How “Bad” Does the Pain Have to Be? AQualitative Study Examining Adherence to PainMedication in Older Adults With OsteoarthritisJOANNA E. M. SALE,1 MONIQUE GIGNAC,2 AND GILLIAN HAWKER3

Objective. To explore the experience of adherence to pain medication in older adults with osteoarthritis (OA).Methods. Individuals were recruited from an existing cohort (n � 1,300) of persons with disabling hip and knee OA.Twenty-seven individuals who reported previous physician visits for their arthritis, spoke English, were Torontoresidents, and were receptive to in-depth interviews were approached by the cohort telephone interviewer to discuss theirexperiences with prescribed painkillers for OA. Semistructured face-to-face interviews were conducted by a qualitativeresearcher in participants’ homes.Results. Nineteen adults (10 women, 9 men) ages 67–92 years were interviewed for 1–3 hours. Participants varied in theirsocioeconomic status and education levels. Most had comorbidities, such as heart disease and diabetes, for which theywere also being treated. Findings indicated that adherence to pain medication differed from that of other prescribedmedications. Participants were reluctant to take painkillers, and when they did, they generally took them at a lower doseor frequency than prescribed. This behavior did not reflect their recommendations for others, who they expected to betreated appropriately for pain and to adhere to pain medication. Perceptions and attitudes to pain played an integral rolein participants’ adherence to painkillers. Despite obvious physical limitations, participants minimized their pain andclaimed to have a high pain tolerance.Conclusion. These findings suggest that reevaluation of the prescription of pain medication for OA is warranted and thatthe effectiveness of pain management in OA needs to account for adherence behavior in older adults.

KEY WORDS. Osteoarthritis; Pain; Qualitative research; Adherence; Painkillers; Medication; Attitudes.

INTRODUCTION

Osteoarthritis (OA) is the most common type of arthritis(1,2), with the number of individuals experiencing OA

expected to double by 2020 (3). Although there is no curefor OA, there are efficacious drug and lifestyle treatmentsthat can reduce pain and improve physical functioning.Failure to adhere to treatment recommendations meansthat individuals may achieve suboptimal symptom reliefand incur personal, health, and economic costs (4).

Studies have found that �50% of individuals withchronic disease are adherent to medication recommenda-tions from health professionals regardless of disease, treat-ment, or age (5). Adherence to arthritis medication is im-proved when individuals are older (6), female (6,7), have ahigher perceived health state (7), and when there is a clearstatement of the drug’s purpose (8). Greater comorbidityinterferes with adherence (9), as do higher and more fre-quent dosing regimens and concerns about addiction (7).However, previous research has been based mostly ondrug studies in clinical samples. Few studies have exam-ined treatment adherence from the perspective of those inthe community living with OA. Even fewer studies haveexamined adherence to OA pain medication using quali-tative methods. In this study, we explored the experienceof adherence to pain medication in older adults with OA

Supported by the Canadian Institutes of Health Research(CIHR) and the Canadian Arthritis Network (postdoctoralresearch awards) and the Orthopedic & Arthritis Institute(project-related costs). Dr. Hawker received support as aCIHR Scientist and as the F. M. Hill Chair in AcademicWomen’s Medicine.

1Joanna E. M. Sale, PhD: Sunnybrook & Women’s CollegeHealth Sciences Centre, Toronto, Ontario, Canada; 2MoniqueGignac, PhD: The University Health Network & Departmentof Public Health Sciences, The University of Toronto, To-ronto, Ontario, Canada; 3Gillian Hawker, MD, MSc: Sunny-brook & Women’s College Health Sciences Centre, and TheUniversity of Toronto, Toronto, Ontario, Canada.

Address correspondence to Joanna E. M. Sale, PhD, Ca-nadian Osteoarthritis Research Program, Women’s CollegeAmbulatory Care Center, Sunnybrook & Women’s CollegeHealth Sciences Centre, 76 Grenville Street, Room 812B,Toronto, Ontario, Canada M5S 1B2. E-mail: joanna.sale@sw.ca.

Submitted for publication August 2, 2005; accepted inrevised form November 3, 2005.

Arthritis & Rheumatism (Arthritis Care & Research)Vol. 55, No. 2, April 15, 2006, pp 272–278DOI 10.1002/art.21853© 2006, American College of Rheumatology

SPECIAL ARTICLE: RHEUMATIC DISEASE THROUGH THE LIFESPAN

272

because the greatest burden of symptomatic OA is in thisage group.

