Foot pain is common, yet few studies have examined the condition in relationship to shoewear. In this cross-sectional study of men and women from the population-based Framingham Study, the association between foot pain and type of shoewear was examined.
Data were collected on 3,378 members of the Framingham Study who completed the foot examination in 2002–2008. Foot pain (both generalized and at specific locations) was measured by the response to the question “On most days, do you have pain, aching or stiffness in either foot?” Shoewear was recorded for the present time and 5 past age categories, by the subject's choice of the appropriate shoe from a list. The responses were categorized into 3 groups (good, average, or poor shoes). Sex-specific multivariate logistic regression models were used to examine the effect of shoewear (average shoes were the referent group) on generalized and location-specific foot pain, adjusting for age and weight.
In women, compared with average shoes, those who wore good shoes in the past were 67% less likely to report hindfoot pain (P = 0.02), after adjusting for age and weight. In men, there was no association between foot pain, at any location, and shoewear, possibly due to the fact that <2% wore bad shoe types, making it difficult to see any relationship.
Even after taking age and weight into account, past shoewear use in women remained associated with hindfoot pain. Future studies should address specific support and structural features of shoewear.
Many podiatric clinicians note that foot problems are common in older persons (1, 2), yet very little epidemiologic information exists on foot pain, especially related to shoewear in older persons. National data reveal that foot and toe symptoms are among the top 20 reasons for physician office visits among patients ages 65–74 years (3). Prior research indicates that not enough attention has been given to foot pain and other foot disorders since historically these factors have not been regarded as important health risks (1). Women are more likely to have foot pain than men (4, 5) but it is unknown whether this is due to a higher prevalence of foot deformities, underlying disease, shoewear, or other lifestyle choices.
Foot pain and foot disorders are serious burdens for many older individuals, especially those with rheumatic diseases (6–8). While foot pain is considered to be a very common musculoskeletal symptom in the US adult population, relatively little is known of the prevalence or cause of foot pain in older Americans in a population-based sample. Examining the association between shoewear and foot pain may lead to a greater understanding of this relationship, which is important when considering strategies to prevent sequelae due to foot pain in older adults.
Previous studies have examined the influence of footwear on foot problems and other diseases in older adults, but have focused on small numbers of older adults or patients with specific diseases. A 1993 study by Frey et al (9) presented descriptive information on shoe trends and their effect on the development of foot deformities and pain in 356 women, ages 20–60 years. They observed that the majority of women wore shoes that were too small for their feet and had foot pain; however, no statistical analysis was presented. In a 2005 study by Menz and Morris, the relationship between footwear characteristics (length, width, and area) and the prevalence of common forefoot problems were examined in 176 older adults residing in a retirement community (10). Similar to the results of the study by Frey et al, it was found that incorrectly fitting footwear was common and associated with forefoot disease and foot pain. Menz et al have also examined the relationship between footwear characteristics and the risk of indoor and outdoor falls in the same cohort of older adults and found that there was no association between the type of shoe and the risk of falling either inside or outside the home (11). Harrison et al reported an assessment of the fit of footwear in 100 patients with diabetes mellitus in 2007 and found that many diabetic patients wear shoes that are too narrow for their feet (12). Garrow et al conducted a population-based foot pain and disability survey in 3,417 adults that examined age and sex differences in a variety of foot symptoms, but did not contain any information on shoewear (13).
In addition, the preventative effects of footwear, like plantar pressure relief, have been shown to be important in previous studies. Lobman et al studied the effect of preventative footwear on foot pressure in 81 older patients with diabetes mellitus (14). This type of footwear was shown to be successful in reducing plantar pressure. Also, Burnfield et al found that in 20 healthy, older adults, faster walking speed and walking barefoot result in higher foot pressures (15).
To evaluate the relationship between foot pain and shoewear in a large, population-based cross-sectional study, we collected information on the major type of shoe worn both currently and in the past, along with a self-report of foot pain by men and women participating in the Framingham Study. In addition to the previously-mentioned information, the Framingham Study has extensive data collected on many clinical and other factors, notably medical conditions, comorbidities, and other potential confounders. The purpose of this study was to evaluate the relationship between type of shoe worn and foot pain, accounting for other possible risk factors, in men and women of the Framingham cohort.
