Self-reported non-adherence toimmune-suppressant therapy in livertransplant recipients: demographic,interpersonal, and intrapersonal factors

Patient adherence to a multifaceted regimen offollow-up care is extremely important in achievingoptimal long-term outcomes after organ transplan-tation, including liver transplantation (LT) (1, 2).Immune-suppressant medications are essential toachieve optimal graft health, and many interper-sonal factors such as communication, relation-ships, and social support may contribute toimmune-suppressant adherence (2). Adherence toa medical regimen may be defined broadly to coverall aspects of patient conformity to medical advice,and this is usually an active, voluntary, andindependent choice of the patient (3). Adherence

to immune suppressants may therefore be definedas the consistent use of medications as prescribedby a physician. Various intrapersonal factors likereligious convictions, traditional and personalperceptions, and coping strategies may furtherinfluence adherence (2). Literature reveals varyingrates of non-adherence to immune suppressants insolid organ transplantation. The non-adherencerates are much higher in renal recipients (36/100person years or py) when compared with heart(14.5/100 py) and liver recipients (6.7/100 py) (2).Issues related to the underlying chronic disease,different population characteristics, etc. possibly

Lamba S, Nagurka R, Desai KK, Chun SJ, Holland B, Koneru B. Self-reported non-adherence to immune-suppressant therapy in liver transplantrecipients: demographic, interpersonal, and intrapersonal factors.

Abstract: Adherence to immune suppressants and follow-up care regimen isimportant in achieving optimal long-term outcomes after organtransplantation. To identify patients most at risk for non-adherence, thiscross-sectional, descriptive study explores the prevalence and correlates ofnon-adherence to immune-suppressant therapy among liver recipients.Anonymous questionnaires mailed consisted of the domains: (i) adherencebarriers to immune suppressants, (ii) immune suppressants knowledge, (iii)demographics, (iv) social support, (v) medical co-morbidities, and (vi)healthcare locus of control and other beliefs. Overall response was 49%(281/572). Data analyzed for those transplanted within 10 yr of study reveal50% (119/237) recipients or 9.2/100 person years reporting non-adherence.Non-adherence was reported highest in the 2–5 yr post-transplant phase(69/123, 56%). The highest immune-suppressant non-adherence rates werein recipients who are: divorced (26/34, 76%, p = 0.0093), have a history ofsubstance or alcohol use (42/69, 61%, p = 0.0354), have mental healthneeds (50/84, 60%, p = 0.0336), those who missed clinic appointments (25/30, 83%, p < 0.0001), and did not maintain medication logs (71/122, 58%,p = 0.0168). Respondents who were non-adherent with physicianappointments were more than four and a half times as likely (OR 4.7, 95%CI 1.5–14.7, p = 0.008) to be non-adherent with immune suppressants. Inconclusion, half of our respondents report non-adherence to immunesuppressants. Factors identified may assist clinicians to gauge patients�non-adherence risk and target resources.

Sangeeta Lambaa, RoxanneNagurkab, Kunj K. Desaic, Shaun J.Chuna, Bart Hollandd and BaburaoKonerue

aDepartments of Surgery and Emergency

Medicine, bDepartment of Emergency Medicine,cDepartment of Surgery, dDepartment of

Preventive Medicine & Community Health andeDivision of Transplant Surgery, Department of

Surgery, University of Medicine and Dentistry of

New Jersey-New Jersey Medical School,

University Hospital, Newark, NJ, USA

Key words: adherence – compliance – immune

suppression – immune-suppressant – liver

transplantation

Corresponding author: Sangeeta Lamba, MD,

Assistant Professor, Surgery and Emergency

Medicine, University Hospital, 150 Bergen Street,

WM-203, Newark, NJ 07101, USA.

Tel.: +1 973 972 5128; fax: +1 973 972 6646;

e-mail: lambasa@umdnj.edu

Conflict of interest: None.

Accepted for publication 20 May 2011

Clin Transplant 2012: 26: 328–335 DOI: 10.1111/j.1399-0012.2011.01489.xª 2011 John Wiley & Sons A/S.

328

contribute to these differences and highlight theneed to focus on each subset of solid organrecipients as a unique group. A pilot study in LTconcluded that one in ten LT deaths was secondaryto non-adherence to medications (4). Additionally,economic models in renal transplant suggest ahigher cost per quality adjusted life year in non-adherent patients as compared with adherentpatients (5). Thus, non-adherence to immune-suppressant medications increases morbidity andmortality, as well as the economic cost of caring fortransplant recipients.

