- Email: [email protected]
Testing a community level research utilization intervention
Testing a Community Level Research Utilization Intervention Martha Shively, Barbara Riegel, Diane Waterhouse, David Burns, Kathleen Templin, and Tanna Thomason
Research supports changing practice from heparinized to saline flushes for adults, yet heparin continues to be used in clinical practice. The primary aim of this study was to test the effectiveness of a community level innovation diffusion intervention as a method of stimulating research utilization at three acute care facilities in one community simultaneously. The change advocated was the flushing of intermittent intravenous devices (IID) with saline rather than heparin. Pilot IID units were chosen on which to test the change from the use of heparin to saline flushes. The innovation focused on staff empowerment and was implemented in five phases. The change to saline flushes was successful and has been maintained.
Copyright © 1997 by W.B. Saunders Company
ANY INVESTIGATORS have studied the benefit of heparinized saline over normal saline flushes for peripheral intermittent intravenous devices (IID) or "heparin locks." Heparin is not a benign medication; heparin-induced thrombocytopenia and drug incompatibilities are the major reasons for discontinuing the use of heparin if it is unnecessary. Research supports changing practice from heparinized to saline flushes for adults, yet hepariu continues to be used in clinical practice (Epperson, 1984; Goode, Titler, et al., 1991). The purpose of this study was to test a community level intervention for using this research base in clinical practice.
M
BACKGROUND
A Community Level Research Utilization Framework Several models of research utilization have been published (Funk, Tornquist, & Champagne, 1989; Horsley, Crane, & Bingle, 1978; Stetler, 1994), but none of these models takes into account the power of the community (rather than individuals) to facilitate change and maintain change. A community-based approach views the individual as part of a larger system with the power to influence behavior through four major mechanisms: (a) values and norms, (b) legitimization of new behaviors, (c) participation of community leaders, and (d) individual responsibility, control, and ownership of the planned change (Murdaugh & Vanderboom, 1997). Successful community level change models incorporate both social change and communication in implementation and evaluation of an intervention or innovation. Social change involves a molding of Applied Nursing Research,Vol. 10, No, 3 (August), 1997: pp 121-127
community values and norms (Stokols, 1992) and is maintained through communication. Change is thought to occur when a community is empowered to discuss the need for change and potential strategies to accomplish change (Wallerstein, 1992). Innovation diffusion is a commonly used model of communicating new ideas in a community. Innovation refers to an object, process, practice, or knowledge that is new to the potential adopter (Rogers & Shoemaker, 1971). Characteristics of From the VA Medical Center and San Diego State University; Sharp HealthCare; Children's Hospital, San Diego, CA; Palomar Medical Care, San Diego, CA; VA Medical Center, San Diego, CA; and Sharp Memorial Hospital, San Diego, CA. Martha Shively, PhD, RN, VA Medical Center and San Diego State University; Barbara Riegel, DNSc, RN, CS, FAAN San Diego State University and Sharp HealthCare; Diane Waterhouse, MS, RN, PNE Children's Hospital, San Diego, CA; David Bums, MN, RN, CCRN formerly of Palomar Medical Center, San Diego, CA; Kathleen Templin, MN, RN, CCRN, VA Medical Center, San Diego, CA; Tanna Thomason, MS, RN, CCRN, Sharp Memorial Hospital, San Diego, CA. Funded by grants from San Diego State University and the San Diego State University Institatefor Nursing Research. The authors thank Nancy Donaldson, DNSc, RN, and Suzanne Prevost, PhD, RN, CCRN, for reviewing the intervention protocol and Nancy Donaldson, DNSc, RN for reviewing a prior draft of this manuscript. Dierdre Thornlow, MN, RN is acknowledged for her assistance in developing the data collection methods. Ira Amargo, BSN, RN, Laura Benson, RN, Florentina Pascual, RN, Lisa Wagner, RN, Vicki DeBaca, DNSc, RN, and Ramona Browne, RN are acknowledged for their assistance with data collection and implementation. Address reprint requests to Martha Shively, PhD, RN, Nursing Service~Research, VA Medical Center (118), 3350 La Jolla Village Dr, San Diego, CA 92161-4189. Copyright © 1997 by W.B. Saunders Company 0897-1897/97/1003-000655.00/0 121
122
the innovation influence the speed and extent of adoption. An innovation with relative advantage, low complexity, and high visibility is relatively likely to be adopted as are innovations that can be tested on a limited basis and those that are congruent with existing norms and values (Murdaugh & Vanderboom, 1997). In a 1991 editorial, Downs argued that change in practice is "a function of how long a practice has been in place, the faith clinicians have in its efficacy, and lack of obvious feedback from its use" (p. 323). Change does not occur automatically as the result of understanding a research report advocating an innovation. In the Rogers (1995) model of innovation diffusion, the innovation adopter goes through five stages of change in a time-ordered sequence and can decide not to use the innovation at any time during the process. The five stages are: (a) knowledge-gain understanding about the innovation; (b) persuasion-acquire more information and form attitudes about innovation; (c) decisiontrial innovation and decide to accept or reject; (d) implementation-use innovation and possibly modify it to fit needs; and (e) confirmation-seek reinforcement of innovation decision (Figure 1). According to this framework, characteristics of the individual and the community influence the process of innovation diffusion and research utilization. Therefore, an intervention was designed to maximize community influence and empower individuals to direct their practice. It was hypothesized that such an intervention would lead to successful change in practice.
The Innovation: Flushing Solution Heparin continues to be used to flush IID despite a decade of research urging nurses to change practice. In the first large, well-controlled, doubleblind study, Epperson (1984) examined the effectiveness of 0.9% sodium chloride injection alone compared to normal saline injection with 10 or 100 units of heparin/ml. A total of 412 consecutive patients were randomly assigned by nursing unit to receive all three solutions. No significant differences were found in patency loss or associated phlebitis. This early study supported that large doses of heparin were not required to maintain IID patency or avoid associated phlebitis. Several subsequent studies were conducted, most of which focused on a comparison of normal saline versus 10 units heparirdml (e.g., Garrelts,
SHIVELY ET AL
Knowledge
(staqe 1)
•
posters
•
bulletin boards
•
journal clubs
•
article copies
•
short summaries
P e r s u a s i o n (staqe 2) •
site coordinator
•
position paper
D e c i s i o n fstaae 3~ •
trial change
•
monitor outcomes
I m p l e m e n t a t i o n fstaae 4) •
yes or no
Confirmation •
tstage 5)
site coordinators worked to maintain the change
Figure 1. Intervention modeled on the process of innovation.
