Does Systemic Lupus Erythematosus Care Provided in a Lupus Clinic Result in Higher Quality of Care Than That Provided in a General Rheumatology Clinic?

Abstract

Purpose:

We compared the quality of care received by SLE patients in two settings within the same academic institution (Lupus Clinic, LC and General Rheumatology Clinic, GRC) using validated SLE quality measures (QM).

Methods:

150 consenting, consecutive SLE patients receiving longitudinal care at Rush University GRC (73) and subspecialty LC (77) were recruited. An updated QM survey and retrospective medical chart review was used for evaluation of each QM. The overall and individual (20) QM performance was calculated and compared between two groups. Data on number of SLE patients seen by each rheumatologist was collected to assess the relationship between SLE patient volume and QM.

Results:

Overall QM performance was significantly greater among LC SLE patients [85.8% vs. 70.2%, P= 0.001]. Differences noted among the two groups were in sunscreen counseling (98.7% vs. 83.6%, P=0.001), antiphospholipid antibody testing (71.4% vs. 37%, P< 0.001), pneumococcal vaccination (84.8% vs. 48.8%, P< 0.001), bone mineral density testing (94.2% vs. 54.5%, P<0.001), drug counseling (92.2% vs. 80.8%, P=0.04), steroid sparing agent and ACE inhibitor use (100% vs. 82%, P< 0.007 and 94.4% vs. 58.3%, P=0.03, respectively) and cardiovascular disease risk assessment (40.3% vs. 15.1%, P=0.01). There was a moderate correlation between the rheumatologists’ number of SLE patients seen and QM performance (rho 0.48, P<0.001).

Conclusion:

SLE patients seen in dedicated LC had better QM performance in this cross sectional single center study. In our health system, we also found indicators to suggest that rheumatologists with higher SLE patient-volume provide better quality of care.

INTRODUCTION

Systemic lupus erythematosus (SLE) is a chronic, multisystem autoimmune disorder with significant health impacts. Disease mortality has decreased over the past five decades, however morbidity remains high(1). Patients may develop other comorbidities, irreversible organ damage or experience toxic side effects from medications(2). Care for individuals with SLE is complicated by the extreme heterogeneity of clinical manifestations both in terms of diversity of organ involvement and differences between individuals. For instance, it is known that African American, Asian and Hispanic SLE individuals tend to develop severe disease, exhibit a greater number of symptoms and accumulate damage from lupus rapidly as compared to white patients(3). Higher quality of care (QOC) in SLE, as assessed by quality measures, is associated with improved patient outcomes over time by affecting accumulation of disease damage, a known predictor of mortality in lupus (4, 5).

Although a positive relationship between patient volume and quality or outcomes has been explored for many surgical procedures and for discrete events (e.g. hospital admissions), this relationship is less explored in the ambulatory care of chronic conditions, including SLE(6-8). In SLE, study examining the relationship between volume and outcomes found lower risks of in-hospital mortality for patients seen by high volume SLE physicians(9). Lately, subspecialty disease focused clinics such as Lupus clinics are being offered at Academic centers with the belief that the subspecialist’s experience, care processes, inter-disciplinary networks and increased awareness of disease pertinent guidelines and literature may potentially lead to better QOC and/or health outcomes, but this has not been formally assessed.

In this study a validated SLE quality measure (QM) set was used to assess and compare the quality of care received by SLE patients in two settings – a Lupus Clinic (LC) and a General Rheumatology Clinic (GRC) within the same academic institution.

METHODS

A Cross-sectional and retrospective chart review study was conducted at Rush University Medical Center, Chicago, Illinois, USA after approval by the Institutional Review Board. English speaking patients who met the American College of Rheumatology (ACR) classification criteria for SLE (10) and received longitudinal outpatient care in either LC or GRC were approached before or after their routine clinic visit. 150 consecutive patients who agreed to participate in the study by signing an informed consent form were recruited.

