Decision-Making in Multidisciplinary Tumor Boards in Breast Cancer Care - An Observational Study.

PURPOSE: Healthcare providers decide on recommendations for further treatment of patients with cancer in multidisciplinary tumor boards (MTBs). As such, communicative processes during MTBs are assumed to influence decision-making and, thereby, treatment planning. The aim of this exploratory study is to gain insights into decision-making during MTBs.
METHODS: Case discussions from MTBs in breast cancer centers in North Rhine-Westphalia, Germany, were observed and audiotaped. The transcripts of the audio recordings were analyzed by procedures of conversation analysis.
RESULTS: Based on 38 case discussions from 15 MTBs in four breast cancer centers, an action scheme for decision-making in MTBs in breast cancer care was developed. In addition, the conversational practices used by the participants during interactions were analyzed.
CONCLUSION: Analysis indicated that conventions in MTBs were observed during individual phases of decision-making but not for the entire process. Although exchanging multidisciplinary knowledge is an essential aspect of MTBs, this exchange does not always seem to occur in practice. The extent to which recommendations are derived from consensus during MTBs remains unclear. Thus, the study suggests developing standards for communication during MTBs to optimize decision-making and, thus, the quality of recommendation.

Introduction

Multidisciplinary Tumor Boards (MTBs) are regular meetings of a multidisciplinary treatment team, in which the diagnosis and treatment of patients with cancer are discussed and evidence-based recommendations are decided upon.1,2 Professionals from various disciplines are involved in cancer care due to the increased complexity of diagnosis and treatment.3 Thus, MTBs are established as a standard worldwide.4 Previous studies demonstrate that a multidisciplinary approach to treatment for breast cancer, where different treatment modalities are coordinated in a timely manner,5 leads to better results for patients and healthcare providers.6–12 Positive effects of treatment decision-making can be observed,9,13 indicating that patients are more frequently diagnosed and treated on the basis of relevant evidence-based guidelines.14

Multidisciplinary reasoning is an essential aspect of MTBs. In the medical context, the term refers to joint decision-making among different disciplines.15 Previous studies find that, in 2–52% of cases, team decisions established within MTBs relatively differ from individual decisions made by healthcare providers in terms of diagnosis and therapy.16–19

Communication is an essential factor during MTBs as a prerequisite to decision-making,20 which is a process that precedes the mere statement of a decision and is traceable by means of a transcript.21 The expected outcome of decision-making during MTBs is a treatment recommendation, which exerts a major impact on further therapeutic procedures. The recommendation influences the subsequent steps of the treatment process on the one hand and the quality of life8 of patients and further course of the disease on the other hand.

A previous study in cancer care using conversation analysis revealed that case discussions during MTBs can generally be categorized into the following phases through analysis of the transcript: opening, case presentation, provision of additional information, discussion, articulation of treatment plan, pre-closure, and closure.22 In another study, also using conversation analysis, 4% of case discussions in MTBs in breast cancer care directly started with the discussion, and in MTBs in gynecological cancer care, in 4% of case discussions the decision was not articulated.23

Several factors indicate that poor communication during MTBs leads to an insufficient exchange of multidisciplinary knowledge.24 Research indicates that certain circumstances surrounding and within MTBs can impede open communication and decision-making, such as the setting itself or the manner of presenting treatment options to patients. In a study conducted in Australia, approximately one-third of the interviewed physicians did not rate the MTB setting as conducive for discussion, especially if the prevailing opinion was evident.25 Similarly, a study in Germany observed several case discussions and reported that only one or a few high-ranking physicians contributed to such discussions.26 Further research demonstrates that if healthcare providers fail to agree on a treatment recommendation, then available options are less frequently presented in a neutral manner, thus resulting in the framing effect.27 As such, patients find difficulty in deciding against a physician’s preferred option.28

The majority of existing evidence related to MTBs are derived using quantitative methods29 or qualitative interviews that focus on the perspective of healthcare providers. To gain a deeper insight into the decision-making of MTBs, the study employs procedures of conversation analysis, which is a valuable method for analyzing verbal communicative practices used by individuals in naturally occurring interactions.30 Furthermore, the current study aims to analyze the conversational practices of MTBs during case discussions for decision-making.

