Developing a Framework for Ankle Function: A Delphi Study

We used an observational, nonexperimental Delphi design. The Delphi method is a technique in which expert opinion is gathered to reach a consensus on a topic through multiple rounds of individual interviews and surveys.¹² 

Although no published guidelines specifically define an expert or the ideal number of experts who should participate,¹³  general guidelines state that the panel of experts should be heterogeneous to allow for a variety of opinions.¹⁴  Therefore, for the scope of this project, we determined that a panel of 16 experts was sufficient because it could include experts from several fields. Our content experts had experience and expertise in at least 1 of the following fields: ankle function, ankle-injury care, general outcomes assessment or outcomes research, biomechanics, disablement models, functional movement assessment, or general sports medicine. We identified potential experts through a record of scholarship (>5 peer-reviewed publications in a content area), professional recommendations, and clinical experience (>5 years in their field). Once a person was identified as an expert, we e-mailed him or her an invitation to participate; those who agreed received an electronic consent document and the interview questions. Recruitment ended once we had a panel of 16 (Table 1), which included experts from athletic training (n = 8; 50%), physical therapy (n = 2; 12%), athletic training and physical therapy (dually credentialed, n = 4; 25%), general medicine (physician, n = 1; 6%), and strength and conditioning (n = 1; 6%). Experts were currently practicing (n = 4; 25%), serving in an academic position (n = 9; 56%), or both (n = 3; 19%). They resided in the United States (n = 14; 88%), Australia (n = 1; 6%), and Ireland (n = 1; 6%).

We collected the initial round (round 1) of data using structured telephone interviews. We recorded the telephone interviews (model WS-210S digital voice recorder; Olympus America, Center Valley, PA). The questions addressed the definition and characteristics of and assessment techniques for a healthy/normal/noninjured ankle, unhealthy/acutely injured ankle, FAI, and CAI. We combined and reduced the experts' responses and returned them in the form of electronic surveys (Survey Monkey, Palo Alto, CA) in subsequent rounds (rounds 2–5).

The following items were addressed:

1. The definition of a healthy/normal/noninjured ankle

2. The functional characteristics of someone with a healthy/normal/noninjured ankle

3. The functional characteristics of someone with an unhealthy/acutely injured ankle

4. The assessment of functional characteristics

5. The definition of FAI

6. The functional characteristics of someone with FAI

7. The assessment of functional characteristics of someone with FAI

8. The appropriateness of the term FAI

9. The definition of CAI

10. The functional characteristics of someone with CAI

11. The assessment of functional characteristics of someone with CAI

12. The appropriateness of the term CAI

After an orientation, we interviewed the experts via telephone. We read their responses back to them and allowed them to review each answer before proceeding to the next question. We then compiled and reduced the responses qualitatively to detect common themes and to categorize the data. We calculated frequencies quantitatively. To minimize bias, all members of the research team assisted in the data reduction and review of the final product. A sample interview question was “How do you define a healthy/normal/noninjured ankle?” All interview responses then appeared in the subsequent electronic survey (round 2).

We presented the reduced data back to the panel in the form of an electronic survey: they could agree or disagree with each of the 228 items. We excluded assessment techniques for the functional characteristics from this round because consensus had not been established yet. An example of a survey item was “Please indicate if you agree or disagree that each of the following items should be included in the definition of a healthy/normal/noninjured ankle.”

When addressing consensus, no criteria are universally accepted. However, 80% agreement is often considered acceptable¹⁵  and has been used in other Delphi studies.¹⁶  After round 2 and all subsequent rounds, we retained items that had reached 80% or greater agreement and removed those that did not.

Based on round 2 feedback, we created a list of items the experts identified and which they agreed should be included in the definition of a healthy/normal/noninjured ankle. From that list, we derived the initial definition of a healthy/normal/noninjured ankle. We sent that definition to the experts via Survey Monkey (round 3) and asked them to agree or disagree. The subsequent rounds (rounds 4 and 5) consisted of a list of compiled, agreed-on items, followed by a brief survey.

In the round 4 survey, we addressed the assessment techniques that were identified in the initial interview.

In the round 5 survey, we addressed the definitions of FAI and CAI. We concluded data collection after round 5 because the panel's responses were no longer moving toward agreement.

