Baldassin V, Gomes CR, Beraldo PS. Effectiveness of prefabricated and customized foot orthoses made from low-cost foam for noncomplicated plantar fasciitis: a randomized controlled trial.
To evaluate the effectiveness of prefabricated and customized foot orthoses made from low-cost foam (ethylene vinyl acetate [EVA]) in plantar fasciitis.
Double-blinded randomized controlled trial.
Public rehabilitation referral medical center.
One hundred forty-two adults (75% women) with plantar fasciitis, without anatomical alterations in the feet. Seventeen subjects (12%) were lost during the follow-up.
Prefabricated and customized foot orthoses, both made from EVA, used for 8 weeks.
Main Outcome Measures
The primary outcome was pain (modified subscale of the Foot Function Index, [FFI] pain). The secondary outcomes were pain elicited by palpation in the medial calcaneal tuberosity and modified FFI total. Each participant was reviewed in the 4th and 8th weeks of follow-up.
One hundred twenty-five participants returned to at least 1 of the follow-up evaluations (63 in the prefabricated and 62 in the customized groups). There was a significant improvement in both groups (P<.05), but there was no difference of modified FFI pain between intragroup differences at 4 (4.03 points; 95% confidence interval [CI], −4.2 to 12.3) and 8 weeks (3.93 points; 95% CI, −4.6 to 12.5).
The low-cost prefabricated and customized foot orthoses, as used in this trial, had similar effectiveness in the treatment of noncomplicated plantar fasciitis after 8 weeks of use. Our results were similar to other trials, although those trials did not use orthoses made from EVA. Thus, EVA prefabricated inserts may be the best choice for the treatment of plantar fasciitis without complication.
There are many densities and thicknesses of rubber, features that are pertinent in this issue and have not been properly addressed. EVA is a low-cost foam with similar properties of rubber, which is rarely cited in the studies. EVA combines the mechanical properties required for orthoses and a low price.
Furthermore, foot orthoses for the treatment of plantar fasciitis can be prefabricated or custom-made. In our facility, the former is based on a mold of average feet, whereas the latter is made specifically for the patient's feet. The use of either type of orthosis can be suggested unless there is an important anatomic change in the feet. Other differences between the insoles are their flexibility, thickness, and cost, which can vary according to the construction process. A prefabricated foot orthosis costs much less than a customized orthosis.
The foot orthoses used in these trials were made in different ways and from different materials, although none of them were made from EVA. Furthermore, 2 of the 5 trials were underpowered, only 1 assessed function (disability), 3 had losses to follow-up of at least 15%, and only 2 explicitly analyzed by intention to treat. Considering the mechanical properties of each material and the methodologic flaws of these studies, their results cannot be generalized.
Thus, we conducted a randomized trial to compare prefabricated and customized foot orthoses made from EVA to treat noncomplicated plantar fasciitis. We hypothesized that both types of orthoses made from EVA would present similar results in terms of pain relief.
Participant Flow and Follow-Up
Patients with plantar fasciitis consecutively admitted from May 2004 to July 2005 in the Orthotics and Prosthetics Department at the SARAH Network of Rehabilitation Hospitals, Brasilia, Brazil, were recruited. The inclusion criteria consisted of adults, age 18 or older, with the clinical diagnosis of plantar fasciitis characterized by a complaint of first-step pain and maximal tenderness over the medial calcaneal tuberosity.
Patients with complicated plantar fasciitis, characterized by anatomic or alignment alterations in the feet, were excluded. In general, the other exclusion criteria were patients who were pregnant, and patients with a history of malignancy or inflammatory arthritis, and patients who already had received foot orthoses. Further exclusion criteria and other reasons to not participate in the study are listed in figure 1.
One hundred forty-two patients out of 248 evaluated were randomly allocated to receive a prefabricated or customized foot orthoses (both 95% EVA) within 2 days. One hundred seventeen subjects attended the first review after using the orthoses for 4 weeks, and 105 subjects attended the second review after 8 weeks of treatment. The progression of participants through the trial is shown in figure 1. The research protocol was approved by the ethics committee of the hospital, and all subjects gave their written informed consent to participate in the trial.
After the eligibility was determined, all subjects were evaluated by one of the authors (VB) who obtained baseline characteristics data of the participants. If both heels were painful, only the data from the most symptomatic foot were analyzed. If the level of pain was similar in both feet, only the right foot was considered. Otherwise, the same orthosis of the allocated group was manufactured to the contralateral foot. The customized foot orthosis was made from a nonplaster casting technique. This implies that the molds from the patients' feet were taken in the proper foam (EVA), and it was not necessary to make a negative cast. The orthoses' manufacturing process followed in this trial is described in Appendix 1.