The study poses the following research questions: 1) How do MTB members produce decisions for treatment recommendations through communication? 2) What insights can be gained regarding the decision-making of MTBs using procedures of conversation analysis?

Methods

Conversation analysis is based on ethnomethodology and focuses on ordinary interactions.31 Such an approach can reveal existing patterns and conversation analysis can lead to new insights into medical interaction and communication.30

Using conversation analysis, other authors created an action scheme for MTBs.22 Thus, we initially proceeded in a deductive manner and examined the extent to which this action scheme fits our sample, then adapted it inductively. In contrast to the origin of the method, the study focused on practices during the individual phases instead of conversation sequences30,32,33 throughout the process. The reason for this approach is that the study aimed to investigate typical conversational practices used during such individual phases.

Study Design and Sample

Data were collected from a multicenter, non-interventional study entitled ”Patient involvement in multidisciplinary tumor conferences in breast cancer care” (PINTU34), which was conducted on six breast and gynecological cancer centers in North Rhine-Westphalia, Germany’s most populous state. The study mainly aimed to answer the following research questions: 1) How do MTBs with and without patient participation differ in terms of organization, interaction, and patient orientation? 2) How do patients experience participation and what are the immediate effects of participation? 3) How do healthcare providers evaluate patient participation in MTBs in terms of feasibility and quality of decision-making? The study was funded by the German Cancer Aid (grant number 70112286), and the Ethics Committee of the University of Cologne gave a favorable vote (reference number: 17–405). The study observed and audiotaped case discussions during MTBs in certified breast cancer centers in North Rhine-Westphalia. The research team consisted of four researchers, who signed a confidentiality agreement with the personnel administration. The medical staff in the centers recruited patients for the study, except for one center, where the researchers recruited the patients. The inclusion criteria were (a) at least 18 years old, (b) diagnosed with breast cancer according to the International Statistical Classification of Diseases and Related Health Problems,35 and (c) undergoing treatment in one of the participating centers. In addition, German language skills were required to ensure understanding of the written informed consent. All MTB members and patients, whose case discussions were audiotaped, provided written informed consent. The case discussions constitute a PINTU subsample without patient participation.

Data Collection

Two researchers at a time carried out participant observations and audiotaped conversations during MTBs. In addition, field notes were taken, including the duration of case discussions and number of discussed cases. The author transcribed audio data using transcription conventions based on the talk-in-interaction system,36 which integrates the conventions of Jefferson.37 The transcripts were anonymized. Essential aspects related to the research questions were noted in the transcription.

Data Analysis

After developing a description of decision-making based on parts of the first material, it was applied to the entire data set. First, the structure of overall conversations in the transcripts was analyzed. An action scheme was deductively developed using the phases of MTBs proposed by Dew et al.22 On the basis of the material, we inductively adapted this action scheme. Second, micro-analysis was performed by examining conversational practices. Typical patterns of conversational practices observed across cases as well as deviations from such patterns were documented. Two researchers separately conducted the process. For validation, identified activities and conversational practices were discussed until a consensus was reached through regular data sessions among three scientific assistants. Additionally, the entire research team held consultations to embed the results in the context of health services research.

Results

The study presents a five-phase action scheme for decision-making during MTBs conducted in the treatment of patients with breast cancer followed by a description of the conversational practices used by healthcare providers during their interactions in various phases.