LaRocca, Ast, Smith, & Sweet, 1989; Lombardi, Gundersen, Zammet, Walters, & Morris, 1988; Taylor, Hntchison, Milliken, & Larson, 1989). Dependent or outcome variables usually included IID patency and associated phlebitis, but some investigators also looked at cost (Cyganski, Donahue, & Heaton, 1987; Garrelts et al., 1989; Taylor et al., 1989), nursing time (Taylor et al., 1989), clots (Garrelts et al., 1989; Taylor et al., 1989), device longevity (Lombardi et al., 1988), infiltration (Taylor et al., 1989), and leaking (Taylor et al., 1989). All studies but one were conducted in an adult, medical-surgical population (Lombardi et al., 1988). Few differences were demonstrated by any investigative team. Taylor et al. found that IID flushed with saline lasted longer than those flushed with heparin. Garrelts et al. found more clots with saline than heparin but no associated adverse outcomes. Cyganski et al. found that IID maintained with saline required twice as malay restarts as those maintained with heparin. Two meta-analyses of this research base were published almost simultaneously in 1991 (Goode, Titler, et al., 1991; Peterson & Kirchhoff, 1991).
RESEARCH UTILIZATION
Both teams concluded that saline was as effective as heparin in maintaining patency, preventing phlebitis, and increasing duration in IID. These benefits continue to be shown by subsequent investigators (Geritz, 1992; Shoaf & Oliver, 1992). METHODS
The primary aim of this study was to test the effectiveness of a community level innovation diffusion intervention as a method of stimulating research utilization at several acute care facilities in one community simultaneously. The change advocated was the flushing of IID with saline rather than heparin.
Settings and Subjects An informal survey of all hospitals in the San Diego vicinity revealed that three hospitals used heparin for IID flushes and were not in the process of changing from heparin to saline. All three hospitals participated in the study. A nonrandom sample of influential staff nurses was recruited from three designated pilot units. These staff were identified by coinvestigators who worked at the hospitals and on the recommendation of others as integral to the change/decision process on that unit. Participation was anonymous and voluntary.
Intervention A community level model of innovation diffusion was used to plan the intervention. The investigators, who are considered community leaders, legitimized the proposed innovation by advocating that change be considered. The change itself was advantageous (e.g., heparin-induced thrombocytopenia could be avoided) and easy to do. The recently published meta-analyses had given the issue high visibility. Other hospitals in the community had already changed so the innovation was congruent with community norms and values. An opportunity to test the change before making a commitment was advocated. Diffusion of the innovation focused on staff empowerment. In the first phase staff were helped to gain understanding about the innovation (Figure 1). The research related to flushing IID with heparin versus saline was provided in the form of posters, bulletin board presentations, journal clubs, article copies, and short written summaries of pertinent studies. Persuasion, the second phase of the innovation
123
diffusion, involved use of a site coordinator to champion the change and help the staff form positive attitudes about the innovation. A written recommendation/position paper was presented to the appropriate formal approval board (e.g., Pharmacy and Therapeutics Committee) at each hospital. The decision phase was facilitated by a trial of the innovation with monitoring of IID patency, associated phlebitis, and duration before and after the change in practice. Implementation, the fourth phase, occurred when the staff chose whether or not to use saline routinely to flush IID. The final phase of confirmation or reinforcement of the decision was facilitated by the site coordinators who were encouraged to help maintain the change in practice when the study was over. Instruments
Demographic data collected included hospital size, unit size, hospital teaching status, work role, access to a clinical nurse specialist, years as a nurse, years on current unit, education, age, gender, and race. Individual innovativeness was predicted to influence willingness to change (Rogers, 1995). Innovativeness was theoretically defined as a personality trait characterized by a willingness to change and measured by the Innovativeness Scale (Hurt, Joseph, & Cook, 1977). The Innovativeness Scale consists of 20 declarative statements expressing a viewpoint on innovativeness, such as "I am generally cautious about accepting new ideas." Each of the 20 items is measured using a Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). Possible scores range from 20 to 100; higher scores indicate a higher degree of innovativeness. A split-half reliability of .94, and construct and predictive validity were shown in previous research (Hurt et al., 1977). For this study, the coefficient alpha was 0.81. Five adopter categories for innovativeness have been identified: innovators, early adopters, early majority, late majority, and laggards (Rogers & Shoemaker, 1971). Innovators are eager to try new ideas. Early adopters are sought after because they are informed about the innovation. The early majority adopt new ideas just before the average member of a social system and may deliberate a long time before adopting a new idea. The late majority do not adopt a strategy until most others in their social system have done so. The laggards are the last to adopt an innovation. One's adopter
124
category is thought to be a trait that is not necessarily amenable to change. The Innovativeness Scale measures these adopter categories. Factors helping and hindering change were measured using an investigator-designed descriptive checklist entitled "Preference for Changing Practice." Eleven categories of factors thought to help or hinder the process of change were generated from theory, the research utilization literature, and personal experience. Those categories are listed with an example of each: • Comparisons ("other hospitals have changed practice to saline only"; 4 items) • Research ("critique of research findings"; 9 items) • Administrative/organizational ("administrative pressure"; 6 items) • Key individuals ("colleague or peer"; 10 items) • Meetings/discussions ("staff involved in decision-making process"; 9 items) • Change process ("steps of change protocol clearly described"; 10 items) • Information media ("bulletin board presentation"; 10 items) • Unit activities ("Quality Assurance Committee"; 6 items) • Patient/workload ("effect on patient outcomes"; 6 items) • Economics ("cost of heparin flush solutions"; 4 items) • Personal ("My input was sought"; 11 items). After identifying each of the 85 items as a help, a hindrance, neutral, or not applicable, subjects were asked to name the one factor that most helped and the one factor that most hindered the change process on their unit. Frequency distributions were analyzed for each item. Face validity was evaluated by having the checklist reviewed for completeness, readability, and ease of use by five nurses knowledgeable in research utilization theory. The checklist was revised based on the suggestions of the reviewers before being used in the study.
