Developing Content Weights for the 2026 Sports Medicine Blueprint

Survey

The survey was comprised of the 231 clinical activities grouped in the clinical content domains as specified above. For each activity, respondents were instructed to provide two ratings. The first was to rate how often they expect to perform the activity using a 5-point Likert scale (Daily, Weekly, Monthly, Few Times Per Year, Rarely or Never).¹ The second was to rate the general level of risk to the patient if the condition is misdiagnosed or not managed properly, using a 4-point Likert scale (Minimal, Moderate, Considerable, Extreme).¹

Sampling Frame

A sampling frame is a comprehensive list representing the population from which a sample is randomly selected. In this study, the sampling frame included all currently certified Sports Medicine Diplomates from five ABMS member boards: ABFM, ABEM, ABIM, ABPMR, and ABP. Physicians were excluded if they were initially certified in 2024 or later, scheduled to lose certification by December 31, 2024, clinically inactive (ABFM only), declined email contact (ABFM only), or lacked a valid email address (ABFM only). After applying these criteria, the final sampling frame consisted of 4,791 eligible physicians.

Sample

From this sampling frame, a random sample of 800 diplomates was selected and invited to participate in the survey. Prior to distributing invitations, demographic characteristics of the selected sample—including race/ethnicity, certifying board, gender, medical school origin (US or International), medical degree (MD, DO, or other), and initial certification year—were compared against the full sampling frame to ensure representativeness.

Survey Administration

Invitations to participate in the survey were emailed to the selected diplomates on January 22, 2025. The survey was estimated to require approximately 30 to 45 minutes to complete. Incentives varied by certifying board: ABFM diplomates received 5 certification points toward their certification along with a $200 honorarium; ABEM diplomates received a $200 stipend; no incentives were offered to ABPMR or ABP diplomates. A total of six reminder emails were sent between January 22 and February 24, 2025, to diplomates who had not yet responded. Additionally, the survey deadline was extended three times to improve response rates.

Institutional Review Board (IRB)

The study procedures were reviewed by senior ABFM executive staff to verify compliance with ABFM privacy policies. Additionally, the American Academy of Family Physicians Institutional Review Board determined that the study met criteria for IRB exemption.

Rater Quality Control

To ensure data quality, raters whose responses substantially deviated from model expectations were identified using an outfit mean square statistic (≥ 2.75) on the risk-of-harm ratings. Both the frequency and risk-of-harm datasets were reanalyzed after excluding these misfitting raters.

Adjusting the Weights for Each Clinical Activity

The baseline percentage of questions recommended for each of the four clinical content Blueprint domains was initially determined by counting the number of clinical activities within each domain, with equal weight assigned to each activity. To refine these weights, the survey collected ratings on both the frequency of performing each clinical activity and the associated risk of patient harm from a random sample of Diplomates. The sports medicine advisory committee voted on the weights. The expert opinion on the advisory committee had a consensus that equal weights should be assigned to both the frequency of performing each clinical activity and the associated risk of patient harm with each clinical activity. The conclusion was that neither frequency nor risk of harm could be judged to be more important than the other.

From these data, three additional sets of clinical activity weights were created:

1. Frequency-based weights (FI): Frequency ratings from the survey were analyzed using the Rasch model, producing a frequency scale. Since the Rasch-derived scale had an arbitrary zero origin, a constant was added to ensure all resulting weights were positive. These adjusted values were normalized (divided by their sum and multiplied by 100) so that the total FI weights equaled 100.

2. Risk-of-harm-based weights (IoH): The risk-of-harm ratings were similarly analyzed with the Rasch model to create a risk scale. Like the FI weights, a constant was added to shift the scale positively, and the resulting values were normalized, again summing to 100.

3. Combined FI and IoH weights: A third set of weights was calculated by averaging the normalized FI and IoH weights, equally balancing (50%) considerations of frequency and risk of harm.

These three empirically derived sets of weights were subsequently adjusted to fit within a 95% weighting framework, accommodating the predefined 5% allocated to the Foundation of Practice domain. This adjusted framework provided alternative approaches to baseline weighting, further enhancing alignment of exam content with actual clinical practice patterns and patient safety considerations.