A review of immune-suppressant adherence intransplant recipients reveals extensive analyses inrenal transplantation, but there remains a paucityof literature among liver recipients. A recentcomprehensive meta-analysis evaluated 147 fulllength publications, regarding patient adherencepost-transplantation, of which only 20% includedLT recipients (2). Immune-suppressant medicationadherence was addressed in only seven of these LTstudies and remains an understudied topic (6–12).Literature also exhibits wide variability in thedefinition used to describe immune-suppressantadherence with no consensus to date (13). Severalmethods have been used to assess non-adherenceincluding indirect measures like electronic pillcounters and blood medication levels, retrospectivehistorical chart reviews, interviews with health careproviders/family, and direct measures like patientself-report/interview. While no single method isconsidered the gold standard, patients� self-reportscan easily and effectively measure adherence,especially when administered anonymously (13–17). The goal of this cross-sectional, self-reportstudy was to determine the prevalence of non-adherence with immune-suppressant therapy andto further explore the potential factors of influence(demographics, social, belief patterns, etc.) in asocio-economically and ethnically diverse popula-tion of liver recipients at an urban public hospital-based LT program.

Methods

This study was a self-report anonymous question-naire.

Sample and setting

The target population consisted of all adult(>18 yr of age) LT recipients at our urban publichospital from January 1989 to April 2005. Thecenter currently performs an average 50–55 LT/yr.The study population included all living adult LTrecipients, who had received a minimum of one yr

of follow up after LT. Our recipients comprised60% Caucasian, 22% Hispanic, 12% AfricanAmerican, and 6% Asian.

Data collection procedures

Survey instrument and administration. The surveyinstrument (in both English and Spanish) consistedof a cover letter and 30 questions in three double-sided pages. The survey domains consisted of: (i)adherence to immune suppressants, (ii) patientknowledge and information about immunesuppressants, (iii) demographics, socio-cultural,and substance use, (iv) perceived social support,(v) medical co-morbidity issues, and (vi) healthcarelocus of control and other beliefs. The survey wasrevised with the help of a panel of transplantphysicians, nurses, and social work coordinators.To enhance the response rate and improve theaccuracy of self-report, the surveys were adminis-tered anonymously. The principal investigator andthe addressee to whom the surveys were returned(SL) were independent from the transplant team.Also, questions directly addressing adherence werefocusedmainly on information gathering and used anon-judgmental tone (4, 18, 19). The study protocolwas approved by the Institutional Review Board,with an exemption from informed consent. Surveyswere mailed with a cover letter and an anonymoussalutation. The cover letter informed the subjects ofthe voluntary nature of their participation andassured confidentiality. A total of twomailings wereperformed, four months apart, after verification ofundeliverable addresses. No incentives were offered.

Variables and measurement of response.Non-adherence to immune suppressants: We classi-

fied respondents as ‘‘non-adherent’’ if theyanswered (i) ‘‘yes,’’ to the question, ‘‘Have youmissed or not taken doses of your transplantmedications?’’ and/or (ii) answered the question,‘‘If you have missed any doses of medications,what has been the frequency in the past three -months?’’ This approach was used to captureunder-reporting based on a single direct question.Admittedly, this yields a liberal definition for non-adherence (see Discussion). To determine thetimeframe of risk of non-adherence after LT,subjects were asked, ‘‘When did you tend to missthe doses of transplant medications?’’(early 1styear, middle-2–5 yr post-LT, or late-greater than5 yr). The Immune-Suppressant Therapy BarrierScale (ITBS) has been utilized to assess barriers tomedication adherence in renal transplant patients,where it achieved high reliability (Cronbach�s alpha0.91) and validity (p < 0.01) (20). We used the

Non-adherence to immune suppressants

329

ITBS to identify the ‘‘controllable’’ and ‘‘uncon-trollable’’ barriers.Knowledge and information about immune sup-

pressants: While transplant professionals may feelthat they provide adequate and uniform instruc-tions to all patients regarding immune suppres-sants, patient perceptions may differ. This variableknowledge of medications may impact non-adher-ence (21). Respondents were therefore asked abouttheir perception of adequacy of information pro-vided regarding the importance of LT medications.We assessed their specific knowledge aboutimmune suppressants with respect to the durationof action and perceived side effects.Demographics, socio-cultural, and substance use:

