Journal Logo

Patient-reported Outcomes

Research Priorities for Optimal Use of Patient-reported Outcomes in Quality and Outcome Improvement for Total Knee Arthroplasty

Franklin, Patricia D. MD, MBA, MPH

Author Information
JAAOS - Journal of the American Academy of Orthopaedic Surgeons: February 2017 - Volume 25 - Issue - p S51-S54
doi: 10.5435/JAAOS-D-16-00632
  • Free


Established with a $12 million program project award from the Agency for Healthcare Research and Quality (AHRQ), Functional and Outcomes Research for Comparative Effectiveness in Total Joint Replacement (FORCE-TJR), centered at the University of Massachusetts Medical School, is a national prospective cohort of more than 28,000 patients with TJR from more than 150 surgeons. FORCE-TJR was initially established to define national outcome norms, serve as a registry for outcome monitoring, and conduct comparative effectiveness research. Parallel to the current US surgical practice, 75% of patients under FORCE-TJR were enrolled from community-based surgeons, including fellowship-trained and general orthopaedists in urban and rural locations, as well as teaching and nonteaching hospitals. Representatives from community orthopaedic practices across 28 US states stratified by geography (eg, urban/rural) and clinical volume were participated to ensure a balanced representation of diverse patients and practices. FORCE-TJR collects preoperative data from patient questionnaires completed during a preoperative physician encounter. Patients' self-reported data include demographic information (age, gender, race, marital status, education, insurance, household income, smoking status, height, and body mass index), modified Charlson comorbidity score, and musculoskeletal comorbidity (pain in the lumbar spine and nonoperative knees and hips). In addition, all patients complete standardized patient-reported outcomes (PROs) before and after surgery, including joint-specific measures (the Hip Disability and Osteoarthritis Outcome Score [HOOS] or Knee Injury and Osteoarthritis Outcome Score [KOOS]) and global function measures (Short Form 36 [SF-36], Physical Component Score [PCS], and Mental Component Score [MCS]). Postoperative joint pain and function and the occurrence of any adverse event or revision are assessed at 6 months, 12 months, and annually.

With a postoperative PRO response rate exceeding 85%, FORCE-TJR now has the largest representative US cohort of comprehensive total knee arthroplasty (TKA) outcomes including PROs and has emerged as a leader in the orthopaedic community in PRO collection and interpretation.13 Using the FORCE-TJR cohort, the US national norms for 30- and 90-day complication rates and PROs at 6 and 12 months post-TJR were established.

FORCE-TJR provides risk-adjusted, comparative feedback to TJR surgeons to support quality improvement (QI) efforts. These data can be used to meet regulatory requirements such as the Centers for Medicare & Medicaid Services Patient Quality Reporting System (FORCE-TJR is a Qualified Clinical Data Registry), for maintenance of certification with the American Board of Surgeons, and to meet value-based payments for accountable care. Surgeons and hospitals are already using these data in managing bundled payment contracts (

Surgeons have access to outcome data for their patients with TJR including pain relief, functional gain, quality of life, and 90-day adverse events and 30-day readmissions at all hospitals. In addition, surgeons receive comparative reports of preoperative patient mix and patients' risk factors compared with other sites. Hospitals and surgeons can compare their patients with patients at other sites within FORCE-TJR and with the US national norms established by FORCE-TJR on pre-TJR pain and function and on key risk-adjustment factors.4

After reviewing their risk-adjusted data, surgeons understand how their patients differ when compared with those in the other sites and national norms. The next logical question is to learn how to address the patient's modifiable risks before surgery and optimize patient-care pathways to assure optimal, consistent outcomes. This paper outlines research strategies and priorities to use QI principles to interpret PROs when designing quality- and outcome-improvement activities.

Quality Improvement in Health care

Quality improvement methods are based on the principles of continuous process improvement.57 Deming and Juran's work in the 1960s–1970s in the manufacturing industry defined the principles of continuous QI that was subsequently adopted by the healthcare industry in the 1980s and 1990s. Over time, many modifications or revisions emerged including Lean initiatives, Six Sigma, and Total Quality Management. Despite the differences, at least two strategies transcend all models: (1) the use of data to define opportunities to improve consistent inputs and processes, and (2) the goal of reducing outcome variation. Translating QI principles to health care, specifically to TKA, we can extrapolate this model as follows: consistent inputs (or patients) and consistent processes (eg, surgical techniques, postoperative care pathways, and implant selection) will result in reduced variation in the outcome and improved patient outcomes. Research is needed to define the role of PROs in identifying preoperative patients' risks, tailoring perioperative care to patients' risk profiles, and reducing variation in primary outcomes including pain relief and functional gains.

Quality Improvement in Total Knee Arthroplasty

High-quality TKA care is achieved when the patient reports maximal knee pain relief and gains in physical function after safe and technically excellent surgical and postoperative care. In TKA QI, the inputs, or patients, vary by sociodemographic, behavioral, medical, musculoskeletal, and emotional health factors. Before TKA, the surgeon and patient must identify modifiable factors to reduce the patient's risk of adverse events. Hospitals and clinical teams define consistent patient-care procedures addressing surgical technique, prophylactic medications to avoid infections and clotting, and implant design to assure consistent best practices. The transition to home includes plans for progressive ambulation, rehabilitation, and pain management. Ultimately, the patients' primary outcomes are maximal knee pain relief and functional gain, followed by uniform implant performance and longevity.

Total knee arthroplasty is a successful surgical procedure with excellent mean or average outcomes in terms of pain relief and improved physical function, but variation in patient-reported functional gain persists.8 This variation defines the opportunity to improve consistent TKA care practices. For example, Figure 1 demonstrates the variation in pre-TKA function (KOOS activity of daily living [ADL]) and pain (inputs) and 6 months post-TKA ADL and pain in a sample of patients under FORCE-TJR. Mean preoperative ADL and pain scores represent severe functional limitations and significant knee pain. However, the figure also illustrates wide variation in pre-TKA levels of knee disability (ADL) and pain. After TKA, mean improvement in both pain and function is dramatic. While the majority of patients have ADL and pain scores in the 90–100 range (representing minimal-to-no pain or disability), some patients linger below a score of 70. What changes in patient preparation for TKA (inputs) or perioperative care procedures, such as rehabilitation and/or medication management, could reduce variation in outcomes?

Figure 1
Figure 1:
Preoperative (black) and 6 months postoperative (green) KOOS ADL and KOOS Pain score distributions of patients under FORCE-TJR illustrate wide variations in preoperative patient symptoms and postoperative ADL and pain scores. ADL = activity of daily living, KOOS = Knee Injury and Osteoarthritis Outcome Score, TKA = total knee arthroplasty.

FORCE-TJR collects a wide array of preoperative risk factors that can be examined to assess how variation in patient (inputs) comorbidity affects variation in outcomes (Figure 2). For example, poor emotional health (SF-36/MCS) has been associated with poorer ADL after TKA.911 How can we support low MCS patients differently in the pre- and postoperative period to improve ADL outcomes? One approach may be to tailor postoperative pathways to augment patient support for low MCS patients. FORCE-TJR collects discharge location (eg, skilled nursing facility and a home with family) that may provide some answers. In summary, research priorities include the following:

  • (1) Does postdischarge location and/or patient support affect post-TKA ADL?
  • (2) How do we tailor perioperative care for subgroups of patients at risk for poor functional gain to improve outcomes?

Figure 2
Figure 2:
FORCE-TJR measures include preoperative risk factors (inputs) and PROs, perioperative care and safety (process), implant, and PROs. KOOS = Knee Injury and Osteoarthritis Outcome Score, MSK = musculoskeletal, PRO = patient-reported outcome, PROMIS = Patient-Reported Outcomes Measurement Information System, VR12 = Veterans RAND 12-Item Health Survey.

Consistent Patients' Pre–total Knee Arthroplasty

Much research focuses on reducing patient risk factors before surgery to reduce the risk of complications and readmissions after TKA. Current strategies include careful preoperative evaluation to optimize risk factors such as the HgA1c for diabetics or fluid balance for patients with cardiac conditions. Methicillin-resistant Staphylococcus aureus screening and prophylaxis are attempts to minimize infection risks. Finally, home assessments in advance of discharge aim to optimize support systems when patients are directly discharged to home. To date, it is unclear whether these strategies will influence post-TKA pain relief and functional gain. Research is needed to understand what patient-preparation interventions will reduce variation in PROs.

For example, FORCE-TJR research and others have documented impaired levels of function (SF-36/PCS or ADL) preoperatively, but actual patient scores vary widely.12 Patients with very low preoperative PCS do not achieve the same level of function (PCS) after TKA as patients with higher preoperative scores. Research is needed to determine whether we can identify optimal functional scores to achieve maximal post-TKA benefits.

Minority race (African American and Hispanic) has been associated with poorer outcomes. Recently, FORCE-TJR analyses have identified that more advanced preoperative musculoskeletal disease (osteoarthritis) exists among minority patients.13 In particular, African-American patients reported both significantly greater pain in the knee at the time of TKA (KOOS mean pain) and greater prevalence of moderate-to-severe pain in nonoperative knees and hip.

Risk-adjustment models of predictors of post-TKA function identify varied demographics, socioeconomic status, preoperative PCS, medical comorbidities, body mass index, and musculoskeletal comorbidities as significant predictors of both readmission and post-TKA function.14 Research is needed to understand which modifiable factors can be addressed before surgery to reduce variation in “inputs” (patients) to improve outcomes.

