Medication Discontinuation at the End of Life: A Questionnaire Study on Physicians' Experiences and Opinions

Abstract

Background:

Many patients who are in the last phase of life use multiple medications that are continued until shortly before they ultimately die. Little is known about physicians' opinions and experiences regarding medication discontinuation at the end of patients' lives.

Objective:

To explore physicians' opinions and experiences regarding medication discontinuation during the last phase of life, and to identify factors influencing the continuation of potentially inappropriate medications.

Design:

Questionnaire study, including a vignette and several statements.

Setting/Subjects:

A random sample of general practitioners and clinical specialists working in three regions in the Netherlands that represent more than half of the Dutch population.

Results:

Questionnaires were returned by 321 physicians (response rate: 37%). The majority of them (73%) agreed with the statement that patients who are in the last phase of life use too many medications. When presented with a vignette of a patient with end-stage COPD with different limited life expectancies, preventive medications would be stopped right early before death. Medications to relieve symptoms would not be stopped or stopped right before death. For medications to treat (chronic) illnesses, there was a huge interphysician variability. All statements about possible reasons why medications are continued in the last phase of life, for example, patients feeling abandoned and lack of time, were agreed upon by a small minority of respondents.

Conclusions:

Although physicians agree that patients use too many medications at the end of life, they quite regularly seem to give patients medications for chronic diseases, for which the benefit at the end of life may be debatable. More scientific evidence on whether or not these types of medication might be discontinued in the last phase of life is needed.

Introduction

Many patients who are in the last phase of life use multiple medications that are continued until shortly before they ultimately die.^1–5 On the one hand, this seems inevitable, because patients' symptom load generally increases when their death approaches.^6,7 Furthermore, patients who are in the last phase of their life often have multiple comorbidities for which different medications are needed.⁸ On the other hand, potential disadvantages may outweigh the benefits of several types of medication in the last stage of life.⁹ In practice, however, such potentially inappropriate medications (PIMs) are often continued until death.^4,10

There is increasing awareness that medications at the end of life should be reconsidered.^11,12 The scarce literature on medication discontinuation in the final phase of life suggests that when a patient has a life expectancy of only months to weeks, physicians should especially reconsider medications that are not aimed at symptom relief.^12,13 However, the clinical effects of the discontinuation of many PIMs are not well known.^11,14

There is little information about physicians' opinions regarding medication discontinuation at the end of patients' lives. In a previous interview study we identified several reasons why physicians may fail to stop PIMs, including lack of priority and fear of negative medical consequences.¹⁵ Until now, there is a lack of generalizable quantitative data about the opinions of physicians regarding medication discontinuation in the last phase of life.

The purpose of this study was to determine physicians' opinions on and experiences with medication discontinuation during the final phase of life, and to identify factors influencing the continuation of PIMs at the end of a patient's life.

Materials and Methods

Study design and participants

We conducted a questionnaire study among general practitioners and clinical specialists. We randomly selected physicians from a database (Cegedim) of addresses of physicians. Eligibility criteria were: (1)

physicians working in the region of Amsterdam, Nijmegen, or Rotterdam. These regions represent more than half of the Dutch population;

(2)

physicians working as general practitioners or as clinical specialists in geriatrics, cardiology, pulmonology, medical oncology, or neurology.

The selection contained 500 general practitioners and 100 physicians from each clinical specialty. The invitation letters and paper questionnaires were sent in May 2014. A web-based reminder was sent by e-mail to nonrespondents. Recruitment was completed in February 2015.

Questionnaire

We developed a questionnaire based on insights from a literature search on this topic. The questionnaire consisted of three sections. Section one included questions on the respondents' characteristics. Section two included a vignette about a patient with multimorbidity (Box 1). In the first part of the vignette, the patient's life expectancy was unspecified; subsequently it was added that the patient's life expectancy was three months and one week, respectively. For each situation, respondents were asked whether or not they would discontinue medication, and, if yes, which types of medication. Respondents were invited to comment on their answer. Section three of the questionnaire comprised several statements about medication discontinuation for patients with a life expectancy of three months or less. Respondents were asked whether they agreed with each statement on a 5-point Likert scale: strongly disagree, disagree, neutral, agree, and strongly agree.