Quality measures:

Twenty measures covering specific aspects of SLE care including diagnosis, general preventive strategies (e.g., vaccinations, sun avoidance counseling, and screening for cardiovascular disease), osteoporosis prevention and treatment, drug toxicity monitoring, renal disease management and follow up, and aspects of reproductive health care have been validated for potential of improving patient outcome(11). Yazdany et al have previously evaluated thirteen of these twenty measures through data obtained from telephonically administered surveys (12). We updated and modified this survey for self-administration and collected data in person. These measures were used as such with the exception of updating the methotrexate monitoring measure to exclude chest radiography; consistent with more recent measure sets for methotrexate screening in rheumatoid arthritis (13).

The self-report form was pretested prior to use in the study and recruited patients completed it in paper format with questions pertaining to receipt of above quality measures before or after their routine clinic visit (LC or GRC). Medical chart review (up to 5 years) was performed for measures that were not amenable to self-report and for confirmation of receipt of quality measures already reported in the questionnaire by the patient since self-report can be limited by recall. Patient’s eligibility to qualify for each QM was also assessed through chart review. For example only those SLE patients who are post menopausal or have received prednisone equivalent to ≥ 7.5 mg/day for ≥ 3 months would qualify for bone mineral density evaluation (DEXA). The receipt of care for each quality measure (patients were eligible for) was scored nominally (Yes/No). Information on demographics (age, gender, education, ethnicity), number of ACR criteria met, disease activity, damage, medications and steroid dose was also gathered from chart review. Disease activity was assessed by SELENA–SLEDAI scale which is a 24-item scale that measures the presence of symptoms, laboratory values and conditions at the time of the visit or in the preceding 10 days (14) and disease damage was evaluated by SLICC/ACR Damage Index which measures damage during the preceding 6 months and the severity of injury in patients with SLE(15). These were assessed on the day of the study visit by the Principal investigator (MJ) and GRC rheumatologists did not partake in this.

Data on number of patients seen in the last 4-month duration by each rheumatologist and their corresponding diagnoses were collected to assess if there was any association between physician volume of SLE patients with the QOC rendered in SLE. Total number and proportion (individual rheumatologists’ total patient volume for that duration as the denominator) of lupus patients seen by each rheumatologist was thereafter calculated.

STATISTICS

Data on quality measures was analyzed for 150 participants using IBM SPSS software (Version 19, New York). Performance on quality measures was defined as percentage of times that care consistent with the measure was received among eligible individuals. Overall performance on the quality measures was examined by dividing the total number of QMs met divided by the total number of QMs that the participants were eligible for. The methodology used to evaluate performance of quality measures (Chart review-CR, Patient-self report-PS, or both) is detailed in the Results. Overall and individual performance on quality measures were then compared between LC and GRC patients using non-parametric tests. In these analyses, each measure was given equal weight rather than making assumptions about the relative importance of the individual measure since the empirical and theoretical basis for such judgments are lacking.

We also calculated spearman correlation coefficients between SLE patient volume (total number) with the QOC overall performance rate. Total patient volume for each of the 11 rheumatologists in the practice was obtained. Three of the 11 rheumatologists provided care within the LC. A correlation coefficient of <0.3 was considered as low, 0.3 to 0.5 as moderate and >0.5 as strong correlation. P-value of less than 0.05 was considered as significant for all results.

RESULTS

150 patients with SLE completed the study: 77 patients were receiving their care in the LC and 73 in the GRC. Mean (SD) age was 46 (15) years. This represents 30.30% of total number of SLE patients seen in the clinic (based on data collected for 4 month duration). Our sample is representative of 24.52% and 40.33% of SLE patients seen in LC and GRC, respectively. Ninety percent were women. Ethnic composition was as follows: Caucasians 31%, African Americans 46%, Hispanic 16%, Asian 5% and others 2%. There were no age or gender differences seen among the patients when stratified by LC or GRC (Table 1). Disease activity and damage scores (SELENA-SLEDAI and SLICC-ACR/SLE damage index), proportion of patients with primary care physician, use of hydroxychloroquine and use of other immunosuppressive agents were also similar in the two groups. Patients seen in the LC had longer disease duration and had more ACR criteria. Fewer patients {25 (32.5%)} from the LC were on steroids currently as compared to those from GRC {35 (47.9%), P=0.05}.

