Conceptual framework

We used the Weiner Organizational Theory of Implementation Effectiveness (Figure 1) [8, 9], as a conceptual model to guide the identification of determinants for organizational readiness for change (ORC) [8, 9]. ORC refers to organizational members’ shared resolve to implement a change and their collective ability to do so; respectively, these dimensions are known as change commitment and change efficacy. Briefly, this model posits that organizational readiness for change is the product of two constructs: change valence, or the degree to which organizational members value the proposed change (e.g., perceived need for change, perceived advantage to change, and perceived fit); and informational assessment, or the degree to which organizational members know what tasks are involved in the change, have enough resources available to implement the change (e.g., people, money and materials), and view positively situational factors such as the timing of change and the time available for implementation. Both change valence and informational assessment can be predicted by contextual factors, the broader conditions that affect an organization’s readiness for change, which may serve as both a facilitator and barrier. When organizational readiness is high, the theory posits, organizational members are more likely to initiate change, exert greater effort in support of change, and exhibit greater persistence in the face of obstacle or setbacks during implementation. The likely outcome is greater consistency and quality of intervention delivery (i.e., effective implementation). This theory was used as a guide for interview guide development, survey questions, and qualitative data analysis.

Intervention overview

HTN-IMPROVE is a nurse-delivered tailored telephone intervention in which the emphasis is initiating and maintaining specific health behaviors related to hypertension. The efficacy of the intervention has previously been demonstrated through rigorous clinical trials [10–15]. The intervention is organized as telephone encounters that occur monthly for one year. At each telephone encounter, trained nurses use the intervention software to gather medical and behavioral information. This information is analyzed by the software program and activates behavioral and educational modules within the software that address hypertension related issues such as medication adherence, social support, and healthy behaviors such as limiting alcohol use, smoking cessation, weight loss, and adequate exercise [11, 13, 15].

The three VA intervention sites agreed to deliver the program to at least 500 individuals over a one-year period with an expectation that the patients would be contacted monthly for one year. Access to the computerized software and related technical assistance were provided by the Durham VA Medical Center Health Services Research & Development Center of Excellence staff study team. The intervention sites and the site principal investigator (PI) were responsible for ensuring the program was implemented (e.g., determining ways in which patient referrals occurred; working with information technology to ensure nurse interventionist needs were met). Details on implementation parameters, software operation, and patient enrollment are previously reported in the protocol [6].

Setting

With IRB review and approval from each participating site, we implemented HTN-IMPROVE at three geographically diverse intervention sites in separate Veterans Integrated Service Networks (VISNs) which re-present three geographic regions across the US. Each participating location is affiliated with an academic medical center. Primary care is provided by attending physicians with academic appointments, non-physician providers (i.e., physician assistants and/or nurse practitioners), and resident physicians. The following four criteria were used to select the intervention sites: facilities perceived that could benefit from improving their level of BP control; to increase generalizability of evaluation results, demographics had to vary (e.g., rural versus urban, minority proportion); investigators had to have established collaboration with leaders of these VISNs; and each intervention site agreed to provide 0.5 full time equivalent employees (FTEs) of nursing time to conduct the program.

Site A includes one academically affiliated medical center and four community-based clinics. In fiscal year (FY) 2011, there were approximately 90,000 primary care encounters among 20,000 among patients. Site B includes two academically affiliated medical centers where the intervention was offered. In FY 2011, there were approximately 140,000 encounters among 61,000 patients. Finally, site C implemented HTN-IMPRVE in one academically affiliated medical center and nine community-based clinics. There were approximately 98,000 primary care encounters among 45,000 unique patients in FY 2011.

Design

To support our qualitative assessment and to further understand organizational readiness for change across all three intervention sites, we administered a 13-item computer-based survey, the Organizational Readiness to Change Survey (Additional file 1). This survey was based on concepts from Weiner’s Organizational Readiness to Change Theory [9]. The survey has been pilot tested and modified for use in other settings [19–21]. We selected relevant items from Dr. Weiner’s survey item data bank [9]. This survey examined perceptions of organizational-level change efficacy and commitment to the HTN-IMPROVE intervention. Survey responses were used to objectively examine ORC as a two-dimensional construct encompassing change commitment and change efficacy. We examined responses to single-item measures hypothesized to be determinants of ORC such as task knowledge, perceived resource availability, and competing priorities, among others. All responses were anonymous. The ORC survey was administered to 336 primary care physicians, non-physician providers, nurses, and information technology professionals through the VA Intranet from the three intervention sites, and a total of 102 responses were obtained (a 30% response rate).

Qualitative analysis

Support for the hypothesized relationships was assessed by using three criteria proposed by Trochim [23] and Miles and Huberman [24]. First, we looked for the overall covariance of the constructs (e.g., whether VA intervention sites exhibiting strong implementation climate have supportive administration). Second, we looked for explicit attributions or the identification of plausible mechanisms to link the two constructs (e.g., participants attribute a strong implementation climate to the deployment of appropriate implementation policies and practices). Lastly, we applied the same criteria across the cases to determine if cross-case variation in implementation was consistent with the hypothesized relationships in the model.

Quantitative analysis

Survey responses used a five-point Likert scale (e.g., ranging from 0 = ‘not at all confident’ or ‘don’t know’, to 4 = ‘very confident’). We performed descriptive statistical analysis to calculate percent, mean and standard deviation for the ORC Items and Scales using Stata version 10.0 (Stata Corporation, College Station, Texas) and SAS version 9.2 (SAS Institute, Inc., Cary, North Carolina). Surveys with partial responses were included.