Box 1: Vignette

Mrs. Bruin is an 88-year-old woman with COPD Gold stage III–IV. A week ago she came back from hospital, where she had been treated for an exacerbation of her COPD, the third admission within two months. In addition to COPD, her medical history includes a cerebrovascular accident (two years earlier), hypertension, diabetes mellitus type 2, and severe arthralgia. During the most recent hospital admission she was also diagnosed with deep venous thrombosis in her left lower leg. Besides shortness of breath, she has no other complaints. Her blood pressure is 135/70 mmHg and her blood glucose level is 6.2 mmol/L (112 mg/dL).

Her medication list includes salmeterol/fluticasone 50/500 μg BID one inhalation; prednisolone QD 5 mg; enalapril QD 20 mg; hydrochlorothiazide QD 12.5 mg; metformin TID 500 mg; paracetamol TID 1000 mg; simvastatin QD 40 mg; carbasalate calcium QD 100 mg; omeprazole QD 40 mg; acenocoumarol.

We coded the questionnaires to guarantee anonymity. The study was approved by the Medical Ethics Review Committee from the VU University Medical Center.

Statistical analyses

Frequencies, proportions, and mean with standard deviations were calculated where appropriate. To examine differences between subgroups we used chi-square tests and McNemar's test. p Values of less than 0.05 were considered to indicate statistical significance. We recoded responses to the statements into disagree, neutral, and agree. We used the Statistical Package for Social Sciences (SPSS), version 24.0, for all analyses.

Results

Respondent characteristics

In total, 122 of the selected physicians had retired from working, did not work as a physician, or could not be reached because the right contact address was lacking. The final sample therefore included 878 physicians. Three hundred twenty-one physicians completed the questionnaire, resulting in a response rate of 37%. Of these physicians, 174 (54%) were general practitioners and 147 (46%) were clinical specialists (Table 1).

Table 1.

Background Characteristics of Respondents

	General practitioners		Clinical specialists		Total
	N	%	N	%	N	%
All	174	100	147	100	321	100
Age^a (years)
<40	124	71	50	34	174	54
40–49	27	16	36	25	63	20
≥50	23	13	60	41	83	26
Gender
Male	50	29	89	61	139	43
Female	124	71	58	39	182	57
Specialty
Medical oncologist			33	22
Geriatrician			35	24
Cardiologist			25	17
Pulmonologist			25	17
Neurologist			29	20
Working experience (years)^a
0–9	137	80	66	45	203	64
10–19	13	7.6	43	29	56	18
20–29	12	7.0	28	19	40	13
≥30	9	5.3	9	6.2	18	5.7

Missing values range from 1 to 10 physicians.

Vignette

Table 2 shows the preferences of respondents with regard to the discontinuation of medications for the patient presented in the vignette. For patients with a limited but unspecified life expectancy, a majority of respondents preferred stopping the cholesterol-inhibiting drug (simvastatin) (71%) and the anticlotting drug (carbasalate calcium) (62%). Most frequently mentioned reasons to stop cholesterol-inhibiting drugs were medical futility because of the patient's limited life expectancy and potential side effects of the medication.

Table 2.

