408 Journal of Pain and Symptom Management Vol. 30 No. 5 November 2005

Original Article

Patient Reports of Symptoms and TheirTreatment at Three Palliative Care ProjectsServicing Individuals with HIV/AIDSDaniel Karus, MS, Victoria H. Raveis, PhD, Carla Alexander, MD,Barbara Hanna, MD, Peter Selwyn, MD, MPH, Katherine Marconi, PhD,and Irene Higginson, MD, PhDMailman School of Public Health (D.K., V.H.R.), Columbia University, New York, New York; Institute

of Human Virology (C.A.), University of Maryland Medical Center, Baltimore, Maryland; Health

Services Center, Inc. (B.H.), Hobson City, Alabama; Montefiore Medical Center (P.S.), Bronx,

New York; Office of Science & Epidemiology (K.M.), HIV/AIDS Bureau, Health Resources & Services

Administration, Rockville, Maryland, USA; and King’s College (I.H.), London, United Kingdom

AbstractSelf-reports of 32 symptoms and their treatments were obtained from patients of threepalliative care programs that provide services to seriously ill HIV patients ($95% AIDS) inAlabama (n5 47), Baltimore (n5 91), and New York City (n5 117). On average, patientsreported 10.9 (SD5 7.6) to 12.7 (SD5 6.2) symptoms. Pain, lack of energy, andworrying were reported by a majority of patients at all sites, often with a high level ofassociated distress. For only four symptoms (pain, nausea, difficulty swallowing, and mouthsores) did half or more of patients at all sites experiencing the symptom also report treatment.Less than a third of patients experiencing 12 symptoms (five of six comprising a psychologicalsubscale) reported treatment. Results show that despite the availability of more efficacioustreatments, many HIV/AIDS patients continue to experience significant physical andpsychological symptomatology. Many of those experiencing symptoms, however, do not perceivetheir symptoms as being treated. J Pain Symptom Manage 2005;30:408--417. � 2005U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved.

Key WordsHIV/AIDS, palliative care, symptom evaluation, underserved populations

IntroductionThe development of more efficacious treat-

ment regimens has resulted in increased

Address reprint requests to: Daniel Karus, MS, Centerfor the Psychosocial Study of Health and Illness,Mailman School of Public Health, Columbia Uni-versity, 100 Haven Avenue, Suite 6A, New York,NY 10032, USA.

Accepted for publication: April 18, 2005.

� 2005 U.S. Cancer Pain Relief CommitteePublished by Elsevier Inc. All rights reserved.

survival and reduced incidence of various co-morbid conditions for many living with HIV/AIDS.1--2 However, there remain thousands ofindividuals for whom these regimens havebeen less successful due to a variety of factors,including initiation of treatment after the pa-tient’s immune systemhad already been severelycompromised, nonadherence to treatment regi-mens, an inability to tolerate treatment(s),and/or nonresponse to treatment.Although a large number of studies have in-

cluded symptomatology of HIV patients as

0885-3924/05/$--see front matterdoi:10.1016/j.jpainsymman.2005.04.011

Vol. 30 No. 5 November 2005 409HIV/AIDS Patient Reports of Symptomatology

either a dependent or an independent variablein analyses, most have either used only sum-mary measures (e.g., a count of symptoms re-ported) or restricted analysis and discussion toone or two specific symptoms. Several reportshave described the prevalence of a wider rangeof individual symptoms experienced by HIV/AIDS patients; however, the symptoms de-scribed remain those specifically associatedwith disease progression or a poor prognosisfor survival and are based on information frompatients living with the disease prior to theadvent of highly active antiretroviral therapy(HAART) and other treatment advances.3--4

Specifically, these studies have documenteda high prevalence of fatigue, weight loss, cough,memory loss, insomnia/sleep disturbance, de-pression/sadness, and paresthesias in thispopulation.

Not much is known about the prevalence, se-verity, and level of distress associated withsymptoms that can adversely impact the qualityof life of individuals living with HIV/AIDS,even though they may not be life threateningor associated with a poor prognosis. One studythat described the prevalence and experienceof 32 symptoms in a sample of 504 ambulatoryAIDS patients5 found that patients reported anaverage of 16.7 symptoms, with 19 or moresymptoms reported by at least half of the sub-jects. About two-thirds or more of patients re-ported each of the six symptoms comprisinga psychological subscale. However, the patientsin that study were accrued prior to the wide-spread availability of HAART.