PATIENTS AND METHODS

The study of experiences lends itself to phenomenology(10,11). An eidetic phenomenologic approach was utilizedto describe the essence of the phenomenon (adherence topain medication) as the participants experienced it in theirdaily lives (12). “Bracketing,” setting aside one’s judge-ments, biases, and preconceived ideas about the phenom-enon (13), is one component of eidetic phenomenology.Having conducted an extensive literature review on treat-ment adherence in OA, it was important for us to suspendpreconceptions derived from existing research and recog-nize that adherence may be a reasoned decision by pa-tients (14). For example, medication adherence is influ-enced by variables such as fear of addiction. However, wedid not introduce this topic when questioning partici-pants. Instead, when the issue of fear of addiction wasraised by a participant, it was probed for further detail.The interviewer also remained neutral throughout the in-terviews and did not reinforce or discourage any topicsdiscussed that were related to medication usage.

Participants were recruited from an existing cohort of1,300 individuals with hip and/or knee OA residing inOntario, Canada (15). Consistent with qualitative research,sampling was purposeful. Potential candidates were iden-tified by 2 telephone interviewers who acted as key infor-mants. These interviewers had established a relationshipwith participants through interviews over the previous 5years. Candidates were considered suitable for interview-ing if they had been articulate in previous interviews andexpressed interest in participating in additional research.Twenty-seven men and women who resided in metropol-itan Toronto, spoke English, and had reported previousphysician visits for their OA were identified. Individualswho had undergone a hip or knee replacement in the prioryear were not eligible for recruitment.

Data were collected through face-to-face interviews overa 4–5-month period. Because of their age and health, par-ticipants were interviewed in their homes where they werelikely to feel relaxed and comfortable (16). Before eachinterview, participants were asked to gather all medica-tions (prescription and nonprescription) and the inter-viewer (JS) discussed with them the doses and timing ofdoses for each medication. The structured component ofthe interview started with the following question: “Tell meabout your OA. What is it like for you at this time?” Thisintroduction was followed by, “What are you doing foryour OA (e.g., what medications are you taking)? Whatother medications are you taking?” A number of probesencouraged participants to elaborate on their experienceswith adherence to pain medication. These included: “Doyou take/use the medication in the way the health profes-sional suggested? Why/why not?”; “Do you take your med-ications regularly? Why/why not?”; “Some people havetold us that they don’t take their OA medications rightaway or in the dosage prescribed. What is your advice topeople who adjust the timing/dosage of medications?”

Interviews were audiotaped and transcribed verbatiminto Microsoft Word (Microsoft, Redmond, WA). As rec-ommended by Kvale (17), the transcripts were verifiedagainst the tapes and downloaded in NVivo (18), a quali-tative program with flexible features that helped organize,code, and retrieve data.

Analysis of the transcripts began after the first couple ofinterviews and was an iterative process whereby codeswere identified immediately and then revised as moreinterviews were conducted. Analysis was conducted byone author (JS) as is customary with phenomenologic stud-ies (19,20). However, discussions regarding additionalprobes and emerging themes were reviewed by all theauthors as data collection and analysis progressed. Datawere analyzed according to Giorgi’s procedures (19,21). Atotal of 47 codes emerged from participants’ descriptionsof their adherence to pain medication (e.g., “prescribedmedication,” “nonprescribed medication,” “barriers to ad-herence,” “comorbidities,” “pain,” “altering doses,” “min-imizing OA”). These codes were organized into distinctthemes.

RESULTS

Nineteen participants (10 women and 9 men) ages 67–92years were interviewed (7 refused and 1 was too ill toparticipate). As is common in phenomenologic research, 5individuals were interviewed twice for a total of 24 inter-views lasting 1–3 hours each. Additional interviews clar-ified topics discussed in the first interviews. In severalcases, participants had a relative or aid present. We en-couraged these persons to give us privacy during the in-terview but did not insist that they leave. By the 24thinterview, no new thematic information was introduced.This sample size is similar to those recommended forphenomenologic studies (22,23). All participants werewhite but varied in their education level (high schooleducation or less to postsecondary education); most hadcomorbidities, such as heart disease and diabetes, forwhich they were also being treated. Although our inter-views focused on adherence to pain medication, partici-pants also talked about other strategies they used for painmanagement, which included use of health professionalssuch as chiropractors and physiotherapists, applying heatand/or ice, using devices, resting, having a positive atti-tude, learning to live with pain, and modifying certainactivities/movements. Ten participants were currentlyprescribed both a pain medication and a nonsteroidal an-tiinflammatory drug (NSAID) for OA, 1 participant re-ported no prescribed medication for OA, 6 were prescribedpain medication only, and 2 were prescribed NSAIDs only.

