Are multidisciplinary teams in secondary care cost-effective? A systematic review of the literature

Objective To investigate the cost effectiveness of management of patients within the context of a multidisciplinary team (MDT) meeting in cancer and non-cancer teams in secondary care. Design Systematic review. Data sources EMBASE, MEDLINE, NHS EED, CINAHL, EconLit, Cochrane Library, and NHS HMIC. Eligibility criteria for selecting studies Randomised controlled trials (RCTs), cohort, case–control, before and after and cross-sectional study designs including an economic evaluation of management decisions made in any disease in secondary care within the context of an MDT meeting. Data extraction Two independent reviewers extracted data and assessed methodological quality using the Consensus on Health Economic Criteria (CHEC-list). MDTs were defined by evidence of two characteristics: decision making requiring a minimum of two disciplines; and regular meetings to discuss diagnosis, treatment and/or patient management, occurring at a physical location or by teleconferencing. Studies that reported on the costs of administering, preparing for, and attending MDT meetings and/or the subsequent direct medical costs of care, non-medical costs, or indirect costs, and any health outcomes that were relevant to the disease being investigated were included and classified as cancer or non-cancer MDTs. Results Fifteen studies (11 RCTs in non-cancer care, 2 cohort studies in cancer and non-cancer care, and 2 before and after studies in cancer and non cancer care) were identified, all with a high risk of bias. Twelve papers reported the frequency of meetings which varied from daily to three monthly and all reported the number of disciplines included (mean 5, range 2 to 9). The results from all studies showed mixed effects; a high degree of heterogeneity prevented a meta-analysis of findings; and none of the studies reported how the potential savings of MDT working may offset the costs of administering, preparing for, and attending MDT meetings. Conclusions Current evidence is insufficient to determine whether MDT working is cost-effective or not in secondary care. Further studies aimed at understanding the key aspects of MDT working that lead to cost-effective cancer and non-cancer care are required.


Introduction
Multi-disciplinary teams (MDTs) are an integrated approach to healthcare in which medical and allied health professionals consider relevant treatment options and collaboratively develop an individual treatment plan for patients [1,2]. Despite some scepticism about the effectiveness of MDT working in health care [2][3][4], the concept of MDT working is now widely accepted [5] and considered to be good practice in many parts of the world for the successful management of chronic disease [6][7][8]. Teamwork is recommended because of increasing complexity in patient management and subspecialisation of health professionals, and because of changes in working patterns in the health care sector with reduced hours of work [9].
Areas where MDT working is now commonly adopted in secondary care include care of the elderly, heart failure, stroke, mental health, and critical care. The area that has seen the most increase in MDT working, however, is in oncology where MDT working has been widely recommended including in North America [10], Australia [11,12], and continental Europe [13,14]. Indeed it is now mandated as part of the NHS Cancer Plan in the UK [6]. Studies have shown that MDT working in cancer care is associated with decreased time from diagnosis to treatment [15], more accurate pathological staging of disease [16], increased number of patients treated with radical intent [17,18], improved survival [19], and greater patient satisfaction [20]. However, these studies are largely retrospective, and there is a lack of well designed randomised controlled trials and prospective studies in this field [1,[21][22][23]. There is also a need to consider the costs of administering, preparing for, and attending MDT meetings in any evaluation.
MDTs demand considerable organisation, management infrastructure and funding to ensure that the correct personnel are present, that the relevant patient details are collated, and that the pathological and radiological materials are reviewed by the appropriate specialists before and during the meeting [1,24,25]. Indeed, it has been shown that if critical information is absent at the MDT meeting, decisions may need to change following the meeting and this can delay the start of treatment [26]. In the UK enormous investment has been made to ensure the functioning of cancer MDT meetings and it is estimated to cost the National Health Service (NHS) around £50 million a year for preparation and a similar amount for attendance time [23]. Whether this substantial investment is justified, however, is unknown and has not been previously considered. The aim, therefore, of this paper was to systematically summarise economic studies of MDT working in secondary cancer and non-cancer care.

