Therapeutic futility in cancer patients at the time of palliative care transition: An analysis with a modified version of the Medication Appropriateness Index

Abstract

Background:

Palliative Care professionals are often confronted with therapeutic futility, consisting in inappropriate strategies that do not add any advantage to the patient and may actually increase adverse events. Scientific literature concerning this issue is lacking. This article is one of the first to study therapeutic futility specifically at the time of transition to the palliative care setting.

Aim:

To study the phenomenon of pharmacologic therapeutic futility at the time of transition of a cancer patient to palliative care.

Design:

The pharmacological prescriptions at the time of the first appointment at an oncological palliative care unit during a time period of 2 months were prospectively collected and characterized using the Medication Appropriateness Index.

Participants:

The sample comprised 71 patients with a mean age of 68.2 years.

Results:

The most common pharmacological groups were analgesics (n = 121; 19.2%), psychoactive drugs (n = 89; 14.1%), and antihypertensives (n = 51; 8.1%). A total of 61 patients (85.9%) consumed 5 drugs or more at the time of the first appointment. The mean number of daily medications decreased significantly after the palliative care team intervention, from 7.15 to 5.73 (p < 0.05). The principal causes of inappropriateness were absence of indication for the drug (23.0% “inappropriate”), the drugs’ adverse interactions (11.1%), and inadequate dosage (9.9%). After the first consultation in the palliative care setting, 28.2% of the drugs were suspended.

Conclusion:

This article tried to evaluate the main causes of therapeutic futility at the palliative care transition. The principal causes of inappropriateness were absence of clinical indication, clinically significant drug–disease/comorbidity interactions, and incorrect dosage/posology.

Keywords

Therapeutic futility palliative care palliative care transition bioethics Medication Appropriateness Index

Introduction

Palliative medicine has as its most fundamental purpose the prevention, treatment, and relief of physical, psychological or existential symptomatology in patients with advanced stage pathologies. It is generally applied (though not exclusively) in patients with limited life expectancy and in highly complex clinical situations.

Disease’s progression is characterized by an increase in symptom intensity, mandating a progressive adoption of preventive measures aiming at minimizing discomfort. At the same time, a progressive discontinuation of disease-modifying strategies and long-term prophylaxis is warranted, as these treatments become gradually more superfluous. Failing to do so may lead to interventions that are not beneficial, since they do not have an impact in the patient’s overall clinical condition. These interventions may, in fact, induce pain or damage, besides amounting to an unnecessary burden to healthcare systems. This is termed “Therapeutic futility,” a frequently unrecognized and underrated kind of iatrogenesis.

However, current literature concerning therapeutic futility is still scarce.¹ There are not many instruments to quantitatively measure it and studies using those instruments applied them to the geriatric population, not the oncologic patient population.¹ Other studies with cancer patients (like Riechelmann et al.² in Canada or Fede et al.³ in Brazil) did not use quantitative instruments to assess therapeutic futility¹ (a notable exception being the American study by Suhrie et al.).^1,4 Papers concerning palliative care patients focused on the terminally ill, not on the palliative care transition.¹

This article is one of the first to study therapeutic futility in oncological patients specifically at the time of transition to the palliative care setting (a time in which physicians are more prone to therapeutic futility). Also, a quantitative questionnaire was used (the Medication Appropriateness Index (MAI)) to ascertain why the medications were inappropriate.

Knowing the reasons for medication inappropriateness at the palliative care transition impacts the way physicians treat cancer patients, allowing optimization of the management of these patients, both from an economic and a bioethical perspective.

Objectives

The main objective of this article is to study the phenomenon of therapeutic futility at the time of transition to a palliative care unit. To attain this purpose, the following specific objectives were stipulated:

Ascertain the medications consumed at that time.

Assess the trends concerning the types of drugs consumed before and after a palliative care team’s first intervention.

Calculate the incidence of suspended medications after that intervention.

Appraise if the degree of appropriateness is significantly related to the type of medication consumed.

Methods

This prospective observational study used the MAI⁵ to assess therapeutic futility at the time of a first appointment at a tertiary palliative care unit. The MAI (Table 1) is a questionnaire consisting of 10 questions, to each of which the physician assigns a score, ranging from 1 (“appropriate”) to 3 (“inappropriate”).