Two themes characterized the adherence experience(Table 1). First, adherence to pain medication differedfrom adherence to other prescribed medications. Partici-pants were reluctant to take painkillers, and when theydid, they generally took them at a lower dose or frequencythan was prescribed. Second, perceptions of and attitudestoward pain played an integral role in participants’ adher-ence to painkillers. In general, despite their physical lim-itations, participants minimized their pain and claimed tohave a high pain tolerance.

Pain Medication in Osteoarthritis 273

Theme 1: adherence to pain medication differed fromthat of other prescribed medications. Pain medication forarthritis ranged from over-the-counter treatments to pre-scription strength and were often labeled “take 4–6 timesa day, or as needed.” With the exception of 4 participantswho took their pain medications as prescribed (and re-ported varying pain relief levels), the remaining partici-pants treated their pain medications differently from othermedications. Specifically, participants were focused onthe “take as needed” or “as required” instruction of theprescription. Participants were generally adherent to theirother medications (although participants admitted thatthey sometimes forgot to take them); however, they pur-posefully did not take their OA pain medication as pre-scribed. When they took their pain medication, they took itin a lower dose or frequency than prescribed. Behaviorsunique to pain medication were also exhibited. For exam-ple, unlike medications for other conditions, pain medica-tions were not included in pill organizers. One participantfilled her prescription for Percodan and then threw awaythe bottle. Another participant recorded on paper everyplain Tylenol she took. This participant also filled a higherdose bottle (Extra Strength Tylenol, 500 mg) with a lowerdose pill (plain Tylenol, 325 mg) to keep her pain medi-cation intake to a minimum. Several participants regularlyrationed their pain medication, especially when theamount of medication was getting low. They did not dothis with other medications.

Participants cited several reasons for not wanting to takepainkillers. Most claimed that they did not like to takepills in general (several were taking �20 pills per day) andthat not taking their pain medication signified 1 less pill intheir regimen. However, 18 of the 19 participants weretaking at least 1 herbal remedy and/or vitamin for theirarthritis, and 2 participants took 1 aspirin per day for“general health” reasons. Some participants claimed thatthey did not take their over-the-counter pain medicationbecause it was not relieving their pain. However, none ofthese individuals took the maximum dose allowed.

Fear of addiction to painkillers was a concern for severalparticipants. Participants stated, “I don’t take things likethat” (Percodan); “That is a hard drug . . . I think of thedruggies on the street” (Percocet); “When you take anykind of pill for any length of time, your system gets used tothem and they more or less expect them . . . it gets to be ahabit more than a necessity” (Tylenol 3 with codeine).Some participants were not as forthcoming and soughtsupport for their decisions from the interviewer. For ex-ample, 2 participants who claimed to not like taking painmedication on a regular basis said to the interviewer, “. . .you know what I mean?” (Tylenol 2) or, “. . . you under-stand what I’m saying?” (Tylenol 3 with codeine). Oneparticipant did not want to admit to her son that she wasafraid of addiction. She told her son that she did not takeher Tylenol 2 with codeine as prescribed because it wasnot effective and that she was afraid of constipation. When

Table 1. Organization of codes (“meaning units”) into themes*

Theme 1: Adherence to pain medication differs from otherprescribed medications

Theme 2: Perceptions and attitudes toward pain playan integral role in adherence to pain medication

Participant discusses: Participant discusses:Prescribed medication for OA PainPrescribed medication for other conditions Participant’s description of painNonprescribed medication for OA Pain toleranceNonprescribed medication for other conditions Minimizes painAltering doses Uses humor

Increases doses Pain as a part of lifeAttitude toward altering doses Fighting the painSeeks approval for altering doses Treating pain only after it is experienced

