Work-as-imagined model in the Follow Me programme
In the analysis of work-as-imagined in the Follow Me programme, 18 functions were identified, of which 8 were selected as foreground functions (Figure 2).
Data collection and standardisation
Standardisations of data measurement instruments were considered to be Preconditions for evaluating outcomes (function 5 in Figure 2) and for conducting research (function 6 in Figure 2). In turn they were to enable creation of a knowledge base on disease, treatment and follow-up outcomes. The functions that generated data collection and their aspects are shown in Additional file 4.
Allocation of key stakeholder roles and responsibilities
The key stakeholder in the work-as-imagined model was the patient, surrounded by a multidisciplinary team providing standardised follow-up care and personalised care. Patient organisations were to have a Control, or monitoring, aspect in maintaining a patient-oriented approach. Academic centres had a Control aspect in benchmarking treatment outcomes and were seen as learning partners in these complex, often rare, patient groups. A dedicated project team for the Follow Me programme initiated outcome evaluation sessions.
Building control and feedback loops
The functions (5) to evaluate outcomes, (6) to conduct research, (7) to standardise treatment protocols and (8) to standardise the follow-up programme were to be used to build control and feedback loops designed to improve treatment and follow-up protocols. Preconditions within feedback loops were the systematic collection of standardised PROMs and the performance of standardised measurements and data collection on medical, developmental and psychosocial outcomes by the multidisciplinary team during the follow-up. To establish and continuously improve follow-up protocols, the standardisation process of the follow-up programme was to make use of research output, evaluations of follow-up outcomes, the experience of the multidisciplinary team, and the cooperative relationships with patient organisations and with other academic centres. Such interrelations formed a cycle of collection of long-term outcomes, and these were to feed into the insights that would provide feedback on treatment and follow-up protocols.
Work-as-done at the neonatal ICU, Location AMC
In the analysis for the work-as-done model of the AMC neonatal ICU, 12 functions were identified, of which 10 were selected to be foreground functions (Figure 3).
Data collection and standardisation
Standardisation of data measurement instruments, including standardised data collection time points, was based on national guidelines in place since the 1990s (12). The functions that generated data collection and their aspects are shown in Additional file 4.
Allocation of key stakeholder roles and responsibilities
A patient defined by the national guidelines as belonging to an at-risk group was offered the standardised follow-up programme in two parts, one with the neonatologist and one with the psychologist. There was an alignment between the neonatologist and the psychologist, with their functions interlinked by consecutive time slots and patient handover. A patient could receive additional or adjusted care, referral or guidance from both the neonatologist and the psychologist individually.
Building control and feedback loops
In this model, small feedback loops were seen that monitored previous functions or Preconditions for downstream functions. In addition, the face-to-face handover between the neonatologist and the psychologist incorporated a Control aspect to make sure a consent form for data use for research purposes had been signed and collected (a Precondition for conducting research). Standardised regular internal evaluation sessions on the outcomes of follow-ups did not take place at the AMC neonatal ICU at the time of this research, and were therefore not modelled.
Work-as-done at the neonatal ICU, Location VUMC
The analysis for the work-as-done model of the neonatal ICU, Location VUMC, included 17 functions, of which 13 were defined as foreground functions (Figure 4).
Data collection and standardisation
Standardisation of data measurement instruments, including standardised data collection time points, was based on national guidelines in place since the 1990s (12). All the health care providers involved prepared the follow-up separately, but were aided by a summary of the patient’s history constructed by a medical psychology trainee with the help of a template. That summary was only available to all the involved health care providers if cognitive development was also assessed (not done in all age groups). The functions that generated data collection and their aspects are shown in Additional file 4.
Allocation of key stakeholder roles and responsibilities
Patients defined by the national guidelines as belonging to an at-risk group were offered the standardised follow-up care involving four consultations. The first was with the psychologist (if the patient was in the age group offered assessment of cognitive development) in a separate but nearby outpatient area. Subsequent sessions were planned with the medical assistant, the neonatologist and the physiotherapist in another outpatient area. All consultations were scheduled consecutively, thus minimising patient waiting time. The follow-up care was coordinated by a designated medical assistant with a host function towards the patient. The assistant welcomed the patient and provided guidance on which questions to pose to which health care provider. In case the next planned provider was not ready to conduct the follow-up when the patient was available, the assistant tried to switch around consultations, thus reducing delays for the patient and providers alike. The assistant also provided brief patient handovers in person to the health care providers, highlighting certain concerns the patient might have and giving an update on consultations so far. The role and responsibilities of the host function were confined to the follow-up programme within the department. In our observations, patients who had received treatment in two different academic neonatal ICUs during hospital stays would now be attending two similar follow-up programmes, one in each academic centre, as both centres were compliant with national guidelines. There was a stronger interlinkage between the neonatologist and the physiotherapy disciplines; those functions were partially performed simultaneously in the same room. This facilitated communication between the two professions and allowed for the parents of the patient to be questioned by the intensivist, while the physiotherapist observed the infant or child in terms of motor development.