SUBJECTS AND METHODS
The study sample, the Framingham Foot Study cohort, was derived from 2 large, population-based samples of residents of Framingham, Massachusetts. The majority of the cohort was comprised of members from the Framingham Study Original Cohort and the Framingham Offspring Cohort. The Framingham Study Original Cohort was formed in 1948 from a two-thirds sample of the town of Framingham, MA in order to study risk factors for heart disease (16). This cohort has been followed biennially since that time. The Framingham Offspring cohort, formed in 1972, consists of adult offspring who had a parent in the Original Cohort, and the spouses of the offspring (17). This group has been followed every 4 years since cohort inception to study familial risk factors for heart disease. Members of the Framingham cohorts were examined for the current study either at their scheduled Framingham clinic examination or at a call-back examination.
The second population-based group used for the study was a new population sample that was derived from census-based, random-digit dialing within the Framingham community by the Center for Survey Research at the University of Massachusetts, selecting subjects who were ≥50 years old and ambulatory. This group was added to the Framingham Foot Study recruitment to increase participation by minority persons and other community members of the Framingham catchment area (using a targeted random-digit dialing schema in selected Framingham census tracts). Persons contacted via the random-digit dialing methodology who were interested in being part of a multiphasic physical examination (foot, osteoarthritis, bone health, and general health), received a written letter of invitation to join the study and a followup phone call to schedule a study appointment.
The collection of study data and information from these Framingham Foot Study participants followed strict, well-established protocols that are available upon request from the Framingham Study. All Framingham Foot Study participants have given informed consent for the data collection and this study has undergone institutional review by both the Hebrew SeniorLife and the Boston University Medical Center Institutional Review Boards. The data are extensive, systematically collected, of high quality, and gathered by trained personnel.
From these combined population-based cohorts, the Framingham Foot Study conducted a physical examination of the foot, and collected participant history, performance measures, and other data via questionnaire.
Foot assessment clinical tool.
A validated foot examination was used with specific criteria to assess foot pain, foot symptoms, and the presence of foot disorders. Trained clinical examiners performed all foot examinations. All participants were ambulatory and cognitively intact (as indicated by Mini-Mental score screening to identify qualified study subjects who would be able to give symptom information about their feet).
Foot pain was assessed in 1,477 men and 1,901 women in the population-based Framingham Foot Study between 2002–2008. Generalized foot pain was measured using the following National Health and Nutrition Examination Survey-based query about foot pain: “On most days, do you have pain, aching, or stiffness in either of your feet?” Possible responses were no; yes, left foot only; yes, right foot only; yes, both feet; yes, not sure what side; and unknown. For this analysis, responses were collapsed into 2 groups: yes, pain in one or both feet; or no, no pain in either foot. All subjects responded with either yes or no. Foot pain at specific locations was also assessed. Participants were given a picture of the top and bottom of the foot and were asked to point out any areas with pain, aching, or stiffness on most days (Figure 1). The identifiable areas were the nails, forefoot, hindfoot, heel, arch of the foot, and ball of the foot. As with generalized foot pain, the responses to the location-specific foot pain were classified as yes or no pain.
Participants were asked about their 1 type of shoe that was worn the most regularly, currently and during 5 age periods in their past. Given the time constraints of the multiphasic Framingham examinations, we collected self-report questionnaire data on general shoe type. To assess current and past shoewear, subjects were handed a list of 11 shoewear types and asked, “From this list of shoes, could you tell me which one type of shoe you currently wear most frequently?” The question was then asked for each age category (20–29 years, 30–44 years, 45–64 years, 65–74 years and 75 years and older) and recorded by the examiner. If the subject had not yet reached a given age, no response was recorded for that age group. Current and past shoewear were then classified into 3 groups (good, average, and poor shoes). The poor group consisted of high-risk shoewear that lack support and sound structure, including high-heeled shoes, sandals, and slippers. The average, referent group included mid-risk shoes such as hard- or rubber-soled shoes, special shoes, and work boots. The good group consisted of low-risk shoes, including athletic and casual sneakers, because these shoes have the characteristics that theoretically make them safer, (namely rigid heel counters, fixation, or firm nonflexible soles) (18, 19). Past shoewear, across the 5 age groups, were further summarized into a single worst past shoe type, as the most high-risk shoe worn at any age. In an analysis of footwear style and falls in older adults, Koepsell et al created similar shoewear groups as those that were chosen here (20). Although our use of this simple, self-report did not use a validated instrument for shoewear (and we are unaware if one exists) we believe this self-report of shoewear provides insight into an important measure of exposure.