In summary, research priorities to efficiently address consistent patient preparation (inputs) include the following:

  • (1) Which comorbidities have the greatest impact on functional gain or on pain relief?
  • (2) How can clinicians best manage patients' factors before surgery to optimize function?
  • (3) Can we improve patients' preoperative preparation programs (motivation, engagement, and exercise) to optimize function or pain relief?

Consistent Care Processes for Total Knee Arthroplasty

How do processes of care (Figure 2) relate to variation in outcomes?

Different patient care procedures are implemented in the operating room, hospital, and in the postdischarge location. Several types of varied care processes may influence the variation in outcomes such as surgical techniques (eg, approach, implant fixation, and ligament balancing), medications (eg, pain management at home), and ambulation in the hospital and rehabilitation phase. Data from the New Zealand registry and FORCE-TJR suggest that implant selection may be associated with varied PROs. Patients with significant pain at 6–12 months after TKA report greater implant revision rates at 6 years. Possible explanations for the varied pain and implant failure could include different patient demands (eg, younger and more active) or different implant performances. Additional research is needed to better understand the role of PROs in implant performance surveillance.

Post-TKA pain management is important to recover from TKA as good pain control allows full engagement in rehabilitation. Preliminary (unpublished) data from FORCE-TJR show that while the majority of patients report significant reduction in pain at 2 weeks after TKA, approximately 30% of patients report persistent pain between 2 and 8 weeks.

Research is under way to evaluate the profile of patients with persistent pain to design strategies to tailor pain management and rehabilitation. In another study,15 the level of the self-managed home activity varies widely and influences outcome. Measures of activity (steps/day) post-TKA document significant difference in the level of mean steps/day by sex, with women reporting fewer steps. Women also report lower function at the time of TKA and poorer overall functional outcomes. Further research is needed to define the optimal levels of activity at home and strategies to use phone accelerometers or proprietary activity measures to engage patients during recovery.

In summary, research is needed to define best care processes and strategies for patients with TKA to tailor the practices to patient subgroups to optimize outcomes. For example, preoperative assessments could be used to guide personalized perioperative pain and activity prescriptions to optimize pain relief and functional gain at home after TKA.

Optimal Total Knee Arthroplasty Outcome Measurement

While PRO implementation is spreading rapidly, research is needed to define which PRO measure is more useful to assess TKA outcomes. No consensus exists as to which PRO measure (eg, long versus abbreviated versions) or calculation (eg, pre-to-post change, absolute level post-TKA) best captures surgery impact. PROs are used to assess individual patient's pain and function and at aggregate levels to sum all patients within a hospital or surgical practice. Abbreviated PROs may be acceptable for aggregate-level analyses but may not be precise enough at the patient level. More research is needed to evaluate what PRO version is best in what situation.

Patients' Factors and Functional Gain After Total Knee Arthroplasty

Today's patients with TKA are 3 years younger on average than patients in the late 1990s (mean age in 1997 = 69 years; mean age in 2013 = 66 years). Yet, we do not understand whether younger and older patients have similar functional trajectories, nor have we defined whether the recent patients' shift toward greater body mass index and more medical comorbidities are influencing the function after TKA. Figure 3 shows data from FORCE-TJR and compares trajectories of function (SF-36/PCS) of patients of different age groups at 1 and 2 years after TKA. Patients younger than 75 years have higher postoperative PCS than those older than 75 years at 1 and 2 years. Also, patients above 75 years tend to peak and decline at 2 years, which may be related to more comorbid conditions. However, knee-specific function (KOOS ADL) shows similar improvements at 1 and 2 years, across all ages.

Figure 3
Figure 3:
Function measured using the SF-36/PCS and KOOS ADL by age group at pre-TKA, 1 year, and 2 years. Although patients above 75 years report poorer global function after TKA compared with younger patients, all age groups report comparable knee-specific function over time. ADL = activity of daily living, KOOS = Knee Injury and Osteoarthritis Outcome Score, PCS = Physical Component Score, SF-36 = Short Form 36, TKA = total knee arthroplasty.

In summary, research priorities to understand how patients' factors influence PROs after TKA should answer questions such as

  • (1) Why do pain/function outcomes vary by age, sex, race/ethnicity, and clinical profiles?
  • (2) Will tailored preoperative interventions, surgical care, or implants reduce variability in outcome? If not, can we give patients tailored estimates of outcomes to inform TKA decisions?

