The 2016 HIGh Heels: Health effects And psychosexual BenefITS (HIGH HABITS) study: systematic review of reviews and additional primary studies

High-heeled shoes (high heels) are a form of footwear that raises the heels of the wearer’s foot substantially above the level of the toes, making the wearer taller. They have a long history in human civilisation [1]. An earlier form of high heels were worn by men in the medieval period, when their function was largely practical, that is to say to raise the wearer above the filth of medieval streets [2]. In modern society, high heels have become a fashionable symbol of modern (heteronormative) female sexuality [3], and are regularly worn by a considerable proportion of women [4]. Fashion operates by capitalizing on concepts of social compliance and conformity [5, 6], potentially mediated through celebrity influence [7, 8] and the expectation to perform normalised gender roles [9, 10]. Many feminist theorists have argued that gendered violence occurs when women become disabled by their clothing and footwear [11], when standards of beauty define dimensions of physical freedom [12, 13].

Although many women like wearing high heels for various reasons, this cultural context raises the possibility of some women wearing high heels when they have not chosen freely to do so. One online opinion poll estimates that the proportion of women whose high heel wear is mainly due to social expectation rather than their own free choice may be around a third [14]. This would represent a substantial minority, and is around the same as the proportion this poll reports have ever been required to wear high heels at work.

Moreover, political features and approaches have been shown to be an important determinant of population-level health outcomes [15]. One way in which such influences may operate is through the extent to which a government decides to intervene in societal matters. One major policy controversy in some localities recently has been whether employers are legally allowed to, or indeed should be allowed to, stipulate that female staff wear high heels to work. In the United Kingdom, the Government has published a response [16] stating that it will work to develop guidelines and raise awareness of this issue, as others have also done, to ensure women are not discriminated against in the workplace. Nevertheless, there has been some media confusion about the details of this response and some commentators have criticised the decision to rely on the Equality Act (2010) rather than propose new legislation. In comparison, the Canadian province of British Columbia has chosen to make an amendment to existing legislation to specifically prohibit employers from requiring female staff to wear high heels [17]. Current policy discussions have largely not considered issues relating to entertainment venue or licensed social event admission policies, although the policy of the Cannes Film Festival has been very controversial [18].

The popular view that wearing high heels increases women’s attractiveness to men is supported by scientific evidence [19, 20]. Although this is not a health outcome, it is socially relevant and may be of interest to the public health community since it is the tension between psychosexual benefits and negative musculoskeletal (MSK) health effects that makes high heels a topic of substantial public health interest and also of interest with regard to the study of the social determinants of health outcomes. A large number of studies have been conducted to examine the relationship between high heels and MSK health. There are two dominant approaches as discussed before [21] each with their associated strengths and limitations: the experimental biomechanical approach and the epidemiological approach. For example, the former approach provides clearer insight into causality, while the latter does not rely on proxy markers of clinical outcomes.

Thorough review articles have been published from both perspectives. The most recent biomechanical review [22] found evidence of qualitatively consistent alterations to the neuromechanics of walking gait and the kinetics and kinematics of bodily structures from the toes to the spine in ways that could be seen as biomechanical markers of MSK conditions such as hallux valgus (HV) and osteoarthritis (OA). This was a narrative non-systematic review published in 2014 with the most recent included study being published in 2012. Meanwhile, the most recent epidemiological review found evidence of an association between high heel wear and MSK pain, HV and first-party injury. Nevertheless, associations with OA and second-party injury were inconclusive. This was a narrative systematic review published in 2016 with a final search date of July 2015.

