Beliefs about medicines in gout patients: results from the NOR-Gout 2-year study

Objective Adherence to urate-lowering therapy (ULT) in gout is challenging. This longitudinal study aimed to determine 2 year changes in beliefs about medicines during intervention with ULT. Method Patients with a recent gout flare and increased serum urate received a nurse-led ULT intervention with tight control visits and a treatment target. Frequent visits at baseline and 1, 2, 3, 6, 9, 12, and 24 months included the Beliefs about Medicines Questionnaire (BMQ), and demographic and clinical variables. The BMQ subscales on necessity, concerns, overuse, harm, and the necessity–concerns differential were calculated as a measure of whether the patient perceived that necessity outweighed concerns. Results The mean serum urate reduced from 500 mmol/L at baseline to 324 mmol/L at year 2. At years 1 and 2, 85.5% and 78.6% of patients, respectively, were at treatment target. The 2 year mean ± sd BMQ scores increased for the necessity subscale from 17.0 ± 4.4 to 18.9 ± 3.6 (p < 0.001) and decreased for the concerns subscale from 13.4 ± 4.9 to 12.5 ± 2.7 (p = 0.001). The necessity–concerns differential increased from 3.52 to 6.58 (p < 0.001), with a positive change independent of patients achieving treatment targets at 1 or 2 years. BMQ scores were not significantly related to treatment outcomes 1 or 2 years later, and achieving treatment targets did not lead to higher BMQ scores. Conclusion Patient beliefs about medicines improved gradually over 2 years, with increased beliefs in the necessity of medication and reduced concerns, but this improvement was unrelated to better outcomes. Trial Registration ACTRN12618001372279

Gout is a chronic inflammatory disease where patients with high levels of urate may experience painful disease flares of arthritis (1).One of the fundamental goals of gout management is urate-lowering therapy (ULT), to prevent the formation of crystal deposits and thereby also painful gout flares.
Adherence to medical treatment continues to be a major clinical problem worldwide (2) in patients with gout (3)(4)(5), as in other rheumatic conditions treated with drugs and modification of lifestyle (6,7).
Barriers to achieving optimal gout treatment may be due to patient factors or physician factors (8).In a Swedish study, the persistence on ULT with allopurinol was only 21-25%, without a statistically significant increase over time (9).When adherence to ULT is high, prevention of gout flares and pain is observed (10).Older age and the presence of other comorbidities such as hypertension and diabetes are factors related to good adherence (4).The severity of gout, such as flares, seen from the patient perspective is complex, with a number of interconnecting factors which should be captured in long-term gout studies (11).
Reasons for lack of adherence can include factors such as forgetting, but the importance of intentional factors has also been studied, along with patients' beliefs and levels of motivation and activation (12).The Beliefs about Medicines Questionnaire (BMQ) (13) is a measure that addresses these factors.
Some factors are able to predict a good treatment outcome, with achievement of the target of low serum urate after the first year of ULT: high self-efficacy, low alcohol consumption, and patients having low levels of belief in the overuse of medicines (14).
Beliefs about medications using the full BMQ have not previously been reported for gout.The objective of this study was to describe possible changes in beliefs about medicines in gout patients over a 2 year period during a treat-to-target intervention with ULT supervised by nurses in a rheumatology clinic, and to investigate whether certain beliefs were related to achieving a good treatment outcome.

