Socioeconomic status and fertility treatment outcomes in high-income countries: a review of the current literature

Abstract The association between socioeconomic status and fertility is a subject that has received much attention. Yet, little is known as to whether the socioeconomic status has an impact on the outcomes of fertility treatment. This systematic review aims to assess any possible relationship between socioeconomic deprivation and treatment outcomes. A database search was conducted of all publications in this field up to March 2021. Eleven studies were identified and six of these specifically investigated the impact of socioeconomic status on fertility treatment outcomes. Children conceived following assisted conception are more likely to be born to mothers of a higher socioeconomic status than those conceived naturally. Of the few studies investigating the impact of socioeconomic status on fertility treatment outcomes and the results are conflicting, making it difficult to draw robust conclusions as to its effect. It is unknown which, if any, marker of socioeconomic status is the most significant for fertility patients: whether it is the characteristics of the individual or that of their surroundings. Further research is urgently needed.


Introduction
There have been great advances in fertility treatment that can offer many the hope of achieving a pregnancy, but there remain significant social, economic, and geographical barriers as to who can access and benefit from it in many countries. Socioeconomic status (SES) can be considered as the position of a person in society based on a combination of economic (income, education, occupation) and sociological factors (social status and class) in relation to others (American Psychological Association, n.d.; Baker, 2014). SES has been shown to have a powerful impact on general health: those with higher SES generally have better health than those with lower status (Moor et al., 2017). This relationship has been well established in conditions such as mental health disorders (Businelle et al., 2014), cardiovascular disease (Rosengren et al., 2019) and stroke (Marshall et al., 2015). Although there is evidence that SES can affect fecundity (Colleran et al., 2015) and influence fertility treatment-seeking behaviour (Swift & Liu, 2014), there is a scarcity of data on socioeconomic factors in fertility treatment and assisted reproductive technology (ART) outcomes. In this review, we explore what is currently known of the relationship between SES and ART success in terms of live birth rates.
There are many metrics used to ascertain a person's SES. Compositional measures relating to the individual include education (e.g. years of formal education, International Standard of Education, (ISCED)), occupation (e.g. the International Standard Classification of Occupations, (ISCO)), income (e.g. above and belowaverage income/poverty line) and assets (e.g. car ownership, number of bedrooms). Contextual measures relate to an individual's environment, including geographical factors (e.g. urban/rural environments, the English Index of Multiple Deprivation) and composite measures combine several SES factors. There are several reviews whereby methods of measuring SES in developed countries are thoroughly explored (Galobardes et al., 2007;Pr€ ag et al., 2016). These factors can then be used to separate individuals into groups from the most to the least socially and economically advantaged. Importantly, many SES measures are not comparable between countries or cultures as there are international differences in the way different populations are educated and receive their income. Lower SES groups may not be able to access treatments in the same way (scarcity of service provision, fewer resources for purchasing medications, or information regarding service availability). Additionally, those with lower SES may not have the education to understand what constitutes good health and the steps required to maintain it.
The population treated by fertility services is dependent on those who seek treatment rather than all couples who are having difficulties conceiving. This is determined by service and treatment availability as well as the likelihood that those who require fertility advice and treatment actually obtain it. As such, outcomes of fertility treatments are affected by the existing SES of those couples that seek treatment in the first place. The main disparities influencing access to fertility treatment are socioeconomic status, geographical region, ethnicity, religion, income, and sexual orientation. Therefore, those patients who undergo fertility treatment are not necessarily reflective of the wider population who are experiencing infertility (Ter€ av€ a et al., 2008). In Britain, it is estimated that only 57% of women and 53% of men with infertility seek medical care (Datta et al., 2016). There are substantial variations in the availability of ART among healthcare systems in different countries. This is dependent on the breadth of coverage of publicly funded healthcare for fertility treatments which itself is influenced by the individual country's cultural, religious, and moral standpoint, as well as economic pressures (Pr€ ag & Mills, 2017). These influences need to be considered when interpreting the results of fertility treatment outcomes concerning SES.

Materials and methods
A systematic review (PROSPERO ID 231756) was conducted to answer the question 'does socioeconomic status influence the outcomes of patients undergoing assisted conception treatment for infertility?' Outcomes were defined as pregnancy and live birth rates after undertaking fertility treatment as this is the current leading metric for fertility treatment success (Human Fertilisation and Embryology Authority, 2020). The decision to keep the review focussed on highincome countries was to try and seek evidence of any discrepancy in healthcare systems that had more established fertility treatment provision rather than resource-poor countries where fertility treatment may be largely unavailable or inaccessible to most of the population (Allahbadia, 2013).
