- Email: [email protected]
Designing nursing interventions
Collegian (2008) 15, 19—25
available at www.sciencedirect.com
Designing nursing interventions Sanchia Aranda, PhD, MN, BAppSc (Adv Nurs), RN a,b,∗ a b
School of Nursing, The University of Melbourne, Victoria, Australia Peter MacCallum Cancer Center, Victoria, Australia
Accepted 7 November 2007
KEYWORDS Nursing; Intervention design; Research
Summary The development of nursing interventions that demonstrate the link between nursing actions and patient outcomes is a high priority for nursing research. The development of intervention research frequently focuses on the methods used to test the intervention while less attention is placed on rigor in intervention development and design. The purpose of this paper is to provide thinking points for researchers considering the development of nursing interventions. The thinking points were developed from the limited literature on this topic in synthesis with the authors own experiences of designing nursing interventions. Adoption of a systematic approach to intervention testing is advocated along with a step-wise intervention development process. This process calls for attention to problem definition, conceptual underpinnings, desired outcomes and measures and evidence-based content along with careful consideration of delivery methods, dose and attention to protecting the integrity of the intervention during testing. The approach advocated will help to ensure that nursing intervention research makes a useful contribution to the development of nursing practice. © 2007 Royal College of Nursing, Australia. Published by Elsevier Australia (a division of Reed International Books Australia Pty Ltd). All rights reserved.
Background The role of nurses in the delivery of health services is becoming increasingly complex as the burden of chronic disease in our aging community begins to outstrip available resources. The real constraints on health funding mean that nurses and other health professionals need to demonstrate the benefits of the services they provide in direct relationship to patient outcomes. However, this relationship is complex and mediated by factors associated with the
∗
Correspondence address: School of Nursing, The University of Melbourne, Victoria, Australia. E-mail address: [email protected].
environment, the patient population and the model of care delivery. Attempts to develop best evidence approaches to common patient problems expose significant deficits in current nursing intervention research largely as a result of a non-systematic approach to the development of nursing interventions. The purpose of this paper is to explore the characteristics of well-developed nursing interventions in order to guide researchers seeking to contribute to this area of nursing research and practice development. The focus of the paper is on intervention design specifically but a development approach utilizing methods testing in pharmacological research is advocated. Establishing how nursing interventions impact on patient outcomes is an identified research priority in nursing (Hinshaw, 2000), yet limited attention has focused on the design of interventions (Conn, Rantz, Wipke-Tevis, & Maas,
1322-7696/$ — see front matter © 2007 Royal College of Nursing, Australia. Published by Elsevier Australia (a division of Reed International Books Australia Pty Ltd). All rights reserved.
doi:10.1016/j.colegn.2007.11.002
20 2001). Several authors (Conn et al., 2001; Whittemore & Grey, 2002) outline systematic approaches to nursing intervention development, including use of those advocated in the testing of pharmacological interventions. Conn et al. (2001) also outline key features of design that help researchers and clinicians to develop effective interventions. Still other authors have characterized how interventions might reflect nursing values such as being patient-centered (Lauver et al., 2002). This paper attempts to provide a synthesis of this work together with the experiences of conducting nursing intervention research in the clinical environment of a cancer centre. The aim is to provide guidance to researchers considering intervention research.
Drawing on approaches in pharmacological research The development of pharmacological interventions follows a well-recognized path from phase 1 to phase 4 studies. This path helps to determine the safety, efficacy and effectiveness of new interventions from the time the drug is first used in humans. Each phase in pharmacological research can be applied to nursing intervention testing in helpful ways. Nurse researchers are encouraged to adopt this approach to developing and testing interventions in order to increase the credibility and strength of the findings. This approach cautions against leaping quickly into randomized controlled studies and advocates a planned program of work that may move several times between phases 1 and 2 until the intervention is ready for formal evaluation. Phase 1 pharmacological studies seek to determine the maximum dose of a drug including establishment of side effect and toxicity profiles. In other words they assess patient tolerance and safety of the intervention. In nursing intervention research a similar concept would be to establish the acceptability of the intervention to patients, the level of burden involvement in the intervention has and the likely attrition rates. In nursing such studies would involve small numbers of patients and data collection would focus on qualitative experiences of the intervention to assist with refinement of the approach. Phase 2 pharmacological research seeks to determine the response rate, side effect profile and feasibility of use, e.g. does toxicity outweigh benefit. The nursing research equivalent would be to determine the amount of benefit the patient gains from the intervention and to assess the degree to which the patient believes the effort of involvement is worth the benefits gained. Phase 2 studies would also seek to determine environmental factors that might impact on delivery of the intervention and other issues of feasibility such as patient recruitment. Phase 2 intervention experiences also assist in refinement of the intervention and protocols for delivery and determines the ability to measure intervention effects, providing data helpful for sample size calculations. For example, we have recently conducted a phase 2 study to develop a cognitive-behavioural intervention for women with bothersome hot flushes following breast cancer treatment (Sheeran, Ftanou, Tremblay, & Aranda, 2007). The intervention was identified by women as highly acceptable and appropriate but the study identified signif-
S. Aranda icant feasibility issues. The intervention was group based and recruitment of sufficient numbers of women to groups was problematic. Not only did this flag a potential problem in recruitment for the planned phase 3 trial but also raised important dissemination challenges if the intervention proved to be effective. Phase 3 pharmacological studies involve prospective randomized trials to assess the efficacy of the intervention in controlled situations. The conditions of the intervention delivery are tightly managed as are the characteristics of the patients recruited to the study. Features of these studies include randomization and other elements that seek to remove the potential for bias in study outcomes. While nursing interventions are much more complex and inherently difficult to control, this approach remains the gold standard of intervention testing and should be used where possible. The purpose of phase 3 studies is to test if the intervention is effective against the outcome to which it is targeted. Conn et al. (2001) argue that initial phase 3 studies must test a clinically effective intervention and it is in subsequent studies that efforts might be made to achieve the same effect with a less intense intervention or using different delivery methods. Phase 4 pharmacological studies assess the effectiveness of interventions in the real world. These studies seek to determine if the efficacy of the intervention is sustained in uncontrolled use. This is a major area of neglect in nursing research and one where attention is required. The factors that may contribute to reduce effectiveness of nursing interventions in real life include differences in nursing skill in delivery of the intervention, loss of integrity of the intervention as nurses adapt it to local environments and differences in effect due to differences in the model of care delivery. Phase 4 studies might also consider factors such as cost effectiveness. For example, we are currently testing a patient education intervention for patients commencing chemotherapy (Aranda, Schofield, Jefford, & Yates, 2007). The intervention has been found to be slightly longer than the usual care condition, implying additional cost. While the phase 3 study will determine if this additional cost is balanced by patient benefit it can only be determined if the intervention is cost effective if the phase 4 study shows acceptable costs and a retention of the benefits.