Criteria for considering studies for this review Types of studies
We reviewed economic evaluations of MDT working compared to no MDT in the provision of secondary care which includes hospital-based and community-based care ( Table 1). The types of study designs eligible for inclusion were randomised controlled trials (RCTs), cohort, case-control, before and after, cross-sectional, and decision analysis modelling studies.

Types of participants
We included studies which enrolled people with any disease type, receiving secondary care and we classified these as cancer and non-cancer.

Types of interventions
MDTs were defined by evidence of both of the following characteristics: 1. Team members from a minimum of two disciplines making decisions; 2. Regular team meetings to discuss diagnosis, treatment and/or patient management, occurring at a physical location or by teleconferencing.
Where team members came from different grades within one discipline, this was considered to represent a single discipline (e.g. a consultant surgeon and a specialist surgical registrar were counted as one discipline).
Studies that evaluated MDT decision making at the patient bedside during ward rounds were excluded because we felt that they were substantially different from the type of meeting room based MDT work advocated in recent policy initiatives (e.g. for cancer services) [6,27]. The key differences between the two are that meeting room based MDTs require a coordinator, and include a systematic review of information about the patient without the patients' input or interruptions. If a paper did not state whether the MDT meeting was ward-based or meeting room based, this was inferred from the description of the meeting.

Types of outcome measures
Full economic evaluations (i.e. cost-effectiveness analysis, cost-utility analysis, cost-benefit analysis and costconsequences analysis) [28] were considered for inclusion.

Types of costs
We included studies that reported on the costs of supporting an MDT meeting or the subsequent direct medical costs of care, non-medical costs (e.g. patient travel), or indirect costs (e.g. productivity losses). The costs of supporting an MDT typically include the costs of team members preparing and attending MDT meetings; the latter could include travel expenses, or costs of teleconferencing, as well as administrative support provided by a team coordinator.

Search methods for identification of studies
We searched the databases EMBASE and MEDLINE (both via the Ovid interface), NHS EED, CINAHL, EconLit, Cochrane Library, and NHS HMIC for relevant publications published from the date of inception of each database to December 2011. Grey literature was identified through the NHS HMIC database as well as based on specialised knowledge of one of the authors (JMB). The search strategy consisted of combinations of free text and MeSH terms related to the economic evaluation of an MDT (see Appendix 1). The search included journal articles and grey literature published in the English language. Reference lists from included studies and other relevant publications, including reviews, were manually checked for citations missed by the electronic search.

Data collection and analysis Selection of studies
Initially, one author (KMK) selected reports fulfilling the first inclusion criterion (see Table 1) of this review based on title and abstract. Full articles of possibly relevant studies were retrieved for more detailed evaluation. A final selection of included studies was decided by discussion with the whole team.

Data extraction
A data extraction template, developed using the guidelines provided by the Centre for Reviews and Dissemination [29], was used to extract the following data alongside the critical appraisal of original studies: country of investigation; objective of the study; study intervention and comparator; study design and setting; target population characteristics; sources and quality of clinical data, if applicable; sources and quality of cost data; methods for dealing with uncertainty; and study results. The following data detailing the characteristics of MDT meetings were also extracted: frequency of meetings, number of disciplines contributing, meeting venue, description of leadership style, method of dissemination of decision, and administrative support. Two authors (KMK, SS) independently extracted data. All discrepancies were reviewed and consensus achieved by discussion.
Due to the varied nature of the studies, we provide a narrative summary of the study results for our review rather than a formal meta-analysis of results.

