Intended for healthcare professionals


Malignant cerebral glioma—II: Perspectives of patients and relatives on the value of radiotherapy

BMJ 1996; 313 doi: (Published 14 December 1996) Cite this as: BMJ 1996;313:1512
  1. Elizabeth Davies, clinical research fellowa,
  2. Charles Clarke, consultant neurologista,
  3. Anthony Hopkins, directorb
  1. a Directorate of Neurosurgery and Clinical Neurosciences, St Bartholomew's Hospital, London EC1A 7BE
  2. b Research Unit, Royal College of Physicians, London NW1 4LE
  1. Correspondence to: Dr E Davies, Research Unit, Royal Colege of Physicians, London NW1 4LE.
  • Accepted 8 October 1996


Objective: To explore the experiences of patients and relatives after the diagnosis and treatment of malignant cerebral glioma.

Design: Two year prospective study with home interviews.

Setting: Six neurosurgery and radiotherapy centres in London.

Subjects: 75 patients and 66 close relatives interviewed at diagnosis, 58 patients interviewed after radiotherapy, and 27 interviewed after recurrence.

Main outcome measures: Awareness of likely prognosis, distress, dissatisfaction with radiotherapy, and perception of severe problems in everyday life.

Results: As they began radiotherapy most patients understood that they suffered from a brain tumour (95%; 71/75), but only one quarter (19/75) seemed fully aware of the poor prognosis. Others were unaware (43%; 32/75) or only partly aware (32%; 24/75). The more aware patients were more distressed. Relatives were three times more likely to be aware of the prognosis (67%; 44/66) and were more distressed. Although 39% (29/75) of patients initially made negative comments about radiotherapy, only 17% (13/75) were completely dissatisfied. The decision to accept radiotherapy could be discussed directly with 19 fully aware patients. Twelve found radiotherapy acceptable if it were medically advised or if it improved survival. Assessed by their own reports of problems only 40% of patients achieved a period of stability or remission, yet dissatisfaction with treatment did not increase.

Conclusion: Most patients with malignant glioma initially seemed unaware or only partly aware of the poor prognosis. Relatives were more aware, more distressed, and often concerned to protect patients from full awareness, which made it difficult to explore with patients directly the possible trade off between quality and length of life. Conceptualising the question as a rational choice ignores the social and emotional context of life threatening disease.

Key messages

  • In general, relatives are more distressed than patients and more likely to be aware of the poor prognosis but tend to want to protect patients from this knowledge

  • Communication with patients needs to take into account their need to make informed decisions about treatment and the implications of the disclosure of such information for communication with families

  • Patients who initially have better WHO clinical performance status (0–1) are most likely to report improving or remaining stable after radiotherapy. Overall only 40% of a non-trial population thus achieve a remission


In some medical settings doctors may perceive treatment to be unwarranted as to prolong a considerably diminished quality of life or to reduce this further by adverse effects may not be justified. Considerations concerning quality of life as opposed to length of life underpin much of the current debate over setting priorities for treatment,1 2 and various formulations of how to resolve such conflicts have been published.3 4 5 6 Whether such formulations would gain much favour in the population at large when faced with the disease in themselves or those close to them is uncertain. Some patients with cancer, for instance, view chemotherapy characterised by side effects and a small chance of success as more worth while than do medical staff or a sample from the general population.7 8 Moreover, patients with cancer (and those with other serious diseases) may report the quality of their lives in unexpectedly positive terms.9 10 11 12 13

We considered the tension between quality and length of life for patients with malignant cerebral glioma, a disease for which the prognosis is almost uniformly poor.14 15 16 In the previous paper in this issue (p 1507) we described a cohort of 105 patients, 92 of whom were treated with radiotherapy. In common with previous studies, we found a median survival with radiotherapy after diagnosis of 10 months.16 The median time free from serious disability was only four months. One quarter of those living more than six months after the diagnosis suffered clinical deterioration or disability apparently associated with treatment. A further 42% experienced severe tiredness.16 This paper explores patients' experiences after the diagnosis and in particular their acceptance of radiotherapy. We also describe the views of close relatives concerned in their care.