What is a nursing intervention? According to McCloskey and Bulechek (1996) ‘‘A nursing intervention is any direct care treatment that a nurse performs on behalf of a client. These treatments include nurse-initiated treatments resulting from nursing diagnoses, physician-initiated treatments resulting from medical diagnoses, and performance of the daily essential functions for the client {who} cannot do these’’ (p. 21). Interventions can be simply or complex. A simple intervention might consist of a single action by the nurse in response to common patient problem. For example where the patient is getting up for the first time after surgery and the nurse provides medication to ensure adequate pain control to enable the patient to mobilize. A complex intervention would normally be in response to a complex problem and might be delivered over more than one interaction, such
Designing nursing interventions as teaching a patient self-management of a central venous device.
Intervention design The following elements of intervention design are useful starting points when beginning the design process. Not all will be applicable to all intervention studies but offer a reflective framework to guide the design process.
Defining the patient problem The design of a nursing intervention clearly depends on the nature of the patient problem. Thus the important starting point is a thorough analysis of the patient problem of interest. While a literature review is a major component of defining the patient problem it is also useful to explore the nature of the problem in the local setting where the research is to be carried out. In particular, this combined process will confirm the applicability of international descriptive and intervention studies to the local context. Simple strategies can be used to achieve this. For example, Hudson, Aranda, and McMurray (2002) conducted focus groups with family members to compare their experiences of caring for a dying person at home to those identified in the literature and used this combination to design a carer intervention applicable in the local context. Other strategies can include involving clinicians and consumers in describing the problem to ensure that areas of significance to those most affected by the intervention are included. Defining the problem will also assist in clarifying the specific outcomes or patient problems that the intervention seeks to target.
Determining a conceptual framework Careful consideration of the patient problem and the features that determine its prevalence in the patient population provides understanding of the dynamics of the problem in the local context. This will in turn assist with considering a relevant conceptual framework to guide the intervention elements. For example, Yates et al. (2004) used Green’s PROCEED—PRECEDE (Green & Kreuter, 1991) model of health behaviour to guide an intervention to address attitudes and barriers to effective pain management. This model suggests that a health behaviour, such as taking opioid analgesics, will be influence by three constructs—–factors that predispose the person towards the behaviour (e.g. attitudes and beliefs about the problem), factors that enable the person to carry out the behaviour (e.g. health related skills) and reinforcing factors that promote use of the behaviour in real life (e.g. a key family member). Conn et al. (2001) also argue that content of the intervention should also be guided by the theoretical framework and defines this as construct validity, suggesting that the intervention should target key elements of the conceptual framework. Dodd and Miaskowski (2000) and colleagues have worked for many years on a symptom management intervention research program. This work was initially grounded in Orem’s self-care theory and over time has been adapted and shaped into a conceptual framework called PRO-SELF. This framework has
21 now been utilized in several intervention studies in areas such as pain (Miaskowski et al., 2004) and mucositis (Larson et al., 1998) and allows a strong conceptual link between the target of improving self-care behaviours (outcome) and the intervention elements (coaching, academic detailing and key message content).
The desired outcome Significant work is being undertaken internationally to define those outcomes that are most sensitive to nursing interventions. For example, the Oncology Nursing Society in the United States has identified a set of nurse sensitive outcomes in cancer nursing utilizing a nursing framework developed by Jennings, Staggers, and Brosch (1999). The ONS define five categories of nurse sensitive outcome: 1. Symptom experience—–e.g. pain and nausea. 2. Function status—–e.g. activity tolerance and nutritional status. 3. Safety—–e.g. infections and falls. 4. Psychological distress—–e.g. anxiety and depression. 5. Economic—–e.g. length of stay and readmission rates. A nursing intervention might cover one or multiple patient outcomes. For example, the pre-chemotherapy education we are currently conducting addresses outcomes related to the first four ONS outcomes using key informational messages and advice and coaching about prevention, early detection and self-care. • Reduction of distressing symptoms associated with chemotherapy. • Retention of functional status during treatment. • Prevention and early detection of life threatening sepsis. • Reduction of anxiety, particularly pre-treatment. Defining the outcomes of interest is critical to ensuring you can adequately measure the impact of the intervention. While satisfaction is often used as a measure of the patient’s response to an intervention, measures of satisfaction are seriously flawed (Avis, Bond, & Arthur, 1995) and may not indicate an actual impact on a patient problem. Crow et al.’s (2002) systematic review of satisfaction measurement found consistent evidence that the patient—practitioner relationship was the critical factor affecting satisfaction. This is important in nursing intervention research, as we need to be sure that the effects measured are not simply a product of the intervention nurse’s personality but are a direct result of the intervention.
Defining measures Each chosen outcome will need a corresponding means of measurement. Selected instruments should be psychometrically valid and ideally previously demonstrated as suitable for use in the chosen population. Evidence of appropriateness for use includes the capacity of the instrument to distinguish between groups of participants and preferably evidence that the instrument will be sensitive to changes in the target outcome as a result of the intervention.
22
Defining the target population For initial testing of the intervention the target population should be one in which the identified problem is a common or serious concern and where characteristics of interest are featured. Making the population too homogenous can result in difficulties in recruitment while being too inclusive can reduce the specificity of the intervention and reduce the likelihood of seeing an intervention effect. For example, Miaskowski et al. (2004) and Yates et al. (2004) conducted similar studies of educational interventions in people with metastatic cancer. In the first study the intervention affected pain outcomes and in the second it did not. There were two key differences between the studies, dose (discussed below) and population characteristics. Miaskowski et al.’s population was limited to patients with bone metastases, a group known to have significant pain and therefore an increased likelihood of intervention effects. Yates et al. (2004) included patients with a range of pain origins, potentially lowering the specificity of the intervention to the target group. A further useful strategy is to use phase 1 and 2 studies to test recruitment capacity to ensure the required sample size can be achieved and to determine if the inclusion and exclusion criteria are appropriate. It is also recommended that a log of screening and recruitment data is kept as recommended by CONSORT (Rennie, 1996) not only to report on recruitment but also to facilitate analysis of the types of patients being excluded where amendments to recruitment strategies might increase the target population.