Methodological quality assessment
The methodological quality of the included studies was assessed using the Consensus on Health Economic Criteria (CHEC-list), a 19-item assessment tool [30]. The CHEC-list represents a minimum set of methodological criteria that address aspects of the internal and external validity of individual economic evaluation studies. For each criterion, a study received a 'yes' if it had sufficiently addressed that aspect of the study, and a 'no' if it had not. Two authors (KMK, FEC) independently evaluated the The six CHEC-list criteria that are regarded as most important in assessing risk of bias are appropriate choice of an time horizon, measurement of costs, valuation of costs, measurement of outcomes, valuation of outcomes, and discounting of future costs and outcomes [31]. Therefore, the risk of bias among the included studies would be high if the number of studies that fulfilled these six criteria is low.

Methodological quality assessment
Only four of the nineteen criteria (i.e. research question posed in answerable form, appropriate economic study design, appropriate time horizon, and identification of important and relevant outcomes) were met by all the studies ( Table 3). The choice of an appropriate time horizon is one of the six CHEC-list criteria that are regarded as important in assessing risk of bias [31]. None of the other five criteria were fulfilled by all the studies. In particular, the criteria of appropriate valuation of costs and appropriate discounting of future costs were met by very few studiesfour and two respectively. The reasons for a high number of studies not meeting the criterion of appropriate valuation of costs were either a lack of information on the method of valuation, the reference year, or the use of a top-down approach in the apportionment of costs a . There were four studies in the review that should have discounted future costs and outcomes because their analytic horizon was more than a year. However, only two [42,43] stated that discounting was undertaken in the analysis. Therefore, given that five of the six criteria that had an impact on the assessment of risk of bias were not met by all studies, there is a relatively high risk of bias in the results of the reviewed studies.

Characteristics of MDT meetings
The amount of detail provided about MDT meetings varied. Twelve papers reported the frequency of meetings which varied from daily to once every three months ( Table 4). The average number of disciplines in each MDT was 5 ranging from 2 to 9. Only one study stated the venue of the MDT meeting [44]. Five of the studies gave details of who led the meetings. No information was provided in any study about the amount of administrative support required for an MDT meeting. Finally, 11 reported how the MDT decisions were disseminated.

RCTs assessing cost-effectiveness of MDTs in non-cancer care
There were 11 RCTs that reported full economic evaluations of MDT care (Table 5). Four evaluated geriatric care, and the results were mixed. One RCT found that MDT directed care resulted in lower costs as well as improvement in some outcome measures. This study included two sites, and it was found that for one of the sites, total costs were lower by approximately 15%, and there were fewer complications during hospitalisation (between 3 to 12 fewer for different types of complications) [34]. However, two other RCTs of geriatric care concluded that there was no evidence of differences in costs or outcomes between MDT and no MDT care [39,44]. The fourth RCT in geriatric care showed that MDT care was associated with higher costs because the cost of implementing MDT care b for 6 months offset the reduction in medication costs, but there was no evidence of differences in functional status and survival between the two groups [40]. Similarly, there were inconsistent results from two RCTs of MDT care for patients suffering from heart Before and After 2

Type of economic evaluation
Cost-consequences analysis 12 Cost-utility analysis 2 Cost-benefit analysis 1 Disease type   [32]. However, the other RCT showed that there was no evidence of a real cost difference but quality of life improved when comparing MDT and no-MDT care [38]. Two RCTs evaluated MDT care amongst patients with terminal conditions; the results were mixed. One reported that costs were lower for the MDT group by about 30%, but there was no evidence of any differences in measures of symptom control, survival, and quality of life between the MDT and no MDT groups [36]. In contrast, the other study showed that there was no evidence of a real cost difference but some outcome measures such as dyspnea, anxiety, spiritual well-being, improved when comparing MDT and no-MDT care [41].
The last three RCTs, each of people with a different disease, also showed varied results. One study relating to patients with mental health problems found that MDT care was cost-effective with an incremental cost effectiveness ratio of €1,267/QALY (£1,039/QALY, £1 = €1. 22, rate as at 20 April 2012) [45]. However, a RCT concerning stroke patients concluded that there was no evidence of real cost or outcome measures differences [46]. Furthermore, in a study that investigated the impact of MDT care versus no MDT care among patients diagnosed with rheumatoid arthritis, there was no evidence of differences in outcome measures but costs were higher in the MDT care group by approximately 30% to 50% [43].
Other study designs assessing cost-effectiveness of MDTs in non-cancer care A prospective cohort study based in Sweden which considered the effect of MDT on vocational activity among patients with chronic minor diseases and long-term absence from working life, reported a total discounted cost of 7.2 million SEK (£684,410, £1 = 10.52 SEK, rate as at 20 April 2012), total benefits of 35.1 million SEK (£3.3 million), and hence, a positive cost-benefit ratio of 4.9 [42]. No estimate for cost per person was provided. In addition, a 'before and after' study found that the rate of infection per 1000 patient days after the implementation of MDT antibiotic management care fell by between 25% to 70% and costs decreased by about 30% [33].

