Vol. 43 No. 6 June 2012 Journal of Pain and Symptom Management 993

Original Article

The Prevalence of Bowel Problems Reportedin a Palliative Care PopulationKatherine Clark, MB BS, MMed, FRACP, FAChPM, Joanna M. Smith, BPsych,and David C. Currow, BMed MPH, FRACPDepartment of Palliative Care (K.C.), Calvary Mater Newcastle, Waratah, and School of Medicine

and Public Health (K.C.), University of Newcastle, Callaghan, New South Wales; Silver Chain

Nursing Association (J.M.S.), Perth, Western Australia; and Discipline, Palliative and Supportive

Services (D.C.C.), Flinders University, Daw Park, South Australia, Australia

Abstract

Context. Constipation and other disturbances of bowel function are distressing

problems for people with specialist palliative care needs. Recent observationssuggest that such problems may worsen as people become more unwell, but thechanges in intensity over time are not well documented.

Objectives. The objectives of this work were to understand the prevalence,intensity, and progression of self-reported bowel disturbances across a communitypalliative care population, which included people with cancer and noncancerdiagnoses.

Methods. All people referred to a community-based palliative care service overa period of 6.3 years had their bowel problem scores reported, using a numericalrating score at every clinical encounter until their death, at four discrete timepoints, namely, 90, 60, 30, and seven days before death. This allowed change overthree time periods to be considered. At the same time, other symptom scores werecollected including nausea, fatigue, pain, appetite problems, and breathingproblems. Patients were categorized according to the underlying disease thataccounted for their referral to palliative care, namely, cancer diagnoses (uppergastrointestinal cancers, lower gastrointestinal cancers, cancers of the associateddigestive organs, and other cancers) and nonmalignant diagnoses. Groupdifferences over the time periods were assessed using analysis of variance. Bivariateanalysis was used to explore the relationship between bowel disturbances andother symptoms using Spearman’s Rho correlation.

Results. For 7772 patients, data were collected an average of 22.5 times,generating 174,783 data collection points over an average of 98.6 days on theservice. At the time of referral to the service, 3248 (42.4%) people had disturbedbowel scores, 548 (7.2%) of whom described these as severe. Only 1020 (13.1%)people never described disturbed bowel function over their time in palliative care.At each time point, approximately one-third were experiencing disturbed bowelfunction, with proportionally greater numbers of people experiencing more

Address correspondence to: Katherine Clark, MB BS,Department of Palliative Care, Calvary MaterNewcastle, Edith Street, Waratah, New South

Wales 2298, Australia. E-mail: katherine.clark@calvarymater.org.au

Accepted for publication: July 12, 2011.

Crown Copyright � 2012 Published by Elsevier Inc. on behalfof U.S. Cancer Pain Relief Committee. All rights reserved.

0885-3924/$ - see front matterdoi:10.1016/j.jpainsymman.2011.07.015

994 Vol. 43 No. 6 June 2012Clark et al.

significant problems as death approached (X2 (9)¼ 119.3; P< 0.001). Mostreferrals to the service were because of cancer diagnoses, with no significantdifferences noted between the bowel disturbance scores of those with cancerdiagnoses compared with those with nonmalignant disease. Associations betweenbowel problem score and appetite problems, nausea, breathing problems, fatigue,and pain were explored. Although weak, there were statistically significantassociations between all symptoms and bowel problem scores except for breathingproblems.

Conclusion. In conclusion, disturbed bowel function consistently remainsa problem for people under the care of palliative care services, with the proportionof people with severe problems increasing as death approaches. This is despitethe time and number of interventions currently used to palliate theseproblems. J Pain Symptom Manage 2012;43:993e1000. Crown Copyright � 2012Published by Elsevier Inc. on behalf of U.S. Cancer Pain Relief Committee. All rights reserved.

Key Words

Bowel problems, prevalence, intensity, palliative care

IntroductionThe prevalence of constipation is quoted to

be between 30% and 90% in palliative carepatients, with such wide ranges likely to bereflective of several issues including the lackof agreed definition of constipation betweenhealth professionals and patients,1 and thefact that most reports do not define the stageof illness that people are inwhendata collectionwas undertaken. To improve the quality of infor-mation that describes bowel problems in pallia-tive care, these and other issues require greaterclarification.