Determining intervention content The content of the intervention seeks to address the relationship between the patient problem and the desired changes in the related outcomes. For example, in Yates et al. (2004) pain intervention the problem identified related to the various patient barriers known to impact on successful pain management. Thus the intervention was designed as a patient education program specifically addressing the known patient barriers to pain management such as fear of addiction and concerns about side effects. The measured outcome was patient rated barriers to effective pain management rather than pain because there is only an indirect link between barriers and actual pain experiences. In determining the educational content for the intervention Yates et al. (2004) developed evidence-based messages for delivery to patients during the intervention. Thus the content of the intervention related to key messages about effective pain management that are known to relate to barriers expressed by patients. In contrast, Larson et al. (1998) designed an intervention to prevent mucositis in patients receiving cancer chemotherapy. This intervention focused on good oral hygiene as a strategy potentially linked to reduced incidence and severity of oral mucositis in that population. The resulting intervention was a patient education intervention that focused on self-care behaviours rather than information provision, although information and patient education where clearly components of the approach. This intervention actively coached patients to
S. Aranda use the self-care strategies of brushing, flossing and rinsing throughout their treatment. Where possible the content of an intervention should itself be evidence-based. When designing the prechemotherapy education intervention our group undertook a systematic review of self-care strategies effective in managing chemotherapy side effects and developed these as summaries of the evidence in a patient format (ref to submitted paper). This intervention also drew on the clinical practice guidelines for the psychosocial care of adults with cancer (National Breast Cancer Centre & National Cancer Control Initiative [NBCC & NCCI], 2003), in particular those aspects focused on general interaction skills and the preparation of patients for potentially threatening medical procedures. These guidelines were central to both the content of the intervention and the specific ways in which the intervention nurses were educated.
Degree of person-centeredness Lauver et al. (2002) link the nursing value of patientcentered care to how interventions are constructed and delivered. Drawing on Mead and Bower (2000), Lauver et al. (2002) define patient-centered care as including a biopsychosocial perspective, seeing the patient as a person, involving the development of a therapeutic alliance and featuring shared power and responsibility between the nurse and patient. In applying these concepts to nursing interventions Lauver et al. (2002) usefully explore differences in interventions depending on the number of patient characteristics that can be customized to the individual patient. Customization can focus on content and/or the mode of delivery. Four terms are used in the literature, often interchangeably, that relate to person-centeredness—–personalized, targeted, tailored and individualized. Personalized interventions feature only minimal customization to the individual and are the least patientcentered of the four. An intervention that might fit this definition would be comparing take up of a Living with Cancer program between those receiving a letter of invitation by name from their treating doctor versus a generic invitation handed out in the clinic. Targeted interventions are customized to a group sharing a common characteristic (e.g. being female) rather than to an individual. A targeted intervention might also focus on those who experience the problem of focus at similar levels. For example, Kristjanson, Cousins, Smith, and Lewin (2005) designed a targeted bereavement intervention that scaled up the dose of intervention depending on the level of risk identified at entry to the study. Tailored interventions often feature greater complexity because the intervention nurse is required to make decisions about certain intervention elements depending on specific characteristics of the individual patient. However, in most situations the elements are predetermined in advance according to a defined set of likely areas requiring individualization. In Yates et al. (2004) the intervention was tailored according to the specific barriers to effective pain management patients reported in baseline questionnaires. Intervention nurses were provided with this information
Designing nursing interventions ahead of the intervention consultation and then customized their education to suit. At the most patient-centered end of the continuum are individualized interventions. These may be so customized that no two patients receive the same content (Lauver et al., 2002). The intervention is designed to fit the individual. At one end the interventionist will use clearly delineated elements and procedures to guide customization and include manuals to ensure critical elements are covered. At the other extreme the intervention will be entirely dependent on the interventionist. Defining the degree of customization is critical to ensure the specificity of the intervention and the more tightly defined the more likely it is that intervention effects can be attributed to the intervention. Studies suggest that some degree of customization to the individual enhances effectiveness of the intervention if well described with structured procedures (Conn et al., 2001).
Simple versus complex interventions A simple intervention is one focused on a single outcome and utilizing a single intervention strategy, e.g. provision of a medical referral to a support group to increase take up of the group. Any change in take up of the support group by the target population is highly likely to be caused by the letter when all other changes to the usual referral processes remain controlled and unchanged. A complex intervention may target one or more patient outcomes but the intervention will contain more than one intervention element. The PRO-SELF intervention approach (Dodd & Miaskowski, 2000) is an example of a complex intervention as it features information provision, patient skill building and nurse coaching. Conn et al. (2001) suggest that many patient problems of interest to nurses require complex interventions. The key limitation of complex interventions is the difficulty in determining the contribution of any individual element to the specific patient outcomes. The strength of such interventions is that they more closely capture the reality of patient experiences and the contexts in which care is provided. Researchers need to balance these competing issues and aim to keep interventions as simple as possible while still addressing the complexity of patient problems.
Delivery methods Most nursing interventions are delivered in face-to-face mode to individual patients. There is very little existing research to determine if different modes of delivery produce different outcomes. Programs of research would usefully focus on first assessing the impact of a face-to-face intervention and then assessing if similar effects can be gained by changing the delivery mode. It should be clear whether the study is testing the mode of delivery, the content of the intervention or both (Conn et al., 2001). Mode changes fall into two main categories: • Use of technology—–developments in information technology provide very useful opportunities for interventions to be delivered by telephone, DVDs, websites, interactive
23 CD-ROMS, Podcasts and more. As the population take up of technology is high the barriers against such interventions are rapidly reducing. • Alterations in patient numbers—–group interventions may be a useful strategy to deliver nursing interventions, particularly when nursing resources are short and patients are easily recruited. A group easy to recruit might be patients about to start a particular treatment (e.g. those scheduled to have a knee replacement) while a difficult to recruit group would be those with infrequent contact with the health service (e.g. survivors of cancer who have returned to work). Delivery methods should be carefully assessed during pilot work where factors such as feasibility of group recruitment can be tested and feedback from patients about the experience of the delivery can be assessed. Mode of delivery can also relate to the dose of intervention, with different methods being of potentially different strength despite similar content.
Dose The ability to achieve a clinically significant intervention effect requires an appropriate balance between the amount of intervention, the severity of the problem and the burden on patients. If the dose is too low a potentially useful intervention will be discarded when potentially effective (Conn et al., 2001). Dose concepts include the amount, frequency, duration and intensity of the intervention (Conn et al., 2001). Amount is the actual time it takes for the delivery of the intervention. For example, in Yates et al. (2004) pain intervention the treatment group received a single face-toface session of up to 30 min and two telephone follow-up calls of up to 10 min each. Thus the total intervention time was less than 1 h. Frequency and duration refer to the number of intervention sessions and the time span over which they were delivered. In the Yates et al. (2004) example, the intervention was delivered in three sessions each a week apart (frequency) in duration of 3 weeks. Some of the effects in this study were sustained out to the 3-month follow up while others were not. Thus a further question to be addressed is whether changes in the amount, frequency and duration of the intervention would lead to more sustained intervention effects. The intensity of the intervention refers to a combination of these factors and is best understood as the amount of burden there is for the patient. More intense interventions usually require more time and more input (e.g. homework) by the patient. The goal with increasing the intensity is to increase the desired effect on outcomes but may also lead to increased attrition if patients judge the burden to be too high (Conn et al., 2001). Important questions related to nursing interventions are those that seek to increase the dose of an intervention in ways that do not require more nursing time in the context of a workforce shortage. This is where the relationship between dose and method of delivery is important. In the previously mentioned chemotherapy education study the
24
S. Aranda
intervention consists of a single face-to-face session with an intervention nurse with a single follow up phone call 2 days later. The dose of intervention is enhanced in two key ways. Firstly, patients are provided with an educational DVD (Schofield et al., in press) prior to their education appointment along with a question prompt sheet to enhance the patient directedness of the education session. Secondly, print materials including patient drug sheets and self-care leaflets for a range of commonly experienced problems are given to patients during the session to reinforce content and to encourage self-care.
to improved patient outcomes. Much of this work occurs in complex health environments and there are significant challenges in such research. While some challenges are managed by rigorous study design, it is equally important to pay sufficient attention to intervention design. Intervention research is expensive, time consuming and demanding of both patients and nurses and thus it is an ethical responsibility to ensure maximum benefits are achieved. The thinking points outlined in this paper will provide enhanced capacity for nurses to demonstrate how they can make a difference.
