1.1 Study aims and objectives

It is not well understood the overall changes that MDTs have had to make in response to the COVID-19 pandemic, nor the impact that such changes, in addition to the other challenges faced by MDTs, have had on decision-making, communication, or participation in the context of their MDT meetings specifically. Furthermore, with the pandemic in relative remission, it is time cancer services established which changes brought about because of the COVID-19 pandemic will or should be retained as beneficial adaptations in the future (post-pandemic), and where services need to return to a pre-pandemic status quo.

This study starts to address these questions in the context of MDT meetings specifically. The aim of the study was to explore qualitatively (objective 1) and quantitatively (objective 2) the views of MDT members on the impact of COVID-19 on their MDT meetings, while eliciting the responses on any related changes that emerged concerning (1) access to MDT meetings, (2) meeting organisation and logistics, (3) patient representation at the meetings, and (4) quality of case reviews at the meetings.

2.1 Study design

This was a prospective cross-sectional survey study.

2.2 Study setting

The study took place in the United Kingdom between September 2020 and August 2021.

2.3 Participants

Survey participants were recruited using a snowball sampling method, using established networks with participants encouraged to share the invitation to participate through the NHS England and NHS Improvement's communication channels such as Cancer Alliances, specialty organisations, and within institutions.²⁵ The invitation was aimed at MDT clinical leads, and MDT coordinators. These members were chosen as both have roles with an overview of how their team had adapted to COVID-19, and what technology/infrastructure had been used. They were sent an introductory email (See File S1) with the web address of the survey and were encouraged to circulate the details to other MDT members. The study was registered as service evaluation or audit in each participating site. Consent was implied upon return of the survey, and response to the survey was voluntary.

2.4 Survey design

Our survey consisted of two complementary parts (See File S2). The first part was qualitative, eliciting free-text responses to four questions focused on the impact of COVID-19 on the MDT meetings specifically, including favourable changes to MDT meeting to be maintained going forward, and changes that highlight more pressing issues limiting MDT working (in the context of MDT meetings). The second part of our survey was quantitative, with 10 questions across four domains including: access to the MDT meeting (two questions), MDT meeting organisation and logistics (three questions), patient representation at MDT meetings (two questions), and case discussions at MDT meetings (three questions). Responses to the survey questions were provided on a sliding scale marker ranging from 0 to 100. Scores below 50 represented improvements from before COVID-19 in the four survey domains. Scores above 50 represented deteriorations from before COVID-19. A score of 50, that is, mid-point on the sliding scale marker, represented no change from before COVID-19. Further description of the individual variables and their scales can be found in the Additional File 2. This mixed-methods approach increases the richness and relevance of our findings regarding MDT workings and highlights improvement opportunities.

2.5 Qualitative and quantitative analyses

To address objective 1, we used NVivo 12 to analyse qualitative responses inductively and deductively using a thematic approach. We used a systematic classification process of coding and identifying themes/patterns in the data with the main themes and sub-themes formed/shaped inductively by the data.^{26, 27} Two assessors analysed the data and reviewed the themes (TS and BWL).

To address objective 2, we used SPSS IBM SPSS Statistics 24 where we examined group differences in survey responses by comparing descriptive statistics using boxplots, and by inferential statistics using t-test. The nature of sampling (snowball) precluded calculation of a formal response rate, as the denominator (i.e. the number of cancer professionals who received the invite to participate) remained unknown throughout the study.

3.1 Participant characteristics

The participants were MDT members (N = 423) from England (n = 401), Scotland (n = 16), Wales (n = 4), and Northern Ireland (n = 2). The members represented different professional groups: Allied Health Professionals (AHP; n = 10), Cancer Nurse Specialists (CNS; n = 45), Coordinating, Administrative and Management (MDT-CAM; n = 105), Oncologists (n = 67), Pathologist (n = 7), Physician (n = 36), Radiologist (n = 12), and Surgeons (n = 135). The average years of work experience in these roles was 7 with a minimum of 1 and maximum of 10 years (n = 82 of 1-2 years; n = 118 of 3-7 years; n = 202 of >8 years). A Kruskal–Wallis test revealed no statistically significant differences in the survey responses depending on the seniority levels (p > 0.05 on all survey questions). The participants covered 8 different cancers: Breast (n = 45), Colorectal (n = 25), Gynaecological (n = 103), Head and Neck (n = 23), Lung (n = 26), Skin (n = 22), Upper GI (n = 21), Urology (n = 102) and Other (n = 39). There were also no statistically significant differences in the responses between different cancer specialties (p > 0.05 on all survey questions). Majority of participants were from more specialist MDTs (SMDTs) hosted largely at University or Teaching Hospitals (n = 291), and somewhat smaller number was from local-MDTs (LMDTs) hosted largely at District Hospitals (n = 129); there were no statistical differences between the two (p > 0.05 on all survey questions).