We examined potential confounders in our analyses including: age, sex, weight, height, and smoking status. Age in years at the time of examination was recorded. Weight in pounds was measured using a standardized balance beam scale and recorded to the nearest half pound. Height (without shoes) was measured in inches using a calibrated stadiometer and recorded to the nearest one-quarter inch. A participant's smoking status was assessed via questionnaire as current cigarette smoker (smoked regularly in the past year), former smoker, or never smoked. Comorbidities were queried with each study participant at the clinic visit.
Descriptive statistics were generated overall and separately for men and women. Multivariate logistic models were performed to examine the effect of the potential risk factors on foot pain, with current and past shoewear as our main focus. Each foot pain outcome was modeled separately. Generalized foot pain, and all location-specific foot pain variables were modeled as functions of current shoewear, past shoewear, and other risk factors.
Because sex is a strong confounder for both foot pain and for types of shoewear, all analyses were performed as sex-specific. All analyses were conducted using the SAS statistical analysis package, version 9.1 (SAS Institute, Cary, NC).
Of the 3,378 participants in the Framingham Foot Study, 1,472 men and 1,900 women had complete foot examination data and were included in this analysis. The mean age for subjects was 66 years (range 36–100 years) and the sample was 56% female. The mean height and weight were 65.5 inches and 174.0 pounds, respectively. Table 1 shows the characteristics of the study sample by sex. All subjects had valid measurements for age, weight, height, sex, and reported their typical shoewear type. Smoking status response data were missing for 65 subjects.
Table 1. Descriptive characteristics of men and women in the Framingham Foot Study who completed the foot examination (2002–2008)
Men (n = 1,472)
Women (n = 1,900)
Age, mean ± SD years
65.7 ± 10.3
65.9 ± 11.0
Weight, mean ± SD pounds
194.1 ± 34.7
158.5 ± 35.9
Height, mean ± SD inches
68.7 ± 2.8
63.1 ± 2.7
Current smoker, no. (%)
Generalized foot pain, no. (%)
Toe or nail pain, no. (%)
Forefoot pain, no. (%)
Hindfoot pain, no. (%)
Heel pain, no. (%)
Arch pain, no. (%)
Ball of foot pain, no. (%)
Of the 3,372 subjects under study, 25% of participants reported the presence of generalized foot pain on most days. Figure 1 shows the 6 possible anatomic locations of foot pain for the total sample. Table 1 also shows the sex-specific distribution of the report of foot pain, both generalized and at specific locations. Generalized foot pain was reported by 19% of the men and 29% of the women.
The distribution of current shoewear and past shoewear is shown in Table 2 for the original 11 categories. We collapsed these categories into 3 major groupings, as shown in Table 2 and Figure 2. Rubber-soled shoes were the most commonly reported shoe, and were worn by 28% of the men and 32% of the women.
