of specific complications of illness, as well as physiologic measures such as blood pressure, serum glucose, or cholesterol levels—that is, the kinds of clinical measures that appear to matter more to physicians and other clinicians. More broadly, however, health outcomes encompass functional abilities (both physical and cognitive), pain or discomfort, energy and vitality, and mental and emotional well-being—in short, aspects of health status that matter most to patients and their families. This committee embraces the broad view of outcome measures taken by prior IOM committees (IOM, 1990), that is, health-related quality-of-life variables regarding physical, social, and emotional heath, cognitive and physiologic functioning, and overall well-being.
Table 5.1 shows an illustrative list of measures pertinent to nursing quality in the inpatient hospital setting. Although the items in the table are selective and illustrative only of major areas of interest, the committee is cognizant of the fact that data would not be easy to acquire for many of these measures.
Measurement of Quality
In the current health care environment, the attention to, and importance placed on, quality of care are increasing at a notable rate. This is predicated on several factors: the belief that institutions and plans in a competition- or market-oriented health care sector will have to compete on more than price; the vastly greater ability than before of those in the health sector to measure both processes and patient outcomes reliably; and the markedly improved understanding of how to implement effective programs for improving quality. More effort is also being directed at understanding the effectiveness of services and outcomes for both individuals and populations.
Recent years have seen important advances in measuring quality of patient care at the individual patient and population levels,3 involving both process and outcome measures. From the vantage point of this study, however, existing work has not typically focused on isolating the contribution of nursing care in measuring the quality of patient care in hospitals. Hegyvary (1991), for instance, notes that the literature about productivity in nursing services does not address results in patient care and almost without exception does not raise the question of quality of care.
A fairly wide array of quality indicators has been developed including mortality, unanticipated hospital readmission, hospital-acquired complications, and nosocomial infections. Hospital-specific mortality rates have received particular
TABLE 5.1 Illustrative Measures of Quality of Care in Inpatient Hospital Settings, with Specific Attention to Nursing Care
Structural Measures of Quality
attention in the past few years, beginning with the release of such information by the Health Care Financing Administration (HCFA) in the late 1980s and subsequent efforts by various states to do the same (at least for selected types of admissions or operating procedures such as coronary artery bypass graft surgery). Many of these efforts have been carried out using large administrative databases. In general, major questions remain about the reliability, validity, and generalizability of the information from these types of mortality rate studies.
For example, Thomas and colleagues (1993) reviewed the seminal work on risk-adjusted mortality rates and conducted detailed validity studies for three conditions (cardiac disease, acute myocardial infarction [AIM], and septicemia) to determine whether death rates appear to relate to quality of care as evaluated by an experienced physician peer reviewer. They report that of 9,721 cases involving angina and cardiac surgery, 1,103 (11 percent) were considered poor quality; the figures for AMI were 776 of 6,004 cases (13 percent), and for septicemia, 285 of 1,709 cases (17 percent) (Thomas et al., 1993, Tables 3, 4, and 5). With respect to the question of using risk-adjusted mortality rates derived from large administrative databases, however, the investigators concluded that this strategy may be inappropriate unless the quality–outcome relationship is explicitly validated.
Especially complex are questions of (a) whether information on hospital performance as measured by death rates or complication rates will be correlated for different clinical services and (b) whether death rates are correlated with complication rates. Recent work (e.g., Iezzoni et al., 1994; Silber et al., 1995) suggests, in particular, that complication rates will not be related significantly to mortality rates. Those problems notwithstanding, it should be clear that for the purposes of this study, such information tells little, if anything, about the precise role of nurse staffing levels or mix in promoting higher-quality patient care. In fact, one study presents data that link mortality with length of stay (and with other resource-use variables) but that do not demonstrate any relationship between mortality and staffing, such as personnel measured in FTEs, total staff per admission, or RN-to-LPN ratio (Bradbury et al., 1994).
In addition, for many years, "generic screens" have been core quality measures for hospital inpatient care.4 Usually based on actual review of patient records or discharge abstracts, these involved items such as adequacy of discharge planning, medical stability of the patient at discharge, unscheduled return to surgery, or trauma suffered in hospital in addition to the broader measures noted. Depending on the level of detail in such quality screens, more information
on the role of nursing staff in producing high (or low) quality of care could be obtained or at least inferred. However, this type of data collection can be very time-consuming and costly for the amount of useful information gained that might point to quality-of-care problems; the federal Peer Review Organization (PRO) program for Medicare, at least, has in recent years turned away from use of record review for generic screens as a basic tactic for quality assessment.