Table 1.

Medication Appropriateness Index (MAI).

Question 1	Is there an indication for the drug?	1: Indicated	2: Moderately	3: Not indicated
Question 2	Is the medication effective for the condition?	1: Effective	2: Moderately	3: Not effective
Question 3	Is the dosage correct?	1: Correct	2: Moderately	3: Incorrect
Question 4	Are the directions correct?	1: Correct	2: Moderately	3: Incorrect
Question 5	Are the directions practical?	1: Practical	2: Moderately	3: Impractical
Question 6	Are there clinically significant drug–drug interactions?	1: Insignificant	2: Moderately	3: Significant
Question 7	Are there clinically significant drug–disease/condition interactions?	1: Insignificant	2: Moderately	3: Significant
Question 8	Is there unnecessary duplication with other drug(s)?	1: Necessary	2: Moderately	3: Unnecessary
Question 9	Is the duration of therapy acceptable?	1: Acceptable	2: Moderately	3: Unacceptable
Question 10	Is this drug the least expensive alternative compared with others of equal utility?	1: Least expensive	2: Moderately	3: Most expensive

Participants

The participants were all the patients who underwent a first consultation at the Palliative Care Unit of a tertiary oncologic Portuguese center (Instituto Português de Oncologia Francisco Gentil, Porto), during a time period spanning from November to December 2011. This palliative care unit admits adult patients treated in that institute, who have an incurable, advanced, and progressive cancer, without an acceptable response to antineoplastic treatments (or refusing said treatments). All first consultations were included, either in the inpatient or in the outpatient setting.

Procedures

The medications consumed at the time of the first consultation were recorded and assigned to a pharmacological group according to the classification of the official Portuguese therapeutic index (“Prontuário Terapêutico”⁶) from the national pharmaceutical regulatory entity. Each medication was also categorized as a regular prescription (i.e. medications administered at fixed intervals) or a rescue dose prescription (i.e. medications to be administered only at the onset of a particular symptom). Other data regarding the patient were also collected, namely, sex, age, race, Eastern Cooperative Oncology Group Performance Status (ECOG PS), and type of cancer. All information was given either by the patient himself/herself or (if the patient had cognitive impairment) from his/her caretakers, or clinical files.

The prescriptions that were suspended after the palliative care team intervention were also registered. The physician who undertook the appointment then applied the questions of the MAI (freely translated from the original) to each drug. Before the beginning of the study, all the physicians involved participated in a meeting where they were introduced to the MAI, so as to limit variability due to different interpretations of the questions’ meanings. At that meeting, it was decided that a modified version of the MAI should be used, so that a “not indicated” response in question 1 rendered questions 2–5 redundant and so, not worth answering. Moreover, question 9 was only applied to non-chronic medications. Question 10 was not ascertained, because it was found that the authors did not have the adequate instruments at their disposal to measure it.

Regular weekly meetings were also held until recruitment was over, so as to evaluate potential difficulties in answering the questions and to further limit inter-rater variability. One of the authors supervised all the other participants as they responded to each questionnaire.

A “not indicated” response in question 1 meant that there was no clinical reason for the use of said drug (e.g. statins in terminal patients, since the protective effect of statins takes years to develop) and a “moderately indicated” response meant that that clinical reason was dubious. As for question 2, a “not effective” response addressed the inefficiency of the drug in order to achieve the clinical benefit for which it was prescribed (e.g. vitamin supplements to treat asthenia or anorexia) and a “moderately effective” response signified that the drug was not sufficiently effective (e.g. antidiabetic drugs that were not controlling glycemia adequately).

In question 3, an “incorrect” response meant that the posology of the drug was inadequate (e.g. a corticosteroid in a higher dose than needed for symptom control) and a “moderately correct” response was used for medications for which the posology could be correct to achieve clinical benefit, but not to avoid iatrogenic events (e.g. tensional targets in terminally ill patients are not as strict as for the general population).

Questions 4 and 5 addressed the instructions given to the patients concerning the drugs, so that the physicians categorized the directions according to their correctness and practicality.

