Measuring therapeutic adherence in systemic lupus erythematosus with electronic monitoring

Abstract

Objective: We used an electronic monitoring system to quantify adherence to prescribed oral therapies by patients with systemic lupus erythematosus (SLE). Methods: Participants were included from a larger longitudinal study cohort of 110 patients recruited from publicly-funded rheumatology clinics, 78 of whom agreed to have their SLE drug therapy electronically monitored for two years with the Medication Events Monitoring System (MEMS®, AARDEX Group). Adherence was determined as the percentage of days (weeks for methotrexate) the patient took the medication as prescribed by the physician. Collected data included SLEDAI; SLICC damage index for SLE (SDI); medical outcome study social support survey (MOS-SSS); Center for Epidemiologic Studies depression scale (CESD); and quality of life (SF-12). Results: Ninety percent of the cohort was female, 45% were Hispanic, and 49% were African-American. Mean age was 36.3 years, disease duration was 5.9 years, SLEDAI score was 3.2, and SDI score was 0.9. Adherence was 62% for all drugs combined and did not differ significantly for individual medications. Patients with more depression (p < 0.02), and higher number of pills taken daily (p < 0.02) were more likely to be non-adherent. Only one-fourth of the patients had an average adherence of ≥80%; these patients had a better mental component score (SF-12) at 24 months than non-adherent patients (p < 0.01). Conclusions: Electronic monitoring demonstrated that only one-fourth of the patients had an adherence rate ≥80%. Polypharmacy and depression were associated with non-adherence.

Keywords

Treatment adherence electronic monitoring systemic lupus erythematosus outcomes

Introduction

Adherence to a medication regimen is ‘the extent to which a patient acts in accordance with the prescribed interval, dose and dosing regimen'.¹ Adherence is frequently reported as the percentage of prescribed doses taken by the patient over a specified period of time. An acceptable level of adherence is often set at ≥80% of the recommended treatment regimen dosages.^1,2 Poor adherence to therapeutic regimens is one of the most common problems in daily practice care especially among patients with chronic conditions and can carry dire economic consequences.

Lack of adherence to medical recommendations has been proposed as an important predictor of poor outcomes in patients with SLE.³^–⁵ The adherence rate has varied from 48% to 93% according to the method used.⁶ Several studies have assessed different aspects of therapeutic adherence in SLE, mostly with self-report questionnaires and pharmacy refill data but none have used electronic monitoring of therapy.⁶^–¹⁵ With the development of nanotechnology more accurate methods of adherence measurement have become available. The Medication Events Monitoring System (MEMS; AARDEX Group, Sion, Switzerland) consists of a microchip placed in the cap of the medicine bottle which records the date and time of every opening. Special software captures information and calculates multiple adherence measures.^16,17

The objective of this study was to quantify adherence to medical therapy among SLE patients using MEMS, to identify the predictors of poor adherence, and its association with clinical outcomes.

Methods

This was a longitudinal cohort study of 110 SLE patients who were recruited from publicly funded outpatient rheumatology clinics in Houston, Texas, USA: two were within the Harris County Hospital District facilities and the other one at the Michael DeBakey Veteran Affairs Medical Center (VAMC). Approval for the study was obtained from our Institutional Review Boards. After permission was given by the treating rheumatologist, patients were approached during a clinic visit. A research assistant was on site to identify potentially eligible patients. All recruited participants signed informed consent forms.

Patient inclusion criteria were: (i) age 18–80 years; (ii) fulfillment of the ACR SLE diagnostic criteria;¹⁸ (iii) disease duration one to ≤15 years; (iv) treatment with steroids, disease-modifying drugs or immunosuppressants at the time of enrollment; (v) adequate cognitive status as determined by the research assistant; and (vi) English or Spanish Language proficiency as determined by the research assistant.

The parent study examined beliefs, attitudes, and self-reported adherence. Participating in the electronic monitoring component of the study was optional. Only those patients who agreed to electronic monitoring of medication are included in this report.