The use of computerized algorithms to screen for possible quality problems, by applying them to hospital discharge abstract data (e.g., large-scale hospital databases), has been a more recent development. Some methodological research (Iezzoni et al., 1992) suggests that computerized indicators show some promise in terms of identifying hospitals that warrant more intensive review for quality-of-care reasons (e.g., untoward complications of care), because they do a reasonably good job of identifying quality problems, although they may also incorrectly point to problems in a rather large number of cases.5 Thus, more work is required to demonstrate the reliability and validity of computerized screens for targeting hospitals for more in-depth quality review. Moreover, as with the earlier generic screen approach, these types of measures tell little about nursing care per se.
Status of Hospital Quality of Care
Quality of patient care is central to the delivery of health care services in hospitals. A major factor precipitating this study were reports, emanating chiefly from nursing groups, that the quality of hospital care is declining, brought about by restructuring and reengineering, and consequent reductions in the proportion of nursing personnel trained as RNs. During the study, the committee heard considerable concern expressed by RNs that hospitals are restructuring and reengineering in increasing numbers that are resulting in smaller proportions of RNs to total nursing personnel, and about the probable negative impact on quality of patient care in those hospitals. Unfortunately, very little recent objective, national data are available that describe the status of quality of care in hospitals and assess if it has been affected in any way by changes in the system of delivery of care. There is virtually no research on the effects of ratios of RNs per bed on patient outcomes. Unlike staffing based on patient acuity, across-the-board staffing ratios assume that all patients can be cared for with the same level and type of resources. The difficulty in establishing staffing ratios appropriate for all settings and situations across the country is obvious. Given the variation that exists in patient acuity and the total patient care environment, the committee believes
that it is neither practical nor desirable to establish specific ratios of nursing personnel enforceable by regulation or law. The committee is not discounting what it heard at the public testimony and site visits that unequivocally expressed concerns about trends in quality, both present and future. However, based on studies conducted in the 1980s, there are indirect indicators that in general quality of hospital care has not declined.
Similar concerns were expressed at the time the Medicare Prospective Payment System (PPS) was implemented in the early 1980s. Many observers predicted major quality problems in the hospital sector. The main reason was simply that the diagnosis-related groups (DRG)-based PPS system turned financial incentives for hospitals completely around (compared to the incentives under traditional cost reimbursement schemes), leading many to believe that hospitals would be forced to scrimp on patient services. An in-depth before-and-after evaluation of the effects of DRG-based PPS was conducted by a team of investigators from the RAND Corporation (Draper et al., 1990; Kahn et al., 1990a,b,c; Keeler et al., 1990, 1992; Kosecoff et al., 1990; Rogers et al., 1990; Rubenstein et al., 1990). The researchers found that in general quality of hospital care did not suffer as a result of PPS implementation, and indeed may have even improved in some areas.6 The researchers found that where the process-of-care and the documentation were improved, the outcomes were improved. The study revealed consistent results between outcome as measured by mortality and the process of care variables (Keeler et al., 1992). Severity-adjusted mortality decreased after PPS.
The RAND researchers also found wide variations in quality among hospitals. Differences between types of hospitals were large, with the lowest group estimated to have four percentage points higher mortality than major teaching hospitals in a cohort of patients with an average mortality rate of 16 percent (Keeler et al., 1990).
Moreover, although quality of care did not suffer after PPS, the RAND study confirmed that more patients were discharged too soon and in unstable condition. Those who were discharged in unstable conditions had significantly higher mor-
tality rates. Kosecoff and colleagues (1990) analyzed the data on the level of patient's medical instability at the time of hospital discharge. Using data about five medical conditions—congestive heart failure, acute myocardial infarction, pneumonia, cerebrovascular accident, and hip fracture—the authors established the following measures of medical instability: fever, new incontinence, new chest pain, new shortness of breath, new confusion, new elevated heart rate, new elevated respiratory rate, high diastolic blood pressure, newly lowered systolic blood pressure, newly lowered heart rate, and new premature ventricular contractions. They found that 17 percent of the patients in the study were discharged with at least one disability; 39 percent were discharged with at least one measure of sickness; 24 percent had an abnormal last laboratory value. The risk of death at 90 days following discharge was 16 percent for patients discharged unstable and 10 percent for patients discharged in stable conditions. Instability at discharge had increased since the introduction of PPS. Pre-PPS, 15 percent of discharged patients were unstable, as compared with 18 percent post-PPS, a 22 percent increase. Most of the increase in instability was concentrated in those patients who were discharged to their homes; after PPS, 43 percent of them were more likely to be unstable than prior to PPS. They comment that given continued reductions in hospital compensation rates, the problem of instability at discharge warrants additional research to answer questions such as, "Are increases in instability caused by inappropriately early discharges, too many tests in a shortened hospital stay, incorrect use of new medications, or by changes in nursing practices (e.g., fewer nurses per patient and less time to talk with the patient or monitor incontinence or disorientation)?" (p. 1982). The study also revealed that one quarter of nursing home patients were admitted from the hospital with an instability. Answers to these questions require new data.