Behaviors unique to pain medication Time and place to treat painPill organizers Other strategies to manage OA painTakes less pain medication than prescribed Activity restrictionsReaction to side effects of pain versus other medication Devices

AvoidanceRationingRestingContradictions about adhering to prescriptionPositive attitudeBarriers to adherenceApplies heat/iceNegative attitude toward medicationsExercisesComorbiditiesHealth care professionals, e.g., physiotherapistFear of addictionOther coping behaviors e.g., cup of teaFear of running out of medication

SleepNegotiations with health care providerHobbiesAdvice to others regarding pain medicationParticipant experiments on ownMisconceptions about treating painDiscusses how healthy they areBoasts about healthCompares self with others

* OA � osteoarthritis.

274 Sale et al

the son recalled this information during the interview, theparticipant leaned towards the interviewer and whispered,“I don’t want to get addicted either.”

Interestingly, the practice of restricting one’s pain med-ication intake was not recommended to friends and/orfamily members. During one interview, a participant whowould not take pain medication reprimanded his wife whoalso had arthritis for not taking her pain medication asprescribed. He told her that if she wanted to relieve thepain, she needed to take more than just 1 aspirin or morethan 2 Tylenol in a day. In an argument that followed heradmission to taking 1 aspirin that day, he said, “What goodis 1 aspirin?”

As part of the interview, we asked participants a hypo-thetical question about persons who alter the doses of theirarthritis medications. Most interviewees claimed that itwas irresponsible to alter medication doses. For example,one woman who was taking 2 of 3 prescribed Tylenol 3 perday told us, “I don’t fool around with my medications. Iknow a few people who have . . . bless them . . . they arenot around any longer.” Upon probing, it became apparentthat all of these participants assumed that altering dosesmeant taking too much, rather than too little pain medica-tion. Taking too little medication was not equated withnonadherence.

Ironically, the treatment of pain medication did notapply to higher-dose NSAIDs such as rofecoxib and cele-coxib. Although participants indicated that they were try-ing to cut down or restrict their use of their pain medica-tions, none indicated that they wanted to cut down onNSAID doses. Participants took these as prescribed andonly altered their NSAID doses or stopped taking themwhen they felt these medications were not working or theirside effects became bothersome. Participants were seem-ingly unaware that their NSAIDs had analgesic properties.

Theme 2: perceptions and attitudes to pain played anintegral role in adherence to pain medication. Percep-tions of pain played an integral role in participants’ adher-ence to pain medication. Despite obvious physical limita-tions, participants belittled their pain, often using humorto lighten previous admissions of suffering. One 92-year-old woman described her pain as so unbearable that itprevented her from tying her shoelaces or raising herhands to style her hair. Later, she claimed that she “wasn’tas bad as lots of people.” Another woman who could notsleep at night because of her OA pain said, “As long as Ican move it [her leg] and hobble around, I’m happy.” One84-year-old man who was confined to a wheelchair be-cause of his OA said about his pain, “My doctor is very,very pretty so I never ask any questions.” After talkingabout how discouraged he and his wife had become abouthis condition, he said, “Other than that, I am in fineshape!” and laughed. A 75-year-old man implied that painbelonged in his life. He said, “. . . that’s how you knowyou’re alive . . . you ache . . . I woke up one morning anddidn’t have an ache or pain in my body . . . I thought I wasdead (laughter).” Another participant said, “I always havepain. That’s part of my life.”

Some participants claimed that their low pain medica-tion intake was due to a high pain tolerance. For example,

“. . . it’s tolerable. You learn to live with it. I’m not crip-pled . . . it just hurts to walk. Then you walk wrong be-cause you’re trying to . . . ease the pain as much as possi-ble so you don’t walk properly balanced. Then the otherknee starts acting up. It’s a losing battle (laughter). But youhave to be tolerant.” Another participant who did not wantthe pain to “rule [her] life” felt that she could tolerate acertain level of pain and that pain medication would onlymask, rather than help, her mobility problem. One partic-ipant said, “I used to take [painkillers] more often. I thinkI was a little softer then . . . I don’t take them unless I amreally upset about my pain.” This woman could not walkand woke frequently at night because of her OA pain; shemodified many of her daily activities so that she couldcope without taking pain medication, e.g., she rarely lefthome, performed daily chores at a much slower pace, satin a comfortable armchair all day. She longed to relax in ahot bath but was unable to do so because her OA made itdifficult to get into and out of the bathtub. Another partic-ipant who had difficulty standing because of her OA (herhomecare worker greeted the interviewer at the front door)told us that she was in pain from the moment she woke upin the morning until she went to sleep at night, but that shewas able to tolerate this pain: “I can stand more pain thanmost people because I started off when I was a little kidwith the polio . . . my bowels were paralyzed and that wasextremely painful . . . I just got over that when I tookscarlet fever and . . . my head swelled up and I had amastoid . . . the surgeon had to come . . . and operated onmy ear on the kitchen table . . . I remember it vividly too. . . so I have had a lot of pain in my life.”