Data collection

Baseline demographic data included self-reported age, sex, ethnicity, education level, employment status, health insurance status, annual household income, number of household members, and marital status. Patient self-reported outcomes and health determinant instruments were assessed at baseline and at 3, 6, 12, 18, and 24 months. These included:

a visual analogue scale (VAS) to indicate patients’ perceived disease activity (0–10 scale, in which 10 represents the highest disease activity)

the Center for Epidemiologic Studies depression scale, 10-item short-form (CESD-10; 0–30 scale, in which 30 represents the greatest depression)¹⁹

the medical outcomes study social support survey (MOS; 0–5 scale, in which 0 represents the least support)²⁰

the 12-item short-form health survey (SF-12; 0–100 scale, in which 0 represents the worst quality of life; mental (MCS) and physical component scores (PCS) were calculated using norm-based methods).²¹

For patients who spoke only Spanish proficiently, all questionnaires, scales, and surveys were translated into Spanish using standard methods: forward/backward translation by independent translators followed by bilingual group evaluation and consensus on item translation. At each patient visit, the physician filled out the patient’s disease activity using:

VAS for disease activity

the Mexican modification of the SLEDAI; 0–32 scale in which 32 represents highest disease activity²²

the SLICC/ACR Damage Index for SLE (SDI; 0–47 scale in which 47 represents greatest chronic damage).²³

Comorbidities and medications were also reviewed and documented at each visit.

Laboratory testing

Laboratory testing was performed at baseline, 12, and 24 months including: erythrocyte sedimentation rate (ESR); C-reactive protein (CRP); complete blood count, creatinine, C3, C4, urinalysis, and anti-double stranded DNA antibodies.

Assessment of adherence

All participants in our study completed the 19-item compliance-questionnaire-rheumatology (CQR)²⁴ self-report to record their adherence to prescribed medications at baseline, 12 months and 24 months.

Patients who agreed to undergo electronic monitoring had their medication intake for prescribed oral SLE medications tracked using MEMS®. Adherence to nonsteroidal anti-inflammatory drugs or therapy for comorbidities was not measured. MEMS® are designed to compile dosing histories of oral medications. In this system, the cap of a standard plastic vial has microcircuitry that records the time and date the bottle is opened. The system assumes that every opening of the bottle signifies the prescribed dose is taken. Specific software (Quick Read® and PowerView®, Aardex) and a MEMS® reader were used to capture and analyze data. Three measures of adherence to oral SLE prescriptions were calculated. ‘Correct dosing’ was calculated as the percentage of days that the patient took the correct number of doses as prescribed: (Total number of days with a recorded event as prescribed)/(Total number of monitored days) x 100.²⁵ ‘Underdosing' was calculated similarly but with fewer doses than prescribed in the numerator. Likewise, ‘overdosing' was calculated with more doses than prescribed in the numerator. For methotrexate prescriptions, the adherence measures were calculated using weeks rather than days. For each patient the average of the percent correct dosing, underdosing, or overdosing was calculated for the entire two-year monitoring period. Adherence estimates were obtained for each individual oral SLE therapy and for all SLE drugs together. Occasional openings of the medication container that were unrelated to drug taking (e.g. opening for refills) were excluded from the analysis. These openings were assessed by patient interview and direct comparison with pharmacy refills. During the initial visit, patients were informed about the characteristics of the MEMS® technology, so they were aware of the monitoring capability of the caps provided.

Statistical analysis

Categorical variables were analyzed using chi-squared or Fisher’s exact tests. Continuous variables were analyzed using Student’s t-test, Pearson’s correlation coefficients, and one-way analysis of variance with Tukey’s honestly significant difference. For the analysis, education was dichotomized into high school diploma or less, and higher educational level. Multiple regression analysis was performed to identify predictors of adherence using a stepwise procedure with entry model criteria of p value ≤ 0.15 and removal criteria of p value > 0.05. A two-sided p value of 0.05 was considered statistically significant. Analyses were carried out using SAS software version 9.2 (SAS Institute, Inc., Cary, North Carolina, USA).