Protocol The protocol was reviewed by appropriate Committees on Protection of Human Subjects. Three pilot units were chosen on which to test the change from the use of heparin to saline flushes. Pilot units were selected based on acute medical-
SHIVELY ET AL
surgical patient population and staff desire to participate. The intervention was provided by the site coordinator with assistance from the project director. Initially, innovativeness data were collected followed by 1 month of monitoring patency, phlebitis, and duration of liD flushed with heparin (usual care). The trial change to saline flushes was then made. No data were collected for approximately 1 month to allow for stabilization of the change from heparin to saline. After implementing the change to saline flushes, all staff were asked about their desire to change practice. That decision was made in a staff meeting; staff were free to decide whether or not to change practice on their unit but the whole unit had to change if that was the majority decision. One month after the change, IID patency, phlebitis, and duration were measured. Factors helping and hindering the change were collected from the participants. To elicit factors that influence success, individual staff members were asked whether or not they had wanted the change to occur when the change was being considered. They also were asked if they expected the saline flush change to be maintained on their unit. The 1 month time frame was determined by staff consensus to be sufficient for implementing and stabilizing the change on the pilot units. Data were analyzed for each hospital and then combined because the differences were not significant. Descriptive statistics were used to describe characteristics (i.e., innovativeness score, helping/hindering factors, and demographics) of the participants. RESULTS
One of the three hospitals decided to change practice based on the literature review prepared by the investigators. Administrators at that hospital chose not to expend the time and resources on the full intervention and were successful in changing practice without the staff empowerment and innovation testing components. Apparently there was little resistance to the new practice from staff but no formal data are available to confirm this observation. At the other two hospitals, a total of 46 nurses agreed to lead the innovation and participate in the study. This number reflects approximately 67% of the staff on all pilot units; 32 from hospital #1 (82% of staff on two pilot units) and 14 from hospital #2
RESEARCH UTILIZATION
(47% of staff on one pilot unit). Hospital #1 is a 350-bed federally funded, university affiliated hospital and Hospital #2 is a 250-bed nonprofit community hospital; both are located in Southern California. The pilot units were adult acute care, medicalsurgical units. Participants in this study were primarily registered nurses working as staff (n = 43; 93%); three were licensed vocational nurses. Demographic data were available only on individuals completing the Preference for Changing Practice survey (n = 25). These subjects were primarily women (22/25, 88%), White (13/25, 52%) or Asian (10/25, 40%), with an average age of 43 years and an average of 18 years experience as a nurse. Education varied widely: vocational nurse (n = 2), associate degree (n = 6), diploma (n = 5), baccalaureate degree (n = 9), master's degree (n = 1), and did not respond (n = 2). The mean innovativeness score for the subjects from hospital #1 was 81 _+ 7 (n = 32) and 76 _+ 8 (n = 14) for those from hospital #2. These scores were not significantly different and thus were combined. The adopter categories for this sample (N = 46) were identified by segmenting the sample distribution according to the obtained standard deviation (Rogers, 1995) as follows: innovators, n = 8 (17%); early adopters, n = 15 (34%); early majority, n = 13 (28%); late majority, n = 10 (22%); and laggards, n = 0 (0%). All three units chose to change practice; both hospitals instituted a housewide change to saline once the pilot units had made the decision to change. Of the 25 nurses who completed the Preference for Changing Practice survey (54% response rate), only 15 (60%) reported wanting to change to saline flushes before the intervention. But after the intervention, 20 of the 25 (80%) predicted that the saline flush practice would be maintained. Almost 2 years later, both hospitals have maintained the practice of flushing IID with saline. The factor from the Preferences for Changing Practice survey that most positively influenced the change process was information about previous research findings. The staff also indicated they were more comfortable making the change knowing that it had been done safely and successfully elsewhere (e.g., "Comparisons"). Workload was the major hindrance to change but several individu-
125
als noted that changing to saline saved them time and thus became a motivator. DISCUSSION AND IMPLICATIONS
The community level research utilization intervention appears to have influenced a practice change at the three hospitals despite the fact that the intervention was conducted with only three representative units, and one hospital changed before full protocol implementation. Further, that change has been maintained for over 2 years. It appears the intervention may have been effective but other factors also may have contributed to the change. Did the intervention itself direct the change in practice? What component of the intervention was most effective? What other factors could have influenced the change? Funk, Champagne, Wiese, and Tornquist (1991) reported barriers to research utilization based on a survey of 1,989 full-time practicing registered nurses. Clinicians reported insufficient authority to change practice, insufficient time to implement changes, and unsupportive administrators, physicians, and staff as the primary impediments to implementing research findings in practice. Based on these results, the current investigators believe that a key component of the community level intervention may have been the staff empowerment component. Prior investigators have used a "top down" approach to changing practice (Goode, et al., 1993; McAllister, Lenaghan, & Tosone, 1993), by first decreasing resistance from upper administration and powerful potential resistors, and then convincing the staff. For this study, the project team sought to convince staff first and then help them decrease resistance of influential others. Giving staff the decision-making power seems to have been an important part of the intervention. Another component of the intervention the staff noted was particularly important was the education about the research findings of others. This is consistent with the findings of Goode, Butcher, et al. (1991), who argued that education is an essential component of the change process. When staff understand that practice changes are based on research they may be less resistant to change. Most models of research utilization advocate data collection or testing the innovation as a powerful method of convincing staff of the need to
126
SHIVELY ET AL
change practice. The data collection itself was not identified as a major motivator in this study because one hospital changed without any data collection. This finding is consistent with that of McAllister et al. (1993) who argued that a formal trial was unnecessary; they were successful in changing the liD flushing solution from heparin to saline without the collection of additional data. Staff innovativeness may also have been an important factor in the success of this intervention as the innovation adopter scores were skewed toward the innovator/early adopter end of the curve. This finding suggests that this sample was composed primarily of innovative individuals willing to consider change. Some staff from the pilot units already used saline flushes for IID based on their own awareness of the research and despite hospital protocol specifying heparin. These informed staff were instrumental in communicating knowledge and advocating for innovativeness and change. Such individuals were targeted as informal champions who could be influential in the process of change. Further testing of a community level research utilization framework is indicated, especially in this era of staffing cuts when clinical nurse specialists and unit level managers may no longer be immediately available to guide unit practice. The model used in this application could be initiated by a core group of community leaders (e.g., faculty) and result in significant changes in several facilities simultaneously. Further research is needed to determine the essential or sufficient steps involved in research utilization. For example, these results and those of McAllister et al. (1993) show that testing the innovation is not essential but the conditions under which the trial could be eliminated remain unclear. The importance of this step may depend on characteristics of the staff (e.g., educational preparation), the research base (e.g., number of prior studies), or the organization (e.g., shared governance) but these hypotheses require testing. Studies in which compo-
nents of a multifactorial intervention are manipulated would allow us to identify necessary and sufficient steps to a successful research utilization intervention.