Patients and methods


Recruitment to the study and follow up by home visit is described in the accompanying paper.16 We did not consider that a questionnaire approach was an appropriate way to broach the sensitive issues we needed to explore. We therefore began the research using open ended interviews with eight patients and relatives. These interviews informed the areas to be explored in the main study, in which we used a semistructured interview format. This method allows respondents to talk at length about any matter that concerns them and yet leaves the interviewer scope to explore difficult issues.17 18 A series of four point rating scales on the basis of interviewer interpretation were developed to cover the range of experiences and meanings expressed. Table 1 shows the principal topic areas. An accompanying manual (available from the first author) defined the rating scale points (“marked,” “moderate,” “some,” and “nil”) for each topic. It also provided anchoring examples on the basis of verbatim interview comments as standards for the level of detail to be obtained. Our rating of distress, for example, took account of all the concerns expressed and the degree of emotion that was conveyed. Patients or relatives constantly upset and overwhelmed by the situation were rated as “highly” distressed while those who felt upset or anxious a good deal of the time but found pleasure in some aspects of life were rated as “moderately” distressed. Those feeling only occasionally depressed, anxious, or dismayed or remaining generally cheerful we rated as showing “little” or “no” distress.

Table 1

Inter-rater reliability study for 13 interviews∗

View this table:

Data collection was shared by three interviewers with social science training, one of whom (ED) was also medically qualified. All interviews were tape recorded and transcribed in full, with notes of emotional content such as pauses, false starts, sighs, and weeping. Table 1 shows results of a reliability study, in which 13 transcribed interviews were rated by each of the three raters. Reliability was good or satisfactory19 particularly when rating scales were dichotomised between “marked/moderate” and “some/nil”'(table 1).


The themes emerging from the first 30 interviews have been briefly described elsewhere.20 These interviews suggested that we should consider patients' views at three stages in the clinical course rather than at set intervals after the diagnosis. These were, firstly, coming to terms with the diagnosis, secondly, adapting to change during any remission, and, thirdly, a final stage of deterioration and increasing disability. The interviews also suggested that we should explore differences in awareness of the likely prognosis between patients and relatives. In our accompanying paper we showed that the patient's initial disability assessed by the World Health Organisation clinical performance status21 was an important prognostic factor.16 In this paper we compared this measure with patients' own views about their problems. We used the χ2 test and χ2 test for trend throughout.


Table 2 summarises some demographic characteristics of the sample. As many as 41% of patients were from professional and managerial classes,22 and just over two thirds were men. Sixty nine percent of carers were women, and most carers were spouses or partners (71/91; 78%).

Table 2

Demographic characteristics of 105 patients and 91 relatives. Figures are numbers (percentages) of subjects

View this table:

Because of the severity of their disability, none of the 13 patients who received conservative treatment with steroids could be interviewed in any depth. The main analysis therefore concerns the 92 patients who received radiotherapy. Of these, 75 could be interviewed within three months of the diagnosis, and 59 could be seen for at least one follow up interview. Twenty seven of these were seen again at a time when they had deteriorated after an initial period of stabilisation or improvement. Death or obvious confusion prevented further patients being seen at this stage of the illness.


At our first interview, as they began radiotherapy, we asked patients to recount in detail what they had been told in their medical encounters and what had been said by relatives and friends. From this it was clear that most patients receiving radiotherapy understood they suffered from a brain tumour (95%; 71/75). Awareness of the prognosis, however, varied considerably. We defined three levels of awareness taking into account the degree to which patients believed that radiotherapy would cure them and whether they seemed to be aware that they might die from their disease. From the tape recordings we considered that only a quarter (19/75) were fully aware of the prognosis, conveying that there was little chance of cure and expressing some thoughts or fears about the possibility that they might die. A 56 year old man with a parietal tumour, for example, spoke frankly about radiotherapy:

“Well, fundamentally it just delays the process. It doesn't stop it. Wish to God it did but it doesn't. We talk about it all the time that I'm going to die, that it's going to happen after Christmas or whenever.”