Study design and participants
NOR-Gout is a prospective, observational single-centre study based in a hospital rheumatology outpatient clinic.Patients had been referred with a suspected gout flare by general practitioners for treatment, for diagnostics, and/ or for consideration of intra-articular steroid injections.They were eligible for inclusion in the study if they had had a gout flare within the past month, serum urate levels greater than the treatment target (>360 μmol/L, 6 mg/dL), and no contraindication for ULT.They were consecutively included according to the study protocol (ACTRN12618001372279).All participants had been diagnosed with gout based on the identification of monosodium urate (MSU) crystals on polarized microscopy after arthrocentesis, and satisfied the American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) classification criteria (15).Important exclusion criteria were unstable medical conditions and known stage 3b or higher chronic kidney disease [estimated glomerular filtration rate (eGFR)/ creatinine clearance < 45 mL/min] (14).Participants were identified during an acute clinical gout flare after examination in the rheumatology outpatient clinic.Those indicating willingness to participate in the study were contacted by a study nurse from the rheumatology outpatient clinic for prescreening (n = 242), received written information, and were scheduled for a baseline rheumatology outpatient visit at Diakonhjemmet Hospital.The study was approved by the Regional Ethics Committee for Medical and Health Research Ethics Eastern Norway, included patient representatives in project planning, and was performed in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines.All patients provided written informed consent.The sponsor of the study was Diakonhjemmet Hospital.
At baseline, patients had an individual consultation with a trained rheumatology research nurse who provided information about the disease and disease process in gout, causes, recommended treatments, and disease control.Furthermore, lifestyle advice on exercise and diet, weight reduction, and alcohol consumption was given.Further advice on diet was provided in a brochure.
Patient expectations were discussed, as well as when to seek additional help and advice on coping.
All patients not already on ULT started as recommended (16)(17)(18) with oral allopurinol 100 mg once per day, which was escalated by 100 mg monthly according to serum urate concentrations until a maximum of 900 mg daily.If a patient was intolerant to allopurinol, febuxostat was started at 40 mg once daily and escalated monthly to 80 or 120 mg as needed.Probenecid or lesinurad could be added, if necessary, but were not used in any patients.Patients received flare prophylaxis, with prescribed colchicine 0.5-1 mg daily, individualized for 3-6 months, as recommended (14).In this treat-to-target approach, ULT was escalated to reach the serum urate treatment target level of < 360 μmol/L (or < 300 μmol/L to achieve fast resolution in patients with tophi) and the dose was maintained when the target was reached.
At every visit, drug use was recorded for nonsteroidal anti-inflammatory drugs (NSAIDs), colchicine, prednisolone, and ULT (allopurinol, febuxostat), registering drug dosage, adverse events, and symptoms of flares.
A study nurse assessed patients at baseline and at 3, 6, 12, and 24 months with a rheumatologist who also performed ultrasound examinations for scoring of the degree of MSU crystal deposits (HBH and LK).Additional scheduled visits with only the study nurse were at 1, 2, and 9 months, but also if necessary monthly, until the treatment target was reached.Telephone contact with review of the serum urate result could substitute for face-to-face visits.During the second year, patients were followed by their general practitioners as needed.

Data collection
The main study outcome variable was change in the BMQ in patients over 2 years.The BMQ was developed by Horne et al (13,19) and consists of 18 items.Each item is answered on a five-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = uncertain, 4 = agree, 5 = strongly agree).The two main categories include general and specific beliefs.The general belief items are grouped into the subscales 'Harm' and 'Overuse'.These subscale scores range from 4 to 20, where a higher score reflects that the patient believes medications to be more harmful or overused, respectively.The specific part of the questionnaire is used to assess the patient's beliefs about the specific medications prescribed for their condition.The specific belief items are divided into the categories 'Necessity' and 'Concerns'.In both categories, the scores range from 5 to 25, and a higher score reflects a higher belief in the necessity or greater concern.
A necessity-concerns differential, subtracting concerns from necessity, was calculated to identify whether patients believed that the necessity of their medication outweighed the concerns about taking it (19).This differential score has been shown to correlate better with adherence than necessity and concerns separately (20), and ranges from −20 to 20.A positive number indicates a belief that the need to take medication is stronger than the concerns about taking the medication.The framework of the BMQ was derived from the Health Belief Model, which describes how beliefs about medicines could affect medication-taking behaviour (19).The BMQ has been found to be valid and reliable in Scandinavian languages, including Norwegian (21,22).
At baseline, patients answered a number of questions on demographics and lifestyle variables, as well as completing health status questionnaires.They reported age, gender, ethnicity, marital status, family history for gout, disease duration, highest level of education, comorbidities, and working status.For comorbidities, the Self-Administered Comorbidity Questionnaire (SCQ) was used (range 0-36) (23); this includes 12 medical problems, allocating 1 point per problem, including presence, receiving treatment, and causing a functional limitation.
Alcohol consumption was assessed with the categories 'Daily', 'Weekly', 'Monthly', and 'Never', then aggregating the categories to daily/weekly and monthly/ never.Daily and previous smoking, consumption of daily glasses of sugar-sweetened drinks, and the frequency of physical activity were reported by patients.
Information on number of flares 'ever' and 'during the last year' (before the recent study entry flare) was collected at baseline, as well as pain severity during the most recent and the strongest flare (0-10 numerical rating scales), with 0 = no pain and 10 = unbearable pain.The occurrence of flares prior to each visit was also recorded.
At all visits, patient-reported outcomes were recorded using questionnaires endorsed by Outcome Measures in Rheumatology (OMERACT) (24), including joint pain, general pain, and patient global assessment of disease activity, on 0-10 numerical rating scales.
Physical function was measured with the Health Assessment Questionnaire Disability Index (HAQ-DI) without adjustment for help or devices (25).Health status was assessed by the 36-item Short Form general health questionnaire (SF-36) (26), reporting the physical and mental component summaries.
Self-efficacy, with subscales for pain (five items) and symptoms (six items), was measured with the Arthritis Self-Efficacy Scale (27).This instrument measures whether patients have confidence in coping with pain, function, and other symptoms due to arthritis (10-100 numerical rating scales, 100 = highest self-efficacy).
Clinical assessments included weight and height, for the calculation of body mass index (BMI), and swollen and tender joint counts including 44 joints.Subcutaneous tophi were counted.