Limitations This study was limited by the multifactorial nature of the intervention which makes it impossible to clearly identify which component was most effective. History is another threat to the internal validity of this study as the intervention was introduced at a time when numerous studies demonstrating the safety and effectiveness of saline flushes were being published. These publications facilitated the community level intervention by legitimizing the innovation and making the need to change more visible. The nurses may have decided to change practice at some point in the future even without the intervention, however, simply based on the amount of data available in the literature. The fact that other hospitals in the community had changed their practices within the previous 5 years whereas these three hospitals had not changed, however, suggests that change was not imminent. Other important potentially confounding variables include dynamics at each institution at the time of the study (e.g., personnel layoffs, administrative shifts, financial considerations) and responses of other staff (e.g., physicians, administrators) to the innovation. Because this was a community level intervention implemented by community leaders and directed toward nursing practice, however, the investigators felt that these variables were not essential to include in data collection. The measures of individual characteristics allowed the investigators to adequately describe the characteristics of staff leaders. A factor not described, however, was role accountability (individual responsibility) for change, which could significantly influence the success of a planned change. This factor was not included on the checklist designed to measure preferences for changing practice, but it should be added for future studies.
REFERENCES Cyganski, J.M., Donahue, J.M., & Heaton, J.S. (1987). The case for the heparin flush. American Journal of Nursing, 87, 796-797. Downs, ES. (1991). How to make a difference [editorial]. Nursing Research, 40, 323. Epperson, E.L. (1984). Efficacy of 0.9% sodium chloride
injection with and without heparin for maintaining indwelling intermittent injection sites. Clinical Pharmacy, 3, 626-629. Funk, S.G., Champagne, M.T., Wiese, R.A., & Tornquist, E.M. (1991). Barriers to using research findings in practice: The clinician's perspective. Applied Nursing Research, 4, 90-95. Funk, S.G., Tornquist, E.M., & Champagne, M.T. (1989). A
RESEARCH UTILIZATION
model for improving dissemination of nursing research. Western Journal of Nursing Research, 11, 361-367. Garrelts, J.C., LaRocca, J., Ast, D., Smith, D.E, & Sweet, D.E. (1989). Comparison of heparin and 0.9% sodium chloride injection in the maintenance of indwelling intermittent i.v. devices. Clinical Pharmacy, 8, 34-39. Geritz, M.A. (1992). Saline versus heparin in intermittent infuser patency maintenance. Western Journal of Nursing Research, 14, 131-141. Goode, C.J., Butcher, LA., Cipperley, J.A., Ekstrom, J., Gosch, B.A., Hayes, J.E., Lovette, M.K., & Wellendorf, S.A. (1991). Research utilization: A study guide [Videotape]. (Available from Horn Video Productions 700 E 2nd St, Ida Grove, IA 51445). Goode, C.L, Kleiber, C., Titler, M., Small, S., Rakel, B., Steelman, V.M., Walker, J.B., & Buckwalter, K.C. (1993). Improving practice through research: The case of heparin vs. saline for peripheral intermittent infusion devices. MEDSURG Nursing, 2, 23-27. Goode, C.J., Titler, M., Rakel, B., Ones, D.S., Kleiber, C., Small, S., & Triolo, RK. (1991). A meta-analysis of effects of heparin flush and saline flush: Quality and cost implications. Nursing Research, 40, 324-330. Horsley, J.A., Crane, J., & Bingle, J.D. (1978). Research utilization as an organizational process. Journal of Nursing Administration, 8 (7), 4-6. Hurt, H.T., Joseph, K., & Cook, C.D. (1977). Scales for the measurement of innovativeness. Human Communication Research, 4, 58-65. Lombardi, T.E, Gundersen, B., Zammet, L.O., Waiters, J.K., & Morris, B.A. (1988). Efficacy of 0.9% sodium chloride injection with or without heparin sodium for maintaining
127
patency of intravenous catheters in children. Clinical Pharmacy, 7, 832-836. McAllister, C.C., Lenaghan, RA., & Tosone, N.C. (1993). Changing from heparin to saline flush solutions: A research utilization model for implementation. Journal of Emergency Nursing, 19, 306-312. Murdangh, C.L., & Vanderboom, C. (1997). Individual and community models for promoting wellness. Journal of Cardiovascular Narsing, 11 (3), 1-14. Peterson, EY, & Kirchhoff, K.T. (1991). Analysis of the research about heparinized versus nonheparinized intravascular lines. Heart & Lung, 20, 631-642. Rogers, E.M. (1995). Diffusion of innovations. (4th ed.). New York: Free Press. Rogers, E.M., & Shoemaker, F.E (1971). Communication of innovations: A cross cultural approach. New York: Free Press. Shoaf, J., & Oliver, S. (1992). Efficacy of normal saline injection with and without heparin for maintaining intermittent intravenous site. Applied Nursing Research, 5, 9-12. Stetler, C.B. (1994). Refinement of the Stetler/Marram model for application of research findings to practice. Nursing Outlook, 42, 15-25. Stokols, D. (1992). Establishing and maintaining healthy environments: Toward a social ecology of health promotion. American Psychologist, 47, 6-22, Taylor, N., Hutchison, E., Milliken, W., & Larson, E. (1989). Comparison of normal versus heparinized saline for flushing infusion devices. Journal of Nursing Quality Assurance, 3 (4), 49-55. Wallerstein, N. (1992). Powerlessness, empowerment, and health: Implications for health promotion programs. American Journal of Health Promotion, 6, 197-205.
