Planning, implementation and operation of a personalized patient management system for subjects with first suspect of cancer (OnkoNetwork): system description based on a qualitative study

The idea of OnkoNetwork first came up in July 2014 when the Strategic Director of the Móritz Kaposi General Hospital set up a small interdisciplinary team to understand the root causes of the unexpectedly high cancer mortality in Hungary [19]. This county hospital is responsible for the comprehensive oncology care of about 500,000 inhabitants in Hungary, in close cooperation with the Kaposvár University Health Center (responsible for CT, MR, PET CT, PET MR diagnostics and also for radiotherapy), with outpatient cancer care departments of the Hospital in other locations (Siófok, Nagyatád, Marcali), and with a radiotherapy unit in Mosdós [4]. The team looked for quantitative indicators with high variability that could be controlled by the above institutions and identified two parameters calling for improvement. First, diagnostic delay in the centre (time from the first presented suspect of cancer in the medical system of the participating institutions to the Tumour Board meeting with final diagnosis) was found to be as long as 6 months in many cases. Secondly, the treatment initiation delay (time from final diagnosis to the first treatment day) was also longer than the optimal time window for many patients. To reduce these time windows in the catchment area of the Hospital, a new system to manage and optimise patient pathways was proposed as a local initiative. The planning phase was based on intensive teamwork involving both the hospital management and the heads of all hospital departments contributing to cancer diagnosis or treatment (listed on Fig. 1). After repeated meetings and strategic discussions, the ultimate objective of OnkoNetwork was framed as “provision of timely and equitable access to comprehensive and integrated oncology care”. Potential economic benefits (such as a reduction in average treatment costs) were not the primary focus in the planning phase. OnkoNetwork was based on several principles. First of all, quality assurance was facilitated by the development and implementation of organ-specific, evidence based clinical protocols. Secondly, tracking of each patient by an identified responsible physician had to be made continuously to prevent patients from being lost in the system. Thirdly, quantitative indicators were defined: the target for a conclusive Tumour Board decision on the proposed therapy was set at a maximum of 30 days of enrolment into OnkoNetwork, and thereafter the treatment should be started within 14 days. To support these principles, development of a new tailor-made IT system was decided for real-time collection and monitoring of the clinical documentation of oncology patients [4]. Given the volume and complexity of the existing medical IT system of the Hospital, the realistic ambition was to develop the new IT tool running in parallel with pre-existing medical IT applications, rather than as a replacement of them.

OnkoNetwork was implemented in a pre-planned change management process. The first implementation step was the development of organ-specific diagnostic algorithms: available clinical protocols that typically contained a couple of optional investigations to be selected along not fully explicit criteria were translated into organ-specific diagnostic algorithms with standardized lists of all necessary investigations by the organs affected. Harmonized process descriptions of hospital departments were also developed. After consensus was achieved in these aspects, the development of a customized IT application (OncoLogistic) was started, in parallel with the thorough reorganization of teams and harmonization of oncology care in all relevant departments [4]. Since no relevant prior experience could be identified in Hungary, the workflow description and the definition of new professional roles were started from scratch, based on thorough discussions involving all stakeholders. An OnkoNetwork Office (see below) was established and OncoLogistic functions were continuously revised and fine-tuned. OnkoNetwork was officially launched in November 2015, after a 1-month test period.

After this brief overview of the planning and implementation phases, the next sections describe the OnkoNetwork model along each component of the SELFIE framework.

The target population of OnkoNetwork consists of adult (≥18 years) patients with new suspect or new diagnosis of solid tumours (an ICD disease code starting with “C” or “D”, except for haematology malignancies, and some rare tumours that are referred to specialized national centres) in the catchment area of the Kaposi Mór General Hospital at Kaposvár. This catchment area roughly overlaps with Somogy county of Hungary with a 500,000 population, including patients visiting the Oncology outpatient units of Municipal Hospitals at Nagyatád and Siófok. Having chronic comorbidities is not an enrolment criterion in OnkoNetwork, the most frequent chronic comorbidities in the enrolled population are cardiovascular diseases, hypertension, and diabetes occurring in about 49, 26, and 11%, respectively. Diagnosis and adequate treatment of comorbidities interfering with cancer care are inevitable in many cancer patients, e.g. pacemakers need to be switched off or removed during high frequency ablation; patients receiving cardiotoxic chemotherapy require cardiologic control; or metformin treatment of diabetic patients need to be discontinued before CT scan to avoid kidney damage, as a few examples.