Fighting the pain allowed interviewees to build up theirpain tolerance. One participant said, “if that’s the way it’sgoing to be, you’re not going to let it knock you down . . .you’re going to fight it . . . I think maybe that’s bettertherapy than all the damned medicines you can take.”Another commented, “I think that you can give into painso easily if you don’t watch it.”

Although many participants had been recommended totake their painkillers in advance of an activity that mightcause pain, and agreed that this was a good recommenda-tion, none reported taking pain medication in this manner.Participants claimed they would take painkillers onlywhen the pain became “very bad.” For example, one 67-year-old man recalled an instance when he took a Tylenol:“I was at a point where I couldn’t stand up in the kitchenyou know, to do any cooking or make a coffee. I washanging on to the counter you know. I couldn’t get my legscomfortable.” One woman reported only receiving 4 hoursof sleep a night because of OA pain. She would wake up at2:00 AM, make a cup of tea, and apply heat to her knee.Only if the pain remained excruciating would she take athird Tylenol 3 as prescribed. One participant with a pre-scription for Tylenol 2 to be taken every 4 hours took 1 pillevery morning. She would only take a second pill onThursdays when returning from shopping. Others had dif-ficulty determining what “bad” was. One woman said,“[before going out] . . . I would probably have the aspirinwith me in my purse and if I started to feel that [the pain]was holding me back, I would take it.” However, sheestimated that she had not taken an aspirin in the last 2

Pain Medication in Osteoarthritis 275

years. Another woman said, “I have seen me when I can’teven get out of bed in the morning. But I don’t take aTylenol then. I just walk to the kitchen. My latest trick is tomake a cup of tea.”

Finally, certain types of pain were believed to warrantmedication. For example, one participant thought her auntwho was screaming on her “death bed” for pain medica-tion should have been given morphine. Other participantshad no opposition to taking pain medications as pre-scribed for short-term reasons such as surgery or recuper-ation from falls.

DISCUSSION

The individuals in this study perceived and behaved to-ward their pain medication differently than other medica-tions. They took it in lower doses and less frequently thanprescribed and appeared invested in taking pain medica-tion only when they felt it was absolutely necessary. Thismay partly explain why few participants appeared to beachieving pain relief. Other research finds that a substan-tial percentage of individuals with OA taking NSAIDs oracetaminophen continue to report pain (24–26); the re-sults of those studies may be due to adherence behaviorssimilar to those demonstrated in our findings. We discussour findings, comparing them with the literature on med-ication testing, activity restriction, and fear of addiction.We then discuss participants’ willingness to take NSAIDsand other herbal remedies that may contain pain-relievingproperties. We close by briefly addressing disease accep-tance and the symbolic role of pain medication as well asthe instruction “take as needed,” which appears on thelabels of many over-the-counter and prescription-strengthpain medication.

Patients with a range of health conditions have beenfound to hold opinions about the value of their medica-tions and use explicit or implicit testing processes to es-tablish their limits and levels of needed relief (27,28).However, the participants in our sample did not appear tobe testing their pain medication so much as choosing notto take it at all. These findings are consistent with those ofRoss et al (29), who found that senior adults with muscu-loskeletal pain considered pain medication a last resort formanagement of pain. To justify low pain medication in-take, participants minimized pain or claimed to have ahigh pain tolerance. This strategy had limited successbecause most participants experienced significant ongoingpain. Many maintained functioning by restricting theiractivities. The failure to accept pain and choice to restrictactivity is a coping response referred to as “subversion”(30). Unfortunately, this response comes at a cost to olderadults as increasing numbers of activity limitations mayhave social and psychological consequences for well-being(31). Moreover, from a clinical perspective, activity restric-tions may be a marker for poor adherence to pain medica-tion. By asking their patients about activity restrictions,clinicians may be able to identify those individuals whoare not managing their pain due to inadequate levels ofpain medication.

