ACR Meeting Abstracts

ACR Meeting Abstracts

Abstract Number: 1422

Pain after Total Knee Arthroplasty: Poverty Modifies the Effect of Race and Education

Susan M. Goodman1, Anne R. Bass1, Michael Parks2, Kelly McHugh1, Meng Zhang3, Joseph Nguyen3, Yuo-Yu Lee3, Mark P. Figgie4 and Lisa Mandl5, 1Rheumatology, Hospital for Special Surgery, New York, NY, 2Orthopedics, Hospital for Special Surgery, New York, NY, 3Epidemiology and Biostatistics, Hospital for Special Surgery, New York, NY, 4Orthopaedics, Hospital for Special Surgery, New York, NY, 5Hospital for Special Surgery, New York, NY

Meeting: 2015 ACR/ARHP Annual Meeting

Date of first publication: September 29, 2015

Keywords: , , ,

Session Information

Date: Monday, November 9, 2015

Title: Orthopedics, Low Back Pain and Rehabilitation Poster

Session Type: ACR Poster Session B

Session Time: 9:00AM-11:00AM

Pain after Total Knee Arthroplasty: Poverty Modifies the
Effect of Race and Education 

Background/Purpose: Race and education are important
predictors of pain after total knee arthroplasty (TKA) in the United States
(US).  However, whether community poverty modifies the effect of race and
education on pain following TKA is unknown.  We assessed the interaction
between poverty and race, and poverty and education on WOMAC pain 2 years after
TKA. 

Methods:  Individual patient level variables were obtained
from a single institution registry of TKA performed between 5/1/2007 and
7/1/2010 including demographics, baseline and 2 year WOMAC pain (0-100; 100
best), HSS Expectations Score (1-100, 100 best) and geocodable US addresses.  Individual
patient level variables were then linked to US Census Bureau data at the census
tract level.  Statistical models including both patient level variables and
census tract level variables were constructed within multilevel
frameworks. Two multivariate linear mixed effect models with separate random
intercepts for each census tract were used to assess the interaction between
race and census tract poverty level, and the interaction between education and
census tract poverty level as predictors of WOMAC pain 2 years after TKA.

Results:  507/4225 (11.3%) of the TKA registry
patients were non-white, and non-white patients were less likely to be college
educated than whites (48.2 vs 62.8%, p-value≤0.0001).   In univariate
analysis, race, education and census tract poverty level (the percent of the
community living below the poverty level) were significantly associated with
WOMAC pain 2 years after TKA. In the first multivariate linear mixed effect
model it was estimated that non-whites had worse WOMAC pain scores than whites
and this difference increased with the increasing census tract poverty level (Table
2). The second multivariate linear mixed effect model showed that patients
without college education had worse WOMAC pain 2 years after TKA than those
with at least some college education, and this effect was magnified by greater
levels of census tract poverty (Table 3).

Conclusion:   Non-white race and a lack of college education
are both associated with worse WOMAC pain 2 years after TKA and this effect is
magnified by higher community poverty level.  The differential impact of
poverty, race, and education on pain after TKA needs more study.

 

Table 1:  Patient characteristics

 

Individual Level Variables

Total

(N=4225)

White

(N=3970)

Non-White

(N=507)

p-value

Age, mean ± std

67.1 ± 9.7

67.2 ± 9.6

65.4 ± 10.0

0.0001

Female, % 

61.8

60.9

70.7

<0.0001

Ethnicity

 

 

 

<0.0001

       Hispanic, %

3.9

1.2

26.7

 

       Non-Hispanic ,%

96.1

98.9

73.3

 

Co-morbidity

 

 

 

<0.0001

       0 Comorbidities, %

71.4

72.8

61.3

 

      +1 Comorbidity, %

28.6

27.2

38.7

 

Education

 

 

 

<0.0001

       College above, %

61.3

62.8

48.2

 

       No college, %

38.7

37.2

51.8

 

BMI, mean ± std

29.9 ± 5.9

29.8 ± 5.8

31.1 ± 6.4

<0.0001

Insurance

 

 

 

<0.0001

       Medicare%

62.2

63.2

50.6

 

       Medicaid%

2.0

0.54

13.3

 

       Other%

35.8

35.8

36.1

 

WOMAC pain at baseline, mean ± std   

54.0± 17.7

54.7 ± 17.5

48.8 ± 19.5

<0.0001

HSS Expectations Score, mean ± std

78.6 ± 18.0

78.5 ± 18.0

78.3 ± 18.0

0.84

 