Some of the issues raised by the authors may be responsive to nursing care. Although the quality of nursing care was measured, details about performance of RNs have not been provided in the reports published to date.7 Inferentially, however, the levels of quality of nursing care must have been well within acceptable limits, given the overall findings of acceptable quality of care and the fact that decisions to discharge patients (unlike the planning for discharge and for care postdischarge) are not made by nurses.
This landmark research illustrates the value of a well-executed and comprehensive study of the effects of an intervention on the quality of hospital care. Unfortunately, such studies are very expensive and can be somewhat dated by the time the data are collected and analyzed, and the results are published. Data from the RAND studies are based on the experience in the late 1980s. Similar studies
have not been undertaken since; hence, no comparable recent data are available to examine if quality of hospital care is improving, deteriorating, or remaining unchanged.
Similarly, with some exceptions, studies suggest that quality of care has remained the same or possibly improved under managed care (see Chapter 3). It should be acknowledged that the studies reviewed were conducted in the 1980s, an era when there were comparatively few managed care organizations. Care delivered by some of the more recent entrants into managed care may be more problematic.
The committee is shocked by the lack of current data relating to the status of hospital quality of care on a national basis, apart from information on indicators such as hospital-specific mortality rates (which HCFA no longer makes easily accessible). The committee, therefore, is unable to draw any definitive conclusions or inferences about the levels of quality of care across the nation's hospitals today. Briefly, quality of hospital care in general did not suffer and may have even improved after implementation of PPS, as shown by the few studies available. At the same time, there may be problem areas with quality of hospital care as suggested by these studies, but the extent of the problem today is not known because of the lack of objective current data. The committee is convinced that investigation of hospital quality of care warrants increasing and immediate attention. Research also needs to move beyond hospital mortality and focus also on the process-of-care problems and conditions that occur during short hospital stays and investigate outcomes over an episode of care.
Relationship Of Nursing Staff To Quality Of Patient Care
The issues surrounding the relationship of staffing levels and staffing patterns of nursing personnel and outcomes have taken on added importance since the committee was established. Hospitals are restructuring and redesigning the organization and delivery of patient care, and the committee heard many reports of reduction of nursing staff and its adverse effects on quality of care. Very little current data are available describing the quality of care in hospitals, and assessing if it has been affected in any way by changes in the system of delivery of care in the hospitals.
The precise questions about the relationship of nursing staff to quality of patient care in hospitals must be clear. One inquiry asks: Do simply the total numbers (or FTEs) of RNs, or the ratio of different types of nurses (usually RNs to LPNs, or RNs to NAs and other ancillary staff), make a difference? For example, are higher ratios of RNs to LPNs correlated in some fashion with better outcome measures (such as lower diagnosis-specific or hospital-specific death or complication rates)? The other question asks: What particular nursing tasks, skills, or ways of organizing teams of nurses (and possibly other health care
personnel) are related to better processes of care and superior patient outcomes? These issues are examined in the remainder of this section.
Nursing Staff and Quality of Care
Differences in mortality rates across hospitals are well documented by several researchers. Literature on RNs' impact on hospital mortality rates is considerable. Prescott (1993) provides a comprehensive review of empirical evidence of the impact of nursing staff levels and mix on quality of patient care in hospitals. Much of the evidence came from regression analyses that used RNs as a share of total hospital nursing employment as an explanatory variable—essentially, a basic ''numbers" variable. Overall, she found "substantial evidence linking RN staffing levels and mix to important mortality, length of stay, cost and morbidity outcomes" (p. 197). While nurse staffing is not the only factor predictive of mortality outcome, it is an important one affecting the quality of hospital care.