Questions 6 and 7 were about the interactions of the drugs with other drugs or with the patients’ condition or comorbidities. A “significant” response in question 6 is exemplified by the interaction between a laxative and an antidiarrheal agent and the same response in question 7 may be illustrated by the interaction of intravenous (IV) sodium chloride with congestive cardiac failure. A “moderately significant” answer in those questions meant that a possible interaction could exist, but that it was dubious.

An “unnecessary” duplication in question 8 could be construed as the duplication of benzodiazepines for insomnia and a “moderately necessary” duplication as the use of an antihistamine for that same reason when a benzodiazepine was already given. An “unacceptable” duration of therapy in question 9 could be construed as the use of an antibiotic for an acute infection for longer than the normal course of therapy, and a “moderately acceptable” duration as the use of an IV fluid therapy for more time than needed to correct severe ionic imbalances (i.e. it was continued when those ionic imbalances were already moderate and had no weight on the patient’s overall status).

Data analysis

The data were analyzed using the IBM SPSS Statistics^® software, version 20. In a first phase, a descriptive analysis was undertaken to characterize the patient and medication samples. Another descriptive analysis, in terms of central tendency measures (mean, median, and mode), was applied to the number of medications consumed (before and after the appointment) and its appropriateness degree, according to the 9 questions from the MAI.

An inferential statistic was applied to test the relation between the variables. As the responses to the MAI questionnaire were concerned, non-parametric tests were used to study the degree of appropriateness of a pharmacological group relative to the others. The use of non-parametric tests was due to the reduced dimension of the several sample groups, with less than 30 elements each (and the absence, in the majority of cases, of a normal distribution). Kruskal–Wallis test was used for each question to detect whether there were general differences between all medication groups. When that difference was statistically significant, the Tukey test was used to see whether the less appropriate pharmacological group on Kruskal–Wallis was significantly more inappropriate relatively to the other groups.

When several drugs of the same kind were used, they were counted as several instances, but the dependence between instances on the same patients was ignored in the analysis, since its impact is likely to be small, as this outcome is binomial.

As for the variation in the mean number of medications before and after the consultation, it included a total number of medications, which exceeded 30 elements. In this case, a parametric test (paired samples t-test) was used, taking into consideration the existence of a normal distribution and the presupposition of the central limit theorem, which postulates that in group samples with more than 30 elements, there is no difference in the use of parametric tests relative to non-parametric ones (cf. Hill and Hill,⁷ Tanis and Hogg⁸). A significance level of 0.05 was used.

Results

Sample characterization

The sample comprised 71 patients (Table 2) with a mean age of 68.2 years (standard deviation (SD): 11.8 years); the majority (60.6%) were male.

Table 2.

Demographic characterization of the sample.

Parameter	N	%
Gender
1. Male	43	60.0
2. Female	28	40.0
Age
Minimum	42
Mean	68.2 (±11.8)
Maximum	90
Race
1. Caucasian	71	100.0
ECOG PS
0	0	0.0
1	10	14.1
2	12	16.9
3	17	23.9
4	25	35.2
N/A^a	4	5.7
Missing	3	4.2
Tumor frequency
1. Colorectal cancer	15	21.1
2. Prostate cancer	12	16.9
3. Breast cancer	8	11.3
4. Lung cancer	7	9.9
5. Head and neck cancer	5	7.1
6. Gastric cancer	4	5.7
7. Bladder cancer	3	4.2
8. Corpus uteri cancer	3	4.2
9. Sarcoma	3	4.2
10. Skin cancer	3	4.2
11. Lymphoma	2	2.8
12. Pancreas cancer	2	2.8
13. Biliary tract cancer	1	1.4
14. Esophagus cancer	1	1.4
15. Occult primary neoplasm	1	1.4
16. Renal cancer	1	1.4

ECOG PS: Eastern Cooperative Oncology Group Performance Status.

Not applicable—patient already disabled before cancer onset (e.g. paraplegia from previous accident).