There is reason to believe that the symptom-atology experienced by HIV/AIDS patientsmay have changed as a result of improvementsin the treatment of the disease and its asso-ciated conditions. Potential toxicities andother side effects have been associated with in-dividual treatments and there is also the poten-tial for unanticipated drug interactions asa result of polypharmacy either for treatmentof HIV/AIDS or, as is often the case, for treat-ment of HIV/AIDS in conjunction with othermedical conditions or diagnoses.6--7 The shiftin the demographics of the disease toward pa-tients diagnosed with comorbidities (e.g., sub-stance abuse or mental illness) and patientsfrom resource-poor and minority communitiesis also likely to have affected the prevalenceand impact of symptoms associated with these

factors among persons living with HIV/AIDSas a whole. The impact of this shift may be par-ticularly evident among patients with advanceddisease, given that persons with a history ofdrug abuse and a high level of homelessnesshave been found in previous studies to receivean HIV/AIDS diagnosis and to enter treat-ment at a later stage of the disease comparedto other groups.8--13

This report compares patient reports ofsymptomatology from three palliative careprograms providing services to seriously illHIV/AIDS patients. Specifically, it describesthe prevalence of symptoms at each site; theprevalence of highly problematic symptom-atology among patients experiencing a symp-tom; and patient reports of whether theyhave received treatment for each symptomexperienced.

MethodsProgram Sites

The patients described are clients of threedemonstration programs funded as SpecialProjects of National Significance through aninitiative of the HIV/AIDS Bureau of theHealth Resources and Services Administration(HRSA) in the Department of Health andHuman Services (HHS) to provide appropri-ate palliative care and services to individualswith advanced HIV/AIDS. Although all proj-ects participated in a multisite study designedto permit comparisons of programmatic expe-riences and patient outcomes, individual pro-grams were developed and delivered at eachsite. Table 1 presents a brief description ofeach project site and its programmatic goals.Similarities and differences among the siteswith regard to their programmatic elementsand client populations have been describedin earlier reports.14--15

MeasuresAll three projects collected a common set

of items and measuresdthe Palliative CareAssessment Forms for HIV (PCAF-HIV). Boththe patient and a staff member completedappropriate forms of the PCAF-HIV at accrualand at regular intervals thereafter (monthlyfor standardized measures, every 3 months

410 Vol. 30 No. 5 November 2005Karus et al.

Table 1Project Sites

Alabama Baltimore New York City

Geographiclocale/servicearea

The agency serves a 14-county,primarily rural area inNortheastern Alabama throughits main clinic in Hobson Cityand regularly scheduled satelliteclinics in other towns.

Downtown, close to areas wherethe homeless congregate andneighborhoods with high ratesof substance abuse and HIVinfection.

A densely populated urbanneighborhood of the Bronx inan area with high rates of HIVinfection.

Institutionalsetting

Service is offered through afreestanding, nonprofit clinicfor persons with HIV/AIDS.

Large university-affiliatedmedical center.

Large medical center with servicesand clinics for HIV/AIDS care.

Target clientpopulation

Terminally ill HIV/AIDS patientswho are poor and usuallyuninsured. In many instances,patients are also homelessand/or substance abusers.

Terminally ill HIV/AIDS patientswho include substance abusers,the homeless, multiplydiagnosed, and the uninsuredurban poor.

Seriously ill HIV-infected patientswho are often poor, uninsured,substance abusers, andhomeless.

Service model Hospice care. Augmented hospice model. Augmented palliative care,consultative service.

Programmaticgoals

Development of a palliative careteam to provide clients in needof such care with medical andsupport services, either at theclinic or in the patient’s home.Development of a communityhospice residence for a limitednumber of clients who areeither homeless or inadequatelyhoused.

Development of a palliative careservice to maintaindthroughactive community outreachand other effortsdan ongoingrelationship with clients thatsupports the utilization ofavailable medical and supportservices (i.e., facilitating themaking and keeping ofappointments encouragingadherence to treatmentregimens), as well as thearticulation and achievementof realistic life goals.

Development of a palliative careteam to provide both palliativecare consultative services(i.e., a resource at thedisposal of hospital staff toprovide assessments of patients’palliative service needs and toidentify appropriate referrals)and ongoing palliative supportservices to clients identifiedthrough outreach at outpatientclinics.

Palliativecare team

Physician, nurse, social worker,addictions specialist.

Physician, nurse, social worker,chaplain/pastor, addictionsspecialist, outreach worker.

Physician, psychiatrist, nursepractitioner, social worker,outreach worker, chaplain/pastor, bioethicist.

Reprinted with permission from Reference 14.

for sociodemographics and service needs/utili-zation assessments). Analyses presented arebased on data from patient interviews con-ducted from June, 2000 to March, 2003. Ap-propriate Institutional Review Board approvalwas obtained for projects at all sites. Includedin the PCAF-HIV are questions eliciting basicsociodemographic information (age, gender,race/Hispanic origin); HIV medical status(HIV/AIDS status, CD41 count, viral load);brief histories of mental illness/treatment(ever, past 3 months) and alcohol/substanceabuse/treatment (ever, past 3 months); anda battery of standardized measures.