In round 1, we asked the experts, “How do you define a healthy/normal/noninjured ankle?” The experts (n = 16; 100%) provided various responses when asked to define a healthy/normal/noninjured ankle during the interview. Most listed components of a healthy/normal/noninjured ankle, rather than providing a definition. To achieve a consensus, we combined and reduced their responses to 24 items (Table 2). Twenty-two items (92%) reached consensus for the definition of a healthy/normal/noninjured ankle. Of those 22 items, we reduced redundant items to 16 (73%). To form the definition, we identified 5 broader, inclusive categories under which each item fit: (1) function, (2) participation, (3) impairments, (4) pathology, and (5) pain. As noted by 3 experts (19%), pain is generally considered an impairment. However, we believed that pain was such an important and unique clinical entity that it should remain an individual component in the definition. Therefore, we established the working definition of a healthy/normal/noninjured ankle as “a person with a healthy/normal/noninjured ankle(s), either through self-report or clinical measurement, presents with full functional capacity and participation status, without pathology, pain, or impairments relative to the ankle(s).” Eighty-eight percent (n = 14) of the experts approved that definition. Therefore, we concluded data collection for the definition of a healthy/normal/noninjured ankle.

In addition to the comments regarding the classification of pain, panelists also emphasized that some components of a healthy/normal/noninjured ankle were not limited to only ankle function. For instance, a patient with a healthy/normal/noninjured ankle may experience limited functional capacity because of another injury. Another expert commented that he or she felt there was a difference between healthy/normal and noninjured and that an abnormality could be present in the absence of injury.

In the interview, we asked each expert to “Describe the functional characteristics of someone with a healthy/normal/noninjured ankle.” We sent the preliminary responses back in a survey, and the experts were asked to agree or disagree (Table 3). Forty-three characteristics reached 80% consensus and were retained in the final list.

In addition, panelists noted that impairments may be required for function, even though the impairments themselves are not functions. For example, 1 expert wrote, “patients can be functional without full range of motion,” or “they can have pain and still be functional.” Another expert suggested adding “as related to the ankle” to each characteristic to clarify the meaning, including the term decreased with some of the characteristics because the reference point for “normal” was not clearly identified and suggested clarifying whether a patient must lack all or just 1 “healthy” characteristic to be considered “unhealthy.”

The third topic we addressed in the initial interview was an unhealthy/acutely injured ankle. We asked the experts to “Describe the functional characteristics or limitations of someone with an unhealthy/acutely injured ankle.” We again collected, combined, and reduced the responses and returned them in a survey. We retained only the items that reached 80% agreement (Table 4).

The fourth topic we addressed in the preliminary interview was assessment techniques. We asked the experts, “How do you assess functional characteristics of the ankle (for unhealthy/acutely injured or healthy/normal/noninjured ankles)?” Of the 87 responses offered by the experts, 33 responses (38%) reached 80% agreement in the subsequent survey; these included items such as “gait analysis,” “figure-8 test,” and “history/subjective report” (Table 5).

The fifth topic we addressed in the interview was the definition of FAI. We asked the experts, “What is your definition of functional ankle instability?” Similar to their responses for the definition of a healthy/normal/noninjured ankle, the experts' tendency was to provide lists of components rather than a concise definition. The initial responses, as well as the subsequent results that reached consensus, are presented in Table 6.

From the components that reached agreement, we established the working definition as “Functional ankle instability is a recurrent sense of giving way of the ankle.” We presented that definition to the experts and 12 of the 15 respondents (80%) agreed with that definition, so we closed data collection and accepted that as the definition of FAI. However, even though the experts reached a consensus for the definition, their specific comments indicated that they had strong opinions about the topic. For example, 1 expert stated, regarding FAI, “if you must use that term,” another wrote “CAI is a bad term” and “FAI is the better term,” and still another would have preferred “not necessarily” rather than be forced to agree or disagree with the definition.

The sixth topic we addressed during the interview focused on the functional characteristics of FAI. We asked the experts, “Describe the functional characteristics or limitations of someone with functional ankle instability; what are the functional costs or consequences?” Forty items were included in the survey and 13 (32%) reached agreement (Table 7). Examples of items that reached consensus included “unable to jump or hop,” “difficulty walking on uneven ground,” and “it depends on what they want to do” (Table 7).

We addressed FAI nomenclature with 3 different interview questions. We asked the experts, “Do you have a different title or name that you use for functional ankle instability?” Based on the lack of consensus from the survey, no other terms were consistently used in place of FAI. We also asked, “Do you feel that there is a name more appropriate than functional ankle instability?” Based on the experts' survey responses, there was no other name more appropriate than FAI.

Additionally, we asked the experts, “Do you feel that functional and chronic ankle instability are the same thing?” Although 9 of the 16 experts (56%) indicated that CAI is an umbrella term and FAI is a component, consensus was not reached on any single response. Additionally, 3 experts (19%) responded no and the 4 remaining experts (25%) responded yes, indicating that they agreed that CAI and FAI were the same.

To address CAI, we asked the experts, “What is your definition of chronic ankle instability?” Similar to the previous definition questions, the experts provided lists of components rather than a single definition (Table 8). We reduced their responses to 14 items and, from the survey, we established the working definition as “Chronic ankle instability is recurrent episodes of ankle instability and sprains.” Only 11 of the 16 respondents (69%) agreed with the definition, so consensus was not achieved, and no definition of CAI was generated.

To address the functional characteristics of CAI, we asked the experts, “What are the functional characteristics of someone with chronic ankle instability; what are the functional costs or consequences?” Although 6 different items were identified through the interview, none were functional. The list included 1 item that could be considered a characteristic (ankle-ligament laxity) and 5 items that were not. The other 5 responses were “I am not sure if there is a performance decrease or cost,” “Same as functional ankle instability,” “Same as functional ankle instability, just a longer period of time,” “It depends on the sport/activity they participate in,” and “Everyone is different.” None of the responses from that interview question reached consensus. The only issue on which the experts did reach consensus was that CAI was different from FAI.

The next interview question addressed CAI nomenclature. We asked the experts, “Do you have a different title or name that you use to describe chronic instability?” Most of the experts (13 of 16; 81%) used the term CAI.

The final interview question presented an alternative to the terms functional or chronic when addressing ankle instability. We asked the following closed-ended question, “Do you feel that ‘persistent ankle instability' is an appropriate substitute for the title or name that you use (eg, functional ankle instability, chronic ankle instability, ankle instability)?” Most of the experts (9 of 16; 56%) indicated that “persistent ankle instability” was not an appropriate substitute, and 4 (25%) indicated that it was appropriate, whereas 3 (19%) indicated that “it all means the same thing, just different words.”

Our purpose was to establish definitions for a healthy/normal/noninjured ankle, FAI, and CAI, as well as to identify the functional characteristics of and assessment techniques for each. Our motivation grew from a need for consistency in the language used to describe ankle injuries and their characteristics as a first step in identifying appropriate measures for assessing outcomes.

To properly measure health care outcomes, the element of health that is of interest must be determined, along with an instrument to measure it.¹¹  Based on the model established by the WHO International Classification of Function,¹¹  the first step is to identify and define health or, in this case, a healthy/normal/noninjured ankle. The WHO defines health as “a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity.”¹⁷  An element of health that is often addressed is level of health, which has been described as function,¹⁸  activity limitation,¹¹  disability,¹⁸  impairment,^11,18  and participation status.¹¹  Models, such as those developed by Nagi¹⁸  and the WHO,¹¹  have been created to provide a common language that can be used by all health care practitioners. Those models provide a broader perspective of the patient's health status to include social well-being, in addition to the physical limitations of the disability.^19,20  The definition of a healthy (normal/noninjured) ankle has not, to our knowledge, been previously established; rather, the literature is focused on injury and instability. Therefore, our first step was to define a healthy ankle. Our definition of a healthy ankle included full functional capacity and participation status. That seems to coincide with the operational definition of an acute lateral ankle sprain presented by Delahunt and colleagues,⁹  which includes disability and deficits in function. This reinforces the concept that the definition of an unhealthy ankle, or in this case, the Delahunt et al⁹  definition of a lateral ankle sprain, should contain language similar to that in the definition of a “healthy” ankle.