Research supports changing practice from heparinized to saline flushes for adults, yet heparin continues to be used in clinical practice. The primary aim of this study was to test the effectiveness of a community level innovation diffusion intervention as a method of stimulating research utilization at three acute care facilities in one community simultaneously. The change advocated was the flushing of intermittent intravenous devices (IID) with saline rather than heparin. Pilot IID units were chosen on which to test the change from the use of heparin to saline flushes. The innovation focused on staff empowerment and was implemented in five phases. The change to saline flushes was successful and has been maintained.
Copyright © 1997 by W.B. Saunders Company
ANY INVESTIGATORS have studied the benefit of heparinized saline over normal saline flushes for peripheral intermittent intravenous devices (IID) or "heparin locks." Heparin is not a benign medication; heparin-induced thrombocytopenia and drug incompatibilities are the major reasons for discontinuing the use of heparin if it is unnecessary. Research supports changing practice from heparinized to saline flushes for adults, yet hepariu continues to be used in clinical practice (Epperson, 1984; Goode, Titler, et al., 1991). The purpose of this study was to test a community level intervention for using this research base in clinical practice.
M
BACKGROUND
A Community Level Research Utilization Framework Several models of research utilization have been published (Funk, Tornquist, & Champagne, 1989; Horsley, Crane, & Bingle, 1978; Stetler, 1994), but none of these models takes into account the power of the community (rather than individuals) to facilitate change and maintain change. A community-based approach views the individual as part of a larger system with the power to influence behavior through four major mechanisms: (a) values and norms, (b) legitimization of new behaviors, (c) participation of community leaders, and (d) individual responsibility, control, and ownership of the planned change (Murdaugh & Vanderboom, 1997). Successful community level change models incorporate both social change and communication in implementation and evaluation of an intervention or innovation. Social change involves a molding of Applied Nursing Research,Vol. 10, No, 3 (August), 1997: pp 121-127
community values and norms (Stokols, 1992) and is maintained through communication. Change is thought to occur when a community is empowered to discuss the need for change and potential strategies to accomplish change (Wallerstein, 1992). Innovation diffusion is a commonly used model of communicating new ideas in a community. Innovation refers to an object, process, practice, or knowledge that is new to the potential adopter (Rogers & Shoemaker, 1971). Characteristics of From the VA Medical Center and San Diego State University; Sharp HealthCare; Children's Hospital, San Diego, CA; Palomar Medical Care, San Diego, CA; VA Medical Center, San Diego, CA; and Sharp Memorial Hospital, San Diego, CA. Martha Shively, PhD, RN, VA Medical Center and San Diego State University; Barbara Riegel, DNSc, RN, CS, FAAN San Diego State University and Sharp HealthCare; Diane Waterhouse, MS, RN, PNE Children's Hospital, San Diego, CA; David Bums, MN, RN, CCRN formerly of Palomar Medical Center, San Diego, CA; Kathleen Templin, MN, RN, CCRN, VA Medical Center, San Diego, CA; Tanna Thomason, MS, RN, CCRN, Sharp Memorial Hospital, San Diego, CA. Funded by grants from San Diego State University and the San Diego State University Institatefor Nursing Research. The authors thank Nancy Donaldson, DNSc, RN, and Suzanne Prevost, PhD, RN, CCRN, for reviewing the intervention protocol and Nancy Donaldson, DNSc, RN for reviewing a prior draft of this manuscript. Dierdre Thornlow, MN, RN is acknowledged for her assistance in developing the data collection methods. Ira Amargo, BSN, RN, Laura Benson, RN, Florentina Pascual, RN, Lisa Wagner, RN, Vicki DeBaca, DNSc, RN, and Ramona Browne, RN are acknowledged for their assistance with data collection and implementation. Address reprint requests to Martha Shively, PhD, RN, Nursing Service~Research, VA Medical Center (118), 3350 La Jolla Village Dr, San Diego, CA 92161-4189. Copyright © 1997 by W.B. Saunders Company 0897-1897/97/1003-000655.00/0 121
122
the innovation influence the speed and extent of adoption. An innovation with relative advantage, low complexity, and high visibility is relatively likely to be adopted as are innovations that can be tested on a limited basis and those that are congruent with existing norms and values (Murdaugh & Vanderboom, 1997). In a 1991 editorial, Downs argued that change in practice is "a function of how long a practice has been in place, the faith clinicians have in its efficacy, and lack of obvious feedback from its use" (p. 323). Change does not occur automatically as the result of understanding a research report advocating an innovation. In the Rogers (1995) model of innovation diffusion, the innovation adopter goes through five stages of change in a time-ordered sequence and can decide not to use the innovation at any time during the process. The five stages are: (a) knowledge-gain understanding about the innovation; (b) persuasion-acquire more information and form attitudes about innovation; (c) decisiontrial innovation and decide to accept or reject; (d) implementation-use innovation and possibly modify it to fit needs; and (e) confirmation-seek reinforcement of innovation decision (Figure 1). According to this framework, characteristics of the individual and the community influence the process of innovation diffusion and research utilization. Therefore, an intervention was designed to maximize community influence and empower individuals to direct their practice. It was hypothesized that such an intervention would lead to successful change in practice.
The Innovation: Flushing Solution Heparin continues to be used to flush IID despite a decade of research urging nurses to change practice. In the first large, well-controlled, doubleblind study, Epperson (1984) examined the effectiveness of 0.9% sodium chloride injection alone compared to normal saline injection with 10 or 100 units of heparin/ml. A total of 412 consecutive patients were randomly assigned by nursing unit to receive all three solutions. No significant differences were found in patency loss or associated phlebitis. This early study supported that large doses of heparin were not required to maintain IID patency or avoid associated phlebitis. Several subsequent studies were conducted, most of which focused on a comparison of normal saline versus 10 units heparirdml (e.g., Garrelts,
SHIVELY ET AL
Knowledge
(staqe 1)
•
posters
•
bulletin boards
•
journal clubs
•
article copies
•
short summaries
P e r s u a s i o n (staqe 2) •
site coordinator
•
position paper
D e c i s i o n fstaae 3~ •
trial change
•
monitor outcomes
I m p l e m e n t a t i o n fstaae 4) •
yes or no
Confirmation •
tstage 5)
site coordinators worked to maintain the change
Figure 1. Intervention modeled on the process of innovation.