Questions addressed demographic characteristicsof the respondents, socio-economic factors, andhealth care insurance, which may impact adherence(3). We also assessed alcohol and substance abusehistory and mental health needs.Perceived social support: When normative beliefs

are scrutinized, psychosocial approval from impor-tant others is valuable, as supported by studies thatindicate that less spousal support is associated withmedication non-adherence (21). The social supportscale used was modified from the Medical Out-comes Study�s survey instrument to assess thetangible support as opposed to affectionate sup-port, because of practical considerations of surveylength (22). The support measures have allreported Cronbach�s alpha >0.91 (22).Medical/co-morbidity issues: Co-morbid condi-

tions, which contribute to total disease and med-ication burden, were addressed. A specific scalewas not used but we looked at associationsbetween each medical illness as related to medica-tion non-adherence.Healthcare locus of control and other beliefs:

Patient beliefs play a central role in their attitudestoward medications and hence non-adherence tosame. Because of survey size limitations, we used amodified scale derived from the MultidimensionalHealth Locus of Control (MHLC) survey (23).This survey measures beliefs about determinants ofa person�s health. The three subscales assess patientbelief in the physician, chance/luck or self as beingresponsible and in control of their health. (Re-ported Cronbach�s alpha 0.60–0.75.)Potential surrogates of non-adherence: In addi-

tion to the six central domains, the survey alsoincluded questions about appointments to the out-patient clinic, for laboratory tests and whetherrespondents had liver biopsy/ies. Because missedclinic and laboratory test appointments correlatewith medication non-adherence in a few studies (4),respondents were asked to indicate whether they

missed such appointments and then quantify themas a few vs. a lot. Also, because immune-suppres-sant non-adherence may lead to rejection and graftdysfunction, information about biopsies wasthought to be relevant.

Statistical analysis

Returned questionnaires were serially numberedand checked for legibility. Data were manuallyentered from questionnaires into an Excel (Micro-soft Corporation, Redmond, WA, USA) spread-sheet. After random auditing of entered data andcleaning, data were converted into SAS files andanalyzed using SAS 9.1 (SAS Institute, Cary, NC,USA). Distributions were examined and summarystatistics generated (mean, median, percentiles),followed by formal analyses and significance tests.Medians and means are reported with range andstandard deviation in parentheses, respectively.Chi-square or Fisher�s exact tests were used fortesting the association of categorical independentvariables to adherence and Wilcoxon�s rank test forcomparisons between continuous variables andtheir impact on adherence. Alpha was set at 0.05,and all tests were two-tailed. Multiple logisticregression was performed with a model thatincluded all variables, for example, ethnicity,education, income, physician office visit adherence,log-keeping, etc. as independent variables andadherence to immune suppressants as the depen-dent variable. As in similar clinical epidemiologiccontext studies where the goal is to compare ratesin the face of differences in observation time, wealso report event rates (Tables 1 and 2) as cases ofnon-adherence per 100 persons per year (or per 100person years [py] of observation) (2).

Results

During the study period, 860 adults received LT.Of those, we excluded 234 because of recipientdeath and 54 because of less than one yr of follow-up. Overall, 572 questionnaires were mailed. Twohundred and eighty-one responded, the majority232/281 (83%) in the first mailing. Eighty-onesurveys were returned undeliverable, despite multi-ple attempts to find/confirm new addresses, includ-ing contacting the next of kin listed. This yielded atrue response rate of 57.2% (281/[572–81]) fromthe delivered surveys (overall response rate 49%).The anonymous nature of the survey precluded anyfurther comparative analysis of respondents vs.non-respondents.

To improve the relevance of the study findings,detailed analyses were performed only in 84%

Lamba et al.

330

(237/281) respondents who received their trans-plants £ 10 yr of the survey. The mean intervalafter transplantation in this subset was 5.4 yr.Prior studies reveal a mean length of follow-up of3.6 yr and a range of 0.3–12.3 yr (2). In addition,we analyzed the recipient subgroup who weretransplanted within the last five yr (171/281) of thestudy as well. The patterns of adherence and trendswere similar in the sub group analyses, but mostvariables did not reach statistical significance,probably because of smaller size (data not shown).

Non-adherence to immune suppressants

Respondents reported an overall non-adherencerate of 50% (119/237) or 9.2 cases per 100 py.When analyzed to quantify the degree of non-adherence in the past three months, 11 respondentshad marked a frequency of 1–20% having initiallyanswered that they never missed medication doses

and these were therefore included in the non-adherent group. We were unable to accurately‘‘quantify’’ the degree of non-adherence during thelast three months because of an inherent flaw in thequestion addressing the same. However, most ofthe question respondents, 91% (83/91), reportednon-adherence frequency of less than or equal to20% within the past three months of survey. Themajority of respondents, 56% (69/123), also iden-tified the 2–5 yr post-LT period as the most likelytime during which they were at risk to skipimmune-suppressant doses compared with theearly/1st year 19% (23/123) or late >5 yr 17%(21/123) post-LT phases.Respondents who reported non-adherence listed

(strongly agree or agree) the following reasons formissing doses: doing things out of their dailyroutine 34% (40/117), difficulty in remembering totake medications 24% (27/113), having too manypills 12% (12/104), and the need to take pills toomany times 13% (13/103). Very few respondentsstated that they did not understand when to takeimmune suppressants, stopped pills when they feltside effects, or when they felt better. Also, few (5%,5/106) agreed or strongly agreed that they misseddoses when finances were running low, or they ranout of pills (The variable numbers for eachquestion are because of the fact that not all sub-categories were uniformly marked by respondents).

