Implementing a Standardized Care Pathway Integrating Oncology, Palliative Care and Community Care in a Rural Region of Mid-Norway

The project was conducted between 2013 and 2019 [21]. An integrated oncology and palliative care outpatient clinic (hereafter: Integrated Clinic) was established in November 2012 in a hospital serving 97,000 inhabitants in rural municipalities. Founded on ideas conceived by authors of the current paper, the clinic was organized to treat patients at all stages of cancer and to integrate cancer and palliative care [21]. Its structure is described in Brenne/Knudsen et al., 2020 [21]. An agreement between the hospital and 13 municipalities described the collaboration for improved cancer care. The municipalities had between 980 and 11,000 inhabitants, and the longest driving time from the municipalities to the hospital was 2 h (130 km). At project start, there was no electronic communication between the levels, or within the municipalities.

For the collection of quantitative outcome variables, adult cancer patients diagnosed with advanced locoregional or metastatic disease, treated with non-curative intent, inhabited in one of the participating municipalities, and able to read and write Norwegian, were consecutively screened and recruited between November 2014 and December 2017, and followed up until death or December 2019 [21]. Healthcare leaders (HCLs) and healthcare professionals (HCPs) exposed to the implementation strategy and eligible for receiving questionnaires were identified by their leaders. Eligibility was fulfilled if being employed in one of the participating municipalities or at Orkdal Hospital, at least 18 years of age, and able to sign an electronic informed consent. Questionnaires were distributed in December 2014 (by email) and January 2017 (by mail). In the fall of 2015, questionnaires on the use of patient-held records (PHRs) were sent by mail to 102 nurses and nurse assistants, identified by their leaders to be working with cancer patients with palliative care needs. For the qualitative evaluation, HCLs and HCPs were interviewed in the spring of 2014 (during the development process of the SCP) and in the fall of 2015 (1 year after the SCP was launched). Patients were interviewed on their use of PHRs in the fall of 2015.

The implementation was conducted on a system level. The implementation strategy consisted of an information plan, appointment of local facilitators, and an education program (Fig. 2) [18]. The planning of the intervention started in 2013 [21]. Between 2014 and 2018 the implementation strategy was directed towards political and HCLs, HCPs, and the public. Political leaders were informed and attended public meetings to increase project focus. HCLs were committed through personal meetings and written information. HCPs were engaged in the development of the SCP and recruited as process facilitators. The facilitators promoted the use of the SCP and identified barriers hampering its use. An educational program was conducted for hospital and community HCPs (Supplementary Material). The program included comprehensive cancer care delivered by the means of an SCP and education on essentials of oncology and palliative care. Physicians, nurses, and other professionals with expertise in oncology and palliative care administered the program. To maximize knowledge dissemination, similar events were held at different locations. After participating in the education program, the HCPs were encouraged to teach their colleagues at their own workplace. In addition, a network of resource nurses facilitated care competence and collaboration [21]. The public, including patients and carers, were informed at meetings, by leaflets and newspaper articles, and online. Due to the slow inclusion rate of patients, the education program was prolonged with 2 years (from 2016 to 2018) without extra funding.

The intervention and the implementation strategy were evaluated according to implementation outcome variables and reported in line with standards for implementation studies [15, 17]. A mixed methods design was applied [26]. Nature and number of activities related to the intervention and implementation strategy were registered prospectively. Semi-structured interviews were performed with three different groups of informants. (1) Patients and HCPs were interviewed individually regarding their experiences with the PHR. The interview guide was developed by two of the authors (TIB and AKK). The questions were formed to explore patients’ and HCPs’ experience with the PHR and how it was used (Table 4, interview guide 1). TIB conducted the interviews. (2) HCLs and HCPs were interviewed individually on the intervention and the implementation strategy. The interviews were conducted during the development process and focused on implementation challenges. MNN and KR developed the interview guide (Table 4, interview guide 2) and conducted the interviews. (3) HCPs were interviewed individually regarding their experiences with the implementation of the SCP. MHJ and AKK developed this interview guide, and topics were HCPs’ experiences regarding success criteria and barriers for implementation, management anchoring and the SCP’s role in reaching central goals in palliative care (Table 4, interview guide 3). MHJ conducted these interviews. The sampling was purposive for all three groups to include a selection of patients, leaders, and representatives from different healthcare professions affected by the project [27]. All interviews were audio-recorded, transcribed, and analyzed using a stepwise-deductive induction approach (MNN and KR) [28] or systematic text condensation (TIB and MHJ) [29]. For the interviews where systematic text condensation was used for analysis, data collection was stopped when saturation was reached.