Results

We screened over 1700 patients; 110 of these had SLE, fulfilled the inclusion criteria and agreed to participate (refusal rate 32%), and 78 agreed to electronic adherence monitoring of their oral therapies using MEMS® caps. Ninety-four percent of the patients were seen at a Harris County District Hospital clinic. Baseline characteristics of participants are shown in Table 1. There was no statically significant difference between patients who accepted electronic monitoring and those who did not with regard to baseline disease activity, chronic organ damage, quality of life measures, depression, comorbidities, number of pills per day, and self-report adherence. Twelve (15.4%) patients using MEMS® discontinued the study vs 12 (37.5%) patients who did not agree to use MEMS® (p = 0.01). We did not find any difference between completers and non-completers regarding age, sex, ethnicity, disease duration, comorbidities, baseline disease activity and self-reported adherence (CQR). The analyses, including two-year outcomes, were performed on all patients with complete follow-up.

Table 1

Baseline characteristics of patients monitored with Medication Event Monitoring System (MEMS®)

	MEMS®
	(n = 78)
Age, years, mean (SD)	36.3 (11.4)
Female sex, n (%)	70 (90.0)
Disease duration, years, mean (SD)	5.9 (5.3)
Ethnicity, n (%)
Hispanic	35 (44.9)
African-American	38 (48.7)
White	3 (3.8)
Other	2 (2.6)
SLEDAI score, mean (SD)	3.2 (3.3)
SDI score, mean (SD)	0.9 (1.2)
CESD-10 score, mean (SD)	12.6 (6.5)
MOS score, mean (SD)	3.9 (0.9)
SF-12 score, mean (SD)
PCS	35.9 (10.5)
MCS	42.3 (10.2)
CQR score, mean (SD)	70.7 (12.6)
Number of comorbidities, mean (SD)	1.4 (1.2)
Number of pills/day for non-SLE conditions, mean (SD)	9.8 (6.1)
Lost to follow-up, n (%)	12 (15.4)

CESD-10: Center for Epidemiologic Studies 10-item depression scale; CQR: compliance-questionnaire-rheumatology; MCS: mental component of SF-12; MOS: medical outcome study; PCS: physical component of SF-12; SD: standard deviation; SDI: SLICC/ACR damage index for SLE; SF-12: short-form health survey; SLEDAI: Mexican modification of the SLEDAI.

The most common SLE medication taken by patients at baseline was prednisone (97%). Other SLE drugs included hydroxychloroquine (87%), mychophenolate mofetil (26%), methotrexate (15%), azathioprine (9%), cyclophosphamide (9%), and sulfasalazine (1%). Overall, a total of 133 oral SLE drug prescriptions taken by 78 patients were electronically monitored for adherence (Table 2). The medications were prescribed for a mean of 1274 days, of which two-thirds were monitored. Across all drugs the average rate of correct dosing was 62.5%, of overdosing was 3.8%, and of underdosing was 33.6%. No statistically significant differences were observed among the individual medications.

Table 2

Adherence measures using Medication Event Monitoring System (MEMS®) dosing of monitored prescriptions

Measure, mean (SD)	All drugs	HCQ	PDN	MTX	MMF	AZA
Measure, mean (SD)	(n = 133)	(n = 52)	(n = 68)	(n = 7)	(n = 4)	(n = 2)
Number of days drug was prescribed for	1274.1	627.7	611.6	426.1	428.2	562.5
Number of days drug was prescribed for	(540.6)	(198.3)	(211.9)	(299.0)	(302.8)	(102.5)
Number of days of MEMS® monitoring	825.0	519.3	484.0	329.9	368.0	323.0
Number of days of MEMS® monitoring	(528.5)	(230.0)	(261.8)	(271.9)	(235.2)	(410.1)
Percentage with correct dosing	62.5	61.2	66.1	73.6	49.8	60.6
Percentage with correct dosing	(24.5)	(26.0)	(24.9)	(19.8)	(16.8)	(25.7)
Percentage overdosing	3.8	3.9	4.0	5.7	3.9	1.6
Percentage overdosing	(3.8)	(5.6)	(4.6)	(4.9)	(6.3)	(2.3)
Percentage underdosing	33.6	34.9	29.9	19.1	46.2	37.8
Percentage underdosing	(25.5)	(27.2)	(25.3)	(24.2)	(21.0)	(28.0)

