Osteoarthritis (OA), the most common form of arthritis, is a heterogeneous disease characterized by multi-tissue failure of the synovial joints and an illness defined by patient reported symptoms (e.g., pain, stiffness) [1]. Symptom management, is important to an individual's quality of life and function, and is the primary focus of current OA treatment [2]. OA symptom management includes both professionally-guided as well as self-guided care interventions [3, 4]. Self-guided care is undertaken based on information obtained from professionals and other personal experiences [3] and has long been recognized for its importance in chronic disease management [5]. Self-guided care for OA includes over-the-counter (OTC) medications, prescription drugs not used for their prescribed purpose, and dietary supplements that individuals take to ameliorate their symptoms [3]. Professionally-guided care typically consists of prescription and OTC medications, injections, and physical therapy prescribed by physicians [3].

In the United States, it is not fully understood how OA patients manage their symptoms with professionally-guided and self-guided pharmacological care strategies. Among patients taking a non-steroidal anti-inflammatory drug (NSAID) for management of their joint discomfort (e.g., rheumatoid arthritis, OA, low back pain), approximately 26% also use other analgesic medications [6]. Specifically among OA patients, it is estimated that 70% take a prescription medication and 44 to 70% take an OTC analgesic to control joint symptoms [3]. While a high percentage of OA patients take prescription and/or OTC analgesics, it is unclear how many OA patients combine medications. In addition, the pattern of use within the same day and at the same time is unknown. As the incidence of OA increases, the use of multiple medications to control for symptoms is likely to increase as well.

In spite of the guidelines and recommendations that exist for the management of OA, inappropriate multiple analgesic medication use among OA patients may represent an additional burden on a health-care system that is struggling to adapt to an increasing prevalence of chronic diseases [7]. In the United Kingdom and United States, 25 and 54% of patients, respectively, are not aware of potential drug side effects and do not understand that analgesics can be taken prophylactically or at the onset of pain [6, 8]. Twenty-six and 8% of patients over medicate with OTC and prescription NSAIDs, respectively. In the United States it has not been determined which medications are being combined among OA patients and if these patients are taking the medications at the same time, potentially placing their health at risk (e.g., gastrointestinal toxicity, hepatic complications, renal complications) [6, 8]. The purpose of our study was to determine the professionally-guided and self-guided pharmacological and supplement use for joint symptom management among knee and hip OA patients in an urban hospital-based outpatient orthopedic practice. The outcome of this study is intended to improve clinical care and education of OA patients.

The type and source of pharmacological and supplement use information is reported in Table 2. The most common class of intervention was NSAIDs (particularly naproxen [29.6%] and ibuprofen [24.1%]) followed by other analgesics (i.e., prescription analgesics), acetaminophen, and nutritional supplements. Half of the participants reported using prescription medications only, while approximately a quarter of the participants used OTC medications and supplements only. Almost 20% of the participants used both prescription and OTC medications and supplements. Most patients (55.6%) reported only professionally-guided recommendations, 24.1% indicated only self-guided recommendations, and 17.3% reported professionally and self-guided recommendations for their medication and supplement use. Results of the logistic regression used to evaluate predictors (i.e., ethnicity, sex, obesity, age, reporting multiple OA joints, and education) of a participant only using self-guided recommendations compared to only using professionally-guided recommendations was not significant (p = 0.149).

Also reported in Table 2 is the pattern of medication use information. Less than half of the participants were always able to recall their medication dose, while the balance was either unable or partially able to do so. Twenty-nine percent of the participants reported never or only sometimes using their medication and/or supplement as directed, while the balance reported they always took their medication and/or supplement as directed. Results of the logistic regression used to evaluate predictors (i.e., ethnicity, sex, obesity, age, reporting multiple OA joints, and education) of a participant not always using their medication as directed was not significant (p = 0.139). Approximately half of the participants reported taking their medication and/or supplement consistently for 2 weeks whereas the others did not.

Results of the logistic regression to evaluate predictors of an individual being a same day dual-medication user were statistically significant. The Homer-Lemeshow goodness-of-fit test showed a good model fit with the data (p = 0.912). The overall accuracy of the model was 75.8% with a sensitivity of 44.4% and specificity of 77.7%. Participants reporting OA at multiple joints were more likely to be same day dual-medication users (odds ratio [95% confidence interval; CI] = 2.48 [1.03 to 5.96]) after controlling for covariates (i.e., ethnicity, sex, obesity, age, and education). In contrast, participants with less than a high school education were 3.82 times less likely to be same day dual-medication users (odds ratio [95% CI] = 0.26 [0.08 to 0.83]) than participants with a high school education or greater.

Although a majority of participants reported using professionally-guided recommendations, over 40% reported using at least one self-guided pharmacological and/or supplement intervention to manage their knee OA joint symptoms. Of the 15 participants who reported dual-use at the same time, 11 were professionally-guided, 5 were professional and self-guided, and 1 was solely self-guided based recommendation. Even with the combinations of professionally-guided, self-guided, and multiple sources of recommendation almost 17% of participants reported their intervention as ineffective or were prescribed a stronger medication. The only variable associated with participants reporting their intervention as ineffective (or requiring stronger medication) was not always taking their medication consistently for 2 weeks.

Our study revealed that nearly 20% of participants tried prescription and OTC medications to manage their joint symptoms. Among patients taking multiple interventions in the same day, 40.5% reported using prescription and OTC interventions. In Canada, 41% of patients recruited from urban primary care practices (n = 78 out of 190) reported using prescription medication and self-care products to manage their joint symptoms during one week [4]. It could not be determined if the Canadian patients were taking their medication combinations on the same day but this is likely since patients were asked to recall their medication use only for the week prior to the survey [4].