All subjects were advised to test the orthosis on their footwear. If a participant complained of discomfort, he/she was advised to use the orthosis for a minimum of 1 week. If the orthosis still bothered the participant, he/she was advised to go back to the hospital for an evaluation. For ethical reasons, no access to other interventions was vetoed. However, during the follow-up, subjects were asked to report the use of cointerventions.
The primary outcome was pain measured by a subscale of FFI pain. The original FFI pain scale provides a valid assessment of foot discomfort with a moderate test-retest reliability (intraclass correlation coefficients of 0.695, 95% CI, 0.53–0.81) and a high degree of internal consistency (Cronbach α, 0.946).
The FFI total is composed of visual analog scales that consider 3 dimensions: activity limitation, disability, and pain. The pain subscale measures the level of foot pain in various situations. The subscale originally contained 9 items, but we included 1 more question evaluating pain upon getting up in the morning. The internal consistency of the modified FFI pain improved to 0.958 after this approach. This questionnaire was blindly applied by 1 of 4 trained physiotherapists at baseline and at the 4- and 8-week reviews.
The secondary outcomes were pain elicited by palpation in the projection of the medial calcaneal tuberosity and modified FFI total. The first of these was evaluated by one of the authors (VB) using the visual analog scale. All scales were graded from 0 (no pain/trouble) to 100 points (worst imaginable pain/trouble).
After the initial appointment, subjects were randomly assigned to receive prefabricated or customized orthoses. The randomization was performed in blocks of 8 by using opaque sealed envelopes to ensure that the contents were concealed. The randomization procedure was conducted by one of the authors (VB) who was not blinded to the allocation group thereafter.
Each outcome was analyzed according to the intention-to-treat principle. In 1 situation, a participant randomized to the prefabricated group accidentally received a customized foot orthosis, but the data from that participant were analyzed in the primary allocated group.
To maintain blinding, the customized procedure was used in both groups. It means that the subjects allocated to the prefabricated group went through the same process as those in the customized group; however, they received a prefabricated orthosis. In our institution, the time spent with patients is usually greater in the preparation of the customized foot orthoses, and this factor could interfere in the results. All insoles were made to look as similar as possible, given the materials used, and subjects were blinded to which one they had received.
To guarantee blinding, each participant was instructed not to show his/her orthosis to the physiotherapists who applied the modified FFI questionnaire at the 2 reviews. Furthermore, the physiotherapists did not have access to the previous evaluations. To evaluate the effectiveness of this blinding after 8 weeks of treatment, subjects were asked to guess which orthosis they were using. The same question was asked of the examiners who applied the modified FFI.
the difference of 13 to 15 points (or mm) is not significant from a patient's point of view, regardless of the pain level. In other words, the sample size calculation considered a detectable difference that was too small to be clinically meaningful. Thus, a minimum of 48 subjects per group would be necessary in this study to achieve a difference in modified FFI pain of 13 points, with an SD of 20, an α of 0.05 (2 tailed), and a β of 0.10. We randomized 142 participants into the 2 groups, estimating a loss to follow-up of at least 15%. All subjects who attended at least 1 of the reviews were included in the analysis.
Data are expressed as proportions and means (SD) of outcomes and absolute (Δ) differences of intragroup measurements (reviews – baseline). A chi-square test was used to compare the opinions of the subjects and the investigators regarding the type of orthosis used; it was also used to analyze differences between the remaining/lost subjects and the distribution of the cointervention between both groups. The internal consistency of the modified FFI pain was tested with Cronbach α.
To evaluate whether there was any effect of the type of orthosis used over the follow-up period, despite a difference at baseline, a linear regression approach (mixed model) was applied.
The dependent variable was the modified FFI pain, and the independent factors were the treatment groups and the time of follow-up (4th and 8th weeks). The baseline measure for the modified FFI pain was considered as the only covariate. However, in a post hoc analysis, we also included other demographic variables that were different between groups at baseline. The statistical analyses were conducted by using SPSS.
Version 13; SPSS Inc, 233 S Wacker Dr, 11th floor, Chicago, IL 60606.
Statistical significance for these analyses was established at P<.05.
Subjects in both groups had similar descriptive characteristics, with a slight difference in body mass index and duration of symptoms (table 1). In general, the participants were women, sedentary people, low-heeled shoe wearers, and cement walkers. Seventeen subjects (12% of randomized subjects: 9 in the prefabricated and 8 in the customized groups) did not attend any of the subsequent follow-ups. In terms of anthropometric and clinical characteristics, there was no difference between the remaining and lost subjects.