Like other symptoms, the experience of con-stipation is highly personal. However, it is oftennot clear from reports as to whether the inci-dence and prevalence figures have been extra-polated from self-reporting symptom scales,criteria such as Rome criteria,2 the frequencywith which bowel movements are charted, or,most simply, thenumber of laxatives prescribed.As noted, the experience of bowel disturbancesis highly subjective, suggesting that the optimalapproach to define the scope of the problem isfrom the patient’s perspective.

Like some other symptoms, recent observa-tions suggest that, as people become sicker,bowel problems, particularly constipation, maybecomemore problematic.3,4 In some respects,this is unsurprising as the numbers of factorsthatmay contribute to bowel problems are likelyto be numerous in any individual receiving

palliative care at any one time point in their dis-ease trajectory. The evidence that underlies riskfactors for bowel problems in palliative care isrobust for opioids but less so for other factors.5

Medications with anticholinergic adverse ef-fects have been identified as increasing the like-lihood that more laxatives will be prescribed.6

Other factors include deteriorating perfor-mance status,7 reduced oral intake,8 metabolicdisturbances,9,10 and the sites of the primarycancer and metastases.Constipation and other disturbances of

bowel function have been identified as prob-lems that may lead to considerable distressfor the individual. Furthermore, bowel carehas been identified as a task that occupies sig-nificant amounts of palliative care health pro-fessionals’ time.11 These observations suggestthat the problem requires careful examinationwith well-executed studies to 1) objectively de-fine the magnitude of the problem, 2) definethe success or otherwise of symptom relief in-terventions, and 3) allow informed workforceplanning.The first aim of this study was to better un-

derstand whether self-reports of bowel dis-turbances change as death approaches, in aconsecutive cohort ofpeople cared forby a largeregional metropolitan palliative care service.The second aim was to begin exploration, usingbivariate analysis, of whether there were associa-tions between constipation and other factors

Vol. 43 No. 6 June 2012 995Bowel Problems: Prevalence and Intensity

including stage of illness and other symptoms:pain, breathing problems, nausea, and fatigue.The objectives of the study were to understandthe prevalence, intensity, and progression ofself-reported bowel problems across a commu-nity palliative care population, which includedpeople with cancer and noncancer diagnoses.

The primary null hypothesis was that therewere no differences in the way that bowel prob-lems were experienced over the time that peo-ple spent under the care of a palliative careteam. The secondary null hypothesis was thatthere is no relationship between the othersymptoms and bowel problems.

MethodsStudy Setting

Silver Chain Hospice Care Service (SCHCS)is the sole regional community palliative careprogram covering all the metropolitan areaof Perth, Western Australia. The SCHCS usesan interdisciplinary model of care provision.The team comprises general practitioners, re-gistered nurses, care aides, volunteers, coun-selors, and pastoral care workers. Registerednurses are available 24 hours per day, sevendays per week, with support available afterhours from the general practitioners and clin-ical nurse consultants. The service receivesapproximately 1500 referrals annually for palli-ative care and is free of charge to patients withlife-limiting illnesses (charges covered by stateand federal health care funds).

Study PopulationThe study used data from a consecutive co-

hort of 7772 patients seen by SCHCS over a pe-riod of 6.3 years (until April 2010). The studywas approved by the Silver Chain Human Re-search Ethics Committee.

Data Collection From SCHCS RecordsAll data were collected contemporaneously

as part of routine practice for each face-to-face visit by a health professional. Deidentifieddata used for this project included demogra-phic characteristics (age, gender, and date ofdeath), clinical data describing the primary ill-ness that accounted for referral to specialistpalliative care, and clinical data recorded ateach contact with the patient.

Patient-reported symptom scores for bowelproblems, pain, insomnia, nausea, breathingproblems, appetite problems, and fatiguewere collected using the validated SymptomAssessment Scale.12 This scale uses a numericrating scale score to report the severity ofsymptoms experienced, where 0¼ no prob-lems experienced and 10¼ the worst imagin-able problems experienced. The SymptomAssessment Scale does not provide an in-depth assessment of individual symptoms butserves as a screening tool to identify trouble-some symptoms that warrant more attentiveand immediate clinical investigation and com-prehensive assessment.13

Other clinical data were collected at con-tacts with patients that took place duringeach phase of illness. Illness phase is a validateddescriptive tool to communicate how stable orotherwise a person is in his/her disease trajec-tory; four clinical phases have been identified:stable, unstable, deteriorating, and terminal.Definitions are summarized in Table 1.14