3.2 Objective 1: Thematic analysis of free-text survey responses

The main themes for each question are presented in Table 1 alongside their frequency, example quotes, and their representation across the participating professional groups. We found hybrid working and possibility of virtual attendance was the most frequent theme in response to what changes to MDT working will you maintain. IT-related issues and slower meetings, longer lists and delays were the most common theme in response to the question on whether changes have caused any specific problems. Virtual meetings and increased attendance/availability of clinicians was the most common response to the question concerning changes that provided solutions. Improving IT infrastructure, its use and access were cited as the most common change that now seems more pressing than before.

3.3 Objective 2: Group comparisons of structured survey responses

Figure 1 shows boxplots with the median survey responses and interquartile ranges (IQR) for the 10 questions (scored on the 0–100 rating scale) grouped into four (colour-coded) domains.

Staff responses were relatively less variable (smaller IQR) on some questions, indicative of more consensus on the perception of ‘no change from before COVID-19’ (median rating = 50). Specifically, staff consensus was clearer on access to MDT meeting (domain 1; q1 and 2), patient representation at MDT meetings (domain 3; q6 and 7), and case discussion at MDT meetings (domain 4; q8 and 9). On the other hand, despite a general perception of no change, staff responses were more variable (larger IQR) on MDT meeting organisation and logistics (domain 2; q3 and 4), and case discussion at MDT meetings (domain 4; q10). Amidst less consensus around the general perception of no change, we observed a wider range of staff ratings indicative of perceived improvement in MDT meeting organisation and logistics (domain 2; q3 and 4) since before COVID-19. In contrast, we observed a wider range of staff ratings indicative of perceived worsening in case discussion at MDT meetings (domain 4; q10) from before COVID-19 (Figure 1).

In line with Figure 1, t-tests, performed on the composite domain scores, showed significantly lower scores, that is, improvement from before COVID-19, for MDT meeting organisation and logistics (domain 2; M = 45, SD = 20) compared to the other three domains, that is, access to MDT meetings (domain 1; M = 50, SD = 12, t(390) = 5.028, p = 0.001), case discussions at MDT meetings (domain 4; M = 50, SD = 14, t(373) = −5.104, p = 0.001), and patient participation (domain 3; M = 50, SD = 12, t(382) = −4.537, p = 0.001), which showed no change from before COVID-19 (i.e. all Ms = 50, which is the mid-point). Breakdown of descriptive statistics for the composite domain scores including the individual survey items can be found in Table 2.

4.1 Implications and generalizability

Our findings are important to consider in the current climate with multiple pressures on healthcare teams, such as financial pressures,^{14, 15} to staff shortages,¹⁷ increasing cancer incidence,^{14, 16} growing workload,¹⁸ and human factors.^19-24 Given the large estimated cost of MDTs (over £100 million per year in the UK) understanding what impacts its performance and how it can be made safer and more efficient is critical to continuous quality improvement efforts.^{35, 36} Indeed, just prior to the pandemic, NHS England completed a pilot study of the effect of streamlining MDT meetings and published national guidance regarding the implementation of streamlining in MDT meetings in order to enable cancer MDTs in England to adapt to increasing demand. This guidance encourages MDTs to be selective about which cases are discussed in full at the MDT meeting, favouring complex cases with simpler cases being treated according to an agreed standard of care. The results of the present study suggest that the ability to discuss cases that benefit most from MDT review, that is, complex patients, might have improved with the changes to remote access and online meetings. Importantly, our qualitative findings point to an increasing volume and complexity of oncology patients that need to be addressed further supporting a more streamlined approach. The present study also suggests that IT infrastructure will be critical to support these changes and facilitate further improvement with innovations such as streamlining.