Physicians' Opinions on Medication Discontinuation for a Vignette of a Patient with a Limited Life Expectancy

	Patient's life expectancy is unspecified, N (%)	Patient's life expectancy is three months, N (%)	p^a,b (N = 321^c)	Patient's life expectancy is one week, N (%)	p^d (N = 321^c)
Physician would prefer to stop
Acenocoumarol	43 (13)	85 (26)	0.000	253 (79)	0.000
Salmeterol/fluticasone	3 (0.9)	3 (0.9)		39 (12)	0.000
Prednisolone	16 (5.0)	26 (8.1)	0.006	100 (31)	0.000
Enalapril	85 (26)	158 (49)	0.000	285 (89)	0.000
Hydrochlorothiazide	123 (38)	191 (60)	0.000	278 (87)	0.000
Metformin	57 (18)	104 (32)	0.000	224 (70)	0.000
Omeprazole	31 (10)	68 (21)	0.000	163 (51)	0.000
Simvastatin	227 (71)	289 (90)	0.000	293 (91)	0.344
Carbasalate calcium	199 (62)	239 (74)	0.000	280 (87)	0.000
Paracetamol	14 (4.4)	21 (6.5)	0.016	59 (18)	0.000
None of the above	34 (11)	15 (4.7)	0.005	2 (0.6)	0.000

McNemar.

Comparison between uncertain life expectancy with life expectancy of three months.

One missing value.

Comparison between life expectancy of three months versus one week.

When the life expectancy of the patient was stated to be about three months, 90% of the respondents preferred stopping the cholesterol inhibitor and 74% preferred stopping the anticlotting drug. In addition, antihypertensive treatment (enalapril and/or hydrochlorothiazide) would be stopped by 70% of the respondents, 26% would stop anticoagulant therapy (acenocoumarol). Most frequently mentioned reasons to discontinue enalapril and hydrochlorothiazide or acenocoumarol were medical futility because of the patient's limited life expectancy and potential side effects of the medications. Another reason to discontinue enalapril and hydrochlorothiazide was that the patient's blood pressure was already low enough.

When the patient's life expectancy was stated to be limited to one week, the physicians preferred stopping the majority of medications. The most frequently mentioned reason to stop medications in case of a life expectancy of one week of life was that they would not contribute to improved quality of life.

Statements about medication discontinuation in the last phase of life

The majority of the respondents (73%) agreed with the statement that patients who are in the last phase of life use too many medications (Table 3). In addition, most of the respondents (79%) agreed that patients' views about possible discontinuation of drugs that are no longer medically necessary are very important to them.

Table 3.

Physicians' Views on Medication Discontinuation at the End of Life

	N (%)^a,b
	321 (100)
	Disagree (%)	Neutral (%)	Agree (%)
Patients with a limited life expectancy use too many medications	37 (12)	49 (15)	235 (73)
Evaluating a patient's use of medications together with the patient provides me with a good opportunity to discuss his/her impending death	86 (27)	110 (35)	121 (38)
Patients' views about the discontinuation of medications that are not medically necessary are very important to me	19 (6.0)	47 (15)	251 (79)
Relatives' views about the discontinuation of medications that are not medically necessary are very important to me	67 (21)	100 (32)	150 (47)
Nurses' views about the discontinuation of medications that are not medically necessary are very important to me	101 (32)	108 (34)	107 (34)

Missing values range from 0 to 5 physicians (0.0%–1.6%).

All statements concern patients with a life expectancy of less than three months.

Experiences with medication discontinuation in the last phase of life

In Table 4, respondents' experiences with medication discontinuation in patients with a limited life expectancy are presented. Virtually all respondents stated that they regularly check if symptom treatment is still adequate for patients in the last phase of life (96%); about 80% stated that they regularly check which drugs may be discontinued.

Table 4.