Data AnalysesThe demographic data are reported using

descriptive statistics. To conduct a longitudinalanalysis, the time points of interest were cal-culated backward from death. This alloweddata to be analyzed at the same points beforedeath regardless of the time before death thepatients were referred to the palliative care ser-vice. This also allowed prevalence and inci-dence of reports of disturbed bowel habits tobe charted at discrete time points, permittingchanges over time to be evaluated. Four timepoints were chosen as this resulted in threediscrete time periods over which to considerchange with sufficient numbers of people aliveat each point to have data available. The dis-crete time points were 90, 60, 30, and sevendays from death, with symptom scores incorpo-rated into the analysis if 1) a visit had occurredexactly on the day that proved to be 90, 60, 30,or seven days from death, and 2) a visit hadoccurred three days either side of the daythat proved to be 90, 60, 30, or seven daysfrom death.

The intensity of bowel problems also wasplotted backward from death using a numeri-cal zero to 10 score where 0¼ no symptomsand 10¼ the worst imaginable problem. Theintensity was substratified into four levels:

Table 1Phase Definitions14

Phase Type Phase Description

Stable Symptoms are adequately controlled by established management.Further interventions to maintain symptom control and quality of life have been planned.The situation of the family/carers is relatively stable, with no new issues apparent.

Unstable The patient experiences the development of a new unexpected problem or a rapid increase in the severityof existing problems, either of which requires an urgent change in management or emergency treatment.

The family/carers experience a sudden change in their situation requiring urgent intervention by membersof the multidisciplinary team.

Deteriorating The patient experiences a gradual worsening of existing symptoms or the development of new butexpected problems.

These require the application of specific plans of care and regular review but not urgent or emergencytreatment.

The family/carers experience gradually worsening distress and other difficulties, including social and practicaldifficulties, as a result of the illness of the person.

Terminal Death is likely in a matter of days and no acute intervention is planned or required.Typically in this phase, people are weak, essentially bedbound, drowsy for extended periods, disoriented fortime and have a severely limited attention span, increasingly disinterested in food and drink, often findingit difficult to swallow medication.

This requires the use of frequent, usually daily, interventions aimed at physical, emotional, and spiritual issues.

996 Vol. 43 No. 6 June 2012Clark et al.

0 (none), 1e3 (mild), 4e6 (moderate), and7e10 (severe).

Group differences were assessed using anal-ysis of variance. Bivariate analysis was used toexplore the relationship between bowel prob-lem scores and other symptoms, using Spear-man’s Rho correlation. Analyses between thebowel problem score and each of the symp-toms at each time point were undertaken sep-arately. Data were analyzed using Stata 11.2software (StataCorp LP, College Station, TX).

Table 2Demographic Details

Variable Descriptive Statistics

Age at referral Mean 68.7, SD 14.1, minimum0, maximum 105

Time of referral untildeath (days)

Mean 98.6, SD 106.5,minimum 14, maximum1530

Number of visits Mean 22.5, SD 24, minimum 1,maximum 634

Gender, n (%)Female 3411 (43.9)Male 4356 (56.1)Not specified 5 (0.1)

Diagnosis, n (%)Noncancer 473 (6.1)Cancer 7299 (93.9)

ResultsDescriptive Data

All referrals to SCHCS were included exceptthose clients who had received this service forless than two weeks before death. These peo-ple were omitted from the analysis. The result-ing sample included the records for 7772clients. On average, patients were seen 22.5times between referral and death by the com-munity palliative care team, with the averagetime from referral to death being 98.6 days(SD �106.5; range 14e1530), thus generating174,783 data collection points.

Demographics are summarized in Table 2.Just over half the people referred were male,with the mean age at referral being 68.7 years(SD �14.1; range 0e104 years). Most peoplereferred had a cancer diagnosis (94%), withmost of these cancers arising from outsidethe gastrointestinal (GI) tract.

Disturbed Bowel Function at Referral toPalliative CareFrom the total population of people refer-

red to this palliative care service, at the timeof referral, 42.2% (n¼ 3248) had disturbedbowel function, as summarized by the bowelscores. Of these, 548 (7.2%) described thebowel disturbance as severe. Regardless ofthe time referred before death, people withcancers arising from the lower GI tract had sig-nificantly higher bowel problem scores on ad-mission than the other groups examined.

Prevalence and Severity of Bowel DisturbancesOver TimeOf the cohort of people who survived 90 days

while under the care of the palliative care

Fig. 1. Collated bowel problem scores over time.