There were 42 patients (59.2% of the total) with an ECOG PS ⩾ 3 (as is shown in Table 2). These were likely to be terminal patients in the last weeks of like. Even so, 22 patients (31.0%) had an ECOG PS < 3, with a higher expected survival. The most frequent primary tumors were colorectal (n = 15; 21.1%), prostate (n = 12; 16.9%), and breast (n = 8; 11.3%) (Table 2).

A total of 631 medications were recorded for 71 patients. The most common pharmacological groups (Table 3) were analgesics (n = 121; 19.2%), psychoactive drugs (n = 89; 14.1%), antihypertensives (n = 51; 8.1%), and antacids (n = 49; 7.8%).

Table 3.

Most consumed drug classes.

Rank	Drug class	Frequency
1	Analgesics	121 (19.2%)
2	Psychoactive drugs	89 (14.1%)
3	Antihypertensives	51 (8.1%)
4	Antacids	49 (7.8%)
5	Hormonal drugs	43 (6.8%)
6	Laxatives	42 (6.7%)
7	Antiemetics	30 (4.8%)
8	Anticoagulants	28 (4.4%)
9	Antidiabetics	24 (3.8%)
10	Antibiotics	23 (3.6%)
11	Antiasmatics	22 (3.5%)
12	Intravenous fluid therapy	19 (3.0%)
13	Antiepileptics	12 (1.9%)
14	Vitamins/food supplements	10 (1.6%)
15	Antidyslipidemics	8 (1.2%)
16	Antianemics	7 (1.1%)
17	Anti-urinary retention drugs	6 (1.0%)
18	Hemostatics	6 (1.0%)
19	Antispasmodics	5 (0.7%)
20	Vasodilators	5 (0.7%)
21	Bisphosphonates	4 (0.6%)
22	Antidiarrheals	3 (0.5%)
23	Antiglaucoma drugs	3 (0.5%)
24	Antiparkinsonians	3 (0.5%)
25	Oxygen	3 (0.5%)
26	Others (less than 2 drugs)	15 (2.4%)
	Total	631 (100.0%)

Polypharmacy

A total of 61 patients (85.9%) consumed 5 drugs or more at the time of the first appointment at the palliative care unit. In all, 19 patients (26.8%) consumed 10 drugs or more.

After the first consultation in the palliative care setting, 178 drugs (28.2%) were suspended. The most frequently suspended medications were psychoactive drugs (n = 24; 13.5% of all the suspended drugs), analgesics (n = 22; 12.4%), laxatives (n = 17; 9.6%), antiasmatics (n = 14; 7.9%), antihypertensives (n = 14; 7.9%), anticoagulants (n = 13; 7.3%), antibiotics (n = 12; 6.7%), antacids (n = 12; 6.7%), IV fluid therapies (n = 11; 6.2%), and vitamins/food supplements (n = 7; 3.9%).

There was no significant decrease in the mean number of medications before and after the palliative care first consultation (Figure 1). Yet, when only regular dose medications were considered into the analysis (and not rescue dose medications), there was a significant decrease (t = 3.512; p < 0.05) in the mean number of drugs consumed before (M = 7.15; SD = 3.64) and after (M = 5.73; SD = 3.66) the palliative care intervention (Figure 2).

Medication appropriateness

Regarding the questions from MAI, the findings are as follows (cf. Tables 4 and 5):

Question 1. Approximately one quarter of the drugs (n = 145; 23.0%) were found as not having a clinical indication in the palliative care setting (adjusted to the patient’s condition and expected survival). There were statistically significant differences as to the different kinds of medications on the Kruskal–Wallis test (p < 0.05). By applying the Tukey test to the most inappropriate class (hemostatic drugs), it is possible to verify that these were considered significantly more inappropriate than analgesics, antidiabetics, antihypertensives, antacids, antiemetics, antiepileptics, anti-urinary retention drugs, hormonal drugs, and psychoactive drugs (all p < 0.01).

Question 2. Of the drugs, 7.2% (n = 35) were found as not being effective to fulfill their indication. There were statistically significant differences between the different drugs’ appropriateness (Kruskal–Wallis p < 0.05) in this regard, but antidyslipidemics (the most ineffective drugs) were not statistically different than the other drugs on Tukey test (p < 0.05).