The evidence base has remained fragmented as no prior review article has integrated both the epidemiological and biomechanical perspectives. Moreover, no prior review has considered the psychosexual benefits, which provide important social context to this important area of public health debate. Conducting an integrated review of these perspectives has been identified [21] as a research priority. Indeed, we were aware of a number of prior reviews on different aspects of this topic that we could synthesize. Therefore, we present the first systematic review of reviews on high heels. Our review is the first to consider both the epidemiological and biomechanical perspectives on the negative MSK health effects of high heels. In addition, our review is the first to present evidence on the psychosexual benefits associated with high heel wear, thereby providing essential context in understanding the public health challenge of high heels. To further strengthen our evidence synthesis, we added additional primary studies on areas in which no review had been conducted or a marked lack of prior research had been identified.

The 2016 HIGh Heels: Health effects And psychosexual BenefITS (HIGH HABITS) study used a systematic review approach following PRISMA [23] as appropriate. In light of the existence of a number of prior reviews on aspects of this topic, albeit a stark lack of integration of different aspects, our systematic review initially sought to synthesise evidence from available reviews. Subsequently, additional primary evidence was added on areas identified to be lacking in research and/or prior reviews. The search strategy was designed by MSB (lead reviewer) and JB, who conducted proportionate quality control on review procedures. Any discrepancies were resolved by discussion.

MEDLINE (Ovid), EMBASE (Ovid), AMED (Ovid), PsycINFO (Ovid), Cochrane Central (Ovid), CINAHL (EBSCO) and Web of Science (Thomson Reuters, excluding patents) were searched up to November 2016. As per Barnish and Barnish [21], we applied the following keywords to all databases: “positive heel inclination”, “high heel”, “high-heeled”, ‘“high heeled”, “wedge heel”, “platform heel”, “platform shoe”, “stiletto” and “elevator shoe”. The only change we made was to add quotation marks to increase search specificity. Supplementary searches were conducted in Google Scholar (GS), Directory of Open Access Journals (DOAJ) and bibliographies of relevant articles.

We included review studies of any design (non-systematic review, systematic review or meta-analysis) that were:

We also considered individual primary studies that:

The following data were extracted for review articles:

The following data were extracted for primary studies:

Heel height measurements expressed in inches were converted to cm at 1 in. = 2.54 cm. Risk of bias assessment could not be conducted in a way that would offer useful comparability due to a wide range of different designs.

Database searches yielded 923 records (MEDLINE 223, EMBASE 251, AMED 54, PsycINFO 30, Cochrane Central 12, CINAHL 2 and Web of Science 351), which yielded 492 unique records following deduplication. Supplementary searches yielded an additional 14 records (3% of unique records), providing a grand total of 506 unique records. Ten reviews proceeded to full-text screening and seven met the inclusion criteria for our review – of these, two [21, 24] were systematic reviews without meta-analysis and five [22, 24‐28] were non-systematic reviews. There were no meta-analyses. All seven reviews were published in English, although one [28] was also available in Portuguese. Ninety-six unique publications were reviewed across these seven review articles.

Seventeen additional primary publications proceeded to full-text screening and 13 met the inclusion criteria for our review, all published in English. Eight publications [29‐36] were case reports of second-party injury reporting a total of 10 cases from three countries (UK, Denmark and Ireland). Five publications [19, 20, 37‐39] reported psychosexual studies from three countries (France, UK and Egypt). There were no further non-case report studies on second-party injury. All reports were considered separate studies since different analyses were reported and our approach did not involve meta-analysis. The article selection process is shown in Fig.1, study results in Table 1, study characteristics in Table 2 and studies excluded at full-text review with reasons in Table 3.

Two reviews provided epidemiological data. Riskowski et al. [26] assessing literature up to 2010, did not focus primarily on high heels and only included three studies on heel height, reaching inconclusive conclusions regarding pain and OA. Meanwhile, a systematic review by Barnish and Barnish [21] up to July 2015 demonstrates how the field has progressed to provide clear evidence that high heels are associated with MSK pain, HV and first-party injury, including to children. There remained no clear epidemiological evidence of a link with OA and second-party injury. Only one descriptive study [40] had presented second-party injury data, the toll was relatively low and these were excluded from further analysis in the paper by Williams and Haines.

