Phase I

A detailed description of methods used in the qualitative work conducted during the first phase of this study has been published elsewhere [19, 20]. In summary, healthcare providers working in remote First Nations communities in the Sioux Lookout Zone in north-western Ontario, Canada were selected using snowball and criterion-based sampling to participate in focus groups or interviews about the barriers to diabetes care. The Sioux Lookout Zone is home to approximately 16,000 people in 28 isolated and semi-isolated Cree and Ojibway communities. Larger communities have nursing stations with two to four permanent nurse staff, while physicians visit three to ten days a month, depending on the community's size. Otherwise, physicians are based in the Zone Hospital in Sioux Lookout (population 5,000), where they take phone calls, review lab results, and have occasional patient-related teleconferences. Data collected from the interviews and focus groups were coded and thematically analyzed using NVIVO 2.0 software. All participants provided informed consent and ethics approval for this portion of the study was received from the Research Ethics Board (REB) at the University of Toronto, Health Canada, and the Meno-Ya-Win Hospital in Sioux Lookout.

Phase II

The qualitative work in Phase I informed the development of a survey of healthcare providers as part of the CIRCLE study, a three-year national cross-sectional chart audit and survey study that documented the quality of diabetes care and rates of complications in 19 First Nations communities in Canada [6, 7]. For the purposes of this paper, quality of care is defined as adhering to the treatment recommendations and targets outlined by the Canadian Diabetes Association 2008 Clinical Practice Guidelines [18]. High-quality care can help prevent diabetes-related complications [21] although it is recognized that both social determinants and genetics may also play an important role [22, 23].

Phase I

Twenty-four nurses, doctors, and community health representatives (CHRs) participated in qualitative interviews and focus groups. Detailed results arising from the interviews and focus groups have been previously published [19, 20] and build on those demonstrated by Brown et al [25]. In summary, providers perceive barriers as falling into four categories: patient, provider, systemic, and environmental factors. Patient factors are the degree to which patients assume responsibility and control over their diabetes. Provider factors include barriers and facilitators to provider knowledge and skills, the implementation of clinical practice guidelines, and practice (re)organization. Lastly, systemic factors are defined as barriers and facilitators that relate to service accessibility and funding. An additional (non-modifiable) factor revealed from our study, and potentially distinct to Aboriginal care settings, was environmental factors as evident from the following quote - "Where else do people not have running water, unpaved roads, no vegetables or they're three times the normal price?"

The previous qualitative work was limited in that it did not allow us to determine the relative importance of these different barriers, nor was it able to identify how providers would prioritize strategies to overcome them. Such information is needed to inform the design of effective, provider-supported quality improvement interventions for diabetes care in First Nations communities.

Phase II

Ranking of Strategies

Addressing patient factors such as changing lifestyle, seeking preventive care and improving medication adherence was considered the most important approach to improving diabetes care. While one has to be careful to interpret provider perceptions of the best strategies with those that will actually impact care, this finding suggests that providers will be supportive of patient-focussed interventions. However, providers' support for patient strategies may also reflect a desire to blame patients for challenges associated with diabetes management in First Nation communities. A study conducted by Sonnenberg of gastroenterologists suggests that providers sometimes resort to blaming their patients for their medical conditions rather that admitting to the failures of disease management, thus making patient blaming an easy "exit strategy" for challenging or complex medical conditions [26]. Sonnenberg notes that this strategy is viewed as being the most productive (highest expected outcome for the physician) for single-visit interactions with patients. However, when patients are seen on a regular basis it becomes more obvious that cooperation and co-management is more productive. Since many rural and remote First Nation communities experience a high turnover of staff, it is possible that this clinical context enables a culture of patient blaming [27].

The impact of physician attitudes and perceptions on their management of diabetes patients has been noted elsewhere in the literature [28–30]. In one study, physicians reported that: "patients do not have a sense of urgency or understanding about treating their diabetes and thus are less likely to adhere to provider recommendations" [28]. This perception diminishes the knowledge and perspective of patients and suggests that providers do not (or could not) play a role in patient education. It is clear elsewhere in the literature, however, that health outcomes for patients with diabetes are improved when empowered patients work with informed providers [28, 31].

The lack of support for strategies within the provider factor -- the strategies over which providers potentially have the most control -- suggests that providers may not be aware of the potential impact that quality improvement initiatives can have on diabetes care. Successful initiatives from the United States [4, 32] and Australia [33] can serve as useful teaching points and models for the development of local, regional, and national quality improvement programs.

Variation by Isolation

The communities involved in the CIRCLE study were distributed across isolation levels, with only slightly more isolated communities than non- and semi-isolated communities. We hypothesized that the degree of isolation would have a significant impact on provider responses. While isolation level did have an impact, it was not as strong or consistent as was expected and the association was sometimes counterintuitive. Two results were particularly surprising. First, "address staff shortages," which was grouped under the patient factor, was ranked as the third most important strategy by semi-isolated communities, but did not factor in the top three for isolated communities (which are generally remote, fly-in communities). The reasons for this may be due to differences in the recruitment and retention systems for isolated and semi-isolated communities. Isolated communities tend to attract and retain staff using subsidies, attractive salary packages, and ample vacation time [34]. Semi-isolated communities may not have as attractive compensation packages compared to isolated ones.

Second, non-isolated communities were significantly more concerned with access to post-clinical services (49%) than isolated communities (23%). This is surprising given the remote location of isolated communities. Providers working in isolated communities generally provide a wide range of services to community members, as access to off-reserve care is often expensive and limited. Providers working in semi- and non-isolated communities, on the other hand, have a more diffuse population that may access services through other providers. Services are also more diffuse, meaning that providers rely more heavily on referrals. This combination of factors may actually make accessing consistent services in semi- and non-isolated communities more difficult and may make follow-up with post-clinical services less coordinated than in places with only one entry point for care.