Building control and feedback loops
A Control function was created that reminded the patient of the follow-up and the PROMs questionnaire to be completed two weeks beforehand. The prepared summary and the preparation functions for each follow-up function involving health care professionals (functions 6, 9, 10 and 11) served as Controls for the data availability needed for the follow-up programme. Individual results from the professionals’ follow-ups were discussed in a multidisciplinary meeting straight after the follow-ups to align thoughts on future treatment. Standardised regular internal evaluation sessions on the outcomes of follow-ups were not being conducted in the VUMC neonatal ICU department at the time of this research, and were therefore not modelled.
Work-as-done in the paediatric ICU follow-up
The analysis for the paediatric ICU’s work-as-done model included 11 functions, of which 8 were defined as foreground functions (Figure 5).
Data collection and standardisation
Standardisation of the follow-up programme was based on the national guidelines for follow-up after paediatric intensive care (13). Patients had problems with delivering the patient-reported outcomes in time for the follow-up programme; in one of our observations, four out of four patient families had failed to meet that goal. That resulted in fewer Resources for the psychologist to prepare for the follow-up and could lead to inadequate Resources or bias for outcome evaluation and research. In our observations, patients reported receiving the invitation to the online portal for the PROM questionnaires on paper, whereas they preferred digital invites to access the questionnaires directly by clicking on a link whilst at the computer. Another observed problem was confusion in a patient’s parents about which questionnaire was needed for the current follow-up; they referred to other evaluation questionnaires received after the hospitalisation and a questionnaire from a children’s support service outside the scope of the department. All data on medical indicators were collected using standardised templates in the patient EHR. The functions that generated data collection and their aspects are shown in Additional file 4.
Allocation of key stakeholder roles and responsibilities
The modelled work-as-done started during the discharge procedure from the paediatric intensive care unit. With the help of a decision tree, a decision was made by a paediatrician-in-training, supervised by a paediatric intensivist. In addition to patients included in line with the decision tree, other patients might be included in the follow-up programme by exception on advice of clinicians and nurses who had treated the patient. If a positive decision was made, the patient was offered the standardised follow-up in three parts, first seeing the medical assistant of the outpatient clinic, then the paediatric intensivist, and then the psychologist. If necessary, a paediatric pulmonologist, cardiologist, neurologist or neuropsychologist could be included in the follow-up, but such consultations were not needed for the patients we observed. Key stakeholders were the paediatric intensive care specialist and the psychologist. The interlinkage between the functions of these stakeholders was reinforced by close proximity, consecutive time slots and face-to-face patient handovers. In our observations, however, a face-to-face handover was not always possible due to misalignments of available time caused by delays in consultations.
Building control and feedback loops
Both the paediatric intensivist and the psychologist prepared for the follow-up in order to ensure Preconditions for the physical follow-up with the patient. No Time aspect interlinked these functions, however; hence, if necessary information, such as PROM results, was found to be missing, there was no opportunity to rectify that in time for the follow-up consultations. One patient in our observations had undergone a lung function test the week before the follow-up, but the test results from the pulmonologist were scheduled for after the follow-up clinic visit. This suggests possible misalignment between the follow-up programme and routine medical check-ups. The observations detected no Control function to check alignment with routine medical follow-up procedures. The Follow Me programme in the paediatric ICU convened a yearly improvement session, attended by all the participating health care providers, patient organisations and the departmental management team, to evaluate outcomes, protocols and organisational indicators.
Work-as-done model in the paediatric surgery follow-up
The analysis for the work-as-done model in the paediatric surgery department’s outpatient clinic identified 15 functions, of which 12 were defined as foreground functions (Figure 6).