Before OnkoNetwork, the organization of diagnostics was a time consuming and stressful administrative burden on physicians who had to fight for early visit dates for their patients. OnkoNetwork gave priority status for patients with suspected cancer throughout the Centre and passed this organization task to selected department administrators. Department physicians allocate their spared hours to more extensive patient communication, to the treatment of more patients in a day, or to other professional tasks. On the other hand, the programme limited the independence of physicians and forced them to collaborate with each other along personalized multidisciplinary diagnostic and treatment plans. To ensure the continuity of care, all enrolled patients have a named responsible physician who is competent to decide on their diagnostic tests and is responsible for any delay in diagnosis and therapy initiation. The responsible physician changes over time in the patient path: is typically a physician from the relevant organ-specific department in the diagnostic phase, a surgeon in the perioperative stage, and an oncologist during chemo- and radiotherapy episodes. The responsible physician must be available for discussions with supervisor physicians, and is also available for departmental OnkoNetwork administrators, but shall not be contacted by OnkoNetwork Office administrators (see the description of these new roles below).

One to three assistants or administrators have been selected, appointed and trained in the relevant departments of the Hospital to become departmental OnkoNetwork administrators. Their related tasks are 1) case finding, i.e. to invite and enrol candidate patients from the target population; 2) to fill in the OnkoNetwork questionnaire at programme entry based on a patient’s interview; 3) to schedule visits for protocol driven and responsible physician initiated diagnostic tests; 4) to record data into OncoLogistic, including the responsible physician’s explanation for any delay or deviation from the protocol; and 5) to monitor therapy initiation after the Tumour Board decision. Most OnkoNetwork administrators in the departments received their new roles on top of existing ones, except for departments with highest enrolment rates where the OnkoNetwork administrator function can be a full-time position.

In contrast to departmental OnkoNetwork administrators described in the previous section, OnkoNetwork Office administrators do not belong to any hospital department. Interestingly, most OnkoNetwork Office Administrators did not have prior education or experience in healthcare service provision but were hired based on their good organization and communication skills. Their main role is monitoring and managing patient pathways directly and through contacts with OnkoNetwork administrators in linked departments, as supported by the tailored IT application OncoLogistic. In a typical day, OnkoNetwork Office administrators identify about 40–50 OnkoNetwork candidate patients based on their recorded ICD codes, and systematically check whether they are invited and enrolled into OnkoNetwork by the departmental OnkoNetwork administrators. They also look after all enrolled patients with a delay in their pathway on a daily basis. When they identify a patient with an apparent delay, they first clarify whether it is due to missing data transfer across medical IT systems. If not, they turn to the OnkoNetwork administrator(s) of the relevant department for clarification. They always accept the received justification but secure that it is recorded in OncoLogistic. OnkoNetwork Office administrators notify the supervisor physicians (see below) about any unsolved issue.

One supervisor physician is working at the Hospital and another one at the Kaposvár University Health Centre. Their primary role is to negotiate those OnkoNetwork related issues with department physicians that could not solved at the non-physician level. Any remaining unsolved cases are referred to the Strategic Director of the Hospital on a weekly basis. The supervisor physicians do not interfere with medical decisions of department physicians, but request justification for unexplained delays in diagnostics and care. The selected supervisors have a calm and helpful personality that efficiently relieves the psychologic stress on the involved non-physician team members.

Since no relevant prior experience could be identified in Hungary, the workflow description and the definition of new professional roles were started from scratch. Involvement of all stakeholders in the planning, and building consensus agreements were important success factors in the planning phase. Training for the new roles included interactive plenary lectures and practice oriented technical presentations on model objectives and workflows. Written guidance to the custom IT tool OncoLogistic was also distributed, but its technical language could not be easily understood by several staff members. However, a few months of learning by doing experience was sufficient for all administrators to get fluent in OncoLogistic.

At implementation, some surgeons were irritated as they could not make independent decisions on the treatment (e.g. when to take a histology sample). Their resistance was overcome by the involvement of department heads in consensus agreements, strong commitment of the management, and early perceived benefits of the model, e.g. scheduling the diagnostic visits by OnkoNetwork administrators. A further barrier came from the culture of many physicians who were not prepared for the new professional roles of non-physician colleagues. OnkoNetwork did not change this attitude of physicians but circumvented this situation by the establishment of the supervisor physician role in the model. Supervisor physicians are much more accepted than non-physician team members as partners by physicians.

In Hungary, patients with more complex medical needs have long waitlists to their definitive care even within a single institution [4, 5]. Accordingly, key opinion leaders of OnkoNetwork decided to focus first on the integration of care processes that were under their control in the region, i.e. hospital care and secondary outpatient care provision. Nevertheless, they recognized that integrated care should cover the holistic needs of patients including prevention, primary healthcare, social care, psychological support, physiotherapy and dietary advice. The OnkoNetwork Management identified primary care and cancer screening programmes as key contributors to early detection of suspected cancer. Note that cancer screening programmes in Hungary are organized by special health promotion offices, without data linkage to oncology centres and without follow-up system of the identified positive screening cases. Regional integration of OnkoNetwork with cancer screening programmes is an important ambition of the OnkoNetwork management.