The remaining reviews focused on biomechanical aspects. Generally, across reviews, there was good evidence of considerable kinetic and kinematic alterations associated with high heel wear that can be considered markers of increased first-party injury risk and MSK conditions such as OA, HV and MSK pain. Silva et al. [28] provide valuable insights into postural disorders related to high-heel wear among adolescent girls (aged 10–19). Increased heel height and decreased heel width were identified as key predictors of negative outcomes. Looking specifically at lumbar lordosis, considering literature up to 2010, Russell [27] found inconclusive results. Meanwhile, Cronin [22] found qualitatively consistent alterations in kinetics and kinematics from the spine to the toes, although some studies were not in agreement on lumbar lordosis. More detail on the findings of each review can be found in Table 1 Part A.

There were no prior reviews about the psychosexual benefits associated with high heel wear. Five studies addressed this aspect. All five studies presented evidence of such psychosexual benefits. One study from Egypt [39] assessed women’s own views as to whether wearing high heels increased their attractiveness, finding this effect among businesswomen, female employees and female university students alike. The remaining studies assessed judgements of attractiveness made by others or influence upon the behaviour of others. High heels were generally found to have a much greater effect than medium heels. A series of studies from France [20, 37, 38] found that attractiveness ratings, made by both male and female participants with regard to a photograph of a woman, were influenced by high-heel-associated postural changes, even though her shoes were not visible, whereas whether or not a man was willing to help a woman was only influenced by the confederate’s (a person who participates in a psychological experiment pretending to be a participant but is working for the researcher) heel height in male participants. Morris et al. [19] also noted that raters were more likely to misclassify a female walking in light-based silhouette form in a point-light task as male if she was wearing flat shoes as opposed to high heels. More details on the findings of the psychosexual studies can be found in Table 1 Part B.

The only eligible additional studies identified on second-party injury were case reports [29‐36]. Eight case reports were included, presenting 10 cases. Injuries requiring emergency department attention were sustained of both deliberate and accidental etiology. Accidental causes were being stepped on by a stiletto heel in the eye after falling on a dancefloor, two unspecified cases of being stepped on by stiletto heels, a young girl suffering foot injury requiring emergency department attention after being stepped on by a stiletto heel when someone else was getting off the bus, and a young woman being stepped on by another woman’s stiletto heel at a bar. More detail on each case can be found in Table 1 Part C.

Here we present the first systematic review of reviews about high heels. We address both the biomechanical and epidemiological perspectives and offer the first review of the psychosexual literature, which provides essential context in which to situate the findings on negative health effects, and understand the public health issues and dilemmas they pose for society. The biomechanical literature is now clear that wearing high heels causes substantial kinetic and kinematic alterations in the MSK system, ranging from the spine to the toes. These alterations increase the risk of MSK conditions such as OA, HV and MSK pain as well as first-party injury. The epidemiological literature also now clearly says that wearing high heels is associated with increased risk of HV, MSK pain and first-party injury, although the risk of first-party injury requiring emergency department attention is at most moderate. However, there is still no clear epidemiological evidence of an association between high heel wear and OA. Therefore, the suggestion from a large volume of biomechanical evidence dating back to the seminal paper in the Lancet in 1998 by Kerrigan et al. [41], that high heel wear may predispose women to OA has not been confirmed in population-based studies. Evidence from one observational study [40] and a series of case reports [29‐36] shows that high heels, and in particular stilettos, can cause second-party injuries requiring emergency department attention. However, current evidence does not allow a robust quantification of the magnitude of risk.