AZA: azathioprine; HCQ: hydroxychloroquine; MMF: mycophenolate mofetil; MTX: methotrexate; PDN: prednisone; SD: standard deviation.

For the analyses examining predictors of adherence, we considered all drugs together and prednisone or hydroxychloroquine independently. Lower numbers of patients received mychophenolate mofetil, methotrexate and azathioprine, and therefore these drugs were not evaluated individually but were added to the overall adherence analyses.

Bivariate analysis revealed several significant differences in correct dosing for all drugs by baseline patient characteristics (Table 3). Lower adherence was associated with a higher number of daily number of pills for non-SLE conditions (r = −0.29; p = 0.01), worse patient-perceived disease activity (r = −0.25; p = 0.03), and worse depression (r = −0.28; p = 0.01). No statistically significant differences in correct dosing were observed between African-American and Hispanic groups. When correct dosing for all drugs was set as the dependent variable, the final multiple regression model demonstrated that worse depression (β = −1.04; p = 0.01), and higher number of pills taken daily (β = −1.29; p = 0.001) continued to be statistically significant predictors of correct dosing. Similar patterns were observed when prednisone or hydroxychloroquine were analyzed on their own.

Table 3

Bivariate analysis of baseline sociodemographic variables and clinical outcomes with percentage of correct dosing taken, for all monitored oral systemic lupus erythematosus (SLE) prescriptions, hydroxychloroquine (HCQ), and prednisone (PDN)

Categorical variables	All drugs (n = 78)			HCQ (n = 52)			PDN (n = 68)
Categorical variables	n (%)	Correct dosing¹	p *	n (%)	Correct dosing¹	p *	n (%)	Correct dosing¹	p *
Ethnicity
Hispanic	35(44.9)	64.4(24.0)	≥0.20	25(48.1)	63.2(23.8)	≥0.20	31(45.6)	67.4(24.1)	≥0.20
African-American	38(48.7)	60.1(26.0)		22(42.3)	58.9(29.3)		32(47.1)	63.6(27.1)
White	3(3.8)	72.2(7.3)		3(5.8)	69.3(9.4)		3(4.4)	75.1(7.5)
Other	2(2.6)	60.3(33.4)		2(3.8)	47.4(43.5)		2(2.9)	73.2(23.4)
Education level
≤High School	48(61.5)	67.1(19.2)	0.07	33(63.4)	78.8(21.6)	0.17	41(60.3)	80.4(20.1)	0.07
>High School	30(38.5)	54.4(30.4)	0.07	19(36.5)	68.9(30.2)	0.17	27(39.7)	68.2(30.8)	0.07
Language
English	47(60.3)	60.7(24.8)	≥0.20	29(55.8)	58.6(27.6)	≥0.20	39(57.4)	65.3(25.8)	≥0.20
Spanish	20(25.6)	66.0(22.6)		15(28.8)	59.3(26.3)		19(27.9)	69.7(20.1)
Both	11(14.1)	63.9(28.4)		8(15.4)	73.8(16.6)		10(14.7	62.6(30.6)
Continuous variables		r^†	p		r^†	p		r^†	p
Age		−0.02	≥0.20		−0.11	≥0.20		−0.03	≥0.20
Disease duration		−0.09	≥0.20		−0.02	≥0.20		−0.13	≥0.20
SLEDAI score baseline		0.16	0.16		0.10	≥0.20		0.22	0.07
SDI score baseline		−0.03	≥0.20		0.02	≥0.20		−0.04	≥0.20
Depression (CESD-10)		−0.28	0.01		−0.21	0.13		0.35	<0.01
MOS scale		0.10	≥0.20		0.12	≥0.20		0.10	≥0.20
SF-12 PCS		0.02	≥0.20		−0.12	≥0.20		0.11	≥0.20
SF-12 MCS		0.22	0.06		0.001	≥0.20		0.30	0.01
VAS activity physician (10=worst)		−0.13	≥0.20		−0.11^#	≥0.20		−0.15^#	≥0.20
VAS activity patient (10=worst)		−0.25^#	0.03		−0.25	0.1		−0.29^#	0.02
Number of comorbidities		−0.19	0.10		−0.28	0.05		−0.20	0.10
Number of non-SLE pills/day		−0.29^#	0.01		−0.33	0.02		−0.24	0.05