Question 3. Dosage/posology inappropriateness was found in 9.9% (n = 48) of the drugs. Again, the difference of appropriateness between the drugs was statistically significant on Kruskal–Wallis (p < 0.05), but the most inappropriate class of drugs in this question (Antianemics) was not significantly different than the rest on Tukey.

Question 4. The instructions in the prescriptions were incorrect in 5.8% of the drugs (n = 28). There were no statistically significant differences between the drugs in this question (p > 0.05).

Question 5. Similarly, the prescriptions’ instructions were impractical in 6.6% of the drugs (n = 32). Again, there were significant differences between the drugs on Kruskal–Wallis (p < 0.05), even though the class with most impractical instructions (Antibiotics) was not found to be significantly more impractical than the other drugs (p > 0.05).

Question 6. Only 2.9% of the drugs (n = 18) were found by the palliative care physician as having clinically significant drug–drug interactions. The drug classes did not differ significantly (Kruskal–Wallis p > 0.05).

Question 7. Of the drugs, 11.1% (n = 70) were considered as having clinically significant drug–disease/comorbidity interactions. The drugs differed significantly between themselves (p < 0.05), but again, antidyslipidemics were not significantly more inappropriate than the rest of the drugs on Tukey test (p > 0.05).

Question 8. There was unnecessary drug duplication in 6.2% of cases (n = 39). Again, there were significant differences between drug classes (p < 0.05), even though the most inappropriate class (food supplements) on this question did not differ from the rest of the drugs on Tukey test (p > 0.05).

Question 9. The duration of therapy was not appropriate in 25.5% of the non-chronic drugs (n = 47). There were statistically significant differences as to the different kinds of medications (p < 0.05), antibiotics being considered significantly more inappropriate than IV fluids (p < 0.05).

Table 4.

Answers from MAI.

Question	Appropriate		Moderately appropriate		Inappropriate		Total		Missing
	n	%	n	%	n	%	n	%	n	%
1. Indication	436	69.1	50	7.9	145	23.0	631	100.0	0	0.0
2. Effectiveness	408	84.1	42	8.7	35	7.2	485	76.9	146	23.1
3. Dosage	374	77.3	62	12.8	48	9.9	484	76.7	147	23.3
4. Correct instructions	443	91.3	14	2.9	28	5.8	485	76.9	146	23.1
5. Practical instructions	436	89.7	18	3.7	32	6.6	486	77.0	145	23.0
6. Drug–drug interactions	606	96.0	7	1.1	18	2.9	631	100.0	0	0.0
7. Drug–disease interactions	540	85.6	21	3.3	70	11.1	631	100.0	0	0.0
8. Duplication	588	93.2	4	0.6	39	6.2	631	100.0	0	0.0
9. Duration	35	74.5	0	0.0	12	25.5	47	7.4	584	92.6

MAI: Medication Appropriateness Index.

Table 5.

Answers from MAI according to drug class.