We rated just under a third of patients (32%; 24/75) as being partly aware of the likely outcome. One 40 year old man with a frontal tumour, for example, remarked:

“The family are still frightened that something might happen. Do you know what I mean? Because you read about it in the papers everyday don't you—this trouble? Cancer—person dies from brain tumour. So I suppose it's at the back of your mind as well.”

Such patients usually conveyed some fear of dying but also that they had a reasonable chance of cure. As many as 43% (32/75) of patients, however, conveyed in the interviews practically no awareness that they might die. One 40 year old man with an occipital tumour, for example, when asked about the chance of cure said,

“I didn't specifically ask the surgeon. He spoke as if it were pretty certain. I can see that people who had only biopsy and growths in awkward places, that there would be a question mark over how complete treatment would be. But in my case, because it seemed to be something that was on the surface…then you should be able to wipe it out completely.”

We had initially thought that patient optimism might be explained by high dose steroids20 and obtained data on the patients' doses at the time of interview in 62 cases. A comparison of low dose (0–2 mg) with moderate (3–8 mg) and high doses (9–32 mg) of steroid, however, showed this had no influence on awareness.

The relatives

Sixty six of these 75 patients had relatives who could also be interviewed at diagnosis. These relatives were three times more likely than the patient to be aware of the poor prognosis (67% (44/66) v 21% (14/66); table 3). The wife of the unaware patient quoted above said:

Table 3

Comparison of awareness of patients and relatives or carers of likely prognosis shortly after diagnosis

View this table:

“They told my husband it was a tumour, but they didn't tell him about the prognosis or anything like that…. I caught the surgeon in the corridor and he said that my husband could have been dead in a month if he hadn't had the operation. He said that some people did well—it could be 18 months, or even 5 years but not really—5 years was quite rare. It knocked me for six.”

The implications of such news was also often spelt out. A 50 year old husband remarked:

“From now on we're living with a death sentence. The doctors said the tumour could come back quite easily, and I got the feeling they were expecting it to.”

Only 9% (6/66) of relatives therefore were unaware of the likely prognosis. The agreement between awareness of patients and relative was poor ((kappa) 0.20). Moreover, there was a consistent tendency for patients to show less awareness (39 out of 66 comparisons between patients and relatives).

On the whole, relatives tended to see their role as one of remaining strong to protect the patient in the hope that they would be among the most fortunate. It was not uncommon for them to convey that to be explicit about the prognosis would be harmful:

“If you say you've got six months or whatever, it does mentally act as a sentence to that person.”

As the illness progressed, slightly more patients became fully aware of the prognosis. Most patients interviewed after deterioration were therefore aware that they were dying (52%; 14/27), although 26% (7/27) still conveyed no awareness of this possibility and the 22% (6/27) remaining were partly aware. At all stages of the disease most patients felt highly supported by their relatives (92% (61/66) at diagnosis and 77% (21/27) at deterioration).


Despite the seriousness of the condition patients were less distressed than might have been expected. Of the 75 patients who could be interviewed within three months of the diagnosis, we rated only 9% (7/75) as highly distressed. Six patients seemed severely depressed or anxious, and one expressed extreme anger. One fifth (16/75) showed a moderate level of distress, but more than two thirds of the patients (54/75) felt only occasionally depressed, anxious, or dismayed and remained generally cheerful or confident. The level of the patient's distress was moderately related to their awareness of the prognosis ((gamma) 0.53; table 4). Fifty eight percent (11/19) of the fully aware showed marked or moderate distress compared with one quarter of those partly aware (6/24) and 16% of those unaware (5/32) (χ2 for trend 9.53; 1 df; P<0.01).