Symptomatology. Patients completed the Me-morial Symptom Assessment Scale (MSAS),a comprehensive symptom assessment mea-sure developed by Portenoy et al.16 for use withcancer patients, which has also been used withambulatory HIV/AIDS patients,5 as well as

with AIDS inpatients receiving long-term careat a skilled nursing facility.17 The MSAS elicitsinformation regarding the presence, frequency(if relevant), severity, and distress associatedwith each of 32 symptoms during the past weekand yields a separate score for each symptomas well as four summary measures: a total score(total MSAS) (Cronbach’s alpha5 0.86--0.93in present samples); a psychological symptomssubscale (PSYCH) (alpha5 0.69--0.81); a physi-cal symptoms subscale (PHYS) (alpha5 0.79--0.82); and a global distress index (GDI) (al-pha5 0.72--0.83). Scores for all summarymeasures can range from 0 to 4, with higherscores indicative of greater problems.Based on discussions among the investi-

gators and with approval of an author of theoriginal scale (A. Kornblith, personal commu-nication), two modifications were made to theoriginal MSAS form for the purposes of thisevaluation. First, a symptom in the original

Vol. 30 No. 5 November 2005 411HIV/AIDS Patient Reports of Symptomatology

MSAS specifically associated with cancer treat-ment but not HIV/AIDS (‘‘hair loss’’) was re-placed with a symptom more relevant topatients with advanced HIV/AIDS (‘‘fever last-ing more than 2 weeks’’). This substitutiondoes not impact individual symptom scoresfor three of the summary measures (PSYCH,PHYS, or the GDI), since the symptom re-placed does not factor into the computationof these measures. Impact on the total MSASis minimal given that only 1 of the 32 symp-toms being assessed was changed.

Another modification implemented for thepurposes of the multisite evaluation, was theaddition of an item to assess whether the pa-tient was being treated for each symptom re-ported. This item was added because it wasfelt that information regarding the patient’sperceptions of whether his or her symptomswere being addressed was important, not onlyfor understanding the patient’s condition butfor identifying potential gaps in care orpatient-staff communication. Whereas atAlabama and Baltimore the MSAS was adminis-tered to patients by a nurse practitioner, inNew York City (NYC) the measure was admin-istered by a nurse practitioner, a social worker,or a community outreach worker.

Statistical MethodsOne-way analysis of variance was used to de-

termine whether means of continuous ordinalvariables, including scale scores, significantlyvaried among the three projects. When differ-ences among projects were noted, Scheffe’stests were conducted to identify significant dif-ferences between each pair of projects. Com-parisons of the distribution of categoricalvariables among sites were made using c2 tests.The criteria for significance used for all statis-tical tests were P# 0.05.

ResultsThe findings presented reflect patients’

symptomatology at the time of their first assess-ment. Of 376 patients accrued at the three sites(52 at Alabama, 94 at Baltimore, and 230 atNYC), 255 completed an MSAS. Completedforms were obtained from 90% (n5 47) ofpatients at Alabama, 97% (n5 91) of patientsat Baltimore, and 51% (n5 117) of patientsat NYC. The proportion of patients completing

an MSAS was particularly low at NYC due, atleast in part, to the consultative nature of theprogram and the accrual of many patients hos-pitalized within days or weeks of death. Forthese patients, completion of forms was unfea-sible or even unethical given the severity oftheir medical condition. All patients witha completed MSAS were HIV positive, with95% or more at all three sites meeting the cri-teria for an AIDS diagnosis. Among patientsfor whom lab results were also available, 81%at NYC, 85% at Alabama, and 98% at Baltimorenot only met the Centers for Disease Controland Prevention (CDC) criteria for an AIDS di-agnosis but also reported a CD41 count of!200 cells/mm3 at last assessment. At leasthalf of the patients at all sites had viral loadshigher than 50,000.

Patient CharacteristicsCharacteristics of patients at the three sites are

presented in Table 2. Amajority of patients at allthreesiteswere in their30sor40s.Althoughama-jority at all sites identified themselves as male,this proportion varied significantly among sites(c25 9.684, df5 2, P5 0.008) and was particu-larly high at Alabama (83%). The racial/ethnicdistribution of patients was also significantlyassociated with site (c25 117.710, df5 6,P! 0.001). Whereas at Alabama most patientswere either Black non-Hispanic (46%) orWhitenon-Hispanic (44%), at Baltimore the vast ma-jority was Black non-Hispanic (92%), and atNYC patients were primarily Hispanic (50%)or Black non-Hispanic (34%).