Where, When, and How to Collect Patient-Reported Outcomes

FORCE-TJR found that direct-to-patient dissemination of PROs after TKA achieved higher PRO collection rates. When PROs were sent to patients at their home (by email or mail), 85% returned the PRO in a narrow window of time. However, when PROs were administered in the office, at the time of the patient's visit, the rate of return averaged 60% because not all patients returned to the office. Registries in the UK and Sweden send PRO questionnaires to the patients at home.16 Another method of collecting PROs is through a phone application. US health systems must determine whether direct-to-patient PRO technologies best serve complete PRO collection, while the measures themselves can be stored in the patient's health record.17

In summary, to achieve consistent TKA outcomes, the role for consistent inputs (patients) and processes of care needs to be clarified. We need a better understanding of how heterogeneity in patient factors influences pain, function, and implant performance. Research must explore how to reduce patients' risk factor variation before TKA and how to tailor care to minimize the influence of patient factors on outcomes.

Ultimately, real-time scored preoperative patient risk profiles, including PROs, can help tailor care and improve outcomes. Patient-reported outcomes can inform shared decisions for surgery, help tailor perioperative and postdischarge care, and monitor in-home and rehabilitation care.

Research is needed to understand what patient, surgical, postoperative, and home care procedures optimize function and pain relief, and how these procedures are best tailored to patient profiles. Finally, research is needed to find the best measure(s) of outcomes (pain, function, and implant performance) to guide care decisions and quality for patients with TKA.


1. Ayers DC, Franklin PD: Joint replacement registries in the United States: A new paradigm. J Bone Joint Surg Am 2014;96:1567–1569.
2. Franklin PD, Harrold LR, Ayers DC: Incorporating patient reported outcomes in total joint arthroplasty registries: Challenges and opportunities. Clin Orthopaedics Relat Res 2013;471:3482–3488.
3. Franklin PD, Allison JJ, Ayers DC: Beyond joint implant registries: A patient-centered research consortium for comparative effectiveness in total joint replacement. JAMA 2012;308:1217–1218.
4. Zheng H, Li W, Harrold L, Ayers DC, Franklin PD: Web-based comparative patient- reported outcome feedback to support quality improvement and comparative effectiveness research in total joint replacement. EGEMS (Wash DC) 2014;2:1130.
5. Deming WE: The New Economics for Industry, Government, and Education. Cambridge, MA, The MIT Press, 2000.
6. Deming WE: Out of the Crisis. Cambridge, MA, Massachusetts Institute of Technology Center for Advanced Engineering Study, 1986.
7. Juran J: Quality Control Handbook. New York, NY, McGraw-Hill, 1951.
8. Franklin PD, Li W, Ayers DC: The Chitranjan Ranawat Award: Functional outcome after total knee replacement varies with patient attributes. Clin Orthop Relat Res 2008;466:2597–2604.
9. Ayers DC, Franklin PD, Ring DC: The role of emotional health in functional outcomes after orthopaedic surgery: Extending the biopsychosocial model to orthopaedics: AOA critical issues. J Bone Joint Surg Am 2013;95:e165.
10. Ayers DC, Franklin PD, Trief PM, Ploutz-Snyder R, Freund D: Psychological attributes of preoperative total joint replacement patients: Implications for optimal physical outcome. J Arthroplasty 2004;19(7 suppl 2):125–130.
11. Ayers DC, Franklin PD, Ploutz-Snyder R, Boisvert CB: Total knee replacement outcome and coexisting physical and emotional illness. Clin Orthop Relat Res 2005;440:157–161.
12. Ayers DC, Li W, Harrold L, Allison J, Franklin PD: Preoperative pain and function profiles reflect consistent TKA patient selection among US surgeons. Clin Orthop Relat Res 2015;473(1):76–81.
13. Ayers DC, Li W, Oatis C, Rosal M, Franklin PD: Patient-reported outcomes after total knee replacement vary based on preoperative coexisting disease in the lumbar spine and other nonoperatively treated joints: The need for a musculoskeletal comorbidity index. J Bone Joint Surg Am 2013;95(20):1833–1837.
14. Ayers DC, Fehring TK, Odum SM, Franklin PD: Using joint registry data from FORCE-TJR to improve the accuracy of risk-adjustment prediction models for thirty-day readmission after total hip replacement and total knee replacement. J Bone Joint Surg Am 2015;97:668–671.
15. Rosal MC, Ayers D, Li W, et al. A randomized clinical trial of a peri-operative behavioral intervention to improve physical activity adherence and functional outcomes following total knee replacement. BMC Musculoskelet Disord 2011;12:226.
16. Rolfson O, Rothwell A, Sedrakyan A, et al. Use of patient-reported outcomes in the context of different levels of data. J Bone Joint Surg Am 2011;93(suppl 3):66–71.
17. Lavallee DC, Chenok KE, Love RM, et al. Incorporating patient-reported outcomes into health care to engage patients and enhance care. Health Aff (Millwood) 2016;35:575–582.
Copyright 2016 by the American Academy of Orthopaedic Surgeons.