Rank	Question 1: Is there an indication for the drug?				Question 2: Is the medication effective for the condition?				Question 3: Is the dosage correct?				Question 4: Are the directions correct?				Question 5: Are the directions practical?
	Drug class	MR	x ² kw	p	Drug class	MR	x ² kw	p	Drug class	MR	x ² kw	p	Drug class	MR	x ² kw	p	Drug class	MR	x ² kw	p
1	Hemostatics^a	528.5	163,013	0.000	Antidyslipidemics	362.0	52,961	0.000	Antianemics	331	40,348	0.001	Antianemics	319.8	26,222	0.064	Antibiotics	331.5	52,899	0.000
2	Food supplements	487.9			Antianemics	356.9			Food supplements	279.5			Laxatives	248.9			Antianemics	316.3
3	Antidyslipidemics	466.7			Food supplements	303.3			Hormonal drugs^b	266.3			Analgesics	244.4			IV fluid therapy	265.2
4	IV fluid therapy	455.1			Antiasmatics	302.9			Analgesics	258.7			Psychoactives	233.6			Analgesics	240.9
5	Antianemics	445.3			IV fluid therapy	280.8			Antiemetics	254.1			Antiepileptics	233.2			Laxatives	238.5
6	Antibiotics	422.5			Antibiotics	273.3			Antibiotics	232.5			Antibiotics	231.1			Antiemetics	237.3
7	Antispasmodics	401.1			Analgesics	242.9			IV fluid therapy	228.0			Hormonal drugs^b	227.0			Psychoactives	231.0
8	Antiasmatics	400.7			Anticoagulants	235.6			Antiepileptics	223.1			Antiemetics	220.5			Antacids	221.7
9	Anticoagulants	332.0			Antihypertensives	229.1			Laxatives	216.2			Antacids	218.0			Hormonal drugs^b	214.2
10	Antacids	303.9			Laxatives	227.0			Antihypertensives	215.6			Antihypertensives	217.2			Antihypertensives	213.8
11	Laxatives	293.4			Psychoactives	226.8			Antiasmatics	214.0			Antidiabetics	211.5			Antidiabetics	208.5
12	Antihypertensives	293.3			Hormonal drugs^b	217.3			Antidiabetics	212.3			Antiasmatics	211.5			Antiasmatics	208.5
13	Psychoactives	282.8			Antiemetics	204.9			Psychoactives	210.6			Anticoagulants	211.5			Anticoagulants	208.5
14	Antiemetics	273.7			Antidiabetics	196.0			Antacids	195.2			Antidyslipidemics	211.5			Antidyslipidemics	208.5
15	Antiepileptics	263.1			Antacids	196.0			Anticoagulants	176.5			Antispasmodics	211.5			Antiepileptics	208.5
16	Analgesics	234.1			Antiepileptics	196.0			Antidyslipidemics	176.5			Anti-UR	211.5			Antispasmodics	208.5
17	Hormonal drugs^b	233.5			Antispasmodics	196.0			Antispasmodics	176.5			IV fluid therapy	211.5			Anti-UR	208.5
18	Antidiabetics	232.9			Anti-UR	196.0			Anti-UR	176.5			Food supplements	211.5			Food supplements	208.5
19	Anti-UR	210
Rank	Question 6: Are there clinically significant drug–drug interactions?				Question 7: Are there clinically significant drug–disease/condition interactions?				Question 8: Is there unnecessary duplication with other drug(s)?						Question 9: Is the duration of therapy acceptable?
	Drug class	MR	x ² kw	p	Drug class	MR	x ² kw	p	Drug class	MR	x ² kw		p		Drug class	MR	x ² kw		p
1	Antispasmodics	343.7	14,974	0.664	Antidyslipidemics	407.5	55,276	0.000	Food supplements	337.4	22,498		0.043		Antibiotics^c	25.13	7.673		0.006
2	Hemostatics	334.1			IV fluid therapy	397.1			Laxatives	334.6					IV fluid therapy	17.11
3	Antidyslipidemics	322.1			Hemostatics	356.3			Antiepileptics	327.4
4	Food supplements	316.1			Anticoagulants	351.1			Hemostatics	327.4
5	Antiasmatics	312.8			Antianemics	329.1			Psychoactives	310.7
6	Antiemetics	306.0			Antihypertensives	323.7			IV fluid therapy	309.0
7	Laxatives	300.0			Food supplements	315.4			Antiemetics	307.5
8	Psychoactives	299.4			Psychoactives	307.2			Analgesics	294.7
9	Antibiotics	299.1			Antiepileptics	305.2			Antibiotics	290.6
10	Analgesics	298.4			Antibiotics	294.0			Antacids	287.7
11	Antacids	298.3			Antiasmatics	293.9			Antihypertensives	289.3
12	Anticoagulants	296.7			Hormonal drugs^b	287.7			Anticoagulants	287.5
13	Antihypertensives	291.7			Analgesics	283.0			Antidiabetics	277.5
14	Antidiabetics	286.0			Antidiabetics	279.6			Antianemics	277.5
15	Antianemics	286.0			Antiemetics	273.0			Antiasmatics	277.5
16	Antiepileptics	286.0			Laxatives	268.7			Antidyslipidemics	277.5
17	Anti-UR	286.0			Antacids	260.3			Antispasmodics	277.5
18	Hormonal drugs^b	286.0			Antispasmodics	254.0			Anti-UR	277.5
19	IV fluid therapy	286.0			Anti-UR	254.0			Hormonal drugs^b	277.5

MAI: Medication Appropriateness Index; MR: mean rank; IV: intravenous; UR: urinary retention.