TABLE 1:

DESCRIPTIVES OF STUDY COHORT

	LUPUS CLINIC	GENERAL RHEUMATOLOGY CLINIC	P-value
Number of study patients (n ; %)	77 ; 51.3	73 ; 48.6
Age (mean ± SD) years	47.2 ± 14.9	45.0 ± 15.1	0.37
Gender: Female (n ; %)	67 ; 88.3	68 ; 91.8	0.48
Ethnicity/Race (n;%)
White	28;36.8	18;24.7
Black	28;36.8	41;56.2
Asian	5;6.6	3;4.1
Hispanic	13; 17.1	11; 15.1
Other	2;2.6	0;0	0.09
Insurance Status (n;%)
PPO	33;42.9	23;31.5
HMO	1;1.3	4;5.5
Medicare	20;26	17;23.3
Public aid	12;15.6	16;21.9
Other	10;13	13;17.8	0.29
Primary Care Physician: Yes (n ; %)	73 ; 94.8	68 ; 90.4	0.36
Disease Duration (mean ± SD) years	9.8 ± 7.1	7.0 ± 6.2	0.01
Number of ACR criteria met (mean ± SD)	5.3 ± 1.6	4.5 ± 1.0	0.01
Frequency of ACR criteria met:
Malar Rash (n ; %)	47 ; 61	26 ; 35.6	0.002
Discoid Rash (n ; %)	18 ; 23.4	7 ; 9.6	0.03
Photosenstivity (n ; %)	38 ; 49.4	24 ; 32.4	0.05
Oral ulcers (n ; %)	34 ; 44.2	26 ; 35.6	0.38
Arthritis (n ; %)	56 ; 72.7	64 ; 87.7	0.02
Serositis (n ; %)	24 ; 31.2	16 ; 21.9	0.2
Renal disorder (n ; %)	19 ; 24.7	16 ; 20.5	0.55
Neurological disorder (n ; %)	8 ; 10.4	2 ; 2.7	0.06
Hematological disorder (n ; %)	29 ; 37.7	24 ; 32.9	0.54
Immunologic disorder (n ; %)	70 ; 90.9	63 ; 86.3	0.37
SELENA-SLEDAI(mean ± SD)	3.9 ± 2.4 (n=52)	3.7 ± 1.8 (n=50)	0.75
SDI (mean ± SD)	1.6 ± 0.9 (n=30)	1.9 ± 1.0 (n=22)	0.39
Medications:
Steroids- ever (n ; %)	63 ; 81.8	61 ; 83.6	0.78
Steroids- current (n ; %)	25 ; 32.5	35 ; 47.9	0.05
Steroid Dose (median ; IQR)	12.4 ± 8.5	10.4 ± 8.5	0.28
Hydroxychloroquine(n ; %)	66; 85.7	61 ; 83.6	0.72
Methotrexate(n ; %)	16 ; 20.8	9 ; 12.3	0.17
Azathioprine(n ; %)	8 ; 10.4	9 ; 12.3	0.71
Mycophenolate mofetil(n ; %)	20 ; 26	12 ; 16.4	0.15
Other medications*(n ; %)	11 ; 14.3	7 ; 9.6	0.38

• Other medications include cyclophosphamide, rituximab, belimumab, cyclosporine, quinacrine, tacrolimus and ustekimumab