LaRocca, Ast, Smith, & Sweet, 1989; Lombardi, Gundersen, Zammet, Walters, & Morris, 1988; Taylor, Hntchison, Milliken, & Larson, 1989). Dependent or outcome variables usually included IID patency and associated phlebitis, but some investigators also looked at cost (Cyganski, Donahue, & Heaton, 1987; Garrelts et al., 1989; Taylor et al., 1989), nursing time (Taylor et al., 1989), clots (Garrelts et al., 1989; Taylor et al., 1989), device longevity (Lombardi et al., 1988), infiltration (Taylor et al., 1989), and leaking (Taylor et al., 1989). All studies but one were conducted in an adult, medical-surgical population (Lombardi et al., 1988). Few differences were demonstrated by any investigative team. Taylor et al. found that IID flushed with saline lasted longer than those flushed with heparin. Garrelts et al. found more clots with saline than heparin but no associated adverse outcomes. Cyganski et al. found that IID maintained with saline required twice as malay restarts as those maintained with heparin. Two meta-analyses of this research base were published almost simultaneously in 1991 (Goode, Titler, et al., 1991; Peterson & Kirchhoff, 1991).
RESEARCH UTILIZATION
Both teams concluded that saline was as effective as heparin in maintaining patency, preventing phlebitis, and increasing duration in IID. These benefits continue to be shown by subsequent investigators (Geritz, 1992; Shoaf & Oliver, 1992). METHODS
The primary aim of this study was to test the effectiveness of a community level innovation diffusion intervention as a method of stimulating research utilization at several acute care facilities in one community simultaneously. The change advocated was the flushing of IID with saline rather than heparin.
Settings and Subjects An informal survey of all hospitals in the San Diego vicinity revealed that three hospitals used heparin for IID flushes and were not in the process of changing from heparin to saline. All three hospitals participated in the study. A nonrandom sample of influential staff nurses was recruited from three designated pilot units. These staff were identified by coinvestigators who worked at the hospitals and on the recommendation of others as integral to the change/decision process on that unit. Participation was anonymous and voluntary.
Intervention A community level model of innovation diffusion was used to plan the intervention. The investigators, who are considered community leaders, legitimized the proposed innovation by advocating that change be considered. The change itself was advantageous (e.g., heparin-induced thrombocytopenia could be avoided) and easy to do. The recently published meta-analyses had given the issue high visibility. Other hospitals in the community had already changed so the innovation was congruent with community norms and values. An opportunity to test the change before making a commitment was advocated. Diffusion of the innovation focused on staff empowerment. In the first phase staff were helped to gain understanding about the innovation (Figure 1). The research related to flushing IID with heparin versus saline was provided in the form of posters, bulletin board presentations, journal clubs, article copies, and short written summaries of pertinent studies. Persuasion, the second phase of the innovation
123
diffusion, involved use of a site coordinator to champion the change and help the staff form positive attitudes about the innovation. A written recommendation/position paper was presented to the appropriate formal approval board (e.g., Pharmacy and Therapeutics Committee) at each hospital. The decision phase was facilitated by a trial of the innovation with monitoring of IID patency, associated phlebitis, and duration before and after the change in practice. Implementation, the fourth phase, occurred when the staff chose whether or not to use saline routinely to flush IID. The final phase of confirmation or reinforcement of the decision was facilitated by the site coordinators who were encouraged to help maintain the change in practice when the study was over. Instruments
Demographic data collected included hospital size, unit size, hospital teaching status, work role, access to a clinical nurse specialist, years as a nurse, years on current unit, education, age, gender, and race. Individual innovativeness was predicted to influence willingness to change (Rogers, 1995). Innovativeness was theoretically defined as a personality trait characterized by a willingness to change and measured by the Innovativeness Scale (Hurt, Joseph, & Cook, 1977). The Innovativeness Scale consists of 20 declarative statements expressing a viewpoint on innovativeness, such as "I am generally cautious about accepting new ideas." Each of the 20 items is measured using a Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). Possible scores range from 20 to 100; higher scores indicate a higher degree of innovativeness. A split-half reliability of .94, and construct and predictive validity were shown in previous research (Hurt et al., 1977). For this study, the coefficient alpha was 0.81. Five adopter categories for innovativeness have been identified: innovators, early adopters, early majority, late majority, and laggards (Rogers & Shoemaker, 1971). Innovators are eager to try new ideas. Early adopters are sought after because they are informed about the innovation. The early majority adopt new ideas just before the average member of a social system and may deliberate a long time before adopting a new idea. The late majority do not adopt a strategy until most others in their social system have done so. The laggards are the last to adopt an innovation. One's adopter
124
category is thought to be a trait that is not necessarily amenable to change. The Innovativeness Scale measures these adopter categories. Factors helping and hindering change were measured using an investigator-designed descriptive checklist entitled "Preference for Changing Practice." Eleven categories of factors thought to help or hinder the process of change were generated from theory, the research utilization literature, and personal experience. Those categories are listed with an example of each: • Comparisons ("other hospitals have changed practice to saline only"; 4 items) • Research ("critique of research findings"; 9 items) • Administrative/organizational ("administrative pressure"; 6 items) • Key individuals ("colleague or peer"; 10 items) • Meetings/discussions ("staff involved in decision-making process"; 9 items) • Change process ("steps of change protocol clearly described"; 10 items) • Information media ("bulletin board presentation"; 10 items) • Unit activities ("Quality Assurance Committee"; 6 items) • Patient/workload ("effect on patient outcomes"; 6 items) • Economics ("cost of heparin flush solutions"; 4 items) • Personal ("My input was sought"; 11 items). After identifying each of the 85 items as a help, a hindrance, neutral, or not applicable, subjects were asked to name the one factor that most helped and the one factor that most hindered the change process on their unit. Frequency distributions were analyzed for each item. Face validity was evaluated by having the checklist reviewed for completeness, readability, and ease of use by five nurses knowledgeable in research utilization theory. The checklist was revised based on the suggestions of the reviewers before being used in the study.