Vol. 43 No. 6 June 2012 997Bowel Problems: Prevalence and Intensity

service, only 13% (n¼ 1020) consistently hadscores of zero over this period. Ninety days be-foredeath, 2274bowelproblemsymptomscoreswere collected. Of these, 26.8% (n¼ 602) wereexperiencing bowel symptoms of any severity.At 60 days before death, 28.24% (n¼ 928)were reporting bowel problems of any severity.Between 30days before death and sevendays be-fore death, the total numbers of people with anyproblems rose to the almost identical figures of33.7% (n¼ 1875) and 35.2% (n¼ 1920), re-spectively (Fig. 1).

Although the total percentages of peoplereporting any bowel disturbance remained rea-sonably consistent, closer examination of thesefigures revealed that the proportion of visitswhere people were reporting bowel distur-bances in themoderate-to-high range increased

Fig. 2. Trajectory of Symptom Assessment Scale bowel pro

as people neared death, with a smaller propor-tion of people reporting zero and a higher pro-portion of people scoring moderate-to-highscores (Χ2 (9)¼ 119.3; P< 0.001) (Fig. 1).

When only the severe bowel problem symp-tom scores were examined from individualvisits, at 90 days before death, 2.33% (n¼ 53)described such problems. At 60 days, thisrose to 2.5% (n¼ 83). This figure continuedto climb, as people being visited became weak-er, to 3.9% at 30 days and 4.1% at seven daysbefore death. This is highlighted in Fig. 2,where the trajectory of severe scores is remark-ably consistent over time. It is notable fromthis graph that the numbers of those experi-encing severe scores do not trend downwardlike the mild-to-moderate scores.

Disturbed Bowel Function by DiagnosisMost referrals to this palliative care service

were because of a cancer diagnosis. Overall,there was no significant difference in the sever-ity of bowel disturbances experienced by thosewith cancer compared with those with nonma-lignant disease.

The cancer diagnoses were broken downinto subgroups. There were cancers deemedmost likely to affect the gut, that is, cancersof GI origin and the affiliated organs of diges-tion, such as pancreatic cancers or cancers ofthe liver and biliary tree, and other cancersarising external to any part of the GI tract.

blems from 90 days until death in each category.

Fig. 3. Change in Symptom Assessment Scale bowelproblem score (range 0e10) over time by diagnosis.GI¼ gastrointestinal.

coresan

dOther

Sym

ptomsd

AllPatients

for

ion

SAS

blem

Other

dAll

Days

eath

%Exp

eriencing

Symptom

ofAny

Severity

30Days

Before

Death

PScoresfor

Correlation

Betwee

nSA

SBowel

Score

Problem

and

Other

Symptomsd

All

Patients

30Days

Before

Death

%Exp

eriencing

Symptom

ofAny

Severity

SevenDays

Before

Death

PScoresfor

Correlation

Betwee

nSA

SBowel

Problem

Scoresan

dOther

Symptomsd

All

Patients

Seven

DaysBefore

Death

39.98

0.00

a73

.40

0.00

a

24.80

0.00

a24

.07

0.00

a

54.00

0.27

60.26

0.10

95.94

0.00

a96

.30

0.00

a

77.76

0.00

a75

.58

0.00

a

998 Vol. 43 No. 6 June 2012Clark et al.

At referral to the service, it was notable thatthe group statistically most likely to have high-er bowel problem scores compared with theother cancer and noncancer diagnoses werethose recorded from people with lower GIcancers (P¼ 0.003). A similar observation wasnoted from visits made at 60 days before death.However, at 30 days and seven days beforedeath, the mean bowel problem scores forboth upper and lower GI cancer stabilized,whereas the other diagnoses’ scores continuedto rise (Fig. 3).

Table3

PSco

resforCorrelationBetwee

nSASBowel

S

Symptom

%Exp

eriencing

Symptom

ofAny

Severity

90Days

Before

Death

PScoresfor

Correlation

Betwee

nSA

SBowel

Problem

Scoresan

dOther

Symptomsd

All

Patients

90Days

Before

Death

%Exp

eriencing

Symptom

ofAny

Severity

60Days

Before

Death

PScores

Correlat

Betwee

nBowel

Pro

Scoresan

dSymptoms

Patients

60Before

D

Appetiteproblems

55.23

0.00

a59

.97

0.00

a

Nau

sea

22.06

0.00

a13

.10

0.0a

Breathingproblems

49.05

0.62

26.26

0.98

Fatigue

69.15

0.00

a94

.08

0.00

a

Pain

82.95

0.00

a83

.40

0.00

a

SAS¼Symptom

Assessm

entScale.

aP<0.00

1.