Estimated costs of the intervention and the implementation strategy were based on knowledge of funding and structure of the local healthcare services.

Change over time in healthcare providers’ knowledge and skills, and implementation and use of the SCP for palliative care, were secondary and exploratory outcomes of the Orkdal Model Study (ClinicalTrials.gov Identifier: NCT02170168; primary outcomes published elsewhere). Therefore, power calculations for these outcomes were not done. Pre study, approximately 1300 HCPs were estimated to be asked for study participation, and with an anticipated response rate of up to 50%, a maximum of 650 healthcare providers would be included [30]. Descriptive statistics calculating frequencies were performed, and the Fisher’s exact test was used for group comparison. IBM SPSS Statistics 25 (Statistical Product and Service Solutions) and Stata Statistical Software (Release16.0; StataCorp LP, College Station, TX, USA) was applied.

The regional committee for medical research ethics approved the study (2014/212). All participants gave their consent to participate. HCLs and HCPs interviewed during the spring of 2014 did not provide written consent, as these interviews were part of a quality assurance project at St. Olavs Hospital, Trondheim University Hospital [31, 32]. The participants signed separate consent forms for the quantitative and qualitative data collection.

Three hundred and nine cancer patients were potentially eligible for the SCP. Of these, 231 were approached for inclusion. One hundred and one patients declined participation and one withdrew consent. For data analysis, 129 patients were included. By the time of inclusion, 106 (82%) received anti-cancer treatment. Mean age was 70 years (range 38–92) and 81 (63%) were men. One hundred and eleven (86%) had metastatic disease, and 99 (77%) had a Karnofsky performance status of 80% or more [33]. Most common cancer diagnoses were prostate (23%), colorectal (13%), urinary organs except prostate (11%), upper gastrointestinal except pancreas (10%) and pancreas (9%). By the end of follow-up, 23 patients were alive.

Fifteen patients were interviewed regarding their experience with the PHR, six men and nine women. Mean age was 76 years, and all had Eastern Cooperative Oncology Group (ECOG) performance status I or II [34].

In total, 1320 HCPs were identified by their leaders to be exposed to the implementation strategy and received questionnaires, of which 119 (9%) were employed in specialist care. Seventy-six (5.8%) were leaders, 69 (5.2%) physicians, 391 (30%) nurses, and 784 (59%) nurse assistants. Three hundred and fifty-five (of which 318 (87%) were from community care) signed informed consent to participate (response rate 27%). There were 21 drop-outs after the first round and 36 drop-ins after 2.5 years (Table 5). Thirteen percent of the participants from community care worked in Orkdal municipality, the municipality where the hospital was located. To assess adherence to PHR, separate questionnaires were sent to 102 HCPs working with cancer and palliative care in the region.

Altogether 60 HCPs were recruited as informants for semi-structured interviews, of which 43% were employed in specialist care. Fifteen percent were leaders, 35% physicians, 42% nurses, and 8% nurse assistants. Of these, 19 HCPs (10 from specialist care and nine from community care) involved in the development of the SCP underwent semi-structured face-to-face or telephone interviews in the spring of 2014. In the fall of 2015, 21 HCPs (eight from specialist care and 13 from community care) were interviewed face-to face regarding their experience with PHR, and 20 HCPs were interviewed (8 from specialist care and 12 from community care) about the implementation of the SCP.