Table 4

Awareness of likely prognosis and distress in patients shortly after diagnosis. Figure are numbers (percentages) of patients

View this table:

In clear contrast with patients, nearly two thirds of relatives experienced a marked or moderate level of distress (64%; 42/66). In most instances, relatives were in fact more distressed than the patients they cared for (58%; 38/66); in only a few instances did the patient seem more distressed (14%; 9/66). Too few relatives remained unaware of the prognosis to explore whether this factor was associated with lower levels of distress.

VIEWS ABOUT RADIOTHERAPY Initial dissatisfaction with radiotherapy

A substantial number of patients (39%; 42/66) made “marked” or “moderate” negative comments about radiotherapy at the initial interviews, expressing their dislike of the procedure, adverse effects, or doubt about its value. This compared with 29% (22/74, data missing in one case) of patients who expressed equal dissatisfaction with surgery. Such negative comments about radiotherapy were more common among patients with marked or moderate distress (59% (13/22) v 30% (16/ 53) with little or no distress; χ2 4.33; P<0.05), but such comments did not relate to awareness of the prognosis or initial clinical performance status. We rated only one fifth of the patient sample (17%; 13/75) as being dissatisfied in an overall sense, however, as they showed considerable reluctance about undergoing radiotherapy. This level of dissatisfaction tended to be more common among patients who were fully aware of the prognosis (32% (6/19) v 11% (6/56) of those partly or unaware; χ2 3.17; 1 df; P<0.1), but this difference was not significant.

We thought it appropriate to explore the decision to accept radiotherapy only with the 19 patients fully aware of the prognosis. Of these, six accepted treatment with reluctance because they saw no alternative but remained dissatisfied. Four accepted treatment as it had been medically advised, and nine believed they could increase their survival by fighting the disease, having accepted radiotherapy for the chance of survival it brought:

“I began to fight it, just began to start the battle…. What I am doing is living for the future. I would like to think that I've got at least a year or so, or more. I'm thinking if you like of a strategy in that I've got to get planned what I can do.”

Patients' improvement, deterioration, and subsequent dissatisfaction with radiotherapy

Patients' initial perceptions of problems in everyday life (table 1) correlated, as might be expected, with ratings of disability based on the WHO clinical performance status (table 5). In case dissatisfaction after radiotherapy related to the patient's perception of not improving after treatment, we first determined which patients could be classed, in their own terms, as having achieved a period of stability or a remission. Overall we found that less than a half (40%; 34/85) of those patients treated with radiotherapy on whom we had follow up data achieved either a period of stability or a remission. The most severely disabled patients, whether judged by their own perceived problems or the WHO clinical performance status, were least likely to achieve a period of stable or improved function (table 5). Patients who achieved neither tended to make more negative comments about treatment at their best follow up interview (50% (20/40) v 22% (4/18); χ2 2.89; 1 df; P<0.10), although this was not significant; and patients making such comments were more likely to be markedly or moderately distressed (58% (15/26) v 21% (7/32); χ2 6.37; 1 df; P<0.01). The proportion of patients dissatisfied in an overall sense with radiotherapy, however, remained stable at one fifth (21%; 12/58). Dissatisfaction at follow up did not relate to awareness of the prognosis, distress over hair loss, or level of distress. At deterioration, fewer patients made highly negative comments about radiotherapy (18%; 5/27) than on completion of radiotherapy (50% (20/40); χ2 5.55; 1 df; P<0.01). Only one patient entirely regretted having undergone treatment, although in such serious circumstances at this stage of the illness it did not seem appropriate to probe too deeply. Overall we could not therefore find evidence that most patients looked back over the illness feeling dissatisfied that they had undergone radiotherapy.

Table 5

Initial disability, perception of disability, and proportion of patients reporting Figures are numbers (percentages) of patients

View this table:


In this study we interviewed 75 patients with malignant glioma and 66 of their relatives to explore the social context of one disease in which treatment can only extend survival. Our findings show some of the difficulties of obtaining clear cut answers to questions about the trade off between quality and length of life. We suggest these difficulties relate to the very nature of such severely threatening situations.