Table 1Baseline Characteristics
Type of Orthosis
Sex, n (%)
Body mass index, kg/m2
Mean duration of symptoms, months
Hours per day on feet, n (%)
2 to 4
4 to 6
More used footwear, n (%)
Type of floor, n (%)
Physical activity, n (%)
NOTE. Values are means ± SD unless stated otherwise.
In general, during the follow-up, both groups presented similar improvement considering the components of modified FFI (fig 2) and pain to palpation. In regards to modified FFI pain, there was a significant improvement in both groups (P<.05, F=5.18) without a difference between them (table 2). On both occasions, the prefabricated group exhibited higher scores of modified FFI pain, but the differences, adjusted to baseline values, between groups at the 4th and 8th weeks were negligible (4.0 points [95% CI, −4.2 to 12.3; P=.337) and 3.9 points (95% CI, −4.6 to 12.5; P=.363), respectively (see table 2). The minimal clinically important difference used for this evaluation was 13 points. During all follow-up (4th and 8th weeks), the upper limits of the CI excluded this minimal clinically important difference.
Table 2Primary and Secondary Outcomes From Subjects Who Returned for at Least One of the Reviews
Because some subjects missed at least 1 of the 2 follow-ups, a sensitivity analysis was conducted for the modified FFI pain. Thus, when no data were available, the previous evaluated score was considered. The result was the same as that obtained with the previous analysis. Even when considering body mass index and the duration of symptoms in addition to baseline modified FFI pain as covariates, the results were similar. There was no statistically significant difference of time, group, and interaction between them; all were covariate adjusted (baseline values) to modified FFI total and pain to palpation (see table 2).
There was no statistical difference between groups in their belief of the type of orthotic provided; 81% from the prefabricated and 84% from the customized group believed they were using a customized foot orthosis (χ2=0.01, P=.92). However, physiotherapists were more evenly divided in their belief of orthotic type; 58% correctly identified the prefabricated group and 48% correctly identified the customized group (χ2=0.48, P=.49).
Cointerventions were used 67% of the time; 40% of them performed stretching exercises for the Achilles' tendon, and 28% used other types of cointerventions, such as ice and anti-inflammatory treatment. No difference was found between the groups concerning these cointerventions. The absence of difference in modified FFI pain between the groups remained when analyzed among participants with or without cointervention. No side effects, such as blisters, were registered among the groups.
This randomized controlled trial compared the use of prefabricated and customized foot orthoses made from EVA by patients with noncomplicated plantar fasciitis for 8 weeks. In terms of the outcomes considered, the study did not show any statistically significant difference in pain regardless of the type of foot orthoses used after 4 or 8 weeks.
Study Strengths and Limitations
With respect to these results, 2 items need to be noted. The first of these involves the validity of the results, and the second concerns the consistency with other studies. The primary outcome was evaluated by independent examiners who did not have access to the previous assessment. Furthermore, there was no difference in the percentage of guesses about the allocated group, showing that the blinding of the participants and researchers to the treatment group was effective. The anthropometric characteristics of the groups were nearly similar, indicating that the randomization was efficient. Moreover, the custom molds were made for all subjects, regardless of the group to which they were allocated. All of these factors point to a high internal validity.
Regarding the external validity, the subjects showed characteristics similar to typical patients with plantar fasciitis.
It should be stressed that pain was the primary outcome considered in this and other studies because it is a clinically relevant characteristic in plantar fasciitis. In addition, the material used to make the orthoses (EVA) is an easily accessible and inexpensive alternative to rubber. Another factor that also favored the external validity was the “intention-to-treat” principle and the fact that no kind of cointervention was restrained. The types of cointerventions used were the same as reported in the literature.
Although there was a 12% loss to follow-up over the course of the trial, this loss was equally balanced between the groups. One potential criticism of our results could be based on the cointervention, allowing for the introduction of confounders. However, this does not seem to be the case because there was no cointervention proportion difference between the groups. Furthermore, the absence of outcome differences remains, even when only the subjects without cointerventions were considered. The study showed that the groups improved similarly. Whether this improvement was associated exclusively to the 2 orthoses is a causal inference that is difficult to answer in our experimental design with random assignment. Besides the types of insoles, many other factors can be contributed to the improvement, mainly the cointervention, Hawthorne effects, reactivity of measurement, and maturation effects.
Aside from the cointerventions that we analyzed, all others factors, if present, probably were also balanced in both groups. In fact, the main message is that both groups improved and not differed. Finally, the study was originally designed as an equivalence trial. But it is an active control trial, where the new intervention is the prefabricated and the standard intervention the customized orthosis, with the hypothesis of noninferiority (1-sided test formulation).