Many factors potentially contribute to high quality of care in any setting; Prescott's review appropriately emphasized studies that considered the role of multiple determinants. Because hospitals are very complex organizations, isolating the role of a single input is difficult, and one cannot eliminate the possibility that, even in a well-controlled analysis, variables such as RN share is surrogates for other, unmeasured quality determinants (Hegyvary, 1991). For example, hospitals with a high proportions of RNs to other staff, may also have the attributes of greater status, autonomy, and control by RNs (Aiken, 1994); these attributes, in turn, may be the determining variable for better quality nursing care and thus better patient outcomes. On the other hand, the ratio of RNs to total nursing employment may be high in hospitals with comparatively few total nursing personnel; if the total number of nursing staff is low enough, the high proportion of RNs may not be correlated with particularly good quality levels.
The performance of a system is determined as much by the arrangement and interaction of its parts as by the performance of the individual components (Scott and Shortell, 1983). Isolating the specific contribution of nursing personnel to the quality of patient care may not be feasible because of the way care is delivered in a hospital, involving contributions of a wide array of staff—nurses, therapists, physicians, and other allied health personnel.
Knaus and colleagues (1986) compared predicted and actual mortality rates for treatment in 13 ICUs. Predicted death rates for each ICU were derived from the acute physiology and chronic health evaluation (APACHE) scores of individual patients. The authors found that although RN staffing levels and positive physician–nurse communication were important factors in achieving lower than expected mortality, the difference between predicted and actual mortality rates were more related to the process of care (e.g., the level of coordination and communication among care givers) than to the structural attributes of the ICUs
(e.g., therapies offered, teaching versus nonteaching status). Mitchell and colleagues (1989) in a demonstration project to document fiscal costs and patient care effectiveness of critical care nursing in a unit characterized by valued organizational attributes, reported findings similar to those of Knaus and colleagues. Units characterized by a high perceived level of nurse-physician collaboration, highly rated objective nursing performance, and significantly more positive organizational climate were associated with desirable clinical outcomes such as low mortality ratio, no new complications, and high patient satisfaction.
In a later analysis based on the APACHE III study, Zimmerman and colleagues (1993) report on the difficulty of measuring ICU performance with currently available measures. The authors attempted to identify and evaluate those organizational and management factors that might be associated with ICU effectiveness and risk-adjusted mortality rates. Nine ICUs participating in the APACHE III study were identified. These units varied significantly in their risk-adjusted mortality rates. Structural and organizational data were collected, and on-site observations of the nine units were conducted. The authors report that the on-site analysis failed to identify the units with significantly high or low performance levels and that organizational and management practices are not sufficient to identify levels of ICU performance as measured by risk-adjusted patient survival rates. High levels of team orientation among care givers was associated with ICU efficiency, but not with ICU risk-adjusted mortality rates. Furthermore, units with lower levels of performance exhibited some excellent organizational practices and high-performing units exhibited some poor organizational practices.
Hartz and colleagues (1989) analyzed Medicare discharges in 1986 and found that the percentage of nursing personnel who were RNs was one of the five significant predictors of hospital mortality rates. Specifically, hospitals with a higher percentage of RNs and hospitals with a higher staffing levels had lower adjusted mortality rates.
Verran (Part II of this report)8 prepared for the committee a detailed review of the literature on quality of hospital care, organizational variables, and nurse staffing. On the basis of both published research and research recently reported at an invitational conference, she concluded that (1) the proportion of RNs on a nursing staff has a positive influence on severity-adjusted Medicare mortality rates; (2) a professional practice environment (defined as a unit-level self-management model including participant decision making, use of primary nursing, peer review, and a salaried status for RN staff) has a beneficial influence on severity-adjusted Medicare mortality rates, over and above the influence of staffing mix; and (3) implementation of a professional practice model is cost neutral.
The IOM committee commissioned this paper from Joyce Verran. The committee appreciates her contributions. The full text of the paper can be found in Part II of this report.
Verran discussed several methodological problems with existing studies: low sample size, in part reflecting the high cost of data collection, at the unit of the hospital in which nurses practice; the lack of nurse-sensitive patient outcome measures; the fact that many pertinent outcomes occur after the patient leaves the hospital and, relatedly, the lack of longitudinal data on patients postdischarge; and the inconsistency of outcome measures among studies, which makes broad generalizations difficult (Verran, Part II of this report).