Data collection and standardisation
The time spent between patient and health care providers was largely being used for data collection purposes. One family reported during our observations that it was not clear to them whether PROMs about parental psychological outcomes were available to specific health care providers; as the PROMs included personal information on psychosocial factors involving the parents, they found it odd not to know who was aware of that information. PROMs on psychosocial factors were in fact reviewed by the psychologist only, who drew on them to provide input in multidisciplinary meetings after the outpatient visit. Other health care providers did have access to the PROMs before the outpatient care began, but, in our observations, use was not made of that access. The functions that generated data collection and their aspects are shown in Additional file 4.
Allocating key stakeholder roles and responsibilities
The standardised follow-up of the paediatric surgery department could consist of up to seven parts (Figure 6, functions 3 to 9), depending on the patient’s age, disease or treatment. The multidisciplinary meeting between health care professionals that took place after the physical follow-up (not in the patient’s presence) was the medium used to exchange professional viewpoints. Interlinkages between professionals whilst a patient was present at the outpatient clinic were few. One interlinkage was observed between the clinical nurse specialist and the paediatric surgeon via brief patient handovers, or when that nurse was physically present while the paediatric surgeon was in consultation with the patient. The order in which professionals saw a patient was not standardised, but was determined by the most efficient schedule. Thus, health care providers involved in the follow-up would see their patients at different stages of the follow-up (they might be the first, second, third or last to see the patient). What previous functions had been performed might be obvious from a patient’s perspective, but complicated to oversee for each individually involved health care provider, even though a schedule for the day was accessible to all involved. In the post-surgery work-as-done, individual care needs of patients and their families were addressed in fragmented ways, with each care provider discussing important medical observations for their field independently. If an aberration in one function occurred, it was dealt with mostly within the discipline in question.
Building control and feedback loops
Few Control aspects were observed in the surgery work-as-done model. Health care providers were guided by standardised templates in a patient’s electronic health record (EHR). The Follow Me programme of the paediatric surgery department convened a yearly improvement session, attended by all the participating health care providers, patient organisations and the departmental management team, to evaluate outcomes, protocols and organisational indicators.
Gaps between work-as-imagined and work-as-done models; learning across case studies
In all four work-as-done models, the implemented data collection was aided by templates in the patient EHRs (such as medical assessments) or by standard measurement instruments (such as PROM questionnaires or Bayley Scales of Infant Development), using multiple entry points, EHR modules and software programs. Varying by case studies, the choices regarding data collection on follow-up care mirrored the provision of ‘standard’ care, with the follow-up model of the paediatric surgery department being more specialised and more fragmented, whereas that of the paediatric ICU was more encompassing and less fragmented.
In all four case studies, data were collected and evaluated on medical, developmental and psychosocial outcomes; however, data collection and reporting on the effectiveness of the follow-up was limited. Indicators were missing, for example, to show the number of patients seen for follow-up as a proportion of those that should have been seen according to national guidelines or internal protocols. Also missing were indicators providing insights into changes made to treatment plans due to the follow-up programme. Process indicators on the follow-up programme in the studied departments were limited to the number of no-shows (highlighting efficiency in planning and success in informing patients about the importance of the programme); they were not available on the effectiveness of the follow-up programme.
In practice, implementation of a multidisciplinary team approach came in many forms, including (1) a predetermined consecutive order of consultation for the care providers involved (e.g. AMC neonatal ICU and paediatric ICU); (2) access to multiple interdisciplinary data collection points by an individual provider (e.g. paediatric ICU, neonatologist); (3) team-up by providers for data collection (e.g. physiotherapist and neonatologist together in one consultation at the VUMC neonatal ICU); (4) face-to-face patient handovers by paired care providers (neonatologist and psychologist at the paediatric ICU); or (5) a hosting function created to facilitate patient handovers (e.g. VUMC neonatal ICU). Responsibilities for coordinating the follow-up were allocated in different ways in the multidisciplinary teams, for example via the medical assistant (VUMC neonatal ICU), via the lead clinician (AMC neonatal ICU and paediatric ICU) or via the clinical nurse specialist (paediatric surgery).
Yearly evaluation sessions to improve treatment and follow-up, and multiple small control loops to check whether Preconditions were being met, were present in most work-as-done models. However, poor timeliness of control and feedback loops and the lack of process indicators limited the ability of programmes to make adjustments when discrepancies were found. In practice, control and feedback loops operated solely within the scope of the follow-up programmes. This resulted in unplanned interactions with ‘routine care’ outside the department or organisation in almost all work-as-done models (VUMC neonatal ICU, paediatric ICU, paediatric surgery).