Continuity of care experienced by patients diagnosed with pancreatic cancer: nationwide multicentre study

As pancreatic surgery in the Netherlands has been centralized, patients with pancreatic cancer can undergo diagnostic work-up and treatment across multiple hospitals within a pancreatic cancer network. Patients may undergo surgery in the pancreatic expert centre and receive adjuvant chemotherapy in a non-pancreatic expert centre. This fragmentation of care may affect continuity of care (COC). COC refers to the degree to which a series of healthcare events is experienced as coherent, connected, and relevant to a patient’s needs and context1. COC is particularly important given its association with increased patient satisfaction2. Two Delphi studies on patient-reported outcome measures (PROMs) in pancreatic cancer care indicated that satisfaction with care and care organization should be included in a core PROM set3,4. Thus, COC could be considered an important PROM for quality of care. Yet, studies describing patient experiences of health services, such as COC, remain scarce, despite their potential to inform quality improvement. The aim of this study was to assess COC experienced by patients with pancreatic cancer in the Netherlands. Secondary aims were to identify factors associated with COC, examine the relationship between COC and quality of life, and to assess variations between regional networks. This was a cross-sectional analysis of a prospective multicentre cohort study, the Dutch Pancreatic Cancer Project (PACAP-PROMS). Patients with pancreatic cancer included in PACAP-PROMS were asked to complete a validated questionnaire on COC (Nijmegen Continuity of Care questionnaire5 (NCQ); Likert scale 1–5) between February and April 2022. The NCQ measures COC experienced in the setting of one clinician (personal continuity), between clinicians in one centre (within hospital), and clinicians in multiple centres (between hospitals). Additional data from PACAP-PROMS were scores from the EORTC Quality of Life Questionnaire (QLQ-C30) and the EORTC QLQ-PAN26, for pancreatic ductal adenocarcinoma. Corresponding patient data, including age, sex, performance status, histopathological diagnosis (tumour characteristics such as tumour stage, location), and the type, timing, and location of primary treatment (expert centre versus non-expert centre), were derived from the Netherlands Cancer Registry. The results for the primary outcome, the COC experienced, are presented descriptively. Secondary outcomes (associations between NCQ scores and patient or treatment-related characteristics) were analysed using multivariable linear regression analysis (supplementary methods). In all, 418 patients with pancreatic cancer from all 15 Dutch pancreatic cancer networks responded to the questionnaire (response rate 55%) (Fig. S1). Most patients first visited a non-expert centre (119, 71%) and underwent multicentre care (299, 72%). The mean(standard deviation) NCQ scores were 3.61(0.75) for personal continuity, 3.84(0.76) for within hospital continuity, and 3.66(0.86) for between-hospital continuity (Table 1). Patients who received multicentre care had significantly lower mean NCQ scores within the network compared with a single centre (3.66 versus 3.83; 95% confidence interval (c.i.) 0.071 to 0.248; P < 0.001). Systemic therapy in an expert centre was associated with higher NCQ scores (Coefficient 0.394; 95% c.i. 0.070 to 0.719; P = 0.018). Lower healthcare satisfaction scores were significantly associated with lower NCQ scores (P = 0.016). Personal continuity of care and cross-boundary continuity of care, expressed as NCQ scores (Likert scale 1–5; n = 418) Values are n (%) unless otherwise indicated. *Excluding 14 patients who indicated the general practitioner was the most important doctor. †Patients were able to answer the question with ‘I do not know’, indicated by the ‘?’. ‡Patients were asked whether they were referred to a different hospital during their patient journey; only those who answered ‘yes’ could answer B2. NCQ, Nijmegen continuity of care questionnaire; s.d., standard deviation. There were no statistically significant differences in NCQ scores between networks and networks as a factor were not associated with NCQ scores (Tables S1–S4). This study provides important insights into how patients with pancreatic cancer experience COC in pancreatic cancer networks. The mean scores for COC can be interpreted as moderate. Patients who underwent multicentre network care evaluated COC in the network significantly lower than COC in one hospital. Factors associated with higher COC included systemic therapy in an expert centre and greater healthcare satisfaction. The results of this study should be interpreted in light of some limitations, such as the under-representation of patients who were solely treated in a non-expert centre. The findings of this study highlight the need to improve care coordination, especially in view of the quality of care paradox, namely that further centralization of complex cancer care may inadvertently increase fragmentation of care. The authors have no funding to declare. Conceptualization: Hopstaken, Augustinus, HWM van Laarhoven, Stommel; Data curation: Hopstaken, Augustinus, Vissers; Formal analysis: Hopstaken, Vissers; Funding acquisition: n.a.; Investigation: Hopstaken, Vissers; Methodology: Hopstaken, Augustinus, Vissers, HWM van Laarhoven, CJHM van Laarhoven, Stommel; Project administration: Hopstaken, Stommel; Resources: Hopstaken, Vissers; Software: n.a.; Supervision: Hopstaken, Augustinus, HWM van Laarhoven, CJHM van Laarhoven, Stommel; Validation: Hopstaken, Augustinus, de Hingh, Besselink, van Santvoort, van Laarhoven, Bonsing, Festen, de Groot, Homs, Liem, Manusama, de Meijer, Schreinemakers, de Vos-Geelen, Wumkes, van Laarhoven, Vissers, Stommel; Visualization: Hopstaken; Writing—original draft: Hopstaken, Augustinus, Vissers, Stommel; Writing—review & editing: Hopstaken, Augustinus, de Hingh, Besselink, van Santvoort, van Laarhoven, Bonsing, Festen, de Groot, Homs, Liem, Manusama, de Meijer, Schreinemakers, de Vos-Geelen, Wumkes, van Laarhoven, Vissers, Stommel. The authors declare no conflict of interest. Supplementary material is available at BJS Open online. The study data may be made available upon request.