Background
Differences in clinical presentation of interstitial lung diseases (ILD) can be subtle, yet their natural history and response to therapy may display striking disparity [
1,
2]. For example, idiopathic pulmonary fibrosis (IPF), a progressive and fatal subtype of idiopathic interstitial pneumonia (IIP), displays clinical, radiological and histological features that have considerable overlap with other subtypes. While IPF historically possesses a significantly worse prognosis [
3], anti-fibrotic therapies are now available that delay its progression [
4,
5], making its accurate identification critical. With the use of a multidisciplinary meeting (MDM), the differentiation of IPF from other diagnoses occurs with improved interobserver agreement along with improved recognition of rare diseases [
6]. As a result of this, recent consensus statements regarding IPF and ILD have advocated that a multi-disciplinary approach be used for their diagnosis [
2,
7,
8].
By definition, the MDM requires input from a variety of specialties, but the most appropriate form of MDM constitution and governance remains unclear. The constitution of ILD MDMs may have important diagnostic and therapeutic implications, Flaherty et al. demonstrating that academic physicians in a multidisciplinary setting display better diagnostic agreement and consider a broader range of diagnoses, compared to community physicians [
9]. The ILD MDM has potential for bias depending on the expertise and number of its attendees, nature of clinical inputs and quality and quantity of data provided. Such biases may have profound effects should they lead to diagnostic error, given that resultant therapeutic decisions may, for example, lead to the use of agents for indications where efficacy is not established, or even for those where they may cause harm [
10]. We sought to determine the constitution and governance of ILD MDMs in expert centres to provide information on current standard of care.
Discussion
Our survey has shown that significant heterogeneity exists in MDM constitution, governance, data input, diagnostic process and information provision. The impact of these differences is unclear, but potentially of crucial importance were it to lead to inaccuracy in disease recognition and resulting therapeutic choices. Also, as MDM diagnosis becomes the clinical benchmark, with access to medications restricted by MDM diagnosis, it is important that we recognise the limitations of the MDM and highlight methods to improve consistency and accuracy.
No randomised trial has ever been performed, or is ever likely to be, that demonstrates MDM based diagnosis results in improved patient survival. Its utility is instead based on the “linked evidence approach” [
11] whereby benefit is derived through improved diagnostic accuracy and consequent alterations in therapeutic approach. It is clear that there is an increasing range of therapeutic choices for ILD, including anti-fibrotic therapy for IPF [
4,
5], antigen avoidance for chronic hypersensitivity pneumonitis [
12,
13], immune suppression for inflammatory and connective tissue disease related disease [
14‐
20], lung transplantation [
21] and palliative approaches [
7]. Many of those therapies have robust evidence for their benefit when applied to specific diagnoses, such as the anti-fibrotic therapies nintedanib and pirfenidone [
4,
5] which are now recommended therapies in the revised IPF consensus statement [
22,
23]. Given the established evidence for MDM’s association with improved diagnostic accuracy [
2,
6,
24], our finding that MDMs are a ubiquitous feature among the expert centres surveyed in our study is not a surprising finding.
Despite the clear utility and importance of ILD MDMs, the constitution and governance of these meetings has never been explored or addressed. Our survey shows that expert centres’ MDM demonstrate several points of consistency in their meetings, being constituted at a minimum by thoracic physicians, radiologists and pathologists - although contribution from non-thoracic clinicians varied - and considering a similar set of clinical and investigative data. However, within this range of broad similarity, significant heterogeneity was seen in MDM governance. This was apparent in the extent of information provided, the presentation style and the approach to resolution of diagnostic dilemmas, which varied between a consensus approach and deferral to the clinician responsible for the case. While the effects of such heterogeneity amongst ILD MDMs are not clear, it could be hypothesised that those with less information presented and less group input with regard to final diagnosis would behave little differently to an individual clinician. Further research that explores the effects of such heterogeneity on MDM diagnostic performance, considers MDM models that provide the most accuracy and concordance, validates the utility of MDM versus other models of diagnosis, and longitudinally explores outcomes, is of paramount importance if MDM is to adequately perform a role as a final arbiter of diagnosis.