Variation by Provider Type

Analysis by provider type found some significant differences in perceptions of important diabetes strategies among doctors, nurses and CHRs. There was no significant variation in responses of providers who were classified as clinic mangers, diabetes educations, and dieticians/nutritionists. For diabetes educators and dieticians/nutritionists this is likely due to the small sample size, with only 6 and 4 respondents respectively. It was more surprising that the 29 clinic managers did not significantly differ from nurses, doctors, and CHRs. It is possible that clinic managers group with the provider type reflective of their training (i.e. grouping with nurses if they are a registered nurse or physicians if they are an MD) or that their voices were diffuse and did not group cohesively.

The variations between the responses of doctors, nurses, and CHRs are interesting and require further discussion. For instance, being a CHR or a doctor affected respondents' perceptions of the importance of communication: CHRs viewed communication between patients and providers as significantly more important than doctors. This finding is particularly interesting given providers' strong overall focus on patient strategies and some of the concerns raised previously in this section about the potential influence of patient blaming by providers. The differences identified between CHRs and other providers is worth considering further as CHRs are generally Aboriginal and from the community and, thus, potentially more likely to have similar life experiences or be more empathetic to the perspectives and needs of patients than to the culture and background of physicians and nurses (who, for the most part, are not Aboriginal) [35, 36].

The literature on patient-provider communication links good communication with improved patient health outcomes [37] and notes that barriers to effective patient-provider education stem from social and cultural differences between patients and providers. These include: lack of provider understanding of the history of Aboriginal people and the negative impacts of this history on life today (i.e. poverty, poor housing, high unemployment, and expensive food in the north); limited time for patient-provider dialogue; and, cultural differences in verbal and non-verbal communication [38–41]. Teaching healthcare providers effective communication skills and educating them about the history of Aboriginal people (including the negative impacts of colonization and the residential school system), the importance of trust-building, and how to make time for effective communication has the potential to improve provider communication with Aboriginal patients, which may positively influence patient outcomes [41]. However, the literature also notes that changing provider behaviour is difficult: in one study, the participation of doctors in intensive communication skill training did not improve their adherence to communication best practice [42]. Overcoming the underlying patterns of social distancing (and sometimes 'Othering') that stems from unprepared providers working in remote communities with social conditions different from those normally experienced by healthcare providers is another significant challenge [43].

An alternative strategy is to focus on enhancing the role of community health representatives. Their unique position in the community, along with their desire for more specialized training, suggests that with greater support CHRs could serve as a much-needed bridge between the community and the clinic.

Strengths and Limitations

Developing surveys from rich qualitative work is an effective way to gain a holistic perspective of a particular issue in health research [44]. This study adopted this approach by building a survey of strategies for improving diabetes care in First Nations communities from a qualitative study in which providers identified three thematic categories of the different barriers to care. This well-grounded survey provided interesting and relevant results about provider perceptions of the relative importance of different strategies and variation of provider responses across isolation level and provider type. To our knowledge, this is the first survey to include a national sample of healthcare providers working in 19 geographically and culturally diverse First Nations communities in Canada.

There were also some limitations with this design. The first concern is the representativeness of the sample: the CIRCLE study was based on a sampling frame where communities were recruited based on pre-existing relationships with researchers. While this may have caused some bias in the results, the lack of heterogeneity of the responses suggests that it was not a major concern; key issues seemed to be common across communities. Secondly, the heavier weighting of nurses and doctors in the sample meant that the results likely reflected the views of these two types of healthcare provider types. Comparisons enabled us to reveal the statistically significant differences between different provider types (i.e. doctors versus CHRs) and, thus, reflect the heterogeneity within our sample. Because of the small number of some provider types (i.e. dieticians and diabetes educators), no significant comparisons could be made. Finally, the CIRCLE study focussed on assessing the quality of diabetes care in First Nations communities. While this clinical focus was important, it did not permit the inclusion of a patient voice. We believe that the viewpoint of patients and how they perceive the current context of care is absolutely essential to complete the picture.

Practice

Robust findings show that patient factors are an issue of great importance to providers, which suggests that providers may be receptive to quality improvement strategies and chronic care programs focused on patient engagement. One concern with providers' focus on patient factors is that it may be a guise for patient blaming or for off-loading responsibility. However, providers can play a role in patient engagement, such as understanding patient perspectives, local history and culture, culturally appropriate self-management support programs, and broader quality improvement initiatives. Such initiatives could include clinic reorganization with increased training of CHRs so that they can offer peer-counseling in diabetes management, and help entire families or indeed the entire community to adopt healthier lifestyles [45–47]. Changes such as this may help refocus providers on their own role in diabetes management.

Policy

Policy-makers play a key role in the implementation of research into practice. This study identified two particular areas where the knowledge and resources of policy-makers could facilitate greater understanding and potential solutions to identified challenges: (1) Staff Shortages - policy-makers should investigate staffing levels in semi-isolated communities and develop solutions to overcome potential human resource challenges in these communities; (2) CHR Training - policy-makers should explore the potential for further CHR training and support.

Research

This study focused on provider perspectives to diabetes care in an effort to better inform quality improvement interventions. Research incorporating patients' perspectives about factors that hinder or improve diabetes care is absolutely needed to complete this picture. Comparing these viewpoints may also help facilitate deeper conversations and analyses about patient-provider communication and the roles of patients, families and communities, as well as their providers in diabetes treatment and management.