About half or more of patients at all threesites reported a history of substance abuse.However, this proportion varied significantlyamong sites (c25 10.487, df5 2, P5 0.005),as did the proportion of patients reportingsuch a problem in the past 3 months(c25 6.579, df5 2, P5 0.037). In both instan-ces, the proportion reporting a problem washighest at Baltimore and lowest at Alabama.It should also be noted that while staff at Ala-bama indicated that substance abuse amongtheir patients was primarily associated with al-cohol use, staff at Baltimore and NYC reportedthat although many of their patients abused al-cohol, a high proportion also had histories ofabusing other illicit substances, often intra-venously. The proportion of patients with a his-tory of substance abuse who had ever received

412 Vol. 30 No. 5 November 2005Karus et al.

Table 2Patient Characteristics by Project Site

Alabama(n# 47)

Baltimore(n# 91)

New York City(n# 117)

P-valueMean (SD) % Mean (SD) % Mean (SD) %

Age 40.5y (8.8) 40.9z (7.2) 44.4y,z (8.9) P# 0.01!30 8 7 5 ns30--39 45 41 2840--49 34 42 3850--59 11 10 22601 2 1 6

Gender: male 83 59 61 P# 0.05

Race P# 0.001Black, non-Hispanic 46 92 34White, non-Hispanic 44 8 10Hispanic 9 d 50Other/multiple races 2 d 6

Ever dependent upon/addicted to alcohol/another drug 49 79 65 P# 0.01

Dependent upon/addicted to alcohol/another drugin past 3 months

17 38 24 P# 0.05

Ever treated for drug/alcoholdependencya 59 72 86 P# 0.05

Ever diagnosed as having a mental health condition/illness 34 22 31 ns

Mental health condition/illness in past 3 months 31 12 24 P# 0.05

Ever treated for a mental health condition/illnessb 92 88 91 ns

ns5 not significant.aRestricted to those ever dependent upon/addicted.bRestricted to those ever diagnosed with condition.Note: Means sharing a subscript differ significantly from each other (P # 0.05).

treatment for it also varied among sites rang-ing from only 59% at Alabama to 72% at Balti-more and 86% at NYC (c25 7.154, df5 2,P5 0.028).

Although the proportion of patients whoreported having ever been diagnosed with amental health condition or illness did not varysignificantly among sites, the proportion expe-riencing such a condition in the past 3 monthsdid, ranging from only 12% at Baltimore to24% at NYC and 31% at Alabama(c25 6.691, df5 2, P5 0.035). The vast major-ity of patients with such a diagnosis reportedreceiving professional care or counseling fortheir condition.

Symptom PrevalenceGiven the important differences noted be-

tween sites with regard to the palliative careprograms established and the client popula-tions served, as well as the imbalance amongsites with regard to the number of clientsinterviewed and the proportion of clientswho completed an MSAS, it was deemed bothimportant and methodologically necessary to

present results separately for each site. Meanscores of patients did not differ significantlyamong sites for any of the four summary meas-ures (Table 3).The proportion of patients at each site re-

porting individual symptoms in the past weekis presented in Table 4. For presentation pur-poses, the symptoms are ordered by the pro-portion of patients in all three total samplescombined who reported the symptom duringthe past week. Symptoms comprising the

Table 3Mean Scores for Summary MSAS Measures

by Project Site

Alabama(n5 47)

Baltimore(n5 91)

New York City(n# 102)

P-valueMean (SD)Mean (SD) Mean (SD)

Total MSASscore

1.1 (0.6) 1.0 (0.6) 0.9 (0.7) ns

PsychologicalIndex

1.2 (1.0) 1.1 (0.9) 1.2 (1.0) ns

PhysicalIndex

1.1 (0.7) 1.1 (0.7) 1.0 (0.8) ns

GDI 1.3 (0.8) 1.2 (0.8) 1.2 (0.9) ns

ns5 not significant.

Vol. 30 No. 5 November 2005 413HIV/AIDS Patient Reports of Symptomatology

Table 4Symptom Prevalence, Proportion for Whom Symptom Is Frequent, Severe, and Distressing,

and Proportion Receiving Treatment

Symptom

Among Patients Reporting Symptom:

Symptom PrevalenceAmong All Patients

Symptom Frequent,Severe and Distressinga Receiving Treatmenta

Alabama Baltimore NYC

P-value

Alabama Baltimore NYC

P-value

Alabama Baltimore NYC

P-value% % % % % % % % %

Lack of energyb 77 71 70 ns 47 66 53 ns 44 16 32 P# 0.01Painb 64 74 69 ns 53 72 61 ns 73 83 67 nsWorryingc 51 55 53 ns 38 72 46 P# 0.01 27 6 23 P# 0.05Shortness of breath 57 58 43 ns 26 57 34 P# 0.05 44 69 52 nsDry mouthb 62 55 45 ns 17 54 31 P# 0.01 29 19 33 nsFeeling sadc 43 50 50 ns 20 36 44 ns 26 26 26 nsWeight changeb 43 56 39 ns 25 61 39 P# 0.05 30 31 44 nsCough 64 50 32 P# 0.001 27 42 26 ns 40 52 48 nsNauseab 57 46 42 ns 26 40 40 ns 67 66 49 nsDifficulty sleepingc 57 47 40 ns 52 49 64 ns 33 32 44 nsFeeling drowsyb 51 40 35 ns 25 44 32 ns 8 9 21 ns‘‘I don’t look like myself’’ 45 37 41 ns 43 65 55 ns 19 3 24 P# 0.05Lack of appetiteb 47 32 41 ns 32 52 56 ns 23 22 33 nsNumbness/tingling in