Organizational readiness to change (ORC)

ORC refers to the extent to which organizational members are prepared as a group to make the changes in organizational policies and practices that are necessary to implement and support innovation use (change commitment) and their perceived ability to do so (change efficacy) [25]. Administration was committed to implementation as demonstrated by their devotion of the required 0.5 FTE. Information technology staff members were confident in their ability to implement the software due to their previous experience with similar research and that there was buy-in from administration.

Still, the commitment by individual nurses and nursing administration was difficult to gauge. There was reluctance by the nurse interventionists due to uncertainty of time commitment and fear of creep in the scope of the time needed for the nurse-delivered telephone calls. There was also a fear that due to the length of implementation time, buy-in was waning, and the devoted nursing time would be permanently shifted to other tasks they had taken up during the pre-implementation phase. One nurse administrator said:

I would just say, with something like this, when it’s a research project, when it first comes out and the leadership is told we’re going to try this, we’re going to implement this here and then now we’re almost a year later. The timeliness of ‘we have an idea, we have a project we’d like to implement’, if it could happen within a three-month timeframe, I think buy-in and support would happen a lot better.

With regard to change commitment and change efficacy, organizational members were favorable toward the task demands required of HTN-IMPROVE, resources were adequate in the immediate future for implementation, and situational factors supported the program. Existing situational factors (e.g., telehealth), helped support buy-in and the view that the intervention could be implemented successfully. This is supported by responses to the survey (Table 2). Nearly 70% reported that the core group of people leading the implementation wanted to put HTN-IMPROVE into practice ‘very much’. Similarly, 69% said that the clinicians expected to support the program were either ‘fairly’ or ‘very committed’ to implement the program. Regarding motivation of the clinicians using the program, nearly one-quarter reported that they were ‘very motivated’ to implement the program, and 36% indicated that that they were ‘fairly motivated’. Of respondents, 39% reported that their clinical work group ‘very’ much wanted to implement the program, and another 43% reported that they ‘somewhat’ wanted to implement the program. Overall, the positive factors associated with implementation outweighed the negative, and the organizations expressed readiness for change.

Change valence

Change valence refers to the value that organizational members attribute to a proposed change [9, 25]. Change valence was largely positive with stakeholders expressing many benefits for patients including increased access to care, the ability to telecommute, both of which add to and complement current care. Value in the new program was expressed in large part because HTN-IMPROVE is an evidence-based intervention that has demonstrated efficacy and cost-effectiveness [7] and would fit with the organizations’ missions, goals and values of improving patient care. Ninety-three percent of survey respondents agreed or strongly agreed that ‘self-management programs fit with our approach to patient care’ (Table 2). Clinicians and administrators also perceived that the telephone-based aspect of the intervention was of particular benefit to patients; it is convenient with no need for patients to commute and would allow the intervention sites to provide cost-effective additional care to patients. One clinician stated, ‘I think the qualitative evidence should be highly stressed. You know, how are the veteran’s feeling? How are the clinicians feeling? Is there a sense of … pride in taking part in really helping the veterans out in this way and equally, if not more important, are the veterans really happy that they’re being reached out to a little more frequently?’

From a professional role perspective, nurses often expressed that the nurse-delivered intervention allowed them to practice to a fuller extent of their license and training; nurses would feel ownership of their patients, independence in their job, and increased job satisfaction. For example, HTN-IMPROVE allowed the nurses to move beyond routine clinic work and be increasingly engaged in their patients’ care, education and self-management.

Informational assessment

Professional role

As the interventionists and a principal stakeholder, the perception by the nurses that the nurse-delivered intervention would allow them to practice to a fuller extent of their license and training may have helped to facilitate collective efficacy of the intervention. The VA is in the process of developing protocols and policies expanding the RN role as a member of PACT teams. An organizing principle for these care teams is to utilize personnel at the highest level of their skill set. This aligns with the recommendations by the Institute of Medicine that nurses should practice to the full extent of their education and training and should be full partners, with physicians and other health care professionals, in redesigning US healthcare [29].

Including the patient as an important stakeholder

Interviewed stakeholders expressed the importance of the intervention for the patient population and the benefits it would bring in improving patient care. However, many logistical barriers were noted as well, such as being able to successfully contact patients and the perceived utility of the intervention by patients. Nurses were largely concerned that patients would view the intervention as a time to talk about topics beyond the scope of the intervention and would use the contact number excessively. These speculations highlighted the importance that a patient be represented as an important stakeholder to include from the planning and pre-implementation phase. Perceived value and pragmatic elements could be more easily resolved and answered with this important stakeholder at the table.

Limitations

Generalizability is limited to the organizational environment of the US Veterans Affairs Healthcare System. In retrospect, data analysis should have occurred concurrently with the interviews so that questions could be adjusted and added as needed. However, due to our own resource availability, we were limited in time commitment to be able to transcribe and analyze interviews while implementing the intervention. The response rate of the survey was around 30%, and interviews were self-selected.

Research was noted as a part of the intervention sites’ culture. The three VA intervention sites were all affiliated with research-oriented academic medical centers where clinicians and resources, particularly research, are often shared. Thus, buy-in for this translational intervention may have been easier and higher than it would be at other non-academic or research medical facilities. Furthermore, the patient was not included in the interviews and was an important stakeholder. Despite these limitations, our findings suggest many lessons learned with relevance to non-VA facilities that are moving forward with innovative changes.

Lessons learned