Protocol The protocol was reviewed by appropriate Committees on Protection of Human Subjects. Three pilot units were chosen on which to test the change from the use of heparin to saline flushes. Pilot units were selected based on acute medical-
SHIVELY ET AL
surgical patient population and staff desire to participate. The intervention was provided by the site coordinator with assistance from the project director. Initially, innovativeness data were collected followed by 1 month of monitoring patency, phlebitis, and duration of liD flushed with heparin (usual care). The trial change to saline flushes was then made. No data were collected for approximately 1 month to allow for stabilization of the change from heparin to saline. After implementing the change to saline flushes, all staff were asked about their desire to change practice. That decision was made in a staff meeting; staff were free to decide whether or not to change practice on their unit but the whole unit had to change if that was the majority decision. One month after the change, IID patency, phlebitis, and duration were measured. Factors helping and hindering the change were collected from the participants. To elicit factors that influence success, individual staff members were asked whether or not they had wanted the change to occur when the change was being considered. They also were asked if they expected the saline flush change to be maintained on their unit. The 1 month time frame was determined by staff consensus to be sufficient for implementing and stabilizing the change on the pilot units. Data were analyzed for each hospital and then combined because the differences were not significant. Descriptive statistics were used to describe characteristics (i.e., innovativeness score, helping/hindering factors, and demographics) of the participants. RESULTS
One of the three hospitals decided to change practice based on the literature review prepared by the investigators. Administrators at that hospital chose not to expend the time and resources on the full intervention and were successful in changing practice without the staff empowerment and innovation testing components. Apparently there was little resistance to the new practice from staff but no formal data are available to confirm this observation. At the other two hospitals, a total of 46 nurses agreed to lead the innovation and participate in the study. This number reflects approximately 67% of the staff on all pilot units; 32 from hospital #1 (82% of staff on two pilot units) and 14 from hospital #2
RESEARCH UTILIZATION
(47% of staff on one pilot unit). Hospital #1 is a 350-bed federally funded, university affiliated hospital and Hospital #2 is a 250-bed nonprofit community hospital; both are located in Southern California. The pilot units were adult acute care, medicalsurgical units. Participants in this study were primarily registered nurses working as staff (n = 43; 93%); three were licensed vocational nurses. Demographic data were available only on individuals completing the Preference for Changing Practice survey (n = 25). These subjects were primarily women (22/25, 88%), White (13/25, 52%) or Asian (10/25, 40%), with an average age of 43 years and an average of 18 years experience as a nurse. Education varied widely: vocational nurse (n = 2), associate degree (n = 6), diploma (n = 5), baccalaureate degree (n = 9), master's degree (n = 1), and did not respond (n = 2). The mean innovativeness score for the subjects from hospital #1 was 81 _+ 7 (n = 32) and 76 _+ 8 (n = 14) for those from hospital #2. These scores were not significantly different and thus were combined. The adopter categories for this sample (N = 46) were identified by segmenting the sample distribution according to the obtained standard deviation (Rogers, 1995) as follows: innovators, n = 8 (17%); early adopters, n = 15 (34%); early majority, n = 13 (28%); late majority, n = 10 (22%); and laggards, n = 0 (0%). All three units chose to change practice; both hospitals instituted a housewide change to saline once the pilot units had made the decision to change. Of the 25 nurses who completed the Preference for Changing Practice survey (54% response rate), only 15 (60%) reported wanting to change to saline flushes before the intervention. But after the intervention, 20 of the 25 (80%) predicted that the saline flush practice would be maintained. Almost 2 years later, both hospitals have maintained the practice of flushing IID with saline. The factor from the Preferences for Changing Practice survey that most positively influenced the change process was information about previous research findings. The staff also indicated they were more comfortable making the change knowing that it had been done safely and successfully elsewhere (e.g., "Comparisons"). Workload was the major hindrance to change but several individu-
125
als noted that changing to saline saved them time and thus became a motivator. DISCUSSION AND IMPLICATIONS
The community level research utilization intervention appears to have influenced a practice change at the three hospitals despite the fact that the intervention was conducted with only three representative units, and one hospital changed before full protocol implementation. Further, that change has been maintained for over 2 years. It appears the intervention may have been effective but other factors also may have contributed to the change. Did the intervention itself direct the change in practice? What component of the intervention was most effective? What other factors could have influenced the change? Funk, Champagne, Wiese, and Tornquist (1991) reported barriers to research utilization based on a survey of 1,989 full-time practicing registered nurses. Clinicians reported insufficient authority to change practice, insufficient time to implement changes, and unsupportive administrators, physicians, and staff as the primary impediments to implementing research findings in practice. Based on these results, the current investigators believe that a key component of the community level intervention may have been the staff empowerment component. Prior investigators have used a "top down" approach to changing practice (Goode, et al., 1993; McAllister, Lenaghan, & Tosone, 1993), by first decreasing resistance from upper administration and powerful potential resistors, and then convincing the staff. For this study, the project team sought to convince staff first and then help them decrease resistance of influential others. Giving staff the decision-making power seems to have been an important part of the intervention. Another component of the intervention the staff noted was particularly important was the education about the research findings of others. This is consistent with the findings of Goode, Butcher, et al. (1991), who argued that education is an essential component of the change process. When staff understand that practice changes are based on research they may be less resistant to change. Most models of research utilization advocate data collection or testing the innovation as a powerful method of convincing staff of the need to
126
SHIVELY ET AL
change practice. The data collection itself was not identified as a major motivator in this study because one hospital changed without any data collection. This finding is consistent with that of McAllister et al. (1993) who argued that a formal trial was unnecessary; they were successful in changing the liD flushing solution from heparin to saline without the collection of additional data. Staff innovativeness may also have been an important factor in the success of this intervention as the innovation adopter scores were skewed toward the innovator/early adopter end of the curve. This finding suggests that this sample was composed primarily of innovative individuals willing to consider change. Some staff from the pilot units already used saline flushes for IID based on their own awareness of the research and despite hospital protocol specifying heparin. These informed staff were instrumental in communicating knowledge and advocating for innovativeness and change. Such individuals were targeted as informal champions who could be influential in the process of change. Further testing of a community level research utilization framework is indicated, especially in this era of staffing cuts when clinical nurse specialists and unit level managers may no longer be immediately available to guide unit practice. The model used in this application could be initiated by a core group of community leaders (e.g., faculty) and result in significant changes in several facilities simultaneously. Further research is needed to determine the essential or sufficient steps involved in research utilization. For example, these results and those of McAllister et al. (1993) show that testing the innovation is not essential but the conditions under which the trial could be eliminated remain unclear. The importance of this step may depend on characteristics of the staff (e.g., educational preparation), the research base (e.g., number of prior studies), or the organization (e.g., shared governance) but these hypotheses require testing. Studies in which compo-
nents of a multifactorial intervention are manipulated would allow us to identify necessary and sufficient steps to a successful research utilization intervention.