In our study, detailed interviews with patients shortly after the decision to have radiotherapy had been made showed that only one quarter (19/75) seemed to be fully aware of the strong possibility that they would die from their disease and that the best they could expect from radiotherapy was some prolonging of life. Assessed by their own reports of problems, only 40% of patients achieved a period of stability or a remission after radiotherapy. In our accompanying paper16 we described the extent of adverse effects associated with radiotherapy, yet in terms of this, and the eventual outcome, relatively few patients conveyed that they regretted having undergone treatment. There was a suggestion that patients who had not achieved a remission made more negative comments about radiotherapy, but little more than one fifth of the sample were dissatisfied in an overall sense, and dissatisfaction did not increase with time.

Social classes I and II are disproportionately represented in our sample. This may reflect the distribution of the disease in the population around London or more probably differences in the referral of patients from different social classes to specialist centres. We do not believe this bias reflects selective recruitment as we identified all available patients in a consecutive manner at each centre. Despite this, however, relatively few patients demonstrated the level of awareness that would have allowed us to explore their thoughts about the kind of trade off they might have been making. Other studies have avoided this problem by asking patients about hypothetical situations rather than their own, and this indirect method has shown that cancer patients are prepared to consider more intensive treatment regimens than members of the general population.7 8 That most of the patients whom we could directly question were prepared to undergo radiotherapy for the chance of longer survival is consistent with the results of such studies.

The possibility that the patients' lack of awareness of the prognosis and their acquiescence to radiotherapy was directly due to their brain disease is made unlikely by the fact that severely confused patients were excluded. The rest were interviewed at length and gave informed and interesting answers to many of our questions. Lack of awareness of the prognosis was also unrelated to the patients' dose of steroids at the time of interview. Part of the explanation for limited awareness seemed to lie in the fact that many had been protected by relatives and medical staff. Patients' relatives, for instance, were three times more likely to be aware of the probable outcome and showed more distress in the interviews. Some patients may have had subtle neuropsychological deficits affecting the way they attributed significance to new information, and in some cases such factors may have reinforced the desire of relatives to protect patients from distress. Another recent interview study found that only 11 out of 30 patients with malignant brain tumours associated their tumours with death23 and described different psychological strategies that patients used to maintain hope. In our study, patients who were aware of the poor prognosis, however, were more distressed. This finding is consistent with the suggestion in some studies that when the prognosis is very poor, some patients may prefer not to know24 or hold on to the belief that they are being treated for a cure.25 It is clear that issues of communication are important in this disease, particularly as it is now generally accepted that communication with cancer patients should be straightforward and open.26 27 28 Our findings underline that the consequences of the communication style adopted by medical staff need to be considered in terms of distress for both the patients and their families, particularly when many families seem to adopt a coping style which is not based on the continuous and explicit discussion of the truth.29 30 31

One explanation that would make sense of our findings is that the question of a trade off between quality and length of life is one more often raised by concerned observers than by those facing the possibility of imminent death or bereavement. This observer perspective seems to emphasise the possibility of a rational choice at the cost of largely ignoring associated emotions. In reality many of the patients and their relatives took hope from the apparent element of clinical uncertainty and from the fact that some treatment was being attempted. This strategy may have allowed them to cope with the day to day problems that the disease caused and, by keeping their distress within manageable levels, helped to maintain the quality of their life for a short period. Observers tend to interpret such coping as creating only “false hope”, but an alternative perspective is to accept that this need for hope is an objective feature of such situations which cannot be rationalised away. Once accepted, it could be debated whether this need for hope should be incorporated into quality of life measures or into calculations about the cost of interventions to the NHS. In short, the dilemma is how much can health services afford costly interventions which provide principally hope, even though that hope improves the quality of life?

We thank the patients and relatives who agreed to talk to us. Maureen Bannon and Sue Hall helped with the development of the interview and rating scales and with data collection. Rachel Warrick and Valerie Pottinger gave administrative support. We are also grateful for the comments of three anonymous reviewers.


  • Funding Cancer Research Campaign grant number CP 1017.

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.