This study was a randomized controlled trial (RCT) titled the Partnership Project [10]. The protocol, details of this study, and findings on primary outcomes (eg, single-item global assessment of intersectoral cooperation) have been published previously [5,10,11]. This study presents secondary outcomes on patient perceptions of care coordination using the Australian Cancer Care Coordination Questionnaire (CCCQ) [12] and satisfaction with received information using the European Organisation for Research and Treatment of Cancer Quality of Life Information Questionnaire 25 (EORTC QLQ-INFO25) [13] from a 7-month follow-up survey in a shared video-based consultation.

Patients diagnosed with any type of cancer and starting their first chemotherapy treatment at the Department of Oncology, Lillebælt Hospital, University Hospital of Southern Denmark, were considered newly diagnosed and invited to participate in this study. The eligibility criteria included (1) being aged >18 years, (2) being proficient in speaking and reading Danish, and (3) having an oncologist’s estimate of a survival time exceeding 7 months.

This trial was concluded upon reaching the predetermined sample size for patient inclusion.

The sample size was based on estimates from a previous Danish RCT aiming to improve GP involvement in cancer follow-up. To detect a clinically meaningful difference with 90% power and a 5% significance level, 194 participants were required. Considering an anticipated dropout rate of 30%, the recruitment goal was set at 278 patients. Detailed information on sample size has been published previously [5,11].

Patients in the intervention group received a shared video consultation involving their GP, an oncologist, and themselves, in addition to usual care. These consultations were scheduled within 12 weeks of inclusion and could be held at either the GP’s or oncologist’s office based on patient preference. The oncologist typically chaired the session, supported by an oncology nurse, and both clinicians received a consultation guide beforehand.

The control group received usual care, which consisted of standard communication between the department of oncology and primary care, including electronic summary letters sent to the GP after each oncology visit, optional phone contact, and patient access to their GP.

A detailed description of the intervention and control groups, consultation structure, and implementation logistics has been published previously [5].

Patients were assigned in a 1:1 ratio through block randomization, with block sizes and sequences managed by REDCap (Research Electronic Data Capture; Vanderbilt University) [14] data manager. A project nurse conducted the randomization and patient enrollment following informed consent. Allocation was transparent to patients, GPs, and oncologists, but blinding was maintained during baseline data collection. Data analysts remained blinded throughout. GPs of control group patients were not formally informed until they received the survey. Detailed information can be found in our previous publication [5].

Patients were asked to complete questionnaires at baseline and at 4 and 7 months. Upon arrival at the department of oncology, they received information about the study, their perceived role, and the possibility of leaving the study as well as a consent form and a paper-based baseline questionnaire, which was collected by an outpatient nurse after enrollment. Follow-up questionnaires were distributed electronically via REDCap [14]. An overview of the secondary outcomes in this study is provided in Multimedia Appendix 1. The outcomes included items and subscales from CCCQ [12] (22 items) and EORTC QLQ-INFO25 [13] (25 items).

For this study, the CCCQ was translated from English to Danish following the guidelines for cross-cultural adaptation of self-report measures proposed by Beaton et al [15]. The translation procedures were documented in separate reports using the template developed by the American Academy of Orthopaedic Surgeons and recommended by Beaton et al [15].

The CCCQ evaluates patients’ perceptions of cancer care coordination over a recall period of 3 months [12]. The CCCQ consists of 22 items: 2 global items, including global rating of coordination of care (item 21) and global rating of the quality of the received care (item 22) and 20 items that form 2 subscales, including the communication subscale (1-13; ranging from 13 to 65) and navigation subscale (items 14-20; ranging from 7 to 35). Responses are provided on 5-point Likert scales, ranging from “never” (1) to “always” (5), except for the global items, which use a 10-point Likert scale, ranging from “very poor” (1) to “very good” (10).

EORTC QLQ-INFO25, consisting of 25 items, assesses various aspects of patient satisfaction with the information provided during cancer treatment. Although no specific recall period was defined, patients were asked to consider their current cancer trajectory. The questionnaire includes a total score, 4 subscales covering information about the disease (items 1-4), medical tests (items 5-7), treatment (items 8-13) and other services (items 14-17) as well as 8 single items (items 18-25) addressing information about different care locations, self-help resources, written material, audiovisual formats, satisfaction with the information received, the need for more or less information, and the overall helpfulness of the information. In total, 4 additional yes-or-no questions explored the formats and quantity of information received by participants. Specifically, they asked whether participants had received written materials or audiovisual content (eg, CD, tape, or video), whether they wished to receive information (with an open-ended follow-up on preferred topics), and whether they felt they had received too much information (also with an option to specify topics). Responses to the core items were given on a 4-point Likert scale, ranging from “not at all” (1) to “very much” (4). All scores were linearly transformed to a scale ranging from 0 to 100 according to the scoring manual [13]. For the 4 items (written information, item 20; information on CD tape/video, item 21; wish to receive more information, item 23; and wish to receive less information, item 24) that have dichotomous response options (yes=1 or no=0), scores were treated as binary indicators.