Both hospital and community HCPs acknowledged the SCP as a method to improve palliative cancer care and reported improved quality of care after the implementation. A GP said, “It has improved, it is probably because of the pathway, also for us who live outside of Trondheim, it is because the cancer outpatient clinic is here. Then they manage to implement this pathway much better.” In specialist care, the relevance and perceived fit were related to increased competence and improved structure. In community care, this view was counterbalanced by a worry for decreased attention to other diagnoses and needs for local adaptations. However, personnel in community care endorsed the patient-centered focus on function and individual needs. The workgroup agreed on a common strategy for the development and content of the SCP, and on the suitability both for everyday use and for reaching central goals in palliative care. Furthermore, the SCP was accounted feasible, provided sufficient training in its use. Factors facilitating the practicality included safe admission and discharge routines and patient involvement. The use of checklists and instant transfer of medical information were considered important factors for the intervention’s actual fit, whereas incompatible ICT systems were regarded as a challenge.

Fifteen cancer patients were interviewed on their experiences with the PHR. The patients reported limited use and limited benefit of the PHR. They preferred to use systems they were familiar with or did not feel the need to write. A patient said, “You take so good care of me that there is nothing special to follow up, so far”. Still, some patients appreciated the availability of an updated medication list.

Forty-five HCPs responded to a questionnaire regarding their experience with PHR, of which 32 were nurses and 13 nurse assistants (response rate 44%). Sixty-one percent were familiar with the PHR, but only one third took care of patients using it. The PHR use was restricted by a perception of limited benefit and uncertainty regarding accuracy, especially for medication lists.

Two hundred and eighty-seven HCPs (response rate 22%) reported on SCP use 2 months after project start and 152 HCPs (response rate 41%) after 2.5 years (Table 5). Fifty-three (18%) used the SCP at 2 months and 43 (28%) at 2.5 years (p = 0.02). During the same period, data from 325 hospital admissions were registered for the 129 patients. For 158 admissions (49%), relevant medical information was transferred the same day as discharge. At 3 years (in 2017), five out of 13 municipalities reported that they used the SCP nurse checklists.

The SCP was developed without dedicated funding. A task force revised the SCP three times during the study period based on revised guidelines and practices, user feedback, and personal experiences.

Interviews with HCLs and HCPs revealed that the implementation strategy was perceived as agreeable and relevant (Supplementary Material). Because of different perspectives, structures, and priorities in hospital and community care, dissimilar components of the implementation strategy were emphasized. Still, understandable relevant information, access to local experts, and SCP-related education were perceived useful. In addition, the informants considered purposeful interaction across levels of care and user involvement important for the feasibility of the implementation strategy. Local facilitators were regarded important for promoting patient safety through smoother transitions between places of care and more predictability for the patient and carer. Finally, the patient-centered focus was considered important for the actual fit of the implementation strategy.

Three major themes contributing to success or failure with the implementation of the SCP emerged from analysis of interviews with HCPs: “competence”, “coordination”, and “patient and carer” (Supplementary Material). HCPs reported increased competence, coordination, and quality of care after the implementation. The network of resource nurses and decentralized education were considered important for the successful dissemination of the SCP. However, the degree of implementation strategy success varied at different locations. Factors related to lack of success were suboptimal competence in palliative care, lack of process ownership, fear of giving one patient group priority, and limited involvement from the management. In addition, restricted funding and lack of an electronic SCP search function hampered the implementation.

Prospectively registered data regarding activities related to the intervention and implementation strategy showed that the implementation strategy was delivered as intended. The invited HCLs and HCPs attended the meetings, and the education program was conducted as planned. The number of participants of the education program was given in Supplementary Material. The majority of the participants were from community care and all municipalities had HCPs attending the education program. GPs from seven of 12 municipalities (the smallest municipality was served by GPs from the neighbor municipality) attended the education program. Thirty-seven process facilitators were recruited, five from hospital and 32 from community care. The SCP was revised yearly, because every 6 months turned out to be too resource demanding.