Bivariate AnalysisAssociations between bowel problem scores

and appetite problems, nausea, breathing prob-lems, fatigue, pain, and phase change wereexplored. Not surprisingly, this was a sympto-matic group of people with the most frequentlyreported symptom being fatigue, regardless ofthe time before death (Table 3).

There was a statistically significant associa-tion between each of the symptoms and bowelproblems, except for breathing problems. Al-though statistically significant, the relation-ships were weak (Table 3), with the strongestrelationships noted for appetite problemsand nausea.

As seen in Fig. 4, bowel problem scores weresignificantly higher when patients were in anunstable phase compared with the otherphases (F (3, 161,187)¼ 884.3; P> 0.001).This was apparent for both GI cancers andother diagnoses. Additionally, on average, cli-ents with GI cancers had significantly higherbowel scores at each phase except for the ter-minal phase, where clients with other diagno-ses were noted to have significantly highermean bowel disturbance scores (Fig. 4).

Fig. 4. Mean bowel problem score by phase of ill-ness. GI¼ gastrointestinal.

Vol. 43 No. 6 June 2012 999Bowel Problems: Prevalence and Intensity

DiscussionThese findings reflect the prevalence and in-

tensity of self-reported bowel problems in peo-ple under the care of a specialist palliative careteam over time. Regardless of the time beforedeath, the data presented here consistentlysuggest that approximately one-third of thepopulation suffered at least a degree of bowelproblems. Furthermore, as people becamesicker and more unstable, those who were ex-periencing bowel symptoms were more likelyto report these as more severe than thosethey had experienced at other times in theirdisease trajectory. This is concerning as it sug-gests that, despite the amount of time andnumber of interventions routinely prescribed,people continue to have bowel disturbancesymptoms. Given the detrimental effects thatsuch symptoms may have on overall quality oflife, there is an imperative to explore ways tobetter palliate these problems.

The descriptive data presented here offer lit-tle insight as to why bowel symptoms worsenrather than improve over time. It is highlylikely, however, that the reasons are multitude.Other data report that as people become moreunwell, the number of medications prescribedthat impose anticholinergic effects increa-ses.15,16 Furthermore, people become less in-dependent, less mobile, and their oral intaketends to lessen considerably.17 Whatever thecauses, it is still unclear how reversible or oth-erwise these changes in bowel function are,what interventions are likely to afford thebest palliation, and for how long cliniciansshould persist in trying to restore bowel func-tion as life shortens.

At the time of referral to palliative care ser-vices, it was interesting to note that peoplewith lower GI cancers had significantly higher

bowel scores. However, closer to death, theproblems became more pronounced in upperGI cancers. It is not possible to comment onwhether this is an inherent feature of upperGI cancers or whether this is secondary toother problems likely to be experienced bypeople with upper GI cancers such as nausea,anorexia, and weight loss. However, it was nota-ble that the data confirmed an association be-tween bowel disturbance symptoms andreports of nausea and impaired appetite.

These preliminary observations of symptomclustering around GI problems require greaterscrutiny. In particular, there is a need to ex-plore how each individual symptom in thisgroup may impact on the other symptoms.This becomes even more important when thepossibility that modifying one problem maybeneficially impact another, allowing explora-tion of alternative therapeutic approaches.The challenge when exploring this furtherwill be to understand to what degree gut dis-turbances occur in context with other symp-toms, what the impact is of the adverseeffects of medications on the development ofnew symptoms, and which symptoms occur asa result of the underlying disease that resultedin referral to palliative care.

The highest scores overall for bowel distur-bances were reported when people were re-ported as unstable. This is not surprising asunstable phases are a time where existingsymptoms worsen or new symptoms develop.However, as previously observed, it is worryingthat all these problems exist in people underthe care of teams whose main role is to mini-mize the impact of symptoms (physical, psy-chological, and spiritual) on the lives ofpeople with advanced and incurable illnesses.