Subscale score calculations were performed as follows:

The direction of the answer scale varied by item. For the CCCQ Navigation subscale, a low score indicated a positive attitude toward the question, whereas for all other CCCQ subscales and single items, a high score indicated a positive attitude. For analysis purposes, all scores were aligned so that higher values indicated a positive attitude toward the questions. Specifically, item scores for items 14 to 20 were reversed coded to ensure consistent interpretation, such that an original score of 5 was recoded to 1, 4 to 2, 3 remained unchanged, 2 to 4, and 1 to 5.

Outcomes were measured at baseline and at 4 and 7 months after baseline. Coding was performed separately for each time point.

Demographic data for patients, including age, gender, education, marital status, having children, employment status, comorbidities, and cancer diagnosis or type, were collected via a questionnaire completed by patients at baseline.

A deviation from the initial analysis strategy was due to substantial missing data across all outcome variables at the 7-month follow-up (refer to the sample size reported for each outcome variable in the Results section). Detailed information on the original statistical analyses can be found in the published protocol [10] and our previous publication [5]. The revised statistical analysis is described subsequently.

For each secondary outcome, we compared the change from baseline to 7 months between the 2 groups using a linear regression model, following an intention-to-treat principle. The model, applied to measurements at both baseline and 7 months, used generalized estimating equations to account for within-patient clusters, with robust variance estimation. The group difference was modeled as a time-by-group interaction. No additional covariates were included in this analysis. We assumed that missing data were missing completely at random, meaning that the probability of missingness was unrelated to unmeasured factors after accounting for treatment group and baseline outcome. The generalized estimating equation approach ensured robustness in the presence of missing data. In addition, we calculated the individual changes from baseline to follow-up, restricted to complete cases, and presented Cohen d as an effect size measure and P values based on a 2-sample 2-tailed t test (Multimedia Appendix 2).

Data analyses were performed using Stata (version 18; StataCorp) [16], with a significance level set at 5% (P<.05).

Patients were enrolled in this study between June 2016 and November 2019. A total of 278 patients were randomly assigned to the intervention (n=139, 50%) and control (n=139, 50%) groups. However, due to GP-related issues (n=22, 15.8%), administrative (n=8, 5.8%) and technical issues (n=15, 10.8%), and clinical (n=3, 2.2%) or patient-related issues (n=8, 5.8%), only 80 (57.6%) patients received the intervention as intended. Figure 1 shows the participation flowchart.

Table 1 presents the baseline characteristics of patients in the intervention and control groups. Patients in both groups exhibited similar baseline characteristics; however, comorbidities were more prevalent in the control group (n=81, 58.3%) compared to the intervention group (n=65, 46.8%).

Multimedia Appendix 3 provides an overview of the missing data regarding patient care cooperation (CCCQ subscales) and information outcomes (EORTC QLQ-INFO25) at various time points for both the control and intervention groups. While the number and proportion of missing data vary across each variable, a consistent pattern was observed across all variables, where missing data were minimal at baseline but increased substantially at the 4- and 7-month follow-ups.

Table 2 presents patients’ perceptions of care coordination and satisfaction with the information received between the primary sector and the department of oncology. No statistically significant differences were observed in the changes over time between the intervention and control groups in any outcome.

The estimated within-group changes between baseline and 7-month follow-up for the total CCCQ score were –0.84 (95% CI –3.27 to 1.59; P=.498) in the control group and 0.26 (95% CI –2.15 to 2.67; P=.83) in the intervention group. The between-group difference was estimated as 1.11 (95% CI –2.32 to 4.53; P=.53). The estimated within-group changes between baseline and 7-month follow-up for the overall INFO25 score were –3.53 (95% CI –6.34 to –0.72; P=.01) in the control group and –2.04 (95% CI –5.52 to 1.44; P=.25) in the intervention group. The between-group difference was estimated as 1.49 (95% CI –2.98 to 5.96; P=.52). The estimated difference in the change from baseline to 7 months for the CCCQ communication subscale was –1.49 (95% CI –1.33 to 4.31; P=.30) and for the navigation subscale was –0.03 (95% CI –1.52 to 1.46; P=.97).

This study found that the addition of a cross-sectoral video consultation to usual care did not lead to statistically significant improvements in patients’ perceptions of care coordination and satisfaction with the information received over 7 months. There were no significant differences in changes over time between the control and intervention groups for any of the secondary outcomes.