Experiences of Physicians with Medication Discontinuation at the End of Life

	N (%)^a,b
	321 (100)
	Disagree, N (%)	Neutral, N (%)	Agree, N (%)
What is your general policy for patients with a limited life expectancy
I check regularly which medications may be discontinued because of a patient's limited life expectancy	14 (4.4)	50 (16)	255 (80)
I check regularly if symptom treatment is still adequate	4 (1.3)	10 (3.1)	305 (96)
I regularly discontinue a medication in patients with a limited life expectancy if
That medication has no therapeutic benefit	6 (1.9)	21 (6.6)	292 (92)
Side effects may be expected more frequently	26 (8.2)	59 (18)	234 (73)
The patient is not capable to take his/her medication	6 (1.9)	23 (7.2)	290 (91)
Costs of that medication are not proportional to the expected benefits	268 (84)	0	51 (16)
I consider discontinuation of a certain medication in patients with a limited life expectancy if
Limited life expectancy is sufficiently certain	18 (5.6)	48 (15)	253 (79)
Treatment goals have changed (i.e., from curative to palliative)	12 (3.8)	37 (12)	270 (85)
Side effects occur	7 (2.2)	10 (3.1)	302 (95)
The patient feels burdened by the number of medications he/she has to take	7 (2.2)	36 (11)	276 (87)
If I propose to stop certain medications to a patient with a limited life expectancy, it regularly happens that
The patient accepts this proposal	3 (.9)	25 (7.9)	290 (91)
The patient asks for explanation before he/she will accept this proposal	29 (9.1)	91 (29)	198 (62)
The patient does not accept this proposal	244 (83)	39 (13)	10 (3.4)
I continue medications more regularly than I discontinue them because
It is not known what the consequences of discontinuation are	177 (56)	90 (28)	50 (16)
It is difficult for me to talk with patients about their impending death	276 (87)	31 (9.8)	10 (3.2)
Discontinuation may give the patient the feeling that he/she is abandoned	209 (66)	66 (21)	42 (13)
Discussing medication withdrawal may result in patient's emotional reactions	255 (80)	39 (12)	23 (7.3)
Due to lack of time, I am often not able to evaluate the list of medications	174 (55)	76 (24)	67 (21)

Missing values range from 2 to 28 physicians (0.6% to 8.7%).

All statements below devote to patients with a limited life expectancy (i.e., less than three months).

All statements about possible reasons why medications are continued in the last phase of life were agreed upon by a small minority of respondents.

Discussion

The results of our study suggest that physicians believe that patients use too many medications at the end of life. In a vignette of a patient with a limited life expectancy many physicians preferred stopping medications, particularly preventive medications. However, preferences varied widely: some physicians indicated that they would continue some medications for which potential disadvantages may outweigh the benefits regarding patient's limited life expectancy.

Discontinuing medications in a vignette

We found that in a vignette, physicians preferred stopping several medications if the patient's life expectancy was limited, which is in line with recommendations for medication discontinuation in the final phase of life.^11,12 The vignette made clear that there are roughly three groups of medications that would be stopped at different times before death. First, preventive medications, for example, statins, which would be stopped relatively early. Second, medications patients frequently use for a long time to treat (chronic) illnesses, for example, metformin and enalapril. We found a huge interphysician variability in preferences regarding if and when such medications should be stopped. Third, medications that are prescribed to alleviate symptoms. Physicians preferred not stopping these medications or to stop them right before death. Especially for the second type of medication, more scientific evidence on whether or not they might be discontinued is needed. Until now, physicians have to rely on a handful of so-called implicit criteria that state if and when medications should be reconsidered, based on, for example, their potential benefit compared with the risk of current or future harm.^12,13 In addition to such implicit guidelines, an expert opinion-based guideline is available that provides explicit guidance on how to proceed when patients with diabetes mellitus are in the final phase of life.¹⁶ This guideline states that if a patient's life expectancy is only weeks to months, metformin can be discontinued, as it barely reduces glucose levels and is especially aimed at preventing long-term complications.

Reasons for continuing medications

The finding that physicians indicated that they prefer stopping medications is in contrast with the reality of patients continuing medications up until death.^1–5 Only a small minority of physicians recognized some potential drivers of continuing potentially futile or harmful medications in the last phase of life. It thus remains unclear why such continuation is common practice. Possibly, continuation of PIMs is not considered or recognized as a problem in clinical practice. This lack of awareness also emerged from our interview study.¹⁵

Limitations

Our study has some limitations. First, the response was low and therefore there is a possibility of nonresponse bias. Second, there is a chance of social desirability bias, that is, that respondents have answered questions based on their views of what is socially most acceptable. Third, the vignette may contain too little clinical information to guide the respondents in their decision-making preferences. For example, laboratory test results (such as blood glucose levels) were not available.