Studies assessing cost-effectiveness of MDTs in cancer care
There were two non-randomised studies that reported results of MDTs in cancer care (Table 5). In a retrospective cohort study of MDT versus no MDT care for patients with malignant melanoma, the costs of health care were 33% to 50% lower in patients whose management decisions were made by an MDT [35]. However, there was no evidence of any differences in wound complication and 5-year survival rates. Hagiwara et al. (2011) [37] reported clinical outcomes before and after the introduction of a multidisciplinary nutritional team for patients with haematologic malignancies. It was found that incidences of hepatic complications, hyperglycemia, and central venous catheter infection were lower in the 'after' group than in the 'before' group, and costs fell by about 20%.

Summary of findings
We identified fifteen primary studies on the costeffectiveness of MDT decision making in secondary care, of which only two concerned cancer and the others covered a wide range of disease types. The characteristics of MDT meetings were diverse. The existing evidence was limited and there was a relatively high risk of bias amongst the studies. Whilst both studies relating to cancer care showed that MDTs were associated with lower costs and better outcomes or outcomes comparable to no MDT care, neither study was a RCT. The results from eleven RCTs assessing the cost-effectiveness of MDTs in non-cancer care were very varied. Only three of these showed that MDT working in secondary care had a positive effect on outcomes and costs. However, the degree of this positive impact varied. Moreover, the positive impact of MDT on some costs components might be offset by the full cost of administering, preparing for and attending an MDT meeting, for which no estimates were reported in the included studies. Therefore, based on existing evidence, it is not possible to determine if MDT working is cost-effective or not in secondary cancer and non-cancer care.

Strengths and weaknesses of this systematic review
To our knowledge, we conducted the first systematic review on the cost-effectiveness of MDT working in secondary care, and we explicitly attempted to capture all types of costs, especially the significant costs of supporting MDT working. The limitations of this review need to be borne in mind. Firstly, as with any systematic review, publication bias might be a problem if studies with null findings have not been published. Secondly, this review was not restricted to a particular country or disease type. While this approach is valuable in describing the available evidence from a wide perspective, it restricts the comparability across studies. Given that the studies were diverse, particularly in the type of care offered in the comparison groups, the findings may not be applicable across different healthcare settings where different clinical practices and geographical constraints operate. The operationalising of MDT may also differ in different settings and together with the variability in reporting, there may be studies that have been conceived as MDT, but based on our interpretation, were not. However, we tried to be inclusive in our definition of MDT working so that we did not exclude potentially relevant studies.