Taken together, these findings suggest the changes that have been necessary to continue service during the COVID-19 pandemic might have the unintended but desirable effect of facilitating a streamlined agenda. Failure of healthcare organisations to provide the necessary IT infrastructure could compromise the ability of MDTs to function normally during the COVID-19 period, let alone to make the necessary steps to successfully improve the effectiveness and efficiency of MDT working. In particular, the MDTs went through several agile reorganisations during COVID-19 to keep diagnostic and treatment pathways functioning. As a result of this, the governance function of MDTs may have taken on more importance to ensure excellent patient oversight and tracking was still maintained when other healthcare processes were more fluid, and resources diverted towards COVID-19. This requirement for a higher level of governance may have made MDTs less likely to introduce other changes the teams themselves wanted to prioritise, given the unusual circumstances of a world pandemic raging around them. But going forward, teams need to be supported to address areas important to them such as streamlining. There are various teamworking tools^{36, 37} available to support this change but the involvement of hospital management is paramount early on, to facilitate this.

Replication and assessment of generalizability of the findings to other cancers and health systems needs to be examined to determine the extent to which they apply to them. Indeed, other national healthcare systems such as the Veterans Health Administration in the United States, have explored the impact of COVID-19 on virtual oncology care, including tumour boards, finding both opportunities and challenges corroborating our findings to some extent.³⁸ Similarly, Sniderman et al utilised mixed methods to describe impact on paediatric cancer providers during the pandemic, capturing some themes that overlap with our findings; most notably around staffing challenges.³⁹

4.2 Limitations

The results of the study should be interpreted in light of potential limitations. First, this study relies on self-report obtained via a snowball sampling approach with most responses coming from England, UK, which may reduce representativeness of findings and reduce wider generalizability. Snowball sampling is a non-probability sampling method where new participants are recruited by other participants to form part of the sample. Advantages of this method include low-cost and ability to target potential participants unknown to the researchers through existing networks. The number of responses exceeded that which the research team would have been able to achieve through direct recruitment, thus increasing the sample and representation of the population. However, limitations must be acknowledged: as the sample is not chosen through random selection of the population being studied, it may not be entirely representative. Moreover, researchers have little or no control over the sampling process (other than the instructions in the brief sent to potential participants) and rely mainly on referrals from already-identified participants. Since people refer others whom they know (and share traits with), snowballing can have a potential selection bias.

Second, the total number of participants is certainly only a fraction of the thousands of healthcare professionals working in the thousands of MDTs in the United Kingdom, and may favour those with a tendency to respond to such a survey (e.g. those with strong opinions). Third, the availability of demographic data on participants is limited. Forth, it should be acknowledged that our subgroup quantitative analyses were not undertaken according to whether the respondents were from district hospital and university/teaching hospitals—this is because we did not find statistical significance between the two groups (as reported in the Results/Participant Characteristics above). We cannot, therefore, comment on any possible quantitative differences in the results according to the situation of respondents. Lastly, the study aimed to explore perceptions of changes caused by the pandemic relating specifically to the MDT meetings, and not to the comprehensive working of MDTs in patient care as whole. Although MDT meetings represent one important segment of the care pathway, there are many other areas of MDT working that could have been impacted by COVID-19. These limitations, when taken together, could limit the generalizability of the findings.

4.3 Further research

Further research is needed to validate current MDT quality improvement tools and approaches^{36, 38} on the virtual/hybrid MDT meetings to enable continuous measurement and improvement going forward to maximise value for all stakeholders, including patients. For example, the indication from the current study that the ability to participate in the meeting case discussion may have gotten worse since COVID-19 and the introduction of virtual/hybrid MDT meetings should be followed up using, for example, observational methodology and currently available tools for assessing the focus and quality of MDT patient reviews.²⁷ In particular, given that virtual MDT meetings can improve attendance,^{29, 30} further research should explore, using the current tools, the quality of contribution to discussion between face-to-face and virtual (or hybrid) meetings. Future research should also further explore the quality of decision-making and communication in MDT meetings pre and post COVID-19 pandemic. This could be achieved using a retrospective-prospective approach with the post-pandemic data collected prospectively and benchmarking it against the existing literature, or to the pre-pandemic MDT datasets available of data repositories (e.g. decision-making, interactions and complexity across three cancer multidisciplinary teams, Zenodo).⁴⁰ Lastly, the impact of COVID-19 may have been different according to the situation of the Hospital, and whether they were a specialist-MDT, at the hub, or a local-MDT, at the spoke end of the organisation—although we did not find statistical significance between the two groups, further research could nonetheless investigate this hypothesis in more detail.