Knowledge and information about immune suppressants

The majority of respondents stated that theyreceived enough information about the immunesuppressants (92%, 218/236). However, a higherproportion of non-adherent respondents (12%, 14/

Table 1. Univariate association of patient characteristics with non-adher-ence to immune-suppressant therapy

Patientsa

N = 237AdherentN (%)

Non-adherentN (%)

Non-adherentPPY

Age20–49 42 18 (43) 24 (57) 12.350–82 195 100 (51) 95 (49) 8.7

GenderMale 152 75 (49) 77 (51) 9.4Female 82 42 (51) 40 (49) 8.8

EthnicityAfrican American 18 8 (44) 10 (56) 11.8Asian/Asian Indian 12 9 (75) 3 (25) 4.5Hispanic 35 18 (51) 17 (49) 7.8Caucasian 169 82 (49) 87 (51) 9.5

Insurance coverageat time of surveyYes 212 104 (49) 108 (51) 9.3No 12 7 (58) 5 (42) 8.5

Marital statusMarried 157 84 (54) 73 (46) 8.3Single 30 18 (60) 12 (40) 8.3Widowed 12 6 (50) 6 (50) 8.6Divorced* 34 8 (24) 26 (76) 14.9

Employment historyNever employed** 12 10 (83) 2 (17) 3.0Been employed(pre-transplant)

194 91 (47) 103 (53) 9.7

Social supportAdequate 222 113 (51) 109 (49) 9.1Inadequate 12 5 (42) 7 (58) 10.0

EducationCollege educated 83 37 (45) 46 (55) 10.3Non-college educated 152 81 (53) 71 (47) 8.5

p Values <0.05 are considered significant: *p = 0.0093, **p = 0.0171.PPY, per 100 person years.aMissing numbers are because of some respondents not answering allquestions.

Table 2. ‘‘Sentinel signs’’: univariate analyses of potential clinical surro-gates of non-adherence among patients receiving immune-suppressanttherapy

Patients Adherent Non-adherent

N = 237 N (%) N (%) PPY

Missed appointments with transplant and/or primary care provider(p < 0.0001)

Yes 30 5 (17) 25 (83) 14.4No 206 113 (55) 93 (45) 8.4

Patients keeping a log of medications (p = 0.0168)Yes 102 59 (58) 43 (42) 8.1No 122 51 (42) 71 (58) 10.1

History of substance or alcohol abuse (p = 0.0354)Yes 69 27 (39) 42 (61) 11.5No 168 91 (54) 77 (46) 8.3

Mental health needs (p = 0.0336)Yes 84 34 (40) 50 (60) 11.1No 153 84 (55) 69 (45) 8.2

PPY, per 100 person years.

Non-adherence to immune suppressants

331

118) stated that they did not receive adequateinformation about immune suppressants whencompared with those who were adherent (3% [4/118], p = 0.025). Of the 57 respondents whowould have liked more information, most preferredthe group setting as a desired modality (equaldistribution among video, written material, and abrief meeting with a healthcare provider).Respondents (41%, 96/237) believed that side

effects ranging from low white cell count, lowblood count, swollen feet, muscle aches to sadthoughts were because of immune suppressants.However, the non-adherent proportion in thegroup that perceived side effects were medication-related did not differ significantly from those whodid not (52% vs. 49%, p = 0.634). Majority (87%,188/215) of respondents correctly identified thatmost immune suppressants lasted less than a day inthe body and reported being on tacrolimus as theprimary immune-suppressant. However, non-adherence was not higher in the group whoperceived the effects to be longer lasting. Thenon-adherence rates were also similar among therespondents taking various immune suppressantsand did not increase with increasing pill burden.