Managing the integrity
References
Intervention integrity refers to the extent to which the intervention is delivered as planned. It is important to ensure that outcomes of a study can be formally linked to the intervention and to ensure that an absence of intervention effect is not caused by a failure of delivery. Maintaining the integrity of an intervention across the life of an intervention study is a serious challenge. This is particularly critical where more than one nurse is involved in delivering a complex intervention. Strategies to enhance integrity need to be built into intervention design and include the following:
Aranda, S., Schofield, P., Jefford, M., & Yates, P. (2007) A randomized controlled trial of an evidence-based chemotherapy education intervention. Current NHMRC Study at Peter MacCallum Cancer Centre. Avis, M., Bond, M., & Arthur, A. (1995). Exploring patient satisfaction with out-patient services. Journal of Nursing Management, 3(2), 59—65. Conn, V. S., Rantz, M. J., Wipke-Tevis, D. D., & Maas, M. L. (2001). Designing effective nursing interventions. Research in Nursing & Health, 24, 433—442. Crow, R., Gage, H., Hampson, S., Hart, J., Kimber, A., Storey, L., et al. (2002). The measurement of satisfaction with healthcare: Implications for practice from a systematic review of the literature [Monograph]. Health Technology Assessment, 6, 32. Dodd, M., & Miaskowski, C. (2000). The PRO-SELF program: A self-care intervention program for patients receiving cancer treatment. Seminars in Oncology Nursing, 16(4), 300—308. Green, L., & Kreuter, M. (1991). Health promotion planning (2nd ed.). Mountain View: Mayfield Publishing Company. Hinshaw, A. S. (2000). Nursing knowledge for the 21st century: Opportunities and challenges. Journal of Nursing Scholarship, 32, 117—123. Hudson, P., Aranda, S., & McMurray, N. (2002). Intervention development for enhanced lay palliative caregiver support: The use of focus groups. European Journal of Cancer Care, 11(4), 262—270. Jennings, B. M., Staggers, N., & Brosch, L. R. (1999). A classification scheme for outcome indicators. Journal of Nursing Scholarship, 31(4), 381—388. Kristjanson, L., Cousins, K., Smith, J., & Lewin, J. (2005). Evaluation of the bereavement risk index (BRI): A community hospice care protocol. International Journal of Palliative Nursing, 11(12), 610—618. Larson, P., Miaskowski, C., MacPhail, L., Dodd, M. J., Greenspan, D., Dibble, S. L., et al. (1998). The PRO-SELF(C) mouth aware program: An effective approach for reducing chemotherapyinduced mucositis. Cancer Nursing, 21(4), 263—268. Lauver, D., Ward, S. E., Heidrich, S. M., Keller, M. L., Bowers, B. J., Brennan, P. F., et al. (2002). Patient-centered interventions. Research in Nursing & Health, 25, 246—255. McCloskey, J. C., & Bulechek, G. M. (1996). IOWA nursing interventions classification (2nd ed.). St Louis, MO: Mosby-Year Book. Mead, N., & Bower, P. (2000). Patient-centeredness: A conceptual framework and review of the empirical literature. Social Science & Medicine, 51, 1087—1110. Miaskowski, C., Dodd, M., West, C., Schumacher, K., Paul, S., Tripathy, D., et al. (2004). Randomized clinical trial of the effectiveness of a self-care intervention to improve cancer pain management. Journal of Clinical Oncology, 9(1), 1713—1720. National Breast Cancer Centre & National Cancer Control Initiative. (2003). Clinical practice guidelines for the psychosocial care of adults with cancer. Camperdown, NSW: National Breast Cancer Centre.
• Structured training of intervention nurses helps to ensure a common knowledge base and the attainment of a core set of skills required for the intervention delivery. For example in our chemotherapy education study the intervention nurses received specific training in evidence-based communication skills and information giving, education on chemotherapy side effects and self-care management strategies and an orientation to the methodology of the study and the threats to intervention integrity that they were required to control. The training included role plays with simulated patients to create real-life practice in the intervention. • Evidence-based support materials to guide intervention nurses. In some cases these might be also developed as patient information brochures for use in the intervention. • An intervention manual that structures the responses given to each factor likely to be encountered. Such manuals can also provide quick access to the evidence behind particular responses as a memory aide. • Separation of intervention nurses from involvement in usual care conditions to prevent diffusion of the intervention into usual care. This is applicable where clinical nurses undertake the intervention but are part of the unit staff where the intervention is being delivered. • Pilot work can also assist in maintaining intervention integrity by allowing a closer scrutiny of what the intervention nurses are doing and allowing feedback and correction of actions that are outside of the intervention protocol. It can also be helpful to audiotape intervention sessions to allow an independent assessment of adherence to the intervention protocol and subsequent correction.
Conclusion Intervention studies in nursing are a critical part of developing an evidence base to support nurses’ contribution
Designing nursing interventions Rennie, D. (1996). How to report randomized controlled trials: The CONSORT statement. Journal of the American Medical Association, 276, 649. Schofield, P., Jefford, M., Carey, M., Baravelli, C., Evans, M., & Aranda, S. (in press). Preparing patients for threatening medical treatments: Effects of a chemotherapy educational DVD on anxiety, unmet needs and self-efficacy. Supportive Care in Cancer. Sheeran, L., Ftanou, M., Tremblay, A., & Aranda, S. (2007). Pilot study of a cognitive behavioural group intervention for women
25 with bothersome hot flushes. Unpublished research report, Peter MacCallum Cancer Centre. Whittemore, R., & Grey, M. (2002). The systematic development of nursing interventions. Journal of Nursing Scholarship, 34(2), 115—120. Yates, P., Edwards, H., Nash, R., Aranda, S., Purdie, D., Najman, J., et al. (2004). Randomized controlled trial of a nurse administered educational intervention for improving cancer pain management in ambulatory settings. Patient Education and Counseling, 53(2), 227—237.
