CONFLICTING COMPARATIVE EFFECTIVENESS EVIDENCE OF HYALURONIC ACID USING REAL WORLD DATA EVIDENCE. AN UPDATE USING A HIGH DIMENSIONAL PROPENSITY SCORE MATCHING APPROACH

OBJECTIVES: Recent evidence from large observational studies suggest that intraarticular HA injections are associated with a delay in total knee replacement among patients with knee osteoarthritis (OA). However, this recent industry sponsored evidence is limited considering residual confounding and...

Full description

Saved in:
Bibliographic Details
Published inValue in health Vol. 20; no. 5; p. A142
Main Authors Shewale, AR, Barnes, L, Fischbach, LA, Ounpraseuth, ST, Painter, JT, Martin, BC
Format Journal Article
LanguageEnglish
Published Lawrenceville Elsevier Science Ltd 01.05.2017
Subjects
Arthritis
Corticosteroids
Disease management
Drugs
Falsification
Hyaluronic acid
Intervention
Joint replacement surgery
Knee
Knee replacement
Observational studies
Osteoarthritis
Pain
Propensity
Sensitivity analysis
Surgery
Online AccessGet full text

Cover

More Information
Summary:OBJECTIVES: Recent evidence from large observational studies suggest that intraarticular HA injections are associated with a delay in total knee replacement among patients with knee osteoarthritis (OA). However, this recent industry sponsored evidence is limited considering residual confounding and potential selective loss to follow up. The objective of this study was to evaluate the effectiveness of intra-articular HA injections for knee OA management. METHODS: A nested cohort of persons with knee OA visiting a specialist and a recent history of pain medication use was created using the Lifelink Plus claims (2010-2015) to compare the risk of composite (any) knee surgical interventions, total (TKR) / unicompartmental knee replacement (UKR) and TKR only among HA users and two comparison groups: corticosteroid (CS) users and HA/CS non-users. A high dimensional propensity score (hdPS) was used to match HA users with HA/ CS non-users and with CS users on background covariates. The risk of surgical interventions among HA users relative to the comparison groups was assessed using Cox proportional hazard models. A wide range of sensitivity analyses were also conducted including falsification tests. RESULTS: Among 13,849 patients, 797 were HA users, 5,327 were CS users, and 7,725 were HA/CS non-users. After hdPS matching, the risk of composite surgical interventions did not differ between HA users and HA/CS non-users (HR=0.88, 95% CI 0.67-1.16) and CS users (HR= 0.89, 95% CI 0.65-1.12). A sensitivity analysis that restricted the sample to patients who had any type of knee surgery (an approach used in recent studies), showed a significant lower risk of surgery for those using HA relative to HA/CS non-users (HR=0.87, 95% CI 0.79-0.95) CONCLUSIONS: There were no significant differences in the risk of surgical interventions among HA users compared to HA/CS non-users and CS users after accounting for residual confounding using a high dimensional propensity score.
ISSN:1098-3015
1524-4733