Improving Health Care for Patients with Multimorbidity: A Mixed-Methods Study to Explore the Feasibility and Process of Aligning Scheduled Outpatient Appointments through Collaboration between Medical Specialties

Setting

The study was conducted at Silkeborg Regional Hospital in Denmark, a development hospital specialising in adaptation and creation of integrated patient pathways. It has a catchment area of 93,000 inhabitants. Through tax financing, the Danish healthcare system provides free universal access to public healthcare, including hospitals, general practice, and selected partners in primary and secondary care [3940].

Study design

This was a feasibility study with a process evaluation of a complex intervention tailored for multimorbid patients seen in the different outpatient medical clinics. The data was collected from 15 August 2018 to 14 August 2019 using convergent mixed methods. The study was guided by the frameworks of the Medical Research Council (MRC) on the development and evaluation and on process evaluations [4142]. For reporting, the TIDieR checklist was followed [4344].

Development

Available evidence was reviewed on targeted care management for patients with multimorbidity [22293238454647]. To facilitate broad support and involvement in designing a care model, a co-creating participatory approach was applied [484950]. Available details in Appendix 1 and 2 [4551].

Intervention theory

The MRC framework recommends a clear description of the intervention, and how it is expected to work [42]. Logic modelling was used to synthesise and describe the complexity to help clarify causal assumptions [45] (Figure 1). For theoretical basis for reorganising health services, we were inspired by existing evidence, consulted stakeholders (Appendix 1), and drew elements of Wagner’s Chronic Care Model and the SELFIE framework.

Figure 1

The intervention

The MOP consisted of three elements (Figure 2). First, nurse care managers (experienced with coordinating care) received a monthly list of forthcoming patients and their medical appointments with nurses or physicians in outpatient clinics. Participants were identified proactively, as registered in the electronic health records [52]. Outpatient appointments were consecutively scheduled to take place on a certain date. The booking was arranged with the individual outpatient clinics. Schedules for care professionals were coordinated and time slots were reserved in appointment calendars and released fourteen days before the date of reservation, if not used. Laboratory personnel were made aware of the need to coordinate testing. Second, the delivering specialty wrote a summary with care-related information (rapidly passed) to the receiving clinic. Third, afterwards, the involved physicians and nurses attended a conference, resulting in a joint care plan with oral feedback and notice of modifications to the patient and a written summary to the GP. This was formed as a supplementary summary listing any modifications of the individual specialty summaries.

Figure 2

As new candidate appointments for alignment entered the monthly lists, previously ineligible appointments could re-enter. The intervention was launched as optional to patients. The program was refined progressively with weekly evaluations and adjustments during the first months.

Participants and place

Data were extracted on outpatient appointments scheduled less than six weeks apart; this was done continuously every month. These data lists constituted the basis for potential candidates. We excluded appointments related to other research projects, new referrals for diagnosing purposes, and procedures conducted without a subsequent care appointment. Adult patients with two or more chronic conditions (duration for at least six months) with scheduled appointments in two or more medical outpatient clinics were included. The intervention was restricted to five specialties: cardiology, pulmonology, rheumatology, nephrology, and endocrinology. The selection was based on highest outpatient activity across specialities while considering the relevance for integration [153]. Existing outpatient clinics were used, and the intervention took place each week on the same two weekdays.

Contextual factors

were considered, acting as barriers or facilitators to the implementation. The outpatient clinics had different ways of arranging appointments in the hospital’s booking system, and arrangements were made by different secretaries and nurses. Appointments were booked months in advance, leaving little opportunity to negotiate rescheduling. The alignment was challenged by; subspecialisation, patients were often followed by the same healthcare professionals, and a national strategy [54] mandated the allocation of a primarily responsible physician. The intervention needed to accommodate these factors. Outpatient clinics were located in proximity to each other, which facilitated to movement between clinics. At this hospital, staff were experienced in reorganising patient pathways.

Patient characteristics and PROMS at baseline

Data on age, gender, civil status, distance to hospital from home address, diagnoses, and outpatient trajectories were retrieved through the Patient Administrative System (PAS). The Danish Multimorbidity Index [55], based on the International Classification of Diseases, 10^th revision, was used to estimate the frequency of diagnoses. Patients were sent a questionnaire before entering the MOP. This included the Short-Form Health Survey 12-item (SF-12) for assessment of health-related quality of life. It uses two weighted scales (mental and physical functioning) ranging from 0–100 (low-high)[56]. Care services were assessed through the 20-item Patient Assessment of Chronic Illness Care questionnaire (PACIC), which considers whether patients receive patient-centred care. Response categories range from 1–5 (never to always) aggregated into an overall score and five subscales: patient activation, delivery system design/decision support, goal setting/tailoring, problem solving/contextual and coordination/follow-up. Higher scores indicate high quality of care [57]. Also, the 10-item Multimorbidity Treatment Burden Questionnaire (MTBQ), which consists of five response categories; presented as a global score grouped into four categories, scores ranging from 0 (no burden) to ≥ 22 (high burden)[58] (Table 1).