Weaknesses of the evidence
We have identified four key weaknesses in the existing literature on the cost-effectiveness of MDT working. First, there was no standard definition or operationalisation of MDT working evident in the literature. Most studies did not provide key details of the MDT meeting venue, leadership (i.e. who leads the meeting and or how are decisions made), or administrative support. Without a common definition or clear description of MDTs, it was difficult to compare results between studies or draw conclusions on the cost-effectiveness of MDT decision making. Second, the full costs of administering, preparing for and attending an MDT meeting were not estimated in the included studies. Given that the MDT meeting is a key feature of MDT working, the omission of the associated costs in any cost-effectiveness analysis of MDT could lead to an underestimation of costs, and hence bias in the results. Third, the methodological quality assessment has demonstrated that there was a relatively high risk of bias amongst the studies. This meant that the reliability and reproducibility of the included studies were compromised, and the ability to determine conclusions from our systematic review was restricted. Fourth, the available evidence was limited so that it was not possible to draw firm conclusions as to whether MDT working in cancer or non-cancer care would lead to cost-effective care or not.

Comparison with other studies
In their review to determine the critical elements of effective team-working in patient care teams, Bosch et al. [47] concluded that current literature provided little insight into the underlying mechanisms. They reported results relating to costs and resource utilisation but not on cost-effectiveness of MDTs. In comparison, our systematic review focused specifically on full economic evaluations of MDT working, and sought to include studies that estimated the costs of supporting MDT working. In another review examining the association between cancer MDTs and survival, the authors also highlighted the lack of a consistent definition of MDT [21]. This directly hampers any investigation of its effectiveness and costs. While none of the included studies in this review had accounted for the full costs of administering, preparing for and attending an MDT meeting, an audit conducted at a large teaching hospital/cancer centre estimated the average annual cost of MDTs in salaries alone was about £1 million [48]. Time spent preparing for MDT meetings could be considerable. A study examining the participation of pathologists and radiologists in MDT meetings found that the former spent about an average of 2.4 hours, and the latter 2 hours in preparation [49]. Although our review did not find evidence that MDT working is cost-effective in cancer care, there appears to be a growing consensus that telemedicine delivered MDT cancer meetings are more cost-effective than 'in-person' meetings [50][51][52][53].

Future research
Based on the findings of our systematic review, we propose that three issues need to be addressed in future research in this area. Firstly, future research should ensure a clear definition of MDT working and detailed reporting of its characteristics. In our systematic review, we found that the description of the two key features of MDT working was so vague that it was difficult to determine whether MDT working had taken place or not. We recommend that studies should provide the characteristics of the MDT meeting as shown in Table 4. Secondly, it is essential to estimate the full costs of administering, preparing for and attending an MDT meeting. Such attempts should take into account two issues. One, all relevant costs should be identified. We recommend that the following costs should be accounted for: time spent by each team member in preparing for the MDT meeting and this should include time spent by team coordinator in providing administrative support to the MDT meeting, time spent on travelling to and from meetings by any team member who is located on a different site than the meeting venue, costs of using teleconferencing if applicable, and time spent by each member at each meeting. Two, these costs should be measured as accurately as possible. The standard methods used in health services research for workflow assessments are work sampling, time efficiency, and time-and-motion. The first two methods are less resource intensive as compared to the third method but there is evidence to show that the time-and-motion technique may produce the most accurate results [54][55][56]. Thirdly, it remains unclear what models of MDT are cost-effective in secondary care. Therefore, future studies using strong study designs should compare the cost-effectiveness of different models or aspects of MDT working c . For this to be performed accurately there is a need for agreement in the clinical community about which outcomes would be core in the evaluation of MDT working. Whilst some clinical outcomes are important, the inclusion of patient-centred outcomes and satisfaction of health professionals as well as assessment of the quality of decision-making and MDT decision-implementation may also be important [2,57,58].

Policy implications
Without further research that addresses the three key issues highlighted in the previous section, it would be hard to make firm policy recommendations. The costeffectiveness of MDT working in secondary care is likely to vary according to many factors, such as cost of supporting MDT working, disease type, treatment selected, and geographical location. The costs of supporting MDT working are directly linked to factors that affect management decisions of MDTs, such as inclusion of time to prepare for MDTs into team-members' job plans, making team and leadership skills training available to team-members, and systematic input from nursing personnel [59]. The commitment of resources to these factors will differ across settings. Therefore, it is likely that there is no 'one-size-fits-all' model of MDT working to suit all diseases, treatment modalities and locations.