She notes that much of the research to date has not addressed the association between nurse staffing skill mixes and numbers and quality of care. The studies reported at the conference, for instance, were conceptualized and funded in response to the nursing shortage of the late 1980s. Consequently, the variables used in those studies are more focused on nurse retention than on the effects of nurse staffing on patient outcomes. Moritz (1995) also gives an exhaustive review of outcomes and effectiveness research that bears on questions of quality of care and the link between nursing processes or outcomes and broader patient outcomes. For example, in reviewing the available information on the importance of organizational models, she draws attention to the growing body of work on clinical outcomes and health status as it applies to nursing research. Her paper does not mitigate the view, however, that at present the evidence of the impact of nurse staffing and mix on quality of hospital care should be viewed as, at best, suggestive of a relationship, but not conclusive.
In investigating the relationship of nursing staff and patient outcomes, hospital organization has received comparatively little attention even though a substantial body of research documents a relationship between organizational attributes of hospitals and nurse satisfaction and turnover. Far less attention has been given to the relationship between nursing organization and patient outcomes (Kramer and Schmalenberg, 1988a,b; Aiken et al., 1994). When institutional attributes or characteristics are the focus of hospital mortality studies, many organizational correlates are examined, of which nursing often is one (Shortell and Hughes, 1988; Hartz et al., 1989). Nurse-to-patient ratios or nurses as a percentage of total nursing personnel are sometimes found to be significant correlates of patient mortality rates, but usually these studies give little consideration to the mechanisms by which staffing ratios might affect patient outcomes.
Several recent studies point to the organization of nursing within hospitals as the operant mechanism by which nurse staffing affects patient outcomes. Shortell and colleagues (1994) studied the role of nurse staffing and managerial factors as determinants of performance of hospital intensive care units (ICU). Although their sample size was small (42 ICUs), this research is important in its attempt to isolate the role of various organizational variables and staffing on hospital performance. In particular, controlling for patient case-mix, the authors found that the availability of state-of-the-art technology was a statistically significant determinant of risk-adjusted mortality. The quality of the care giver interaction affected risk-adjusted ICU length of stay. When these and other factors are held constant,
the ICU nurse-to-patient ratio had no effect on any of the outcomes analyzed, giving weight to the suspicion that variables other than nurse staffing per se affect patient outcomes in the hospital context.
In a study examining Medicare mortality rates, Aiken and colleagues (1994) found that magnet hospitals (that is, hospitals with low RN turnover and vacancy rates and high levels of RN satisfaction) have lower patient mortality than control hospitals. In a presentation to the IOM committee (October, 1994), Aiken summarized the study findings. These findings indicate that lower Medicare mortality rates, as well as improved work-related well-being for RNs, are linked to hospital organization characteristics that result in RNs having: (1) more autonomy to provide care in their professional roles and within their areas of expertise; (2) greater control over what other care givers do in the patient care environment and over resources; and (3) well-documented and well-developed professional relationships with physicians.
Aiken and colleagues also conducted a study of specialized AIDS care units. The combined results of these two studies provide interesting information concerning the organization of nursing care. The magnet hospital study indicates that the preferred organizational structure is one in which the hospital management sees its primary responsibility as delivering patient care, and therefore both places a high value on the quality of nursing services and actively supports the professional role of nursing services. The research on specialized AIDS care units demonstrated how, in the absence of the preferred hospital-wide organization of nursing services, unit-level organization of care can help create environments where the RN autonomy and control that promote lower mortality rates can develop. Specifically, Aiken and colleagues found that AIDS care units foster RN autonomy and control through RN specialization (which promotes autonomy and interaction with physicians based on mutual expertise) and the correlation between patients' high care needs and RNs' areas of specialization. Furthermore, these two studies confirm that the same factors that lead hospitals to be identified as effective from the standpoint of the organization of nursing care are associated with lower mortality among Medicare patients.
Aiken and colleagues concluded that although RN-rich staffing ratios are sometimes associated with improved outcomes, the results of their research indicate that such staffing ratios are essentially a proxy measure for other organizational attributes of hospitals that grant nurses autonomy over their own practice and control of the resources necessary to deliver patient care and create good relationships with physicians. The committee concurs with the findings of Aiken and colleagues (1994, p. 783), "that the mortality effect derives from the greater status, autonomy and control afforded nurse in the magnet hospitals, and their resulting impact on nurses' [RNs'] behaviors on behalf of patients—i.e., this is not simply an issue of the number of nurses, or their mix of credentials."
Clearly, one of the research challenges in determining the relationship between staffing and quality of care has been the difficulty of isolating the factors
(and the relative importance of these factors) that are involved in producing improved patient outcomes. Aiken and Salmon (1994, p. 324) contend that "[n]urses place considerable importance on having high nurse-to-patient staffing ratios and that, in general, they have been reluctant to think critically about other strategies that could be as, or more, important in achieving good patient outcomes." The issue is not just a matter of staffing ratios. In the authors' view, what RNs do and how they do it are both more important than simply how many RNs there are. These studies suggest that RNs may have overestimated the value of staffing ratios and skill mix in hospitals and underestimated the importance of the organization of nursing. The committee is of the view that more attention to organizational factors can lead to more efficient use of RNs and other nursing personnel and at the same time improve patient outcomes.