available at www.sciencedirect.com
Designing nursing interventions Sanchia Aranda, PhD, MN, BAppSc (Adv Nurs), RN a,b,∗ a b
School of Nursing, The University of Melbourne, Victoria, Australia Peter MacCallum Cancer Center, Victoria, Australia
Accepted 7 November 2007
KEYWORDS Nursing; Intervention design; Research
Summary The development of nursing interventions that demonstrate the link between nursing actions and patient outcomes is a high priority for nursing research. The development of intervention research frequently focuses on the methods used to test the intervention while less attention is placed on rigor in intervention development and design. The purpose of this paper is to provide thinking points for researchers considering the development of nursing interventions. The thinking points were developed from the limited literature on this topic in synthesis with the authors own experiences of designing nursing interventions. Adoption of a systematic approach to intervention testing is advocated along with a step-wise intervention development process. This process calls for attention to problem definition, conceptual underpinnings, desired outcomes and measures and evidence-based content along with careful consideration of delivery methods, dose and attention to protecting the integrity of the intervention during testing. The approach advocated will help to ensure that nursing intervention research makes a useful contribution to the development of nursing practice. © 2007 Royal College of Nursing, Australia. Published by Elsevier Australia (a division of Reed International Books Australia Pty Ltd). All rights reserved.
Background The role of nurses in the delivery of health services is becoming increasingly complex as the burden of chronic disease in our aging community begins to outstrip available resources. The real constraints on health funding mean that nurses and other health professionals need to demonstrate the benefits of the services they provide in direct relationship to patient outcomes. However, this relationship is complex and mediated by factors associated with the
∗
Correspondence address: School of Nursing, The University of Melbourne, Victoria, Australia. E-mail address: [email protected].
environment, the patient population and the model of care delivery. Attempts to develop best evidence approaches to common patient problems expose significant deficits in current nursing intervention research largely as a result of a non-systematic approach to the development of nursing interventions. The purpose of this paper is to explore the characteristics of well-developed nursing interventions in order to guide researchers seeking to contribute to this area of nursing research and practice development. The focus of the paper is on intervention design specifically but a development approach utilizing methods testing in pharmacological research is advocated. Establishing how nursing interventions impact on patient outcomes is an identified research priority in nursing (Hinshaw, 2000), yet limited attention has focused on the design of interventions (Conn, Rantz, Wipke-Tevis, & Maas,
1322-7696/$ — see front matter © 2007 Royal College of Nursing, Australia. Published by Elsevier Australia (a division of Reed International Books Australia Pty Ltd). All rights reserved.
doi:10.1016/j.colegn.2007.11.002
20 2001). Several authors (Conn et al., 2001; Whittemore & Grey, 2002) outline systematic approaches to nursing intervention development, including use of those advocated in the testing of pharmacological interventions. Conn et al. (2001) also outline key features of design that help researchers and clinicians to develop effective interventions. Still other authors have characterized how interventions might reflect nursing values such as being patient-centered (Lauver et al., 2002). This paper attempts to provide a synthesis of this work together with the experiences of conducting nursing intervention research in the clinical environment of a cancer centre. The aim is to provide guidance to researchers considering intervention research.
Drawing on approaches in pharmacological research The development of pharmacological interventions follows a well-recognized path from phase 1 to phase 4 studies. This path helps to determine the safety, efficacy and effectiveness of new interventions from the time the drug is first used in humans. Each phase in pharmacological research can be applied to nursing intervention testing in helpful ways. Nurse researchers are encouraged to adopt this approach to developing and testing interventions in order to increase the credibility and strength of the findings. This approach cautions against leaping quickly into randomized controlled studies and advocates a planned program of work that may move several times between phases 1 and 2 until the intervention is ready for formal evaluation. Phase 1 pharmacological studies seek to determine the maximum dose of a drug including establishment of side effect and toxicity profiles. In other words they assess patient tolerance and safety of the intervention. In nursing intervention research a similar concept would be to establish the acceptability of the intervention to patients, the level of burden involvement in the intervention has and the likely attrition rates. In nursing such studies would involve small numbers of patients and data collection would focus on qualitative experiences of the intervention to assist with refinement of the approach. Phase 2 pharmacological research seeks to determine the response rate, side effect profile and feasibility of use, e.g. does toxicity outweigh benefit. The nursing research equivalent would be to determine the amount of benefit the patient gains from the intervention and to assess the degree to which the patient believes the effort of involvement is worth the benefits gained. Phase 2 studies would also seek to determine environmental factors that might impact on delivery of the intervention and other issues of feasibility such as patient recruitment. Phase 2 intervention experiences also assist in refinement of the intervention and protocols for delivery and determines the ability to measure intervention effects, providing data helpful for sample size calculations. For example, we have recently conducted a phase 2 study to develop a cognitive-behavioural intervention for women with bothersome hot flushes following breast cancer treatment (Sheeran, Ftanou, Tremblay, & Aranda, 2007). The intervention was identified by women as highly acceptable and appropriate but the study identified signif-
S. Aranda icant feasibility issues. The intervention was group based and recruitment of sufficient numbers of women to groups was problematic. Not only did this flag a potential problem in recruitment for the planned phase 3 trial but also raised important dissemination challenges if the intervention proved to be effective. Phase 3 pharmacological studies involve prospective randomized trials to assess the efficacy of the intervention in controlled situations. The conditions of the intervention delivery are tightly managed as are the characteristics of the patients recruited to the study. Features of these studies include randomization and other elements that seek to remove the potential for bias in study outcomes. While nursing interventions are much more complex and inherently difficult to control, this approach remains the gold standard of intervention testing and should be used where possible. The purpose of phase 3 studies is to test if the intervention is effective against the outcome to which it is targeted. Conn et al. (2001) argue that initial phase 3 studies must test a clinically effective intervention and it is in subsequent studies that efforts might be made to achieve the same effect with a less intense intervention or using different delivery methods. Phase 4 pharmacological studies assess the effectiveness of interventions in the real world. These studies seek to determine if the efficacy of the intervention is sustained in uncontrolled use. This is a major area of neglect in nursing research and one where attention is required. The factors that may contribute to reduce effectiveness of nursing interventions in real life include differences in nursing skill in delivery of the intervention, loss of integrity of the intervention as nurses adapt it to local environments and differences in effect due to differences in the model of care delivery. Phase 4 studies might also consider factors such as cost effectiveness. For example, we are currently testing a patient education intervention for patients commencing chemotherapy (Aranda, Schofield, Jefford, & Yates, 2007). The intervention has been found to be slightly longer than the usual care condition, implying additional cost. While the phase 3 study will determine if this additional cost is balanced by patient benefit it can only be determined if the intervention is cost effective if the phase 4 study shows acceptable costs and a retention of the benefits.