The overall combination of prescription and OTC interventions in our study may be lower than in the Canadian study [4] because they had fewer participants consuming only prescription medications (11% Canada versus 49% current study). The use of only OTC interventions was similar between studies (33% Canada versus 28% current study). The high prevalence of prescription medication use in the current study is similar to a cohort study conducted in Pittsburgh, Pennsylvania that had a high prevalence of prescription (54%) and OTC (63%) medication use during a 30-day period [3]. This study did not indicate whether participants were patients were taking a combination of prescriptions and OTC medications and supplements were included as a different class of interventions. Participants in the Pittsburgh study were requested to "bring in" their medications to be recorded on a medication inventory form. Based on these data, the authors reported that 70% of the participants were consuming prescription medications for OA symptom management. This is similar to our overall prescription medication use (69%) which was based on an interview-questionnaire format.

In the Pittsburgh study [3], among OTC medications, NSAIDs (58%: including COX-2 selective inhibitors) were the most commonly used drug class followed by other analgesics (55.7%). Our findings, although not limited to just OTC medications, were similar with NSAIDs being the most commonly used drug class utilized by 65% of participants followed by other analgesics (57%). These findings contradict current guidelines and recommendations that suggest acetaminophen should be the first line of defense over NSAIDs [2, 11]. However, it is not known if the participants in our study had previously failed with acetaminophen treatment and therefore progressed to NSAIDs. Future research should include assessments of whether NSAIDs are commonly recommended after failure with other analgesics.

Approximately 23% of participants reported dual-use in the same day, while 9% of participants reported dual-use at the same time. While these data include OTC and prescription medications, the findings are similar to those reported by Albert et al [3] in which 23% of the patients used prescription medications in 2 or more classes. The extent to which the dual medication use occurred within the same day is not reported. Participants were more likely to use multiple medications on the same day if they reported that more than one joint had OA or if they had a high school education or greater. Future research should explore if patients are taking more than one medication because they believe each medication is for a particular joint or if one medication is only helpful for certain joints and not others. Although dual-use is not necessarily bad or against current guidelines and recommendations, four of five organizations currently offer no statement on dual-use [2, 12–14]. The American College of Rheumatology is the only organization to offer a statement advocating dual-use for the treatment of OA [11]. While future research should evaluate medication dosage by using medication inventory forms, the data from our study indicates that the same-time dual-use combinations tend to be low-risk combinations (e.g., medications with supplements, NSAIDs with analgesics); however, recent research has raised concerns about the safety of combining NSAIDs with acetaminophen [15].

The current study identified several behaviors that have not been well studied in past OA research and may warrant further investigation. Thirteen (8%) participants reported using alternative methods to achieve symptomatic relief (e.g., misusing a prescription medication or consuming marijuana). Similarly, 26 participants reported stopping a medication because it was ineffective or they were prescribed a stronger medication. Based on these two findings it may be hypothesized that almost 1 in 5 patients in this cohort perceived one of their interventions to be ineffective. Participants that reported not taking medication consistently for 2 weeks were more likely to report their medication as ineffective. While longitudinal studies may be required to infer causality, it may be beneficial to suggest to patients that they should try new medications consistently (as prescribed) for at least 2 weeks. An additional 17 participants stopped taking their intervention because their prescription "ran out". We hypothesize that the reasons for this may have been due to socioeconomic status, access to healthcare, or effectiveness; but are yet to be determined as this was not the focus of the current study. These results highlight the need for clinicians to discuss with patients their satisfaction with interventions and what they should do if they believe a treatment is ineffective or their medication runs out.

The current study had several limitations. First was the sample size, which resulted in a ± 7.84% error rate. Another limitation of the study was that all participants were recruited from one inner-city orthopedic practice, narrowing the external validity of the study. The study also relied solely on self-reporting of medication use, which leaves the potential for misclassification of pharmacological and supplemental interventions. The Pharmacological and Supplement Use Assessment Questionnaire used in this study may have reduced the risk of misclassification because it provided a list of common NSAIDs and other interventions. Self-reported medication use was selected because of limitations within the current health system, a desire to capture over-the-counter medication use, and a preference to include any patient entering the clinic compared to a potentially biased sample of people that were willing to return with their medication bottles. We hypothesize that self-reporting was more likely to bias the results towards less medication use and safer patterns of medication use. Ostensibly, our data may represent a conservative reflection of how patients in an inner-city orthopaedic clinic manage their symptoms. Future larger, multi-center, epidemiology studies should be pursued. Although inclusion of medication inventory forms may help gather more information about current medication uses, the current instruments helped evaluate the behavior and influences on decisions concerning patient medication use. More research is needed to evaluate patient motives and alternative treatment strategies pursued through nontraditional mechanisms.

In conclusion, in a small cohort of inner-city orthopaedic patients trying to manage their joint symptoms a large percent of patients initiated self-guided care, used OTC medications, were unable to recall their medication dose, and reported not taking their medication as directed. Furthermore, participants who reported never taking their medication consistently for 2 weeks were more likely to report their medication as ineffective. Some patients stopped taking their medications because their prescription ran out, sought additional symptomatic relief by taking multiple medications at the same time, and reported inappropriate behaviors to achieve symptomatic relief. Almost one in four patients reported taking multiple medications or supplements during a day to manage their symptoms. Patients reporting OA at multiple joints and a higher education were more likely to consume multiple interventions during a day to manage their symptoms. Fortunately, combinations of multiple medications at the same time appear to be low risk (e.g., acetaminophen with NSAIDs). Based on this study, it may be important for clinicians to discuss how to manage multiple joint symptoms, the importance of taking a medication as prescribed, and what they should do if they believe a treatment is ineffective or their medication runs out.