Table 2:  Estimation based on a Linear Mixed Effects Model Assessing the Effect of Interaction between Race and Poverty at Census-tract Level on WOMAC Pain at 2 Years after TKA*

Race

Below poverty**

Estimated WOMAC Pain at 2 years

Standard Error

p-value

Estimated difference in 2 year  WOMAC Pain: (Non-white vs. White)

Standard Error

p-value

Non-white

10%

83.9

0.96

<0.0001

-3.01

 

1.02

0.0033

White

10%

86.9

0.42

<0.0001

Non-white

30%

80.4

1.54

<0.0001

-5.13

 

1.99

0.0104

White

30%

85.6

1.29

<0.0001

Non-white

50%

76.9

2.85

<0.0001

-7.25

3.65

0.0472

White

50%

84.2

2.32

<0.0001

Non-white

70%

73.5

4.25

<0.0001

-9.37

5.39

0.0822

White

70%

82.9

3.36

<0.0001

Non-white

90%

70.1

5.68

<0.0001

-11.49

7.15

0.1081

White

90%

81.6

4.41

<0.0001

TKA = total knee arthroplasty

*Linear mixed effect model controlling for age, gender, BMI, race, education, baseline WOMAC pain, expectations, Charlson-Deyo comorbidities, Percentage of population below poverty at census tract level (census tract poverty level)  and interaction of census tract poverty level and race (Census tract poverty level x race); estimates were calculated based on the following: WOMAC pain=53.73; age=66.51; BMI= 29.99; expectation score= 78.49; gender=female; education=college-above; comorbidities=0

** percentage of population in census tract below poverty level

 

Table 3:  Estimation based on a Linear Mixed Effects Model Assessing the Effect of Interaction between Education and Poverty at Census-tract Level on WOMAC Pain at 2 Years after TKA*

Education

Below poverty**

Estimated WOMAC Pain at 2 years

Standard Error

p-value

Estimated difference in 2 year  WOMAC Pain: (college- above vs. Non college)

Standard Error

p-value

College-above

10%

87.4

0.60

<0.0001

4.05

0.69

<0.0001

No college

10%

83.4

0.64

<0.0001

College-above

30%

86.8

1.49

<0.0001

6.92

1.99

0.0005

No college

30%

79.9

1.35

<0.0001

College-above

50%

86.3

2.66

<0.0001

9.80

3.59

0.0065

No college

50%

76.5

2.44

<0.0001

College-above

70%

85.7

3.85

<0.0001

12.67

5.22

0.0154

No college

70%

73.0

3.57

<0.0001

College-above

90%

85.1

5.05

<0.0001

15.54

6.86

0.0237

No college

90%

69.6

4.72

<0.0001

TKA = total knee arthroplasty

*Linear mixed effect model controlling for age, gender, BMI, race, education, baseline WOMAC pain, expectations, Charlson-Deyo comorbidities, Percentage of population below poverty at census tract level (census tract poverty level) and interaction of census tract poverty level and education (Census tract poverty level x education); estimates were calculated based on the following: WOMAC pain=53.73; age=66.51; BMI= 29.99; expectation score= 78.49; gender=female; race=non-white; comorbidities=0

** percentage of population in census tract below poverty level

 

Disclosure: S. M. Goodman, None; A. R. Bass, None; M. Parks, Zimmer, Inc, 5,Orthopeadic Research and Education Foundation, 6,American Academy of Orthopaedic Surgery Orthopaedic Learning Center, 6,New York State Society of Orthopaedic Surgeons, 6; K. McHugh, None; M. Zhang, None; J. Nguyen, None; Y. Y. Lee, None; M. P. Figgie, None; L. Mandl, Up To Date, 7,Annals of Internal Medicine, 9.

To cite this abstract in AMA style:

Goodman SM, Bass AR, Parks M, McHugh K, Zhang M, Nguyen J, Lee YY, Figgie MP, Mandl L. Pain after Total Knee Arthroplasty: Poverty Modifies the Effect of Race and Education [abstract]. Arthritis Rheumatol. 2015; 67 (suppl 10). https://acrabstracts.org/abstract/pain-after-total-knee-arthroplasty-poverty-modifies-the-effect-of-race-and-education/. Accessed .

« Back to 2015 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/pain-after-total-knee-arthroplasty-poverty-modifies-the-effect-of-race-and-education/