StrengthsThis study was conducted with data pros-

pectively collected at each clinical encounterusing validated tools that allow people to self-report their problems. As this was routinedata collection, it allowed data to be comparedat discrete time points before death regardlessof the time before death people entered theservice. This study offers a unique longitudinalassessment of self-reported problems, whichcould then be analyzed in the framework ofthe patient’s illness phase. This not only allowsthe trends over time to be mapped but also

1000 Vol. 43 No. 6 June 2012Clark et al.

takes into account how the patient’s changingphysical status may impact the severity of symp-toms experienced.

LimitationsAlthough phase data were included in this

analysis, a measure of performance status wasnot. Whereas analyzing people within the ill-ness phase allows a degree of comparison oflike with like, including an objective measureof performance status would strengthen thisapproach. Furthermore, the data presentedhere are routine point-of-care data that donot codify potential contributing etiologies.However, the data presented here continueto strengthen the base from which other re-search to better quantify the impact of factorslikely to impair bowel function may beconducted.

Implications for ResearchThere remains a need to better understand

how the physiology of the colon and otherstructures of defecation change with progres-sive illness. There is still very little understand-ing of how much changed bowel functionreflects the use of medications and howmuch other factors such as diet and functionalstatus impact the problems that people experi-ence; therefore, it remains difficult to moreclearly define the problems. Further researchis needed to develop an unambiguous defini-tion of the problem and tailor treatments.

References1. Clark K, Urban K, Currow D. Current ap-proaches to diagnosing and managing constipationin advanced cancer and palliative care. J Palliat Med2010;13:473e476.

2. Longstretch G, Thompson W, Chey W. Func-tional bowel disorders. Gastroenterology 2006;130:1480e1491.

3. Fallon MT, Hanks GW. Morphine, constipationand performance status in advanced cancer pa-tients. Palliat Med 1999;13:159e160.

4. Clark K, Lam L, Currow DC. Exploring the rela-tionship between the frequency of documentedbowel movements and prescribed laxatives in

hospitalized palliative care patients. Am J Hosp Pall-iat Care 2011;28:258e263.

5. Davis MP. Cancer constipation: are opioids re-ally the culprit? Support Care Cancer 2008;16:427e429.

6. Clark K, Lam L, Agar M, Chye R, Currow DC.Retrospective analysis of contributing factors to lax-ative prescription in hospitalized palliative care pa-tients. Palliat Med 2010;24:410e418.

7. Fallon M. Constipation in cancer patients: prev-alence, pathogenesis, and cost-related issues. Eur JPain 1999;3(Suppl 1):3e7.

8. Leung F. Etiologic factors of chronic constipa-tion: review of the scientific evidence. Dig Dis Sci2007;52:313e316.

9. Camilleri M. Gastrointestinal problems in diabe-tes. Endocrinol Metab Clin North Am 1996;25:361e378.

10. Wu MJ, Chang CS, Cheng CH, et al. Colonictransit time in long-term dialysis patients. Am J Kid-ney Dis 2004;44:322e327.

11. Wee B, Adams A, Thompson K, et al. How muchdoes it cost a specialist palliative care unit to man-age constipation in patients receiving opioid ther-apy? J Pain Symptom Manage 2010;39:644e654.

12. Eagar K, Green J, Gordon R. An Australian case-mix classification for palliative care: technical devel-opment and results. Palliat Med 2004;18:217e226.

13. Samar M, Monteresso L, Kristjanson L,McConigley R. Measuring symptom distress in palli-ative care: psychometric properties of the SymptomAssessment Scale. J Palliat Med 2011;14:315e321.

14. Palliative Care Outcomes Collaborative(PCOC). PCOC data definitions and guidelines, ver-sion 2. Wollongong, NSW, Australia: University ofWollongong, 2009. Available from http://ahsri.uow.edu.au/content/groups/public/@web/@chsd/@pcoc/documents/doc/uow090306.pdf. AccessedMay 14, 2012.

15. Agar M, Currow D, Plummer J, et al. Changes inanticholinergic load from regular prescribed medi-cations in palliative care as death approaches. PalliatMed 2009;23:257e265.

16. Agar M, To T, Plummer J, Abernethy A,Currow DC. Anti-cholinergic load, health care utili-zation, and survival in people with advanced cancer:a pilot study. J Palliat Med 2010;13:745e752.

17. Clark K, Lam LT, Agar M, Chye R, Currow DC.The impact of opioids, anticholinergic medicationsand disease progression on the prescription of laxa-tives in hospitalized palliative care patients: a retro-spective analysis. Palliat Med 2010;24:410e418.