Furthermore, we considered an effect size of 0.65 (13/20), which would be considered moderately large to detect differences between the groups. From this point of view, it can be considered underpowered. However, it is important to stress that the small differences noted between orthoses until 8 weeks of follow-up never approached the level of reduction believed necessary to make a clinical difference.
Over the last 10 years, 5 randomized controlled trials were published involving at least 2 groups of patients with the diagnosis of plantar fasciitis who used prefabricated and customized foot orthoses.
investigated the effect of foot orthoses in 85 subjects. After 3 months, the treatment with functional foot orthoses made from an undescribed material proved to be more effective than either anti-inflammatory treatment or an accommodative viscoelastic heel cup. In a multicenter trial, Pfeffer et al
compared 5 treatments in 236 patients. After an 8-week follow-up, they concluded that, when combined with a stretching program, the prefabricated shoe inserts were equivalent to the custom orthotic devices. Fifty-five patients completed the study by Turlik et al,
who evaluated the effectiveness of ¼-inch thick urethane heel pads and functional foot orthoses in relieving symptoms of heel spur syndrome. After 3 months, they found that functional foot orthoses were associated with better outcomes. However, the researchers did not show the validity or reliability of the outcome measures that they used. Furthermore, they did not describe some of the methodologic issues that could have led to bias. Martin et al
assigned 193 patients to 1 of 3 groups: custom-made orthoses, over-the-counter arch supports, or tension night splints. After 3 months, no significant difference among the 3 groups was observed with respect to final outcomes based on first-step pain or pain felt during the day. However, there were some methodologic failures in this study, including the high dropout rate (24%) and the absence of information about the adoption of the intention-to-treat principle. Finally, Landorf et al
showed that prefabricated and customized foot orthoses had similar effectiveness in the treatment of plantar fasciitis, with 3- and 12-month reviews. One hundred thirty-one subjects returned for a review after 12 months of treatment, and the authors concluded that foot orthoses did not have long-term beneficial effects compared with a sham device, although small 3-month benefits in function and pain were observed.
Thus, an intervention would be necessary, helpful to relieve the pain, and improve the individuals' quality of life. In our institution, the typical costs associated with supplying and fitting EVA prefabricated orthoses, were roughly half of those of the EVA customized orthoses as used in this trial. In a cost minimization analysis, the consequences of 2 or more interventions being compared are equivalent. The analysis, therefore, focuses on costs, and promotes the choice of the less expensive option. Since the prefabricated inserts appeared to be at least as effective in reducing the heel pain as a custom orthotic device, we conclude that the prefabricated inserts should be selected in cases of noncomplicated plantar fasciitis. This is especially true when these orthoses are made from EVA. With increasing demands on limited resources, worldwide health systems must pay attention to cost saving medical devices.
aVersion 13; SPSS Inc, 233 S Wacker Dr, 11th floor, Chicago, IL 60606.
Appendix 1: Orthoses Manufacturing Process
Three plates of EVA were used for making the customized foot orthoses. These plates were cut in a rectangular shape. The width, length, and thickness of each plate were 13 × 33 × 0.5 cm, 13 × 20 × 0.5 cm, and 13 × 20 × 1.0 cm, respectively. The first plate of EVA was heated to 100°C for 1 minute in an electric oven. This hot EVA plate was placed on the top of 2 plates of foam, 1 low-density plate with a thickness of 4 cm, and another greater density plate with a thickness of 5 cm. During the procedure, the participant remained seated with 1 foot placed on this surface for approximately 5 minutes. The expert made sure that the temperature of the foam was tolerable for the participant. During this modeling process, the expert held the subtalar joint in a neutral position. The next step consisted of joining the 2 remaining plates of EVA under the single foam plate, already molded and cooled, in water to prevent any undesirable material deformity. Afterward, the expert cut and shaped the insole according to the participant's footwear.
The prefabricated foot orthosis was made from a positive foot model based on the size of the participant's footwear. The positive foot model was taken from a healthy foot without anatomic deformities. Two plates of EVA were heated to about 100°C in an electric oven for 1 minute. The width, length, and thickness of each plate were 13 × 20 × 0.5 cm and 13 × 20 × 1.0 cm, respectively. They were placed and pressed by the positive foot model.
An additional plate made from low-density foam of a commercial brand was placed in the posterior region of both insoles to promote greater comfort for the participant. This additional plate was used for both customized and prefabricated orthoses.
Trial registration: Australian New Zealand Clinical Trials Registry, ACTRN012607000319471 (www.anzctr.org.au).
No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated.