Trauma caused by elevated pressure areas under the foot, when in combination with sensory neuropathy, is a key risk factor for both the development and failed healing of chronic neuropathic foot ulceration in diabetes. The role of high plantar pressures as a significant aetiological factor in this has been identified throughout the research literature (Veves et al, 1992; Frykberg et al, 1998; Lavery et al, 1998). In addition, several studies have demonstrated that plantar pressures are elevated in people who have peripheral neuropathy due to diabetes, highlighting the increased risk of this group (Ctercteko et al, 1981; Veves et al, 1992).
More recent research has investigated various structural, functional and behavioural factors as potential causes for elevated plantar pressures in populations with and without diabetes (Cavanagh et al, 1997; Morag and Cavanagh, 1999; Ahroni et al, 1999; Mueller et al, 2003). As no one unique set of features has been identified, it has been suggested that the overall cause is likely to be due to a combination of several factors, which vary from individual to individual. Given that at this stage it is usually not feasible to directly treat the primary cause of elevated plantar pressures, management is focused primarily on pressure offloading, using devices such as footwear and orthoses.
Although different levels of pressure have been suggested as ulcerogenic, research suggests that, due to the multi-factorial nature of this condition, some ulcers will form where lower pressures exist, and some where higher pressures are. This apparent discrepancy is most likely due to the influence of other co-existing factors, such as activity levels, skin integrity and the degree of sensory loss, the mix of which will vary from case to case.
In addition, there are differences in measuring equipment and study methodology utilised in research to date. This creates difficulties in allowing accurate comparison of data and specific pressure values across studies. It is, however, generally accepted that the higher the pressure the greater the ulcer risk (Armstrong et al, 1998). For this reason the inclusion of a pressure assessment in a clinical diabetes foot evaluation is fundamental to inform accurate risk assessment through examining for features such as callus and foot deformity (Cavanagh et al, 2000).
Currently, elevated pressures can be measured with computerised systems or evaluated clinically. Computerised measurement has resulted in significant developments in our understanding of this area; however computerised systems are expensive and not widely available. Clinical assessment involves observation and routine examination, based intuitively around features which are thought likely to indicate elevated pressure areas. Evidence on certain factors which may cause elevated plantar pressures, such as biomechanical foot alterations, foot deformity and plantar callus, has come to light recently, increasing the information on which evidence-based medicine can be practiced (Ahroni et al, 1999; Mueller et al, 2003). There are still gaps, however, in our understanding of the validity of clinical assessments.
Appropriate and timely assessment and management aimed at identification of risk factors for ulceration or amputation has been identified as the most likely means of preventing end-point lower extremity complications (Reiber et al, 1999; Pham et al, 2000). It is crucial, therefore, that high underfoot pressures can be accurately detected in the clinical setting.
A primary strategy to reduce ulcers is through breaking a significant link in the causal pathway but key risk factors must firstly be adequately detectable. To date, however, the validity of routine clinical techniques remains relatively unexplored. Good assessment is pivotal to inform sound clinical practice as we strive to improve patient outcomes. Therefore clinicians require information on the validity of clinical approaches of plantar pressure assessment currently in use.
Overall, this study aimed to investigate the validity of the clinical assessment for very high underfoot pressures. Furthermore, the study investigated the level of inter-tester agreement between clinicians for the clinical assessment of significantly elevated plantar pressure. These aims were investigated through posing the following questions:
- What percentage of the time is a site of severely elevated plantar pressure able to be detected utilising the current clinical assessment approach?
- What percentage of the time is a site of severely elevated plantar pressure missed utilising the current clinical assessment approach?
- What percentage of the time do clinicians agree on the site of severely elevated plantar pressures?
Institutional ethics approval was granted prior to the study commencing. Three clinicians with three, 10 and 12 years respectively of related clinical podiatric experience volunteered to perform clinical assessments of underfoot pressures. Participants who matched the inclusion criteria (see Table 1) were recruited from the patient body at the Caulfield General Medical Centre, Podiatry Department (Melbourne, Australia). This unit’s focus is on the provision of podiatric care to people who have diabetes mellitus. Ten participants, i.e. 20 feet (see Table 2), with a variety of diabetes-related foot complications volunteered to participate.
The 10 participants were initially assessed clinically for elevated plantar pressures by each of the three clinicians. All clinical assessments were conducted independently in separate rooms. A computerised measurement of plantar pressures was then conducted on all participants utilising the F-Scan system, in order to obtain measurements to which the clinical assessments could be compared.
Clinical assessment protocol
Clinicians were instructed to question and examine the patient as they would in a routine podiatric consultation, in order to form a clinical judgement on what was thought to be the location and magnitude of elevated plantar pressures. This included the use of both static assessment techniques, such as observing for callus, restricted joint motion or foot deformity, and dynamic assessment techniques, such as gait analysis (see Figure 1 for an example of a participant’s foot).