What is a nursing intervention? According to McCloskey and Bulechek (1996) ‘‘A nursing intervention is any direct care treatment that a nurse performs on behalf of a client. These treatments include nurse-initiated treatments resulting from nursing diagnoses, physician-initiated treatments resulting from medical diagnoses, and performance of the daily essential functions for the client {who} cannot do these’’ (p. 21). Interventions can be simply or complex. A simple intervention might consist of a single action by the nurse in response to common patient problem. For example where the patient is getting up for the first time after surgery and the nurse provides medication to ensure adequate pain control to enable the patient to mobilize. A complex intervention would normally be in response to a complex problem and might be delivered over more than one interaction, such
Designing nursing interventions as teaching a patient self-management of a central venous device.
Intervention design The following elements of intervention design are useful starting points when beginning the design process. Not all will be applicable to all intervention studies but offer a reflective framework to guide the design process.
Defining the patient problem The design of a nursing intervention clearly depends on the nature of the patient problem. Thus the important starting point is a thorough analysis of the patient problem of interest. While a literature review is a major component of defining the patient problem it is also useful to explore the nature of the problem in the local setting where the research is to be carried out. In particular, this combined process will confirm the applicability of international descriptive and intervention studies to the local context. Simple strategies can be used to achieve this. For example, Hudson, Aranda, and McMurray (2002) conducted focus groups with family members to compare their experiences of caring for a dying person at home to those identified in the literature and used this combination to design a carer intervention applicable in the local context. Other strategies can include involving clinicians and consumers in describing the problem to ensure that areas of significance to those most affected by the intervention are included. Defining the problem will also assist in clarifying the specific outcomes or patient problems that the intervention seeks to target.
Determining a conceptual framework Careful consideration of the patient problem and the features that determine its prevalence in the patient population provides understanding of the dynamics of the problem in the local context. This will in turn assist with considering a relevant conceptual framework to guide the intervention elements. For example, Yates et al. (2004) used Green’s PROCEED—PRECEDE (Green & Kreuter, 1991) model of health behaviour to guide an intervention to address attitudes and barriers to effective pain management. This model suggests that a health behaviour, such as taking opioid analgesics, will be influence by three constructs—–factors that predispose the person towards the behaviour (e.g. attitudes and beliefs about the problem), factors that enable the person to carry out the behaviour (e.g. health related skills) and reinforcing factors that promote use of the behaviour in real life (e.g. a key family member). Conn et al. (2001) also argue that content of the intervention should also be guided by the theoretical framework and defines this as construct validity, suggesting that the intervention should target key elements of the conceptual framework. Dodd and Miaskowski (2000) and colleagues have worked for many years on a symptom management intervention research program. This work was initially grounded in Orem’s self-care theory and over time has been adapted and shaped into a conceptual framework called PRO-SELF. This framework has
21 now been utilized in several intervention studies in areas such as pain (Miaskowski et al., 2004) and mucositis (Larson et al., 1998) and allows a strong conceptual link between the target of improving self-care behaviours (outcome) and the intervention elements (coaching, academic detailing and key message content).
The desired outcome Significant work is being undertaken internationally to define those outcomes that are most sensitive to nursing interventions. For example, the Oncology Nursing Society in the United States has identified a set of nurse sensitive outcomes in cancer nursing utilizing a nursing framework developed by Jennings, Staggers, and Brosch (1999). The ONS define five categories of nurse sensitive outcome: 1. Symptom experience—–e.g. pain and nausea. 2. Function status—–e.g. activity tolerance and nutritional status. 3. Safety—–e.g. infections and falls. 4. Psychological distress—–e.g. anxiety and depression. 5. Economic—–e.g. length of stay and readmission rates. A nursing intervention might cover one or multiple patient outcomes. For example, the pre-chemotherapy education we are currently conducting addresses outcomes related to the first four ONS outcomes using key informational messages and advice and coaching about prevention, early detection and self-care. • Reduction of distressing symptoms associated with chemotherapy. • Retention of functional status during treatment. • Prevention and early detection of life threatening sepsis. • Reduction of anxiety, particularly pre-treatment. Defining the outcomes of interest is critical to ensuring you can adequately measure the impact of the intervention. While satisfaction is often used as a measure of the patient’s response to an intervention, measures of satisfaction are seriously flawed (Avis, Bond, & Arthur, 1995) and may not indicate an actual impact on a patient problem. Crow et al.’s (2002) systematic review of satisfaction measurement found consistent evidence that the patient—practitioner relationship was the critical factor affecting satisfaction. This is important in nursing intervention research, as we need to be sure that the effects measured are not simply a product of the intervention nurse’s personality but are a direct result of the intervention.
Defining measures Each chosen outcome will need a corresponding means of measurement. Selected instruments should be psychometrically valid and ideally previously demonstrated as suitable for use in the chosen population. Evidence of appropriateness for use includes the capacity of the instrument to distinguish between groups of participants and preferably evidence that the instrument will be sensitive to changes in the target outcome as a result of the intervention.
22
Defining the target population For initial testing of the intervention the target population should be one in which the identified problem is a common or serious concern and where characteristics of interest are featured. Making the population too homogenous can result in difficulties in recruitment while being too inclusive can reduce the specificity of the intervention and reduce the likelihood of seeing an intervention effect. For example, Miaskowski et al. (2004) and Yates et al. (2004) conducted similar studies of educational interventions in people with metastatic cancer. In the first study the intervention affected pain outcomes and in the second it did not. There were two key differences between the studies, dose (discussed below) and population characteristics. Miaskowski et al.’s population was limited to patients with bone metastases, a group known to have significant pain and therefore an increased likelihood of intervention effects. Yates et al. (2004) included patients with a range of pain origins, potentially lowering the specificity of the intervention to the target group. A further useful strategy is to use phase 1 and 2 studies to test recruitment capacity to ensure the required sample size can be achieved and to determine if the inclusion and exclusion criteria are appropriate. It is also recommended that a log of screening and recruitment data is kept as recommended by CONSORT (Rennie, 1996) not only to report on recruitment but also to facilitate analysis of the types of patients being excluded where amendments to recruitment strategies might increase the target population.