MCS: mental component of SF-12; MOS: medical outcome study; NA: not applicable; PCS: physical component of SF-12; SD: standard deviation; SDI: SLICC/ACR damage index for SLE; SF-12: short-form health survey; SLEDAI: Mexican modification of the SLEDAI; VAS: visual analogue scale.

Percentage of correct dosing as mean (SD).

p value determined using Student’s t-test or analysis of variance. Bold font denotes statistically significant (p≤0.05).

†

r values were determined using Pearson correlation coefficients.

Two patients were missing.

Relationship between adherence and outcomes at two years

The only variable that statistically significantly changed between baseline and two years was the SLEDAI score, showing a decrease in SLE disease activity. However this change was not significantly associated with adherence. At two years the only variable significantly positively associated with adherence was a higher score for mental component of quality of life (p = 0.02) (Table 4). When considering the conventional cut-off of ≥80% for an acceptable level of adherence, we observed that only 16 patients (24%) had achieved that goal. At two years these 16 patients had a statistically significantly higher quality of life mental component score than did the 59 patients who had a compliance rate of <80% (50.4 vs 41.2; p < 0.01) (Figure 1). No significant differences were observed for SLEDAI score (1.4 vs 1.6; p = 0.15), SDI (1.0 vs 1.6; p ≥ 0.20), SF-12 physical component score (37.2 vs 37.7; p = 0.90), or proportion of visits with flares (59% vs 53%; p ≥ 0.20).

Figure 1

Mental component of quality of life between patients with all drugs >80% treatment adherence, as determined using Medication Event Monitoring System (MEMS®). SF-12: short-form health survey.

Table 4

Correlation between outcomes at 24 months and percentage of correct doses taken for all monitored oral systemic lupus erythematosus (SLE) prescriptions, hydroxychloroquine, and prednisone

	Percentage days with correct dosing
		All drugs (n=66)^#		HCQ (n=44)		PDN (n=59)
Outcomes	Mean (SD)	r ^†	p	r ^†	p	r ^†	p
SLEDAI score
24 months	1.4 (2.5)¹	−0.001	≥0.20	0.17	≥0.20	−0.03	≥0.20
24 months - baseline	−1.7 (4.0)¹	−0.10	≥0.20	0.06	≥0.20	−0.19	0.19
SDI score
24 months	1.2 (1.2)	0.08	≥0.20	0.12	≥0.20	0.09	≥0.20
24 months - baseline	0.3 (0.7)	0.15	≥0.20	0.07	≥0.20	0.21	0.14
SF-12 PCS
24 months	37.3 (11.3)	0.06	≥0.20	−0.06	≥0.20	0.15	≥0.20
24 months - baseline	1.8 (10.3)	0.08	≥0.20	0.06	≥0.20	0.08	≥0.20
SF-12 MCS
24 months	43.4 (12.4)	0.30	0.02*	0.20	0.19	0.42	0.001*
24 months - baseline	0.7 (11.2)	0.18	0.14	0.19	≥0.20	0.26	0.05*
Proportion of mild flares/patient
24 months	53.4 (25.9)	−0.13	≥0.20	−0.13	≥0.20	−0.11	≥0.20
Proportion of severe flares/patient
24 months	11.2 (14.9)	−0.06	≥0.20	−0.03	≥0.20	−0.01	≥0.20