A literature search was conducted across PubMed, EMBASE, Medline and Cochrane from inception to March 2021. Search terms of assisted conception, IVF, assisted reproductive technology, insemination, ovulation induction, socioeconomic, deprivation, education, occupation, income, geographical, pregnancy, live birth, IVF outcome and treatment outcome were used (See Supplementary Online Material S1). Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed. After removing duplicates, publications were independently screened for inclusion criteria by two authors (RI and SG). Those included were all peer-reviewed studies that linked socioeconomic factors to fertility treatment outcomes in high-income/developed countries (World Bank, 2021

Results
A total of 11 studies were suitable to be included in this review. Six of these studies were found to investigate the relationship between SES and fertility outcomes with a further five investigating the SES of the population who have already conceived using fertility treatment (Table 1).

SES and fertility treatment pregnancies
Five population studies were identified that analyzed the socioeconomic status of parents who had children as the result of fertility treatment. Both Murray (2014) and Goisis et al. (2020) used household income as a marker for SES, Goisis et al. (2020), K€ all en et al. (2005) and Tierney and Cai (2019) used educational attainment and Raisanen et al. (2013) used maternal occupation as their primary measure of SES. Murray (2014) analyzed interview data from the Irish 'Infant Cohort of Growing Up' which included those who already had a child 10 months of age: 456 infants born following any fertility treatment (ovulation induction in 31.5% and IVF in 28.1%) between 2007 and 2009 were identified. Household income alone was used as the primary measure of SES. It was found that a disproportionate 62% of pregnancies conceived using fertility treatment were in patients in the top two income quintiles. Although it is well established that household income increases with age (largely due to employment progression), logistic regression showed older mothers in these higherincome quintiles did not explain the effect. Women in the lowest income quintile had a lower rate of fertility treatment pregnancy compared to the highest quintile (OR 0.32, CI 0.22-0.48). Goisis et al. (2020) used national databases in Norway andused the household income to determine the socioeconomic characteristics of parents who conceived through ART (32,580 live births), registered as those conceived through IVF, ICSI or unknown ART procedure between 1985 and 2014. Intrauterine insemination (IUI) cycles were not included. Like Murray (2014), they also found that parents who had a live birth after ART were more likely to be from a higher income bracket. Income was determined by the parents' taxable income in the 2 years before the birth of each child and stratified into quartiles. In the period 2010-2014, ART births accounted for 6.2% of all births in the highest income quartile compared to 1.4% in the lowest. Parents in the top income quartile were 4.2% (95%CI 4.1-4.3) more likely to have conceived through ART than those in the bottom quintile but this finding was attenuated by 35% when adjustment was made for maternal age. In addition, both mothers and fathers of children born through ART were more likely to have higher levels of education (tertiary education or masters/Ph.D.) compared to births not conceived through ART. Between 2010 and 2014 less than 2% of births were conceived by women who used ART and had only a primary school education compared to over 4.5% who had a master's or Ph.D.
Similar to Goisis et al. (2020), K€ all en et al. (2005) found a clear significant association between the number of IVF births and maternal education levels (OR 1.46 for IVF and 1.92 for ICSI in mothers with graduate studies compared to gymnasium/upper secondary school education). Their study looked at 12,160 women giving birth after IVF/ICSI in Sweden between 1982 and 2001. Interestingly, after adjusting for maternal age, the effect of higher education decreased the probability of having an IVF birth, but when maternal age at first identification of subfertility was adjusted for, the trend re-appeared.