Determining intervention content The content of the intervention seeks to address the relationship between the patient problem and the desired changes in the related outcomes. For example, in Yates et al. (2004) pain intervention the problem identified related to the various patient barriers known to impact on successful pain management. Thus the intervention was designed as a patient education program specifically addressing the known patient barriers to pain management such as fear of addiction and concerns about side effects. The measured outcome was patient rated barriers to effective pain management rather than pain because there is only an indirect link between barriers and actual pain experiences. In determining the educational content for the intervention Yates et al. (2004) developed evidence-based messages for delivery to patients during the intervention. Thus the content of the intervention related to key messages about effective pain management that are known to relate to barriers expressed by patients. In contrast, Larson et al. (1998) designed an intervention to prevent mucositis in patients receiving cancer chemotherapy. This intervention focused on good oral hygiene as a strategy potentially linked to reduced incidence and severity of oral mucositis in that population. The resulting intervention was a patient education intervention that focused on self-care behaviours rather than information provision, although information and patient education where clearly components of the approach. This intervention actively coached patients to
S. Aranda use the self-care strategies of brushing, flossing and rinsing throughout their treatment. Where possible the content of an intervention should itself be evidence-based. When designing the prechemotherapy education intervention our group undertook a systematic review of self-care strategies effective in managing chemotherapy side effects and developed these as summaries of the evidence in a patient format (ref to submitted paper). This intervention also drew on the clinical practice guidelines for the psychosocial care of adults with cancer (National Breast Cancer Centre & National Cancer Control Initiative [NBCC & NCCI], 2003), in particular those aspects focused on general interaction skills and the preparation of patients for potentially threatening medical procedures. These guidelines were central to both the content of the intervention and the specific ways in which the intervention nurses were educated.
Degree of person-centeredness Lauver et al. (2002) link the nursing value of patientcentered care to how interventions are constructed and delivered. Drawing on Mead and Bower (2000), Lauver et al. (2002) define patient-centered care as including a biopsychosocial perspective, seeing the patient as a person, involving the development of a therapeutic alliance and featuring shared power and responsibility between the nurse and patient. In applying these concepts to nursing interventions Lauver et al. (2002) usefully explore differences in interventions depending on the number of patient characteristics that can be customized to the individual patient. Customization can focus on content and/or the mode of delivery. Four terms are used in the literature, often interchangeably, that relate to person-centeredness—–personalized, targeted, tailored and individualized. Personalized interventions feature only minimal customization to the individual and are the least patientcentered of the four. An intervention that might fit this definition would be comparing take up of a Living with Cancer program between those receiving a letter of invitation by name from their treating doctor versus a generic invitation handed out in the clinic. Targeted interventions are customized to a group sharing a common characteristic (e.g. being female) rather than to an individual. A targeted intervention might also focus on those who experience the problem of focus at similar levels. For example, Kristjanson, Cousins, Smith, and Lewin (2005) designed a targeted bereavement intervention that scaled up the dose of intervention depending on the level of risk identified at entry to the study. Tailored interventions often feature greater complexity because the intervention nurse is required to make decisions about certain intervention elements depending on specific characteristics of the individual patient. However, in most situations the elements are predetermined in advance according to a defined set of likely areas requiring individualization. In Yates et al. (2004) the intervention was tailored according to the specific barriers to effective pain management patients reported in baseline questionnaires. Intervention nurses were provided with this information
Designing nursing interventions ahead of the intervention consultation and then customized their education to suit. At the most patient-centered end of the continuum are individualized interventions. These may be so customized that no two patients receive the same content (Lauver et al., 2002). The intervention is designed to fit the individual. At one end the interventionist will use clearly delineated elements and procedures to guide customization and include manuals to ensure critical elements are covered. At the other extreme the intervention will be entirely dependent on the interventionist. Defining the degree of customization is critical to ensure the specificity of the intervention and the more tightly defined the more likely it is that intervention effects can be attributed to the intervention. Studies suggest that some degree of customization to the individual enhances effectiveness of the intervention if well described with structured procedures (Conn et al., 2001).
Simple versus complex interventions A simple intervention is one focused on a single outcome and utilizing a single intervention strategy, e.g. provision of a medical referral to a support group to increase take up of the group. Any change in take up of the support group by the target population is highly likely to be caused by the letter when all other changes to the usual referral processes remain controlled and unchanged. A complex intervention may target one or more patient outcomes but the intervention will contain more than one intervention element. The PRO-SELF intervention approach (Dodd & Miaskowski, 2000) is an example of a complex intervention as it features information provision, patient skill building and nurse coaching. Conn et al. (2001) suggest that many patient problems of interest to nurses require complex interventions. The key limitation of complex interventions is the difficulty in determining the contribution of any individual element to the specific patient outcomes. The strength of such interventions is that they more closely capture the reality of patient experiences and the contexts in which care is provided. Researchers need to balance these competing issues and aim to keep interventions as simple as possible while still addressing the complexity of patient problems.
Delivery methods Most nursing interventions are delivered in face-to-face mode to individual patients. There is very little existing research to determine if different modes of delivery produce different outcomes. Programs of research would usefully focus on first assessing the impact of a face-to-face intervention and then assessing if similar effects can be gained by changing the delivery mode. It should be clear whether the study is testing the mode of delivery, the content of the intervention or both (Conn et al., 2001). Mode changes fall into two main categories: • Use of technology—–developments in information technology provide very useful opportunities for interventions to be delivered by telephone, DVDs, websites, interactive
23 CD-ROMS, Podcasts and more. As the population take up of technology is high the barriers against such interventions are rapidly reducing. • Alterations in patient numbers—–group interventions may be a useful strategy to deliver nursing interventions, particularly when nursing resources are short and patients are easily recruited. A group easy to recruit might be patients about to start a particular treatment (e.g. those scheduled to have a knee replacement) while a difficult to recruit group would be those with infrequent contact with the health service (e.g. survivors of cancer who have returned to work). Delivery methods should be carefully assessed during pilot work where factors such as feasibility of group recruitment can be tested and feedback from patients about the experience of the delivery can be assessed. Mode of delivery can also relate to the dose of intervention, with different methods being of potentially different strength despite similar content.
Dose The ability to achieve a clinically significant intervention effect requires an appropriate balance between the amount of intervention, the severity of the problem and the burden on patients. If the dose is too low a potentially useful intervention will be discarded when potentially effective (Conn et al., 2001). Dose concepts include the amount, frequency, duration and intensity of the intervention (Conn et al., 2001). Amount is the actual time it takes for the delivery of the intervention. For example, in Yates et al. (2004) pain intervention the treatment group received a single face-toface session of up to 30 min and two telephone follow-up calls of up to 10 min each. Thus the total intervention time was less than 1 h. Frequency and duration refer to the number of intervention sessions and the time span over which they were delivered. In the Yates et al. (2004) example, the intervention was delivered in three sessions each a week apart (frequency) in duration of 3 weeks. Some of the effects in this study were sustained out to the 3-month follow up while others were not. Thus a further question to be addressed is whether changes in the amount, frequency and duration of the intervention would lead to more sustained intervention effects. The intensity of the intervention refers to a combination of these factors and is best understood as the amount of burden there is for the patient. More intense interventions usually require more time and more input (e.g. homework) by the patient. The goal with increasing the intensity is to increase the desired effect on outcomes but may also lead to increased attrition if patients judge the burden to be too high (Conn et al., 2001). Important questions related to nursing interventions are those that seek to increase the dose of an intervention in ways that do not require more nursing time in the context of a workforce shortage. This is where the relationship between dose and method of delivery is important. In the previously mentioned chemotherapy education study the
24
S. Aranda
intervention consists of a single face-to-face session with an intervention nurse with a single follow up phone call 2 days later. The dose of intervention is enhanced in two key ways. Firstly, patients are provided with an educational DVD (Schofield et al., in press) prior to their education appointment along with a question prompt sheet to enhance the patient directedness of the education session. Secondly, print materials including patient drug sheets and self-care leaflets for a range of commonly experienced problems are given to patients during the session to reinforce content and to encourage self-care.
to improved patient outcomes. Much of this work occurs in complex health environments and there are significant challenges in such research. While some challenges are managed by rigorous study design, it is equally important to pay sufficient attention to intervention design. Intervention research is expensive, time consuming and demanding of both patients and nurses and thus it is an ethical responsibility to ensure maximum benefits are achieved. The thinking points outlined in this paper will provide enhanced capacity for nurses to demonstrate how they can make a difference.