HCQ: hydroxychloroquine; MCS: mental component of SF-12; MOS: medical outcome study; PCS: physical component of SF-12; PDN: prednisone; SD: standard deviation; SLEDAI: Mexican modification of the SLEDAI; SDI: SLICC/ACR damage index for SLE; SF-12: short-form health survey.

12 missing patients at the end of the study.

†

r values were determined using Pearson correlation coefficients.

SLEDAI was the only variable that decreased significantly across the study (3.2 vs 1.4, p = 0.0007), as determined using paired Student’s t-test.

denotes statistically significant (p ≤ 0.05).

Discussion

The primary objective of this study was to quantitatively measure adherence to oral therapies in patients with SLE using electronic monitoring over a two-year period. To our knowledge, this is the first treatment adherence study of SLE patients using MEMS®. Of the available methods for measuring adherence, MEMS® is the closest to gold standard since it allows the day and precise time of every bottle opening to be obtained.²⁶ Our sample was comprised of mostly ethnic minority patients receiving care at publicly funded clinics. We found that only 20% of patients had an adherence rate of ≥80%. Polypharmacy and depression were associated with lower adherence.

Previous studies examining adherence with questionnaires or pharmacy refills also showed poor adherence among SLE patients, varying between 48% and 93%.^6,7,9,10,13^–^15,27 In our study, adherence to specific drugs did not significantly differ, which is consistent with data reported by Koneru et al.¹⁰ for prednisone, hydroxychloroquine, and other immunosuppressants.

Because SLE is a chronic disease that requires lifelong treatment, certain problems can hinder treatment adherence. In this study polypharmacy and depression were the only statistically significant characteristics associated with lower adherence rates. The association between education level and adherence in lupus patients has been studied previously.^7,9,11,14 but only Koneru et al. showed that low education was related to poor adherence.⁷ Treatment adherence is a complex process that results from the interaction of multiple factors. An individual’s health-related behavior depends on his/her perception of four critical areas: the severity of a potential illness, the perceived susceptibility to the effects of that illness, the benefits of taking an action, and the barriers to taking that action.²⁸ Beliefs and attitudes related to these domains could vary among patients and could potentially explain adherence variation.

Many patients with lupus have comorbidities such as hypertension, dyslipidemia, depression, coagulopathies, and osteoporosis, each of which may require one or more drugs for adequate control. This prescription medication burden can result in an increased incidence of adverse events and health-related costs which can also impact patient adherence.² In our study, the number of pills taken was a statistically significant predictor of non-adherence, pointing to the relevance of discussing the importance of each specific drug in patients receiving polypharmacy therapy.

Depression was another independent risk factor associated with non-adherence in our study. Our finding is consistent with that from a previous study, which showed that among patients with SLE, depression is associated with medication forgetfulness.⁹ Evidence shows that depressed patients have poorer overall self-care and lower self-efficacy, favoring poorer treatment adherence and lower quality of life.²⁹ Psychological factors play an important role in patient outcomes.³⁰ The present findings also suggest that better treatment adherence is associated with better mental health among people with SLE. However, the direction of this effect cannot be clearly established. On the one hand patients with better mental health status might be more adherent but it is also possible that adherence leads to increased self-efficacy and better mental health status.

