Article Text

Download PDFPDF

OP0338 Relationship of provider density on total joint replacement outcomes
  1. B. Mehta1,2,3,
  2. J. Szymonifka4,
  3. S.A. Dey3,
  4. I.Y. Navarro-Millan2,3,
  5. L.A. Mandl2,3,
  6. A.R. Bass2,3,
  7. L.A. Russell2,3,
  8. M.L. Parks5,6,
  9. M.P. Figgie5,6,
  10. J.T. Nguyen4,
  11. S.M. Goodman2,3
  1. 1Columbia Mailman School of Public Health
  2. 2Rheumatology, Weill Cornell Medicine
  3. 3Rheumatology
  5. 5Orthopaedics, Weill Cornell Medicine
  6. 6Orthopaedics, Hospital for Special Surgery, New York, USA


Background The proportion of providers in a geographical location (provider density) has been associated with improved surgical outcomes in hand and wrist surgery, appendicitis and other high-volume procedures, demonstrating the importance of access to care.

Objectives The purpose of this study is to assess the association of provider density with Total Knee Replacement (TKR) and Total Hip Replacement (THR) outcomes.

Methods Individual patient-level variables were obtained from a single institution TKR and THR registry between 5/1/07 and 2/1/11. Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain and function scores at baseline and 2 years after elective TKR and THR were collected and retrospectively analysed. We included patients living in New York, Connecticut and New Jersey which is the catchment area for the hospital. Individuals’ geocodable addresses were used to obtain the number of primary care physicians (PCPs) and specialists in each census tract using the ArcGIS software. Provider density was calculated and broken down into percentage quartiles. Comparisons were made using Kruskal-Wallis tests.

Results A total of 3606 TKR, and 4295 THR patients were included. (figure 1) Mean number of PCPs were 5.4(SD 9.5) and specialists were 15.8(SD 41.1) per census tract. The median number of PCPs were 2 (IQR 1, 6) and specialists were 3 (IQR 1, 13). Neighbourhood poverty or education correlated poorly with the number of PCPs and specialists. Baseline WOMAC pain and function scores (table 1) are statistically significantly better in neighbourhoods with a higher proportion of specialists, but not PCPs. These differences were no longer present 2 years after surgery.

Abstract OP0338 – Table 1

Association of specialist proportions in neighbourhoods with WOMAC* scores

Abstract OP0338 – Figure 1 Map of New York, New Jersey and Connecticut with census tract specialist provider density and baseline WOMAC pain scores of individual patients.

Conclusions Patients from neighbourhoods with fewer specialists seek arthroplasty with worse baseline pain and function than those from neighbourhoods with more specialists. However, once these patients receive arthroplasty (i.e. specialty care), these differences resolve by 2 years. These data suggest that once a patient can access specialty care in the health care system, their outcomes improve despite worse baseline pain and function.

Disclosure of Interest B. Mehta: None declared, J. Szymonifka: None declared, S. Dey: None declared, I. Navarro-Millan: None declared, L. Mandl Grant/research support from: Boehringer-Ingelheim, A. Bass Grant/research support from: Abbott, Pfizer, L. Russell: None declared, M. Parks Grant/research support from: Zimmer, Consultant for: Zimmer, M. Figgie Shareholder of: Mekanika, J. Nguyen: None declared, S. Goodman: None declared

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.