When Value Metrics Miss the Mark

One goal of value-based care is to achieve high-quality patient outcomes while incentivizing cost containment by using alternative payment models (APMs). Total knee arthroplasty (TKA) has been a major focus of APMs because it is a high-volume and costly surgery to the U.S. Centers for Medicare & Medicaid Services (CMS). Traditionally, cost reduction has been the primary factor used to improve value in health care, whether in capitated insurance payment programs developed by health maintenance organizations (HMOs) or subsequent APMs such as the Comprehensive Care for Joint Replacement (CJR) model, which achieved substantial cost savings in TKA mostly by reducing post-acute care1. Incentivizing improvements in quality of care has been more complicated, with many APMs focusing on a predefined set of complications (for example, myocardial infarction) or process improvements (timely antibiotic administration)2. The downside of these quality measures is that they are not patient-centered. For example, patients may not have complications or readmissions, yet they may not return to their desired level of function or be satisfied with the procedure. One mechanism to provide a patient-centered measure is the use of patient-reported outcome measures (PROMs). There are a number of validated PROMs that capture various domains of patient experience, function, and health status after TKA, including the Knee Injury and Osteoarthritis Outcome Score (KOOS). Use of previously validated instruments such as KOOS can have difficulty in achieving high response rates across large populations because of survey length (42 items in the case of KOOS), but improvements in technology and the development of short forms such as the Knee Injury and Osteoarthritis Outcome Score, Joint Replacement (KOOS JR), with 7 items, have expanded the ability to collect these surveys across large populations of patients. The wider applicability of these PROMs has culminated in the CMS incorporating them into future APMs as a pay-for-performance quality measure, for example, the mandatory Patient-Reported Outcome-Based Performance Measure (PRO-PM) evaluation, which requires KOOS JR collection and reporting as a CMS quality measure for hospitals3. Although PROMs provide a patient-centered quality outcome, defining “success” using these measures is more complex. For example, within the PRO-PM, the CMS defines a successful procedure as achieving a ≥20-point increase, considered substantial clinical benefit (SCB), in the KOOS JR from before to after TKA3. Patient satisfaction is an alternative way of defining a patient-centered outcome, as described by Zheng et al. in their article. Ideally, patient satisfaction with the procedure and PROMs would correlate with each other, so that patients who achieved excellent physical function and quality of life would also be satisfied. Unfortunately, as shown by Zheng et al., commonly used PROMs do not always serve as a proxy for patient satisfaction in TKA, with 83.5% of patients being satisfied but only 64.3% meeting the CMS definition of surgical success. This suggests that using the CMS definition of a successful result on the KOOS JR may underestimate the number of patients satisfied by the procedure and may not represent the quality, and hence the value, being created. One of the most important reasons for this finding appeared to be the ceiling effects of the postoperative KOOS JR in the setting of TKA: patients with higher baseline scores (≥60) had more difficulty meeting the CMS definition of surgical success, despite being satisfied with the procedure (88.3% were satisfied, but only 34.6% had successful surgeries based on the CMS criteria). There are a number of possible solutions to this problem. Previous APMs have illustrated the importance of risk adjustment when measuring quality. For example, risk-standardized complication rates are used in the Hospital Inpatient Quality Reporting Program to avoid hospitals deliberately selecting healthy patients who are less likely to have an adverse outcome4. Similar mechanisms may apply to PROMs, and Zheng et al. show that the inclusion of the baseline KOOS JR in the risk adjustment measure improves its discrimination of patient satisfaction. Other possible ways to improve the use of PROMs as a quality measure for TKA would be using alternative definitions of KOOS JR improvement such as the minimum important change (MIC) or patient acceptable symptom state (PASS), using a score with fewer ceiling effects, or including an additional direct measure of patient satisfaction as part of the APM. Value-based care that aligns the incentive of cost containment and achievement of quality outcomes is important, but incentivizing quality without unintentionally penalizing hospitals and physicians remains challenging.