Since an “inappropriate” answer has a score of 3, while an “appropriate” answer has a score of 1, drug classes with the higher MR are more inappropriate.

Hemostatics in question 1 are significantly more inappropriate than analgesics (p < 0.001), antidiabetics (p < 0.001), antihypertensives (p < 0.001), antacids (p < 0.001), antiemetics (p < 0.001), antiepileptics (p < 0.001), anti-urinary retention drugs (p < 0.001), hormonal drugs (p < 0.001), laxatives (p < 0.001), and psychoactive drugs (p < 0.001) on Tukey test.

Hormonal drugs include corticosteroids.

Antibiotics in question 9 are significantly more inappropriate than IV (p = 0.006).

Discussion

Therapeutic futility consists in every intervention that is unlikely to be of any benefit to a particular patient in a particular medical situation. Futile treatments increase the risks to the patient without adding substantial benefits. They are a grave sign of malpractice.

Even so, there is still a paucity of studies about therapeutic futility, namely, at the time of the transition to the palliative care paradigm. At this transition, healthcare interventions are mainly focused at the patient’s quality of life, that is, at reducing unnecessary suffering. So, futility should be particularly avoided at this time.

In our study, polypharmacy (usually defined as consuming more than 5 drugs^1,9) was a very common finding (85.9% of the patients) at the time of transition to palliative care, when prescriptions should focus on symptomatic relief, aiming at therapeutic simplicity. Even when a threshold of 10 drugs was used, there were still 26.8% of patients who consumed such amount.

More than one quarter of the drugs (28.2%) was suspended by the palliative care physician at the first appointment. A preliminary analysis showed no significant difference between the mean number of medications taken before and after the consultation (Figure 1). This finding agrees with the existing literature on this regard.^2,10 However, it should be taken into account that palliative care intervention is not limited to minimize futile medication, but also promotes a shift to a more comprehensive therapeutic plan, which includes prescriptions of its own. Furthermore, many of the palliative care drugs are prescribed as rescue medication, not as regular, allowing the patients and their caregivers to act in the presence of new (albeit predictable) symptoms or in the case of loss of control from previously known symptoms. So, when all rescue medications were excluded, a significant reduction in the number of medications was indeed observed (Figure 2).

Psychoactive drugs and analgesics were the most suspended drug classes. However, they were also the most prescribed, signaling an adjustment to a more effective and secure care plan.

Approximately 23% of all the drugs did not have any clinical indication in the first place. As a result, these prescriptions simply increased the risk of adverse events, without adding any substantial benefit to the patient’s quality of life. This finding can be partially explained by the fact that the transition to the palliative care paradigm eliminates many drug indications that exist in the curative/preventive care setting. A multidisciplinary approach, aimed at improving the communication between the palliative care agent, the oncologist/surgeon/general practitioner, and the patient/caregiver would be helpful in ensuring a smoother transition in the care plan.

A high percentage of inappropriateness concerned the drugs’ adverse interactions with the disease, comorbidities or general status of the patient (11.1% “inappropriate” answers in question 7). Other important reasons for the inappropriateness of a drug were related with the dosage/posology of the drug (9.9% “inappropriate” in question 3) or ineffectiveness of the medication (7.2% in question 2). A regular revision of all the medications, in accordance with the patient’s status and goals, is paramount to avoid all these situations.

Medication revision should, therefore, (a) identify potentially inappropriate medications at the end of life, (b) eliminate medications of diminishing value, and (c) add palliative preventive and therapeutic drugs.

Even though there were many questions in which drug classes differed significantly inappropriateness between themselves (questions 1, 2, 3, 5, 7, 8, and 9), only in 2 questions (1 and 9) did the most inappropriate drug class differ significantly from others. Hemostatics were significantly more inappropriate regarding the existence of an indication regarding, among others, analgesics, antidiabetics, and antihypertensives. Antibiotics had significantly more inappropriate durations of therapy than IV fluid therapy.