Two hundred and eighty-seven HCPs reported on training in use of the SCP 2 months after project start and 153 after 2.5 years (Table 5). Response rates were 22% and 41%, respectively. Eighty-three (29%) reported that they were trained in using the SCP at 2 months and 63 (41%) at 2.5 years (p = 0.01). One hundred and eight HCPs (response rate 8.2%) reported on collegial teaching 2 months after project start and 155 (response rate 42%) after 2.5 years. Thirty-four (31%) had performed collegial teaching at two months and 31 (20%) at 2.5 years (p = 0.04). For confidence with opioid treatment, 168 HCPs reported at two months and 95 after 2.5 years (nurse assistants excluded; response rates 37% and 47%, respectively). Seventy-two percent felt confident after 2 months and 85% after 2.5 years, respectively (p = 0.02), and for confidence with end-of-life care, 77% and 84%, respectively (p = 0.06; response rates 22% and 41%). Two hundred and ninety-three HCPs (response rate 22%) reported on use of symptom assessment tools 2 months after project start and 152 (response rate 41%) after 2.5 years. One hundred and forty-six (50%) applied symptom assessment tools at 2 months and 94 (62%) at 2.5 years (p = 0.02, Table 5).

The costs of the implementation strategy included grants for hospital employment of two nurses and one assistant for 12 months and two community care physicians for 19 months. In addition, community care nurses from ten municipalities received job training at the local hospital.

An SCP across levels of care represented the intervention introduced to palliative cancer patients in a rural region of Mid-Norway by a comprehensive implementation strategy. Aspects of the intervention and the implementation strategy were evaluated using a mixed method approach. SCP as a method to improve palliative cancer care was found to be feasible among the HCPs. Despite different perspectives of care in specialist and community care, there was agreement on a common strategy for the SCP. Focus on patients’ function and needs, and not diagnoses, were endorsed. Sufficient training in its use was regarded important for the success of the SCP. Different organizational structures and ICT systems challenged implementation. The implementation strategy, running over a period for 4 years, was delivered as intended. Information and educational plans, appointment of local facilitators, and close collaboration with the regional network of resource nurses were essential parts of the strategy. Elements facilitating care integration were access to assessment tools, nurse checklists for care coordination, patient-held records, available treatment recommendations, and internet access. Nearly one third of the HCPs reported use of the SCP 2.5 years after the intervention was implemented. The use of essential elements of the SCP varied, and for only half of the admitted patients, medical information was transferred the same day as discharge. Forty-one percent of the HCPs reported that they had received training in using the SCP 2.5 years after the implementation was launched. Both HCLs and HCPs reported improved quality of care in the wake of the implementation. Limited project involvement and knowledge seemed to restrict the success of the implementation strategy.

Already eleven years ago, the beneficial effects of early palliative care in cancer care were reported [1]. However, these first reports were from specialist care [1, 3, 35]. Since then, there has been a growing attention to the need of strengthening community palliative care to reach all in need of palliative care [36]. In our model, specialist care and community care cooperated through active use of the SCP making the trajectories as seamless as possible. The GPs, together with the community nurses, played an essential role in providing palliative care to the patients at home, strongly supported by the decentralized Integrated Clinic [21].

There is little research on care pathways in palliative medicine [12]. A clinical care pathway for palliative cancer patients may result in reduced symptom intensity [37]. However, a Cochrane review found no evidence to support end-of-life care pathways [38], and a cluster randomized trial on a generic care pathway for elderly patients in need of home care services was inconclusive [22]. Integrating care pathways into clinical practice is complex and depends on contexts both inside and outside of the organization [39]. To measure how and why a care pathway is working may be difficult [39]. In our study, the education program was pointed out as an important part of the implementation strategy. Participation in the education program was high. A Canadian project evaluating integration of early palliative care into routine oncology also highlighted the importance of training of HCPs to succeed with integration [40].

In addition to improved patient outcomes, an SCP may enhance competence and quality of care [41]. A systematic review and meta-analysis on integrated outpatient oncology and palliative care services reported improved quality of life, symptom burden, and survival [4]. In the current study, even with the low response rates, the quantitatively evaluated implementation strategy outcomes indicated an increase in HCP-reported confidence with opioid treatment and use of assessment tools. However, increase in confidence with end-of-life care did not reach statistical significance and there was a decrease in collegial teaching during the study period. Pre-study level of confidence and competence might influence the results.

A multinational cluster randomized trial studying an in-hospital care pathway found no significant effect on the primary outcome, which measured the communication and relationship between HCPs [41]. This may be explained by the time and complexity of interventions needed to improve interprofessional relations [41]. In our study, after implementation medical information was transferred at discharge only for half only the evaluated patients. This despite the emphasis put on communication in transitions of care in the SCP.