Tierney and Cai (2019) conducted a large population study of ART births in the United States from 2010 to 2017. They also used women's education level (less than high school/high school/some college/4year degree/more than a 4-year degree) as a marker of SES. Consistent with the studies by Goisis et al. (2020) and K€ all en et al. (2005), women with more than a 4-year degree were found to have an ART birth incidence rate of 2.08 times that of women with a 4-year degree after multivariate analysis. Those who have less than a 4-year degree were significantly less likely to have an ART birth than those with a 4-year degree. In terms of ART usage, they noted that women with less than a high school degree and with a high school degree had a lower incidence of ART births than women with higher educational levels across all age groups. There was no statistical difference in the risk of ART births between women with a 4-year degree and women with some college education between the ages of 20 and 29. However, for women aged 30-34, the predicted incidence rate for women with some college education was 0.54 ART births per 1,000 compared with 0.87 per 1,000 in those with a 4-year degree. The difference was even more pronounced in women aged 35-39 with predicted incidence rates of 0.57 per 1,000 and 1.10 per 1,000 respectively. A large Finnish population-based cohort study of 5,647 singleton births in 2006-2010 achieved by IVF (representing 1.9% of singleton pregnancies in Finland) was conducted by Raisanen et al. (2013) and used maternal occupation rather than income or education as the sole indicator of SES. They found that 85% of all women giving birth after IVF belonged to the two highest SES groups. They also found that women who gave birth after IVF were older (33.4 vs. 29.5 years) compared to those who conceived spontaneously.
These studies, however, do not reveal whether women of higher SES are more likely to conceive and have a live birth with fertility treatment compared to those in the lower SES groups, or whether they are over-represented in the population that use fertility treatment to conceive.

SES and ART outcomes
There were six studies that looked specifically at the relationship between socioeconomic factors and ART outcomes. Smith et al. (2011), Mahalingaiah et al. (2011), Huang et al. (2012 and Hansen et al. (2016) all analyzed the relationship between educational levels and fertility treatment outcomes in the United States, with Smith et al. (2011) and Hansen et al. (2016) also focussing on household income as a socioeconomic variable. Barzilai-Pesach et al. (2006) looked at the occupation and Richardson et al. (2020) used the Index of Multiple Deprivation (IMD), a contextual measure that uses postcodes to determine relative deprivation scores. Smith et al. (2011) recruited 391 Californian women from eight reproductive endocrinology clinics and followed them for 18 months. They included women who had no treatment (21.5%), ovulation induction (4.4%), IUI (21.2%) and IVF (52.9%). After controlling for confounding factors, they noted significant trends for educational levels and the use of fertility treatment and the chance of pregnancy. College-educated women were more likely to use IUI and IVF compared to no treatment (RR2.0 (95%CI1.0-4.0) and RR2.3 (95%CI 1.2-5.3) respectively), were more likely to have >3 cycles and had double the chance of pregnancy compared to those who were not college-educated (OR 2.06, 95%CI 1.22-3.47, p ¼ 0.01).
Three further studies conducted in the United States conversely found no association between educational levels and fertility treatment outcomes. Mahalingaiah et al. (2011) looked at the association between female educational levels and first IVF/ICSI cycle outcomes on a prospective cohort of 2,569 women in the United States from 1994 to 2003. After adjustment for age and clinic, they observed lower odds of cycle cancellation (most frequently due to poor ovarian response) in those with higher educational attainment (OR 0.60 (95%CI 0.42-0.84) and 0.52 (95%CI 0.36-0.76) p ¼ 0.001 for those with college degrees and graduate school attendance compared to women with no college degree). Of those women who reached embryo transfer, there was no statistically significant difference in pregnancy outcomes based on educational levels.
Two studies analyzed education levels and treatment outcomes by undertaking secondary analysis of data from randomized controlled ovarian stimulation (COS) and intrauterine (IUI) insemination trials in the United States. Huang et al. (2012) looked at 664 couples who were randomized to four treatment groups (intracervical insemination (ICI)/COS with ICI/IUI/COS with IUI) and had completed a lifestyle questionnaire. After bivariate analysis, they found no statistically significant association between female or male education (stratified as high school/college/post-college) and pregnancy and live birth rates. Hansen et al. (2016) analyzed data from participants in the AMIGOS clinical trial (randomized clinical trial looking at outcomes after COS-IUI with clomiphene citrate, letrozole or gonadotrophins). As with Huang et al. (2012), they found no statistically significant association of educational status with any pregnancy outcome after IUI treatment. However, an income of >$50,000 was more common in those who did have a live birth compared to those with an income of <$50,000 (26.4% vs. 16.4%, p ¼ 0.33). Much of this difference was attributed to a higher rate of pregnancy loss in those with incomes <$50,000. Using bivariate analysis, it was shown that an income of >$50,000 was associated with greater odds of conception and live birth compared with those earning <$50,000 (OR 1.53 (95%CI 1.04-2.25) p ¼ 0.031 and 1.83 (95%CI 1.15-2.91) p ¼ 0.11 respectively).