Our study failed to show an association between treatment adherence and traditional SLE outcomes such as disease activity and organ damage. We believe this was partially due to the fact that we included in our study both patients who had been recently diagnosed and patients with established disease. The latter may have fewer changes in disease progression over a period of time of two years. Furthermore, our sample size only had power to detect moderate to large associations and smaller effects may have been missed. A limitation of our study is that MEMS® caps only record the time the patient opens the bottle and not direct intake. Nevertheless, this is also a limitation with all other measures of adherence such as pharmacy refills or self-report, and MEMS® technology is considered the gold standard for adherence, only surpassed by direct intake observation which is impractical for daily oral therapies. While patients may open their bottles and not take their medications, such a behavior over several months would require a persistent routine and continued malfeasance, which are unlikely.¹⁶

Adherence to treatment is crucial to the successful management of chronic diseases. Several interventions have been developed to improve adherence with long-term therapy but many are complex and only a few have been shown to improve adherence.³¹ An important determinant of adherence is communication between the health provider and the patient. There is evidence that patients with SLE who participate more actively in their physician visits have better outcomes.³² Effective communication about the nature of SLE, patient participation in decision-making, and motivating patients can conceivably improve therapeutic adherence.

Our study shows that almost 40% of SLE patients did not take their medications as prescribed. In addition, we also showed that patients with better adherence to oral SLE medications had significantly better mental health at two years. Depression and a high pill burden predicted non-adherence. Physicians and allied health care providers should attempt to identify patients at risk for non-adherence, and take the time to ask if a patient is taking the prescribed medications properly, and explore if any additional factors such as depression or other comorbidities may interfere with the adequate management of SLE.

Footnotes

Acknowledgements

We would like to thank Ms Vanessa Cox for initial data management.

Funding

This publication was made possible by Grant Number R01 AR47858 from NIAMS/NIH. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIAMS or NIH. Dr. Suarez-Almazor is a recipient of a K24 award from NIAMS.

Conflicts of interest statement

None declared.

References

Cramer

Roy

Burrell

. Medication compliance and persistence: terminology and definitions. Value Health 2008; 11: 44–47.

Osterberg

Blaschke

. Adeherence to medication. N Engl J Med 2005; 353: 487–497.

Petri

Perez-Gutthann

Longenecker

Hochberg

. Morbidity of systemic lupus erythematosus: role of race and socioeconomic status. Am J Med 1991; 91: 345–353.

Adler

Chambers

Edwards

Neild

Isenberg

. An assessment of renal failure in an SLE cohort with special reference to ethnicity, over a 25-year period. Rheumatology 2006; 45: 1144–1147.

Bruce

Gladman

Urowitz

. Factors associated with refractory renal disease in patients with systemic lupus erythematosus: the role of patient nonadherence. Arthritis Care Res 2000; 13: 406–408.

De Achaval

Suarez-Almazor

. Treatment adherence to disease-modifying antirheumatic drugs in patients with rheumatoid arthritis and systemic lupus erythematosus. Int J Clin Rheumatol 2010; 5: 313–326.

Koneru

Kocharla

Higgins

. Adherence to medications in systemic lupus erythematosus. J Clin Rheumatol 2008; 14: 195–201.

Ward

Lotstein

Bush

Lambert

van Vollenhoven

Neuwelt

. Psychosocial correlates of morbidity in women with systemic lupus erythematosus. J Rheumatol 1999; 26: 2153–2158.

Julian

Yelin

Yazdany

. Depression, medication adherence, and service utilization in systemic lupus erythematosus. Arthritis Rheum 2009; 61: 240–246.

10.

Koneru

Shishov

Ware

. Effectively measuring adherence to medications for systemic lupus erythematosus in a clinical setting. Arthritis Rheum 2007; 57: 1000–1006.

11.

Chambers

Raine

Rahman

Hagley

De Ceulaer

Isenberg

. Factors influencing adherence to medications in a group of patients with systemic lupus erythematosus in Jamaica. Lupus 2008; 17: 761–769.

12.

Chambers

Raine

Rahman

Isenberg

. Why do patients with systemic lupus erythematosus take or fail to take their prescribed medications? A qualitative study in a UK cohort. Rheumatology 2009; 48: 266–271.