We applied a mixed methods design. A mixed methods design combines quantitative and qualitative approaches and is considered particularly suitable for implementation research [17]. The combination of methods provided opportunities for measuring the implementation of the intervention and for understanding the process [26]. The design also made possible a comparison of data collected by different methods [26]. In addition, the design gave an opportunity to monitor the implementation process over time and to compare anticipated and experienced challenges regarding the implementation strategy [42]. Our study described recommended intervention and implementation strategy outcomes, according to standards for reporting implementation studies [15]. However, the evaluation was conducted among small fractions of the included patients and affected HCPs, and self-reported data were used. This raises the question whether the samples were representative for the populations from which they were drawn, as few measures were made to reduce the risk of selection bias [43, 44]. Additionally, the relative response rates dropped during the study period. The questionnaire response rate was 27% at the first survey, and only those who signed informed consent received the second survey. Retrospective review of patient records regarding the use of the SCP may have provided a more reliable result [45]. The number of HCPs was tenfold the number of patients, limiting the plausible contact with patients included in the SCP for each respective HCP. Furthermore, less than one percent of the nurse assistants were interviewed. This group of HCPs constituted almost 60% of the recipients of the implementation strategy. Nurse assistants may represent HCPs benefiting from the implementation of the SCP. Finally, no carers were included in the evaluation of the intervention or the implementation strategy.

The effectiveness of a patient-held record is dependent on it being used both by patients and HCPs [25]. In our study, patients and HCPs reported limited use of the PHR. Based on these findings, no definitive recommendations regarding the benefit of PHRs in integrated oncology and palliative care pathways can be provided. With electronic transfer of medical information, the benefit of a paper-based PHR could further be questioned.

Checklists are applied to improve standardization and reduce preventable errors, and the use of palliative care checklists may improve documentation [46]. Effective implementation of checklists is dependent on a coordinated effort from both HCLs and HCPs [47]. In the current study, only five out of 13 municipalities reported use of the nurse checklists after 3 years. Thus, the implementation strategy provided limited sustainability on this matter. Possible reasons for this attrition were that the municipalities had their own checklists or an incompatible ICT system. Future work must address the creating of more enduring structures supporting the implementation of new standards [47]. This may be challenging due to different priorities and standards in specialist and community healthcare services, as reported in our study and by others [14].

Efforts to implement complex interventions interact with individuals, collectives, organizations, and political and economic systems [48]. Reasons for failure may be related to lack of effect within the context tested, not meeting the needs of the stakeholders, or related to contextual instability, such as key staff replacement and insufficient funding [48]. In our study, the SCP was developed bottom-up, which one could believe would increase its use [49]. We observed an overall absolute improvement in the quantitatively collected implementation strategy outcomes of about 10% (except for collegial teaching). The clinical relevance of these improvements can be questioned. We did not predefine a satisfactory program fulfilment rate. A retrospective study evaluating the process of implementing a care pathway integrating oncology and palliative care at Oslo University Hospital found that the care pathway was not used as intended in terms of patient-reported outcome measures (PROMs) [45]. The authors of this study recommended a more complex strategy to implement the use of PROMs in oncology. In our study, interviews with HCPs revealed that especially different ICT systems and variable management anchoring were barriers that should be more strongly addressed in the future. Enhanced follow-up of the process facilitators may also have improved the outcome, and one may speculate if change in key staff and lack of funding contributed negatively. The SCP was created mainly by staff from specialist care (Table 1), another factor possibly contributing negatively to implementation. Still, both HCLs and HCPs reported quality improvements after the implementation of the SCP, underlining the potential benefits of the approach. Future research must address the essentials of an integrated pathway in palliative cancer care, the necessary content of a successful implementation strategy, and measurement properties of instruments measuring integrated care [50].

Almost 40% of the included municipalities reported interest in continuing the project after formal closure (ATB, personal communication, 2020). In the aftermath of the current study, a national cluster randomized controlled trial with a complex intervention of compulsory early integration of palliative care was developed in Norway [9]. An SCP, an education program, and systematic symptom assessment constitute the essentials of the intervention. Looking further into the future, an SCP with integrated decision support may represent a promising approach [37].