Statistical analysis
Descriptive measures of baseline variables are presented using absolute and relative frequencies, as well as means with standard deviation or median with interquartile range, as appropriate.Differences between groups were explored using the independent samples t-test, chi-squared test, or Fisher's exact test, and Pearson correlation coefficients; differences within groups were assessed using the paired t-test.Longitudinal repetitive measurements of BMQ were analysed with linear mixed effect models.These analyses combine multiple measurements per patient, use all available follow-up data, take into account missing values, and correct for within-patient correlations.The influence of other variables was studied, and time interaction terms were tested in the model.
Answers to the individual BMQ items were categorized as endorsing a belief after collapsing the answer categories 'Agree' and 'Strongly agree'.
A value of p < 0.05 was defined as statistically significant, and all analyses were performed with IBM SPSS statistics (version 27).

Results
Of 211 patients included in the study, 186 (88.2%) completed the visit at 1 year and 173 (82.0%) at 2 years.After 1 year of intensive ULT, mean serum urate had decreased from 500 µmol/L to 311 µmol/L.At 1 and 2 years, 85.5% (159/186) and 78.6% (136/173) of patients, respectively, had achieved the main serum urate treatment target of < 360 µmol/L.Baseline demographics and disease-related variables are shown in Table 1.Patients not completing the 2 year follow-up were not statistically different from completers in terms of demographics and disease-related variables, including baseline BMQ scores.
The degree to which patients agreed with the 18 belief statements in the BMQ at baseline is shown in Table 2, ordering the items with the highest support grouped across the four subscales.In general, there was less agreement with statements on necessity than on concerns among the statements.
Results from the BMQ over 2 years are shown in Table 3.There was a gradual, numerical, and statistically significant increase in the beliefs for necessity and decline for concerns over 2 years, while results for the other general subscales, overuse and harm, remained stable.
The necessity-concerns differential as a measure for positive perception in beliefs is given in Figure 1 with boxplots for all time-points.There was a statistically significantly (p < 0.001) increase in marginal means, from 3.52 to 6.58, over 2 years, indicating that the positive perception by patients of their gout medication increased.Results from the BMQ linear mixed model analyses after the inclusion of other candidate variables were similar to the main analysis with time as a random factor, and no new factors were identified predicting the BMQ differential (data not shown).
We then studied whether previous BMQ results were related to subsequent 1 and 2 year treatment success, i.e. whether patients achieved the treatment target or not.The BMQ necessity-concerns score at baseline or at 1 year was not related to reaching the treatment target at 1 or 2 years (Table 4), but increased in both groups over year 1 (p < 0.001).Finally, we investigated a possible inverse relationship, i.e. whether achieving the treatment target at 1 year was predictive of the BMQ necessity-concerns score at year 2, but found no association (Table 5).Furthermore, serum urate values at baseline were not related to BMQ necessity-concerns at year 1.In sensitivity analyses, the 35 tophaceous patients followed the BMQ pattern for non-tophaceous patients without major differences.

