Withdrawal of anticancer therapy in advanced disease: a systematic literature review

Searches yielded 8368 citations, 371 abstracts and 81 full papers were assessed. A
total of 42 papers met the inclusion criteria and were included in the final analysis.
Fig. 2 summarises this process.

Fig. 2. Flow diagram illustrating systematic search process

Numbers of included papers are indicated in square brackets [ ]. Thirty-two papers
(76 %) investigated decisions to stop treatment for a range of solid and haematological
cancers. Ten (24 %) examined specific malignancies: lung 4], pancreas 2] and 1 each for ovarian, breast, lung and colorectal, colorectal and breast cancers.
Most examined chemotherapy in general rather than specific drug regimens. Two papers
discussed the use of MTAs in the final months of life 38], 39], and a further three briefly commented on MTAs 40]–42]: none described decision making processes concerning stopping MTAs close to the end
of life. Papers came from 14 different countries, some studies comprised evidence
from multiple countries: USA 15], Netherlands 5], UK 4], Japan 3], 2 each from New Zealand, Canada, Italy and Sweden and 1 each from Germany, Australia,
Taiwan, Belgium, Korea, UK with USA, and Australia with Israel. Twenty-four of these
studies contained primarily qualitative data, and 18 contained primarily quantitative
data. Additional file 1 [see Additional file 2] summarises the main characteristics of included studies. The sensitivity analysis
revealed that removing studies rated as low quality did not alter the overall findings;
no themes or topics were lost, and only a small amount of the ‘depth’ or ‘thickness’
of the findings was lost.

(1) How are decisions made?

The process of treatment withdrawal was described as a shared progression of many
different conversations rather than a one-off decision 11], 38], 40], 42]–50]. It involved a trajectory of repeated reassessments and trying different treatment
options 13], 38], 40], 42]–50], including treatment breaks before starting subsequent new regimens 4], 38], 40], 42], 50]. MTAs contribute to this process by providing more options before cessation of all
treatment 2], 38], 39]. Decisions about withdrawing treatment were described as an integral part of patient
care 45], strongly influenced by patients’ views and wishes 51]. Collusion in continuing inappropriate treatment may occur unless both oncologist
and patient are ready to discuss stopping treatment and both acknowledge that the
end of life is near 38].

Oncologists commonly report finding discussions about stopping treatment difficult
6], 38], 42], 44], 46], 50], 52] and emotionally challenging 3], 38], 49], 52].

“What I like least [about my job] is giving people bad news. It’s just terrible. It’s just horrible. It’s so sad…It’s
just horribly sad.” (Oncologist) 38]

Clinicians reported that discussions around stopping treatment are commonplace and
challenging, with a limited evidence base outside of clinical trial protocols available
for guidance 38]. The ‘grey area’ in which patients want another line of chemotherapy and their condition
could permit it, but on the balance benefits and burdens it would be inadvisable,
is particularly challenging 52]. The process of withdrawing treatment involves switching role from that of an “advocate”
for curative or disease modifying treatments to a supportive “counselling” role for
palliative, end of life care 49]. This role switch may create “role dissonance” for oncologists who often perceive
themselves as “treaters” who “do something” 38]. Clinicians may become “stuck” on an “institutional script for treatment” 46] which prevents them from discussing alternative supportive interventions, “watchful
waiting” 53], or discussions concerning death and dying 46], 54].

Stopping treatment may be raised by clinicians as a new direction of treatment in
which the patient’s emotions are directly addressed 42], or by presenting it as a “biological fact” that leaves little room for discussion:

“And so we started you off with the standard treatment. . . .We didn’t quite see what
we wanted with that, we moved on to something else. One thing I just want to make
sure you realize is that we don’t save our best weapons for the end. We use them up
front. So the chances of you responding to other agents, are even less than the chances
were with the two other regimens you’ve already gotten”. (Oncologist talking to patient) 42]

Some oncologists describe a “contemplation stage” during which evidence is gathered
concerning disease progression and treatment options and consideration given on how
to present this to the patient 38], 49], 52].

Few studies addressed the involvement of patients and their supporters in treatment
cessation decisions. Oncologists expressed the view that patients need to be beginning
to accept their terminal status before they can be ready to consider stopping treatment
3], 44], 47], 51], for some this acceptance may not happen until very close to death 44]. Patients with greater experience of treatment may be more confident about negotiating
treatment options and being involved in decisions 48]. Those with strong belief in the effectiveness of the treatment may have it continued
longer 55]. Some patients withdraw from the decision-making process as their disease progresses
and a surrogate decision-maker takes over 51], 53]. A perception that doctors do not wish to involve family members in discussions may
cause considerable distress 56].

(2) Why are decisions made?

Decisions to stop treatment are complex and multifaceted, with four major factors
identified from the literature.

(a) Disease and clinical factors: Key factors in decisions to stop treatment included worsening patient condition or
functional status 10], 38], 43], 44], 49], 52], 55], 57]–60] disease progression or advanced stage disease 9], 38], 40], 43], 44], 47], 49], 52], 57], 60] and treatment side effects 7], 38], 49], 52], 55], 59]–61]. Patients who received their diagnosis when only in advanced disease, and patients
who had not received treatment earlier in the course of illness for other reasons,
often had their therapy extended much longer than patients who had received earlier
treatments 5], 38], 62]–65].