In summary, the committee concludes that literature on the effect of RNs on mortality and on factors affecting the retention of RNs is available. But there is a serious paucity of recent research on the definitive effects of structural measures, such as specific staffing ratios, on the quality of patient care in terms of patient outcomes when controlling for all other likely explanatory or confounding variables. Part of the problem lies in the area of severity of illness and risk adjustment, where patient acuity is a significant factor. Across-the-board staffing ratios tend to assume that in some measure all patients are "alike" and can be cared for with the same level and type of resources. Equally difficult is the task of establishing ratios that will be appropriate for all settings and situations.
At least one committee member strongly supports mandated minimum staffing levels specific to different types of acute care units and facilities but recognizes that specifying any particular minimum level was beyond the scope and competency of the committee. All committee members support the current federal requirements and accreditation standards for nursing services and support the need for hospitals to maintain the highest possible standards for nursing care. Moreover, the committee agrees that hospitals should develop improved methods for matching patient needs (severity-of-illness or acuity measures) with the level and type of nurse staffing. The committee supports efforts to improve systems for planning appropriate nursing care as well as monitoring the outcomes of that care.
The committee believes that high priority should be given to obtaining empirical evidence that permits one to draw conclusions about the relationships of quality of inpatient care and staffing levels and mix. Such data should focus on nursing care and quality of care across institutions and within given institutions, and across departments and services. Existing work has not typically focused on isolating the contribution of nursing care in measuring the quality of patient care in hospitals.
Thus, the committee is convinced that more rigorous research on the relationship between nursing variables, broadly defined, and quality of care would
have significant payoffs for policymakers, nursing educators, and hospital administrators.
RECOMMENDATION 5-5: The committee recommends that the National Institute of Nursing Research (NINR) and other appropriate agencies fund scientifically sound research on the relationships between quality of care and nurse staffing levels and mix, taking into account organizational variables. The committee further recommends that NINR, along with the Agency for Health Care Policy and Research (AHCPR) and private organizations, develop a research agenda on staffing and quality of care.
Several other agencies of the U.S. Department of Health and Human Services also have a major role in health workforce data collection and research, including the Division of Nursing in the Bureau of Health Professions, the Health Resources and Services Administration; the National Center for Health Statistics (the principal health statistics agency of the federal government) in the Centers for Disease Control and Prevention; and the Health Care Financing Administration. Finally, private organizations could be partners in these research programs, for example, hospital associations (e.g., Hospital Research and Education Trust, the American Hospital Association [AHA]; and state hospital associations, some of which [e.g., in Maryland and New York State] conduct work of this sort already) and private philanthropic and research foundations, particularly those with long-standing interests in health personnel, organization of the health care sector, or quality of care.
A major part of any such research agenda might call for elaboration of the actual variables—in terms of structure, process, and outcome—that warrant high priority attention in studies of the relationship of nursing care, staffing patterns for nursing, to patient outcomes. As discussed below, for example, the American Nurses Association (ANA) has been developing quality indicators that warrant further investigation.
During its work, the committee heard a great deal about the need for more information on hospital quality of care to be made available to policymakers and the public. The reasons are several: to improve the workings of a competitive health care market, to enable the public to make better choices about health care plans, and generally to reflect the nation's expanding interest in generating and using information to help improve the quality and cost-effectiveness of health care. One notable indication of this movement is the growing belief in "report cards"—that is, summary collections of indicators or measures of health care
providers' performance. Such report cards may serve as conduits of information about the quality of inpatient care in individual institutions and facilities or entire health care systems.9 Despite the surge of interest and activity in report cards, however, they are still in the early stages of design and implementation. The ultimate feasibility of linking report card use to actual improvements in the quality of care remains to be seen.
One example of a report card approach that relates to the central questions before this committee—the role of nursing in quality of care and patient outcomes—is that being developed by Lewin-VHI for the ANA (ANA, 1995b; Lewin-VHI, 1995). Because very few of the report cards under development document the specific effects of nursing on the quality of care delivered in hospitals, ANA felt the need for quality indicators that would clarify elements of quality of care from a nursing practice perspective. The organization thus commissioned Lewin-VHI, in 1994, to develop performance indicators for nursing care in hospital settings.