Table 2. Distribution of current shoewear patterns of the men and women in the Framingham Foot Study (2002–2008)
Men, no. (%)
Women, no. (%)
Hard-soled leather shoe
Other/not applicable/ unknown
Heels or pumps
Tables 3 and 4 show the covariate-adjusted results of foot pain modeled on current and past shoewear. All models were adjusted for age and weight. Exploratory models showed that height and smoking had no relationship with foot pain, so those risk factors were not included in the analysis. Table 3 shows the results for sex-specific analysis of generalized foot pain and current shoewear. There was no association between generalized, toe, forefoot, ball of the foot, heel, or arch of the foot pain and current shoewear in women or in men. We also examined the relationship between generalized foot pain and past shoewear and similarly found no associations for this relationship or for most of the specific locations of foot pain. The exception was a statistically significant association in women who reported pain at the hindfoot and past shoewear. Table 4 shows the associations in women between past shoewear and hindfoot pain, adjusting for age and weight. Compared with women who wore average shoes, women whose worst past shoe was in the good category were 67% less likely to report hindfoot pain, after adjusting for age and weight (adjusted P = 0.022; crude P = 0.026). In men, no significant relationship was found between foot pain at any location and shoewear groups, whether current or past shoewear.
Table 3. Odds ratios (ORs) and 95% confidence intervals (95% CIs) for the association between current shoewear and generalized foot pain in the men and women of the Framingham Foot Study (2002–2008)
In our study of community dwelling older adults, past shoewear among women was a predictive, statistically significant factor for hindfoot pain; however, no significant associations were seen in men. In women, we found an increased risk between foot pain at the hindfoot location and shoewear. This was the only site-specific association in women, and none were seen among the men in our sample, or among generalized foot pain. In our study, weight was significantly associated with foot pain for both men and women. Age was not significantly associated with foot pain in our sample, but we included it as a covariate in order to compare with other studies and to address possible residual confounding. Even after taking into account age and weight, past shoewear in women remained associated with hindfoot pain.
We found that 25% of the participants (19% of men, 29% of women) reported the presence of generalized foot pain on most days, which is similar to the results of other studies examining foot pain. Similar to our study, Garrow et al found that 20% of men and 24% of women reported foot pain (13). Menz et al also found that more women report foot pain than men (10, 11).
In men, <2% of our population reported wearing poor shoe types. This could make seeing any possible relationship between foot pain and shoewear very difficult due to sparse statistical power to detect a possible difference. We also observed several protective, nonsignificant associations in men between shoewear and foot pain. Despite the large number of men in our study, very few reported wearing good or poor shoes. It may well be that other studies with larger numbers of men reporting good or poor shoe types (57% and 92% of men in our study reported average shoes for current shoewear and past shoewear, respectively) could have sufficient power to examine these possible protective effects in more detail.
No statistically significant associations were found in a study by Manna et al (21) that evaluated the relationship between foot troubles and type of footwear (shoe, sandal, or slipper) in 300 men and women. Although this study was limited by their definition of foot troubles, their null results between foot troubles and shoe, sandal, or slipper use are in agreement with our results. A study by Dawson et al (22) examined the association of the age at which heels were first worn for different heel heights and maximum heel height in relation to foot pain and other foot problems in women. Although no statistically significant associations were observed between the age at which heels were first worn and foot pain, the authors found statistically significant associations between maximum heel height worn for going out socially and for dancing with foot pain (P < 0.05). These associations of high-heel use and foot pain were not seen in our study, which found no association between shoewear and generalized foot pain. It is quite interesting to note, however, that poor shoewear (including high heels) was associated with hindfoot pain.
It is interesting to note that we observed an association with foot pain at specific locations in the foot but not with generalized foot pain. The definition of good shoes (athletic shoe or casual sneaker) used in this study implies a shoe design with better fit, foot posture, and shock absorption characteristics. Each heel strike during walking may produce a biomechanic shock of 3–7 g (1 g = 1 times the acceleration due to gravity). Good shoes often have softer out-soles, mid-soles, or insoles that may use elements of gel, foamed polyurthethane, or air chambers that serve to smooth (low pass filter) the raw shock wave. Attenuation of shock could be responsible for the reduction in perceived pain at the hindfoot. It is important to note that the hindfoot receives the largest shock wave within the foot at each and every heel strike. Therefore, it makes sense that good shoes will protect for pain within the hindfoot. It is also possible that the single association seen at the hindfoot is due to the tightness of the heelcords that might result from sustained use of high heels. If this is the case, it is possible that interventions with stretching exercises could mitigate the influence of the poor shoes.