Conclusion
The widespread adoption of MDT working in cancer and non-cancer care has occurred despite a lack of evidence on cost-effectiveness. The available evidence is limited and at relatively high risk of bias. It is not possible to determine if MDT working in cancer and non-cancer care is cost-effective or not. Further studies aimed at understanding the key aspects of MDT working that lead to cost-effective care, and the costs of supporting MDT working are required.
Endnotes a Costs can be derived by a bottom-up or top-down approach. In a bottom-up approach, costs are determined by examining individual utilisation. In a top-down approach, costs are apportioned across individuals, and this may be less accurate due to questionable assumptions used in the apportionment of costs. b Personal communication with the lead author has confirmed that these costs did not include the full costs of administering, preparing for and attending an MDT meeting. c As MDT working is now so prevalent in secondary care, especially cancer care, it might not be possible to compare the cost-effectiveness of no-MDT versus MDT working.

Appendix 1
Search Strategy EMBASE #1. multidisciplinary OR "multidisciplinary team" OR interdisciplinary OR "interdisciplinary team" OR team* OR "patient care team" OR "health care team*" OR "medical care team" #2. "cost savings" OR "cost-effective*" OR "cost-benefit" OR "cost-utility" OR "economic evaluation" OR economic* OR "cost analysis" OR "health care costs" #3. hospital or "hospital setting" OR "hospital based" OR "secondary care" #4. #1 AND #2 AND #3 #5. limit to (human and English language) AND (article) MEDLINE (MeSH terms used) #1. (Patient Care Team) #2. ((Costs and Cost Analysis) OR (Economic*) OR (Cost-Benefit Analysis) OR (Cost Savings) OR (Health Care Costs) OR (Cost of Illness) OR (Health Services Accessibility) OR (Patient Transfer) OR (Travel) OR (Transportation of Patients) #3. #1 AND #2 #4. limit #3 to (English language and humans) NHS EED (multidisciplinary OR "multidisciplinary team*" OR interdisciplinary OR "interdisciplinary health team*" OR team* OR "patient care team*" OR "health care team*" OR "medical care team*") (english:la) CINAHL #1. multidisciplinary OR "multidisciplinary team*" OR interdisciplinary OR "interdisciplinary health team*" OR team* OR "patient care team*" OR "health care team*" OR "medical care team*" #2. "cost savings" OR "cost-effective*" OR "cost-benefit" OR "cost-utility" OR "economic evaluation" OR economic* OR "cost analysis" OR 'health care costs' OR access* OR travel OR burden #3. #1 AND #2 #4. limit to English Language; Peer Reviewed; Research Article; Exclude MEDLINE records. EconLit #1. multidisciplinary OR "multidisciplinary team*" OR interdisciplinary OR "interdisciplinary health team*" OR team* OR "patient care team*" OR "health care team*" OR "medical care team*" #2. "cost savings" OR "cost-effective*" OR "cost-benefit" OR "cost-utility" OR "economic evaluation" OR economic* OR "cost analysis" OR 'health care costs' OR access* OR travel OR burden #3. #1 AND #2 #4 limit to (English language) AND (article or conference paper) Cochrane Library multidisciplinary OR "multidisciplinary team*" OR interdisciplinary OR "interdisciplinary health team*" OR team* OR "patient care team*" OR "health care team*" OR "medical care team*" in Title, Abstract or Keyword and "cost savings" OR "cost-effective*" OR "cost-benefit" OR "costutility" OR "economic evaluation" OR economic* OR "cost analysis" OR 'health care costs' OR access* OR travel OR burden in in Title, Abstract or Keyword NHS HMIC (multidisciplinary OR "multidisciplinary team*" OR interdisciplinary OR "interdisciplinary health team*" OR team* OR "patient care team*" OR "health care team*" OR "medical care team*") (english:la)