13.

Costedoat-Chalumeau

Amoura

Hulot

. Very low blood hydroxychloroquine concentration as an objective marker of poor adherence to treatment of systemic lupus erythematosus. Ann Rheum Dis 2007; 66: 821–824.

14.

Garcia-Gonzalez

Richardson

Garcia Popa-Lisseanu

. Treatment adherence in patients with rheumatoid arthritis and systemic lupus erythematosus. Clin Rheumatol 2008; 27: 883–889.

15.

Rojas-Serrano

Cardiel

. Lupus patients in an emergency unit. Causes of consultation, hospitalization and outcome. A cohort study. Lupus 2000; 9: 601–606.

16.

Urquhart

. The electronic medication event monitor. Lessons for pharmacotherapy. Clin Pharmacokinet 1997; 32: 345–356.

17.

Feinstein

. On white-coat effects and the electronic monitoring of compliance. Arch Intern Med 1990; 150: 1377–1378.

18.

Tan

Cohen

Fries

. The 1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1982; 25: 1271–1277.

19.

Andresen

Malmgren

Carter

Patrick

. Screening for depression in well older adults: evaluation of a short form of the CES-D (Center for Epidemiologic Studies Depression Scale). Am J Prev Med 1994; 10: 77–84.

20.

Sherbourne

Stewart

. The MOS social support survey. Soc Sci Med 1991; 32: 705–714.

21.

Ware

Jr Sherbourne

. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992; 30: 473–483.

22.

Guzman

Cardiel

Arce-Salinas

Sanchez-Guerrero

Alarcon-Segovia

. Measurement of disease activity in systemic lupus erythematosus. Prospective validation of 3 clinical indices. J Rheumatol 1992; 19(10): 1551.8–1551.8.

23.

Gladman

Ginzler

Goldsmith

. The development and validation of the systemic lupus international collaborating clinics American College of Rheumatology damage index (SLICC = ACR). Arthritis Rheum 1996; 39: 363–369.

24.

De Klerk

Van der Heijde

Van der Tempel

van der Linden

. Development of a questionnaire to investigate patient compliance with antirheumatic drug therapy. J Rheumatol 1999; 26: 2635–2641.

25.

De Klerk

Van der Heijde

Landewe

. Patient compliance in rheumatoid arthritis, polymyalgia rheumatica, and gout. J Rheumatol 2003; 30: 44–54.

26.

De Klerk

Van der Heijde

Landewe

. The compliance-questionnaire-rheumatology compared with electronic medication event monitoring: a validation study. J Rheumatol 2003; 30: 2469–2475.

27.

Sailler

Puissant

Meliani

. Blood concentrations of hydroxychloroquine and its desethyl derivative correlate negatively with the percentage of CD45RO+ cells among CD4+ lymphocytes in hydroxychloroquine-treated lupus patients. Ann N Y Acad Sci 2007; 1108: 41–50.

28.

Johnson

. The medication adherence model: a guide for assessing medication taking. Res Theory Nurs Pract 2002; 16: 179–192.

29.

Kyrius

Nankervis

Reddy

. The relationship of depression to treatment adherence, quality of life and health outcomes in type 1 diabetes mellitus. E J Appl Psychol 2006; 2: 3–14.

30.

Carballo

Cadarso-Suarez

Carrera

IFJ

. Assessing relationships between health-related quality of life and adherence to antiretroviral therapy. Qual Life Res 2004; 13: 587–599.

31.

Haynes

Ackloo

Sahota

McDonald

Yao

. Interventions for enhancing medication adherence. Cochrane Database Syst Rev 2008; CD000011–CD000011.

32.

Ward

Sundaramurthy

Lotstein

Bush

Neuwelt

Street

Jr . Participatory patient-physician communication and morbidity in patients with systemic lupus erythematosus. Arthritis Care Res 2003; 49: 810–818.