Quality Measures

Receipt of care consistent with quality measures was significantly higher among LC patients as compared to patients seen in GRC {85.8% (IQR: 19.6%) vs. 70.2% (IQR: 11.5%); P=0.001}. Patients seen in the LC were more often offered preventive measures such as counseling for use of sunscreen (LC: 98.7% vs. GRC: 83.6%; P=0.001), recommendation for pneumococcal vaccine if on immunosuppressive medication/s (LC: 84.8% vs. GRC: 48.8%; P<0.001), prescription of angiotensin converting enzyme (ACE) inhibitor if proteinuric (LC: 94.4% vs. GRC: 58.3%, P=0.03) and assessment and counseling for cardiovascular disease risk (LC: 40.3% vs. GRC: 15.1%, P=0.01, Table 2). Over seventy percent of patients received testing for antiphospholipid antibodies within 6 month of lupus diagnosis in LC as compared to only 37% in the GRC (P<0.001). LC patients were also more likely to get bone densitometry and prescription of steroid sparing agent if on steroids for more than three months as compared to patients seen in GRC (LC: 94.2% vs. GRC: 54.5%, P<0.001 and LC: 100% vs. GRC 82%, P< 0.007, respectively).

TABLE 2:

PERFORMANCE ON QUALITY MEASURES (QM)

		LUPUS CLINIC			GENERAL RHEUMATOLOGY CLINIC			P- Value
QM No.	Description of QM	QM eligible (N)	Met QM (n)	PP (%)	QM eligible (N)	Met QM (n)	PP (%)
1 (CR)	ANA, CBC, Platelet, Creatinine, UA at diagnosis of lupus	77	77	100	73	72	98.6	0.49
2 (CR)	AntidsDNA, C3/4, APL within 6 months of diagnosis	77	55	71.4	73	27	37.0	<0.001
3 (CR+PS)	Counselling for use of sunscreen	77	76	98.7	73	61	83.6	0.001
4 (CR+PS)	Infleunza vaccine in last year if on ISM	46	45	97.8	43	38	88.4	0.1
5 (CR+PS)	Pneumococcal vaccine if on ISM	46	39	84.8	43	21	48.8	<0.001
6 (CR+PS)	DEXA if have received ≥7.5 mg/day CS for ≥3 months	52	49	94.2	44	24	54.5	<0.001
7 (CR+PS)	Calcium and Vitamin D if have received ≥7.5 mg/d CS for ≥3 months or is post-menopausal	59	46	78	57	41	71.9	0.45
8 (CR+PS)	Antiresorptive agent if have received ≥7.5 mg/d CS for ≥1 month & central T score ≤2.5 or h/o fragility fracture	12	12	100	7	7	100	N\A
9 (CR+PS)	Counselling about risks/benefits of new medication	77	71	92.2	73	59	80.8	0.04
10 (CR)	Baseline tests at initiation of drugs	74	73	98.6	72	67	93.1	0.11
11 (CR)	Monitoring for drug toxicity	73	66	90.4	69	60	87	0.6
12 (CR+PS)	Steroid sparing agent if have taken ≥10 mg/day CS for ≥3 months	46	46	100	37	30	81.1	0.002
13 (CR)	Follow up tests (UA, CBC, Creatinine) done for LN at every 3 months	19	14	73.7	10	6	60	0.68
14 (CR)	Treatment with ISM & CS within 1 month of diagnosis of Class 3/4 LN	13	13	100	14	14	100	N\A
15 (CR)	Antihypertensive if have proteinuria ≥ 300 mg/d or GFR < 60 ml/min & ≥ 2 BP readings > 130/80	15	14	93.3	15	15	100	1.00
16 (CR+PS)	ACE inhibitor or ARB if have proteinuria ≥ 300 mg/d	18	17	94.4	12	7	58.3	0.03
17 (CR+PS)	Assessment of CVD risk & counseling	77	31	40.3	73	11	15.1	0.01
18 (CR)	Tests in pregnancy (AntiSSA/SSB, APL)	9	6	66.7	7	2	28.6	0.32
19 (CR)	Treatment of APS in future pregnancies	1	1	100	1	1	100	N\A
20 (CR+PS)	Reproductive health counseling	29	26	89.7	20	13	65.0	0.07