Response Rate

After 34 days, the survey window closed (Figure 1). Among the 800 invited participants, 12 had invalid email addresses, and 2 became ineligible as they lost their certification during the survey window, leaving 786 eligible invitees. Of these, 337 completed the survey. However, a quality control check using misfit analysis identified five respondents whose answers suggested careless responding by misfit statistics. After excluding these cases, the final adjusted response rate was 42.2% (Table 1).

• Download figure
• Open in new tab

Figure 1. Email Open & Click Rate by Day for the Sports Medicine Blueprint Survey.

Rater Quality Control

A quality control check was performed on the risk-of-harm ratings by comparing each rater’s responses to the overall risk hierarchy, which was stable across most physicians. Because this hierarchy is broadly shared regardless of individual scope of practice, substantial deviations from it suggest inattentive or invalid responses. We defined misfit as an outfit mean square statistic ≥ 2.75, a threshold for detecting extreme outliers in Rasch-based analyses used in developing Sports Medicine Certification Examination (SMCE) blueprint. Applying this criterion led to the exclusion of five respondents.

Although a parallel analysis for frequency ratings was considered, variability in clinical scope makes it difficult to distinguish genuine differences in frequency from careless responses. Consequently, the same five misfitting raters identified through the risk-of-harm analysis were also excluded from the entire dataset, thus reducing the final response rate.

Representativeness of the Responding Raters

A key consideration in practice analysis/job analysis is whether respondents adequately represent the overall sampling frame. To address this concern, we conducted χ² tests of independence comparing survey respondents to the sampling frame across Race, Ethnicity, Gender, IMG Status, and Degree Type. None of these demographic comparisons reached statistical significance, indicating that survey respondents closely matched the sampling frame on these variables. The only statistically significant difference occurred for the certifying board, reflecting the differential incentives described in the Methods section (Table 2). Consequently, the 322 raters who provided complete and valid responses appear broadly representative of the sampling frame, suggesting minimal nonresponse bias for the evaluated demographic variables.

Overview

This manuscript outlines the methodology used by the ABFM to refine content domain classifications and establish exam question distributions that accurately reflect the current scope of sports medicine practice nationwide. The survey achieved a 42.2% response rate, exceeding the typical practice analysis response rate among physicians (less than 30%), enhancing the robustness of the findings.⁶ A good response rate is necessary to reduce the risk of response bias.⁷

Importantly, respondents were demographically representative of the overall sampling frame across key variables, with the only statistically significant difference observed in certifying board distribution. This high level of representativeness minimizes concerns about nonresponse bias and supports the generalizability of the proposed content domain allocations. To ensure validity and accuracy in determining content domain proportions, the Rasch model was employed.² This methodology eliminates individual rater biases while integrating both the frequency with which clinical activities are performed and their potential risk of harm. By applying this empirically grounded approach, the final domain percentages reflect an objective, data-driven framework for exam content allocation.

Limitations

Three key limitations should be considered when interpreting the findings of this study. First, the Foundations of Care content domain lacks empirical validation for its assigned percentage, as it is primarily knowledge-based rather than grounded in observable clinical activities. Unlike other domains, which were weighted based on data-driven estimates of clinical activity frequency and risk of harm, the Foundations of Care domain could not be empirically assessed in the same manner. As a result, its allocation of 5% of the examination was a policy decision made by the SPMED Advisory Committee, balancing the inclusion of essential foundational topics with the broader objective of basing the exam structure on empirical data. Second, while the survey sample was generally representative across key demographic variables, participation rates varied by certifying board, likely due to differences in incentive structures. This discrepancy may have influenced response rates and was an inherent limitation of the study, as governance constraints prevent the ability to offer uniform incentives across all certifying boards. Despite this, the overall representativeness of the sample on other demographic variables supports the validity of the findings. The third limitation of the study is that it has yet to be externally validated in actual practice; however, it will be challenging due to the difficulty of controlling other cofounding variables related to patient outcomes.

Future work

The practice of Sports Medicine is continuously evolving, particularly in times of transformative change. Advances in technology, such as artificial intelligence and telemedicine, are likely to shape the field by enhancing diagnostic capabilities, expanding access to care, and streamlining clinical workflows. Additionally, policy changes, including regulations governing telemedicine, modifications to medical reimbursement policies, and evolving restrictions on scope of practice, will further influence the delivery of care. As a result, practice analysis activities must be responsive to these ongoing developments to ensure that assessments and certification standards remain aligned with the evolving landscape of Sports Medicine.