Initial analyses suggest shortened timelines for diagnosis and treatment initiation, as expected. The mean time from initial indicative diagnosis in the hospital to Tumour Board meeting is reported to be 21–22 days in OnkoNetwork with a standard deviation of 14 days. Before OnkoNetwork, the mean duration was 64 days with a standard deviation of 40 days in the Hospital. For comparison, the mean duration of cancer diagnosis in other Hungarian hospitals can be as high as 110 days [12].The interviewees expected that OnkoNetwork will result in better patient compliance in the diagnostic phase, less incentives for patients to visit private healthcare providers in order to speed up the diagnostic process, faster diagnosis and treatment of comorbidities, lower frequency of inconclusive Tumour Board meetings, earlier start of treatment, lower proportion of patients with cancer metastases at diagnosis, longer patient survival, and/or better nutritional status of patients with head-neck tumours during therapy. Some interview participants suggested that more patients would receive surgical care due to a shift in cancer stage distribution at diagnosis towards less progressed cases, while the opposite expectation was also mentioned referring to the decreasing dominance of surgeons in multidisciplinary treatment decisions. Future evaluation of clinical outcomes will be facilitated by the ongoing patient-level data collection in OncoLogistic since the completeness, timeliness, and face validity of the entered data is monitored and assured by the OnkoNetwork Office.

The interviewees expected that OnkoNetwork will result in improved patient experience, since OnkoNetwork allows more time for physicians to communicate with their patients, the waiting lists are getting shorter, patients experience less need for travel and shorter waiting times on the day of diagnostics, and the number of patients intending to get enrolled outside from the catchment area of OnkoNetwork has recently increased. However, it was also noted that patients probably cannot link their fairly positive healthcare experience to the OnkoNetwork model, as fortunately they usually do not have relevant personal benchmarks in cancer care. Current monitoring of patient experience is restricted to the collection of anonymous patient satisfaction surveys that are evaluated at the department level and can’t be linked with the patients’ OnkoNetwork enrolment status.

No economic analyses of OnkoNetwork have been conducted so far. From the perspective of the Centre, no significant economic impact is expected. The current running and operating costs of OnkoNetwork are limited to the human resources cost of about 4 full-time employees and the office leader in the OnkoNetwork Office. In general, no change is expected in the number of diagnostic visits: the number of outpatient visits per patient is expected to decrease, but this decrease is believed to be compensated for by slightly higher patient turnover. In contrast, significant economic benefits are expected to occur at the level of the national healthcare payer and the society due to a shift of the cancer stage distribution towards less advanced stages, enabling earlier and therefore less expensive treatment and better clinical outcomes.

OnkoNetwork is a regional initiative without a specific macro-level payment scheme / incentive system around it, except for a prior financial support from EU funds for the development of OncoLogistic [21]. All running and operating costs of the model are funded from regular inpatient and outpatient financing sources received from the national healthcare payer. The National Healthcare Services Center (ÁEEK), which is a governmental institute in charge of state ownership rights at the majority of hospitals in Hungary, is aware of the programme and has recently rewarded it [22]. However, no additional central funds were provided for the maintenance or development of OnkoNetwork. It is unclear whether financial incentives for the management of patient pathways will be established in the Hungarian healthcare financing system in the future.

Although the intellectual property rights of OnkoNetwork belong to the General Director of the Hospital, she considers it as a public good to improve the Hungarian healthcare system. Hence, it is available for extension to other regions subject to a financial contribution to investments in the initial and further development of the system.

OnkoNetwork administrators of the departments and supervisor physicians typically took up the new roles on top of their existing responsibilities and tasks without significant incremental financial compensation for their additional workload. Instead, staff members reported other motivating factors, including 1) participation in an innovative and patient-oriented model, 2) working in well-designed hospital buildings with modern infrastructure, such as dedicated workstations with personal computers, 3) open communication at project meetings, 4) positive behaviour of the supervisor physicians, 5) informal social events e.g. birthday parties, and 6) the perceived improvement in individual patient pathways. Coordinators of pathways expressed that they considered the monitored patients almost as their family members and achieving optimal patient care provided personal satisfaction to themselves. Supervisor physicians were motivated by the importance of this pioneering initiative in the Hungarian healthcare setting, and by the immaterial appreciation from the colleagues and programme management. The management of OnkoNetwork acknowledges that healthcare administrators have a particularly low salary in Hungary by law, and definitely more compensation would be needed for them. In general, salaries in the Hungarian healthcare system are thought to be disappointingly low.