Demographics, socio-cultural, and substance use

The reported median age at LT was 53 (18–73)years (Table 1). Patients aged 50 yr or oldertended to report less non-adherence to immunesuppressants when compared with those who wereyounger. Rates of non-adherence did not differsignificantly between men and women or byethnic group. Asians reported the lowest, whereasAfrican-Americans reported the highest rates ofnon-adherence (p = 0.33). Rates of non-adher-ence also did not differ significantly among thosemarried, single, or widowed. In contrast, thedivorced recipients reported significantly higherrates of non-adherence (p = 0.009). One hundredand ninety-four (82%) respondents stated thatthey worked prior to LT and only 73 reported(31%) that they worked post-LT, and a lessernumber, 61 (26%) reported employment at thetime of survey. Those who were not employedhad lower non-adherence (17% vs. 52%,p = 0.017). We also found a trend for the collegeeducated to be less adherent (55% vs. 47%).However, these differences were not statisticallysignificant.Nearly one-third (29%) of our respondents

reported a history of substance or alcohol use,and non-adherence rate was higher in this groupwhen compared with those with no such history(61% vs. 46%, p = 0.04) (Table 2). One-third

(35%) of respondents reported that they receivedcare from a mental health professional after LT,and 26% reported that they were receiving suchcare at the time of the survey. Non-adherence toimmune suppressants was significantly higher inthose with self-identified mental health needs (60%vs. 40%, p = 0.04) (Table 2).

Perceived social support

Majority, 95% (222/234), of respondents per-ceived their social support to be ‘‘adequate.’’ Atangible social support score (1 [low] to 25 [high])was calculated from responses to five questions.Non-adherence decreased as the perceived socialsupport increased, from 75% at a score of 5, to45% at a score of 25 (p = 0.049).

Medical/co-morbidity issues

Non-adherence rates did not differ significantlyby the underlying indication for LT or presenceof co-morbidities either singly or in combination.

Healthcare locus of control and other beliefs

Respondents agreed or strongly agreed that theywere themselves responsible for their health (89%),that their physician played an important role in theirhealth (76%), and that luck or chance had a role intheir health (27%). Rates of non-adherence did notdiffer significantly among the various groups.

Potential surrogates of non-adherence

Majority (87%) of respondents stated they regu-larly kept their appointments with their health careproviders (Table 2). Importantly, non-adherencewas significantly lower among those who regularlykept their appointments when compared with thosewho did not (45% vs. 83%, p < 0.01). Overall,46% of respondents stated they maintained a logfor their medications. Respondents who kept sucha log were less non-adherent when compared withthose who did not (42% vs. 58%, p = 0.0168).Furthermore, a majority of those who stated theymissed their appointments with their healthcareprovider also did not keep a medication log (79%,23/29; p < 0.01). Self-reports of liver biopsy werenot significantly different among the two groups.

Multivariate analysis

Only one variable, non-adherence to physicianappointments, emerged as having an independenteffect on non-adherence to immune suppressants.

Lamba et al.

332

Respondents who were non-adherent with physi-cian office appointments were more than four anda half times as likely (OR 4.7, 95% CI 1.5–14.7,p < 0.01) to be non-adherent with immune sup-pressants when compared with those who wereadherent. The small numbers in other subgroupsprecluded further meaningful analyses.

Discussion

Our study furthers the understanding of clinicalepidemiology of non-adherence to immune sup-pressants in LT recipients. The salient findingswere: (i) Based on a liberal definition, about half ofall survey respondents (nine cases per 100 py) werenon-adherent by self-report, which may be a biggerproblem than often recognized; (ii) We identifiednon-adherence to office appointments as a poten-tially important ‘‘tip-off’’ in identifying non-adher-ence to medications; and (iii) Our study suggeststhat some subgroups, such as the divorced, thosewith a history of alcohol/substance use and thosewith perceived mental health needs may be athigher risk for non-adherence.

Our survey response rate of 49% falls within therange (49–58%) reported in earlier studies in solidorgan transplantation (24–26). A ‘‘gold standard’’measurement of medication adherence does notexist (18). Each of the prior methods used hasstrengths and limitations. Direct patient interviews(4, 21, 27), while providing a detailed assessment ofpatient belief patterns, run the risk of the patientproviding only the information that the researcherwants to hear. Reports by family member orhealthcare provider (28), while helpful in identify-ing a non-adherent patient, introduce bias and arecounter to the patient-centered approach. Bloodlevels of drugs (4) reflect adherence only at the timeof measurement but not in the interim andelectronic pill counters, while accurate are expen-sive, labor intensive, and subject to some bias (28,29). The strengths of a self-report format are itscost effectiveness, ease of measurement, its prioruse in transplant populations, and the concordancewith electronic monitoring and blood level mea-surements (2, 14). Self-report has therefore beenhighly recommended as the prominent methodol-ogy in both the clinical and research evaluation oftransplant-related medication non-adherence (2).By using an anonymous response format, webelieve we have addressed an important limitationof self-report surveys, that is, respondents mayfalsely communicate high levels of adherence toplease providers (18, 29).