hands/feet45 41 30 ns 43 60 44 ns 35 40 37 ns

Itching 38 46 29 P# 0.05 50 69 53 ns 33 43 52 nsFeeling nervousc 43 31 40 ns 25 39 32 ns 32 22 26 nsChanges in skin 21 50 38 P# 0.01 60 61 46 ns 44 43 70 P# 0.05Difficulty concentratingc 40 34 40 ns d 19 23 ns 21 4 21 nsDiarrhea 53 37 29 P# 0.05 28 54 58 ns 33 66 58 P# 0.05Sweats 38 31 26 ns 22 36 30 ns 28 41 20 nsIrritablec 43 29 28 ns 25 38 39 ns 22 4 33 P# 0.05Vomitingb 28 30 23 ns 8 33 40 ns 62 70 41 nsDizzinessb 36 22 26 ns 23 20 48 ns 18 5 12 nsSwelling of arms/legs 19 25 27 ns 56 35 54 ns 22 64 62 nsChange in tasteb 34 28 20 ns 50 28 33 ns 27 8 25 nsFeeling bloatedb 28 19 26 ns 15 29 46 ns 23 d 40 P# 0.05Difficulty swallowing 19 23 20 ns 56 48 43 ns 56 48 53 nsConstipationb 23 19 22 ns 64 47 50 ns 46 65 54 nsProblems with urination 21 22 16 ns 50 80 41 P# 0.05 30 37 60 nsFever 4 12 28 P# 0.001 d 54 66 ns d 73 86 nsProblems with sexual

interest/activity15 11 19 ns 57 20 33 ns 14 10 6 ns

Mouth sores 2 11 12 ns 100 50 58 ns 100 80 75 ns

ns5 not significant.aRestricted to patients who reported symptom.bSymptom included in the physical (PHYS) subscale.cSymptom included in the psychological (PSYCH) subscale.

PSYCH and PHYS subscales are noted. As withother summary measures, the mean total num-ber of symptoms reported by patients did notvary significantly by site. On average, patientsat NYC reported 10.9 (SD5 7.6) symptomscompared to 11.9 (SD5 6.6) at Baltimoreand 12.7 (SD5 6.2) at Alabama.

Whereas 20 of the 32 symptoms assessedwere reported by at least 25% of patients atall three sites, 8 were reported by at least40% (lack of energy, pain, worrying, shortnessof breath, dry mouth, feeling sad, nausea, anddifficulty sleeping), and 3 (lack of energy,pain, and worrying) were reported by 50--

74%. The prevalence of five symptoms variedsignificantly by site, with two (coughing, itch-ing) reported by a disproportionately lownumber of patients at NYC, one (diarrhea) re-ported by a disproportionately high propor-tion of patients at Alabama, and two(changes in skin, fever) reported by a dispro-portionately low number of patients atAlabama.

In addition to prevalence of symptoms, Ta-ble 4 also presents the proportion of patientswho described each symptom as frequent, severe,and distressing as well as the proportion who re-ported receiving treatment for it. A symptom

414 Vol. 30 No. 5 November 2005Karus et al.

was defined as frequent if it occurred ‘‘frequently,’’‘‘almost constantly,’’ or if it was one of theeight symptom conditions/states for which fre-quency was not assessed because it was as-sumed to be constant. A symptom wasdefined as severe if it was described by the pa-tient as either ‘‘severe’’ or ‘‘very severe.’’ Asymptom was defined as distressing if the pa-tient reported that it caused ‘‘quite a bit’’ or‘‘very much’’ distress. Preliminary analysesshowed that patients experiencing a high levelof problems with regard to one dimensionwere likely to report high levels of problems re-garding other dimensions as well. Therefore,to minimize the presentation of redundant re-sults, we have limited discussion to those pa-tients who described their symptoms asfrequent, severe, and distressing.

A total of 23 symptoms were described by atleast 25% of patients at all three sites who ex-perienced the symptom as frequent, severe,and distressing. For 11 symptoms (lack of ener-gy, pain, difficulty sleeping, changes in appear-ance, numbness/tingling in hands/feet,itching, changes in skin, difficulty swallowing,constipation, problems with urination, andmouth sores), this proportion was at least40%, and for 3 symptoms (pain, itching, andmouth sores), it was at least 50%. The propor-tion of patients describing five symptoms (wor-rying, shortness of breath, dry mouth, weightchange, and problems with urination) as fre-quent, severe, and distressing varied signifi-cantly among sites. In all five such instances,the proportion reporting highly distressingsymptomatology was particularly high atBaltimore, even though the overall prevalenceof these symptoms did vary among sites.