Quantitative implementation assessment

Reach was assessed through participation. Scheduled appointments were monitored by the care managers, who facilitated sequential appointments in agreement with the patient and clinics. From PAS, we retrieved dates for blood sampling in the fourteen days leading up to the appointment date to assess if samples had been collected.

Fidelity: The care managers documented when and why aligning appointments was practicable or not, if patients declined, cancellations, and if the conference was conducted. The time interval between appointments, (from end of dictation summary to start of next appointment), included PAS data on transcript times from physicians in all specialties and nurses in the endocrinology outpatient clinics. PAS data also included when a joint summery was transferred to the GP.

Qualitative assessment: Experiences with the MOP

Acceptability and fidelity: Individual interviews were conducted with ten patients and ten healthcare professionals on their experiences of the intervention, including uncertain aspects of program design and delivery [4549]. Respondents were recruited at 2–4 days after entering the intervention in chronological order of presentation. A semi-structured interview guide was used; one for patients and another for healthcare professionals [59]. Details were elicited by open questions and at the end of each interview, the interviewer summarised key features and allowed the respondent to elaborate and validate. Interviews were audio-recorded and transcribed verbatim. Using Malterud’s systematic text condensation strategy, CB and CWA analysed the transcripts separately for healthcare professionals and patients [60]. This includes four steps of analytic process: 1) total impression – from chaos to themes, 2) identifying and sorting meaning units – from themes to codes, 3) condensation – from code to meaning, and 4) synthesising – from condensation to descriptions and concepts.

Study population

In the first intervention year, 102 patients received the intervention (Table 1). They were aged 71 years (median), 46% were females, 54% lived with a partner, and the median distance to the hospital was 7.7 kilometres (4.8 miles). The patients had a median of five diagnoses, with hypertension and diabetes mellitus being the most frequent. The patients scored an average of 32 points on the physical scale and of 42 points on the mental health of the SF-12. The quality of care was assessed to an average of 2.9 on the PACIC questionnaire. About 43% of the patients reported treatment burden to be medium/high. The vast majority (94%) had 2–3 trajectories in outpatient clinics.

Reach

From the care managers’ lists for potential alignment, 2027 unique appointments (100%) were assessed. As the dataset comprised many lists of appointments retrieved every month, some appointments reappeared in newer lists as new appointments created new combinations for alignment. Therefore, 2367 appointments (116.8%) were evaluated in total. The intervention reached 102 patients, who comprised a 20% inclusion, and 309 appointments (15.2%) were coordinated. In median, patients entered the MOP once (interquartile range (IQR); 1: 4) and had two appointments coordinated (IQR; 2:4). Reasons when alignment was not possible are shown in Table 2. Blood sampling was requested 142 times by different specialties in 80 patients. The number of times that a patient was sampled for blood was reduced by 29.1% to 101 blood samplings. Two requisitions from different specialties were taken simultaneously in 33 cases, and three requisitions were taken simultaneously in four cases.

Fidelity

Often, lack of flexibility in other appointments was the reason why coordination was not possible. When it was possible to coordinate appointments, few patients declined entering the pathway (3.1%). There were few cancellations (4.9%). The intervention was predominantly conducted according to the protocol (Table 2), which was supported by interviews. However, the dictation summaries were largely unavailable for the receiving outpatient clinic before consultation start. The median waiting time between appointments was 35 minutes (IQR 19:54) for the patients. The full intervention was held 83.2% of the time, with a joint summary transferred to general practice in 95.4% of cases.

Experiences through interviews

Respondents’ characteristics are reported in Table 3. The interview duration was in median 26 minutes (IQR 21:35) and 30 minutes (IQR 26:40) for healthcare professionals and patients, respectively.

Main findings

This study describes the formation and implementation of a tailored intervention to patients with multimorbidity with concurrent trajectories in outpatient clinics. The MOP includes patients based on their frequent attendance in outpatient clinics as identified by electronic records. Care coordination was challenged by context, although this hospital was focused on creation of integrated patient pathways. Navigating existing delivery systems, booking schemes, subspecialisation, and accommodating intentions to have patients followed by the same physician or nurse made care coordination challenging. During the first year, 102 patients underwent the MOP, corresponding to a reach of 20% of eligible patients. They were elderly patients, who at baseline assessed that follow-up/coordination could be improved, their physical well-being was low [56], and 43% reported medium/high treatment burden. Age and the combinations of chronic conditions resembled earlier intervention studies targeted patients with multimorbidity [3361]. A total of 15% of all appointments were aligned and blood samples were reduced by 29%. In median, the patients used the intervention one time. The intervention was implemented as intended and conceptual elements were foremost followed accordingly. However, the handover of information in the transition between specialties lacked delivery. Although knowledge sharing already existed, healthcare professionals found a formal collaboration was welcomed but the patient selection required refinement. Patient respondents resembled the study population. They found the MOP endurable and appreciated having appointments and tests aligned. Furthermore, that multidisciplinary collaboration and regular visits to outpatient clinics provide a sense of security.