Given that no correction for multiple comparisons was made, the results should be conservatively interpreted; specifically, note that there is only 1 significant association and thus other studies are needed to confirm or extend these results.
Our study has several limitations that should be addressed. Since our population was predominately white adults, there is limited ability to generalize to other racial populations. The cross sectional study design limited our ability to infer causality. Therefore, we cannot note whether the choice of shoewear caused the foot pain or if the foot pain caused the participant to select a particular type of shoe.
In addition, participants were given a list of 11, nonspecific categories from which to choose the 1 shoe type that was worn most often currently and at past age groupings. Since these categories were broad and inadequately specified, there is most probably misclassification that has occurred in the categorization of footwear. Due to the time constraints of the study, we were forced to limit the choices of categories and were not able to measure this as accurately or specifically as would have been ideal. This probable misclassification would lead to results that are biased towards the null. Sandals, in particular, are likely to be misclassified given the fact that some sandals may actually provide excellent support for the foot. Due to the time constraints in this study, only 1 category for sandals was used. Despite this limitation, it is important to realize that regular sandal use in the Northeastern US is not particularly common (4% in this study), and therefore may not be a major factor. There was also opportunity for recall bias when participants were asked to remember what type of shoe they were wearing as long ago as 60 years.
Also, part of what makes this analysis more complicated may be the precise category in which we placed certain shoe types. For example, a work boot that was categorized as average may be steel toed for construction and very uncomfortable and might actually be better classified as poor. Even an athletic shoe that was classified as good may have pronatory control elements, and if it was worn by a subject with a pes cavus foot (who therefore over-supinates) the shoe would be a poor match for that person. Furthermore, previous work on specific qualities of shoewear has suggested that foot structure and shoe structure interact with foot function (23–26). This implies that the shoe may actually dominate this relationship so the issue is not simply if one is wearing a good shoe but if they are wearing a good shoe for their foot type. We were not able to account for this possibility in our analysis, as we did not have a reliable measure of the subject's foot type.
Despite the previously mentioned limitations, our study also has several strengths. To our knowledge, this is one of the first studies to examine the association between shoewear, beyond just high-heel use, and foot pain. Our study sample included both men and women, which enabled us to generalize the results to both sexes. Foot pain was measured across multiple sites of the foot. We obtained information on nail pain, toe pain, forefoot pain, hindfoot pain, heel pain, pain in the arch, and pain in the ball of the foot, in addition to the measure of generalized foot pain. This allowed us to conduct more specified analyses of a particular pain location. The foot examination was conducted by a trained examiner and was not simply self-report. This information is likely to be more reliable and should result in less misclassification or recall bias than a simple self-report of foot conditions.
Further research is needed to address the specific support features of shoewear such as arch support, toe box width, and toe box depth. Also, future studies should examine the relationship between the severity of foot pain and whether the subject has a pes planus, rectus, or pes cavus foot type. Furthermore, examining the relationship between foot type, shoe structure, the presence or absence of foot pathologies, and associated pain would be of keen interest to the rheumatology and podiatric community.
In conclusion, our study found that in women, past shoewear was a statistically significant, predictive factor for hindfoot pain, but no such association was found in men. In men, less than 2% wore poor shoe types, making it difficult to see any relationship. Given the small percentage of men wearing poor shoes, it appears that shoe type may not be a major factor for developing foot pain in men. Past shoewear in women is associated with hindfoot pain, regardless of age or weight. Thus, young women should make careful choices regarding their shoe type in order to potentially avoid hindfoot pain later in life, or perform stretching exercises to alleviate the effect of high heels on hindfoot pain.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Ms Dufour had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Dufour, Hillstrom, Walker, Kivell, Hannan.
Acquisition of data. Dufour, Broe, Nguyen, Gagnon, Walker, Hannan.
Analysis and interpretation of data. Dufour, Broe, Nguyen, Gagnon, Hillstrom, Walker, Kivell, Hannan.
The authors thank the Framingham Foot Study research team and study participants for the contribution of their time, effort, and dedication, and also the anonymous reviewers and the editors for their helpful comments on our work.