Limitations This study was limited by the multifactorial nature of the intervention which makes it impossible to clearly identify which component was most effective. History is another threat to the internal validity of this study as the intervention was introduced at a time when numerous studies demonstrating the safety and effectiveness of saline flushes were being published. These publications facilitated the community level intervention by legitimizing the innovation and making the need to change more visible. The nurses may have decided to change practice at some point in the future even without the intervention, however, simply based on the amount of data available in the literature. The fact that other hospitals in the community had changed their practices within the previous 5 years whereas these three hospitals had not changed, however, suggests that change was not imminent. Other important potentially confounding variables include dynamics at each institution at the time of the study (e.g., personnel layoffs, administrative shifts, financial considerations) and responses of other staff (e.g., physicians, administrators) to the innovation. Because this was a community level intervention implemented by community leaders and directed toward nursing practice, however, the investigators felt that these variables were not essential to include in data collection. The measures of individual characteristics allowed the investigators to adequately describe the characteristics of staff leaders. A factor not described, however, was role accountability (individual responsibility) for change, which could significantly influence the success of a planned change. This factor was not included on the checklist designed to measure preferences for changing practice, but it should be added for future studies.
REFERENCES Cyganski, J.M., Donahue, J.M., & Heaton, J.S. (1987). The case for the heparin flush. American Journal of Nursing, 87, 796-797. Downs, ES. (1991). How to make a difference [editorial]. Nursing Research, 40, 323. Epperson, E.L. (1984). Efficacy of 0.9% sodium chloride
injection with and without heparin for maintaining indwelling intermittent injection sites. Clinical Pharmacy, 3, 626-629. Funk, S.G., Champagne, M.T., Wiese, R.A., & Tornquist, E.M. (1991). Barriers to using research findings in practice: The clinician's perspective. Applied Nursing Research, 4, 90-95. Funk, S.G., Tornquist, E.M., & Champagne, M.T. (1989). A
RESEARCH UTILIZATION
model for improving dissemination of nursing research. Western Journal of Nursing Research, 11, 361-367. Garrelts, J.C., LaRocca, J., Ast, D., Smith, D.E, & Sweet, D.E. (1989). Comparison of heparin and 0.9% sodium chloride injection in the maintenance of indwelling intermittent i.v. devices. Clinical Pharmacy, 8, 34-39. Geritz, M.A. (1992). Saline versus heparin in intermittent infuser patency maintenance. Western Journal of Nursing Research, 14, 131-141. Goode, C.J., Butcher, LA., Cipperley, J.A., Ekstrom, J., Gosch, B.A., Hayes, J.E., Lovette, M.K., & Wellendorf, S.A. (1991). Research utilization: A study guide [Videotape]. (Available from Horn Video Productions 700 E 2nd St, Ida Grove, IA 51445). Goode, C.L, Kleiber, C., Titler, M., Small, S., Rakel, B., Steelman, V.M., Walker, J.B., & Buckwalter, K.C. (1993). Improving practice through research: The case of heparin vs. saline for peripheral intermittent infusion devices. MEDSURG Nursing, 2, 23-27. Goode, C.J., Titler, M., Rakel, B., Ones, D.S., Kleiber, C., Small, S., & Triolo, RK. (1991). A meta-analysis of effects of heparin flush and saline flush: Quality and cost implications. Nursing Research, 40, 324-330. Horsley, J.A., Crane, J., & Bingle, J.D. (1978). Research utilization as an organizational process. Journal of Nursing Administration, 8 (7), 4-6. Hurt, H.T., Joseph, K., & Cook, C.D. (1977). Scales for the measurement of innovativeness. Human Communication Research, 4, 58-65. Lombardi, T.E, Gundersen, B., Zammet, L.O., Waiters, J.K., & Morris, B.A. (1988). Efficacy of 0.9% sodium chloride injection with or without heparin sodium for maintaining
127
patency of intravenous catheters in children. Clinical Pharmacy, 7, 832-836. McAllister, C.C., Lenaghan, RA., & Tosone, N.C. (1993). Changing from heparin to saline flush solutions: A research utilization model for implementation. Journal of Emergency Nursing, 19, 306-312. Murdangh, C.L., & Vanderboom, C. (1997). Individual and community models for promoting wellness. Journal of Cardiovascular Narsing, 11 (3), 1-14. Peterson, EY, & Kirchhoff, K.T. (1991). Analysis of the research about heparinized versus nonheparinized intravascular lines. Heart & Lung, 20, 631-642. Rogers, E.M. (1995). Diffusion of innovations. (4th ed.). New York: Free Press. Rogers, E.M., & Shoemaker, F.E (1971). Communication of innovations: A cross cultural approach. New York: Free Press. Shoaf, J., & Oliver, S. (1992). Efficacy of normal saline injection with and without heparin for maintaining intermittent intravenous site. Applied Nursing Research, 5, 9-12. Stetler, C.B. (1994). Refinement of the Stetler/Marram model for application of research findings to practice. Nursing Outlook, 42, 15-25. Stokols, D. (1992). Establishing and maintaining healthy environments: Toward a social ecology of health promotion. American Psychologist, 47, 6-22, Taylor, N., Hutchison, E., Milliken, W., & Larson, E. (1989). Comparison of normal versus heparinized saline for flushing infusion devices. Journal of Nursing Quality Assurance, 3 (4), 49-55. Wallerstein, N. (1992). Powerlessness, empowerment, and health: Implications for health promotion programs. American Journal of Health Promotion, 6, 197-205.