The type of cancer influenced treatment decisions 7], 38], 43], 49], 60], 62], 64], 66]. For example, patients with haematological malignancies 4], 38], 43], 62], 63], and advanced lung cancers 2], 38], 64], were more likely to have their treatment continued. Chemo-responsiveness of tumours
was reported to be a key factor in decisions to continue treatment 2], 43], 66], although two population-based studies found no such connection 41], 67]. Six population studies found treatment was more likely to be withdrawn or withheld
in older people 57], 58], 62], 63], 67], 68], although one study found age was not a predictor of discontinuing palliative treatment
60].

(b) Clinician-dependent factors: A range of non-clinical factors influenced decision-making in the face of cancer
progression. Clinicians’ personal approaches, heavily shaped by their personal perspectives
and ethics 6], 38], 43], 49], 52], 69], 70] were significant predictors of whether chemotherapy was continued as disease progressed
2], 57], 66] Doctors’ views of patients’ personalities and circumstances also influenced the decision
4], 38], 43], 49], 70].

“I think instinctively you feel that this is a young patient with a young family you
need to make even more effort to try and help them live for a bit longer”. (Oncologist) 49]

Some doctors reported they would continue treatment within two weeks of death for
even a small chance of possible extension of life 39]. Treatment was also more likely to be continued by younger and less experienced doctors
2], 38], 43], and in the face of uncertainty of clinical benefit 38], 39].

“We are poor predictors of prognosis even in these near death time frame and we therefore
err on the side of more treatment.” (Oncologist) 39]

Treatment was often continued to avoid “taking away hope” 4], 38], 49], 52], 71] and when strong relationships have developed with patients 3], 43], 44], 72].

(c) Patient-dependent factors: Hope for the future was an important driver for patients’ decision-making 3], 44], 47], 48], although this hope at times reflected poor understanding of the palliative rather
than curative aims of treatment 4], 48], 72]–75]. When making decisions, patients strive to balance hope and improvements in quality
of life, with side-effects and the burdens of treatment 5], 51], 55], 61], 63], 68]. Older patients are more likely to discontinue treatment when approaching the end
of life 8], 57], 58], 62]–64], 66], 67], although one study found no such association 60].

(d) Environmental factors: Hospital setting was a key influence on whether or not treatment was continued 6], 38], 39], 57], 60], 62], 76], 77]. Treatment was more likely to be continued in teaching hospitals 62], and those receiving private treatment in physicians’ offices, rather than in general
hospital outpatient clinics 77]. Access to and information about palliative care services was associated with stopping
treatment 3], 39], 57], 60], as was having a supportive care plan in place 54], although hospice referral was not always associated with treatment discontinuation
54].

Treatment costs also influence these decisions, particularly in health systems where
patients have to meet drug costs personally 3], 38], 39], 77]. The availability of new MTAs, although expensive, influence decisions to continue
treatment as they provide additional options for continuation of treatment 3], 38], 39], 43].

(3) When are decisions made?

The literature concerning the timing of decisions to withdraw treatment was very limited.
Clinicians expressed great difficulty over judging the appropriate time to stop 5], 38], 43], 44], 48], 49], particularly when a patient’s condition could justify, but the prognosis was unclear
and the benefit of treatment was uncertain 52]. Patients also expressed uncertainty about the timing of treatment withdrawal. In
one study, all participants unanimously emphasised the difficulty of anticipating
the right time to make the decision to stop all treatment 48].

(4) Who makes the decision?

The majority of studies reporting upon ‘who makes the decision’, found decision-making
was an on-going process based in the interaction between the physician and patient,
and sometimes close relatives 8], 38], 40], 43], 45], 48], 49], 52], 73]. However there was mixed evidence upon who in reality was involved in decision-making
and who made the final decision. One study reported fewer patients to have been involved
in the decision-making process than wanted to be, and that only half of competent
patients were involved 78]. In contrast, another study reported family members to believe that the patients’
preferences had been followed in 78 % of cases 79]. Nurses often have an important supportive role 4], 43], 52], 57], 61], as may the wider multidisciplinary team 49].

A smaller number of studies found decisions were driven by one particular party: four
studies found decision-making was patient driven, two of these studies were based
in the USA, one was based in The Netherlands and the other in Israel and Australia
39], 44], 51], 52]; in contrast three studies found physicians had the strongest influence, one on these
studies was based in the USA, one in New Zealand and one in the Netherlands 43], 52], 71]. The number of studies and the limited high quality evidence on this topic means
that no effect by country or continent could be found.

Doctors exerted influence by choosing which options to present to patients 43], and the way in which those options were presented 52]. Some patients preferred not to be involved in the decision-making at all 48], while for others feeling in control was important, even if it was only control of
day-to-day decisions 44], 48], 80].

“I want to be able to have control or say about my illness, whether I think I should
take chemo or not. The doctors tried to talk me out of it [her decision to stop treatment], and it’s just like, it’s MY body. I feel it’s not going to do anything for me. It’s
making me sicker so why do it? I felt that it was important for me to have control
over that.” (Patient declining chemotherapy in advanced disease) 80].