Limitations of the study

This study used a freely translated and modified version of the MAI, which is yet to be validated in the Portuguese language. Furthermore, the MAI has been developed and validated for the geriatric population, not the oncologic or palliative populations. This study, however, enrolled non-geriatric oncologic patients (Table 2). Also, in order to completely grasp the distinctive particularities of the palliative population, some modifications to the MAI (namely, question 9 and the reliance of questions 2–5 on the answer to question 1) were needed. A specific questionnaire should be developed, if a full analysis of therapeutic futility at the palliative care setting is to be properly carried out.

The questions from the MAI, on the other hand, were answered by palliative care physicians, reflecting their opinions on which medications were futile or not. The other physicians involved in the prescriptions before the palliative care transition were not contacted to ascertain the reasons why they prescribed those drugs. This study does not intend to be an objective marker of therapeutic futility, but solely to demonstrate that many medications are considered futile by the palliative care specialist (according to his/her experience on palliative care) at the time of the patient’s first appointment. Multidisciplinarity and proper communication between different medical specialties should be encouraged to avoid these occurrences.

Finally, this study is not multicentered. The sample belongs to a single palliative care unit of a Portuguese oncological hospital. Caution should be exercised in extrapolating these data to other settings. More studies like this should be carried out in other countries and institutions so that the medical community may truly understand the scope and extent of therapeutic futility across the world.

Awareness campaigns and medical consensuses are possible solutions to this problem, but they mandate a profound evaluation of this phenomenon at each hospital if they are to be effective. Only by knowing how therapeutic futility differs (either quantitatively or qualitatively) from one hospital to the other, or from one country to the other, will physicians be able to grasp what the most prevalent prescription errors are in their region and act upon them.

To do this, appropriate tools and questionnaires should be developed specifically to the palliative care population. Multidisciplinarity and proper communication between the palliative care physicians, other medical specialties, and patients/caregivers is also paramount if the problem of therapeutic futility is to be effectively tackled at this crucial time of the patients’ lives.

Conclusion

This study showed that polypharmacy is the rule at the time of transition to palliative care. After a palliative care intervention, the amount of regular dose drugs was significantly reduced and the number of rescue medications increased.

The principal causes of inappropriateness were unacceptable duration of therapy (in non-chronic medications), absence of clinical indication, clinically significant drug–disease/comorbidity interactions, and incorrect dosage/posology.

More studies about this theme are needed, if the problem of futility is to be acknowledged and engaged.¹

Footnotes

Declaration of conflicting interests

The authors have no conflict of interest to declare.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Maddison

Fisher

Johnston

. Preventive medication use among persons with limited life expectancy. Prog Palliat Care 2011; 19(1): 15–21.

Riechelmann

Krzyzanowska

Zimmermann

. Futile medication use in terminally ill cancer patients. Support Care Cancer 2009; 17: 745–748.

Fede

Miranda

Antonangelo

. Use of unnecessary medications by patients with advanced cancer: cross-sectional survey. Support Care Cancer 2011; 19(9): 1313–1318.

Suhrie

Hanlon

Jaffe

. Impact of a geriatric nursing home palliative care service on unnecessary medication prescribing. Am J Geriatr Pharmacother 2009; 7(1): 20–25.

Hanlon

Schmader

Samsa

. A method for assessing drug therapy appropriateness. J Clin Epidemiol 1992; 45: 1045–1051.

Prontuário Terapêutico on-line. https://www.infarmed.pt/prontuario/index.php (accessed in 2012).

Hill

. Investigação por Questionário. Lisboa: Edicões Sílabo, 2005.

Tanis

Hogg

. Brief course in mathematical statistics. Upper Saddle River, NJ; New York: Pearson Prentice Hall, 2008.

Viktil

Blix

Moger

. Polypharmacy as commonly defined is an indicator of limited value in the assessment of drug-related problems. Br J Clin Pharmacol 2007; 63(2): 187–195.

10.

Koh

Koo

. Polypharmacy in palliative care: can it be reduced? Singapore Med J 2002; 43(6): 279–283.