CORR Insights®: What Are the Minimum Clinically Important Differences in SF-36 Scores in Patients with Orthopaedic Oncologic Conditions?

Where Are We Now?

While most patients with bone and soft-tissue sarcomas will opt for the surgical treatment that maximizes the likelihood of survival and minimizes the risk of local recurrence—en bloc resection with adequate margins, when possible—this choice sometimes can be surprisingly complicated. Sometimes, survival comes at a severe trade-off against health-related quality of life (HRQoL), and vice versa. Two examples of this might be opting for a high amputation over limb salvage in certain pelvic and acetabular malignancies in order to get an adequate margin and avoid a tenuous, inconsistent reconstructive approach (thus choosing length of life over HRQoL), or choosing not to have any surgery for sacral chordoma with sacral plexus involvement (and in so doing, prioritizing HRQoL over length of life).

Treatment goals are different for patients with metastases than for patients with primary bone and soft-tissue sarcomas. The aim of treatment for a patient with a metastatic lesion is to improve HRQoL for the remaining—sometimes short—lifespan, while minimizing the risk of complications.

Regardless of whether one is treating primary or metastatic tumors, measuring patient-reported outcomes is valuable to understand the change in level of functioning or HRQoL [1, 10]. And when one is measuring patient-reported outcomes, one needs to know how small an improvement would be considered large enough for a patient to care about; after all, patients perceive effect sizes, not p values. This concept, called the minimum clinically important difference (MCID), is foundational when interpreting outcomes scores, like the SF-36, following surgery [7].

The current study by Ogura and colleagues [8] is, to my knowledge, the first to generate such a reference using the SF-36 for the orthopaedic oncology patient population, and as such, it is important [6]. As honestly recognized by the authors, heterogeneity due to inclusion of a wide range of tumors in the cohort is indeed a limitation. However, this is largely resolved by separately reporting primary tumors and metastatic disease. Ogura and colleagues [8] determined MCID values ranging from 2 to 6 out of a maximum score of 100 representing best health—depending on the calculation method and patient subgroup—for the SF-36 physical component summary (PCS) and mental component summary subscales. Consider, for example, a patient with a painful bone metastasis in the femoral neck; an improvement of more than 2 points out of a maximum score of 100 on the SF-36 PCS subscale, whether by surgery, radiation, or improved pain management, is likely to be considered by that patient to be a clinically important change.

Where Do We Need To Go?

Validation of the findings by Ogura and colleagues [8] in more-homogeneous populations of patients with musculoskeletal tumors would be of interest. In addition, questionnaires focusing on physical function, such as the Toronto Extremity Salvage Score and the PROMIS Physical Function instrument should be included for calculation of MCIDs [4]. Anchor-based MCIDs should be emphasized over distribution-based MCIDs [7]. The latter uses sample distribution statistics such as half the standard deviation for calculation, but ignores the patients’ perspective [2, 9].

Future studies that assess change in HRQoL, regardless of treatment, using the SF-36 instrument in patients with orthopaedic oncologic conditions should present the proportion of patients that achieved the MCID. Proportions of patients that achieved the MCID can be compared between subgroups or among treatments. These statistics are valuable to inform our patients when discussing treatment options [1].

Shared decision-making requires data about prognosis, risk of complications, durability of a reconstruction, expected function, and HRQoL. Such data not only better inform our patients, but also require a thorough understanding of our patients’ values and preferences [3]. These values and preferences likely diverge for patients with primary tumors versus metastatic disease. In addition, patients with musculoskeletal tumors and their physicians might weigh the relative importance of various outcomes differently [3]. We should explore these values and preferences.

How Do We Get There?

Future research should include homogenous samples like lower-extremity bone metastases, and assess multiple QoL and physical function questionnaires to calculate corresponding MCIDs, focusing on anchor-based MCIDs. Smaller sample sizes may be adequate, and short followup duration (3 months) may be sufficient in for example patients undergoing treatment for bone metastases [11].

Other studies should focus on identifying the subgroups that achieve the MCID in QoL from surgery. This is particularly relevant in the population of patients with bone metastases and would aid preoperative patient selection and counseling.

Considering the variety of treatment options and corresponding morbidity, prognosis, and QoL, choosing treatment can be difficult for both patients and physicians. We should explore patients’ and physicians’ preferences for treatment and decision-making. The Ottawa Decision Support Framework helps improve and evaluate the decision-making process [5]. The decision conflict scale is a useful tool to explore uncertainty in choosing treatment options and assess modifiable factors that contribute to this uncertainty [3, 5]. A survey including patients and orthopaedic oncologists could provide insight in patients values and preferences for treatment [3]. Results should be stratified based on tumor type (primary sarcoma versus metastatic disease), and perhaps by anatomical location (proximal femoral metastases).