The nine cases per 100 py of non-adherence toimmune suppressants in our study are comparable

to the average of seven per 100 person yearsreported in a recent meta-analyses (2). Our ques-tionnaire may have captured all magnitudes ofnon-adherence, some of which may have no clinicalconsequences. For example, missing a medicationonce in several weeks may not be as clinicallysignificant as consistent non-adherence. Muchvariability exists in literature with regards todefining non-adherence to immune suppressants(13). Studies use differing terminology and classifynon-adherence as: deliberate vs. inadvertent, covertvs. overt, major vs. minor, number of doses missed,mistiming of doses, any deviation from prescribeddoses, etc. (13). Defining and measuring medica-tion taking is therefore a persistent challenge inadherence research with no current consensus. Infact, a study that used multiple methods to assessmedication taking behaviors found prevalence ofnon-adherence to medications to be much higher(73%) so the use of a liberal definition is perhapsnecessary to better estimate non-adherence (30).The high and under-reported prevalence of non-adherence is further highlighted by O�Grady et al.when they attempt to validate the consensusstatement ‘‘One should assume that all people arenon-adherent, it is the degree to which they arenon-adherent that is important.’’ In clinical prac-tice, non-adherence to immune suppressants ismore often identified retrospectively when rejectionor graft loss occurs. Many publications thataddress adherence in liver recipients deal withgroups such as those with an etiology of alcoholiccirrhosis (12, 27, 31), ‘‘higher risk’’ patients iden-tified by missed clinic appointments (4), or pediat-ric LT recipients (28, 32, 33), which may not berepresentative of all LT recipients as a group. Incontrast, we include all adult LT recipients regard-less of clinical status, reason for transplant, orperception of their health care providers. There-fore, our findings are more likely to be represen-tative of adult LT recipients and thus may be moregeneralizable.Our finding of increased non-adherence in those

who reported a history of alcohol and substanceabuse and in those with mental health needs issimilar to some prior studies (30, 32, 34, 35). Thisincreased prevalence of non-adherence may berelated to psycho-social factors (patterns of coping,behavioral, and cognitive dysregulation, etc.) thathave been shown to impact adherence (30). Wesuggest a pro-active approach in identifying needsand support intervention strategies in these highrisk groups (30, 32, 34, 35). The increased trend inperceived prevalence of non-adherence in younger,working, and more educated patients is similar toprior reports in solid organ recipients (25, 26, 29).

Non-adherence to immune suppressants

333

Our study, also similarly, suggests that there isincreased non-adherence as more time elapses post-transplant, possibly because of less frequent officevisits and hence lesser reinforcement by health careproviders (34, 36). Unlike other studies, however,financial status of respondents did not impactadherence in our population, perhaps because of ahigh percentage of patients reporting adequatehealth insurance (37, 38).Inherent beliefs about health, physicians, and

medications are expected to impact adherence, butthese beliefs did not attain statistical significancein our study. Most LT respondents are aware ofthe short-lasting nature of their medications.While a large percentage of patients attributethe adverse effects, they experience to theirmedications, this does not seem to influenceadherence to medications.The correlation between missed office appoint-

ments and non-adherence to immune suppressantsfound in our study supports the findings from apilot LT, as well as a study in heart recipients (4,39). In the 2007 comprehensive meta-analysis byDew et al. (2), of the 30 LT publications, noneaddressed adherence to clinic appointments.Hence, our study is among the few that addressesclinic visits as an outcome measure as well as acorrelate to immune-suppressant adherence. Non-adherence to office appointments is multi-factorialand involves patient, provider, and systems-related factors. Liver recipients often havememory impairment, which may contribute toforgetting or being off the daily schedule, whichare cited as the main reasons for non-adherence byone-third of our non-adherent respondents. Keep-ing a medication log/diary may facilitate adherenceto both clinic visits and medications, as is supportedby our data. Systems-based interventions to mini-mize barriers and facilitate frequent, streamlinedoffice appointments may be helpful as well. Thishigher frequency of office visits may translate intobetter medication adherence (40).Our study has some limitations. First, because

our study was performed at an urban publichospital, our findings may not be applicable toother settings. Second, our survey instrument,while a well-accepted tool to measure adherenceto medications, is not the ‘‘gold standard.’’ Thus,estimates of non-adherence may differ from thoseobtained with other methods. Third, non-respond-ers in our study may differ from those whoresponded. It is possible that non-adherence maybe greater among the non-respondents in whichcase our estimates of non-adherence are likely to belower than the true rates. The anonymous natureof our survey precludes us from obtaining any

further information about non-respondents.Fourth, while our study examines patient-relatedvariables and non-adherence, health care providerand system-related issues of non-adherence werenot explored. Finally, as mentioned, our surveyinstrument question design did not allow for betterestimates of the extent of non-adherence and weuse a liberal definition for non-adherence. Whilesuch liberal definitions may illuminate on one handthe overall problem of immunosuppressant adher-ence, the clinical relevance may need to be furtherexamined.