Abbreviations: PP - Performance percentage, ANA - Antinuclear antibody, CBC - Complete Blood Count, UA - Urinalysis, APL - Anti- phospholipid antibodies, ISM - Immunosuppressive medications, CS - Corticosteroids, HCQ - Hydroxychloroquine, MTX - Methotrexate, MMF - Mycophenolate mofetill, LN - Lupus Nephritis, ACE – Angiotensin converting enzyme, ARB - Angiotensin receptor blocker, CVD – Cardiovascular Disease, APS - Antiphospholipid antibody syndrome.

There were no significant differences between the two groups of patients with respect to initial tests done at diagnosis, baseline and follow up testing for drug monitoring and treatment and follow up tests for lupus nephritis. Reproductive health counseling and testing for antiphospholipid antibodies during pregnancy was better in LC patients but did not attain significance because of small number of eligible patients.

Patient volume-outcome relationship

Eleven rheumatologists within the rheumatology practice provided care for SLE patients, of which three provided care within LC. A total of 4,424 patients were seen by the 11 rheumatologists over the 4 month period. Percent of SLE patients seen may not be a true indicator of lupus volume, as the denominator may be variable for various rheumatologists, thus instead we obtained total number of SLE patients seen by the rheumatologists.

Of the 4,424 patients seen by the 11 rheumatologists over a 4 month duration, 495 were SLE patients. Of the 495 SLE patients, 314 SLE patients were seen in the LC. In the GRC, in 4 months period 3,293 patients (mean 412 per rheumatologist) were seen, of which 181 were SLE (Mean 23 per rheumatologist, min-max 0-67). In contrast, in LC 1,131 (Mean 377 per rheumatologist) patients were seen, of which 314 were SLE (Mean 105 per rheumatologist, min-max 78-142). There was a moderate correlation between the rheumatologists’ total number of SLE patients seen and the providers QOC overall performance rate (rho 0.484, P <0.001). On linear regression with overall performance rate as the dependent variable and actual SLE patient volume as the independent predictor, nearly 20% variance in overall performance rate was predicted by the actual SLE patient volume of the rheumatologist (β 0.44, 95% CI 0.09-0.18, P<0.001).

DISCUSSION

We found that patients receiving care in a dedicated Lupus Clinic or by rheumatologists with a higher volume of SLE patients were more likely to receive care consistent with SLE quality measures in this single center, cross sectional study. Providing quality care in SLE is challenging, as patients require ongoing, time consuming, multidisciplinary care; physician comfort and skills with the diagnosis and treatment of SLE may be limited due to its low prevalence; complexities involved in SLE diagnosis and management are variable among patients and there are a paucity of approved, effective or safe therapeutic options.

Previous studies on QOC in SLE are based only on patient reported assessments and thus limited to concepts amenable to self-report (12). Self-report may also be affected by various biases. Our study, utilizing both survey and medical chart review methodology, indicates better quality of care in SLE when being treated in a LC as compared to GRC. There are no known differences in patient scheduling, check in, work processes and flows, check out among patients seen at either of the two clinics, and are housed within the same office space. Therefore plausible explanations for our findings may be (a) greater sensitivity of LC rheumatologists to the importance of preventive care and disease monitoring among SLE patients due to their scholarly interests, (b) improved processes of care, sensitivity or expertise due to increased SLE patient volume in the lupus clinic and, (c) the uncommon occurrence and complex nature of SLE may increase the importance of prior experience in the care of these patients.