Fear of addiction was a concern for some participants,although this concern was not always explicitly articu-

lated. Fear of addiction to pain medications has been re-ported elsewhere (32–34). However, this fear may be un-warranted because prescriptions of stronger painmedications, such as opioids, have not been found to leadto addiction (35,36) and appear to be successful in treatingpain in the elderly (37,38). Interventions to educate per-sons with OA about safe doses of opioids might benefitthose with addiction concerns.

A number of participants claimed they avoided painkill-ers because it meant 1 less pill in their daily regimen.Given the many medications taken and the pain symptomsexperienced, it is unclear why pain medication was sin-gled out as the medication to avoid. It may be that partic-ipants prioritized their health conditions and believedthat, because their pain was not life threatening and wastolerable, it could be sacrificed.

Unknowingly, participants were taking pain medicationas part of their NSAIDs. Because these were taken as pre-scribed, it may partly explain why many patients with OAidentified NSAIDs as more helpful than other analgesics(25). Furthermore, if acetaminophen in maximum dailydoses is less effective than NSAIDs for pain relief (24),then our findings suggest that individuals with OA whorely on acetaminophen alone may achieve even less painrelief because they are not taking it as prescribed. At thesame time, participants were using other herbal and alter-native remedies for their pain. Willingness to take herbalmedicines over pain medication has been reported by oth-ers (14). Ironically, the herbal industry in Canada is notregulated, so participants were not aware of the ingredi-ents of many of these medicines. According to one source(39), some products (e.g., devil’s claw, glucosamine, chon-droitin) have pain-relieving properties and may contain apainkiller and/or NSAID component. Therefore, partici-pants were probably achieving some analgesic effect fromtheir daily doses of herbal medication. Future research isneeded to determine whether pain medication usage isaffected by knowledge that other substances, such asherbal remedies, have analgesic qualities.

Studies show that medication adherence is influencedby subjects’ acceptance of their disease or condition(28,30,33). In one study on asthma, Adams et al (33) foundthat “deniers” (those who did not assimilate their condi-tion as part of their personal identity) claimed that theircondition had no effect on their lives and rarely tookreliever medications. Participants in our study did notdeny having OA, but rather denied that the pain associatedwith their arthritis warranted treatment. Dowell and Hud-son’s research (28) on medication adherence indicated thatunderstanding the symbolic role of pain medication andhow pain challenges an individual’s identity is key toaddressing low adherence. If individuals with OA do notconsider OA pain to be integral to their illness, they maynot recognize the need for medication to specifically treatOA pain. A similar finding has been established in otherresearch where elderly individuals who were potentialcandidates for total joint arthroplasty were unwilling toundergo the procedure partly because they viewed OA asa normal part of aging (40).

Finally, the instruction “take as needed” or “as re-quired” appeared to give license to many participants to

276 Sale et al

alter medication dose and may have reinforced their desireto minimize their pain. Specifically, for most participants,pain was not considered “bad” enough to warrant painmedication despite the fact that it impeded daily activities.This finding suggests that instructions for pain medicationneed to be revisited and the “take as needed” instructionsomitted or clarified by clinicians. Participants might alsobe more likely to take pain medication if they are labeled“arthritis medication” as opposed to “painkillers.”

There are some limitations to our study. Our sample waspurposive. Individuals who were not receptive to inter-views may have articulated different concerns and/or ad-herence behaviors. Future research should pursue the is-sues raised in this research in other samples that include agreater range of ages and cultural backgrounds. We alsodid not consider the physician’s perspective in this study.It is possible that patients’ reluctance to take pain medi-cation reflected their physicians’ concerns about addictionand/or medication tolerance. It is also possible that phy-sicians were not clear or consistent in their instructionsregarding pain medication, especially over-the-counterproducts. Adherence to other therapies for OA (e.g., topi-cal ointments, physical therapy, herbal remedies) was notexamined. Additional questions about the role of healthprofessionals in adherence, as well as use of other types oftreatment would provide additional context for these re-sults. As with all qualitative research, investigators with adifferent perspective or qualitative tradition (e.g., herme-neutic phenomenology, grounded theory) may have dif-fered somewhat in their analysis and interpretation of thedata collected. At the same time, we have provided nu-merous statements in participants’ own words to illumi-nate the themes presented.