The findings of this exploratory study providepotentially important information on issues havinga bearing on adherence to immune suppressants.We think that such research should be extended toa variety of settings and to larger studies. Largerstudies would permit examination of numerousvariables while lessening a limitation of our studynamely that with a relatively modest sample size(and many variables) a Type I error may be theexplanation for some sporadic statistically signifi-cant p-values.

Conclusion

Using a liberal definition, half of our surveyedadult liver recipients report non-adherence to theirimmune suppressants, which may be a biggerproblem than often recognized. Missed physicianoffice appointments may serve as important‘‘tip-off’’ in identifying non-adherence to immunesuppressants. Improving health care systems toincrease the frequency of physician office appoint-ments and pro-actively encouraging log-keepingmay be effective in decreasing non-adherence toimmune suppressants.

References

1. Adam R, McMaster P, O�Grady JG et al. Evolution ofliver transplantation in Europe: report of the EuropeanLiver Transplant Registry. Liver Transpl 2003: 9:1231.

2. Dew MA, DiMartini AF, De Vito Dabbs A et al. Ratesand risk factors for nonadherence to the medical regimenafter adult solid organ transplantation. Transplantation2007: 83: 858.

3. Osterberg L, Blaschke T. Adherence to medication. NEngl J Med 2005: 353: 487.

4. O�Carroll RE, McGregor LM, Swanson V, Mast-

erson G, Hayes PC. Adherence to medication after livertransplantation in Scotland: a pilot study. Liver Transpl2006: 12: 1862.

5. Cleemput I, Kesteloot K, Vanrenterghem Y,De Geest S. The economic implications of non-adherence after renal transplantation. Pharmaco-economics 2004: 22: 1217.

Lamba et al.

334

6. Haustein SV, McGuire BM, Eckhoff DE et al. Impactof noncompliance and donor/recipient race matching onchronic liver rejection. Transplant Proc 2002: 34:1497.

7. Dew MA, Dimartini AF. The incidence of nonadherenceafter organ transplant: ensuring that our efforts at count-ing really count. Liver Transpl 2006: 12: 1736.

8. DiMartini A, Day N, Dew MA et al. Alcohol use fol-lowing liver transplantation: a comparison of follow-upmethods. Psychosomatics 2001: 42: 55.

9. Gayowski T, Singh N, Keyes L et al. Late-onset renalfailure after liver transplantation: role of posttransplantalcohol use. Transplantation 2000: 69: 383.

10. Lucey MR, Carr K, Beresford TP et al. Alcohol useafter liver transplantation in alcoholics: a clinical cohortfollow-up study. Hepatology 1997: 25: 1223.

11. Pageaux GP, Bismuth M, Perney P et al. Alcohol re-lapse after liver transplantation for alcoholic liver disease:does it matter? J Hepatol 2003: 38: 629.

12. Osorio RW, Ascher NL, Avery M, Bacchetti P,Roberts JP, Lake JR. Predicting recidivism after ortho-topic liver transplantation for alcoholic liver disease.Hepatology 1994: 20: 105.

13. O�Grady JGM, Asderakis A, Bradley R et al. Multi-disciplinary insights into optimizing adherence after solidorgan transplantation. Transplantation 2010: 89: 627.

14. Garber MC, Nau DP, Erickson SR, Aikens JE,Lawrence JB. The concordance of self-report with othermeasures of medication adherence: a summary of theliterature. Med Care 2004: 42: 649.

15. Walsh JC, Mandalia S, Gazzard BG. Responses to a1 month self-report on adherence to antiretroviral therapyare consistent with electronic data and virological treat-ment outcome. AIDS 2002: 16: 269.

16. Haynes RB, Taylor DW, Sackett DL, Gibson ES,Baerholz CD, Mukherjee J. Can simple clinical mea-surements detect patient noncompliance? Hypertension1980: 2: 757.

17. Schafer-Keller P, Steiger J, Bock A, Denhaerynck

K, De Geest S. Diagnostic accuracy of measurementmethods to assess non-adherence to immunosuppressivedrugs in kidney transplant recipients. Am J Transplant2008: 8: 616.

