Pearls: How to Make the Most of PROMs in Everyday Clinical Practice

Patient-reported outcome measures (PROMs) are increasingly being used in orthopaedic research [7], and their value in assessing patient-centered outcomes is widely known [6]. But while many arthroplasty registers collect PROMs [6], it is my impression that most hospital systems do not. Even where PROMs have been successfully integrated into health-information systems on a broad scale, their use within these systems in routine clinical practice is challenging for two main reasons: (1) the lack of an established standard on what change in PROMs should be achieved by joint replacement surgery, and (2) the imbalance between providing high data quality and making the data available instantly for routine procedures [5].

In 2014, my institution launched a PROM program that aimed to collect both preoperative and postoperative (at 2 to 3 months and 1 year) PROM data on all patients who undergo hip or knee arthroplasty at our institution. As of this writing, approximately 80% of our patients have completed both pre and postoperative questionnaires (> 3000 patients annually). We use Oxford Hip and Knee Scores [2, 4] because they are valid, reliable, responsive, and easily administered. The Oxford scores also have well-defined minimal clinically important differences (MCIDs) [2] and patient acceptable symptom states (PASS) [3] levels.

I often hear colleagues ask two questions: (1) How is it possible to collect PROMs as a part of everyday clinical work, and (2) does the PROM data provide the hospital with any useful tools for everyday clinical practice?

Regarding the first question, there are numerous ways to collect PROMs [1, 8]. At my institution, PROM collection is integrated into every patient’s clinical pathway. Before the first appointment at the outpatient clinic, all patients receive a letter that contains the PROM questionnaire and a note emphasizing the importance of PROMs as an assessment tool for our patients’ hip and knee conditions. Our patients are offered an opportunity to complete the PROM questionnaire either on a secure website or traditional paper forms (as it turns out, about 85% prefer paper forms).

The followup process is automated. The hospital’s information-technology system informs our subcontractor about planned PROM followups and then the questionnaires are securely sent to them. At the hospital, specialized physiotherapists run the followup program, and they may contact a symptomatic patient or organize an appointment with an orthopaedic surgeon at the outpatient clinic, if necessary. Followup radiographs are routinely seen by orthopaedic surgeons.

In my opinion, this is a practical and cost-effective way to run the followup program for a large patient volume (currently more than 4000 joint replacements annually). What makes it cost-effective? (1) Our hospital owns the followup software that has been integrated both with the electronical medical records and with the picture archiving and communication system, and (2) the software’s algorithms pick up only the symptomatic patients (based on Oxford scores) for the physiotherapists’ review, while all other patients receive a standardized feedback letter with information on the timing of the next followup. Of course, if an asymptomatic patient has clinically significant findings in followup radiographs (such as severe osteolysis), an appointment with an orthopaedic surgeon will be organized for the patient. Thus, very little manual work is needed to run this PROM and followup program, and the vast majority of the followup can be implemented without actual outpatient clinic visits.

The second question—whether PROM data provide the hospital with any useful tools for everyday clinical practice—has not been answered as convincingly.

In my experience, PROMs provide the hospital with a continuous, patient-centered measure of surgical quality. By looking at the preoperative PROM scores across a practice, we can determine whether any surgeons in the practice tend to operate too early (that is, on patients whose baseline scores hardly can improve because they already are so high). PROMs also provide surgeon-level information on outcomes and on the change in PROMs that have been achieved with joint replacement surgery (Fig, 1A-C), which helps to inform surgeons and practices about the effectiveness of the care they deliver. Finally, PROM data are an excellent tool for performance appraisal. During the performance appraisal, we share the surgeon’s personal complication rates, revision rates, dislocation rates, and rates of prosthetic joint infection—all of which can be regarded as measures of surgical quality. To this, we add the surgeon’s PROM results, as it is rewarding for the surgeon to quantitatively see how his or her own patients are benefiting from joint replacement surgery. In our practice, every surgeon gets to see his or her personal results in comparison with the others each year.

Fig. 1.

The distribution of Oxford Hip Scores among 18 orthopaedic surgeons at Coxa Hospital for Joint Replacement in Tampere, Finland is shown. (A) The Preoperative Oxford Hip Scores (shown in the form of Tukey boxplots) represent patients who underwent THA at our institution during 2016-2017. (B) The Oxford Hip Scores of the same patients at 12 months after THA is shown. (C) The change in Oxford Hip Scores in the same patient cohort is shown. The red line refers to the MCID defined for Oxford Hip Score [2]. MCID = minimal clinically important difference.

But as I alluded to earlier, some unsolved problems remain in terms of the practical application of PROMs as tools that can guide everyday practice. Many surgeons quite reasonably fear that a healthcare system or department will take too simple a look at things. For example, imagine that 3% of Surgeon A’s THA patients have dislocated, while that figure for surgeon B is only 1%. It is too easy for a hospital administrator to conclude that surgeon A is better (and one can imagine a similar analysis for Oxford hip scores). Of course, this may not be right; an equally plausible explanation for the observed differences is that Surgeon A may has a more-complex practice. Without a fairly in-depth, granular analysis of practice demographics (age, gender, BMI, for example) and surgical complexity, such comparisons should not be made at all, as they could be easily misinterpreted as favoring surgeons with more-straightforward, less-complex surgical practices.

Even so, seeing year-over-year trends in specific quality indicators (such as readmissions, reoperations, and major complications like dislocations) as well as PROM results is also important for surgeons, as it helps them to continuously assess quality—and perhaps get an early warning about problems as they arise—over time.

When traditional quality indicators are combined with surgeon-level PROM results, surgeons are provided with enough data to form achieve what we might call “performance appraisal 2.0”.