Despite these limitations, this study found that individ-uals with arthritis have clearly articulated reasons for de-ciding to alter their pain medication. These decisions needto be considered when prescribing pain medication for OAand when examining the effectiveness of OA pain man-agement in older adults. Based on our findings, a numberof implications have been discussed. Activity restrictionsmay be a marker for poor adherence to pain medicationand may identify those individuals who are not takingpain medications as prescribed. Addiction concerns maybe alleviated by interventions to educate persons with OAabout safe doses of opioids. Participants were unknow-ingly achieving some analgesic effect from their dailydoses of NSAIDs and possibly their herbal medications.Future research is needed to determine whether knowl-edge that these substances may contain pain-relievingproperties would further affect pain medication usage.Finally, instructions for pain medication need to be revis-ited and perhaps omitted to encourage older adults withOA to take pain medication in doses that provide painrelief.

ACKNOWLEDGMENTThe authors are indebted to the participants interviewedfor this study.

REFERENCES

1. Felson DT, Naimark A, Anderson J, Kazis L, Castelli W,Meenan RF. The prevalence of knee osteoarthritis in theelderly: the Framingham Osteoarthritis Study. ArthritisRheum 1987;30:914–8.

2. Forman MD, Malamet R, Kaplan D. A survey of osteoarthritisof the knee in the elderly. J Rheumatol 1983;10:282–7.

3. Badley EM. Population projections and the effect on rheuma-tology. Ann Rheum Dis 1991;50:3–6.

4. Sokol MC, McGuigan KA, Verbrugge RR, Epstein RS. Impactof medication adherence on hospitalization risk and health-care cost. Med Care 2005;43:521–30.

5. Dunbar-Jacob J, Erlen JA, Schlenk EA, Ryan CM, Sereika SM,Doswell WM. Adherence in chronic disease. Annu Rev NursRes 2000;18:48–90.

6. Viller F, Guillemin F, Briancon S, Moum T, Suurmeijer T, vanden Heuvel W. Compliance with drug therapy in rheumatoidarthritis: a longitudinal European study. Joint Bone Spine2000;67:178–82.

7. De Klerk E, van der Heijde D, Landewe R, van der Tempel H,Urquhart J, van der Linden S. Patient compliance in rheuma-toid arthritis, polymyalgia rheumatica, and gout [publishederratum appears in J Rheumatol 2003;30:423]. J Rheumatol2003;30:44–54.

8. Daltroy LH, Katz JN, Liang MH. Doctor-patient communica-tion and adherence to arthritis treatments [abstract]. ArthritisCare Res 1992;5 Suppl 4:S19.

9. Davis GC, Hiemenz ML, White TL. Barriers to managingchronic pain of older adults with arthritis. J Nurs Scholarsh2002;34:121–6.

10. Schwandt TA. Dictionary of qualitative inquiry. 2nd ed.Thousand Oaks (CA): Sage Publications; 2001.

11. Sokolowski R. Introduction to phenomenology. Cambridge:Cambridge University Press; 2000.

12. Morse JM. Designing funded qualitative research. In: DenzinNK, Lincoln YS, editors. Handbook of qualitative research.Thousand Oaks (CA): Sage Publications; 1994. p. 220–35.

13. Moustakas C. Phenomenological research methods. ThousandOaks (CA): Sage Publications; 1994.

14. Donovan JL, Blake DR. Patient non-compliance: deviance orreasoned decision-making? Soc Sci Med 1992;34:507–13.

15. Hawker GA, Wright JG, Coyte PC, Williams JI, Harvey B,Glazier R, et al. Determining the need for hip and kneearthroplasty: the role of clinical severity and patients’ prefer-ences. Med Care 2001;39:206–16.

16. Rice PL, Ezzy D. Qualitative research methods. South Mel-bourne, Australia: Oxford University Press; 2000.

17. Kvale S. Interviews: an introduction to qualitative researchinterviewing. Thousand Oaks (CA): Sage Publications; 1996.