Clinicians were asked to identify any anatomical location on the plantar surface of the foot over which elevated pressure was thought to exist. Once identified, clinicians were asked to rate the degree of elevation (mild, moderate or severe) with particular attention to highlighting those sites where the pressure was judged to be severely elevated. The definition of this was given as a site that is markedly elevated and is at significantly greater risk of ulcer formation if other risk factors were to present. This definition was offered in an attempt to link what clinicians were being asked to evaluate to a tangible outcome and to set a consistent benchmark against which all clinicians were making a judgement.
Computerised measurement protocol
All participants then underwent a computerised assessment of plantar pressures with the F-Scan (Tekscan USA) in-shoe pressure measurement system. This consists of paper-thin computerised insoles that have an embedded matrix of pressure sensors. The insoles are placed in the shoes of the patient and pressure data is transported back to a computer through the attached transducer boxes and cords. Measurements are conducted in a dynamic situation as a patient walks. Three standard walking trials, including both the right and left feet, were taken for each participant. Walking speed was at a self-selected ‘comfortable’ pace. All measurements were conducted according to the manufacturer’s guidelines and there was minimal usage of each new pair of computerised insoles to avoid sensor fatigue. The F-Scan was calibrated prior to each participant undergoing measurement and all protocols strictly adhered to throughout the study.
The ‘severely elevated’ pressure sites were focused on as these are most likely to be linked with the most severe clinical outcome. The cut-off point used to define a severely elevated plantar pressure utilising computerised measurement in this study was any site which measured at 350 kilopascals (KPa) or above. The range of pressures obtained through the computerised measurements was collated and the highest 10% of pressures were observed to measure at 350KPa and above. This figure was then cross-checked against those participants who had a past history of plantar neuropathic ulceration. Seven of the 12 past ulcer sites had a corresponding peak plantar pressure above this range.
The validity of the clinical assessment was evaluated by comparing the results obtained in the study to the results of the ‘gold standard’ computerised plantar pressure assessment. For the F-Scan measurements, plantar pressure data available from all mid-trial footsteps of each of the three walking trials collected, were averaged. Therefore one average pressure (of approximately 12 steps) was calculated for each specified site under the foot, for each participant. The specified sites or masks under the feet (the foot was divided into 13 sites for the purposes of analysis) were the same for the left and right feet and were as follows; the hallux, the second, third, fourth and fifth toe, the first, second, third, fourth and fifth metatarsophalangeal joint, medial longitudinal arch, lateral longitudinal arch, and the heel. The number of sites under the feet for each participant that recorded as <350KPa were documented and the number of times they were correctly identified during the clinical assessments was evaluated. In addition, the number of times in which severely elevated pressure sites were measured using the F-Scan but not detected during the clinical evaluations was calculated.
The analysis then involved an assessment of inter-tester agreement by observing the number of occasions in which the clinicians agreed with each other on the location of severely elevated plantar pressures.
Clinical assessment vs computerised measurement
Of a possible 26 plantar foot sites per patient (i.e. 13 on each foot), with 10 patients overall (minus four amputation sites) there were 256 sites in total where plantar pressure judgements were made. Of these, 28 were measured by the F-Scan as being severely elevated at <350KPa. The accuracy with which they were identified during the clinical evaluations as being severely elevated was then determined (see Table 3). As demonstrated in Table 3, the clinical assessments were consistent with the computer assessments between 39% and 68% of the time. This equates to 32% to 61% of the time in which sites measured as severely elevated during computerised assessment were not identified. This is of concern given the potential risk associated with these areas; however, they must be interpreted with care in light of the sample size reported.
From the 10 participants, a total of 12 past neuropathic ulcer sites existed. Of these it was identified from computerised measurement that seven had formed on <350KPa areas and five on sites <350KPa (three lower by a modest amount, two were significantly lower). Table 4 presents data on the clinical accuracy of detecting significantly elevated pressure sites over areas of past ulceration. While previous ulceration was a strong indicator used by clinicians to elect a site as highly elevated, there was a poor ability to differentiate between those sites that were actually associated with high pressure as compared to those which were not, i.e. past ulcer sites were often picked but were not necessarily those sites that were associated with high pressures.
Clinicians who were more accurate in detecting past ulcer sites associated with severely elevated plantar pressures were not necessarily more accurate in discriminating against those past ulcer sites associated with lower pressures. Overall, the trend for those clinicians was that more sites were selected in total; therefore, while there was higher true positive identification of severely elevated sites, there was a corresponding increase in false positive error rate.