The study by Smith et al. (2011) also looked at income as a measure of SES. They noted that women with household incomes of >$150,000 per year were more likely to use higher levels of fertility treatments such as IUI (RR 2.4, CI 1.0-5.9) and IVF (RR 5.2, CI 2.4-11.3) compared to those with a household income <$100,000. Those in the highest income households were more likely to have more 1-5þ treatment cycles compared to 0 cycles. They observed a threshold point of fertility treatment expenditure above which there were significant increases in pregnancy rates (constant increase in pregnancy rates after an expenditure of $2,500 up to $45,000 expenditure on fertility treatment after which a decline was noted), even though there was no direct correlation between treatment expenditure and outcome. Only a slight nonstatistically significant increase in pregnancy rates was seen in those with higher household incomes. Barzilai-Pesach et al. (2006) found no significant differences in occupational status between those who did not conceive, those who conceived and those who delivered after IVF treatment. They undertook a prospective observational cohort study in Soroka, Israel, of 75 working women who attended fertility and IVF clinics and underwent a cycle of treatment between 1999 and 2000. No significant differences in treatment outcomes with the type of occupation were seen. There was a tendency for higher pregnancy rates towards those who had more freedom to determine their daily schedule (RR 1.43, CI 0.96-2.11), although this was not statistically significant. Those in full-time employment were significantly less likely to have a live birth after conceiving than those who worked part-time (RR 0.3, CI 0.11-0.96). A non-significant trend towards a higher pregnancy rate in those who worked shifts was seen (RR 1.52, CI 0.89-2.61).
A study by Richardson et al. (2020) conducted at Leeds Fertility (England) was a retrospective cohort study of 3,091 patients undergoing their first NHSfunded single fresh embryo transfer in a 6-year period between 2012 and 2017. Their main findings showed that the clinical pregnancy rate increased significantly from 31.0% in the most deprived women to 38.8% in those who were least deprived (p < 0.01) and the live birth rate increased from 26.8 to 35.4% (p < 0.01). After adjusting for known variables that impact the success rate of IVF success, they found that women who were least deprived had an adjusted relative risk of live birth of 1.18 (95%CI 1.00-1.39). As this confidence interval includes 1.00 it can be argued that the influence of deprivation level on live births may be negligible. Those women in the least deprived group were significantly older than those in the most deprived group (33.1 ± 3.8 years compared to 30.4 ± 4.3 years) despite IVF success rates decreasing with increasing female age (van Loendersloot et al., 2010). The women who were least deprived had a lower BMI compared to those who were most deprived (23.9 ± 3.1 compared to 25.2 ± 3.0, p ¼ 0.02).

Discussion
An extensive literature review on whether socioeconomic factors influence the outcome of assisted conception treatments revealed surprisingly few studies that focussed on this topic. The main findings of the five population studies demonstrated that births resulting from fertility treatment had mothers who were more likely to be in the highest SES groups regardless of whether education, income or occupation was used as the marker for SES. The six studies that focussed on whether SES influenced treatment outcomes had varying results. Smith et al. (2011) andMahalingaiah et al. (2011) both demonstrated that educational levels contributed to a difference in fertility treatment/IVF outcomes, whereas Huang et al. (2012) and Hansen et al. (2016) both found no correlation between educational levels and IUI outcomes. Although Huang et al. (2012) did show an association between income and fertility treatment outcomes, this was not corroborated by Smith et al. (2011) who found no statistically significant difference in fertility treatment outcomes in their study. Barzilai-Pesach et al. (2006) found no differences in occupational characteristics in those that did versus those that did not conceive and Richardson et al. (2020) showed that women in the least deprived areas had a statistically higher chance of achieving a pregnancy and live birth.
There are significant limitations to the scope of this review, most notably the limited number of published studies in developed countries that have looked at SES in the context of fertility treatment outcomes. Those reviewed include a range of methodologies (e.g. prospective and retrospective) and modes of assessing SES, whether on an individual SES level or an environmental SES level, therefore cannot be directly compared in a meta-analysis. The economics and cultures of different countries mean that certain SES factors will have more influence in one country over another, especially in those where it is less likely that a woman will complete her education or work once married. However, these differences are likely to be less striking when comparing high-income countries to each other. It is difficult to ascertain whether each study is truly reflective of the impact of SES on fertility treatment results, especially where only small numbers are studied. The decision to keep the review focussed on high-income countries was to try and seek evidence of any discrepancy in healthcare systems that had more established fertility treatment provision rather than developing, resource-poor countries where fertility treatment is still largely unavailable or inaccessible to most of the population (Allahbadia, 2013).
It is unclear to what extent the population served by fertility units is representative of the population suffering from infertility, especially as SES influences access to, and utilization of, fertility treatments. All studies included in this review were conducted in a variety of different countries with different contextual socioeconomic backgrounds, healthcare systems and ethnicities. Within some countries, national funding of fertility treatment varies considerably. The postcode lottery of fertility treatment in the UK means that in many boroughs access is limited by stringent criteria, and the number of IVF cycles funded maybe none or up to three. In the USA there are only 16 states with a mandate for insurers to offer funding for fertility treatment (including California, but IVF is not covered as standard) with patients accessing treatment in the other states having to finance their care privately ('Infertility, Inequality, and How Lack of Insurance Coverage Compromises Reproductive Autonomy', Insogna & Ginsburg, 2018). Ireland remains one of only two European Union countries not to offer national funding for assisted conception, but by contrast, both Finland and Israel have very generous state funding for fertility treatment. For those who reside in countries without nationally-funded fertility services, individuals from lower SES groups may not be able to seek help if they have difficulties conceiving due to an inability to afford self-funded treatment. By its nature, studies on SES can only be observational and the root cause of any differences described will likely be multifactoral. Patients with higher levels of education are more likely to earn more and be able to spend that income on higher levels of housing and healthcare. A person's SES is dynamic and may also change over time, making it difficult to determine whether early childhood SES and deprivation have any greater impact than a person's current socioeconomic status.
The population studies in this review analyzed the parental characteristics of children already born through fertility treatment. Therefore, they cannot be well placed to understand whether SES in its own right is an influencing factor on fertility outcomes, or whether it is that those in the higher SES tiers are overly represented in undertaking assisted conception treatments. However, these studies benefit from large patient numbers and highlight that children born through IVF were more likely to belong to parents in the top SES groups based on mother's education (Goisis et al., 2020;K€ all en et al., 2005;Tierney & Cai, 2019), income (Murray, 2014;Goisis et al., 2020) and occupation (Raisanen et al., 2013). They are limited in that they only look at treatment cycles that have been successful and resulted in a live birth. Thus, the increased numbers of children born through ART to higher SES mothers may be a reflection that women from these groups access and undertake treatment at a more frequent and higher level, rather than the treatment being more successful in this population. In addition, the study by Murray (2014) included infants born through any fertility treatment including ovulation induction. Access to this kind of treatment is usually much easier (it may in some cases be prescribed by family doctors) and cheaper compared to the more expensive and specialized treatment of IVF. This makes interpretation of the results of this study more difficult to analyze.
In the studies that focussed on treatment outcomes in different SES groups, there were mixed results. The smaller prospective Israeli study by Barzilai-Pesach et al. (2006) looked only at 75 patients and was largely focussed on the effect of female working conditions and occupational strain on the outcomes of fertility treatments rather than socio-economic status specifically. Their study cannot be relied upon to be representative of outcomes for shift-workers on a national or international scale. Conversely, current literature tends to suggest a negative correlation between shift work and menstrual health (Mills & Kuohung, 2019;Stock et al., 2019).
Both Smith et al. (2011) and Richardson et al. (2020) showed a statistically significant increase in pregnancy rates in women with a higher SES although they used very different markers of SES. The study by Smith et al. (2011) recruited their cohort of patients across 8 different clinics and therefore may have reached a wider geographical area and range of patients from differing SES groups. Female education levels seemed to be the biggest influencing factor; this may be because higher educated women are better equipped to communicate and engage with their healthcare providers, or that they have more health literacy and ability to understand the steps they can take individually to improve their chances of success (e.g. optimal weight and diet). Richardson et al. (2020) demonstrated that socioeconomic deprivation had a statistically significant impact on pregnancy and live birth outcomes after IVF. Their study had a large cohort and, as it focussed on cohabiting couples, was reflective of both male and female socioeconomic deprivation. This study found a small but statistically significant increase in the odds of achieving a pregnancy or live birth after IVF treatment in those couples who were in the least deprived groups. This is consistent with the expected finding that those who are in lower SES groups have a reduced rate of treatment success, a conclusion that has been seen across many health conditions (Marshall et al., 2015;Oates & Schechter, 2016). It is well established that higher socioeconomic status infers a degree of health protection above and beyond access to medical care through differences in lifestyle and exposure to environmental stressors (Braveman & Gottlieb, 2014). Individual SES variables such as higher levels of income, education and occupation influence lifestyle behaviours that promote health (e.g. better nutrition and housing) (Chetty et al., 2016). Those who are lower in the SES hierarchy are more likely to live in (or be exposed to) environments that are health-damaging (e.g. traffic pollution). In addition, poorer neighbourhoods are prone to overcrowding which has been shown to negatively influence health (Evans & Repper, 2000).
There are many theories as to how differences in socioeconomic status have an impact on an individual (the 'biology of social adversity') and their response to disease and medical treatments (Boyce et al., 2012). The term 'allostatic load' refers to the cumulative wear and tear throughout life on the body after exposure to chronic and recurrent stress (McEwen & Stellar, 1993). Exposure to psychological chronic stressors, especially in early childhood, can cause epigenetic changes (Lam et al., 2012) that impact lifelong biological responses. Short-term protection from acute stressors provided by hormones such as adrenal steroids and glucocorticoids can, in the long-term, negatively impact health. These factors combined are thought to account for socially mediated discrepancies in behaviour, immune responses (Slopen et al., 2015), inflammatory cascades and cardiovascular and metabolic health (McCrory et al., 2015). In particular, it is thought that environmental factors can influence the epigenome of spermatozoa which may be a factor in any socioeconomic differences in fertility treatment outcomes (Donkin & Barr es, 2018). It is currently not known whether any difference in outcomes between SES groups is because of gamete quality and the subsequent embryos that are created through ART, or the physiological environment to which they are transferred (e.g. pelvic infections are higher in lower SES groups (Leichliter et al., 2013) which may influence implantation).
However, SES can never just be a stand-alone factor in predicting fertility treatment outcomes. SES is a concept derived from socially constructed ideologies that are inextricably tied to race and ethnicity. Structural racism can lead to stark differences in SES and subsequent treatment outcomes and higher SES does not protect Black individuals from the health effects of this (Churchwell et al., 2020). There is evidence, particularly in the US that women from minority groups have lower success rates due to a combination of limited access to care, cost, language, and cultural barriers (Humphries et al., 2016).
The studies included in this review use a variety of factors to determine SES and use different ways in which to measure it. Although there is significant overlap in the methods used (e.g. levels of educational achievement, income brackets), there is no standardized approach that can allow for direct comparison. There are recommendations by the US National Committee on Vital and Health Statistics which are helpful in best practice for obtaining and using socioeconomic factors in research (Centers for Disease Control and Prevention, 2012) and adherence to these methods would prove beneficial when comparing studies investigating the link between SES and health outcomes.
This review highlights the scarcity of data on this subject and exposes the lack of understanding we have in this field. It is difficult to conclude whether SES acts as a true independent factor in ART outcomes or whether it is simply that those in higher SES groups undertake more treatment cycles and there is a bias in the populations that can be observed. At present, it is unknown which, if any, marker of SES is the most significant for fertility patients: whether it is the characteristics of their clinical, social and structural surroundings, or of the individual, whose SES is so heavily influenced by their environment. Infertility is a unique condition where both men and women are treated together so it will be interesting to determine whether male or female influences on treatment outcome are more susceptible to variations in SES. We still do not understand the impact of socioeconomic variation between countries or within a country, and to what extent. If a true link between socioeconomic deprivation and fertility treatment outcomes can be established, then it may have an impact on national health policies and practices. In England, local commissioning groups determine access to, and the number of nationally-funded fertility treatment cycles a couple is eligible. Socioeconomic differences may allow commissioners to interpret success rates based on a clinic's local population and more accurately allocate funding. If we can determine which are the main SES factors that underlie a discrepancy in treatment outcomes then specific populations, as well as individuals, may be able to be targeted from a public health perspective to improve their overall chance of success.

Disclosure statement
No potential conflict of interest was reported by the author(s).