Our survey also demonstrated that the majority of MDMs generate outputs in addition to diagnosis, including management recommendations and treatment aims. Interestingly, while multi-craft group meetings focused both on diagnosis and management are the norm in the treatment of other pulmonary conditions such as lung cancer, in only a minority of ILD MDMs did other craft groups attend that might contribute to forming those additional outputs. This is despite the fact that many ILD diagnoses have co-morbidities, and produce significant impairment and mortality, such that the meeting may provide an excellent opportunity for referral to relevant craft groups, such as transplant [
21] and palliative care services [
7]. The utility of non pulmonary craft group attendees was highlighted in our survey by rheumatology attendees, who contribute frequently or always to discussion in 100 % of meetings they attended.
Clearly, local regulatory and insurer factors will have significant impact on the frequency and nature of the use of MDM in the management of ILD, and any position statement would need to be tailored to reflect regional circumstances. For those regions where MDM is used, we propose a set of core criteria for structure and governance, given our survey’s findings with regards to heterogeneity and the implied impact this may have on diagnosis and subsequent therapeutic recommendation. These are likely to include the features listed in Table
4, which were common to the majority of centres surveyed. A number of other features, including the attendance and influence of other specialties, presentation format, and data outputs other than a diagnosis, were not uniform, and thus are difficult to ascribe as core features. However, in developing any expert statement, we suggest such features be considered and guidance given as to which additional features to core criteria are preferred, particularly those that might decrease potential bias and increase the utility of the ILD MDM.
Table 4
Core criteria for interstitial lung disease multidisciplinary meetings
1. An adequate case-load to enable a frequency of meetings commensurate with the development and maintenance of expertise in ILD diagnosis; 2. Attendance by at least one respiratory physician, radiologist and histopathologist; 3. Data presentation by the clinician directly responsible for the patient’s care; 4. Presentation of a set of routine investigations that include high quality HRCT images, PFT, rheumatological serology and, if available, histology; 6. A consensus approach to diagnosis formulation; 7. The provision of a diagnosis, degree of diagnostic confidence, and differential diagnoses. | |
An important area our study did not explore is the prescription patterns for those patients undergoing diagnosis via heterogeneous MDMs. Our study focused solely on expert centres, but it is impractical to imagine that in most regions there are sufficient expert centres present to provide similarly constituted MDM. It is therefore inevitable that heterogeneity will exist in MDM with regards to expertise, based on whether they are located in community or academic centres. Academic MDMs when compared to community centres have previously been demonstrated to possess different diagnostic performance, with an increased likelihood of alternative diagnoses to IPF and therefore by corollary, a probable decreased frequency of prescription of anti-fibrotic therapy [
9]. A further contributor to such differences in prescription rates may be via regional differences in approaches taken by regulatory authorities to anti-fibrotic therapy prescription and whether or not the MDM has been specifically mandated to control the prescription of anti-fibrotics. Each of these potentially has fundamental health economic implications, with regards the frequency of prescription of anti-fibrotic therapies.
A perceived limitation of our study is its concentration solely on expert centres. These centres however, were deliberately chosen in the belief they would most likely have a common approach to ILD diagnosis and management. Despite being a small and highly selected sample, the centres demonstrated significant heterogeneity. A wider survey of non-academic centres to determine whether even broader heterogeneity in MDM output occurs would be worthwhile. Also, we did not explore any specific patient outcomes but instead inferred using existing literature that heterogeneity in ILD MDMs would likely result in a significant difference in ILD MDM diagnostic performance [
6,
9,
24‐
28]. This hypothesis requires confirmation in further studies that explore differences in diagnosis and treatment recommendations.
Competing interests
Nil relevant to this publication.
Authors’ contributions
HJ analysed the collected data and drafted the manuscript. TC and YM contributed to the interpretation of results and drafting of the manuscript. KL, GW, PH and DC participated in study design, coordination and data collection. IG conceived and designed the study and helped draft the manuscript. All authors read and approved the final manuscript.