The current analysis is based on 33 new case discussions and five re-discussions, among 15 MTBs from four breast cancer centers recorded in January and February 2019. Two of the four breast cancer centers were teaching hospitals. In one center, MTBs were held at a round table, whereas healthcare providers in the three other centers sat in a theater-style seating arrangement. The recorded cases differed in complexity according to clinical stages partly because only non-metastatic cases were discussed within any one MTB in one breast cancer center, whereas non-metastatic and metastatic cases were discussed in the same MTBs in the other centers. Large variations were observed in the numbers of discussed cases for each MTB (range: 9–29; median: 16), healthcare providers (range: 7–23), and speakers involved in discussions (range: 1–8; median: 3). Total audio data lasted for 58 min and 43 s, where individual case discussions varied widely (range: 00:25–05:42 min; median: 1:08 min). The patients in the sample consisted of n=33 women, ages ranging from 29 to 83 years. Table 1 provides a description of the sample.

Table 1

Description of the Sample

Breast Cancer Centers (n = 4)
3 non-metastatic and metastatic cases	1 only non-metastatic cases
2 teaching hospitals	2 non-teaching hospitals
3 theater-style arrangements	1 round table
MTBs (n = 15)(38 Case Discussions)
33 new case discussions	5 re-discussions
Descriptive Results of Case Discussions in MTBs
	Range (Median)
Number of cases discussed for each MTB	9–29 (16)
Number of participating healthcare providers	7–23
Number of speakers involved	1–8 (3)
Duration of individual case discussions	00:25–05:42 min (1:08, Σ 58:43 min)

Action Scheme

Figure 1 illustrates the developed action scheme of the tasks of healthcare providers for decision-making during MTBs.

Figure 1

Action scheme. Colors symbolize variations in decision-making during case discussions.

This action scheme varies by case. Colors are used to represent variations observed from available material. Each case discussion began with an announcement of the case followed by the presentation of a patient’s age and a report on current diagnostic and therapeutic statuses. At least one of two optional sequences followed. In the first, missing information was supplied (ie, diagnostic values), which was important for further decisions. In the second, subsequent procedures were planned. Decision-making was deemed completed by reaching agreement. In several case discussions, other topics, such as technical equipment, were discussed. These topics were noted in the margin because they interrupted the decision-making process.

Conversational Practices

A description of the conversational practices used by healthcare providers during interaction in the various phases of decision-making can be found below. The “supplementation of missing information” is an optional phase, which rarely occurred in the observed case discussions, such that no patterns of conversational practices were found. The presentation of the transcripts is explained in the following text and in Figure 2 to facilitate understanding. Lines are numbered (first column), whereas numbers in brackets refer to transcript number. In the excerpts, time is indicated only if relevant. Speakers are numbered (ie, S1, S2 under column 2) to enable matching of statements. In the excerpts, the bold font represents aspects highlighted in the interpretation.

Figure 2

Example of presentation of transcripts (assignment).

The presented analyses of extracts are illustrative of the analyses throughout the dataset. Table 2 shows the transcription conventions employed.36

Table 2

Selected Transcription Conventions based on Selting et al36

Sequential Structure
[]	Overlap and simultaneous talk
[]
Other Segmental Conventions
Uh, uhm	Hesitation markers, so-called “filled pauses”
Continuers
Hm yes no yeah	Monosyllabic tokens
Hm_hm	Bi-syllabic tokens
Accentuation
SYLlable	Focus accent
Other Conventions
(xxx), (xxx xxx)	Unintelligible syllables and passages

Announcement of the Case

In many observed case discussions, the start of the case was first announced. Frequently, the announcement used sequencers, such as “then” (740), “next”, or “first”, in conjunction with the patient’s name:

01 S2: all right then next we will do ms patient name.

(740)

The majority of case discussions presented an agendum, ie, a predetermined sequence from which deviations occurred for various reasons. In the next excerpt, the order of cases was changed, whereas the announcement as the first step is missing. The patient name is announced relatively quietly, followed by the question of whether S2 would like to introduce the case (line 01). The presenting physician then begins with the presentation of the patient. After approximately 90 s of discussion, another physician asks for the name of the discussed patient (line 52):

{00:00}01 S1: <