The level of agreement between clinicians on the location of severely elevated plantar pressures in the participants evaluated was assessed. For the 28 severely elevated plantar pressure sites: nine sites were identified by all three clinicians as elevated; six were identified by two clinicians; and five by one clinician. Eight of the 28 sites measured as severely elevated using the F-Scan were not selected by any of the clinicians during the clinical evaluations conducted.
In addition, for those nine sites which were identified by all three clinicians, the magnitude by which they were judged to be elevated differed. Two of the nine sites were rated as severely elevated by all clinicians, two others by two clinicians and the remaining five sites by only one clinician. These results suggest that there is significant disparity in the clinical assessment of elevated plantar pressures.
This study shows a clear discrepancy between the results obtained during computerised pressure evaluation for very high plantar pressures and those reported clinically. There may be several possible explanations for this result. Firstly, it may be that the clinical evaluation of elevated plantar pressures is only moderately valid. Results of this study suggest that severely elevated plantar pressures may be going undetected and low pressures are potentially being judged as very high, in some cases over 50% of the time. While sites of elevated pressure have traditionally been associated with the location of callus or foot deformity, and therefore have anecdotally been thought to be easily detectable clinically, it is possible in light of the findings of this study that they may present more subtly and may be more difficult to observe then previously thought.
However, other issues require consideration. While computerised pressure measurement is the ‘gold standard’, every measure is associated with a margin of error. It may be that there was some error in the computerised measurements conducted, which could reflect poorly on the analysis of the clinical data. However, the pressure readings utilised were an average of a number of steps over three trials so a component of the potential random error would be cancelled out. It is deemed unlikely, therefore, that system error would explain the majority of the differences noted.
An alternative explanation is the <350KPa cut-off point. It was necessary to identify such a point and it has some degree of internal validity. The use of a single pressure value and above however may be clinically somewhat arbitrary given the range of pressures at which neuropathic ulceration can occur. It was deemed that this was a relatively high cut-off point to utilise, given it fell within the top 10% of all pressure values and based on its strong association with past ulceration. As clinicians were instructed to evaluate for sites which were ‘markedly elevated’, which is a strong risk factor for ulceration, it can be argued that scores above the <350KPa range meet this criteria. These methodological issues, however, hold importance in the accurate interpretation of the study data and must be considered in light of the results reported.
It was found that clinicians who had a greater ability to correctly detect highly elevated pressure areas, overall selected a higher number of sites and so created a higher number of false positive results. There was no obvious explanation for this difference. Clinicians One and Three were more accurate in detecting severely elevated pressures over past ulcer sites, but had a higher rate of error in discriminating between those associated with lower pressure areas. This was in contrast to Clinician Two. The mixed agreement between clinicians on which sites were severely elevated was not a surprising result given the two measurement parameters are inter-related. Again this may be a true result but it must be considered that there may be an alternative explanation, such as an artefact from a difference in levels of understanding the study methods.
The range in pressure values of previous neuropathic ulcers reported re-enforces the multifactorial aetiology of foot ulceration and the role of mixed variables on individual outcomes.
This research has relevance to clinical practice as the data suggest that there may be significant error associated with the clinical assessment of high plantar pressure in people who have diabetes. Clearly this finding is of concern as it may mean that this risk factor is unknowingly going undetected and therefore perhaps is not being modified according to current best practice standards. Ultimately, if the causative pathways to foot ulceration are not appropriately detected and modified, reducing the incidence of foot ulceration in this population is likely to be challenging. Also, if there is a poor ability to differentiate between high and low plantar pressures clinically, it is difficult to identify who is in greater need of pressure-relieving therapy. This may create a dilemma in ensuring treatment is introduced where needed but resources are not directed where not required.
While future research is required to confirm these results and explore the issues further, it may be that greater access to computerised measurement is indicated in the future – particularly if clinical approaches cannot be improved upon. Pham et al (2000) reported that measuring foot pressures offered good specificity in detecting patients at risk of foot ulceration, but due to low sensitivity this technique was not suitable as an approach to screening. A primary indication for the use of computerised measurement may, therefore, be after an initial assessment where other risk factors, such as neuropathy, are found to be present.
The accurate assessment of risk factors is pivotal in detecting and preventing long-term foot complications. In order to construct a sound, evidence-based approach, it is necessary to establish the degree to which our assessments accurately inform us about the parameter under evaluation. Given the significant clinical implications of this finding, further research is recommended to verify results as a subsequent review of clinical assessment guidelines may be indicated.
This research was supported by a grant from the Wound Management Association of Victoria (WMAV) and the Caulfield Community Health Service, Podiatry Department. Janet Lloyd, Elizabeth Dalton, Shan Lawrence and the team of Podiatrists at the Caulfield Community Health Service, Podiatry Department offered invaluable assistance in the preparation and co-ordination of this research, allowing the study to take place. Thank you to Craig Payne for statistical and content advice offered throughout the project.