Although the experts reached a consensus for the definition of a healthy ankle in 3 Delphi rounds, several factors became apparent. Multiple experts emphasized that some components with which they agreed were not limited to only ankle function and that function could be affected for reasons unrelated to the ankle. Another cautioned that healthy/normal and noninjured were not necessarily synonymous, and there could, in fact, be an abnormality in the absence of injury.

In addition to debating several components that ultimately reached consensus, several experts raised concerns about the inclusion of pain. They emphasized that pain is an impairment and suggested that it not be listed separately within the definitions. However, we decided to leave pain within our definition of a healthy/normal/noninjured ankle for several reasons. First, because no pain was agreed on by 15 (94%) of the experts, we felt it was important enough to include in the definition, even though pain is an impairment. Second, 14 of 16 experts (88%) agreed with the definition, which included both pain and impairment. Third, the WHO International Classification of Functioning, Disability and Health language is not yet universal to all health care providers. Those who are unfamiliar with this language may not appreciate that pain is universally included in the definition of the term impairment. Although this definition may evolve with additional research, it represents progress toward establishing a common definition.

Once the definition of a healthy ankle was established, the experts were asked to define FAI. We expected the panel to reach a consensus for FAI, and they did. Of the 32 components that were initially offered by the experts, however, only “a recurrent sense of giving way of the ankle” reached agreement. Therefore, our definition includes only “recurrent sense of giving way of the ankle.” Our definition is very similar to that of Freeman (1965)²¹  and the operational definition of Delahunt et al (2010),⁹  which included “giving way,” and “frequent episodes of giving way and feelings of ankle joint instability,” respectively. Absent from these definitions are terms related to function or functional capacity. Although the definition of a healthy/normal/noninjured ankle included items that address function, the definition of FAI did not. It would then seem that the diagnostic criteria for FAI should include only 1 question specific to recurrent episodes of giving way. Yet the inclusion criteria for FAI research seem to include multiple impairments¹⁰  that are often not associated with a self-reported sense of giving way. Those impairments include anatomic differences, decreased inversion strength, decreased balance, altered foot mechanics during gait, and jump-landing deficits,¹⁰  all of which represent clinician-measured impairments.

Other authors have described functional instability as “the disabling loss of reliable static and dynamic support of a joint.”^22(p692) Hertel^6(p364) defined lateral ankle instability as “the existence of an unstable ankle due to lateral ligamentous damage caused by excessive supination or inversion of the rearfoot” and CAI as “the occurrence of repetitive bouts of lateral ankle instability, resulting in numerous ankle sprains.”^6(p363) In his own research, Hertel⁶  did not use the term functional ankle instability but instead supported a paradigm in which mechanical and functional insufficiencies were subcomponents of CAI. Tropp⁷  also supported the concept that functional and mechanical instabilities are components of CAI but described FAI as “the subjective feeling of ankle instability or recurrent, symptomatic ankle sprains (or both) due to proprioceptive and neuromuscular deficits.”⁷⁽⁵¹²⁾

Although we successfully defined healthy/normal/noninjured ankle and FAI, we were not able to define CAI. In 2010, through a systematic review, Delahunt et al^9(p2117) provided an operational definition of CAI as “an encompassing term” and included “mechanical and functional instability” as well as “residual symptoms of giving way and ankle joint instability present for a minimum of 1 year post-initial sprain.” During our interviews, the experts offered their own definitions and listed components of chronic instability to include in the definition, as they did for the healthy ankle and FAI definitions. However, unlike for the previous concepts, the experts could not agree on a definition of CAI. Some indicated that they felt that FAI and CAI had the same meaning. Others suggested that CAI was a broader term under which FAI was a component, a concept that was supported by Delahunt et al,⁹  Hertel,⁶  and Tropp.⁷  Still, some insisted that FAI was the appropriate title, whereas others offered the same support for CAI. One expert's comment about about FAI was “if you must use that term.” Yet another stated that CAI “is a bad term” and FAI “is the better term.” Based on the brevity of our definition of FAI, the lack of agreement on its components, and our failure to define CAI, it is apparent that the meanings of FAI and CAI are not clear. Yet “the sense of giving way,” which is exclusively a patient-reported impairment, seemed to be the 1 identifier that was consistently agreed upon in the definition and description of FAI.