With regard to treatment, for only foursymptoms (pain, nausea, difficulty swallowing,and mouth sores) did at least half of the pa-tients who reported the symptom also reportreceiving treatment. Conversely, for 12 othersymptoms (worrying, dry mouth, feeling sad,feeling drowsy, changes in appearance, lackof appetite, feeling nervous, difficulty concen-trating, feeling irritable, dizziness, and sexualproblems), a third or less of patients at allthree sites who reported the symptomreported receiving treatment. Among these12 symptoms were 5 of the 6 that comprisethe PSYCH subscale and 5 of the 12 that com-prise the PHYS subscale.

A comparison of prevalence data with pa-tients’ reports of treatment for their symptomssuggests that many of the most common symp-toms at each site, including the three symp-toms reported by a majority of patients at allthree sites (lack of energy, pain, and worrying),were perceived by the patient as not havingbeen treated. At Alabama, for 5 of the 10 symp-toms experienced by half or more of thepatients, treatment was reported by a third orless of those who reported experiencing thesymptom. Similarly, low levels of treatmentwere reported for five of the nine symptomsreported by at least half of the patients atBaltimore and for three of the four suchsymptoms at NYC.

DiscussionThe data presented in this report from

patients at three diverse programs targetingunderserved populations living with HIV/AIDS demonstrate the feasibility of obtaininga detailed assessment of patients’ symptom-atology from this vulnerable population. Thefindings also document that, despite improvedtreatments for HIV/AIDS, many living with thedisease continue to experience high levels ofphysical and psychological symptomatology.Although the mean number of symptoms

(10.9--12.7) reported by patients at each ofthe three programs described was approxi-mately 25% less than the mean number ofsymptoms (16.7) reported in a study of ambu-latory AIDS patients accrued 1992--1995,5 andless than half the mean number of symptoms(27.67) reported for a sample of 50 inpatientsof an AIDS inpatient unit in a skilled nursingfacility at NYC accrued in 1997--1998 prior toor shortly after initiating protease inhibitoruse,17 the total number of symptoms experi-enced by seriously ill HIV/AIDS patients con-tinues to remain very high. From a clinicalperspective, managing treatment for a patientwith 10--12 symptoms is highly complex, ofteninvolving detailed clinical assessment and re-view, polypharmacy, assessment of drug andsymptom interactions, and the potential forcompound side effects. High levels of symp-tomatology may also impact patients’ abilityor willingness to adhere to various treatmentregimens.

Vol. 30 No. 5 November 2005 415HIV/AIDS Patient Reports of Symptomatology

Eight symptomsdphysical (pain, lack of en-ergy, shortness of breath, dry mouth, and nau-sea) as well as psychological (worrying, feelingsad, and difficulty sleeping)dwere reported by40--74% of patients at all three sites, with abouthalf or more of patients at all sites describingthree of these (pain, lack of energy, and diffi-culty sleeping) as highly frequent, severe, anddistressing. Thus, there remains a need for de-tailed symptom management among seriouslyill HIV/AIDS patients. The prevalence of manyof these symptoms was similar to (and in someinstances worse than) that of terminally ill can-cer patients.18--21 It should also be noted thatalthough less prevalent than nausea, severaladditional symptoms associated with gastro-intestinal problems (weight change, diarrhea,difficulty swallowing, urination, and vomiting)were also common and found at levels similarto those reported for cancer patients.22 Previ-ous research has identified nausea, vomiting,and diarrhea as potential side effects of variousnucleoside reverse transcriptase inhibitors.23

The present analyses also suggest that a sub-stantial proportion of patients at all three sitesperceive that many of their symptoms, includ-ing some of the most prevalent, go untreated.This could occur for several reasons. First, itmay be that many symptoms commonly experi-enced by patients in this population are notroutinely screened for by medical staff. As aresult, some of the symptoms reported bypatients in this report were not previouslyreported to medical staff and, therefore, re-mained undiagnosed and untreated. A secondpossibility is that some symptoms reported asuntreated by the patient are associated witha condition that is being treated, but the pa-tient is unaware of the connection (e.g., pa-tients may perceive that their fatigue is notbeing treated although the physician is treat-ing them for anemia). Third, it could also bethat medical staff are aware of some symptomsbut cannot treat them for medical reasons(e.g., they are side effects of other treatmentsor the condition causing the symptom isuntreatable) or financial constraints (e.g.,funds are unavailable for the treatment). Allthree of these possibilities may be indicativeof insufficient or poor physician--patientcommunication.