18. Denhaerynck K, Manhaeve D, Dobbels F, Garzoni

D, Nolte C, De Geest S. Prevalence and consequences ofnonadherence to hemodialysis regimens. Am J Crit Care2007: 16: 222, quiz 236.

19. Butler JA, Roderick P, Mullee M, Mason JC,Peveler RC. Frequency and impact of nonadherence toimmunosuppressants after renal transplantation: a sys-tematic review. Transplantation 2004: 77: 769.

20. Chisholm MA, Lance CE, Williamson GM, Mulloy

LL. Development and validation of the immunosuppres-sant therapy adherence instrument (ITAS). Patient EducCouns 2005: 59: 13.

21. Russell CL, Kilburn E, Conn VS, Libbus MK,Ashbaugh C. Medication-taking beliefs of adult renaltransplant recipients. Clin Nurse Spec 2003: 17: 200, quiz209–230.

22. Medical Outcomes Study: Social Support Survey.Available at http://www.rand.org/health/surveys_tools/mos/mos_socialsupport.html. Accessed January 20, 2011.

23. Wallston KA. Multidimensional health locus of control(MHLC) scales. Available at http://www.vanderbilt.edu/

nursing/kwallston/mhlcscales.htm. Accessed January 20,2011.

24. Frazier PA, Davis-Ali SH, Dahl KE. Correlates ofnoncompliance among renal transplant recipients. ClinTransplant 1994: 8: 550.

25. Raiz LR, Kilty KM, Henry ML, Ferguson RM.Medication compliance following renal transplantation.Transplantation 1999: 68: 51.

26. Greenstein S, Siegal B. Compliance and noncompliancein patients with a functioning renal transplant: a multi-center study. Transplantation 1998: 66: 1718.

27. Beresford TP, Schwartz J, Wilson D, Merion R,Lucey MR. The short-term psychological health of alco-holic and non-alcoholic liver transplant recipients. AlcoholClin Exp Res 1992: 16: 996.

28. Shemesh E, Shneider BL, Savitzky JK et al. Medicationadherence in pediatric and adolescent liver transplantrecipients. Pediatrics 2004: 113: 825.

29. Drent G, Haagsma EB, Geest SD et al. Prevalence ofprednisolone (non)compliance in adult liver transplantrecipients. Transpl Int 2005: 18: 960.

30. Stilley CS, DiMartini AF, de Vera ME et al. Individ-ual and environmental correlates and predictors of earlyadherence and outcomes after liver transplantation. ProgTransplant 2010: 20: 58.

31. Berlakovich GA, Langer F, Freundorfer E et al.General compliance after liver transplantation for alco-holic cirrhosis. Transpl Int 2000: 13: 129.

32. Lurie S, Shemesh E, Sheiner PA et al. Non-adherence inpediatric liver transplant recipients–an assessment of riskfactors and natural history. Pediatr Transplant 2000: 4:200.

33. Berquist RK, Berquist WE, Esquivel CO, Cox KL,Wayman KI, Litt IF. Adolescent non-adherence: preva-lence and consequences in liver transplant recipients. Pe-diatr Transplant 2006: 10: 304.

34. Hansen R, Seifeldin R, Noe L. Medication adherence inchronic disease: issues in posttransplant immunosuppres-sion. Transplant Proc 2007: 39: 1287.

35. Dew MA, Kormos RL, Roth LH, Murali S, DiMartini

A, Griffith BP. Early post-transplant medical compli-ance and mental health predict physical morbidity andmortality one to three years after heart transplantation. JHeart Lung Transplant 1999: 18: 549.

36. Chisholm-Burns MA, Kwong WJ, Mulloy LL, Spivey

CA. Nonmodifiable characteristics associated withnonadherence to immunosuppressant therapy in renaltransplant recipients. Am J Health Syst Pharm 2008: 65:1242.

37. Kasiske BL, Cohen D, Lucey MR, Neylan JF. Paymentfor immunosuppression after organ transplantation.American Society of Transplantation. JAMA 2000: 283:2445.

38. Willoughby LM, Fukami S, Bunnapradist S et al.Health insurance considerations for adolescent transplantrecipients as they transition to adulthood. Pediatr Trans-plant 2007: 11: 127.

39. Stilley CS, Lawrence K, Bender A, Olshansky E,Webber SA, Dew MA. Maturity and adherence in ado-lescent and young adult heart recipients. Pediatr Trans-plant 2006: 10: 323.

40. Shemesh E, Annunziato RA, Shneider BL et al.Improving adherence to medications in pediatric livertransplant recipients. Pediatr Transplant 2008: 12: 316.

Non-adherence to immune suppressants

335