Once the definitions were established, we then identified the characteristics of a healthy ankle, FAI, CAI, and an unhealthy ankle. Although the experts agreed on 11 of 27 characteristics (41%) of a healthy ankle, several of the 11 characteristics did not represent functions but rather the absence of impairments (ie, range of motion, strength, pain, balance). Although some experts noted and commented on those discrepancies, most did not. These overlapping concepts reinforce our opinion that these terms and language are not yet universal. However, as 1 expert stated, the resolution of impairments may be necessary to achieve full functional capacity, even though the impairments themselves are not truly considered measures of function. Others indicated that “patients can be functional without full range of motion,” and “they can have pain and still be functional.” Those comments highlight the divergent opinions specific to ankle health and injury and reinforce the need for clarity.

An additional trend appeared as we attempted to reach consensus for the final list of healthy ankle characteristics. Many of the impairments that were initially identified during the interviews did not reach consensus. We believe this reinforces 2 points. First, the language common in health care outcomes is not universally understood or applied. Second, our Delphi design and the selection of a diverse panel of experts, for the most part, was successful. Specifically, the inclusion of panelists with backgrounds in general outcomes assessment and outcomes research who did not specialize in ankle-injury care or ankle-injury research, was appropriate and necessary. Those panelists consistently reinforced the use of outcomes terminology, such as function and impairment.

In addition to the classification of the characteristics, the relative specificity of the characteristics of the ankle was questioned. Several panelists emphasized that the functional characteristics were not necessarily exclusive to the ankle. Normal gait, for example, can be affected by any problem linked to the lower extremity functional chain, which raises the larger issue of whether it is appropriate or necessary to have outcomes instruments specific to injuries, regions, or articulations. Because those characteristics are ultimately used to address ankle-injury outcomes, it could be appropriate to add “as related to the ankle” before each item, as was recommended by 1 panelist.

The panel also identified patient values as a key component in addressing function. In contrast to some characteristics of healthy ankles, a comment such as “It depends on what they want to do; it is based on [a] patient's perspectives and goals” is not specific to the ankle or any other injury. It does, however, emphasize each patient's unique values and priorities because it highlights the personal effect that injury can have on function. Considering this aspect is vital when addressing the effects of injury on function and is paramount when monitoring recovery and outcomes.^19,20  We also believe that including this important characteristic reflects the diversity among the experts' perspectives and backgrounds.

Next, we addressed the characteristics of an unhealthy (acutely injured) ankle. One difference between the healthy and unhealthy ankle is that the characteristics of the latter included not only more items but also more impairments. Although impairment does not carry the same meaning as function, including impairments within the list of functional characteristics reinforces the lack of clarity regarding these terms. Several impairments were included in the list of functional characteristics of an unhealthy ankle, and again, many more overall characteristics were included in the final list. “Depends on the patient's desired or normal level of activity” was also on both the unhealthy and healthy lists, which emphasizes the importance of considering the patient's values and goals when addressing the functional cost of an injury. Other comments and concerns pertained to the term decreased. Similar to the list of characteristics of a healthy ankle, the reference point for “normal” was not clearly identified. In addition, it is not clear whether a patient must have all characteristics or just a single characteristic on the list to be considered “unhealthy.” As with the healthy ankle, characteristics such as “decreased confidence” are not specific to the ankle, which again suggests the need to add “as related to the ankle” to each characteristic.

When the experts identified the functional characteristics of FAI, several items on the final list also appeared on the final list of unhealthy characteristics. Because both an unhealthy ankle and FAI are deviations from “healthy,” similarities should be expected. Similar items included “unable to complete sport-specific activities,” “difficulty walking on uneven ground/engaging unpredictable surfaces,” and “decreased confidence.” As with the first 2 lists, “it depends on what they want to do” was included as well. There were, however, a few differences: the most notable were that “repetitive episodes of giving way” and “there is a continuum” appeared on the FAI list but not the unhealthy list.

Another interesting comparison is with previously published FAI inclusion criteria.^23–30  Our final list showed some similarities yet was not an exact match to any of them. Another item that was included in the final list but not in any previous publications or in our “unhealthy” list, was “there is a continuum.” This seems to represent the comments made by 1 expert who suggested that FAI “is a syndrome” and “not all patients are heterogeneous.”