It is encouraging to note that two-thirds ormore of patients at each site experiencing pain

reported receiving treatment for it, althoughone would hope that all patients experiencingpain felt they were being treated for it. Eventhough treatment of pain is more complex inpatients with drug and/or alcohol depen-dence, effective regimens can still be devised.24

Less encouraging are the low proportions ofpatients who reported treatment for symptomswith known effective treatments,22 such asconstipation, vomiting, nausea, and diarrhea.These findings call for improvements in educa-tion among general medical staff and indicatea need for more skilled palliative assessmentand management.

The findings presented also provide empiri-cal evidence of variation among sites with re-gard to the symptoms reported and theproportion of patients who describe varioussymptoms as frequent, severe, and distressing.Significant differences were noted among siteswith regard to the prevalence of five symp-toms. Coughing and diarrhea were most likelyto be reported by patients at Alabama and leastlikely to be reported by patients at NYC. Onthe other hand, fever was most prevalent atNYC, where patients were more likely to havebeen accrued while in the hospital and within1 month of death. Finally, two skin-relatedsymptoms likely associated with substanceabuse (itching and changes in skin) were mostprevalent at Baltimore, where homelessnessand substance abuse were especially high. Pa-tients at Baltimore were also disproportionatelylikely to describe four common symptoms(worrying, shortness of breath, dry mouth,and weight change) as particularly distressing,even though the prevalence of these symptomsdid not differ among sites.

Such differences underscore the importanceof carrying out a population-based needs as-sessment prior to the establishment of pallia-tive care programs for HIV/AIDS patients toensure that as many of the patients’ medicalneeds as possible are identified and addressed.To the extent that they are receiving regularongoing treatment at all, patients enteringprograms in different settings are likely to ar-rive with different comorbidities, treatmenthistories, and are on different treatment regi-mens. These differences are, in turn, likely toimpact the overall prevalence and experienceof individual symptoms. For example, a pro-gram with a majority of patients who enroll

416 Vol. 30 No. 5 November 2005Karus et al.

while on HAART might be expected to reporta higher prevalence of treatment-related symp-toms among its patients, while a program serv-ing patients with histories of erratic treatmentor poor adherence might be expected to re-port a higher prevalence of symptoms relatedto opportunistic infections. Similar differenceswould be expected based on the prevalence ofsubstance abuse and the types of substancesfor which abuse is reported.

Three methodological issues merit discus-sion. First, missing data may have biased thefindings. Second, findings may also have beenbiased as a result of the MSAS being adminis-tered by a registered nurse at Alabama and Bal-timore and by a nurse practitioner, a socialworker, or a community outreach worker atNYC. However, the likelihood of such a biasis minimal since the form elicits patient re-ports regardless of who administers it and, al-though nonclinical staff administered someof the forms at NYC, most were administeredby the nurse practitioner. Third, the generaliz-ability of results is unknown. Although allthree programs described target underservedindividuals living with HIV/AIDS, importantstructural differences between the programsites were noted. Sites differed with regard totheir geographic location, patient populationsserved, services provided by staff, service deliv-ery model, and the level of contact and inti-macy that staff can develop with patientsduring the course of service provision. Evi-dence that there are also differences amongsites with regard to the prevalence of specificsymptoms and whether they are treated, as wellas patients’ perceptions of symptom intensityand treatment, is consistent with the hypothe-sis that patient symptomatology is associatedwith structural factors. Unfortunately, becausethere were only three sites, each with a uniquecombination of structural attributes, it was im-possible to control for important differencesbetween the sites in these analyses.

Overall, despite the trend toward increasedsurvival for patients with HIV/AIDS, many liv-ing with the disease experience a wide rangeof frequent, severe, and distressing symptoms,not all of which are directly associated withHIV/AIDS disease progression or survival.They are also likely to perceive that many oftheir symptoms are not being treated. In com-bination with earlier investigations that have

documented a lack of agreement between phy-sician and HIV/AIDS patients’ reports ofsymptoms,3--4 these findings underscore theimportance of improving patient-physiciancommunication around the patient’s experi-ence of symptomatology and symptom man-agement. It is incumbent upon physicians tomore systematically query HIV/AIDS patientswith regard to symptoms they are experienc-ing; provide patients with a more complete un-derstanding of the treatments they arereceiving; and, foster realistic expectationswith regard to symptom treatability, palliation,and potential side effects.

AcknowledgmentsThis work was supported by a Special Projects

of National Significance grant (H97 HA00148-02) funded by the HRSA, Department of HHS(V.H. Raveis, Principal Investigator).The authors acknowledge and appreciate

the assistance of Monique Carrero and KarynFaber in the preparation of this manuscript.

References1. Detels R, Munoz A, McFarlane G, et al. Effective-ness of potent antiretroviral therapy on time toAIDS and death in men with known HIV infectionduration. JAMA 1998;280:1497--1503.

2. Paella FJ, Delaney KM, Moorman AC, et al. De-clining morbidity and mortality among patients withadvanced human immunodeficiency virus infection.N Engl J Med 1998;338(13):853--860.

