MDT Decision Making

Decision-making about prognoses within Multi-Disciplinary Teams (MDT's) using experimental studies

This study is part of the MCPCRD's programme of research 'Improving care, assessment, communication and training at end-of-life (I-CAN-CARE)' that is running from 2016-2019. The programme of research is funded by Marie Curie.


Previous research has suggested that clinical predictions of survival are both inaccurate and over-optimistic. Our own research has shown that, in patients with advanced cancer, a multi-disciplinary team (MDT) survival estimate is indeed more accurate than either a doctor's or a nurse's estimate alone. However, the factors which expert clinicians use to formulate prognoses and the manner in which these estimates are combined to arrive at an MDT estimate are not well understood. 

The ultimate purpose of this research will be to make recommendations about measures that MDTs can take to improve the accuracy of prognostic judgments. It is recognised that real-world MDT decision-making is inevitably more complicated than our simple experimental set-up will allow. However, it is often the case that more complex situations are best first addressed by investigating their simpler components.

The aims of this study are to investigate the following research questions:

1. Does prognostic accuracy improve with receipt of a second opinion?

2. Is the degree to which prognostic judgments are adjusted dependent upon the   characteristics of the advisor.

Outline of the study

We will work with doctors and nurses with a range of experience and seniority involved in the care of patients with life-limiting illnesses. The methodology we plan to use in this study is borrowed from the Judge-Advisor System (JAS). The JAS is a form of advice taking structure and is considered a subset of decision-making in the social sciences. There are two roles in a JAS, a judge and an advisor. The judge is the decision maker who assesses the information concerning a specific decision and makes the final decision. An advisor is the person who offers advice, information, or suggestions to the judge. A key component of the dynamics in a JAS is the differentiation between the two roles in that while the advisor provides input to the decision, actual decision-making power resides solely with the judge.

This experimental study will involve individual clinicians completing a prognostic judgment task presented to them on a computer. The study will be conducted in two phases. In phase 1 each clinician will be asked to make a prognostic judgment on approximately 15 vignettes. Each vignette is based on patients receiving specialist palliative care in hospitals and hospices that are expected to die within approximately two weeks and for whom the date of death and length of survival will be known. For each case clinicians will be presented with key prognostic information (e.g. diagnosis, age, extent of disease, performance status, nutritional intake, breathing pattern, blood results) and they will be asked to estimate how long they expect the patient to live. Clinicians will give an estimate in 'days' along with the range within which they would expect 90% of patients to fall. During phase 2 clinicians will be presented with the same cases and will be provided with two prognostic estimates; their own and that of another 'advisor'. The characteristics of the 'advisor' will vary and may be a doctor, a nurse or information provided by a prognostic score (e.g. the Palliative Prognostic Index). The clinician will then be given the opportunity to give a revised estimate in the light of the advice received. These data will be used to calculate the degree to which the clinician is inclined to discount the opinions of other MDT members.

Research team

Professor Paddy Stone - Principal Investigator, MCPCRD UCL

Dr Jane Harrington - Senior Research Fellow, MCPCRD UCL

Dr Adam Harris - Lecturer in Experimental Psychology, UCL

Dr Priscilla Harries - HOD Clinical Sciences, Brunel University