18. NVivo. Victoria, Australia: Qualitative Solutions and Re-search Pty Ltd; 1999.

19. Giorgi A. The theory, practice, and evaluation of the phenom-enological method as a qualitative research procedure. J Phe-nomenol Psychol 1997;28:235–60.

20. Crotty M. Phenomenology and nursing research. South Mel-bourne, Australia: Churchill Livingstone; 1996.

21. Giorgi A. Some theoretical and practical issues regarding thepsychological phenomenological method. Saybrook Rev1989;7:71–85.

22. Polkinghorne DE. Phenomenological research methods. In:Valle RS, Halling S, editors. Existential-phenomenologicalperspectives in psychology. New York: Plenum Press; 1989.p. 41–60.

23. Creswell JW. Qualitative inquiry and research design. Thou-sand Oaks (CA): Sage Publications; 1998.

24. Geba GP, Weaver AL, Polis AB, Dixon ME, Schnitzer TJ, andthe Vioxx, Acetaminophen, Celecoxib Trial (VACT) Group.Efficacy of rofecoxib, celecoxib, and acetaminophen in osteo-arthritis of the knee: a randomized trial [published erratumappears in JAMA 2002;287:989]. JAMA 2002;287:64–71.

25. Pincus T, Swearingen C, Cummins P, Callahan LF. Preferencefor nonsteroidal antiinflammatory drugs versus acetamino-

Pain Medication in Osteoarthritis 277

phen and concomitant use of both types of drugs in patientswith osteoarthritis. J Rheumatol 2000;27:1020–7.

26. Silverstein FE, Faich G, Goldstein JL, Simon LS, Pincus T,Whelton A, et al. Gastrointestinal toxicity with celecoxib vsnonsteroidal anti-inflammatory drugs for osteoarthritis andrheumatoid arthritis: the CLASS study: a randomized con-trolled trial. Celecoxib Long-term Arthritis Safety Study.JAMA 2000;284:1247–55.

27. Conrad P. The meaning of medications: another look at com-pliance. Soc Sci Med 1985;20:29–37.

28. Dowell J, Hudson H. A qualitative study of medication-takingbehaviour in primary care. Fam Pract 1997;14:369–75.

29. Ross MM, Carswell A, Hing M, Hollingworth G, Dalziel WB.Seniors’ decision making about pain management. J Adv Nurs2001;35:442–51.

30. Miles A, Curran HV, Pearce S, Allan L. Managing constraint:the experience of people with chronic pain. Soc Sci Med2005;61:431–41.

31. Gignac MA, Cott C, Badley EM. Adaptation to chronic illnessand disability and its relationship to perceptions of indepen-dence and dependence. J Gerontol B Psychol Sci Soc Sci2000;55:P362–72.

32. Pound P, Britten N, Morgan M, Yardley L, Pope C, Daker-White G, et al. Resisting medicines: a synthesis of qualitativestudies of medicine taking. Soc Sci Med 2005;61:133–55.

33. Adams S, Pill R, Jones A. Medication, chronic illness andidentity: the perspective of people with asthma. Soc Sci Med1997;45:189–201.

34. Paice JA, Toy C, Shott S. Barriers to cancer pain relief: fear oftolerance and addiction. J Pain Symptom Manage 1998;16:1–9.

35. Portenoy RK, Foley KM. Chronic use of opioid analgesics innon-malignant pain: report of 38 cases. Pain 1986;25:171–86.

36. Joranson DE, Ryan KM, Gilson AM, Dahl JL. Trends in med-ical use and abuse of opioid analgesics. JAMA 2000;283:1710–4.

37. Freedman GM. Chronic pain: clinical management of com-mon causes of geriatric pain. Geriatrics 2002;57:36–42.

38. Anonymous. AGS clinical practice guidelines: the manage-ment of chronic pain in older persons. Geriatrics 1998;53Suppl:S8–25.

39. Soeken KL. Selected CAM therapies for arthritis-related pain:the evidence from systematic reviews. Clin J Pain 2004;20:13–8.

40. Hudak PL, Clark JP, Hawker GA, Coyte PC, Mahomed NN,Kreder HJ, et al. “You’re perfect for the procedure! Why don’tyou want it?” Elderly arthritis patients’ unwillingness to con-sider total joint arthroplasty surgery: a qualitative study. MedDecis Making 2002;22:272–8.

278 Sale et al