Existing studies

Collaboration on care decisions is considered important for integration of patient care and includes sharing the responsibilities of problem solving and decision making [62]. Evidence on interventions aiming at ensuring multidisciplinary collaboration in a systematic manner, is lacking despite demands for integrated care solutions. Previous studies have no resemblance to the MOP [23293233366364656667], although programs have been identified in the ICARE4EU project [38]. An example is a holistic approach taken in Denmark. To support GPs, a multidisciplinary outpatient team offers a review of medications, care plan, and follow-up recommendations. Patients are selected on medical complexity, interacting diseases, and drug regimes [61], whereas the MOP includes patients based on their frequent attendance in outpatient clinics.

Concerns about care for patients with multimorbidity include disease-centred rather than patient-centred care, lack of attention to comorbidities, patient preferences and needs, and limited care coordination [233568]. Reuben and Tinetti advocated for a wider span of conditions and more focused alignment of treatments toward common goals [69]. Likewise, Bayliss et al. have argued that an ideal process of care is also individualised to support the patient’s unique constellations of problems [28]. Moreover, Boye et al. have stated that healthcare systems need to create flexible systems for multimorbid patients [21]. The MOP accommodates many of these factors, although the approach to patient preferences and priorities are similar to the approach in usual care.

Strengths and limitations

This is the first study to integrate outpatient appointments through intensified interdisciplinary collaboration on patients with multimorbidity. The study was performed in accordance with the well-established methods and prevailing guidelines, which ensured a systematic approach and transparency. Launching the intervention with five different outpatient clinics implied that different booking methods and work schedules could be equally managed, which gave a realistic idea of how a full-scale intervention will work. Data are presented on fidelity, reach, and acceptability and using mix-methods leads to a deeper understand of the intervention and is ideal to discover uncertainties in the intervention design and delivery [4270]. The integration of methods was used when assessing fidelity.

Some limitations should be addressed. Missing data was a limitation. The time consumption among the care managers was difficult to document, we estimated it took one workday going through each monthly list of appointments. The same limitation applied to the treatment changes following MOP, which according to the interviews, seldom happened. Coordination was restricted to two weekdays. Additionally, CB being both interviewer and hospital employee may have affected responses and the analysis.

Perspectives and implications

Despite similarities, patients with multimorbidity form a heterogeneous population with varying needs because conditions vary widely in their manifestations and treatment indications. Patient characteristics (treatment burden, physical well-being, and follow-up/coordination) underpin that these were patients with challenges. Despite this, both delivery system design/decision support and patient activation were assessed as good, which could point to that patient-centred care was already being delivered. Interviews revealed that the patient selection needs refinement because the conference discussions often did not have sufficient content or led to medical changes. This questions whether the intervention targets those most in need of multidisciplinary collaboration. However, alignment of hospital visits may seem relevant for some e.g., patients with mental disparities and patients who are reliant on help from caregivers. Complexity could be considered in the MOP, including condition severity, higher symptom burden, and polypharmacy to improve acceptability [71]. However, this might increase the work burden in the care model. Reach met our expectations, as appointments within the same speciality will have a time interval. Additionally, as intervals between appointments are not the same across specialties because they follow specific disease course programs, this limits how many appointments can be aligned and impact of the intervention.

The elements of the model are feasible and practicalities of organising the intervention are manageable and can be adapted, but this will require time, effort, and awareness of care coordination, patient information, and for healthcare professionals to engage. Self-booking and alignment of appointments by patients as well as adding voice-recognition for dictation summaries, might be considered, to ease hospital workload.

However, the organisational challenge of aligning appointments and setting up a multidisciplinary collaboration, as done in the MOP, must be addressed and efforts should be seen in relation to effectiveness. Our process-evaluation found system- and patient-level benefits gained by the MOP i.e., collaboration and knowledge-sharing across specialties, reduction of blood samples, acceptance of having appointments aligned, and a sense of collaboration among one’s healthcare professionals. Whether the MOP was effective in optimising treatment, led to reduced outpatient activity and fewer trajectories, or improved the patient-experience of care services and treatment burden, calls for investigation. Nonetheless, this study demonstrates the complexity of modifying patient trajectories to fit the existing organisational practices to integrate care for patients with multimorbidity.