It is important to note that all our quality measures are process measures, focusing on what health care providers do in delivering care. They are easy to interpret and very sensitive to detect differences in quality of care, especially when linked to particular outcomes. Yazdany et al. elucidated the impact of quality measures in SLE; processes of care (defined by SLE quality measures) have been strongly linked to subsequent accumulation of disease damage, an important disease outcome (4). Higher disease damage has been showed to be associated with poor survival and is a predictor of mortality in lupus (5). Our study showed a statistically significant difference in 8 of the 20 quality measures between LC and GRC rheumatologists and suggested a trend towards difference in another one. The statistically significant measures pertain mostly to preventive processes and hence may prevent accrual of disease damage. Higher quality of care may also be related to satisfaction from treatment; our work shows that patient’s satisfaction with medical care is associated with better health related quality of life in SLE (16).

Research from surgical and medical care, shows that physician’s higher volume and expertise, and hospital volume has a direct impact on QOC, patient outcomes, and health care utilization (6-8, 17-20). Two studies have evaluated relationship between physician volume and outcome in SLE however these are limited to hospitalized SLE patients (9, 21). We compared average number of SLE patients seen between LC and GRC and found that LC have a much higher volume of SLE patients in comparison to GRC. This suggests that either the differences in QOC may be related to greater volume of SLE patients or the expertise from seeing a greater volume of SLE patients may be associated with better QOC. Both axioms that “practice makes perfect” explanation and the “selective referral” explanation may be operative in this context.

Limitations of our study include the small number of patients studied, which may be accountable for lack of generation of significant differences in some quality measures, for example those pertaining to aspects of reproductive health care in SLE patients. Different proportions of total SLE patients seen in either setting (24.5% in LC vs. 40.3% in GRC) were represented in the study, and this may theoretically impact the results. In context of greater than 60% of lupus patients receiving care in the LC, and the need to balance total study sample size in each arm, skewed enrollment, favoring GRC (in proportion) was expected and unavoidable. Study bias is another possibility as the author was one of the providers in the LC, however patient recruitment, data accrual, chart review and data management were performed by the first author without involvement of rheumatologists to avoid influence on patient selection or responses. Questionnaire administration and completion were done in privacy before or after the clinical visit, in the absence of the treating rheumatologist. It is plausible to have a disjunct between physician’s documentation of having provided a specific QI care or and patients recollection or knowledge of having received it for varied reasons. For our study purpose, quality measure was considered to be met if either patient self-report or medical chart documentation was noted on for that quality measure. Though we did not study this systematically, it was surprising that we seldom found discrepancies between patient self-report and medical chart reviews. This may be partially because (a) the patients had mean disease duration of 8.4 years and (b) medical chart reviews were conducted for the last 5 years duration from the study visit Another fact is that aspects of SLE care are also delivered by other physicians for example influenza vaccination by primary care physicians, treatment and follow up tests for lupus nephritis by nephrologists. These may hence not be representative of processes performed by rheumatologists and such a comparison then, may not apply to them. However, we can expect these effects to be random and equally distributed in both the groups. Lastly, only four month volume data was used as a surrogate for testing volume outcome relationship. This relationship needs to be further tested formally.

Disparities in the quality of care in SLE exist, despite availability of clinical practice guidelines and SLE QOC indices (12, 22). Physician (e.g. interaction and shared decision making) and patient characteristics (e.g. primary language, access to care) are known to be associated with varied practices in SLE (23-26), which in turn have important implications on disease outcome (27). Our study identifies the QOC differences related to the setting of a “disease focused clinic” care model and the volume-outcome relationship in SLE in clinic settings. Though our single center cross sectional study supports better QOC in lupus clinic, larger multicenter, prospective studies are required to further elucidate efficiencies gained (if any) when longitudinal care for SLE is provided either in the setting of a subspecialty Lupus clinic or by rheumatologists with greater volume of SLE patients. It would be pertinent to see if these efficiencies translate over time to measurable gains in health status and health resource utilizations.

Significance and Innovations –

1. Quality of care, an important predictor of patient outcome, is measured in a longitudinal care setting of Outpatient Clinic.

2. Effect of subspecialty disease focused clinic on quality of care in SLE is explored to assess the volume–outcome relationship in the care of SLE patients.