In the first phase of the project, the developers identified 21 categories of measures with an apparent conceptual link to nursing. The candidate measures suffered from significant practical limitations, however: lack of a strong research base linking them to nursing outcomes, lack of specificity to nursing, lack of necessary data collection mechanisms, and lack of applicable risk adjustment systems. Thus, considerable empirical validation of the measures was needed.
The committee is informed that work on the second phase of the project is under way. Seven indicators have been identified for further research because the necessary data collection mechanism exists and/or because of their specificity to nursing: (1) patient satisfaction; (2) pain management; (3) skin integrity; (4) total
nursing care hours per patient; (5) nosocomial infections (urinary tract infection, pneumonia); (6) patient injury rate; and (7) assessment of patient care requirements (Telephone communications with Janet Heinrich, Director, American Academy of Nursing).
The committee commends the ANA for its exploratory efforts to develop a set of nursing care quality indicators. This research can set an important precedent and standard for the development of meaningful quality standards relating to nursing. It offers promise for further evolution of external regulatory quality assurance mechanisms (like those of the Joint Commission on the Accreditation of Health Care Organizations [JCAHO]) and for improved public information efforts. Nevertheless, the committee judges that in the future, a broader set of inputs from the nursing community and other affected parties is desirable. It also believes that efforts based solely in the private sector, with little or no public sector involvement, might be less useful than if federal and state perspectives were taken into account. Therefore:
RECOMMENDATION 5-6: The committee recommends that an interdisciplinary public–private partnership be organized to develop performance and outcome measures that are sensitive to nursing interventions and care, with uniform definitions that are measurable in a uniform manner across all hospitals.
Such a partnership should involve a group comprising the following: various professional associations of nurses (clearly including ANA but not limited to it); leaders in the nursing profession in areas such as quality assessment and improvement, health services research, and nursing education; hospital systems and associations; accrediting bodies that have long experience with setting quality standards and criteria; researchers and experts in administrative databases who are familiar with developing uniform minimum data sets in the health area; and government officials representing the health agencies that pay for care, monitor quality of care, or track education and training curricula, as well as have long experience in developing uniform minimum data sets for national use.
Legislative And Regulatory Requirements
Regulation of hospitals is a long-standing part of government responsibility. States have had their own licensing requirements for hospitals and other facilities since the early part of the century. Regulation has taken many forms, such as certification, licensure, and accreditation.10
Since Medicare and Medicaid legislation was passed in 1965, the Social Security Act has required that providers be certified as a condition of participation in the program. This is accomplished through mechanisms known as Conditions of Participation that are promulgated through specific standards in the Code of Federal Regulations. For hospitals to be so certified for participation, the Social Security Act requires that facilities be licensed and in good standing by the state. In addition, hospitals must meet all federal certification standards, and the federal HCFA is authorized to determine whether hospitals meet these federal requirements. HCFA may conduct on-site inspections to observe care and review records to determine compliance, or it may ask state agencies to carry out these surveys.
Under the Social Security Act, certification may also be based on accreditation, which is in turn based on a concept of deemed status . Hospitals found to meet accreditation standards by the JCAHO are deemed automatically to meet the federal Conditions of Participation in the Medicare program—in effect, they are considered to be certified to receive Medicare (and Medicaid) reimbursement. HCFA performs independent validation surveys of individual hospitals on a sample basis as an assurance that the federal government can rely on the JCAHO approach. In addition, accreditation by the JCAHO is a requirement for hospitals approved to conduct graduate medical education residency programs and is frequently a requirement for payment by health maintenance organizations (HMO) and health insurance companies.
Certification by JCAHO requires that (a) nursing care be provided on a 24-hour-a-day, 7-day-a-week basis; (b) nursing services show evidence to the surveyors that each patient's status is monitored; (c) provision of nursing care is coordinated with the provision of care by other professionals; and (d) specific patient care plans be in place and in use for each patient. Furthermore, JCAHO standards for nursing care place a new emphasis on the role, responsibility, qualifications, and accountability of the nurse executive, including the authority and responsibility for ensuring that standards of nursing practice are in place and meet JCAHO's patient care standards (JCAHO, 1994).