Discussion
In this 2 year observational cohort study, beliefs among gout patients in the necessity of medication increased, and concerns decreased, whereas general beliefs about the overuse and harm of medication remained largely unchanged.The approach with ULT led by a health professional was feasible and patient acceptance of medications with dose escalations increased, independently of whether the treatment target was achieved after 1 or 2 years.
Important factors that could contribute to medication beliefs and achieving treatment target in this study were that the nurse gave initial individual information and education on aspects of gout, including medication, as well as having frequent controls and a clear treat-to-target schedule with escalation of medication until patients had target serum urate levels < 360 µmol/L.We demonstrated  previously that adjustment of medication for optimal ULT, education, and personal interaction with the nurse and rheumatologist demonstrated effective ULT in 85% of patients (14) and reduction of flares over 2 years (28).Thus, one explanation for the lack of an association of the belief in necessity and concerns about medicines with the outcome may be that optimized ULT was so effective in most patients that no additional effect of enhanced beliefs in medicines could be observed.
In a study on short-term interventional education study in rheumatoid arthritis, et al (29) found no association between adherence to disease-modifying antirheumatic drugs (DMARDs) and the BMQ.The BMQ necessity subscale was unchanged, BMQ concerns and overuse changed slightly, and no difference was seen for adherence.In another cross-sectional study, adherence to oral DMARDs was associated with the BMQ necessity scale (30).In our study, we did not measure adherence but rather specific beliefs about medicines, which could indirectly indicate adherence.
The sensitivity to change for the BMQ has not been sufficiently researched, and apart from a study on shortterm intervention (30), there is only one other observational study which examined scores for the BMQ harm and BMQ overuse scales with assessments at baseline and after almost 4 years, with stable results and without a specific intervention (31).Having found no relationship between the BMQ and disease outcomes in NOR-Gout, we cannot tell whether the sensitivity to change in the BMQ is insufficient or whether the BMQ in our study setting is less important than intensive ULT.
In this study, the BMQ was evaluated for the first time in gout patients.The necessity-concerns differential of the BMQ increased steadily over the 2 years of this study, indicating the increasingly positive attitude of patients towards their medicines.There were differences in the degree to which patients supported concepts expressed in the individual items of the BMQ.Patients agreed most with two items each (concerns and harm of medication each).Agreement with the necessity items was generally lower, and at baseline, as few as 3% of patients agreed with the statement 'My medicines protect me from becoming worse'.The low support for this item illustrates that the concept of using ULT to prevent flares and urate depositions needs to be communicated clearly and consistently to gout patients to achieve improved adherence to ULT.Relationships between the BMQ and adherence were described by Horne et al (19), and high levels on the BMQ scales could represent high adherence levels.A systematic review found that adherence to medication is weakly correlated with necessity beliefs and concern beliefs in the BMQ (20).Nevertheless, in our study, the changes in BMQ scores reflect changes in beliefs about medicines rather than adherence, and should be taken as an indication of increased knowledge about gout in our patients.An instrument measuring adherence, such as the Medication Adherence Report Scale (32), may be better able to predict disease outcomes in gout.Low adherence to medication and lifestyle advice is a major challenge in gout (4) and may result in only 50-70% of patients achieving the treatment target after 3-5 years, even when followed up with regular visits (33).Thus, more attention needs to be paid to supporting adherence in gout and treatment uptake, including follow-up by other health professionals, such as nurses or pharmacists (34).Involving patients in treatment decisions, with education as part of treat-to-target ULT (14,(35)(36)(37), will result in the best possible adherence to therapy (37).Good treatment results may also enhance prescriber adherence to guidelines (38) and overcome clinical inertia (39).When considering alternative interventions for improved adherence, methods with attractive visual appeal, such as images and illustration (40) or gout storytelling (41), could be used in gout patients as part of individualized education.Furthermore, factors contributing to patient-perceived disease severity in gout have been examined in a qualitative study (11).
Strengths of this study include the repetitive use of the BMQ over five time-points in a longitudinal study over 2 years.The sample size was large and the population was recruited from clinical practice.For the first time, the full BMQ was applied to people with gout.
Limitations also need to be considered.First, this is a single-centre study and the results need to be replicated before generalization.Secondly, the beliefs were studied during a treat-to-target approach, where patients may be especially stimulated to change beliefs.However, improvements in BMQ outcomes and the necessity-concerns differential were not related to the treatment target at 1 and 2 years.Thirdly, we did not use a specific measure for actual drug adherence to compare with the BMQ results, although this is planned for a future follow-up study.Finally, this is an observational study, precluding any causal inferences.

Conclusion
This study shows that over 2 years with nurse-led ULT in a rheumatology clinic, patient beliefs in the necessity of medicines increased and concerns decreased.The positive perception of specific medication and the necessity-concerns differential gradually increased over 2 years.However, our findings did not show that improved beliefs about medicines also improved disease outcomes.Disease outcomes are good in a study setting such as ours, with frequent control visits and intensive ULT (14,42).Our approach may contribute towards increased knowledge about medication, increase patients' understanding of the necessity of taking ULT, and reduce concerns about medication in patients with gout.

Table 2 .
Results of the Beliefs about Medicines Questionnaire at baseline (N = 211).

Table 3 .
Description of subscale scores in Beliefs about Medicines Questionnaire over 2 years.

Table 5 .
One and 2-year serum urate treatment outcome status and Beliefs about Medicines Questionnaire (BMQ) differential necessity-concerns outcome after 2 years.

Table 4 .
Beliefs about Medicines Questionnaire (BMQ) differential necessity-concerns scores and serum urate treatment outcome after 1 and 2 years.