Conclusion

Although physicians are willing to discontinue medications, they often seem to give their patients medications for which the benefit at the end of life may be debatable. Lack of awareness could be an important reason for continuing medications in clinical practice.

Footnotes

Acknowledgment

The authors would like to acknowledge all respondents.

Author Disclosure Statement

This research study was funded by a grant from the ZonMw, the Netherlands Organization for Health Research and Development (grant number 1151.0036). The authors declare that they have no conflict of interest.

References

McNeil

, Kamal

, Kutner

, et al.: The burden of polypharmacy in patients near the end of life. J Pain Symptom Manage, 2016; 51:178–183.e2.

Morin

, Vetrano

, Grande

, et al.: Use of medications of questionable benefit during the last year of life of older adults with dementia. J Am Med Dir Assoc, 2017; 18:551.e1–551.e7.

Todd

, Husband

, Andrew

, et al.: Inappropriate prescribing of preventative medication in patients with life-limiting illness: A systematic review. BMJ Support Palliat Care, 2017; 2:113–121.

van Nordennen

, Lavrijsen

, Heesterbeek

, et al.: Changes in prescribed drugs between admission and the end of life in patients admitted to palliative care facilities. J Am Med Dir Assoc, 2016; 17:514–518.

Arevalo

, Geijteman

, Huisman

, et al.: Medication use in the last days of life in hospital, hospice and home settings in the Netherlands. J Palliat Med, 2018; 2:149–155.

Chaudhry

, Murphy

, Gahbauer

, et al.: Restricting symptoms in the last year of life: A prospective cohort study. JAMA Intern Med, 2013; 173:1534–1540.

Kehl

, Kowalkowski

: A systematic review of the prevalence of signs of impending death and symptoms in the last 2 weeks of life. Am J Hosp Palliat Care, 2013; 30:601–616.

Nobili

, Licata

, Salerno

, et al.: Polypharmacy, length of hospital stay, and in-hospital mortality among elderly patients in internal medicine wards. The REPOSI study. Eur J Clin Pharmacol, 2011; 67:507–519.

Lipska

, Ross

, Miao

, et al.: Potential overtreatment of diabetes mellitus in older adults with tight glycemic control. JAMA Intern Med, 2015; 175:356–362.

10.

Mullvain

, Kozak

, Moody

, et al.: Statin use in cancer patients with brain metastases: A missed communication opportunity at the end of life. Support Care Cancer, 2015; 23:2643–2648.

11.

LeBlanc

, McNeil

, Kamal

, et al.: Polypharmacy in patients with advanced cancer and the role of medication discontinuation. Lancet Oncol, 2015; 16:e333–e341.

12.

Scott

, Hilmer

, Reeve

, et al.: Reducing inappropriate polypharmacy: The process of deprescribing. JAMA Intern Med, 2015; 175:827–834.

13.

Holmes

, Hayley

, Alexander

, et al.: Reconsidering medication appropriateness for patients late in life. Arch Intern Med, 2006; 166:605–609.

14.

Lindsay

, Dooley

, Martin

, et al.: Reducing potentially inappropriate medications in palliative cancer patients: Evidence to support deprescribing approaches. Support Care Cancer, 2014; 22:1113–1119.

15.

Geijteman

, Dees

, Tempelman

, et al.: Understanding the continuation of potentially inappropriate medications at the end of life: Perspectives from individuals and their relatives and physicians. J Am Geriatr Soc, 2016; 64:2602–2604.

16.

End of Life Diabetes Care: Full Strategy Document Commissioned by Diabetes UK. Second Edition. 2013. http://www.diabetes.org.uk/end-of-life-care (last accessed January 13, 2018).