Managing the integrity
References
Intervention integrity refers to the extent to which the intervention is delivered as planned. It is important to ensure that outcomes of a study can be formally linked to the intervention and to ensure that an absence of intervention effect is not caused by a failure of delivery. Maintaining the integrity of an intervention across the life of an intervention study is a serious challenge. This is particularly critical where more than one nurse is involved in delivering a complex intervention. Strategies to enhance integrity need to be built into intervention design and include the following:
Aranda, S., Schofield, P., Jefford, M., & Yates, P. (2007) A randomized controlled trial of an evidence-based chemotherapy education intervention. Current NHMRC Study at Peter MacCallum Cancer Centre. Avis, M., Bond, M., & Arthur, A. (1995). Exploring patient satisfaction with out-patient services. Journal of Nursing Management, 3(2), 59—65. Conn, V. S., Rantz, M. J., Wipke-Tevis, D. D., & Maas, M. L. (2001). Designing effective nursing interventions. Research in Nursing & Health, 24, 433—442. Crow, R., Gage, H., Hampson, S., Hart, J., Kimber, A., Storey, L., et al. (2002). The measurement of satisfaction with healthcare: Implications for practice from a systematic review of the literature [Monograph]. Health Technology Assessment, 6, 32. Dodd, M., & Miaskowski, C. (2000). The PRO-SELF program: A self-care intervention program for patients receiving cancer treatment. Seminars in Oncology Nursing, 16(4), 300—308. Green, L., & Kreuter, M. (1991). Health promotion planning (2nd ed.). Mountain View: Mayfield Publishing Company. Hinshaw, A. S. (2000). Nursing knowledge for the 21st century: Opportunities and challenges. Journal of Nursing Scholarship, 32, 117—123. Hudson, P., Aranda, S., & McMurray, N. (2002). Intervention development for enhanced lay palliative caregiver support: The use of focus groups. European Journal of Cancer Care, 11(4), 262—270. Jennings, B. M., Staggers, N., & Brosch, L. R. (1999). A classification scheme for outcome indicators. Journal of Nursing Scholarship, 31(4), 381—388. Kristjanson, L., Cousins, K., Smith, J., & Lewin, J. (2005). Evaluation of the bereavement risk index (BRI): A community hospice care protocol. International Journal of Palliative Nursing, 11(12), 610—618. Larson, P., Miaskowski, C., MacPhail, L., Dodd, M. J., Greenspan, D., Dibble, S. L., et al. (1998). The PRO-SELF(C) mouth aware program: An effective approach for reducing chemotherapyinduced mucositis. Cancer Nursing, 21(4), 263—268. Lauver, D., Ward, S. E., Heidrich, S. M., Keller, M. L., Bowers, B. J., Brennan, P. F., et al. (2002). Patient-centered interventions. Research in Nursing & Health, 25, 246—255. McCloskey, J. C., & Bulechek, G. M. (1996). IOWA nursing interventions classification (2nd ed.). St Louis, MO: Mosby-Year Book. Mead, N., & Bower, P. (2000). Patient-centeredness: A conceptual framework and review of the empirical literature. Social Science & Medicine, 51, 1087—1110. Miaskowski, C., Dodd, M., West, C., Schumacher, K., Paul, S., Tripathy, D., et al. (2004). Randomized clinical trial of the effectiveness of a self-care intervention to improve cancer pain management. Journal of Clinical Oncology, 9(1), 1713—1720. National Breast Cancer Centre & National Cancer Control Initiative. (2003). Clinical practice guidelines for the psychosocial care of adults with cancer. Camperdown, NSW: National Breast Cancer Centre.
• Structured training of intervention nurses helps to ensure a common knowledge base and the attainment of a core set of skills required for the intervention delivery. For example in our chemotherapy education study the intervention nurses received specific training in evidence-based communication skills and information giving, education on chemotherapy side effects and self-care management strategies and an orientation to the methodology of the study and the threats to intervention integrity that they were required to control. The training included role plays with simulated patients to create real-life practice in the intervention. • Evidence-based support materials to guide intervention nurses. In some cases these might be also developed as patient information brochures for use in the intervention. • An intervention manual that structures the responses given to each factor likely to be encountered. Such manuals can also provide quick access to the evidence behind particular responses as a memory aide. • Separation of intervention nurses from involvement in usual care conditions to prevent diffusion of the intervention into usual care. This is applicable where clinical nurses undertake the intervention but are part of the unit staff where the intervention is being delivered. • Pilot work can also assist in maintaining intervention integrity by allowing a closer scrutiny of what the intervention nurses are doing and allowing feedback and correction of actions that are outside of the intervention protocol. It can also be helpful to audiotape intervention sessions to allow an independent assessment of adherence to the intervention protocol and subsequent correction.
Conclusion Intervention studies in nursing are a critical part of developing an evidence base to support nurses’ contribution
Designing nursing interventions Rennie, D. (1996). How to report randomized controlled trials: The CONSORT statement. Journal of the American Medical Association, 276, 649. Schofield, P., Jefford, M., Carey, M., Baravelli, C., Evans, M., & Aranda, S. (in press). Preparing patients for threatening medical treatments: Effects of a chemotherapy educational DVD on anxiety, unmet needs and self-efficacy. Supportive Care in Cancer. Sheeran, L., Ftanou, M., Tremblay, A., & Aranda, S. (2007). Pilot study of a cognitive behavioural group intervention for women
25 with bothersome hot flushes. Unpublished research report, Peter MacCallum Cancer Centre. Whittemore, R., & Grey, M. (2002). The systematic development of nursing interventions. Journal of Nursing Scholarship, 34(2), 115—120. Yates, P., Edwards, H., Nash, R., Aranda, S., Purdie, D., Najman, J., et al. (2004). Randomized controlled trial of a nurse administered educational intervention for improving cancer pain management in ambulatory settings. Patient Education and Counseling, 53(2), 227—237.