The included literature benefits from a large number of reviews, although only two were systematic reviews. Compared to non-systematic reviews, systematic reviews have a more reproducible methodology and are less prone to selection bias [42]. The strengths and limitations of the epidemiological and biomechanical approaches have been discussed before [21]. High heel wear is a very common exposure [4] by epidemiological standards, being more common than very common health conditions such as childhood asthma or other negative health-related behaviours such as smoking [43]. It is challenging to find a truly non-exposed group. Moreover, studies have tended to focus on relative rather than absolute risk measures. Moreover, two injury studies [40, 44] included in the Barnish and Barnish review [21] used emergency department databases which may be subject to misclassification and/or underreporting bias and only provide insight into the extreme end of the injury spectrum. Meanwhile, psychosexual findings are limited by a failure to consider platforms and heel shape as well as an upper heel height bound of 9 cm. No study assessed the role of the respective heights of men and women. This could be relevant since there is a preference in heterosexual partnerships for the male to be taller than the female [45, 46], and very high heels may eliminate this difference since young women in Australia, the United States and England [47‐49] are on average less than 15 cm shorter than young men. Indeed, research into this area has a clear heteronormative bias and does not consider alternative sexualities.

High heels are a challenging topic due to the tension between health and psychosexual considerations. The evidence synthesised in our review shows that wearing high heels increases women’s attractiveness to men and can reward female wearers with other psychosexual benefits in terms of male attention and their own view of their beauty. This poses a potential opportunity or a dilemma, which many women face regardless of their sexuality. We hope our review can provide a useful resource for clinicians such as podiatrists and MSK specialists in discussing in gender sensitive ways high-heel related conditions and injuries with which their clients present. It is important that women’s footwear choices are respected and that they are not pressurised into wearing high heels against their will. Unstated social expectation and pressure, mediated by social compliance [5] and celebrity influence [7, 8] is clearly a complex challenge that is difficult to confront. Continuing to raise awareness of the health issues associated with high heels appears the most useful solution to seeking to maximise freedom of choice. Important policy developments have occurred recently in different localities [16, 17], although challenges remain regarding compliance with legislation and also compulsion that occurs beyond the workplace, which may be covered by equality legislation that refers to provision of goods and services.

High heels are currently one of the few ‘adult’ items allowed to be sold to children in countries such as the UK, and as Silva et al. [28] show us, this can lead to substantial health issues. Adolescents are particularly prone to peer pressure [50] and there is evidence that health may play a limited role in female teenagers’ shoe choices in particular [51]. Indeed, increased adolescent high heel wear has been observed in recent years [52]. It has been suggested [53] that education on their health effects, for example as part of Personal and Social Education, could be useful. Moreover, it may be worthwhile considering whether it is appropriate that high heels can be sold to and for the wear of people who are still skeletally immature.

Further research is required into OA risk from an epidemiological perspective, second-party injury (perceptions and observations, related to high heels and potentially other glamorous women’s outfit types) and the competing influences on women’s freedom of choice in shoe selection, potentially paralleling Elliott-Green et al. [54]. It is also important, either from academic studies, or by monitoring the media, to assess as time progresses how issues of compulsion (explicit or implicit) are addressed. Moreover, academics should more frequently take the potential mediating effects of high heels into account when studying or reviewing the MSK health of female participants.

The key scientific strength of our work is that it brings together for the first time three hitherto disparate bodies of evidence – the epidemiological, biomechanical and psychosexual – on a topic of current social interest. Moreover, we believe our work is of methodological value and offers an evidence synthesis approach that we hope will be of interest to academics working in areas in which non-systematic reviews remain the predominant synthesis approach. However, there are certain potential limitations we should address. Our review was limited to seven European languages and seven major databases plus GS and DOAJ, so may not have achieved 100% coverage of global literature. Also, the speed of scientific research and communication remains too slow to offer an effective critique on an ongoing social issue – this work was inspired by an event that was reported in the media in May 2016. In addition, the non-inclusion of grey literature would be considered a limitation by some academics [55, 56]. Nevertheless, Barnish et al. [57] have previously argued against the inclusion of grey literature in certain circumstances, and we believe this is a topic in which it is most appropriate to focus solely on peer-reviewed publications, at least at this stage. Risk of bias assessment could not be conducted in a way that would offer useful comparability due to a wide range of different designs. Therefore, readers should consider their own biases when evaluating this work.