3. Justice AC, Rabeneck L, Hays RD, Wu AW,Bozzette SA. Sensitivity, specificity, reliability, andclinical validity of provider-reported symptoms:a comparison with self-reported symptoms. J AcquirImmune Defic Syndr 1999;21(2):126--133.

4. Fontaine A, Larue F, Lassaumiere J. Physicians’recognition of the symptoms experienced by HIVpatients: how reliable? J Pain Symptom Manage1999;18(4):263--270.

5. Vogl D, Rosenfeld B, Breitbart W, et al. Symp-tom prevalence, characteristics, and distress in AIDSoutpatients. J Pain Symptom Manage 1999;18(4):253--262.

6. Easterbrook P, Meadway J. The changing epide-miology of HIV infection: new challenges for HIVpalliative care. J R Soc Med 2001;94:442--448.

7. Reisler RB, Han C, Burman WJ, Tedaldi EM,Neaton JD. Grade 4 events are as important as AIDS

Vol. 30 No. 5 November 2005 417HIV/AIDS Patient Reports of Symptomatology

events in the era of HAART. J Acquir Immune DeficSyndr 2003;34(4):379--386.

8. Culhane DP, Gollub E, Kuhn R, Shpaner M. Theco-occurrence of AIDS and homelessness: resultsfrom the integration of administrative databasesfor AIDS surveillance and public shelter utilizationin Philadelphia. J Epidemiol Community Health2001;55(7):515--520.

9. DesJarlais DC, Perlis T, Friedman SR, et al. Be-havioral risk reduction in a declining HIV epidemic:injection drug users in New York City, 1990--1997.Am J Public Health 2000;90(7):1112--1116.

10. Hwang SW, O’Connell JJ, Lebow JM, et al.Health care utilization among homeless adults priorto death. J Health Care Poor Underserved 2001;12(1):50--58.

11. Martens WH. A review of physical and mentalhealth in homeless persons. Public Health Rev2001;29(1):13--33.

12. Weinreb L, Goldberg R, Perloff J. Health char-acteristics and medical service use patterns of shel-tered homeless and low-income housed mothers.J Gen Intern Med 1998;13(6):389--397.

13. Somlai AM, Kelly JA, Wagstaff DA, Whitson DP.Patterns, predictors, and situational contexts of HIVrisk behaviors among homeless men and women.Soc Work 1998;43(1):7--20.

14. Karus D, Raveis V, Marconi K, et al. Serviceneeds of patients with advanced HIV disease: a com-parison of client and staff reports at three palliativecare projects. AIDS Patient Care STDs 2004;18(3):145--158.

15. Karus D, Raveis VH, Marconi K, et al. Mentalhealth status of clients from three HIV/AIDSpalliative care projects. Palliat Support Care 2004;2:125--138.

16. Portenoy RK, Thaler HT, Kornblith AB, et al.The Memorial Symptom Assessment Scale: an in-strument for the evaluation of symptom prevalence,

characteristics and distress. Eur J Cancer 1994;30A(9):1226--1236.

17. Brechtl JR, Breitbart W, Galietta M, Krivo S,Rosenfeld B. The use of highly active antiretroviraltherapy (HAART) in patients with advanced HIV in-fection: impact on medical, palliative care, and qual-ity of life outcomes. J Pain Symptom Manage 2001;21(1):41--51.

18. Higginson IJ. Epidemiologically based needsassessment for palliative and terminal care. UK:Radcliffe Medical Press, 1997.

19. Higginson IJ, Constantini M. Epidemiologicalmethods in studies of symptoms in advanced dis-ease. In: Max MB, Lynn J, eds. Symptom research:Methods and opportunities (An interactive text-book). Bethesda, MD: NIH, 2003. Available fromhttp://symptomresearch.nih.gov/chapter_19/index.htm. Accessed May 14, 2004.

20. Koffman J, Higginson IJ, Donaldson N. Symp-tom severity in advanced cancer, assessed in two eth-nic groups by interviews with bereaved familymembers and friends. J R Soc Med 2003;96:10--16.

21. Klinkenberg M, Willems DL, Van der Wal G,Deeg JH. Symptom burden in the last week of life.J Pain Symptom Manage 2004;27(1):5--13.

22. Potter J, Higginson IJ. Frequency and severity ofgastrointestinal symptoms in advanced cancer. In:Ripamonti C, Bruera E, eds. Gastrointestinal symp-toms in advanced cancer patients. London, UK: Ox-ford University Press, 2002: 1--15.

23. Max B, Sherer R. Management of the adverse ef-fects of antiretroviral therapy and medication adher-ence. Clin Infect Dis 2000;30(Suppl 2):S96--S116.

24. Bruera E, Neumann CM. The management ofchronic pain in palliative non-cancer patients. In:Addington-Hall JM, Higginson IJ, eds. Palliativecare for non-cancer patients. London, UK: OxfordUniversity Press, 2001: 66--81.