The hospital organization must provide "a sufficient number of qualified nursing staff members to assess the patient's nursing care needs; plan and provide nursing care interventions; prevent complications and promote improvement in the patient's comfort and wellness; and alert other care professionals to the patient's condition as appropriate" (JCAHO, 1994, p. 521). The "Care of Patients" chapter of the standards manual (part of the patient-focused section) directs attention to a wide set of elements of good care. These include the following:
- "Formulation, maintenance, and support of a patient-specific plan for care, treatment, and rehabilitation;"
- "Implementation of the planned care, treatment, and rehabilitation;"
- "Monitoring the patient's response to the care, treatment, and rehabilitation provided, the actions or interventions taken, and/or the outcomes of the care provided;"
- "Modification of the planned care, treatment, and rehabilitation … based on reassessment, the patient's need for further care, and the achievement of identified goals;" and
- Planning and coordination of the "[c]are, treatment and rehabilitation necessary after the patient's discharge from the organization" (JCAHO, 1994, p. 125).
A large majority of hospitals seek and attain accredited status.11 The JCAHO renders five types of accreditation decisions, depending on the extent to which hospitals are judged (on the basis of institution-specific survey data) to comply with published standards of performance; the five accreditation levels are (1) accreditation with commendation, (2) accreditation, (3) conditional accreditation, (4) provisional accreditation, and (5) not accredited.
Given the continued reliance of the federal government on this approach to certification for hospital reimbursement through federal health programs, the committee is encouraged by the evolution of JCAHO methods and standards in the past few years and by the more sophisticated attention being paid to the role of nursing care in those standards. It also takes note of a new initiative, the Council on Performance Measurement, which will serve as an advisory body for evaluation of performance measurement systems, especially with respect to considering whether they are suitable for incorporating into future accreditation processes.
All in all, therefore, the committee endorses the current federal requirements for hospitals to participate in Medicare, which incorporate the use of voluntary accreditation, to assure the quality of hospital care, and it is particularly supportive of requirements that call for matching nursing resources with patient needs. The committee believes that Congress ought to continue to support this element of assuring the quality of care in hospitals.
Hospital restructuring and redesign of staffing systems are undertaken for a variety of reasons that include controlling costs and adjusting to the dramatic changes in the delivery of health care. Hospitals are restructuring to maintain their economic viability, but they need to do so without adversely affecting the outcomes of the care they provide. The changes now taking place in the hospital sector involve major rethinking of the use of different types of clinical staff, as
well as reconfiguration of units, departments, and care teams. The redesign of nursing services also is leading to changes in the roles and responsibilities of RNs and to increased emphasis on interdisciplinary teams. These developments have prompted uncertainty in employment and great concern among RNs about the potential for erosion of quality of hospital care, and about their own well-being.
In this chapter the committee has examined these concerns, specifically whether quality has deteriorated and whether empirical evidence exists of a link between the number and skill mix of nursing personnel and the quality of patient care. The committee has found that little empirical evidence is available to support the anecdotal and other informal information that hospital quality of care is being adversely affected by hospital restructuring and changes in the staffing patterns of nursing personnel. At the same time the committee notes a lack of systematic and ongoing monitoring and evaluating of the effects of changes resulting from organizational redesign and reconfiguration of staffing on patient outcomes.
Researchers have concluded that although RN-rich staffing ratios are sometimes associated with improved outcomes, the results of their research indicate that they are essentially proxy measures for organizational measures. For quality of care changes, the committee was unable to isolate a number-of-RNs effect from the organizational and related factors attending different levels of staffing. The committee concludes that high priority should be given to obtaining empirical evidence that permits one to draw conclusions about the relationships of quality of inpatient care and staffing levels and mix.
The committee, however, is concerned about the paucity of objective research on the relationship between staffing and quality, and the effects of restructuring. The committee concludes that a clear need exists for a system for monitoring and evaluating the impact of the rapidly changing delivery system on the quality of patient care and the well-being of nursing staff. For this reason, it has advanced several recommendations intended to provide better information on hospital restructuring and to help in delineating those factors that affect patient outcomes. It also calls for the development of a research agenda in this area and for the articulation of reliable, valid, and practical measures of structure, process, and outcome to be used in quality-of-care research as well as quality assurance and improvement programs. A systematic effort is needed at the national level to collect and analyze current and relevant data and develop a research and evaluation agenda so that informed policy development, implementation and evaluation are undertaken in a timely manner.
Some broader issues of changes in nursing services, such as the enhanced responsibilities of advanced practice nurses, and the use of ancillary nursing personnel and their competency, cut across the straightforward issue of the relationship between nurse staffing and quality of hospital care. In reflecting on the role of nursing personnel in the future, therefore, the committee has also proposed recommendations about these specific types of nursing personnel.