In contrast to FAI, none of the functional characteristics of CAI reached consensus. Some of the experts stated that the functional characteristics of CAI were the same as for FAI, whereas others stated that the condition was the same but over a longer period. Interestingly, 3 experts (19%) chose “I am not sure if there is a performance decrease or cost” as related to the functional characteristics of CAI. That implies persistent misconceptions and misunderstandings about CAI. Because the definition of CAI did not reach consensus, it was not surprising that the functional characteristics did not reach consensus.

Although the final lists of techniques for assessing the characteristics of a healthy or unhealthy ankle were not a perfect match, the assessments did generally align with the characteristics of a healthy or unhealthy ankle. When the experts were asked to agree on assessment techniques, pain again was the focus of several discussion points. One expert suggested that “can perform without pain” should be added to the assessments. Another noted that it is possible to function with pain. Perhaps those comments and the debate about pain identify an opportunity to improve our methods of assessing function. The existence or absence of pain could complement existing functional rating assessments and instruments. Regardless, when addressing function, the presence of pain certainly warrants additional attention.

In addition to pain, the importance of self-reported functional assessments was again emphasized. One expert commented, “What I really like about this is that it seems as though laboratory-oriented, clinician-oriented, and patient-oriented measures have emerged as being relevant for assessment. This is a great sign that we're looking for context among the patient, clinician, and researcher!” Similarly, Wiklund³¹  stated that patient-reported outcomes, linked with clinician-reported outcomes, add important information that cannot be collected by using clinician-reported techniques alone. Patient-reported outcomes also facilitate communication between the patient and clinician and may reveal patient concerns or conditions that might have gone unnoticed.³²  Furthermore, discrepancies often appear between the clinician's and the patient's impressions of readiness to return to activity.³³  Laboratory-oriented, clinician-oriented, and patient-oriented assessments did, indeed, emerge throughout the experts' responses to the various questions in our results.

The final question we posed to the panel addressed other terms that they use in addition to or in place of FAI or CAI. Our goal was to achieve clarity and allow the experts to express their opinions. We asked if they felt persistent ankle instability was an appropriate substitute for FAI or CAI or both. The experts did not agree on any term as an alternate to FAI or CAI, but some terms emerged from the comments. They included recurrent ankle instability, recurrent ankle dysfunction, and dysfunction. Some respondents emphasized that the terms FAI and CAI should be changed to something different, whereas others indicated that FAI and CAI were interchangeable terms. One expert commented that that no single term will encompass all patients and that clinical prediction rules should be explored. Ultimately, however, those results indicate that inconsistencies and contradictions about the concept of ankle instability persist, whether it is classified as chronic or functional.

We established a working definition of a healthy ankle, along with the functional characteristics and assessment techniques of healthy and unhealthy ankles. Consensus was also reached regarding the definition of FAI; however, many disagreements remain, as described in the experts' comments. The definition of CAI did not reach 80% consensus, but most experts did agree on the definition. Although the experts voiced mixed opinions regarding the definitions of both FAI and CAI, these results offer a good starting point for the working definitions.

We established the definition of a healthy ankle and identified 11 characteristics. Experts agreed on 16 components of the definition, which we condensed into 5 categories, including pain. These are important steps in creating a unified outcomes model for advancing the care of patients with ankle injuries. Future researchers should build upon the definition, components, and assessment techniques of a healthy ankle and compare the experts' opinions with those of patients with ankle injuries and with existing ankle-outcomes instruments. Furthermore, the effect of pain during function should be explored as a factor in the assessment of function.

In addition, although we were able to establish a definition for FAI, the experts could agree on only 1 of the 32 components (3%) of FAI that they initially identified. Therefore, the definition includes only “recurrent sense of giving way of the ankle.” That definition is surprisingly similar to the original Freeman²¹  definition of 47 years ago. It could then be argued that the only criterion necessary to classify a patient with FAI is the presence of the sense of giving way. This suggestion highlights the complexity of FAI, but it does not clarify how the term should be identified in practice or research. Future investigators should continue to address the nature of FAI specific to its identification and effect on function.