Telemedicine, including video consultation and phone calls, has emerged as a crucial tool for maintaining the continuity of medical care for patients with various medical conditions during the COVID-19 pandemic [17]. However, our study on the innovative health care model was conducted before the COVID-19 pandemic and the implementation of standard video setups. Since then, the technical aspects of video-based communication in health care have significantly improved, making it challenging to compare our findings with post–COVID-19 pandemic studies that use these advanced protocols.

Despite this advancement, recent research highlights that the impact of telemedicine on patient outcomes remains nuanced [18,19]. For instance, a meta-analysis on the efficacy of telemedicine for outpatients found that video consultation was feasible but did not significantly outperform face-to-face care in oncology in terms of patient satisfaction and attendance [18]. Authors in an editorial with a conceptual perspective lens argued that while remote consultation may enhance access, they risk compromising rational continuity, information richness, and shared decision-making [19]. This underscores the importance of context-sensitive consultation approaches, guided by clinical judgment and patient preference rather than policy mandates [19]. We believe that while these 2 studies focus on different dimensions, both highlight limitations of telemedicine and reinforce the need for thoughtful implementation.

A systematic review that analyzed 51 studies across multiple medical disciplines, including primary care, oncology, dermatology, and so on, revealed that telemedicine provided valuable support to traditional medicine during the COVID-19 pandemic [17]. The review highlighted a high level of patient satisfaction with telemedicine, particularly in areas such as convenience, continuity of care, communication, and efficiency [17]. In contrast, our study, which focused on patients’ perception of care coordination and satisfaction with the information received, found no significant improvement in these areas after the follow-up period. This discrepancy might be due to the different variables and methodologies used in each study as well as several limitations in our study, such as many missing values at follow-up, the lack of standard protocols for video consultation due to the timing of the trial, and other implementation challenges, including technical issues or scheduling difficulties, that were comprehensively explained in our previous publication [5].

The other study comparing the content and quality of various modes of consultation (eg, video, telephone, and face-to-face) found that face-to-face consultations enabled more information exchange between GPs and patients compared to video and telephone consultations [20]. The observed difference between our study and the aforementioned study highlights the potential limitations of video consultation in achieving patient outcomes and emphasizes the importance of face-to-face interaction in the health care setting. However, due to several limitations in our study that have been discussed in our previous publication [5], we believe that comparisons with other studies and the generalizability of our findings should be approached with caution. Specifically, the high rates of missing data in our study hinder us from drawing concrete conclusions about the intervention effect on patients’ perception of care coordination and satisfaction with the information received. Instead, we encourage the research community to learn from our trial’s challenges [5] and insights to improve future studies on this research area.

First, a key limitation of this study concerns the use of the CCCQ questionnaire, which was translated from English into Danish for this project. Although the translation followed the established guidelines for cross-cultural adaptation of self-report measures [15], the Danish version had not been validated before this study. This may impact the reliability and cultural appropriateness of the instrument in the Danish context. Second, the considerable amount of missing data observed at the 7-month follow-up reduced statistical power, affecting the ability to detect significant effects. Third, another limitation is the likely ceiling effect observed in both the CCCQ and EORTC QLQ-INFO25. At baseline, scores in both groups were high, leaving limited room for measurable improvement, potentially masking the intervention’s impact. Implementation challenges further limited this study. These included low completion rates of video consultations, technical and administrative issues, and the absence of a standardized video setup during the trial—details previously reported [5]. In addition, the absence of structured fidelity metrics, such as session counts, duration, and quality, limits the interpretability of intervention delivery. Although descriptive data on session uptake were collected, logistical challenges and a large amount of missing data prevented meaningful statistical analyses. Finally, the absence of statistically significant findings likely reflects a combination of factors, including limited intervention uptake, variability in participant engagement, measurement limitations (eg, the use of translated but unvalidated questionnaires), and inconsistent implementation. These findings highlight the need for future studies to use validated outcome measures and incorporate robust implementation and fidelity-tracking strategies.

From the clinical and implementation perspective, the findings suggest that when baseline care is already perceived as high quality, additional interventions may offer limited incremental value. Future efforts to improve care coordination and patient information may be more appropriately targeted toward settings or populations where baseline performance is suboptimal, as there may be more potential for meaningful improvement in these contexts. In addition, future studies should incorporate structured fidelity metrics (eg, session completion rate) to better understand intervention delivery and identify implementation barriers. This would support more adaptation to clinical settings and enhance scalability within established frameworks, such as the Consolidated Framework for Implementation Research and the reach, effectiveness, adoption, implementation, and maintenance framework, which help guide and evaluate the adaptation, effectiveness, and sustainability of health interventions [21-23]. Furthermore, redesigning the intervention to address technical and administrative challenges, such as streamlining, scheduling, integrating automated reminders, and improving digital infrastructure, could enhance feasibility and uptake. Co-designing future iterations with end users may further improve usability and engagement, supporting more effective implementation and transferability to other settings.